m ' - . .. . I "> - ; '-' ,-/.:., , ;; ' " .' | ;. ; " p -.-:'''- -' . ' ,.., :^ i- - - ''^ WORKS OF PROF. J. A. MOVER PUBLISHED BY JOHN WILEY & SONS Descriptive Geometry for Students of Engineering Third Edition. 8vo, viii + 204 pages, 128 figures. Cloth, $2.00. The Steam Turbine A Practical and Theoretical Treatise for Engineers and Designers, including a Discussion of the Gas Turbine. Second Edition, Revised and Enlarged. 8vo, ri + 430 pages, 225 figures. Gloth, $4.00. STEAM TURBINES A PRACTICAL AND THEORETICAL TREATISE FOR ENGINEERS AND DESIGNERS INCLUDING A DISCUSSION OF THE GAS TURBINE BY < JAMES AMBROSE MOVER, S.B., A.M. MEMBER OF THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS, MITGLIED DBS VEREINES DEUTSCHER INGENIEURE, MEMBRE TITULAIRE ASSOCIATION INTER- NATIONALE DU FROID, MEMBER OF THE FRANKLIN INSTITUTE, AMERICAN INSTITUTE OF ELECTRICAL ENGINEERS, ETC. PROFESSOR IN CHARGE OF DEPARTMENT OF MECHANICAL ENGINEERING, THE, PENNSYLVANIA STATE COLLEGE, FORMERLY ENGINEER, WESTING- HOUSE, CHURCH, KERR & CO., AND ENGINEER STEAM TURBINE DEPARTMENT, GENERAL ELECTRIC COMPANY SECOND EDITION, REVISED AND ENLARGED FIRST THOUSAND NEW YORK JOHN WILEY & SONS, INC. LONDON: CHAPMAN & HALL, LIMITED 1914 COPYRIGHT, 1908, 1914 "V JAMES AMBROSE MOVER Stanbopc ipre&s F. H. G1LSON COMPANY BOSTON. U.S.A. Ml Library PREFACE TO FIRST EDITION. THE object of this book is to give in a small volume what I believe, as the result of years of practical experience, engineers and students of engineering want to know about steam turbines. It is intended that it shall be a manual for the practical engineer who is designing, operating, or manufacturing steam turbines rather than a compilation of manufacturers' catalogs combined with a digest of standard books on thermodynamics and mechanics. In a general way the author has tried to explain briefly and directly some of the more important problems about which the qualified steam engineer must have some knowledge. When this book was first planned it was intended primarily for the use of the author's assistants in the experimental and testing departments of one of the large manufacturing companies, but later it seemed that it might be useful in a larger field. The order in developing the subject is the reverse of that adopted by most authors. Instead of discussing the intricacies of blading in the beginning of the book, the more simple problems of nozzle design are presented first. A great deal more is now known about nozzles than there was even very few years ago, and many of the conditions affecting the efficiency of nozzles may now be considered well established. Nozzles are also becoming a more important part of all types of turbines. Even the Parsons turbine is now being modified in America and England so that in many of the latest designs for large sizes, nozzles are used in the high-pressure stages. It is coming to be generally recognized that in the future there will probably be no large installations of reciprocating engines for electric services. A few years ago this might have been considered a bold statement, but it is a fact which is now iii 291520 iv PREFACE generally, although reluctantly, admitted by manufacturers of reciprocating engines. The entropy-total heat chart in the back of the book is laid out with lines of constant superheat instead of lines of constant tem- perature which have been generally used for charts of this kind. For practical engineering work it is very desirable to have lines of constant superheat on such charts, because in America and England guarantees of steam consumption are usually given in degrees of superheat rather than of temperature. When charts made with constant-temperature lines are used, it is always neces- sary to calculate the temperature before the chart can be used. Most of the graduates of our American technical schools are entirely "at sea" with the simplest heat calculations, and one of the reasons for this deficiency is that most of the books on steam engines and especially those on the steam turbine are more devoted to giving a large quantity of facts than to fulfilling a useful purpose. Practical engineers who have had to deal with large numbers of men with an engineering training agree most candidly with Dr. Steinmetz when he says in substance that it seems to cause no concern in some of our large technical schools that the graduates are sent out loaded with a mass of half-understood and undigested subjects, while they are deficient both in the understanding of the fundamental principles and in the ability to think. If this volume can serve the purpose of encouraging students to think it will have accomplished one of its principal purposes, not losing sight of the fact that the book is intended primarily to show how to do things. 4 Nearly all the proof-reading has been done by Professor John F. Pelly of Philadelphia. Because of Professor Pelly's thoroughly practical as well as theoretical knowledge of the subject matter, his conscientious and painstaking work is very greatly appreciated. I take this opportunity to thank Professor Ira N. Hollis and Professor F. Lowell Kennedy of Cambridge for the criticisms and suggestions which I received from them when the manuscript of this book was preparing. I am also greatly indebted to Mr. Walter C. Kerr, president, and Mr. Sidney E. Junkins, vice-president of PREFACE V Westinghouse, Church, Kerr & Company, for their encouragement and for making it possible to finish the book at this time. For placing at my disposal a great deal of information regarding the latest results in steam turbine engineering, which is usually very difficult to obtain, I am particularly indebted to Mr. Richard H. Rice of Lynn, and Mr. J. R. Bibbins of Pittsburg. I wish to thank Professor Arthur M. Greene of Troy and Mr. Albert Stritmatter of Cincinnati for suggestions relating to the subject matter. For various services in the preparation of this book, I should mention also Messrs. Francis Hodgkinson and Harold P. Childs of the Westinghouse Machine Company; C. P. Crissey, S. A. Moss, and W. E. Culbertson of the General Electric Company; C. P. Chasteney of the De Laval Steam Turbine Company; James Wilkinson, president of the Wilkinson Turbine Company; St. John Chilton of the Allis-Chalmers Company; H. H. Wait of the Western Electric Company; Carl S. Dow of the B. F. Sturtevant Company; and J. Clarence Moyer of Philadelphia. Many of the illustrations for the book have been provided, in some cases at considerable expense to themselves, by the Gassier Magazine Company, Westinghouse Machine Company, General Electric Company, De Laval Steam Turbine Company, Rateau Turbine Company, Kerr Turbine Company, Wilkinson Turbine Company, Allis-Chalmers Company, C. H. Parsons & Co., and Brown, Boveri & Co. Throughout the text important words and sentences are brought out by the use of bold-faced type, thus making the subjects of a paragraph visible at a glance. The author is always glad to answer correspondence with teachers relating to questions which inevitably arise in the discussion of designs for steam turbines, all of which cannot, of course, be taken up in detail in any book. JAMES AMBROSE MOYER. 417 WEST n8xH ST., NEW YORK, September, 1908. PREFACE TO SECOND EDITION. SINCE the issuance of the first edition, scientific investigation has not added much to our previous knowledge of the proper- ties of steam, nor have the new types invented in the interval become commercially successful. In fact the trend of things has been rather toward the general adoption of one of four types of steam turbines: (i) a single turbine wheel of the im- pulse type; (2) impulse wheels with two velocity stages in each pressure stage; (3) drum construction with " reaction " blading; and (4) a combination of (2) and (3) called a combined impulse and reaction type. Discoveries like that of Tesla's, claim- ing to have made possible very great simplification of turbine construction with unheard of improvements in economy, have at times attracted the attention of engineers, but always with the final result that the claims have not been made good. The really important developments of the last few years have been in the construction of increasingly large sizes. The largest turbine-generator now ready for installation is rated at 35,000 kilowatts, which is to be compared with a maximum size of 14,000 kilowatts of only three years ago. Certainly these are the days typical of the concentration of power in large units, not only in the turbine and generator room, but also in the boiler room, in the condenser pit, and in the installation of the other plant auxiliaries. Extremely large sizes are being installed be- cause they effect a substantial reduction in the unit cost of power generation. In spite of the general increase in the cost of raw and manufactured materials in the last five years, the application of steam turbines in power plants in the place of reciprocating engines has reduced the total first cost of large first-class power plants from $120 per kilowatt of rated capacity, vii viii PREFACE TO SECOND EDITION which was a fair average value five years' ago, to nearly $60 to-day. In these times when there is such general discussion of con- servation and efficiency, the low-pressure steam turbine takes an important place, because of its innumerable applications for preventing the wasting of any steam to the atmosphere. In nearly every large central station hundreds of pounds of steam exhausted from the auxiliaries at atmospheric pressure in excess of that required in the heaters for heating the feed is lost through the exhaust-heads. Modern methods of turbine application would save and utilize this steam for power. In the most modern practice, therefore, the greatest skill of the engineer is called upon more in connection with the methods of applying the commercial types of turbines already developed rather than in the actual designing of new types of machines. In steam engine designing there have been always unlimited possibilities; in steam turbine designing these are few. Most of the additions made in this edition have been, there- fore, mainly in the line of new applications. The chapter on low-pressure turbines has been rewritten and very much extended to include the latest developments and applications. This chapter should be unusually interesting to all engineers and students. New chapters have been added on Bleeder or Ex- traction Turbines and Mixed Pressure Turbines. Both mark recent successful developments in turbine applications, making it a still more important competitor of the reciprocating engine in the non-condensing field. The chapters on Heat Theory, Steam Flow, Nozzle Design, Blade Design and Reaction Tur- bine Design have been rewritten with the addition throughout the text of many illustrative examples and the inclusion in an appendix of a large number of practical exercises and problems to illustrate important principles, thus making the book con- siderably more serviceable than before as a class-room text. For these exercises the data are selected in most cases so as to simplify the calculations and to avoid taking too "much of the tune of the student or reader with purely numerical work. A PREFACE TO SECOND EDITION ix new entropy-total heat chart has been calculated and engraved, which embodies the most recent and reliable data on the prop- erties of superheated and saturated steam. In the preparation of this edition I am particularly indebted to Professor J. E. Enswiler of the University of Michigan and Professor J. P. Calderwood of The Pennsylvania State College for innumerable suggestions. Many important suggestions and criticisms have been re- ceived from Professor J. V. Ludy, Purdue University; Professor E. A. Fessenden, University of Missouri; Mr. M. Nusim, Gen- eral Electric Company, Lynn, Mass.; Mr. C. P. Crissly, Henry Worthington Co., New York; Mr. C. P. Chasteney, De Laval Steam Turbine Co.; and Mr. N. C. Miller of The Pennsylvania State College. Finally, important mention must be made of assistance in this work received from Mr. H. T. Herr, Vice-President and General Manager of the Westinghouse Machine Co.; Mr. Francis Hodg- kinson of the same company; Messrs. Richard H. Rice and Chas. K. West of the General Electric Company; Dr. E. J. Berg of Union College; Mr. Alfred Rigling of The Franklin Institute; and The Electric Journal. THE AUTHOR. STATE COLLEGE, PA., January i, 1914. CONTENTS CHAPTER PAGE I. INTRODUCTION i II. THE ELEMENTARY THEORY OF HEAT 10 III. NOZZLE DESIGN 29 IV. STEAM TURBINE TYPES AND BLADE DESIGN 58 V. MECHANICAL LOSSES IN TURBINES 115 VI. METHOD FOR CORRECTING STEAM TURBINE TESTS 126 VII. COMMERCIAL TYPES OF TURBINES 140 De Laval, Parsons, Westinghouse, Allis-Chalmers, Curtis, Rateau, Wilkinson, Zoelly, Sturtevant, Riedler-Stumpf, Kerr, Terry, Dake, etc. VIII. GOVERNING STEAM TURBINES 218 IX. LOW-PRESSURE (EXHAUST) TURBINES 247 X. MIXED-PRESSURE TURBINES 2$8a XI. BLEEDER OR EXTRACTION TURBINES 2586 XII. MARINE TURBINES 259 XIII. TESTS OF TURBINES 261 XIV. STEAM TURBINE ECONOMICS 273 XV. STRESSES IN RINGS, DRUMS, AND DISKS 313 Design of Turbine Wheels Critical Speeds of Loaded Shafts. XVI. GAS TURBINES 340 XVII. ELECTRIC GENERATORS FOR TURBINES 358 Direct-current Generators Alternating-current Generators. APPENDIX OF PRACTICAL EXERCISES 363 THE STEAM TURBINE CHAPTER I. INTRODUCTION. THE steam turbine is the most modern as well as the most ancient steam motor. Recently its development has gone by leaps and bounds; and, above all, in its applications it is gaining ground daily. Doubtless it is to be the most important prime mover of the near future. During recent years results have been secured with steam tur- bines that only a short time ago were considered practically unat- tainable. Primarily their great success lies in their adaptability to operation with high vacuums. Steam turbines are, therefore, almost ideally suitable for the conditions of modern engineering practice requiring both high vacuums* and high superheats. To-day in the economical use of steam they are unrivaled; and, because of improved manufacturing methods, marking the tran- sition from the experimental to the commercial stage, first cost is no longer a deciding factor favoring reciprocating engines. Compared with reciprocating steam and gas engines, steam turbines require much smaller and cheaper foundations, occupy less floor space, require fewer attendants, and because no lubrica- tion is required for any parts in contact with the steam, the con- densation becomes directly available for feed water. The highest superheats can be employed without affecting the choice of lubri- cants, and the cost of oil for lubrication is very low. A steam turbine of the simplest type is essentially a wheel similar to an ordinary water wheel, which is moved around by a steam jet impinging on its blades. Steam is directed against the turbine wheel by nozzles or similar passages delivering the * The question of the most profitable vacuum for given conditions is discussed on pages 273 to 281. I THE STEAM TURBINE FIG. i. A Small Modern Steam Turbine with Part of the Casting Removed. steam at mathematically exact angles, calculated to make the steam strike the blades of the wheel most advantageously. Fig. i is an illustration of a modern steam turbine with a part of the casing removed to show the construction. The turbine INTRODUCTION wheel W is shown here with numerous blades on its circumference. The steam comes to the turbine frorn the boilers through a suitable steam main connected to the top of the turbine at M and passes down through the pipe A to the steam-chest B. From this steam- FIG. 2. The Turbine Wheel and Nozzles. chest it is guided through one or more nozzles, from which it escapes at a high velocity to impinge on the blades on the circum- ference of the turbine wheel, which is thus made to rotate, and performs work by moving machinery connected to the shaft. Nozzles frorn which the steam is discharged are located around the periphery of the wheel as shown in Fig. 2 with their enlarged 4 THE STEAM TURBINE M ends, technically called mouths, very close to the blades.* Steam after passing through the blades enters the exhaust pipe at E (Fig. i ) and is discharged into the atmosphere or into a condenser, depending on whether the operation is non-condensing or con- densing. Preliminary to the study of the modern commercial types of steam turbines it is desirable to state briefly some of the most important stages through which this very ancient form of steam motor has passed in its development. Early History. The earliest notices of heat engines of any kind are found in a book by Hero of Alexandria, which was probably written in the second century before Christ. In this book of mechanical contrivances a steam reaction wheel is men- tioned. This first steam turbine is shown in Fig. 3. It is described as consisting of a hollow spherical vessel pivoted on a central axis and supplied with steam through the support M and one of the pivots from a boiler, B, beneath. Steam escaped from the vessel through bent pipes or nozzles N, N, facing tan- gentially in opposite directions. The spherical vessel was revolved by the re- action due to the escaping steam, just as a " Barker's mill " is moved by the water escaping from its arms. Any fluid escap- ing under pressure from a vessel which is free to move causes a "reaction" tending to displace the vessel in the opposite direction from the flow of the fluid. This reaction, although imperfectly understood by Hero, was perfectly applied in his steam turbine which was used to open the doors of temples. Only a few years ago a model of Hero's engine was constructed by a celebrated English engineer,! with, of course, * In this figure one of the nozzles is represented as if transparent to show its shape on the inside, and a part of the steel band around the blades is cut away to show the shape of the blades or vanes, as well as to illustrate the passage of steam from the nozzle into and through the blades. t See page 8. FIG. 3. Hero's Turbine. INTRODUCTION 5 all the advantages of modern machine tools and appliances, with the result that an engine was produced which, in economy, compared well with our elaborate and complicated modern engines. In 1577 a German mechanician, it is said, used a turbine simi- lar to Hero's to rotate reaming and burnishing tools, but from the time of Hero down to the seventeenth century there is no record of progress in the development of steam heat engines. In 1629 Branca, an Italian architect, designed a steam turbine (Fig. 4) resembling a water wheel, which was driven by the impulse from a jet of steam directed by means of a nozzle upon suitable vanes attached to the wheel. Branca's turbine en- gine, however, was not success- ful; and until the end of the nineteenth century, although in the interval many steam tur- bines and other rotary engines were patented, the piston or reciprocating steam engine, FIG. 4 . Branca's Turbine, under the leadership of Watt, had, commercially, an unrestricted field and remarkable results were accomplished. It is interesting to observe that the modern type of impulse turbine with a single row of blades like the one illustrated in Fig. i is practically the same, except for details, as the historic Branca's wheel. The principal difference is that Branca's wheel was not enclosed in a casing. Essential parts the nozzle, the blades, the wheel, and the shaft were practically the same as in some modern machines. Probably if Branca had under- stood the laws of the expansion of steam as we do to-day, he could have made a successful prime mover of his turbine. Those who came after him were aided not only by a superior knowledge but also by the opportunities for scientific investigation and the skill of our present-day workshops. De Laval Type. Dr. Gustaf De Laval, a Swedish scientist, was 6 THE STEAM TURBINE a pioneer in the modern commercial development of steam tur- bines. In 1882 he constructed his first steam turbine, which was similar in principle to Hero's reaction engine. De Laval's first turbine was designed primarily for driving his milk and cream separators, for which there was then a large sale. For other purposes, however, there was no general application, because at the very high speeds for which they were designed, it was difficult to utilize the power; and besides, the steam consumption was practically prohibitive. Later De Laval turned his attention to the development of Branca's steam turbine, and was remarkably successful. After much experimenting, he developed an impulse turbine which is still one of the standard makes. (See Figs. 82 to 86.) This great engineer, after investigating the possibilities of both Hero's and Branca's types and having decided to adopt the latter, began then some strikingly original inventive work, which, in many respects, led the way for the accomplishments of to-day. It should be stated, however, in this connection, that no engineer thinks of belittling De Laval's work because his investigations were mostly in the line of improvements to existing types. Un- questionably he must have the credit for producing the first commercially successful steam turbine. Many of the features of his original designs have actually contributed in no small meas- ure to our knowledge of machine design and thermodynamics, and have become fundamental principles underlying many of the most important modern steam turbine developments.* Parsons Type. With the early work of De Laval, however, the development of steam turbines designed to operate by the reaction principle of Hero's engine was not given up. Almost contemporaneously with De Laval, C. A. Parsons in England began the development of the well-known type which to-day bears his name, and which has made possible the brilliant records of turbine ocean steamers. In April, 1884, this great inventor took * The most important feature introduced by De Laval is that of the diverging nozzle (British Patent No. 7143 of 1889), the principle of which has influenced the development of practically all types of steam turbines. INTRODUCTION / out his first patents on steam turbines. The practicability of the steam turbine he then proposed is a striking feature of even his first patents. His specifications showed, above all, that a great deal of time and thought had been devoted to constructive details. Methods for reducing vibration, preventing leakage of steam, and providing for efficient lubrication contributed very largely to his success. Many of the details of this early turbine are now obso- lete, so that only a very short description will be given here. A section drawing of Parsons* first turbine is shown in Fig. 5. A FIG. 5. Early Parsons Steam Turbine. large central collar, C, is attached to the main shaft, S, which runs the length of the turbine. At the ends of the casing where the shaft passes through it the cross-section is reduced. The main shaft, S, supports a large number of rings which are held in place between the collar, C, and the nuts, N, which are screwed on the reduced section of the shaft at the ends. These rings, around their circumferences, support those turbine blades (b, b) which move with the shaft. There are, however, alternating with them, other rows of blades (c, c) attached to the inside of the turbine casing. Technically the blades b, b are called moving blades, and c, c are called fixed or stationary blades. Steam is admitted to the turbine blades through the annular chamber, A, encircling the collar, C, and then it passes to the right and left through the alternate rows of stationary and moving blades to the exhaust passages E, E one at each end of the turbine. The steam expands in the blades as in a nozzle, and its reaction moves the blades attached to the shaft, just as Hero's turbine was rotated by the steam escaping from its arms. 8 THE STEAM TURBINE By the "double-flow" arrangement in this design by which the steam is passed from the center to the exhaust at both ends there can be very little axial thrust on the shaft. Any thrust that does occur, however, is balanced by the pressure of the exhaust steam in the chambers E, E at the ends of the casing. A slight movement of the shaft toward either end checks the flow of the exhaust steam, and increases the back pressure at that end. This increased pressure then moves the shaft back to its normal position. Usually it is not possible to balance the parts of a rotating mass to make its center of gravity coincide exactly with the geometric center about which it revolves. In any machine like a steam turbine, when these two centers do not coincide excessive vibra- tions of the shaft are produced which at certain speeds * are sufficient to break it. To overcome this difficulty, Parsons in- geniously allowed a little lateral play, or " elasticity," as he called i i FIG. 6. Screw Type of Steam Turbine. it, for the shaft by means of a series of rings of two different diameters, in principle very much the same as the present con- struction of the main bearings of Parsons turbines (see Fig. 100), so that it was permitted to move laterally a certain amount, say a hundredth of an inch, to allow the proper adjustment in passing from rest to the normal speed of running. Among his early experiments Parsons also tried a purely reaction steam turbine, following almost exactly the published designs of Hero. This turbine, running with 100 pounds per * This phenomenon occurs at very definite speeds, called "critical," for every rotating mass. Fuller discussion, with a method for calculating "critical" speeds, is given on page 338. INTRODUCTION 9 square inch steam pressure and 27 inches of vacuum, gave an output of 20 horsepower at 5,000 revolutions per minute. Steam consumption was only 40 pounds per brake horsepower per hour, which was indeed a remarkably good result for that time. Screw Type. Still another kind of turbine, of only historical interest, should be mentioned. A large number of inventors have worked on the development of a screw type like Fig. 6. Hewitt worked for a long time on a turbine of this kind, and finally con- cluded the results were not satisfactory. Steam was admitted to this turbine through the chamber A, and passed through holes in the plates P, P into the helical grooves on the shaft. In these grooves the steam was expanded and then escaped to the exhaust pipes E, E at the two ends. Effective action of the steam was probably obtained only in the first part of the grooves; and after being deflected into a helical course, it rushed through to the exhaust without much additional effect in moving the shaft. Excessive leakage of steam between the helical threads and the casing was another serious difficulty. Recently a somewhat similar arrangement having two " screw wheels " meshing together not unlike spiral or helical gears has been successfully developed by the Buffalo Forge Company (see Figs. 157 and 158, page 2i7a). CHAPTER II. THE ELEMENTARY THEORY OF HEAT. NOTE. This short chapter may well be omitted, in reading, by those who are familiar with the thermodynamics of heat engines and with the use of entropy diagrams. It is intended primarily for practical engineers, who will find it par- ticularly valuable for reference " purposes, as the subject matter is completely indexed. TECHNICALLY the steam turbine must be regarded as a heat engine, that is, a machine in which heat is employed to do mechan- ical work. From the viewpoint of the practical man its function, the same as that of any other heat motor, is to secure as much work as possible from a given amount of steam, or, going a step farther back, from the combustion of a given amount of fuel. Heat theory is, therefore, of first importance. Heat is a form of energy like electrical, chemical, mechanical, potential, and kinetic. No doubt exists about the equivalence of the different forms of energy and their close relation to each other. Each, at will, can be changed into any of the other forms. The relative amount of heat in a body is observed, in common experience, by the sense of touch whether the body is a solid, a liquid, or a gas. By such experience we have learned to recog- nize certain sensations as hot or cold; and then, with more accuracy, to speak of degrees of temperature. Now when a hot and a cold body are brought together their temperatures become equalized. The hotter body always loses heat. The colder body always gains heat.* This experience is the principal basis for all heat calculations. When in the course of time it had been found that a more accurate method than that of the sense of touch was needed for heat determinations, methods utilizing the expansion of liquids * This phenomenon is called the second law of thermodynamics, that "heat energy always passes from a warm body to a cold body. " 10 THE ELEMENTARY THEORY OF HEAT II came to be generally employed. Many substances have a practi- cally uniform rate of expansion between the limits of temperature an engineer has to deal with. A small column of mercury in a glass capillary tube is usually taken as a standard for temperature measurements.* The mercury in an accurate thermometer expands very nearly ^|| of its volume when heated from the freezing temperature of water (32F.) to the boiling point (2i2F.). The expansion between the freezing point and the boiling point of water has therefore been called, arbitrarily, 180 F. For theoretical heat calculations the zero of temperature is taken as 492 F. below the freezing temperature of water; or, 460 below the Fahrenheit zero. This very low temperature is called the absolute zero, and at this point there is theoretically no heat energy. Temperatures measured from the absolute zero are called absolute temperatures and are indicated generally by T, to dis- tinguish them from the ordinary Fahrenheit temperatures, t, as read on a thermometer scale. Using these symbols, we have then in Fahrenheit degrees, T = t + 460. Absolute temperatures are convenient for heat calculations because " perfect " gases, at constant pressure, increase in volume in proportion to the increase in absolute temperature. * The ordinary mercury thermometers can be used to measure temperatures to about 575 F. with accuracy. For higher temperatures the capillary tube over the mercury should be filled with nitrogen or carbonic acid gas under high pressure. Such thermometers can then be used for temperatures up to 1000 F. If the mercury is not throughout its whole length at the same temperature as that being measured, a correction, k, given by the following formula must be added to the observed temperature, /, in Fahrenheit degrees: k = .000,088 D(t t'), where D is the length of the mercury column exposed, measured in Fahrenheit degrees, and t' is the temperature of the exposed part of the thermometer. When long thermometers are used in shallow wells in high-pressure steam pipes this cor- rection is often 5 to 10 F. For experimental data and direct-reading correction curves, see Moyer's Power Plant Testing, 2d edition (McGraw-Hill Book Co.), pages 31-33. 12 THE STEAM TURBINE Heat Units and Specific Heat. The amount of heat required to raise the temperature of one pound of water from 62 to 63 F. is taken arbitrarily as the standard English unit of heat, com- monly called the British thermal unit (B.T.U.).* The ratio of the amount of heat required to raise the temperature of a pound of water or steam one degree to the British thermal unit is called the specific heat.f The specific heat of steam and of gases changes in value accord- ing to the conditions under which the heat is applied. If heat is added to a vapor or a gas held in a closed vessel, with no chance for expansion, no external work is done, and therefore practically all the heat added is used to increase the temperature. This is the condition in a boiler when no steam is being drawn off. In this case the specific heat is symbolized by C v = specific heat at constant volume. If, on the other hand, the pressure is kept constant but the volume is allowed to change to permit expansion and the performing of external work, we say then, C p = specific heat at constant pressure. Heating at constant pressure is the condition that is most interesting to the engineer. When his engines are running the boilers are making steam at constant pressure. The heat energy absorbed by a pound of steam for raising only the temperature must be, obviously, approximately the same, regardless of the conditions of pressure and volume. Since for constant pressure conditions some external work is always done, requiring a larger amount of heat energy than for the case when the volume is constant, it follows that C p is always greater than C v . We should add, further, that an engineer's calculations con- cerning energy transformations in steam turbines are almost * In the C. G. S. system of units the kilogram-calorie, called in German Warmeeinheit (WE), is used as the standard heat unit, i kg.-cal. or i WE = 3.97 or nearly 4 B.T.U. t The specific heat of water at 200 F. is 1.005, an d of superheated steam an average value of .6 is often assumed in rough calculations for steam at the usual boiler pressures in power plant practice for superheats less than 150 F. Mean values of the specific heat of superheated steam are given by the curves in Fig. 30. THE ELEMENTARY THEORY OF HEAT 13 without exception for the condition of constant pressure, and, consequently, only values of C p are generally useful. Most gases have practically constant values for their specific heats. At temperatures near the boiling point, the heating of vapors, like steam, is influenced by molecular attraction, so that their specific heats are variables depending on conditions of tempera- ture and pressure. . The specific heat of superheated steam de- creases with increasing temperatures to a minimum value. The values of specific heat increase slightly, on the other hand, with an increase of pressure.* Mechanical Equivalent of Heat. Heat and work are both forms of energy and are " equivalent," meaning that energy can be transformed into mechanical work, and that work, as a form of energy, can be changed back again into heat. The relation is expressed by i British thermal (heat) unit = 778 foot-pounds (work). HEAT AND WORK. Heat is a form of energy. Each of the various kinds of heat motors, such as the steam engine, the steam turbine, the gas engine, or the gas turbine, is a machine for obtaining mechanical work from heat energy. In the general principles of operation the steam turbine and the reciprocating or " piston " steam engine are essentially similar machines. Both do work according to the same heat relations. The gas turbine is somewhat different. This new motor, which as yet has scarcely reached a practical stage of development, will be discussed in its proper place. In a reciprocating steam engine working " expansively " the steam is admitted at boiler pressure until the point of cut-off; and during the remainder of the stroke the piston is pushed ahead, or does work, by the expansion of the steam shut up in the cylinder. In the steam turbine the heat process is analogous, except that * Knoblauch and Jakob, Zeit. Verein deutscher Ingenieure, Jan. 5, 1907, and an article by the author in Mechanical Engineer (London), Aug. 24, 1907. 14 THE STEAM TURBINE the flow of the steam, instead of being intermittent, is continuous. Steam is continually pushed into the nozzles, or similar steam passages, and expanding, expends its internal energy in producing velocity. Vanes or blades, 'fixed to a rotating wheel, are placed near the nozzles so that the jets of steam are directed against them. These blades or vanes thus set in motion move the wheel and with it the shaft which transmits the power. Theoretically the work from expanding steam behind a piston is exactly the same as that we obtain from a nozzle. The difference is only in the method for making the heat energy available for doing work. Before going farther with the discussion of -how the steam tur- bine converts heat energy into work, the more familiar case of the reciprocating steam engine will be considered briefly, because it is assumed the reader is already more or less familiar with its heat processes. By the static pressure in the steam pipes and in the boiler the steam is pushed into the engine cylinder and causes the piston to move up to the point where the supply of steam is shut off. Then the steam expands, reducing, at the same time, the pressure till the piston has reached the end of the cylinder. On the return stroke the steam is discharged at a nearly constant low pressure into the atmosphere or into a condenser. Now on the " working " stroke when the steam is being pushed into the cylinder,* and when it is expanding, the steam is doing work at the expense of the heat energy put into it by the fires under the boiler. The heat in a pound of steam at a given pressure and temperature represents a definite amount of energy. Expansion of the steam in the cylinder after cut-off is accompanied, therefore, with a reduction of pressure and temperature, and the work done is in proportion to the heat energy lost by the steam. Thus heat energy and work go hand in hand. A loss to one is a gain to the * Until the point of cut-off is reached, all the time that steam is being pushed into the cylinder work is being done at the expense of the boiler pressure. Actually the pressure in the boiler is a little lower after the amount of steam required for a stroke has been taken out than it was before.. When, however, the strokes of the engine come in quick succession, the variation in boiler pressure is not per- ceptible. THE ELEMENTARY THEORY OF HEAT other. Fig. 7 shows a typical steam engine indicator card, repre- senting, diagrammatically, the heat relations that have just been discussed. The horizontal scale of coordinates (abscissas) repre- sents volumes, and the vertical scale (ordinates) represents pressures. It is obvious then that any area included by the lines of this diagram represents work done by the steam. In this figure P t and Vj represent initial pressure and volume, and P 2 and v 2 the corresponding final conditions, meaning the pressure and volume at the end of the "working" stroke. This diagram as it applies to the steam engine may be analyzed briefly as follows: FIG. 7. Pressure-Volume Diagram Showing Work Areas. 1. Area AOiB is the work done in " pushing" the steam into the cylinder against the resistance of the piston to motion. 2. Area I2CB is the work done when the steam is expanding. 3. Area A432C is the work lost in the heat energy discharged in the exhaust.* 4. Area 40123 is the net work done. The discussion given above is, of course, for the ideal case where the cylinder clearance is neglected and expansion to back-pressure (P^) is complete. * If an almost perfect vacuum were attainable this loss would be practically negligible. Actually with the best condensing apparatus it is quite large. 16 THE STEAM TURBINE The same diagram (Fig. 7) can also be used for the analysis of the work done by steam expanding in the nozzles or similar passages * of a turbine. The work done in " pushing " the steam into the engine cylinder has its counterpart now in the work done by the steam in entering the nozzle, so that, 1. Area AOiB is the work done in " pushing " the steam out of the pipes or receiving vessels into the nozzle, f 2. Area I2CB is the work done during expansion at the expense of the heat energy, to give velocity to the steam. 3. Area A432C is work lost by the steam in forcing its way against the external or exhaust pressure. 4. Area 40123 is the work done in producing velocity. The work of " pushing " the steam into the nozzle produces initial velocity^, or " velocity of approach." In all practical steam turbine nozzles this initial velocity, compared with the final velocity after expansion, is very small. For this reason, in the calculations required for the designing of nozzles and blades, this initial velocity is usually neglected. Practical designers, therefore, are interested only in the heat energy of the area 123 and the velocity it represents. In order to secure high efficiency and low steam consumption the designer is always striving to make this area as large as possible, allowing, of course, for other limiting conditions. As the result of the comparison of the heat functions of steam turbines and reciprocating steam engines, \ve should observe, then, that the heat energy in a pound of steam available for per- forming useful work is exactly the same whether the steam goes to the one or to the other. It follows then also that, theoretically, * In some types of turbines there are no nozzles, but instead stationary blades are used 'which are arranged to expand the steam just as in a nozzle. In this chapter, therefore, where the term "nozzle" is used it will be assumed to apply as well to stationary "expanding" blades. t The amount of this work, or the area AOiB, is very small in the case of the turbine compared with that in the steam engine. V 2 t This initial velocity, V , is calculated from the relation P 1 v 1 = , where P l and v are the initial pressure and volume of a pound of steam and g is the acceleration due to gravity (32.2). All velocities are in feet per second. THE ELEMENTARY THEORY OF HEAT 17 the steam consumption for the same conditions of temperature and pressure is the same for the turbine as for any other form of engine. Discussion of the merits of different forms of steam motors with only the theoretical viewpoint in mind is, therefore, useless. Only the conditions in practice affecting the design of commercial machines are of any significance in determining the type of steam motor to be used for given conditions of service. HEAT THEORY RELATING TO THE DESIGN OF NOZZLES AXD BLADES. Diagrams similar to those made on a steam engine indicator (Fig. 7), showing for an engine stroke the conditions of pressure and volume inside the cylinder, are very useful in the design and operation of reciprocating steam engines, but they are of very little use for work relating to steam turbines. In a steam tur- bine it is not practicable to put a measured amount of steam through a nozzle " at a time " as the flow is practically con- tinuous. The pressure- volume diagram has, therefore, a very limited application. Another kind of diagram, the details of which are somewhat more difficult to understand, is universally used by steam turbine engineers. In this diagram, which will now be described, any surface represents accurately to given scales a quantity of heat. Absolute temperatures (T) are the ordinates, and entropies * (0) are the abscissas. * Entropy, which Perry calls the " ghostly quantity," has no real physical significance, so that complete definition is not possible. If dQ is a small amount of heat added to a body, and T is the absolute temperature at which the heat is added, then the change in entropy of that body is -^ , or d = -^ Entropy of saturated steam above the entropy of water at the freezing point (32 F.) is easily calculated. For saturated steam at any pressure, then, = ^r + (or 0), where x is the quality of the steam, r is the latent heat of evaporation or " heat of vapor- ization," T is the absolute temperature, and n (or 0) is the entropy of the liquid (water). All values of latent heat of evaporation, heat of the liquid, total heat, etc., given in steam tables are in heat units above 32 F. The symbols used here are those given in Peabody's Steam and Entropy Tables, published by John Wiley & Sons, New York, and in Marks and Davis' Steam Tables and Diagrams, published by Longmans, Green & Co., New York. i8 THE STEAM TURBINE 000 400- 200 1.0 Entropy (0) FIG. 8. A Simple Entropy- Temperature Diagram. 2.0 Fig. 8 shows a simple heat diagram laid out with absolute temperature and entropy for the coordinates. Steam at a certain condition of temperature and entropy is represented here by the point A. Then if some heat is added, increasing both temperature and entropy, the final condition is repre- sented by the point B, and the area ABCD represents the heat added in passing from the condition at A to the condition at B. Such a diagram is usually called an entropy-temperature dia- gram, although the name heat diagram would prob- ably be more appropriate, since every area represents a definite amount of heat. Another entropy-temperature diagram is shown in Fig. 9, representing by the various shaded areas the heat added to water at 32 F. to completely vaporize it at the pressure P t . The un- shaded area under the irregular curve AB represents the heat in a pound of water at the freezing point (32 F. or 492 in absolute temperature). The area OBCD is the heat added to the water to bring it to the temperature of vaporization, or in other words, this last area represents the heat of the liquid (q) given in the steam tables for the pressure P x . Further heating after vaporiza- tion begins is at the constant temperature T 1 corresponding to the pressure P w and is represented by an increasing area under line CE. When " steaming " is complete, the latent heat, or the heat of vaporization (r), is the area DCEF. If after all the water is vaporized more heat is added, the steam becomes superheated, and the additional heat required would be represented by an area to the right of EF. The use of the entropy-temperature diagram in exhibiting the THE ELEMENTARY THEORY OF HEAT behavior of steam during expansion and the various heat losses and exchanges in the passage of steam through a turbine will now be discussed and illustrated with a practical example. II 2 400 800- 300 /Steaming \ T! and P, \ .5 I'.O ll6 2.0 Entropy(0) FIG. 9. Entropy-Temperature Diagram showing the Total Heat in a Pound of Dry Saturated Steam at the Temperature T r 400 800- 300 200 /Steaming T, \ O M .129 .566 1.528 Entropy (0) FIG. 10. Practical Example Illustrated with an Entropy-Temperature Diagram. Fig. 10 illustrates the heat process going on when feed water is received in the boilers of a power plant at 100 F., is heated and 20 THE STEAM TURBINE converted into steam at a temperature of 400 F., and then loses heat in doing work. When the feed water first enters the boiler its temperature must be raised from 100 to 400 F. before any " steaming " begins. The heat added to the liquid is the area MNCD. This area represents the difference between the heat of the liquid of steam at 400 F. (q c ) and at 100 F. (q) and is about 306 B.T.U. The horizontal or entropy scale shows that the difference in entropy between water at 100 and 400 F. is about 437-* Every reader should understand how such a diagram is con- structed and especially how the curves are obtained. In this case the curve NC is constructed by plotting from the steam tables the values of the entropy of the liquid (usually marked with the symbol n or 6) for a number of different temperatures between 1 00 and 400 F. If now water at 400 F. is converted into steam at that tem- perature, the curve representing the change is necessarily a con- stant temperature line and therefore a horizontal, CE. Provided the vaporization has been complete, the heat added in the " steaming " process is the latent heat or heat of vaporization of steam (r) at 400 F., which is approximately 827 B.T.U. The change in entropy during vaporization is, then, the heat units added (827) divided by the absolute temperature at which the change occurs (400 + 460 = 860 F. absolute) or r 827 = The total entropy of steam completely vaporized at 400 F. is, therefore, the sum of the entropy of the liquid (water) .566 and the entropy of the steam .962, or 1.528.! To represent then by CE this final condition of the steam, the point E is plotted where entropy measured on the horizontal scale is 1.528, as shown in the * As actually determined from Marks and Davis' Steam Tables, pp. 9 and 15, the difference in entropy is .5663 .1295 or .4368. Practically it is impossible to construct the scales in the figure very accurately. t Entropy, like the total heat (#), and the heat of the liquid (q), is measured above the condition of freezing water (32 F.). THE ELEMENTARY THEORY OF HEAT 21 figurec* The area MNCEF represents then the total heat added to a pound of feed water at 100 F. to produce steam at 400 F., and the area OBCEF represents, similarly, the total heat (H in the steam tables) in a pound of steam at 400 F. above that in water at 32 F. Adiabatic Expansion and Available Energy. The practical example illustrated by Fig. 10 (repeated here) will also be used to explain how the entropy-temperature diagram can be used to ,566 1.528 Entropy (0) FIG. 10, Practical Example Illustrated with an Entropy-Temperature Diagram. show how much work can be obtained by a theoretically perfect engine from the adiabatic expansion of a pound of steam. When steam expands adiabatically without a gain or loss of heat its temperature falls. Remembering that areas in the entropy- * The point E is shown located on another curve ST, which is determined by plotting a series 01 points calculated the same as E, but for different pressures. If more heat had been added than was required for vaporization, the area DCEF would have been larger and E would have fallen to the right of ST, indicating by its position that the steam had been superheated. The curve ST is therefore a "boundary line" between the saturated and superheated conditions. This curve can also be plotted from the values obtained from a table of the entropy of dry saturated steam. 22 THE STEAM TURBINE temperature diagram represent quantities of heat and that in this expansion there is no exchange of heat, it is obvious that the area under a curve of adiabatic expansion must be zero ; and this con- dition can be satisfied only by a vertical line which is a line of constant entropy.* For the case in Fig. 10 the expansion curve will lie, therefore, along the line EF, and if the temperature falls to 100 F. the expansion will be from E to G, and during this change some of the steam has been condensed. If now heat is removed from this mixture of steam and water till all the steam is reduced to the liquid state, but without further lowering of the temperature, the horizontal line GN f will represent the change in its condition. The quantity of heat absorbed in this last process - technically known as condensing the steam is represented by the area MNGF, and the heat converted into work is, therefore, the. area NCEG; and this is called the available energy. By means of diagrams like those in the preceding figures, it will now be shown how the available energy of dry saturated steam for any given conditions can be readily calculated from the data given in steam tables. Fig. ii is an entropy-temperature diagram representing dry saturated steam which is expanded adiabatically from an initial temperature T x corresponding to a pressure P x to a lower final * Since in an adiabatic expansion there is no change of entropy, lines of constant entropy, in practice, are often called " adiabatics ". It is very rare in steam turbine work that the expansion in a nozzle departs far from the adiabatic. For this reason other kinds of expansion are not mentioned here. t That the steam might be dry and saturated, the expansion would have had to follow the curve E T and G would have appeared at G'. The heat of the liquid, q, of a pound of steam at 100 F. is represented by OBNM, and the heat of vaporization (r) is MNG'F', so that the total heat (q + r or H) is OBNG'F'. The total heat of wet steam is expressed by q + xr, where x is the quality or relative dryness. In the case of this adiabatic expansion, then, q is as before OBNM and xr is MNGF. It is obvious also that the lines NG and NG' have the same relation to each other as the areas under them, so that line NG area MNGF xr NG line NG' ~ area MNG'F' r ' NG' showing that the quality of the steam at any point, G, on a constant temperature line (which for saturated steam is also a constant pressure line) is determined as in this case by the ratio of NG to NG'. THE ELEMENTARY THEORY OF HEAT 23 temperature T 2 corresponding to a pressure P 2 . The other initial and final conditions of total / heat (H) and entropy (0) are represented by. the same subscripts i and 2. The available energy or the work that can be done by a perfect engine under these conditions is the area NCEG. It is now desired to obtain a somewhat complicated and will not be dis- cussed at this place, because it can be worked out more simply with the aid of the entropy-total heat chart in the appendix.- The method to be used for this case will be discussed therefore after the use of this chart has been explained. (See pages 55 and 57.) The case of initially wet steam is, however, easily treated in the same way as dry and saturated steam. Fig. 12 is an example of r. / Ti and Pi E \E N/ T 2 and 0=0 x l 2 Entropy FIG. 12. Entropy- Temperature Diagram for Initially Wet Steam. the case in hand. At the initial pressure P 1; the total heat of a pound of wet steam (q x + x^) is represented in this, diagram by the area OBNCP"< X . The initial quality of the steam (xj is repre- CE" sented by the ratio of the lines The available energy from CE adiabatic expansion from the initial temperature T\ (correspond- ing to the pressure PJ to the final temperature T 2 (corresponding 26 THE STEAM TURBINE to the pressure P 2 ) is the area NCE"G". If we call this available energy E ou ,, we have E aw = area OBNCEGF + FGG'F'- OBNG'F' - G"E"EG, E aw = H t - H 2 + (fa - &) T 2 - (& - (T t - T 2 ),* E aw = H t - H 2 + (fa - fa) T 2 - ^ (i - xj (T t - T 2 ). (I') M The velocity corresponding to this energy is found by substi- tution in equation (2), just as for the case when the steam was initially dry and saturated. Example. Calculations for the velocity resulting from adia- batic expansion for the same conditions given in the preceding example on page 24, except that the steam is initially 5 per cent, wet, are given below. P! = 165 Ibs. abs. T! = 826.0 F. P 2 = 15 Ibs. abs. T 2 = 673.0 F. HI = 1195.0 B.T.U. H 2 = 1150.7 B.T.U. 0i = 1.5615. <& = 1-7549- ri = 856.8 B.T.U. xi = i. oo - .05 = .95. E aw = 1195.0 - 1150.7 + (i.7549 ~ 1-5615) 673-0 - X.05 (826.0 - 673-0), E aw = 166.53 B.T.U. V =223.7 ^E aw = 2-23.7 X 12.90 = 2886 feet per second. * In general terms, $ = * 4- 0. Here < t = - + QI because x = i. THE ELEMENTARY THEORY OF HEAT 2/ FIG. 13. Example of a Well-designed Nozzle. rffh DeLaval Type. Nozzle Diaphragm Curtis Type. FIG. 14. Examples of Standard Designs of Nozzles. 28 THE STEAM TURBINE It is observed that the theoretical velocity is reduced from 2953 to 2886 feet per second by the presence of moisture in the steam. The percentage reduction in velocity is, however, only about 2 per cent, while the amount of moisture is 5 per cent. The shaded area NCEG in Fig. n is also known as the the- oretical Rankine cycle for the case where the steam supplied is initially dry saturated. The available energy, therefore, as given by equation (i) on page 23, multiplied by 778 gives the maximum theoretical foot-pounds of work that can be accom- plished with this cycle, neglecting losses, from a pound of dry , steam. There are 33,000 X 60 foot-pounds in one horsepower- hour, and hence dividing 33,000 X 60 by E a X 778 we get the theoretical steam consumption (" water rate") of an engine or turbine using the ideal Rankine cycle with steam initially dry saturated. Similarly the area NCE" G" in Fig. 12 shows the available work for the theoretical Rankine cycle when the steam is initially wet, and the theoretical steam consumption of the Rankine cycle for this case is 33,000 X 60 divided by E aw * X 778. Fig. 32, page 57, shows also the Rankine cycle for steam ini- tially superheated. Calculation of theoretical steam consump- tion is similar to the cases already explained. The most important part of the design of a nozzle is the deter- mination of the areas of the various sections especially the smallest section, if the nozzle is of an expanding or diverging type. Various forms of standard nozzles are shown in Figs. 13 and 14. In order to calculate the areas of nozzles we must know how to determine the quantity of steam (flow) per unit of time passing through a unit area. It is very essential that the nozzle is well rounded on the " entrance " side and that sharp edges along the path of the steam are avoided. Otherwise it is not important whether the shape of the section is circular, elliptical, or rectangular with rounded corners. Typical " square," " rectangular " and circular nozzle sections used in different makes of commercial turbines are shown in Fig. 15. * From Equation (i') page 26. THE ELEMENTARY THEORY OF HEAT 28a Example. Calculate the work done in foot-pounds by one pound of steam expanding behind a piston in a reciprocating engine for the conditions given in the example on page 24. (See discussion on page 16.) Ans. 174.2 X 778 ft.-lbs. FIG. 15. Sections of Nozzles Used in Commercial Turbines. Example. If the flow of steam at 165 pounds per square inch absolute pressure from a nozzle with a cross-sectional area of .128 square foot is 200 pounds per second, what is the velocity of the discharging jet? At the pressure stated steam has a specific volume of 2.75 cubic feet per pound (from steam tables). Let V = velocity of discharge (ft. per sec.) A = area of nozzle = .128 sq. ft. AV = volume discharged (cu. ft. per sec.) .128 V = 200 X 2.75 V = 4297 ft. per sec. If turbine blades could be made to transform all this velocity into useful work, how much horsepower could be transmitted to machinery from its shaft? WV 2 Kinetic energy of jet (ft.-lbs. of work per sec.) = - > where W is the weight in pounds of the steam flowing and g is the acceleration due to gravity (32.2 ft. per sec.). ft.-lbs. of work per sec. Horsepower = - 550 200 X (42Q7) 2 .2 X32.2 X 550 = about 104,300 h.p. 28b THE STEAM TURBINE Example. What is the theoretical steam consumption (water rate) of the Rankine cycle for the conditions given in the ex- ample on page 24; that is, for steam initially dry saturated? 33,000 X 60 ; Ans. -^^- 174.46 X 778 Example. What is the theoretical steam consumption of the Rankine cycle for steam initially wet at the conditions stated in the example on page 26? A 33.c X 60 . 166.53 X 778 CHAPTER III. , FLOW OF STEAM AND NOZZLE DESIGN. /Flow of Steam through Nozzles. The weight of steam dis- :harged through any well-designed nozzle with a rounded inlet, (similar to those illustrated in Figs. 13 and 14, depends only on the initial absolute pressure (PJ, if the pressure against which the nozzle discharges (P 2 ) does not exceed .58 of the initial pres- ^sure. This important statement is well illustrated by the follow- ing example. If steam at an initial pressure (PJ of 100 pounds per square inch absolute is discharged from a nozzle, the weight of steam flowing in a given time is practically the same for all values of the pressure against which the steam is discharged (P 2 ) which are equal to or less than 58 pounds per square inch absolute. If, however, the final pressure is more than .58 of the initial, the weight of steam discharged will be less, nearly in proportion as the difference between the initial and final pressures is reduced. (See pages 32 and 33.) The most satisfactory and accurate formula for the " constant flow " condition, meaning when the final pressure is .58 of the initial pressure or less, is the following, due to Grashof,* where F is the flow of steam | (initially dry saturated) in pounds per * Grashof, Theoretische Maschinenlehre, vol. i, iii; Hiitte Taschenbuch, vol. i, page 333. Grashof states the formula, but the formula given in equation (3) is accurate enough for all practical uses. t Napier's formula is very commonly used by engineers and is accurate enough for most calculations. It is usually stated in the form 70 where F, 'P,, and A Q have the same significance as in Grashof' s formula. The following formula is given by Rateau, who has done some very good theoretical 29 30 THE STEAM TURBINE second, A is the area of the smallest section of the nozzle in square inches, and P x is the initial absolute pressure of the steam in pounds per square inch, F-^. ( or, in terms of the area, A - ^rl (3') These formulas are for the flow of steam initially dry and saturated. An illustration of their applications is given by the following practical example. Example. The area of the smallest section (A Q ) of a suitably designed nozzle is .54 square inch. What is the weight of the flow (F) of dry saturated steam per second from this nozzle when the initial pressure (PJ is 135 pounds per square inch absolute and the discharge pressure (P 2 ) is 15 pounds per square inch absolute ? Here P 2 is less than .58 P l and Grashof's formula is applicable, or, 60 ,., .54 X 116.5* F = -^ - = i. 049 pounds per second. 60 When steam passes through a series of nozzles one after the other as is the case in many types of turbines, the pressure is reduced and the steam is condensed in each nozzle so that it becomes wetter and wetter each time. In the low-pressure noz- zles of a turbine, therefore, the steam may be very wet although and practical work on steam turbines, but his formula is too complicated for convienent use: F = .001 4 .Pi [15.26 - .96 (log P l -f log. 0703)]. Common or base 10 logarithms are to be used in this formula. * A curve from which values of ^-^ can be read is given on page 38 (Fig. 19). "\ The flow (F) calculated by Napier's formula for this example is F = , or 1.041 pounds per second. NOZZLE DESIGN 31 initially it was dry. Turbines are also sometimes designed to operate with steam which is initially wet, and this is usually the case when low-pressure steam turbines (see Chapter IX) are operated with the exhaust from non-condensing reciprocating engines a practice which is daily becoming more common. In all these cases the nozzle area must be corrected for the wetness of the steam. For a given nozzle the weight discharged is, of course, greater for wet steam than for dry; but the percentage increase in the discharge is not nearly in proportion to the per- centage of moisture as is often stated. The general equation for the theoretic discharge (F) from a nozzle is in the form* * The general equation for the theoretic flow is / T - ft + l-i F- A V *g*P, \(P*\ k {S*\ k \, A V ( *-I)V,LU/ UJ J where the symbols F, A , P lt and g are used as in equations (2) and (3). P 2 is the pressure at any section of the nozzle, v l is the volume of a pound of steam at the pressure P v and k is a constant. The flow, F, has its maximum value when 2 k + l k is a maximum. Differentiating and equating the first differential to zero gives i P*! * PI U + i P 2 is now the pressure at the smallest section, and writing for clearness P Q for P 2 , and substituting this last equation in the formula for flow (F) above, we have F = Now regardless of what the final pressure may be, the pressure (P ) at the smallest section of a nozzle (^4 ) is always nearly .58 P l for dry saturated steam. Making then in the last equation P = .58 P l and putting for k Zeuner's value of 1.135 for dry saturated steam, we may write in general terms the form stated above, where K is another constant. See Peabody's Thermodynamics of the Steam Engine, page 132; Zeuner's Theorie der Turbinen, page 268 (Ed. of 1899). 32 THE STEAM TURBINE where P is the initial absolute pressure and v l is the specific volume (cubic feet in a pound of steam at the pressure P^. Now, neglecting the volume of the water in wet steam, which is a usual approximation, the volume of a pound of steam is proportional to the quality (X). For wet steam the equation above becomes then The equation shows, therefore, that the flow of wet steam is inversely proportional to the square root of the quality (xi). Grashof's equations can be stated then more generally as (4) These equations become, of course, the same as (3) and (3') for the case where x l = i. Flow of Steam when the Final Pressure is more than .58 of the Initial Pressure. For this case the discharge depends upon the final pressure as well as uponrthe initial. No satisfactory formula can be given in simple terms, and the flow is most easily calculated with the aid of the curve in Fig. 16 due to Rateau. This curve is used by determining first the ratio of the final to the initial p pressure -*, and reading from the curve the corresponding CO- PI efficient showing the ratio of the required discharge to that cal- culated for the given conditions by either of the equations (3) or (4). The coefficient from the curve times the flow calculated from equations (3) or (4) is the required result. Obviously the discharge for this condition is always less than the discharge when the final pressure is equal to or less than .58 of the initial. The actual design of the nozzles for a commercial turbine will NOZZLE DESIGN 33 be taken up in the next paragraph; but before this is done, one other equation used almost continually in nozzle and blade de- signs must be explained. It is to find the quality of the steam after an adiabatic expansion. The initial quality of the steam is usually determined by the conditions in the boiler equipment, or 1.0 Ratio of FiuaLto Initial PressureJijL FIG. 16. Coefficients of the Discharge of Steam when the Final Pressure is Greater than .58 of the Initial Pressure. is given in the engineer's specifications for a new design, but the quality of the steam after each expansion must be calculated. The general equation for adiabatic flow (constant entropy *) is and solving, * 1 (5) where the subscript i attached to the symbols refers to the initial condition, and the subscript 2 to the final. * See footnote on page 17. The terms 6 l and 34 THE STEAM TURBINE are the entropies of the liquid (water) at the initial and final con- ditions, and the other symbols are used as before. To avoid the laborious calculation of equation (5) to determine the quality after adiabatic expansion, curves of steam quality have been calculated and plotted on the entropy-total heat chart in the appendix. To illustrate the use of these curves an example is given below. Example. Steam at 165 pounds per square inch absolute pressure (Pj), which is 4 per cent, wet (^ = .96), is expanded adiabatically in a nozzle to 15 pounds per square inch absolute (P 2 \ What is the quality after expansion? Method. A point is first located on the chart where the quality curve for x = .96 crosses the pressure line for 165 pounds as shown diagranimatically in Fig. 17. A horizontal line of constant entropy drawn through this point shows at its intersection with FIG. 17. Illustrates the Use of the Entropy-Total Heat Chart to Determine the Quality of Steam after Expansion. the pressure line for 15 pounds the quality after expansion. In this case the quality is .837. For practical designing to get satis- factory results the quality should be read to three significant figures. Nozzle Calculations. In the calculations to determine the dimen- sions of a nozzle it is necessary to have given the following data : (i) the weight of steam that is to be delivered through the nozzle to develop the required power in the turbine. NOZZLE DESIGN 35 (2) the initial and final pressures (P x and P 2 )- (3) the quality fa) of the steam supplied. With these data^4 . is then calculated by substitution of these quantities in equation (4'). This is the area at the smallest section or throat as shown in Fig. 18. FIG. 18. A Typical Expanding Nozzle. The area of the nozzle can be determined by simple calculations only at the smallest section or throat. To determine the area at any other section of the expanding portion between the throat and the mouth involves equations of the form of those at the bottom of page 31. It is therefore convenient to determine the sections other than the throat by a proportional method. Now the areas of different sections depend on the following three conditions: (1) the velocity of the steam. (2) the specific volume. (3) the quality or dryness. The essential condition to observe is that the weight of steam flowing per second is the same at every section; and for the same flow the areas are inversely proportional to the velocities at any two sections compared, and directly proportional to both the specific volumes and the qualities. We may then write the equation A V v x ** x /% _ * vy *P vx /\.\ *9f -77" X ~~~ X ~~ ) (y) A o I 36 THE STEAM TURBINE A = area of nozzle at the smallest section in square inches. A x = area of nozzle at any section of expanding portion in square inches. V = velocity of steam at the smallest section in feet per second.* V x = velocity of steam at any section in feet per second.* v = specific volume at the smallest section in cubic feet per pound. v x = specific volume at any section in cubic feet per pound. x = quality of steam at the smallest section. x x = quality of steam at any section. V v x The product - ^ X - X - when calculated for the largest V* v x section or mouth, is often called the expansion ratio (see Fig. 21, page 41), and is very nearly proportional to the ratio of the initial to the final pressure. An example will now be given to show how the actual area of the nozzles of a commercial turbine can be calculated. Example. A test of a De Laval turbine was as follows: Pressure in the steam-chest (Pi) .211.5 pounds absolute Vacuum referred to 30-in. barometer. . 26.6 in. mercury Moisture in steam ..................... 2.2 per cent. Brake horsepower ........................... . . 333 Steam consumption, per brake horsepower- hour as weighed (" wet ") ........ I 5-5 I pounds Number of nozzles open .............. . ........ ... 8 In this case P 2 is given as 2(5.6 inches vacuum, which is less than 2 pounds absolute pressure, and is therefore less than .58 PI and formula (4') is applicable, so that the " throat " area of the eight nozzles is expressed by * Since practically all the loss in a nozzle occurs before the steam " emerges " from the throat, the same coefficient applies to both V and V^ and cancels when expressed in equation (6). The non-expanding nozzles shown on page 51 are no more efficient than equally well made expanding nozzles. NOZZLE DESIGN 37 where x l = .978, P t = 211.5 pounds per square inch absolute, and p = 333__LiU = 1L-5 pounds wet steam per second. 3600 3600 ~ X (211. 5)' 97 3600 ^o = -333 X- ? X .989 = .472 square inches. 3600 The area of the throat of each nozzle is therefore .0590 square inches. The value of - - was read from the curve * of - in (211. 5 )- 97 P/ 7 Fig. 19. The nozzles of most commercial types of steam turbines are made with straight sides as shown in Fig. 18, so that in addition to the area at the throat only one other area must be found to fully determine the expanding portion. This is obviously most easily determined at the mouth, since the velocity must be calcu- lated from the available energy for an adiabatic expansion from P l = 211.5 pounds per square inch absolute to P 2 = 1.67 pounds per square inch absolute (26.6 inches vacuum). This available energy can be calculated by equation (i') for initially wet steam, but the calculation is laborious, and instead the energy will now be read from the entropy-total heat chart in the appendix. The point is first located on the chart where the line for 211.5 pounds pressure crosses the .978 steam quality line (estimated). Read- ing the scale of abscissas at this point we find that the total heat energy in a pound of steam at this condition is 1181 B.T.U. By following a horizontal line from this point across the chart as indicated diagrammatically in Fig. 20 till it intersects the pres- sure line corresponding to 1:67 pounds (estimated), the total heat energy escaping with the exhaust steam after adiabatic expansion * The curve was made in this form to make the final form of the result more convenient for slide-rule or cancellation calculations. THE STEAM TURBINE Initial Pressure, V l (5 to 50 Lbs.) 10 15 20 25 30 40 45 60 80 400 120 140 160 .180 200 220 240 JnitiaMIressure, P^ (50 to 260 Lbs.) FIG. 19. Curves Showing Values of 9 ~ A \ 1181 B.T.U. 874 B.T.U. Total Heat (H) FIG. 20. Illustrates the Use of the Entropy-Total Heat Chart for Deter- mining the Available Energy in a Pound of Steam. NOZZLE DESIGN 39 as read on the scale of abscissas is 874 B.T.U. The difference between the two readings, or 307'B.T.U., is the available energy (E a ). The quality at the end of expansion (x 2 ) as read from the curves is .767. In this way the labor of calculating oc 2 is saved. From the value of the available energy due to expansion, E aw> the velocity V 2 at the mouth of the nozzle is calculated by equation (2), or F, = 223.7 \/Eaw = 223-. 7 V307 = 3919 feet per second. In order to determine the ratio of the area at the mouth of the nozzle (A 2 ) to that at the smallest section (^4 ) by equation (6) the velocity (F ) and the quality (# ) * must be determined. These evaluations are most easily made in the same way as for F 2 and x 2 by means of the entropy-total heat chart. Now the available energy E ao corresponding to the velocity F , must be calculated for adiabatic expansion from P l = 211.5 pounds and x l = .978 to P = .58 Pit = 122.7 pounds. This available energy is 44 B.T.U. and XQ is .939. The velocity Fo is, therefore, 223.7 VE^ 223.7 \/44 = 1483 feet per second. J * For steam initially dry and saturated, the quality after adiabatic expansion (x 2 ) for all practical cases is very nearly expressed, empirically, by the equation and the quality at the throat (# ) may be taken as .965 for all practical cases regard- less of the initial and final pressures. f It is well established by thermodynamic calculations and by actual experi- ment that the pressure P at the 'smallest section of a nozzle is always very nearly .58 of the initial pressure (PJ. J Very elaborate curves of the velocities resulting from the adiabatic expansion of dry saturated steam have been prepared and published in some American books. Considering the several stages in nearly all types of turbines, such curves can be of very little use to practical men, because the condition that the steam admitted to the nozzles is dry and saturated occurs infrequently. That some of the authors neglected to mark the curves "for steam initially dry and saturated" deserves severe criticism. The curves, as given, are very misleading, as they are appar- ently intended for general application for all qualities and superheats. 40 THE STEAM TURBINE In addition to the values already obtained it is only necessary to get V Q and v^ (the specific volumes of dry saturated steam at the corresponding pressures P and P t ) to determine all the terms in the equation for the expansion ratio as already given, and putting now the subscript 2 for x in equation (6) to express the conditions corresponding to the pressure P 2 ; then or x -r X '-** X .01500 = 1.030 square inches (area 3919 3.042 .939 at mouth). The author has found as the result of some investigations regarding the design of nozzles that the expansion ratio ( -)of a \AO/ properly designed nozzle is very nearly proportional to the ratio of the initial pressure (P x ) to the final pressure (P 2 ). The curve shown in Fig. 21 has been calculated on this basis for widely different conditions but for rather small expansions, and has been found to be accurate enough for practical purposes in designing turbines of more than one stage. A similar curve is now being used by the nozzle designers of one of the large manufacturing companies. After the relations shown by Fig. 21 had been worked out, it was found that Zeuner had arrived at a similar result mathematically after making certain assumptions;* but * In Zeuner's Theorie der Turbinen, page 270, the following equation is given to express the ratio of the area at the mouth to that at the smallest section (expan- sion ratio): A^ = .1550 A, ' r -fir where the terms A and P are used as in the equations above. There is probably some error in Zeuner's assumptions, because actually values of are not quite *o P., constant for varying values of ~ ' NOZZLE DESIGN Zeuner's equation itself is practically useless on account of being too complicated.* Shapes of Expanding Nozzles. The inside walls of the expand- ing portion of the nozzle are usually surfaces with straight-line 28 26 24 22 20 & lie 314 1 Sis ^3 10 >-H 3 8 4 2 / / / / / / / / / / / / / / / / / t / / / / / / i / For Steam Initially Dry Saturated or Wet / / / / / * P.l=. 58 fPi \ \ 12345 Expansion of Nozzle 1-35* FIG. 21. Curve Showing the Approximate Relation between Expansion Ratio of a Nozzle and the Ratio of the Initial to the Final Pressure. elements, meaning that in any section of the nozzle along the axis, like Fig. 18, the inner walls are shown by straight lines. * The author's curve in Fig. 21 is expressed by, A more accurate form for pressure ratios greater than 25 is the following: 42 THE STEAM TURBINE Parenty has shown that, for the highest efficiency, theoret- ically, such a section along the axis should be slightly elliptical with the focus in the throat, but practically this shape shows no advantage and is much too difficult to construct. For making nozzles like those in Curtis turbines, where the work is done largely with hand tools, the construction of even the simplest form is very expensive and the cost of an elliptical curvature is practically prohibitive. The shape to give the best expansion curve has been the subject of investigation by various experi- menters.* As the practical results are particularly interesting, it may be well to describe briefly a typical form of apparatus usually employed in these experiments as shown in Fig. 22. The FIG. 22. Searching Tube Apparatus for Determining the Pressures in Nozzles. nozzle to be tested is marked A in the figure. The steam entering the passage B discharges through the nozzle directly into the exhaust pipe E. A small "searching" tube, C, 4s provided which is sealed at one end and has a very small hole, D, a short distance from this end. The other end of the tube is attached to a mercury column or pressure gauge. Suitable means are provided for sliding the " searching " tube with its pressure gauge back and * The conditions of pressure and velocity of steam inside a nozzle are discussed 'very completely from the mathematician's viewpoint in Die Dampfturbinen by Stodola, 3rd edition, pages 42 to 75, and in Zeitschrift fur das Gesamte Turbinen- ivesen, Aug. 10, 1906, pages 325-327. NOZZLE DESIGN 43 180 170 100 150 140 J 130 I 120 ~110 "100 5 90 380 2 70 i GO 50 40 30 20 10 Expansion from 165 Lbs. per q. In. Abs. forth so that pressures can be observed in different parts of the nozzle, corresponding to the position of the hole D. From these observations a curve of pressures may be made, and from this, together with the data of the weight of steam passing per unit of time, a second curve may be developed showing the corresponding velocities. The curves in Figs. 23, 24, and 25 are examples of the results obtained by this method for three very dif- ferent nozzles. The nozzle shown at the top of Fig. 23 has curved lines, nearly elliptical, for its inside walls. A pressure curve is shown beneath the sec- tion drawing, in it:, true relative position corre- sponding to points along the axis of the nozzle. Theoretically this shape of nozzle approaches the ideal for an adiabatic expansion.* Practical conditions, however, as stated * A nozzle with a circular section (perpendicular to the axis) has less surface exposed to the flow of steam than a nozzle of any other form of the same length and expansion. For this reason this form should give minimum friction losses. In practice, however, this type is not often used when the section at the mouth of the nozzle is made rectangular, at least when the nozzles are arranged in groups with the mouths of the several nozzles close together. There are obvious advan- tages from this last construction, as first pointed out by Professor Riedler, because if the nozzle mouths are made rectangular and close together a long continuous "band" of steam is secured which is approximately homogeneous and of constant velocity. The flow from the end nozzles is, of course, affected by excessive eddying and other irregularities just as. single nozzles. Efficiency of the end nozzles is therefore considerably less than that of any of the others in the group. 2 Lbs. per S|q. In. Abs. (26 Ins, \facuum) Distance along Axis of Nozzle FIG. 23. Expansion Curve of a Nozzle with an Elliptical Axial Section. 44 THE STEAM TURBINE 1*0 130 before, make the nozzle shown in Fig. 24 with expanding straight- line walls preferable if the throat and mouth areas are properly designed. Fig. 24, however, is intended to show primarily the effect of using a nozzle for non-condensing service, which was designed to be used condensing. For this reason the expansion in the nozzle is greater than it should be for the pressures with which it is operating; and for this reason the pressure inside the nozzle, as illustrated by the curve, falls below the exhaust pressure. This is called over-expansion or " over-compounding " and is always accompa- nied fa$ a loss in efficiency. In feet, as will be shown again later, the effect of over-expansion, or making a nozzle too large at the mouth, reduces nozzle effi- ciency much more than if it is made the same per- centage too small. (See Fig. 28.) The curves > Distance alang Axis of Nozzle in Fig. 24 show that the FIG. 24. Expansion Curve of a Nozzle with p res sure at the mouth Straight Walls. . g a ^ lower than the atmospheric exhaust, and a partial vacuum is thus secured at the blades opposite the nozzles. When such nozzles are operated non-condensing there is some gain from the reduc- tion of disk and blade friction because the wheel and blades revolve in a less dense medium; but when considering also the increased losses in the nozzle itself because of over- expansion, there is certainly no net gain over having a nozzle g 60 30 \. Atmospheric Line NOZZLE DESIGN 45 designed exactly for the expansion corresponding to the operat- ing conditions. Fig. 25 is intended to show an abnormal but interesting form of nozzle which gives some idea of the behavior of steam when the expansion is not gradual and continuous. It was argued by a designer who made this nozzle that this form should be as efficient as any other. It was re$$j$S$S^-V S 1 -* -^ x his theory that if the areas at the throat and at the .mouth were of the right size, the shape of the walls between was of no consequence, and, in fact, that the steam of itself would take the correct passage. Thus by preventing the steam particles from touching the walls the friction losses in the nozzle should be reduced. It will be observed, how- ^ _^ ^^ee along Axi8 of Nozz)e ever, from the curve FIG. 25. Expansion Curve in an Abnormal Nozzle. in the figure deter- mined from some experiments with this nozzle that the pres- sure first drops abruptly in the throat to .58 of the initial, as in any other nozzle, and then forms a series of waves, from which it appears that the particles of steam strike the walls and rebound, to meet again at a point, as at A, where an increased pressure is produced, and so on till the mouth is reached. The probable path of the steam is shown by the dotted lines in the drawing of the nozzle. These experiments show therefore that the steam will not take the correct passage through a nozzle with- out the provision of properly designed walls of gradually increasing 180 1GO 140 1 100 lao 40 46 THE STEAM TURBINE area corresponding to the expansion required. The importance of careful workmanship in the manufacture of nozzles is therefore obvious. The results shown by Fig. 25 bring up naturally the discussion of the proper length for a nozzle, as the one in this figure was obviously much too long. Probably the best designers of the Curtis types of turbines make the length of the nozzle depend only on the initial pressure. In other words, the length of a nozzle for 150 pounds per square inch initial pressure is usually made the same for a given type regard- less of the final pressure. And if it happens that there is crowding for space, one or more of the nozzles is sometimes made a little shorter than the others. Designers of De Laval nozzles follow practically the same " elastic " method. The divergence of the walls of non-con- densing nozzles is about 3 degrees from the axis of the nozzle, and condensing nozzles for high vacuums may have a divergence of as much as 6 degrees * for the normal rated pressures of the turbine. The author has used successfully the following empirical formula to determine a suitable length, L, of the nozzle between the throat and the mouth (in inches): L = Vis A , (7) where A is the area at the throat in square inches. The design of the nozzle calculated in the example on page 36 can now be completed with the determination of its proper length, L = Vi$ X .059 = .9 inch. The important dimensions of nozzles of circular section suitable * According to Dr. O. Recke, if the total divergence of a nozzle is more than 6 degrees, eddies will begin to form in the jet. There is no doubt that a too rapid divergence produces a velocity loss. When a number of nozzles intended for different initial pressures are supplied for use in the same turbine, the length as determined by the taper is usually made to correspond to the pressure that is to be most used. Inspection of the De Laval nozzle in Fig. 14 shows that it is necessary to make all the nozzles of the same length for a given size of De Laval turbine, so that the nozzles maybe used interchangeably. NOZZLE DESIGN 47 for this De Laval turbine tested by Dean & Main may be tabu- lated as follows: Area at throat (A ), .0590 square inch. Diameter (D ), .274 inch. Area at mouth (A 2 ), 1.008 square inches. Diameter (D 2 ), 1.132 inches. Length of nozzle (L) as determined by equation (7), .9 inch. Length of nozzle assuming a divergence of 12 degrees, 1.9 inches. It will be observed from the last calculation that a de- signer of De Lavai nozzles would make the length about twice that calculated by equation (7). The nozzles of De Laval turbines are made unusually long largely for mechanical rea- sons. There is probably very little loss in this additional length. A nozzle of circular section suitable for these conditions is shown at the top of page 27 (Fig. 13). It will be observed that a rounded entrance to the nozzle has been made, If a well- rounded entrance is not provided the rate of flow through the nozzle may be only 50 to 70 per cent, (depending of course on the sharpness of the corners) of the normal flow calculated from Grashof's formulas given in equations (3) and (4). The effi- ciency is also very much reduced if the steam is not led to the throat along a surface of gradual curvature.* * Jude states that a very large rounded inlet appears to "choke" the nozzle a little. He admits that it gives- maximum discharge "but at the expense of kinetic energy, that is, of the kinetic energy effective in an axial direction." The results of Rateau's experiments seem to show, however, that the efficiency of a convergent nozzle suitably rounded is unity. If any loss does result from a rounded entrance which is too large it is probably of negligible amount. . Some conclusions drawn from Rosenhain's experiments reported in Proc. Inst. Civil Engineers, vol. 140, may be of interest in this connection. A series of experiments was made with vari- ous nozzles working from 20 pounds to 200 pounds per square inch gauge pressure with atmospheric exhaust. The most efficient form of nozzle up to about 80 pounds gauge pressure appears to be a plain orifice in a thin plate, as measured by nozzle reaction (see page 58), but this does not imply that such a form is the best nozzle for a turbine under similar conditions. With this kind of orifice there is too much spreading of the jet, and the internal eddies and whirls are too violent for useful application at the point where the jet strikes the turbine vanes. 4 8 THE STEAM TURBINE It has been shown by Stodola's experiments that the difference in pressure between the outer and inner portions of the jet inside a nozzle of approximately correct design are practically negligible. The conclusion is, therefore, that the jet always completely fills the nozzle, and that there is no " zonal formation," meaning an outer zone moving at a different velocity from the inner one, although there is certainly a considerable amount of., fric- tional dragging of the steam at the surface. Obviously, of course, the statement does not hold for absurdly diverging forms of nozzles, and in such cases the steam leaves the walls with apparently much loss of velocity as in the ex- mple shown by Fig. 25. Stodola observed also that any nozzle the pressure r 60 in 12345 Inches from the end of the Nozzle usually falls in the vicinity of the throat to consider- ably less than the discharge F, G . 2 6_ Experiments with an Expand- with a sudden rise ing Nozzle Showing the Effect of Vary- r ' ing the Final Pressure. immediately after the fall. This effect is shown by pressure curves in Fig. 26, plotted from Stodola's data taken in a divergent nozzle like the one represented at the top of this figure. Similar effects, only more pronounced, observed in a straight non-expanding nozzle with rounded inlet are shown in Fig. 27. Here also a sudden drop below the discharge occurs; and, peculiarly, the point of depression progresses along the axis of the nozzle as the pressure decreases. Very pro- nounced oscillations are set up which extend even into the exhaust space for a distance of about one and a half times the length of the nozzle. The oscillations are apparently most violent for the middle range of pressure, and tend toward a NOZZLE DESIGN 49 1 2 Inches from Mouth minimum when the lower pressure approaches a perfect vacuum. In the divergent nozzle, however, there appear to be no internal oscillations of pressure after those at the throat have died out. The size and most likely also the shape of the external space has a considerable effect on these oscillations of pressure. Jude states in this connection that there is a greater loss in velocity, due to oscillations or eddies, in a square or rectangular nozzle than in a circular one. Recent experience with nozzles of this type does not bear out this statement, except in the case probably of square or rectangular nozzles with no rounding at the edges. An efficiency of 97 per cent, is not unusual for prop- erly designed square and rectangular FIG. 27. Experiments with shaped nozzles without any " square " a Non-expanding Nozzle ... . . Showing the Effect of edges; and circular nozzles have certainly Varying the Final Pressure never given 99 per cent, efficiency. Under- and Over-Expansion. The best efficiency of a nozzle is obtained when the expansion required is that for which the nozzle was designed, or when the expansion ratio for the condition of the steam corresponds with the ratio of the areas of the mouth and throat of the nozzle. A little under-expansion is far better, how- ever, than the same amount of over-expansion, meaning that a nozzle that is too small for the required expansion is more efficient than one that is correspondingly too large.* Fig. 28 shows a * It is a very good method, and one often adopted, to design nozzles so that at the rated capacity the nozzles under-expand at least 10 per cent., and maybe 20 per cent. The loss for these conditions is insignificant, and the nozzles can be run for a large overload (with increased pressures) in nearly all types without immediately reducing the efficiency very much. This applies especially to tur- bines governed by cutting out nozzles in the first stage (see page 221) and with no control of the nozzles in the other stages. Under-expansion due to a throttling governor is also an important condition affecting the efficiency of nozzles. 50 THE STEAM TURBINE % curve representing average values of nozzle loss used by various American and European manufacturers* to determine discharge velocities from nozzles under the conditions of under- or over- expansion. This curve will be referred to again in connection with the design of blades and is very useful to the practical designer. Non-expanding Nozzles. All the nozzles of Rateau turbines and usually also those of the low-pressure stages of Curtis turbines Nozzle Velocity Loss Due to Under-or Over-Expansion 10 t* 05 00 o' / / / / / / / \ /* "^ --*. s* / ~-~. "- ^, -. 1 ^ *^ " 'Percentage Nozzle is too Small A$Jttouth (Under Expansion) Percentage Nozzle is too Large at Mouth (Over Expausion) FIG. 28 Curve of Nozzle Velocity Loss. are made non-expanding; meaning, that they have the same area at the throat as at the mouth. For such conditions it has been suggested that instead of a series of separate nozzles in a row a single long nozzle might be used of which the sides were arcs of circles corresponding to the inside and outside pitch diameters of the blades. Advantages would be secured both on account of cheapness of construction and because a large amount of friction against the sides of nozzles would be eliminated by omitting a number of nozzle walls. Such a construction has not proved desirable, because by this method no well-formed jets are secured and the loss from eddies is excessive. The general statement may be made that the, throat of a well-designed nozzle should have a nearly symmetrical shape, as for example a circle, a square, etc., rather than such shapes as ellipses and long rectangles. The * C. P. Steinmetz, Proc. Am. .Soc. Mech. Engineers, May, 1908, page 628. A. Jude, The Theory of the Steam Turbine, page 39. NOZZLE DESIGN 51 shape of the mouth is not important. In Curtis turbines an approximately rectangular mouth is used because the nozzles are placed close together (usually in a nozzle plate like Fig. 114) in order to produce a continuous band of steam; and, of course, by using a section that is rectangular rather than circular or elliptical, a band of steam of more nearly uniform velocity and density is secured. Fig. 29 shows a number of designs of non-expanding nozzles used by Professor Rateau. The length of such nozzles beyond FIG. 29. Rateau Non-expanding Nozzles. the throat is practically negligible. Curtis non-expanding nozzles are usually made the same length as if expanding and the length is determined by the throat area. The Curtis nozzles made in Germany are a little shorter than the length calculated by formula (7). Materials for Nozzles. Nozzles for saturated or slightly super- heated steam are usually made of bronze. Gun metal, zinc alloys, and delta metal are also frequently used. All these metals, have unusual resistance for erosion or corrosion from the use of wet steam. Because of this property as well as for the reason that they are easily worked with hand tools * they are very suitable materials for the manufacture of steam turbine nozzles. Super- heated steam, however, rapidly erodes all these alloys and also greatly reduces the tensile strength. For nozzles to be used with highly superheated steam, cast iron is generally used, and except * Nozzles of irregular shapes are usually filed by hand to the exact size. 52 THE STEAM TURBINE that it corrodes so readily is a very satisfactory material. Com- mercial copper (about 98 per cent.) is said to have been used with a fair degree of success with high superheats; but for such con- ditions its tensile strength is very low. Steel and cupro-nickel (8 Cu + 2 Ni) are also suitable materials, and the latter has the advantage of being practically non-corrodible. SUPERHEATED STEAM. In the following pages the important properties of superheated steam with which the modern engineer must deal will be briefly discussed. It is generally recognized that a gain in steam economy results from the use of superheated steam in either steam turbines or reciprocating engines, but an accurate analysis of tests for the actual gain in economy of a plant is very difficult because there are so many factors entering. The peculiar circumstance, also, that water can exist indefinitely in the liquid state in the presence of superheated steam, makes conclusions from experimental data often uncertain. Flow of Superheated Steam through Nozzles. The discharge of superheated steam from a nozzle is one of the most important subjects of which the engineering profession generally has no correct data. The author has observed in his practice again and again that the formulas ordinarily given for the flow of super- heated steam were not correct and more reliable data had to be found. The formulas given here were actually determined from the data of Lewicki's experiments with a 3o-horsepower De Laval turbine * but was later checked with a great mass of data in the possession of the General Electric Company. The precision with which the formula applies to Lewicki's data is shown in the table given on the next page. A formula was desired to express the flow of superheated steam discharged from a nozzle in the form of formula (3) for the flow of dry saturated steam, together with a suitable coefficient to * T^eit. Verein deutscher Ingenieure, April 4, 1903, page 494. Mitteilungen uber Forschungsarbeiten, Heft 12 (1904), Zalentafel 25. NOZZLE DESIGN 53 correct for the effect of superheat. A formula of this form is expressed by A P 97 (8) F = 60 (i 4- .00065 D) 60 F (i + .000650) or (80 where F is the weight in pounds of superheated steam discharged per second, A is the area of the smallest section of the nozzle in square inches, P t is the initial pressure in pounds per square inch absolute, and D is the superheat in degrees Fahrenheit.* Lewicki's data for the tests f given below were in metric units but are recorded here in the corresponding English units. Initial pressure P t = 99.25 pounds per square inch absolute. Final pressure P 2 = 14.6 pounds per square inch absolute. Number of Test. 2 463.1 I36.I 837.0 1.089 910.0 3 491.9 164.9 824.0 I.I07 9II.O 4 5 6 : 7 8 Temperature of steam, degrees F. Superheat, degrees F. (D) Flow, pounds per hour (tests) . . (i + . 00065 D) 386.6 59-6 882.0 1.038 917.0 529.7 2O2. 7 804.0 I.I32 910.0 59 2 -7 265.7 778.0 I-I73 914.0 619.7 292.7 776.0 1. 190 910.0 73-4 37 6 -4 735- 1.245 914.0 723.6 396.6 729.0 1.258 916.0 Flow, pounds per hour, corrected by formula (8) to equivalent flow of dry saturated steam . . Volume of Superheated Steam. Thermodynamic relations show that the flow of superheated steam is inversely proportional to the square root of the specific volume, J so that from the author's equation for the flow of superheated steam (8) the fol- lowing formula for the specific volume is easily obtained : = (i + .00065 (9) * It is stated that Mr. A. R. Dodge has shown practically the same results from the Newport tests of a Curtis turbine reported by Mr. G. H. Barrus. t Mitteilungen uber Forschungsarbeiten, Lewicki, Heft 12 (1904), Zalentafel 25. J This relation is discussed by the author in Mechanical Engineer (London), Aug. 24, 1907, page 277, and in the Harvard Engineering Journal, June, 1907, page 36.. Compare with Stodola, Die Dampfturbinen, 3rd ed., page 9. 54 THE STEAM TURBINE where v s is the volume in cubic feet of one pound of superheated steam, v is the volume in cubic feet of one pound of dry saturated steam at the same pressure as for v. s , and D is the superheat in Fahrenheit degrees.* Available Energy of Superheated Steam. In the following paragraphs the significance of the entropy-temperature curves for superheated steam will be explained, and it will be shown also how they are to be used to determine the available energy and the corresponding theoretical velocity resulting from adiabatic expan- sion in a nozzle. Specific Heat of Superheated Steam. In modern practice, superheated steam often enters our calculations and a troublesome modification of the entropy diagram results. The difficulty arises because the specific heat of superheated steam is not at all accurately known. The diagrams in the appendix are calculated for the specific heat determinations by Knoblauch and Jakob, f The specific heat of steam varies with the temperature and pres- sure as shown in Figs. 30 and 31, giving values of the mean and the true specific heat at constant pressure (C p ). True specific heat represents the ratio of the amount of heat to be added to a given weight of steam at some particular condition Of temperature and pressure to raise the temperature one degree to that required to raise the temperature of water at maximum density one degree. The mean specific heat is almost invariably used in steam turbine calculations. Entropy Diagram of Superheated Steam. The graphic repre- sentation of the heat added during the superheating of steam is easily accomplished with entropy-temperature diagrams. Fig. 32 shows the same diagram that represented dry saturated steam in Fig. 12, with the added area EHJF to show the super- heating from the temperature, T t corresponding to the pressure * This formula gives values of specific volume representing a fair average of results obtained from the formulas of Zeuner, Tumlirz, Knoblauch, and Schmidt (based upon Hirn's experiments). f Zeit. Verein deutscher Ingenieure, Jan. 5, 1907. Values of mean specific heat are taken from Mechanical Engineer, July, 1907, and Professor A.M. Greene's paper in Proc. American Society of Mechanical Engineers, May, 1907. NOZZLE DESIGN 55 Pj to the temperature of the superheated steam, T. s . The total heat in a pound of steam above the freezing point is now repre- sented by the area OBCEHJFO. For adiabatic expansion of superheated steam at the temperature T s and pressure P l to a pressure P 2 the available energy is the area CEHKL. Too much calculation is involved in the construction of entropy diagrams to make a new diagram for every particular case from .06 .42 200250300350400450500550 650 700 750 Temperature F FIG. 30. Mean Values of C p Calculated by Integration from Knoblauch and Jakob's Data. the properties usually found in steam tables; but the construction of such diagrams should be understood. From the explanations that have preceded, the construction of all the lines except EH should be obvious. This line is obtained by calculating the entropy of superheated steam for various values of temperature from the following well-known relations: or log. ? = 2.3028 CpJlogio T a - logio Ti\ 5 6 THE STEAM TURBINE where C pm is the mean value taken from the curves in Fig. 30 for the temperature TV General Remarks Regarding Nozzles. Finally it may be stated that there is practically no difference in the efficiency of the nozzles 0.90 0.85 0.45 100 J 200 250 300 Temperature C. 400 FIG. 31. Values of the " True " Specific Heat of Superheated Steam. used in commercial turbines if they have smooth surfaces and are properly designed for the correct ratio of the area at the throat to that at the mouth, and if the length is not made much less than, nor more than possibly twice that calculated by formula (7). NOZZLE DESIGN 57 Whether the nozzle section is throughout circular, square, 01 rectangular (if these last sections have rounded corners) the efficiency as measured by the velocity will be about 96 to 97 per cent., corresponding to an equivalent energy efficiency of 92 to 94 f> s Entropy FIG. 32. Entropy-Temperature Diagram for Superheated Steam. per cent. Speaking commercially, therefore, it does not seem to be w,orth while to spend a great deal of time in the shops to make nozzles very exactly to some difficult shape. Simpler and more rapid methods of nozzle construction should be introduced. In some shops the time of one man for two days is required for the hand labor alone on a single nozzle. Similarly to the equations for available energy on page 23 for dry saturated steam and on page 26 for wet steam, the available energy E as for superheated steam is when the final con- dition is "wet": When at the final condition it is superheated, then . = n l - H 2 + c^ (T, - ro - c_ (T; - r,), where other symbols are used as before and T 8 is initial tem- perature and T/ is the final temperature when superheated. CHAPTER IV. STEAM TURBINE TYPES AND BLADE DESIGN. ALL the types of both water and steam turbines are commonly ^divided into two general classes, designated by the descriptive terms impulse and reaction. Without further explanation, these terms, as they are used in turbine practice, would be very mislead- ing, because practically all commercial types of steam turbines operate by both the impulse and the reaction of steam. Long usage, however, has determined the accepted meaning of these terms and it is useless now to try to change them. Briefly, the Reaction. Force . Impulse Force FIG. 33. Impulse of a Jet Exerted on a Flat Surface. physical phenomena known as impulse and reaction will first be described, to be followed by an explanation of the technical sig- nificance of these terms as they are used by engineers. In all important commercial types of steam turbines the blades 58 STEAM TURBINE TYPES AND BLADE DESIGN 59 are moved by both the impulse and the reaction of impinging steam jets issuing from nozzles (see Fig. 2) or passages essentially equivalent to nozzles. According to the older school of scientists, who have handed down to us the classification of turbines men- tioned above, an impulse is a force acting in a " forward " direc- tion, and a reaction is a " backward " force, relative to the impulse and equal to it in magnitude. Fig. 33 is a simple con- crete illustration of both impulse and reaction. A suspended tank filled with water is shown from which a jet issues through a nozzle and impinges upon a flat board hung opposite. As the Reaction Force Impulse and Reaction Forces FIG. 34. Impulse of a Jet Exerted on a Curved Surface. result of the pressure due to the jet, the board will obviously move to the right. As the jet issues from the nozzle it exerts at the same time a reaction on the tank causing it to move to the left.* * The pressure on the walls of a tank at any point depends on the height of the water above that point ("the head") and upon the density of the fluid. When a fluid escapes from an opening in the tank there is no resistance at that point to pressure, and the unbalanced force exerted on the walls directly opposite will tend to move the tank in the direction opposite to that of the escaping jet. The greater the "head" and the density the greater will be the velocity of the issuing fluid and the reaction on the tank, 6o THE STEAM TURBINE 34 is intended to show the significance of impulse and reaction as they are used in regard to turbines. In this case water from the tank impinges against the curved surface of a wooden block, and before it leaves this surface it is turned back upon itself through an angle of 180 degrees. The block is there- fore acted on by two forces simultaneously, both tending to move it to the right. When the jet first strikes the surface of the block FIG. 35. Impulse Wheel with Blades of " Single Curvature." FIG. 36. Impulse Wheel with Blades of "Double Curvature." an impulse force tends to move.it, and when leaving, there is acting in a " backward " direction a reaction equal to the impulse. If the jets represented in the two figures have the same velocity and density, and frictional losses are neglected, the pressure on the block in Fig. 34 will be twice as great as on the board in Fig. 33. Fig. 35 shows a nozzle and a blade wheel in which the blades have a " single curvature " as compared with the curved surface in Fig. 34; that is, the steam in its passage through the blades is not " turned back on itself," or in other words, the curvature of the blades is less than 90 degrees. If this wheel were held STEAM TURBINE TYPES AND -BLADE DESIGN 6 1 stationary so that the blades could not move, the steam would leave them in a direction nearly parallel to the shaft. The only force, therefore, that is effective for moving the blades is the impulse. Fig. 36, on the other hand, shows blades with nearly 180 degrees curvature which turn the steam back on itself on leaving. The wheel is thus moved first by the impulse force of the steam exerted on the blades in the'direction of flow, and then by its reaction. A blade turning the steam through less than 90 degrees like the one in Fig. 35 will exert only about half as much pressure as one turn- ing the steam through nearly 1 80 degrees like the one in Fig. 36. A turbine wheel which would be called a reaction type is shown in Fig. 37. It differs from the one in Fig. 36 chiefly in the blade section B, shown at the top of the draw- ing. In this type the expan- sion of the steam in the noz- zle is only partial, and the blades are made so that part of the expansion occurs in them. In the types shown in Figs. 35 and 36, on the other hand, all the expansion is in the nozzles, with no ex- pansion at all in the blades.* FIG. 37. Simple Reaction Wheel. The amount of expansion of the steam in the blades marks, therefore, the essential difference between the two important types of steam turbines illustrated by Figs. 36 and 37. In im- pulse turbines there is no expansion in the blades, while in * The turbine wheel illustrated in Fig. 37 is not, however, typical of commercial " reaction " types in which there are often a hundred or more pressure stages. 62 THE STEAM TURBINE reaction turbines " expanding " blades are used, with the result that some of the kinetic energy of the steam is changed to velocity in flowing through them. From the explanation that has preceded it is obvious that both of the types represented by the last two figures operate by both impulse and reaction. Impulse and Reaction of Fluids. The kinetic energy of a fluid jet discharging from a nozzle may be regarded as produced by a constant impulse force I acting upon a weight W of the fluid discharged for one second. During this second the velocity has changed from zero to V feet per second and has gone through a space of V feet. The work done by this force in producing y the kinetic energy (K foot-pounds per second) is I X -, which WV 2 is equal to K or IV WV 2 We have then = > 2 2g T* g '* In practice the principal distinguishing feature of reaction turbines is the applica- tion of stationary blades for partially expanding the steam. The rest of the ex- pansion takes place in the moving blades. It is sometimes stated, although inaccurately, that the angles of the moving blades may be used as a criterion for distinguishing the two types. According to these authorities, the moving blades of impulse turbines are symmetrical like Fig. 36, and those of reaction turbines resemble in contour those of Fig. 37. In many cases the rule could probably be applied, but there are also many exceptions. There are some blades made for Curtis turbines which are not nearly symmetrical, and no one would call a Curtis turbine a reaction type. The difference between impulse and reaction turbines can be very easily shown experimentally by putting a pressure gauge between the nozzle and the wheel. In the impulse type, because the expansion is completed in the nozzle, it will be found there is no drop in the pressure of the steam in passing through the blades; but in the reaction type the gauge will record a higher pressure than that in the casing. As these words " impulse " and " reaction " are used at the present time there is really little connection between the usual meaning of the words and the ideas they are to convey in regard to steam turbines. Actually all commercial steam turbines work by impulse and by reaction. A German writer has used instead of " impulse " and " reaction " the more accurate words, " gleichdruck " and " ungleichdruck," meaning " equal pressure " and " unequ al pressure," which to the author seem much more appropriate. STEAM TURBINE TYPES AND BLADE DESIGN 63 In the first principles of physics it was learned that impulse and reaction were " equal and Opposite," so that if the reaction WV is represented by R in pounds, then R = I = o Example. If the vessel shown in Fig. 33 discharges 10 pounds of water per second at a velocity of 322 feet per second, what is the force I (impulse) pushing the wooden block away from the vessel? Ans. 100 pounds.* Also what is the force R (reaction) pushing the vessel itself toward the left? Ans. 100 pounds. Example. If water is discharged against flat blades of a water wheel made up of vanes similar to the block shown in Fig. 33 at the rate of 32.2 pounds per second at a velocity of 200 feet per second and is spattered from the wooden blocks with a " residual " velocity (leaving the vanes) of 100 feet per second, what horsepower is this water wheel capable of devel- oping? Solution. Calling the " residual " velocity Vz we have W (V 2 - F 2 2 ) 32.2(200 2 -I00 2 ) K = - -^ = ^- L = 15,000 ft.-lbs. per sec., 2g 2X32.2 15,000 or =27.27 horsepower. 55 WV 2 The maximum theoretical horsepower of the wheel is - - > 2 g X 550 if the water is discharged at zero velocity. We have (in this x 32.2 X 2oo 2 20,000 case) ^ = - , or 36.36 h.p. ' 2X32.2X550 550 The efficiency of the (blades of this) water wheel is therefore 27.27 T* J = .75 or 75 per cent. 3 6 -3 6 Example. Steam discharges from a nozzle at the rate of 3.542 pounds per second with a velocity of 4000 feet per second against the blades of a steam turbine and leaves them with a * It is assumed that the water leaves the block with practically no velocity, that is, all the velocity is absorbed in producing the impulse force. THE STEAM TURBINE velocity of 1000 feet per second. Neglecting frictional losses, what is the maximum horsepower that this turbine wheel can develop? Calculate the efficiency (percentage) of the blades in this turbine. Ans. 1500 horsepower; 93.75 per cent. Example. The steam discharging from the blades of the turbine wheel in the last exercise is finally directed upon the blades of a second turbine wheel. Assuming there has been no loss of velocity in passing from one turbine wheel to the other and that the steam leaves the second one at 100 feet per second, calculate the maximum horsepower that could be developed in this second turbine wheel and the efficiency of its blades. Ans. 99 horsepower; 99 per cent. Example. If we consider the two turbine wheels mentioned in the two preceding exercises as combined in a single turbine, what would be the total horsepower of the turbine and the over-all efficiency if frictional and other losses are neglected? Ans. 1599 horsepower; 99.94 per cent. Suggestion. The same result could have been obtained by calculating the total kinetic energy of the combined wheels, using V = 4000 feet per second, V% = 100 feet per second and W = 3.542 pounds of steam. Example. Remembering that impulse and reaction are equal and opposite, what is the force of the reaction against the plate supporting the nozzle required to give a velocity of 4000 feet per second to a flow of 3.542 pounds of steam per second? Ans. 440 pounds. Suggestion. Reaction = impulse (/) = - o Example. The area of a nozzle is .322 square inch. How many pounds of steam per second having a density of .144 pound per cubic foot must be discharged from the nozzle in order to exert a pressure of 90 pounds against a plate suitably designed to turn away the steam with zero velocity? Ans. .966 pound. Suggestion. In this case all the velocity is absorbed in produc- ing the pressure (impulse) upon the plate. STEAM TURBINE TYPES AND BLADE DESIGN 6$b Substituting the values given in the example and substituting in the equation for impulse, we' have ,,, .322 X V X .144 W = 3 * = .000322 F, 144 T WV .000322 F 2 T72 / = - ^ = .ooooi F 2 . 3 2 - 2 and since the impulse is 90 pounds; we have .00001 F 2 = 90 pounds F 2 = 9,000,000 F = 3000 feet per second. Substituting this value of V in the equation at the top of the page, W = .000322 X 3000 = .966 pound per second. Example. Steam of the same density as in the preceding exercise discharges at the rate of 3478 pounds per hour and pro- duces a reaction against the plate into which the nozzle is in- serted of 90 pounds. What is the velocity of discharge? Ans. 3000 feet per second. EXAMPLES OF IMPULSE TURBINES. A simple impulse turbine is represented by diagrammatic draw- ings in Fig. 38. In the shaded drawings in this figure, " Section A " is made by a plane cutting one of the blades and passing through the center of the shaft. The other view, " Section B," shows a section made by a plane parallel to the shaft and passing through the center of one of the nozzles in the turbine. In the same figure, Curve I shows the decreasing pressures in the nozzle and the constant pressure through the blades. Curve II shows similarly the velocity changes. In the nozzle the steam velocity increases as the pressure falls, while in -the blades the velocity of the steam is absorbed in moving the wheel. This simple impulse turbine represented by these diagrams is typical of the De Laval type. These turbines have always a single set of nozzles and one row of blades. Velocity Triangles FIG. 38. Diagrams of a Single- stage Impulse Turbine. SECTION A t 1^1 t ^"SECTION B (6 3 c) Velocity Triangles FIG. 39. Diagrams of an Impulse Turbine with Two Velocity Stages. STEAM TURBINE TYPES AND BLADE DESIGN 63d In Figs. 38, 39, 40, and 41 illustrating the important types of steam turbines, the direction of the flow of the steam is marked by the symbol c/3 and the motion of the blades by w >. The moving blades are shown by solid black to distinguish them from the stationary blades, which are indicated by cross-hatching. A modification of the simple impulse type is shown in Fig. 39. The drawings marked "Section A" and "Section B" show a turbine with two moving blade wheels and a set of stationary " intermediate " blades. The stationary blades are merely guides for changing the direction of the steam so that it will enter the second set of moving blades at a suitable angle. Two blade wheels are used instead of one in order to make it possible to use efficiently a lower peripheral speed for the moving blades. The reasons for this statement will be discussed in another part of this chapter. The curves at the top of the figure show, graph- ically, the relation between pressure and velocity. Curve III shows the sudden fall of pressure in the nozzle and the constant pressure through the three rows of blades. Curve IV shows first the rapid increase in velocity as the pressure falls, and then the gradual loss of velocity in the moving blades as it is given up in doing work. Velocities represented in Curves II and IV are drawn approximately to the same scale. A comparison shows that the reduction in velocity of the steam in the first wheel as represented in Curve IV is only about half that for the single wheel in Curve II. The arrangement of blades represented in Fig. 39 makes possible comparatively low blade speeds with initially high steam velocities. This method of increasing the number of rows of blades is often used with three rows of moving blades and two " intermediate " (stationary) rows; and even four rows of moving blades have been used. Not much advantage, however, has been shown from the use of the third and fourth rows of moving blades, and this construction has been generally abandoned. Turbines of this type are often spoken of as having velocity stages, the number of velocity stages being the same as the number of rows of moving blades. The Curtis turbines, made by the General Electric Company, 64 THE STEAM TURBINE are the best examples of the type illustrated by Fig. 39 with several rows of blades following a set of nozzles. In the latest designs of the larger sizes of these turbines there are two rows of moving blades and one set of " intermediate " blades for each set of nozzles, so that the arrangement shown in Fig. 39 is typical of these designs.* In Fig. 40 another distinct type of steam turbine is illustrated. The left-hand half of this figure represents a single impulse wheel as in Fig. 38 and the right-hand half is practically a duplicate of that on the left. In this construction each of the halves a single nozzle or set of nozzles with the blades following is called a pressure stage, or very commonly it is called simply a stage. The difference between the operation of this turbine and the single impulse wheel in Fig. 38 is best shown by comparing the pressure and the velocity curves at the top of the two figures. In Curve I, showing the pressure for the single impulse wheel, the steam drops from the boiler pressure to that of the exhaust in a single nozzle, that is, in a single stage. In Curve V of Fig. 40 there is about equal reduction of pressure in each of the two nozzles, and the velocity change, as Curve VI shows, is about the same for each of the two stages. This figure represents, diagrammatically, a number of types that are more complex. It should be mentioned here that there are often two or more groups of nozzles and blades, each like Fig. 39, in succession (cf. Fig. 119). Each of these groups is then called a stage. In other words, the first set of nozzles and all the rows of blades up to the next nozzle make the first stage, and so on. This last arrangement is typical of the Curtis turbines with more than one pressure stage and the various Rateau designs. *The blades shown in "Section A" of Fig. 39 have the sahie height on the "entrance" and "exit "sides. It is, however, a very common practice to make the "exit" side of the "intermediate" blades of Curtis turbines a little higher than the " entrance " side so as to increase the cross-sectional area and thus allow for the lessened velocity, due to friction and eddies, and thereby prevent " choking " in the blades. There is therefore a little expansion in these blades. STEAM TURBINE TYPES AND BLADE DESIGN Velocity Triangles FIG. 40. Diagrams of an Impulse Turbine with Two Pressure Stages. 66 THE STEAM TURBINE DrumRotor SECTION A SECTION B VaJ V Velocity Triangles FIG. 41. Diagrams of a Three-stage Reaction Turbine. STEAM TURBINE TYPES AND BLADE DESIGN 6/ The Rateau turbine has from 20 to 40 pressure stages, with a set of nozzles and a single blade wheel for each. The drop in pressure is then, of course, comparatively small in each stage. REACTION TURBINES. The arrangement of blades in the well-known Parsons turbine is illustrated in Fig. 41. This is the typical modern reaction tur- bine. There are no nozzles. The- steam flows from the boiler into the " admission space" of the turbine (see " Section A") with practically no velocity. From this space it enters the first set of stationary blades, where it expands and attains some velocity as the pressure drops. Curves VII and VIII show the change of velocity with change of pressure. When the steam leaves the fixed, blades it enters immediately the first set of moving blades. Here it expands again; but at the same time some of the velocity from the expansion is taken away, or, in other words, the velocity is reduced in moving the blade wheels. The pressure and velocity curves show plainly what happens in turbines of this type as the steam passes alternately through the fixed and moving blades, expanding in every row till it escapes in the exhaust. There is here considerable expansion in the moving blades, and conse- quently because the pressure is not the same on both sides of these blades it is called a reaction turbine. All the other three types (Figs. 38-40) are impulse turbines, because the pressure is practi- cally the same on both sides of the moving blades. We should observe here that all the possible simple combina- tions have been mentioned except the case of expansion only in the moving blades and with no expansion in the stationary parts. Such an arrangement would be feasible but has probably never been used. In a reaction turbine any two rows of blades, the first stationary and the second moving, make a pressure stage. In a Parsons reaction turbine there are sometimes more than a hundred stages. Graphical Diagrams of Steam Velocities. A velocity diagram representing graphically the steam velocities in the passages of each of four types of turbines shown in Figs. 38-41 is represented 68 THE STEAM TURBINE at the bottom of each of these figures. These diagrams, in the shape of velocity triangles, are represented here with the nozzles and blades in their proper order. In practical designing, how- ever, this pictorial effect is omitted and only the triangles are drawn. The lines of these triangles show by their lengths the magnitudes of the blade as well as the steam velocities in the tur- bine. As all of these triangles are drawn to the same scale, they show how different the velocities are in the four types. In each case the blade speed (Vb) is taken at about the value that has been found by experience to give the best efficiency. Such velocity diagrams are used by engineers for determining the best relation between the velocity of the blades and the velocity of the steam. In order to interpret such diagrams intelligently the significance of absolute and relative velocities * of the steam must be clearly * This distinction between absolute and relative velocities should probably be made plainer for those who are unfamiliar with these terms. A thorough under- standing of what is meant by absolute and relative velocities is very necessary to work intelligently with the velocity diagrams on which the whole theory of turbine practice depends. Suppose a train is just moving out of a station at the rate of 30 feet per second, and a man standing in the middle of the track behind the train throws a ball with a velocity of 40 feet per second through the back door of the last car. Then a passenger in the train - will see the ball moving through the car at a velocity of only 10 feet per second. In this case the velocity of the ball, or 40 feet per second, is its absolute velo- city with respect to bodies that are not moving, and 10 feet per second is the relative velocity of the ball in the train. In this connection a slightly different case should also be considered. Sup- pose now the ball is thrown upon a boat moving in a stream at a velocity of 30 feet per second by a man standing on the bank at P as represented in Fig. 42. Let us assume the absolute velocity, or the velocity with which the ball is thrown, as again 40 feet per second, but that now the path of the ball makes an angle of 20 with the direction of the moving boat. Then the relative velocity of the ball (F r ) with respect to the direction of the boat is shown graphically by a triangle of velocities ABC in the figure, where AC is the absolute velocity (VJ of the ball, BC is the velocity of the boat (Vb), and AB is the relative velocity (F r ) of the ball with respect to that of the boat. FIG. 42. STEAM TURBINE TYPES AND BLADE DESIGN 69 understood. An absolute velocity of a body is its velocity with respect to immovable points on the earth. A relative velocity is its velocity with respect to points that are also moving. The direction of the line representing the velocity of the steam relatively to the blades should be such that the lines of flow of the steam enter the blade tangentially to the conven- tionally straight portion of the back of the blade (see Figs. 43, 49, and 50). If the backs of the blades are made to any other angle there will be losses due to impact and eddies. EFFICIENCY OF THE BLADES OF IMPULSE TURBINES. In the velocity diagram in Fig. 38, the initial velocity of the steam entering the nozzle is marked Vi, the velocity in the throat is V , and the absolute velocity of the steam as it leaves the nozzle and enters the blades is V 2 , making an angle a with the direction of motion of the blades. The velocity of the blades V fe , which is the peripheral velocity of the wheel, produces a " relative " velocity of the steam in the blades V^. The angle shows then the theoretical " entrance " angle for the blades that the steam may enter without loss of velocity due to shock or impact. These angles a. and are marked plainly in the draw- ing of " Section B." The relative velocity of the steam leaving the blades is represented by V r3 . Often the blades for impulse turbines are made symmetrical, so that the angle y on the " exit " side of the blades is equal to the angle /? on the " en- trance " side. The absolute velocity of the steam leaving the blades is found by geometrically subtracting again the blade velocity V 6 . The velocity of the blades is always subtracted a second time, because the direction of the steam has been reversed in passing through them. The steam is discharged with the absolute velocity V 3 , which is called commonly the " residual " velocity. Conditions of Best Efficiency. The condition for the highest efficiency of this simple turbine (Fig. 38) will now be discussed. The same velocities represented at the bottom of Fig. 38 are shown again with the addition of an enlarged section of a blade 7 o THE STEAM TURBINE in Fig. 43. The notation is the same as in Figs. 38-41. V 2 and V 3 * are the absolute velocities of the steam entering and leaving the blade, of which a shaded section is shown. V^ and V r3 are FIG. 43. Velocity Triangles for an Impulse Turbine. the corresponding relative velocities of the steam as it passes through the blade. Now the energy in the steam is measured, of course, in terms of its absolute velocity, and is proportional to the square of its velocity, f The energy, then, in a pound of yz y 2 steam entering a blade is and on leaving is The energy taken away by the blades is, therefore, (V 2 2 V 3 2 ). Here g 2g is the acceleration due to gravity (32.2), and for all practical purposes is a constant value. Energy converted into work in a * Observe that V 2 , V 3 , V 4 , etc., indicate absolute velocities, and V&, V r3 , V H , tc., are relative velocities. This relation should be of much assistance in reading the diagrams. The order in the use of subscripts follows the method used for the nozzles in the preceding chapters. The subscript i is still used to represent the initial condition of the steam as it enters the nozzles of an impulse turbine or the first row of sta- tionary blades in a reaction turbine, while the subscript o is for the condition at the throat of a nozzle. The first " discharge " velocity either from nozzles or .stationary blades is therefore represented by the subscript 2. t See discussion of kinetic energy and velocity, page 24. STEAM TURBINE TYPES AND BLADE DESIGN 71 turbine depends then, theoretically, only on the term (F 2 2 F 3 2 ). This term will have its best value, of course, when F 3 is made as small as possible. The best theoretical conditions of blade speed and steam velocity are shown in the following discussion: In practice it is usual to have given (i) the velocity of the steam entering the blades; (2) the " nozzle angle " (the angle at which the steam strikes the blades) ; and usually in impulse tur- bines still another condition, (3) that the entrance and exit angles (|8 and 7) are equal. The velocities that must be considered for these conditions are shown in Fig. 43. Here V 2 is the absolute velocity of the steam entering the blades, the angle a is the " noz- zle angle " and shows the inclination of the nozzle to the plane of the turbine wheel. V 6 is the peripheral velocity of the blades, V r2 and V r3 are the relative velocities of the steam in the blades, and V 3 is its absolute velocity leaving the blades. By the conditions stated, F 2 and the angle a are known, and we are to find the most suitable blade velocity (V 6 ). Also the angle $ is equal to the angle 7, although the value of neither of these angles is assumed. The velocities V 2 , V 6 , and V^ will form one triangle of velocities, and still another triangle is made with V 6 , V r3 , and V 3 . The corners of the latter triangle are marked i, 2, 3, and from the geometry of the figure this triangle is obviously equal to the triangle i, 2', 3, marked by cross-hatching. Now, if we as- sume there is no loss of velocity due to friction and shock in the blades then F r z = F r3 , and the triangle i, 2', 3 can then be in- verted, and, putting the point 2' at 2, it can be made to join up with the triangle o, 2, 3 which shows the initial velocities at the upper end of the blade. The base o, i of the new triangle o, i, 3 is now equal to 2 V b and we can write, by the "Law of Cosines" the equation V 3 2 = V 2 2 + (2 V 6 ) 2 - 2 V 2 (2 V 6 ) COS a, (I i) or F 2 2 - F 3 2 = 4 V 2 V b cos a - 4 V b \ V 2 2 - V 3 2 = 4 V 6 (V 2 cos a - V 6 ). (12) In this equation the term (F 2 2 F 3 2 ), which is a measure of the energy taken away from the steam, is greatest when 4 F 6 72 THE STEAM TURBINE (V 2 cos a V b ) has its largest value; * or we get the maximum energy taken from the steam when V 6 = - V 2 cos a, (13) which is the condition when the line 3, i, or V 3 , is perpendicular to V 6 , that is, when the steam leaves the blade perpendicular to the plane of the wheel. t The condition for which the last set of equations has been worked out represents the usual conditions in practice. That is the " nozzle angle " is usually assumed (about 20 degrees), and the blade angles and 7 are made equal. For this case equation (12), above, represents the best blade conditions, with the abso- lute velocity of the steam entering the blades (F 2 ) and the velocity of the blades (V b ) as the only variables. We can express the efficiency of the action of the blades by dividing the energy taken away in performing work by the energy represented by the velocity of the entering steam; thus, Energy taken away for work, or the actual work done = V 2 2 -V 3 2 ^g Total energy in the steam, which is a measure of the total work V 2 possible = 2g Efficiency^ Actual work done = Vg-V,; g = V 2 total work possible 2g 2g V 2 Now, in equation (12) we have for the best conditions, * If we make the substitution F 2 2 F 3 2 = y, Vb = x, K = F 2 cos a, then for equation (12) we 'can write y = 4 x (K x) = 4 Kx 4 x 2 . For the maximum value of y, x = \K, or V b = i F 2 cos a. f Without the calculus demonstration it is obvious that Vz 2 V is largest for given values of Vz, when Vs is smallest, and this is when the line 3,1 in the triangle 0,1,3 is shortest; or, in other words, when the direction of Vs is perpen- dicular to the direction of Vb- STEAM TURBINE TYPES AND BLADE DESIGN 73 Then substituting this in equation (14), If, further, the-" nozzle angle " a is 20 degrees, as is so com- mon in practice, then Efficiency = b (.940 -J-- (16) The only variable left in this equation is the ratio -r^, and V g and therefore total pressure on the blade is g ~~ Also when = 90 degrees, as is approximately the case in Fig- 33 > cos = 0, and the total pressure is I _ WV g ' Impulse Force Due to Stream Flow Across Moving Blades. When velocities of blades are also considered the impulse of the WV stream entering the moving blade as in Fig. 43 is * cos 0. Similarly the stream leaves with the relative velocity V r3 of which TTTTT the component in the direction of motion of the blade is - -cos g WV of angle 3, i, 2', or - - (V 6 V r3 cos 7). Total impulse is, g therefore, cos + (v b _ Vr3 cos 7). g g For further demonstration see Exercise 6 in Appendix. In Fig. 44 a curve is shown which has been calculated to repre- sent equation (16) for varying values of blade speed (V b ) and with an initial steam velocity (F 2 ) of 3,000 feet per second. The increase in efficiency with increased blade velocity should be STEAM TURBINE TYPES AND BLADE DESIGN 75 observed, and that the highest efficiency is obtained when the blade speed (V fr ) is about half the velocity of the steam discharged from the nozzle (V 2 ). This is a good rough-and-ready rule to remember. If, then, the steam velocity is 2,500 feet per second, the peripheral velocity of the blade wheel, for the highest effi- ciency, should be about 1,250 feet per second. For mechanical reasons it is difficult to construct turbine wheels to run at speeds IPO CO 20 200 400 600 800 1000 1200 1400 1600 1800 20002200 Blade Speed V^ - Ft. per Sec. FIG. 44. Curve of Efficiency of an Impulse Turbine with One Row of Blades and a Nozzle Angle of 20 Degrees for Varying Blade Speeds. much greater than 500 feet per second, so that many designers will generally use low blade speeds to get velocities more suitable for commercial application, knowing well that in this respect they are sacrificing their highest efficiency. In designing blades for turbine wheels the entrance and exit angles (0 and 7) should always be made as nearly as possible of the size determined by the velocity diagrams. If the angles are made much different, there is a sudden change in the direction of the steam instead of a gradual change, with a consequent loss due to shock or impact. Efficiency of Velocity Stages. An impulse turbine with more than one row of moving blades in a single pressure stage (veloc- ity stage type) is represented- by Fig. 39. The energy taken away from the steam for work, as expressed in equation (12), can be readily modified to suit this case. We should have observed that each time steam passes through a moving blade the blade velocity (V b ) is twice taken away (subtracted geometrically) 7 6 THE STEAM TURBINE in the velocity diagrams If there are N rows of moving blades, V 2 2 - W + 2 = 4 NV 6 (V 2 cos a - NV b ).* (12') And similarly (compare with equation 14, page 72), 4 NV b (V 2 cosa-NV b ) V 2 2 Efficiency = and for a 2o-degree nozzle, Efficiency = COS a NV, Vo d7) (18) NV, V 2 V V 2 Efficiency of a Simple Impulse Turbine for Given Blade Speed. In the discussion of the maximum blade efficiency of impulse turbines which has preceded, the velocity of the steam entering the blades was assumed to be known and a suitable blade speed was determined in terms of the entrance and exit angles, which were assumed to be equal. This is the problem which arises when a single-stage impulse turbine is to be designed for given initial and final pressures. When, however, an impulse turbine of more than one stage is to be designed with a fixed blade speed (V b ) of say 500 feet per second,f it is desirable to determine the. * This can be shown geometrically very easily by the method illustrated at the top of Fig. 43 which will be here drawn for three rows of moving blades. As in. the other figures, F 2 is the velocity of the steam entering the first row of blades and Vfi = F r3 ; then in Fig. 45. F 5 2 = V 2 2 + (6 Vb) 2 - 2 V* X 6 V b cosa. F 2 2 _ F 5 2 = I2 Vb (F 2 cos a - 3 F 6 ); and F 2 2 - Flv+ 2 = 4 NV b (F 2 cos a. NVb), if N is the number of rows of moving blades. t Many manufacturers have a standard blade speed and all sizes of turbines are designed for this standard. The blade speeds of impulse turbines vary from 350 to 1200 feet per second. The latter figure, it is stated, has been used success- fully by a European manufacturer. STEAM TURBINE TYPES AND BLADE DESIGN 77 pressure drop in the first stage (and probably also in the second stage, depending on the action 'of the valve gear) to obtain the highest efficiency in this stage. This is because the best results are obtained in most types by getting a larger proportion of work from the first stage than from the other stages.* Efficiency, therefore, is a more important consideration in this stage than in the others. We have thus obtained a very simple form for calculating the efficiency of an impulse turbine; but it must not be overlooked that if the entrance and exit angles are not equal, and in the case of velocity stages if the exit angle of the stationary " interme- diate " blades is not the same as the angle at which the steam is discharged from the preceding blades, these formulas must be considerably modified and the result would not be nearly so simple. It should be observed also that all losses from friction and eddies have been neglected. These more practical con- siderations are discussed in connection with the examples of actual designs of blades on pages 85 to 96. EFFICIENCY OF THE BLADES OF REACTION TURBINES. As in the case of the impulse turbine, the expressions for energy and efficiency will now be derived for the reaction turbine, assuming again that there are no losses to be considered. We must remember that in the reaction turbine there are no nozzles for expanding the steam but that the expansion occurs in both the stationary and the moving blades, so that as the steam goes through the turbine its velocity is gradually and continually changing. We shall first consider a reaction turbine (Fig. 46) with only two sets of blades. As there are no nozzles, the first set is, of course, made stationary. The steam expands in going through * The reason for designing the first stage for the largest amount of work from 25 to 50 per cent, more than in any of the other stages is most apparent in turbines operated by "cut-off" governing like the Curtis and Wilkinson turbines. This method of governing permits a constant standard pressure (presumably that giving the maximum efficiency) in the first stage at all loads, while with fluctuating loads the pressures will vary considerably in the other stages. 78 THE STEAM TURBINE these stationary blades and attains the velocity V^ * when it reaches the first set of moving blades. The relative velocity with which the steam enters the moving blades is V^. Now, in these FIG. 46. Velocity Diagrams for One Stage of a Reaction Turbine. blades the steam is again expanded, so that just before it leaves the moving blades its relative velocity is V r3 , which is greater than V^. The absolute velocity at which it is discharged from the moving blades is V 3 , and we have the following energy relations: V 2 = kinetic energy developed in the stationary blades, or the 2g kinetic energy entering the moving blades. V , 2 Vy> 2 - 5- = kinetic energy developed in the moving blades. 2g V 2 = kinetic energy carried away in the discharged steam.^ 2g The actual work done on the moving blades is W k = (kinetic energy of the steam entering the moving blades) + (kinetic en- ergy developed in the moving blades) (kinetic energy carried away), or V2 2 2 2 W, = IL 2g 2g 2g If the steam had left the moving blades with zero velocity, and, therefore, no energy had been carried away in the discharged steam, the energy available for work would be 2g 2g * See note at the bottom of page-So^ regarding this notation. (20) STEAM TURBINE TYPES AND BLADE DESIGN 79 pffi . actual work done (19) V 2 2 +V r 3 2 -V,3 2 -V3 2 ,. Efficiency = - : -. r = __ 9 .. , 17-; . (21) total work possible (20) V 2 2 +V r3 2 -V r2 2 In the same way the efficiency can be calculated for any number of rows of blades. Equation (21) expresses the efficiency for only two rows of blades one stationary and one moving or, in other words, for one stage. We shall now obtain the efficiency for three stages, that is, for six rows of blades. The correspond- ing velocity diagram is shown in Fig. 47. ,. 47> _ Velocity Diagrams for Three Stages of a Reaction Turbine. '. V 2 = kinetic energy developed in the first stationary blades. y 2 __y 2 = kinetic energy developed in the first moving blades. T-^ = 2 S blades. rr 2 _ kinetic energy developed in the second stationary 1/2 ^ = kinetic energy developed in the second moving blades. V 2 - = kinetic energy developed in the third stationary blades. 80 THE STEAM TURBINE Y .2 _ Y 2 L = kinetic energy developed in the third moving blades. V 2 ^ kinetic energy carried away in the discharged steam. We observe here that the velocities V 03 and V/, 3 are not lost but represent velocities that can be effective in the succeeding stages. For this reason their energies do not enter the discussion of efficiency. The actal work in moving the blades is then '-v-n + rv to! + V - yi 2g J [_2g 2 g J 2 g 2 V 2 V 2-1 V 2 - -Xi. ( 22 ) K g 2g Now, in designing a reaction turbine it is desirable to assume that the blade velocities and the corresponding angles of the blades are the same and that equal steam velocities are developed in each of the three stages, so that, v 2 = v h , = v C2 , y Y V * r2 ~ v n ~ * re> V* - V r , = V n , and W 2 g 3 \-2 g ' 2 g \ The total energy in the steam available for work in this case is [Y 2 V 2 V 2 1 T^^TT^J- The efficiency is then \JJ, V 2j_V 2 V 2 IV 2 ** K _ Y a2 f V r3 v r2 3 V a3 ^ (2*} Wa V a2 2 + V r3 2 - Fr 2 2 It is clear, then, that in the expression for efficiency the last term in the numerator changes its coefficient with the number of stages, and we see in what proportion the efficiency is increased with the number of stages. STEAM TURBINE TYPES AND BLADE DESIGN 8l PRACTICAL DESIGNING OF BLADES. In designing blades for steam turbines we must determine with accuracy, (1) The angles for the edges of the blades. (2) The radial height or length of the blades. From the preceding discussion of velocity diagrams and blade efficiencies it should be clear how the best angles for the edges are obtained. It is first necessary to calculate the velocity resulting from adiabatic expansion between the limits of pressures in the stage for which the blades are intended. Then velocity diagrams must be constructed, varying the blade angles if the blade speed is assumed till the best efficiency is obtained. This will be when the steam leaves the last blades nearly at right angles to the plane of the wheel; that is, when the absolute velocity of the steam leaving the blades is, in the diagram, nearly perpendicular to the line showing the blade velocity (see page 76). Design of Blades for Impulse Turbines. We shall continue with the discussion of the design of blades for an impulse turbine with nozzles and with a single row of blades, assuming now that the entrance and exit angles (/? and y) have been deter- mined. We shall assume also that the total area of the noz- zles at their largest section has been calculated as it has been explained on pages 36 to 41. To avoid losses by eddies, nozzles are often arranged in groups placed symmetrically with respect to the periphery FIG. 48. - Diagram Showing Location of of the blade wheel. Usually Nozzles in a Diaphragm. the nozzles would be arranged in two groups diametrically opposite in a circular plate, called a diaphragm, as in Fig. 48. We shall assume that each nozzle 82 THE STEAM TURBINE group covers one-fourth of the circumference of the blade wheel. Then if the blades in the wheel were removed so that they could not obstruct the flow of steam, the area through which the steam can pass is approximately J nDh for each nozzle group, where D is the mean diameter of the blade wheel and h is the height of the opening from which the blades have been removed. When, however, there are blades on the wheel the height h must be increased, because the effective area for the passage of steam is reduced. Fig. 49 shows two views of a small segment of a blade wheel. The pitch of the blades is marked p and the blade angle is ft. If there are no blades, the area for the passage of steam in a length p is approximately p X h. With the blades in the wheel the area is only p X h sin ft.* It follows then, when we have blades under the nozzle groups, that the effective area under each group is 1 TrDh sin ft. If we call A the total area of the nozzles at the largest cross-section (mouth) we can write A = \ TrDh sin ft + \ TrDh sin ft. A = i TrDh sin ft. ' 2 A , . h = ^ta7' (24) For steam at very high velocity the height of the blades as calcu- lated will be too small for practical working conditions; so that blades less than .25 inch high are not often made. This minimum height is determined most on account of mechanical difficulties; but steam leakage through the clearance outside the blades also becomes excessive when very small blades are used. In impulse turbines with only a few stages no effort is made to make use of the velocity, as such, of the steam leaving the last blades of a stage. This means some loss ; and more experimental work might well be done with the object of showing how this loss can be turned to better account. Fig. 50 shows how impulse turbine blades are laid out. The * It is assumed in such calculations that the thickness of the edges of the blades is practically negligible. STEAM TURBINE TYPES AND BLADE DESIGN SECTION AT CENTER LINE designer must first decide how wide his blades shall be. For turbines of less than 100 horsepower the width of the blades is. often made about i inch, increas- ing this dimension to about 1.5 inches in turbines of 1,000 horsepower. The pitch or cir- cumferential distance between consecutive blades is made about .5 to .6 of the axial width.* In Fig. 50 the pitch is shown by the distance between the points i and 2. Hence, when a drawing FlG 49 ._ Diagram i, lustrating the DC- of blades is Started these points sign of Blades for Impulse Turbines. should first be located. At any point between i and 2 mark another point 3 and through it draw a line 33', making an angle with the horizontal equal to the blade angle y on that side. Draw I through 2 a line perpendic- ular to the last line and intersecting the center line of the blades. Mark this point on the center line 5. Draw through i a line par- allel to 3 3' to intersect 2 5, at 4. With 5 as a center draw an arc tangent to i 4, which completes the lower half of the convex side of the blade. With the same center the concave side of, the next blade is drawn with the arc passing through 2. The arrows in the figure show * The most efficient blade pitch appears to be between the limits of J inch and i inch. Between these two values the efficiency of blades made according to conventional designs is practically constant. The usual blade pitches are f, f, and I inch. Even for very small turbines not much less than J-inch pitch should be used. Designers usually make the pitch and axial width increase a little with the height of the blades. FIG. 50. Diagram Illustrating the Method f6r Laying Out the Blades of an Impulse Turbine. 84 THE STEAM TURBINE plainly the center for these arcs. This construction makes the " perpendicular " width of the steam passage nearly constant. Blade Velocity Losses. Various attempts have been made by experimenters to determine the velocity losses in blades with stationary apparatus, usually by some method of measuring the reaction somewhat in the same way as the coefficients given in Fig. 28 were obtained for nozzles. Such results, however, are not satisfactory for application to designing. Frictional, eddy, and impact losses in moving blades are certainly very different from what they are in stationary blades. Apparently there are only two ways to get good data regarding these losses. Either the velocity must be measured between the blades of an operat- ing turbine with a Pitot tube or similar device, or they must be determined by the " cut and try" method that has been generally followed by designers. The latter method is certainly expensive and a slow one for obtaining results. It seems, there- fore, that more work should be done along the line of the former method by the application of the Pitot tube. In the latest designs of steam turbines there is no difficulty about getting sufficient space for a pressure tube between the blades, as the axial clear- ance in large turbines is often as much as ^ inch. Fig. 51 shows values of the velocity coefficients to be applied in designing steam turbine blades. Curve A is for blades re- ceiving steam from nozzles. Curve B is for stationary blades, or for moving blades receiving steam from stationary blades. Values of both curves vary with the relative velocity of the steam in the blades. The true velocity in the blades is found by mul- tiplying the theoretical relative velocity by the coefficient from the curves.* The values given by these curves may be taken as fairly representative for all the well-known commercial types * Values given by these curves agree well with the determinations made by Rateau, Stevens, and Hobart from the analysis of the losses in operating turbines. Hobart calculated that the blade frictional and eddy losses in a 275-horsepower De Laval turbine are 17 per cent, of the steam velocity which in this case is about 4000 feet per second. He states also that generally in turbines of this type this loss is about 15 per cent, of the relative velocity in the blades. It is stated that designers of Rateau turbines assume a blade velocity efficiency of 96 per cent, at relative velocities of about 600 feet per second. Obviously near zero velocity the STEAM TURBINE TYPES AXD BLADE DESIGN M M .so .81 .82 Ml ^ \ s \ \ \ V \ ^s N \ x N A > \ N X N S B \ \ N X v ^ in which the blades have smooth surfaces and the entrance edges are made comparatively sharp and at a true angle. These curves are intended to be read for only two significant figures. L00 The initial steam velocities in turbines of the Parsons type |- % vary from 150 to 600 feet per |' second, in Rateau turbines from * 500 to 1500 feet per second, in J- 88 Curtis turbines from 1500 to 3000 feet per second, and in the De Laval type from 2 500 to 4500 ~ 500 1000 1500 2000 2500 fppf npr cpronH Relative Velocity of Steam 1Q - in Blades Ft. per Sec. The efficiency of energy con- FJG ^ B , ade Velodty Coefficients> version in the blades of Steam Curve A for moving blades following turbines varies from 60 to 70 per nozzles. Curve B for stationary blades cent.* for sizes from 300 to 3000 or for s blades followin s sta ' tionary blades. kilowattsj and is, roughly about 50 per cent, for smaller sizes of impulse turbines down to about 10 kilowatts. Still smaller sizes may have efficiencies less than 50 per cent., depending largely on the type of construction. For any size of turbine, then, the difference between 100 per cent, and this efficiency of energy conversion is the loss due to disk and blade rotation, leakage, residual velocity, and radiation. In a well-designed turbine of say 300 to 500 kilowatts' capacity > loss is practically zero. Designers of Parsons and Curtis turbines must use some- what larger coefficients (cf. Curve B) for combinations of stationary and moving blades, because stationary blades are not as efficient as nozzles. The data/ for these curves were obtained by measuring with modified Pitot tube apparatus the velocity of steam discharged from stationary blades of various designs. The steam was directed upon the blades from calibrated nozzles. * In stating this efficiency it is assumed that adequate provision is made in these turbines to prevent leakage: in impulse turbines, between the diaphragms and the shaft; and, in reaction turbines, over the ends of the blades .through the radial clearance. This leakage loss is as much as 10 to 15 per cent, in some good commercial turbines. It should be reduced, however, to not more than 5 per cent. f A well-known engineer states that the energy efficiency of the 9000 to 12,000 kilowatt turbines installed in New York and Chicago is as high as 80 per cent. On a conservative basis, however, about 70 per cent, can be assumed for 5ooo-kilowatt sizes and 75 per cent, for io,ooo-kilowatt sizes. 86 THE STEAM TURBINE the sum of the losses due to friction, disk and blade rotation or " windage/' leakage, residual velocity, and radiation losses is, therefore, about 40 per cent. But these are not all actual losses. The energy equivalent of each of these losses, except that due to radiation, which is very small, is immediately converted into heat, and is partially regained in reheating the steam. The sum of these losses actually reheating the steam, expressed as a per- centage of the total available energy, is called the reheating factor. DESIGN OF BLADES FOR AN IMPULSE TURBINE ^Blades are to be designed for a 3oo-kilowatt turbine to operate }yith steam at 50 F. superheat, at an initial pressure of 165 pounds per square inch absolute, and exhausting at i pound per square inch absolute (approximately 28 inches vacuum). Blade speed V 6 is 500 feet per second at the rated speed 3600 r.p.m. It is assumed that the nozzle will be correctly designed for the pressure, so that the nozzle velocity loss is 3 per cent. Governing is to be accomplished by the method of " cutting out nozzles " in the first stage (see page 221). By this method a practically constant steam pressure is maintained in the nozzles of the first stage from light load to overload, and the velocities in this stage are at all loads approximately those giving the best blade effi- ciencies. In the other stages, however, where the number of nozzles open is not controlled by the governor, the velocities are variable. For this reason a large pressure drop is to be used ior this stage,* and to utilize the resulting high velocity efficiently there are to be two velocity stages in this pressure stage. A reasonable value for the first stage pressure is about 35 pounds per square inch absolute. The other stages are to be designed for highest efficiency with a single blade wheel in each pressure stage. Such a design will be a compound type the first stage resembling the Curtis, and the other stages the Rateau turbines. The energy available from adiabatic expansion in the first stage nozzles (as read from the entropy-heat chart) from 165 pounds per square inch absolute and 50 F. superheat to 35 pounds * See footnote on page 7 STEAM TURBINE TYPES AND BLADE DESIGN 8/ per square inch absolute is 122 B.T.U. Disk and blade rotation losses, leakage between the stages at the joint between the shaft and the diaphragm, and residual velocity of the steam leaving the blades amount to 40 per cent.; and it is assumed that all this energy appears again as heat produced by disk and blade friction, steam impact, eddies, and throttling. There is then 40 per cent, of 122 B.T.U., or nearly 49 B.T.U,, going to reheat the steam. This reheating occurs, of course, at the pressure in the first stage (35 pounds). As the result of reheating, the quality of tlje steam in the first stage is changed from .932^0 .985, and the total heat of the steam going to the nozzles of the next stage is increased from 1103 to 1152 B.T.U. Fig. 52 shows graphically this reheating effect. It serves also to show 1225 1152 1138 1103 B.T.U.gcale 892 FIG. 52. Entropy-Heat Diagram for the Design of an Impulse Turbine. the complete energy distribution as required for this design. In each stage, as in the first, the reheating is assumed to be 40 per cent. . Since all the stages after the first are to be of the single wheel impulse type, it is obvious that a large number of stages will be needed in order to absorb the velocity of the steam efficiently. In a stage of the single wheel type the velocity of the steam should 88 THE STEAM TURBINE not be greater than twice the blade speed. Equation (i8b) shows the relation between the steam and blade velocities for the highest efficiency, and this equation can be used for determin- ing quite accurately the best energy distribution. Because a designing coefficient (C) must be inserted to correct for the velocity loss in the blades, this equation will now be written Now the velocity coefficient for single blade wheels is about .95.* The angle /3 is usually about 40 degrees. Blade speed, Vb, is 500 feet per second. Then T7 2 X 500 X COS 20 , K 2 = - - - = 989 feet per second. But from equation (2) we have the relation that the available energy (E a ) in terms of velocity is V223.7/ It is shown then that the required energy per stage is between 19.5 and 20 B.T.U. per stage. The energy distribution with reheating (40 per cent.) was calculated with the help of the chart for 19.5 and for 19.8 B.T.U. per stage, and it was found possible to get almost exactly equal energy distribution with 1.2 stages each of 19.8 B.T.U. between 35 pounds pressure (quality .985) and the exhaust pressure i .o pound? This distribution is shown in Fig. 52. The quality of the steam in each stage is recorded, so that the disk and blade friction can be calculated later from the formulas in Chapter V. * See Fig. 51. To determine an approximate value for this, coefficient the probable relative velocity must be estimated. If a very large error were made in assuming this coefficient it would be discovered as soon as the velocity diagrams are made, as the relative velocity and the coefficients are then accurately deter- mined. STEAM TURBINE TYPES AND BLADE DESIGN 89 Velocity of the steam discharged .from the first stage nozzles is ^2 = -97 X 22 3-7 Vi22 =? 2398 feet per second, and that from each of the other stages is ' V 2 f = .97 X 223.7 Vig.8 = 965 feet per second. The velocity coefficients given in Fig. 51 have been used to lay out the triangles in Figs. 53 and 54. The application can be best illustrated by the triangles in Fig. 53, show- ing the velocities of the first stage. For constructing the triangles in Fig. 53, V 2 is laid off inclined 20 de- grees (the nozzle angle) to the horizontal and made to scale 2398 feet. FIG. 53. Velocity Triangles for Two Velocity FIG. 54. Velocity Triangles for Stages in One Pressure Stage. a Simple Impulse Wheel. To the same scale the blade speed (F&) is laid off for 500 feet, making the relative velocity (V r2 ) in the first row of blades 1938 feet per second, and the entrance angle (B) of these blades is found to be 25^ degrees. The entrance and discharge angles will be made equal, so that the angle C is also 2 5i degrees, determining the slope of the relative velocity (F r s). The velocky coefficient taken from curve A in Fig. 51 for a relative velocity of 1938 feet is .88, so that F r3 = 1938 X .88 or 1705 feet. Vb is again laid off in a horizontal direction, and the absolute velocity of the steam discharged from the first row of 90 THE STEAM TURBINE blades (F 3 ) as read by the scale is 1270 feet, and the true dis- charge angle (D) is 35 degrees. In order that the steam may enter the stationary intermediate blades without shock, the entrance angle of these blades must be also 35 degrees, and the discharge angle (E) will be made 20 degrees, the same as the nozzle angle. The velocity coefficient is now read from curve B in Fig. 51 for 1270 feet,* which is .87, and F 4 is laid off for 1270 X .87 = 1105 feet. Completing the triangles, F r4 is 662, and the entrance and discharge angles F and G are each 35 degrees. The velocity coefficient (read from curve B) is .93, so that F r5 is 615 feet and the final discharge velocity (F 5 ) is 355 feet. Velocities and blade angles are determined in the same way (by applying a velocity coefficient) for the 12 single wheel stages as shown in Fig. 54. Data and results of these velocity triangles are tabulated below for convenient reference: Blade Angles and Velocities of First Stage. First row (moving): entrance and discharge angles 25 J degrees. Intermediate (stationary): entrance angle 35 degrees; discharge angle 20 degrees. Second row (moving): entrance and discharge angles 35 degrees. Vb = 500 F r3 = 1938 X .88 = 1705 Vr 4 = 662 F 2 = 2398 F 3 = 1270 F rs = 662 X .93=615 F r2 = 1938 F 4 = 1270 X .87 = 1105 F 5 - 355. Blade Angles and Velocities of Second to Thirteenth Stages. Single row (moving) : entrance and discharge angles 39^ degrees. F 6 = 500 F r3 = 5 2 5 X .96 = 504. F 2 = 965 F 3 = 339. V* = 525 A slightly higher efficiency could have been obtained if the first stage pressure had not been assumed but had been determined * In stationary blades the absolute and relative velocities are equal. STEAM TURBINE TYPES AND BLADE DESIGN 91 by a "cut and try" method to get the highest efficiency. If the energy for this stage had been a little less, the efficiency would have been increased although an insignificant amount. It is a good rule to remember that with a given blade speed, whenever the line representing the residual velocity slopes toward either side of the vertical, the minimum residual velocity has not been found. A higher efficiency could have been obtained also by reducing the discharge angle of the intermediate blades. This angle is usually made about the same as the nozzle angle (about 20 degrees in most types). If it is made less than 20 degrees, although the apparent efficiency will be increased, there will be probably a greater loss than gain on account of the steam spilling over the blades. Stage Efficiencies. Nozzle efficiency is assumed to be 97 per cent., on the basis of the velocity developed. Efficiency of the energy conversion in the blades can be calculated from the results given by the velocity triangles in Figs. 53 and 54. In the first stage the velocity absorbed in moving the turbine is the initial velocity (F 2 ) less the residual velocity, (F 5 ), and the velocity losses in the blades are (V n F r3 ) and (F r4 F^). Then the energy absorbed in the first stage,* or Work Done = F 2 2 - (F r2 2 - F r3 2 ) - (F 3 2 - F 4 2 ) - (F r4 2 - F r5 2 ) - F 5 2 . Work Done Blade Efficiency = Work Possible F, 2 - F r2 2 + F r , 2 - F 3 2 + V:~ - V r < 2 + VJ - V, 9 V? Blade Efficiency (first stage) (2.SQ8) 2 -(iQj8) 2 + (I7Q.0 8 - (i27o) 2 + (iio.O 2 - (66 2 ) 2 4- (6i.Q 2 - ( 239 8) 2 Blade Efficiency = 75.3 per cent. Nozzle and blade efficiency of the first stage is therefore 75-3 x ^97 * = 74-2 per cent. * When writing efficiency equations, it must be remembered that efficiencies are proportional to the available energies and to the square of the velocities. 92 THE STEAM TURBINE Similarly for the second stage (also third to thirteenth stages) we have, V - V 2 (V 2 V 2 } Blade Efficiency = - - ^ ^^ - rjj - Blade Efficiency I <*# ~ -57 X 2000 X 1.15 (see page 100), or d z must be at least 33.76 inches, which corresponds to a periph- end* speed of 33-?6 X 3.1416^ X 1500 > m ^ ^ ^ ^^ The value selected for this stage (280 feet per second) from the viewpoint of permissible stresses is well above this minimum limit. Blade speeds as high as 400 feet per second are now used in some American designs of steam reaction turbines. Having determined that the conventional blade speeds are very satisfactory for this design the required number of reaction stages will be calculated in the usual manner as follows: If the blade speed of the whole turbine had a constant value of 140 feet per second, then (i4o) 2 X n = 2,560,000; n = 128 (nearly). As, however, only one-fourth of the work is to be done by the first section * operating at this blade speed, the number of stages in the first section is - - = 32. The value of n for the second 4 section is 64; and as one-fourth of the work is done also in this section, the number of stages is 16. For the third section n is 32, and since one-half of the work is done in this section, the number of stages is 16. Each section of the rotor is commonly divided into two or four groups or " expansions." Reaction turbines are usually designed for equal work (energy) per stage for a given section of the rotor. In the high-pressure or first section, one-quarter of the work is done, and the avail- able energy for each of its thirty-two stages is 260 B.T.U. -r- (4 X 32) = 2.03 B.T.U. Similarly the available energy for each stage of the intermediate section is 260 B.T.U. -r- (4 X 16) = 4.06 B.T.U.; and for each stage of the low-pressure section is 260 B.T.U. -*- (2 X 16) = 8.13 B.T.U. It may be assumed * See page 97. STEAM TURBINE TYPES AND BLADE DESIGN IOI that about one-half of the available energy in each stage produces velocity in the stationary blades' and the other half in the moving blades. The theoretical angles are determined from velocity tri- angles, applying the coefficients from curve B in Fig. 51, by the usual methods as explained for impulse turbines. The discharge angles for all the stages except the last groups in the low-pressure section will be assumed to be 20 degrees. The angles for the last stages will be made 45 degrees. It is obvious, of course, that the discharge angle is always the same as the " absolute " angle at which the steam enters the succeeding row of blades. In this design no allowances are made for probable " gauging " of the blades to adjust the thrust on the rotor or for other reasons.* The velocity of the steam leaving the first row of stationary blades in the high-pressure section is about 225 f feet per second. A net area of 28.50 cubic feet -f- 225, or .127 square feet, or 18.2 square inches, is required to pass the steam. As the discharge angles of the blades in the high-pressure and intermediate sec- tions are to be made 20 degrees, that value will be taken for this design, and the actual area of the blade ring will be approximately 1 8. 2 -f- .342, t or 53.3 square inches. The blade speed of the high-pressure rotor is 140 feet per second, so that the mean diameter of the blade ring is * It has been stated that some makers of marine turbines who have not had much experience in building them will often design turbines to give considerably larger output than is intended for the service and then reduce the output to the required rating by " gauging" the blades. t In a reaction turbine the maximum velocity in each stage is attained when the steam is discharged from the stationary blades. Although there is expansion also in the moving blades, more velocity is absorbed in them than is produced, and the velocity of the steam discharged from the moving blades is considerably less than 225 feet per second. J The total area of the annulus for blades with discharge angles of 20 degrees is the net required area divided by sin 20 degrees (see Fig. 49). Practical designers often call the sin of 20 degrees one-third and make the area of the annulus three times the net required area. Manufacturers generally appreciate the gain from operating at high peripheral speeds of the rotor. To-day efforts are directed generally by all makers of direct- connected turbine-generators to improve the mechanical construction of the gen- erator to run at higher speeds. 102 THE STEAM TURBINE 14 X = 1.78 feet, or 21.4 inches, and the height of the 1500 X 3- I 4 l6 first row of blades on the rotor is approximately 53.3 square inches -T- 21.4 X 3.1416 = .80 or nearly^f (see table, page 103) inch. With full rated pressure in the admission chamber about 7 per cent, of the total steam leaks through the " dummies" or balance pistons at the high-pressure end of the turbine. This leakage as well as that around the tips of the blades through the radial clearance is not considered here in the calculations. It is prob- able, however, that the amount of this leakage is quite sufficient to allow for the thickness of the blades on the discharge side.* The volume of the exhaust steam (i pound per square inch abso- lute pressure and .886 quality) is 297 cubic feet per pound. Initially the volume was 3.215 cubic feet per pound, so that the volume in the last row of blades is 92.5 times that at admission. Since one-fourth of the work is done in the blades of the first section, one-fourth of the total expansion occurs in them, or the volume entering the second section is ^92.5^ or 3.10 tunes the original volume. Since the mean diameter is to be made A/2 times that at the high-pressure end and the steam velocity is to be also A/2 tunes as great so as to correspond with the increase in blade speed (which is \/2 times that in the first section. See page 100), the height of the blades in the first row of the inter- mediate section will be 3- 10 = 1.55 times that of the first V 2 X V 2 row in the high-pressure section. Similarly the blade height for the first row of the low-pressure end will be 1.55 times that of the first row of the intermediate section. Each of these sections will be divided into four groups or " expansions." Since the volume is increased four times for each section, the blade height of each * Thomas uses a coefficient of 1.5 to increase the area of the blades to allow for the thickness at the discharge side. If the blades are made thin at their edges, as in good designing, it is not customary to use a coefficient " for the thickness of the blades." t Let v r = volume at end of third section, vi = volume at beginning of first section, x = number of expansions, then v r = vf, and v\ = *ff. STEAM TURBINE TYPES AND BLADE DESIGN 103 of the high-pressure and intermediate groups will be ^3.10, or 1.33 times as large as in the preceding one. The calculated blade heights for each of the four groups of the high-pressure and intermediate sections are given in the follow- ing table: Group Number. i 2 3 4 Blade height, Blade height, high-pressure section intermediate section f \t $ lit Blade heights are adjusted to sixteenths, although in practice the nearest eighth is commonly used. Because of the long blades in the low-pressure section they will be made in eight groups. The height of the first group will be 1.55 times the height of the first group of the intermediate section. The volume entering the third section is ^92.5 X ^92.5 = 9.61 times the original volume, and blade height in first row of third section is 1.55 X 1.55 X height of first row = (i.ss) 2 X .80 = 1.9 inches, or approximately iH inches. Each blade in third section is ^9.61 = 1.33 times height of preceding one, or the height of second row is 1.33 X i|f or 2.57 (approximately 2 T \ inches) . The results are tabulated as follows : Group Number. i 2 3 4 5 6 7 8 Blade height (inches) itf *& 3* 4& 6& 8| ioi Ml Martin states that at the high-pressure end (in turbines for stationary service) it is desirable to limit the blade height to not less than one- twenty-fifth (&) of the drum diameter. If the blades are shorter than this the loss by leakage around the tips may become excessive. In marine turbines the high-pressure bla.des in the first section are only 7^ of the drum diameter. On this basis the blade heights might be slightly increased. 104 THE STEAM TURBINE At the low pressure end of the turbine the length of the blades would be considered excessive in practice. It is a rule generally followed by designers of reaction turbines not to make the greatest blade height more than one-sixth the mean diameter of the blades for the section considered. The mean diameter of the low pressure section is 21.4 X \/2 X \/2 = 42.8 inches. In this case the maximum height would be, therefore, about 7.1 inches. In order to reduce the length of the blades so that FIG. 56. Details of the Design of Reaction Blades. practical requirements shall not be exceeded, the discharge angle of the blades must be made greater than 20 degrees. Such blades with enlarged "exit" angles are called wing blades. The tangent to the curve at the back of the blade on the entrance side becomes about 90 degrees, and at the discharge side 45 degrees instead of 20 degrees. As the result of this change, the net area for the passage of the steam is .71 * (sin 45 degrees) instead of the standard ""f f tne annulus without blades. Wing blades 7 inches long can be used to replace satisfactorily the blades in the 5th group; but as those of the j5th, 7th, and 8th groups must * In the turbines of the steamer Mauretania, wing blades giving a passageway of .86 of the annulus were used, but such a large degree of "-winging " is not adopted in steam turbines for electric generators. t The sin of 20 degrees is .34, but practical designers take.it often for convenience in calculating as J. STEAM TURBINE TYPES AND BLADE DESIGN 105 be made of the same length, the^e blades will be shorter than they should be. This constriction of the steam passage, however, can- not well be avoided without making the rotor in four diameters. Fig. 56 shows how the blades of reaction turbines are laid out. As explanatory of this figure a table is given below showing the corresponding dimensions used by one manufacturer.* In the table data for five standard blades are given for varying discharge angles (0) from 20 degrees to 35 degrees and blade widths (w) of .25, .375, and .50 inch. All the linear dimensions are given in inches. ft is the entrance angle of the blades. Blade Number. < 2 4 5 20 2O 20 30 5 a 10 9 30' 1 4 30' 15 45' 1 8 40' w 0.25 0-375 0.50 0.50 0.50 ft 67 30' 67 30' . 67 30' 60 60 R 0.485 o-555 0.794 0.804 0.810 A -35 0.045 0.068 0.050 0.040 b 0.020 O.O2O O.O2O 0.020 0.020 R, 0.172 0.260 0.342 0.3*3 -34 C 0.049 O.O4O O.I IO 0.147 0.218 R, 0.070 0.109 0.164 0.210 0.212 k 0.008 0.010 0.015 O.O4O 0.056 m 0.123 0.185 0.288 0.280 0.280 n I 0.185 o.i 66 0.280 0.223 0.383 0.282 0.332 0.156 0-33 o 134 H 0.478 0.552 0.770 0.770 0.770 ! Another table is given here showing the principal dimensions of a 4oo-kilowatt reaction turbine with 3, 4, and 5 groups per section. The blade numbers in this table refer to the corre- sponding numbers in the table above. This table is partic- ularly useful for showing values assumed by designers for the blade pitch. It is not considered practicable in this type of blade construction to use a smaller pitch than .177 inch when a calking tool must be inserted between the blades. Manufacturers have usually curve sheets of empirical data from which the pitch and other standard dimensions are obtained. * The Engineer, Dec. 16, 1907. io6 THE STEAM TURBINE FIRST SECTION. Number of Group. Diameter of Section in Feet. Dis- charge Angle. Blade Height in Inches. Blade Number Volume Cubic Feet per Pound. v b + i\ Blade Pitch. Number of Blades. I 0.84 20 0.6875 , 4.08 .62 0.177 179 5-6 0.25 127 2 0.84 20 1. 00 I 5-6 .62 0.1875 169 7.38 0.2475 128 3 0.84 2O '*5 I 7.38 .62 0.172 184 8.92 0.2175 146 SECOND SECTION. I 1.187 20 0.6875 2 8.92 .62 O.2O 180 10.63 0.31 116 2 1.187 20 0-9375 2 10.63 .62 0.215 207 13.8 0.3075 144 3 1.187 20 1.2 2 13.8 .62 0.215 207 4 1.187 20 1-75 2 18.8 18.8 .62 0.3 2 3 0.208 158 214 26.6 0.326 i37 THIRD SECTION. I 1.88 20 0-9375 2 26.6 .62 0.208 340 35-8 0.307 230 2 1.88 20 L3I25 2 35-8 .62 0.208 34 52.8 0-339 210 3 1.88 20 2.00 2 y-\ .62 0.198 358 83.8 0-355 200 4 1.88 30 2.75 4 83.8 .69 0.25 284 161 55 0-53 134 5 1.88 30 4.5 4 "161 .70 0.308 230 3 11 .46 0.54 J3 1 Radial Leakage. As the volume of the steam increases, the area of the annulus of each ring of blades must, of course, increase proportionally. This increased area would be obtained most easily, as with impulse turbines, by increasing the blade heights in each ring. This method, however, would make it necessary to carry as stock in the store-room a great number of blades of different sizes. In order to reduce the stock of blades and to reduce the cost of machining the rotor and casing, it is custom- ary to make a group of several rows of blades of the same height, STEAM TURBINE TYPES AND BLADE DESIGN 1 07 and the required increase in area through each ring of blades is obtained by decreasing the number of blades in each succeeding stage. The two values of volume, pitch, and number of blades given for each group in the preceding table are for the rows at the beginning and at the end of the group. In the discussion of the design of these reaction turbines it has been assumed that each section of the rotor is made of the same diameter from the first to the last group. For theoretical con- siderations this assumption is permissible, but actually for each blade group the diameters of both the rotor and casing are changed so that approximately half the increase in blade height is cut out of the casing and the other half is taken from the rotor. It is usually stated that this is done merely for mechanical reasons, but this method has advantages also in order to secure the best steam flow. It is well known that steam tends to fill completely the passage through which it flows and at the same time expand at right angles to its axis of flow. Now if all the expansion is made on the casing side of the blades the expansion of the steam will increase the leakage around the tips of the blades next to the rotor without materially affecting the leakage at the tips nearest the casing. The leakage of steam around the tips of the blades depends, of course, again upon the amount of the radial clearance. Im- provement in the design of reaction turbines will be largely accomplished (i) by skillful designing and machine work to permit the reduction of radial clearances and (2) by increasing the blade speed. In fact the question of allowable radial clear- ance depends finally on the blade speed. If the blade speed is increased it is possible to use higher steam velocities with larger pressure drop per stage, and consequently fewer stages. This is apparent also from an inspection of the designing formula on page 97. With fewer stages a shorter rotor is required which will also be proportionately stiff er; and with a stiff shaft it is pos- sible to allow very small radial clearances, provided, of course, temperature effects are carefully studied. On the other hand, by increasing the pressure drop per stage the tendency for leakage I08 THE STEAM TURBINE is increased, but there is also a compensating effect in that the number of leakage areas is correspondingly reduced. The reader will have observed that the design of reaction turbines is largely by " cut and try " methods. For this reason it is a financial absurdity for a manufacturer to-day to begin making reaction turbines unless he has practically unlimited resources and can obtain from makers of similar machines at not too large a cost the results of their experiences. The method explained here of determining the important and unique parts in the design of a reaction turbine for a given set of conditions, as regards maximum output, steam consumption, pressure, superheat and vacuum, although very simple in all essentials as regards standard practice, gives results on which it is impossible to improve by the most elaborate mathematical analysis imaginable. In fact all elaborately mathematical anal- yses of the action of steam in a reaction turbine depend finally on the substitution of certain coefficients, most of which have no basis in fact. DESIGN OF A COMBINED IMPULSE AND REACTION TURBINE. The design as regards the general method for a combined impulse and reaction turbine will be similar to that for the design of the impulse turbine (pages 86-95), which had two velocity stages in the first pressure stage with two rows of moving blades or buckets. All the other pressure stages had only one velocity stage and therefore only one row of moving blades. For the combined impulse and reaction (similar to Fig. 107, page 174), the first stage might well be arranged with two velocity stages as in the design referred to, and the other stages could then be designed as a separate two section reaction turbine, assuming the steam to enter the reaction portion of the turbine at the quality determined by adiabatic expansion in the first stage. Assume twice as much work is done in the second reaction sec- tion as in the first. With this understanding it is certainly STEAM TURBIXE TYPES AXD BLADE DESIGN Io8a unnecessary to go through again in detail the details of the de- signing of the blading of the reaction portion. Another method would be to divide up the work to be done by the various sections as in the design of the complete reaction turbine; that is, one-fourth of the work would be done by the first pressure stage having its two velocity stages (as in Fig. 107, page 174), another fourth would be done by the first section of the reaction blading, and the remaining half of the work by the second and last section of reaction blading. In the case of a double-flow turbine in which the low-pressure section is divided into two halves, the equation given on page lood for minimum permissible diameter of the last stage would be found to be approximately d 2 2 = | X .57 X output in kilowatts or d z 2 = .285 X output in kilowatts. In the same connection purchasers of steam turbines should guard well their interests by exercising good business judgment in purchases. Like all other kinds of machinery, there will be " troubles " with new types of steam turbines, and unless the manufacturer is known to be financially responsible and well established in the business, the purchaser should not buy until he has made very careful investigations of the merits of the new machines; and he should always insist on having accurate and complete acceptance tests, made preferably by disinterested engineers of repute. Exercise. Design the blades for a 3Oo-horsepower (maxi- mum output) impulse turbine with two pressure stages and two velocity stages in each pressure stage (Curtis type). Initial ad- mission pressure is 165 pounds per square inch absolute at 100 F. superheat, and the exhaust pressure is i pound per -square inch absolute. Blade speed 500 feet per second. Reheating factor is 50 per cent. Use 8 nozzles and arrange for equal energy dis- tribution in the various stages. Nozzle loss is 2 per cent, of velocity, and take blade losses from curves on page 85. Io8b THE STEAM TURBINE Exercise. Design of the blades for a reaction turbine with 50 stages (Parsons type) for the same conditions of power, pres- sures and superheat as in the preceding example. Exercise. Design the blades of a combined impulse and re- action turbine, having a single pressure stage of the impulse type with two velocity stages (Curtis type) and the usual type of reaction blading. Conditions of power, pressures and super- heat are to be the same as in the preceding exercises. GENERAL COMPARISON OF COMMERCIAL IMPULSE AND REACTION TURBINES. IMPULSE. 1. Few stages. 2. Expansion in nozzles. 3. Large drop in pressure in a stage. 4. Initial steam velocities are in general high (1000 to 4000 feet per second). 5. Blade velocities 400 to 1200 feet per second. 6. Best efficiency when blade velocity is nearly half the initial velocity of the steam. For a single wheel per pressure stage. REACTION. 1. Many stages. 2. No nozzles. 3. Small drop in pressure in a stage. 4. All steam velocities are low (30x5 to 600 feet per second). 5. Blade velocities 150 to 400 feet per second. 6. Best efficiency when the blade velocity is nearly equal to the highest velocity of the steam. STEAM TURBINE TYPES AND BLADE DESIGN 109 The work done per stage is always much greater in current practice in impulse than in reaction turbines. For the same total limits of pressure the work per stage is inversely propor- tional to the number of stages. In general, we may say that mechanical considerations and the speed at which machinery can be conveniently operated determine the size and number of revolutions at which a tur- bine can be run. In a good design about the same total efficiency is obtained, whether the turbine is classified as an impulse or a reaction machine. Radial Blade Clearances. In impulse turbines the radial clearance (between the blade ring and the inside of the casing) FIG. 57. Illustrates Radial and Axial Clearances in an Impulse Turbine. is not important. It is one of the first principles of a good design of an impulse turbine that the blades shall be made long enough to allow the steam to be discharged through them freely without "choking" the flow and "spilling" steam over the outer edges of the blades. Since the pressure is the same on the two sides of the blades, radial blade clearances in impulse turbines can be made of generous dimensions. (See Figs. 57 and 119, in which Curtis designs are shown.) In reaction turbines, on the other hand, it is very necessary to 110 THE STEAM TURBINE make radial clearances as small as is mechanically possible, because in these turbines the steam expands in the moving as well as in the stationary blades and there is a drop in pressure between the two sides of every row of blades. On account of this pressure drop there is a continuous flow of steam around the edges of the blades, which is large or small in amount in pro- portion to the size of the radial clearances. The clearance between the stationary blades fixed to the casing and the surface of the rotor is of course just as important as that between the moving blades and the casing. An American manufacturer of the Parsons reaction turbines states that the radial clearances are from .02 to .10 inch, varying with the diameter of the drum. These limits are given for drums between i foot and 10 feet in diameter. Radial clearances of large sizes of Parsons turbines made by Brown-Boveri & Co. are from 2 to 3 millimeters (.08 to .12 inch). Attainment of minimum safe radial clear- ances is the goal for every designer of reaction turbines. Axial Blade Clearances. Axial clearances with respect to impulse and reaction turbines present conditions just opposite from those for radial clearances. In reaction turbines, axial clearance is not an important factor in the design. Until recently, however, it was considered very important in the design of im- pulse turbines to make the axial clearance between the moving blades, and nozzles or stationary blades, as small as possible; and indeed, unfortunately, some impulse turbines in the early days were built with very small axial clearances, so that the least vibration of the shaft would cause striking of the moving blades against the nozzles. It has been shown, however, by actual experience as well as by experiment that axial clearances can be made as large as .20 inch without appreciable loss; or, in other words, practically as large as in reaction turbines usually about .10 to .20 inch. The difficulties of the designers of the first commercial impulse turbines can well be imagined when it was considered so essential to make the axial clearances not more than .02 or .03 inch. In the case of one small turbine built with three stages STEAM TURBINE TYPES AND BLADE DESIGN III the axial expansion of the shaft in the length included between the high pressure nozzle mouths and the blades of the third stage was .10 inch by actual measurement. To allow for a shift- ing of the blades of .10 inch with only .03 inch axial clearance in a turbine with velocity stages was not an easy problem. Axial clearances in Curtis impulse turbines are .06 to .15 inch for soo-kilowatt sizes, and in larger machines are sometimes as much as .25 inch. In Rateau impulse turbines these clearances are from .12 to .25 inch.* Materials for Blades and Erosion. Rolled steel is a very suitable metal for turbine blades when used for dry or super- heated steam at either high or low velocities if the turbine is kept in practically continuous operation. Wet steam, however, will wear away steel blades very rapidly by erosion, and when a turbine fitted with steel blades is idle for days at a time the blades will corrode, so that when it is started again the particles of iron oxide (rust), will be carried away by the steam to act like a sand blast on the blades in succeeding stages. Steel is an exceptionally good material for blades under favorable con- ditions because it can be rolled cheaply into bars of any shape of section,! and it is unequaled for strength. Copper alloys, known in the trades as " extruded metal," are made into bars of any shape of section by " drawing" as wire is manufactured. No metal has all the physical properties desirable in a blading material. Recently a compound metal known as Monnot or "duplex" metal has been developed. It consists of a steel core covered with a thin copper sheathing chemically .welded to the steel in such a perfect manner that the blades may be drawn cold from the original ingot into the required finished section * In impulse turbines with nozzles discharging radially into blades or buckets on the rim like the Sturtevant, Terry, or Riedler-Stumpf types, it is stated that there is no appreciable change in velocity loss when the radial clearance (between the nozzle and the buckets) is increased from .10 to .40 inch. f Rolled bars are cut up into lengths corresponding to the height of the blade plus an additional length for dovetailing into the rim of the turbine wheel. When this dovetailing method is used (Fig. 63) the blades are separated from each other by " spacing pieces " of suitable shape to fit between the blades. 112 THE STEAM TURBINE without in any way affecting the bond between the copper and the steel. Fig. 58 shows an etched section of a blade of this material from a Westinghouse turbine. Blades like those, for example, in Figs. 59-61, which are too irregular to be rolled or drawn are usually cast of bronze or copper FIG. 58. Etched Section of a Blade made of Monnot Metal (Steel and Copper). alloys. Forked blades (Fig. 59) similar to those used in Wilkin- son turbines are commonly cast with the forks far enough apart so that they will pass over the enlarged section of the rim and are forced together when they are in place. Another method is to FIG. 59. FIG. 60. FIG. 61. Designs of Steam Turbine Blades. cut away the enlarged part of the rim section for a short length, and blades cast with the forks in their normal position can be inserted at this place and can then be pushed around on the rim till all the blades are in place. The parts of the rim cut away must be replaced to secure the blades at that section. The blades of small sizes of Curtis turbines are sometimes cut in the rim of a solid disk by automatic machinery. De Laval STEAM TURBINE TYPES AND BLADE DESIGN 113 FIG. 63. Dovetailed Type of Blade (Curtis). FIG. 64. Typical De Laval Blades. THE STEAM TURBINE blades are made of steel forged into the peculiar shape required for insertion into the disk wheel. (See Fig. 64. ) It is stated that the usual alloy used in England for blades of Parsons turbines is 63 Cu + 37 Zn; but any zinc alloy is quite unsuitable for superheated steam or for high velocities. Fig. 65 * shows the effect of the erosion due to steam on blades FIG. 65. Photograph of Turbine Blades Showing Erosion. made of Delta metal about 60 Cu + 37 Zn + 3 Fe. These blades were held stationary in a steam jet for 128 hours. The blades on the left side of the figure were subjected to steam at 2900 feet per second; and those on the right to steam at 600 feet per second. Low-velocity steam eroded the blades so little that the tool marks put in the blades when they were made are still visible. * The author is indebted to Mr. Francis Hodgkinson for this photograph. CHAPTER V. MECHANICAL LOSSES IN TURBINES. IN the designs of turbines on the preceding pages the nozzle and blade efficiency was first calculated, and then the total, or "over-all," shaft efficiency was obtained by subtracting other losses as follows: (1) Disk and blade friction, or windage, due to rotation in a fluid medium (steam). (2) Leakage of the steam chiefly through the clearance between the shaft and the diaphragms of a multi-stage impulse turbine and through the radial blade clearances in a reaction turbine. (3 ) Bearing and stuffing-box friction losses. (4) Radiation. Of these the first three are, in a way, mechanical losses in the sense that the details of mechanical design largely determine their values. The first of these losses, disk and blade rotation loss, is by far the most important and will be discussed first. Losses Due to Friction of Turbine Wheel Revolving in Steam. Losses due to revolving disks or wheels in steam are very diffi- cult to determine with accuracy. Tests to determine these losses are usually made with the wheel rotating in stagnant steam, and it is practically impossible to have, under these con- ditions, steam of the same quality or superheat in all parts of the casing. A number of formulas have been proposed for the friction losses of disks and blades in dry saturated steam, but there is no good agreement of the results of different experi- menters. In fact no great accuracy can be expected because there is no doubt that the exponents of logarithmic friction "5 IX 6 THE STEAM TURBINE curves plotted from such tests vary considerably with the details of design, and besides, it is very difficult to get good tests.* An important reason why the tests from different designs of turbines do not agree better is that clearances between moving and stationary parts have an appreciable effect. If the clear- ances all around the wheel are very small the wheel and blade friction loss will be somewhat less than for a wheel revolving in large clearance spaces. This effect is most marked at low speeds. When higher speeds are reached there is more tendency for the wheel to "cut through" the surrounding steam without increas- ing the " disturbance" in proportion to the increase in speed. The author has from time to time investigated large numbers of tests to determine the friction losses of wheels and blades of turbines in steam and air, and this experience has shown that the following formulas will give fair average results for forward running in practically stagnant steam. The rotation loss or skin friction of a plain disk | revolving in dry saturated steam is expressed by the following formula in horsepower : (25) where d is the diameter of disk to inner edge of blade in feet. u is the peripheral velocity of disk in feet per second. J y is the density of surrounding medium in pounds per cubic foot (reciprocal of the specific volume). A similar term to determine the rotation loss of one row of blades F 6 (without the disk), in horsepower, is , (26) ioo * The peculiar circumstance that water in the liquid state can exist, indefinitely, in the presence of superheated steam, leading some to propose a vergasungs- wdrme, is one of the greatest difficulties. f Similar to those in Curtis and Rateau turbines. On account of the thick hubs of De Laval disks (Figs. 83 and 84), about 15 per cent, should be added to the results given by equation (25) to allow for the larger surface of these disks. J It is often stated that the disk and blade friction losses vary as the third power of the speed. But this value cannot be stated with any claim to great accuracy. Experimenters do not all agree on this value, and values from 2.5 to 3.5 are given by different authorities. The author, from the result of the experiments he has MECHANICAL LOSSES IN TURBINES 117 where 1 = length of blades in inches excluding the band (if there is one), and d, u, and y are used as before. For a simple turbine wheel with only one row of blades we can write for the total rotation loss F, in horsepower : F, = (.o8d + .Sl 1 ' 5 )^ T' 8 dy. (27) ( IOO ) The density of superheated steam varies with the amount of superheat, so that by adding the following notation, y d = density of dry saturated steam at the pressure of the surrounding medium in pounds per cubic foot, D = superheat in degrees F., v d = specific volume of dry saturated steam at the pressure in the surrounding medium in pounds per cubic foot, and using the following equation for specific volume v of super- heated steam given on page 53, v 5 = (i + .00065 D) 2 v rf , we have the following formulas, taking the place of (25), (26), and (27) above, for superheated steam: / u V 8 ( ) 7d T w = .08 d 2 Q6 p)2 ' (25' ) i^\ F, - (.08 d + .3 I 1 ' 5 ) XI ' 7 4 (27') (i + .00065 D) 2 investigated, considers the 2.8 power a good average value suitable for practically all conditions. In the value of the exponent this rotation loss resembles train and ship resistance. The windage loss of dynamos properly designed for high speeds is a curve of the second power. When the windage loss curve of a dynamo shows an exponent of 3 or 3.5 it must be inferred that the machine was not properly designed for high speeds. It may be interesting to the practical men reading this book to know how the exponent is obtained from a test. This is done most conveniently by plotting on any suitable coordinate paper the logarithms of the loss for the ordinates, and the logarithms of the speed for the abscissas. The tangent of the curve is the value of the exponent if the scales of ordinates and of abscissas are the same. n8 THE STEAM TURBINE Or the curve given in Fig. 68 can be used to correct equations (25), (26), and (27) by means of a coefficient. While the effect of superheating is to reduce these losses, moisture, on the other hand, increases them very appreciably. .70 40 60 100 120 140 160 180 200 220 240 Superheat-Degrees Fahr. FIG. 68. Curve to Correct Rotation Losses for Superheat. 2.60 2.40 2.20 2.00 1.80 1.60 1.40 1.20 1.00 / / / i / / / / / / / / / / X x A !/ ^ s* ^ -"'' e== ' " p- ^- - ****** 2 4 Percent. Moisture in Steam FIG. 69. Curve to Correct Rotation Losses for Moisture. Fig. 69 shows a curve giving the coefficients to be applied to the losses calculated by the above formulas for dry saturated steam to correct for moisture. Example. Calculate the frictional rotation loss of a disk 3 feet in diameter of a non-condensing single stage turbine (steam MECHANICAL LOSSES IN TURBINES 119 pressure 15 pounds per square inch absolute) when the steam is (i) dry saturated, (2) superheated 100 F., (3) 10 percent wet. The speed is 3600 revolutions per minute. Determine also the rotation loss of a single row of blades i inch long on this disk. At high peripheral speeds the rotation loss of a non-condensing turbine with the wheels revolving in steam at atmospheric pres- sure is quite large, as the example above illustrates. This loss decreases, however, very rapidly with increasing vacuum, and is, in fact, nearly proportional to the pressure. This fact is not, however, always appreciated by designers. Of course, when disk and blade rotation losses are being calculated for a series of pressures for the several stages of a turbine, as is usually done before deciding on the nozzle proportions, it is only neces- sary, if the wheel dimensions are constant, to calculate for one pressure and determine the values for the other stages by multi- plying by a constant representing the ratio of the densities. Of all the variables in equations (25), (26), and (27), the density is the only term varying as the first power. For most work it will be allowable to assume, within a small range, the density pro- portional to the pressure ; that is, if the disk and blade loss has been calculated in steam at some given pressure, the correspond- ing friction loss at any other pressure may be found by the ratio of the pressures. The disk and blade rotation losses of a Parsons or other drum type may be calculated with the above formulas by calcu- lating the loss for each group of blades of the same length and diameter and adding to the sum of the blade losses the rotation loss due to disks approximately equivalent to the outside surface of the drum. As the friction loss due to the drum itself is small compared with that of the many rows of blades, no great accuracy need be attempted in this calculation. In small sizes of steam turbine-generators the rotation loss is a considerable percentage of the total output. The disk and blade loss of a single stage turbine with a single row of blades, rated by the manufacturer at about 250 kilowatts at 3600 r.p.m., is shown 120 THE STEAM TURBINE in Fig. 70. The curves show that the rotation or windage loss of the generator alone is about 30 kilowatts and the total rotation loss is 50 kilowatts or 20 per cent, of the rated output. Similarly 2000 FIG. 70. Rotation Loss Curves of 25o-Kilo\vatt Turbine-Generator. the total rotation loss of a 2000 to 3000 kilowatt turbine-gen- erator is from 10 to 15 per cent, of the rated output. Method of Making Tests to Determine Wheel and Blade Rotation Losses of a Steam Turbine. The simplest method for making such a test, and the one commonly employed, is to attach an electric motor to the turbine shaft (sometimes in a direct- connected set the generator is used as a motor) and run it at a number of different speeds. In taking a series of speeds, no observations are made until conditions have become " steady, " and the speed must be held constant for several minutes so that a number of readings can be taken on the electrical instruments measuring the input of the motor. The results give the rota- tion loss of the wheel and blades in steam as well as bearing friction and the rotation or "windage" and electrical losses of the motor. Then the turbine wheel is removed, leaving the packing MECHANICAL LOSSES IN TURBINES 121 at the generator end of the turbine on the shaft, and the motor is run alone. The power now measured is that required to over- come the rotation and electrical losses of the generator and the bearing friction. Curves of power and speed as variables (Fig. 70) are plotted for each set of observations, and the disk arid blade loss is determined by subtracting the ordinates of one curve from those of the other. It may be assumed with sufficient certainty that the weight of the turbine wheel itself would not alter the bearing losses to any considerable extent.* The important fact that all results given here are for disks and blades revolving in a stagnant medium must not be overlooked, and it must not be assumed that the results will be the same under actual operating conditions. It may be a coincidence that the losses are the same in both cases. Under operating condi- tions, the spaces between the wheel blades are filled with steam flowing from the nozzle over the blades and then to the condenser. Now it has been shown by a series of experiments by Laschef of the Allgemeine Electricitat Gesellschaft (Berlin) that increasing the number of nozzles around the turbine wheel reduces the disk and blade rotation losses. These losses in the blades are very largely due to the fan action of the blades which start currents of steam just as a centrifugal fan does. In other words, this is what Stodola calls "ventilation." With steam flowing through the blades, this fan action is largely prevented and the losses are con- sequently reduced. Another reason why the disk and blade rotation losses should be less when the turbine is operating than they are in stagnant steam, is that they are really friction losses, or a conversion of kinetic energy into heat, with the effect of either superheating or drying the steam. In a turbine with more than one stage a part of the heat energy gained as the result of the friction is converted in the next expansion into kinetic energy or velocity. It is usually assumed that about 15 percent. * It may be interesting to observe that since disk and blade friction is pro- portional to the density of the medium, the friction is therefore greater in air than in dry saturated steam at atmospheric pressure. This is shown by experiments published by Lewicki in Zeit. Verein deutscher Ingenieure, March 28, 1903. | Stodola, Die Dampfturbinen, third edition, page 130. 122 THE STEAM TURBINE of the disk and blade losses are regained by the reheating, and that therefore the actual friction losses in an operating turbine are about this amount smaller than in stagnant steam. In cases of full admission true blade friction disappears ; and a proportion- ate reduction will also take place, according to the degree of admission, when it is partial. Investigation of wheel and blade friction losses by the author, using a modification of the method first suggested by Lasche of Berlin, did not show the reduction in these losses to be expected when determined under operating conditions. These results, however, cannot be considered conclusive, as the type of machine used was not well suited for the purpose, and only 25 per cent, of the blades were filled with steam. It has been stated that when a large quantity of steam passes into the casing through a suit- able opening without passing through nozzles and escapes through the exhaust (without increasing the pressure), the disk and blade rotation losses are increased as much as 20 per cent. This apparently is an influence to counteract the effect of filling the blades. In all the analysis that has preceded there are so many uncer- tain variables entering that it is impossible to get agreement, although, apparently, we have a large amount of data from which to draw. It may be stated, however, that all in all, the best data on disk and blade friction seem to show that it is smaller and of less significance than the results of most investigators would show. A little space should be given to Lasche's very interesting method.* A turbine-generator set was used in which the number of nozzles discharging into the turbine could be regulated and the output of the generator was observed for each setting of valves, and tests with varying loads were made at a number of different speeds. The turbine wheel was then removed from the shaft, and by running the generator as a motor the friction losses in the stuffing-box at the generator end of the turbine and in the bearings, as well as the windage loss of the generator, were determined. * Stodola, Die Dampfturbinen, third edition, page 131. MECHANICAL LOSSES IN TURBINES 123 The resistance of the armature and brushes was also measured to calculate the heating (Fr) loss. The sum of these losses was cal- culated for a number of loads (kilowatts) and curves similar to those in Fig. 70 were obtained. Curve A in Fig. 71 shows the elec- 340 I Ren 320 1 | E ases / 280 / i 260 240 220 jjj 200 I I 1 1 / 1 % 1 / I B A O 160 140 120 100 80 1 1 / 1 I 1 / f 1 A- Elec Output ~f 3- -fi ha to utj jut I 1 / i XI P n. 40 20 1 / ! ii 1 / 1 !j- u 7 1 2 3 4 5 6 Nozzles Open FIG. 71. FIG. 72. Curves for Determining Disk and Blade Rotation Losses at Operating Conditions. trical output at 3500 r.p.m. Curve B in the same figure, represent- ing the power delivered to the shaft by the turbine, was obtained by adding to the generator output for each set of nozzles open (curve A) the corresponding generator losses (windage, heating, and bearing friction). The lower portions of curves A and B 124 THE STEAM TURBINE are practically straight lines, and by producing curve B to the horizontal axis, its intersection represents on the scale of abscissas the disk and blade rotation losses of the turbine at the speed of the test and under actual operating conditions. By making a series of such tests at different speeds curves of. rotation losses can be made. Fig. 72 shows typical curves of shaft output for speeds of 3000, 3500, and 4000 r.p.m. Although this method requires very careful experimenting, the same must be said of any other method of obtaining these losses, and most of the results that have been published are very poor. At least it must be admitted that by this method a number of uncertain factors to be considered in the "stagnant steam" method are eliminated. The lines in Fig. 72 are really the same as " Willans lines" and might just as well be plotted for total "flow" of steam per hour as for nozzles open. In fact in turbines where there are no nozzles the "flow" of steam must be used. It is obvious that any load curve of brake horsepower giving the total steam con- sumption can be used to determine the rotation loss by producing the "flow" line to the axis on which the output is scaled. A good check on the results of such rotation loss tests is secured by observing whether the lines for the speeds near the rated speed cross each other at about the rated output. In a good design the speed-output curve will be like the curve in Fig. 80, giving nearly the same output at speeds considerably above or below the rating. The no load steam consumptions of 2000, 5000 and 9000 kilo- watt Curtis turbine-generators are respectively about 14, 12.5, and 8 per cent, of that at full load. In other words these percentages are only from one to two per cent, greater than the sum of the disk and blade rotation and generator windage losses. Generator windage loss is probably about equal to the sum of all the turbine losses. It is generally assumed that the no load steam consumption of a Parsons turbine (without the generator) is about 12 per cent, of that at the normal maximum output. MECHANICAL LOSSES IN TURBINES 125 It is stated* that at no load the steam required for very large reciprocating engines and generators is probably in no case less than 15 per cent, of that used at full load. Leakage Loss. The other important mechanical loss in a steam turbine is that due to the leakage of steam through the passages of the turbine without doing work. In impulse turbines of more than one stage this loss is chiefly caused by the leakage of steam between the shaft and the diaphragms. In a great many turbines no satisfactory packing is provided at these places and the loss is sometimes more than 10 per cent, of the total amount of steam supplied to the turbine. In reaction turbines the loss is due to leakage through the radial clearance passages and is large or small in proportion to the size of these clearances. The loss is usually assumed to be about 5 per cent, in good Parsons turbines. Future improvements in the economy of all types of steam tur- bines will depend largely on the success of designers in reducing these leakage losses. For impulse turbines an improved design has been patented by Wilkinson (page 204). In reaction tur- bines it can be reduced by making a shorter and stiffer shaft. Bearing Friction. This loss is due to the friction of the shaft in its bearings, and in a De Laval turbine the friction of the gears is usually included. An analysis of the losses in a De Laval turbine is given on page 152, where the bearing friction loss is given as one per cent. Bearing friction is also discussed in the footnote on page 92. * Kruesi, Proc. Am. Street and Interurban Rail-way Engineering Association, 1907. CHAPTER VI. METHOD FOR CORRECTING STEAM TURBINE TESTS. Standard Conditions for Steam Turbine Tests. If tests of steam turbines could always be made at some standard vacuum, super- heat, and admission pressure, then turbines of the same size and of the same type could be readily compared, and an engineer could determine without any calculations which of two turbines was more economical for at least these standard conditions. But steam turbines and engines even of the same make are not often designed for and operated at any standard conditions, so that a direct comparison of steam consumptions has usually no signifi- cance. It will be shown now how good comparisons of different tests can be made by a little calculation involving the reducing of the results obtained for varying conditions to assumed standard conditions. The method given here is that generally used by manufacturers for comparing different tests on the same turbine (a "checking" process) or on different types to determine the relative performance. To illustrate the method by an applica- tion, a comparatively simple test will first be discussed. * Practical Example. Corrections for Full Load Tests. The curve in Fig. 73 shows the steam consumption for varying loads obtained from tests of a i25-kilowatt steam turbine operating at 27.5 inches vacuum, 50 F. superheat, and 175 pounds per square inch absolute admission pressure (at the nozzles). It is desired to find the equivalent steam consump- tion at 28 inches vacuum, o F. superheat, and 165 pounds per square inch absolute admission pressure for comparison with "guarantee tests" (Fig. 74) of a steam engine of about the same capacity operating at the latter conditions of vacuum, superheat, and pressure. The manufacturers of the steam 126 METHOD FOR CORRECTING STEAM TURBINE TESTS 127 Steam Consumption Lbs.Per Kw.Hour S g 8 8 8? fe 27.5 I us. Vacuum 50F Superheat 175 Lbs.Per Sq.In.Abs. Pressure < \ \ v \ N \ ^ -* ~~~> - * _ ^ -. >* -~ - - - r ^ ^ - ) 20 40 60 80 100 120 140 .160 180 .200 Output in Kilowatts FIG. 73. Load Curve of a Typical i25-Kilowatt Steam Turbine. 35 15 28 Ins. Vacuum 0F Superheat 165 Lb&Per Sq.In.Abs. Pressure A. Steam Consumption of Engine B. Corrected Steam Consumption of Steam Turbine. 20 40 80 100 120 140 160 180 200 Output in Kilowatts FIG. 74. Comparative Load Curves of a Reciprocating Steam Engine and a Steam Turbine Both of 125 Kilowatts Capacity at Full Load 128 THE STEAM TURBINE turbine have provided the curves in Figs. 75, 76, and 77 showing the change of economy with varying vacuum, superheat, and pressure. With the help of these correction curves, the steam consumption of the turbine can be reduced to the conditions of the engine tests. Fig. 75 shows that between 27 and 28 inches yacuum a difference of one inch changes the steam consumption i.o pound. Fig. 76 shows a change of 2.0 pounds per 100 F. ^superheat, and from Fig. 77 we observe a change of 5.0 pounds :in the steam consumption for 100 pounds difference in admission pressure. Compared with the engine tests the steam turbine was operated at .5 inch lower vacuum, 50 F. higher superheat, and jo pounds higher pressure. At the conditions of the engine tests, then, the steam consumption of the steam turbine should be .reduced .5 pound to give the equivalent at 28 inches vacuum, but is increased i.o pound to correspond to o F. superheat, and ,5 pound more to bring it to 165 pounds absolute admission pressure. The full load steam consumption for the steam tur- bine at the conditions required for the comparison is, therefore, 24.5 .5 + i.o + .5, or 25.5 pounds.* Persons who are not very familiar with the method of making these corrections will be liable to make mistakes by not knowing whether a correction is to be added or subtracted. A little thinking before writing down the result should, however, prevent such errors. When the performance at a given vacuum is to be corrected to a condition of higher vacuum the correction must be subtracted because obviously the steam consumption is reduced by operating at a higher vacuum. When the steam consumption with superheated steam is to be determined in its equivalent of dry saturated steam (o superheat) the correction must be added because with lower superheat there is less heat energy in the steam and consequently there is a larger consumption. Usual * The corrected steam consumption is found to be nearly the same as that which the three correction curves show for the same conditions, that is, about 25.0 pounds. If there had been a difference of more than about 5 per cent, between the corrected steam consumption and that of the correction curves for the same conditions, the " ratio " method as explained on page 130 for fractional loads should have been used also for full load. METHOD FOR CORRECTING STEAM TURBINE TESTS 129 F.Superheat 165 Lbs.Per Sq.Iu.Abs. Ful ILoadC 125 Kw.) 20 21. 22 23 .24 25 26 .27 28 29 Vacuumrinches of Mercury , FIG. 75. Vacuum Correction Curve for a i25~Ki!owatt Steam Turbine. Steam Consumption Lbs. per Kw.Hour Z S 8 % S 28 Ins. Vacuum 165Lbs.PerSq.In.Abs. FuHLoad(125Kw.) mmm Ci __ . . . ^ - ..- ^ ^ - -. > 90 40 60 80 100 120 140 1GO 130 200 . Superheat-Degs.Fahr. "IG. 76. Superheat Correction Curve of a 125-Kilowatt Steam Turbine. 28 Ins. Vacuum 0F. Superheat Full Load ( 125 Kw.) 100 110 120 130 140 150 160 1TO 180 190 200 Steam Pressure-Lbs.Per Sq.In.Abs. FIG. 77. Pressure Correction Curve of a 125-Kilowatt Steam Turbine. 130 THE STEAM TURBINE corrections for differences in admission pressure are not large; but it is well established that the economy is improved by increasing the pressure. Corrections for Fractional Loads. It is the general experience of steam turbine manufacturers that full load correction curves, if used by the following " ratio" or percentage method, can be used for correcting fractional or over loads. This statement applies at least without appreciable error from half to one and a half load, and is the only practicable method for quarter load as well.* Stated in a few words, it is assumed then that the steam consumption at fractional loads is changed by the same percent- age, as at full load, for an inch of vacuum, a degree of super- heat, or a pound of pressure. It will now be shown how this method applies to the correction of the steam consumption of the turbine at fractional loads. Now according to the curve in Fig. 75 the steam consumption at 27.5 inches (25.6 pounds) must 9 C O obviously be multiplied by the ratio -*^- ,f of which the numer- 25.6 ator is the steam consumption at 28 inches and the denominator at 27.5 inches, to get the equivalent consumption at 28 inches vacuum. This reasoning establishes the proper method for making corrections; that is, that the base for the percentage (denominator of the fraction) must be the steam consumption at the condition to which the correction is to be applied. { Similarly the correction ratio to change the consumption at 50 F. .super- *} C O heat to o F. is~-2i., and to correct 175 pounds pressure to 165 24.0 * A very exhaustive investigation of this has been made by T. Stevens and H. M. Hobart which is reported in Engineering, March 2, 1906. f Assuming that this short length of the curve may be taken for a straight line without appreciable error. % In nearly all books touching this subject so important to the practical, consulting, or sales engineer, the alternative method of taking the steam consump- tion at the required conditions as the base for the percentage calculations is implied. By such a method percentage correction curves derived from straight lines like Figs. 76 and 77 would be straight lines and, in application, give ab- surd results. Actually such percentage corrections will fall on curves (see Figs. 87 and 88). METHOD FOR CORRECTING STEAM TURBINE TESTS 131 pounds the ratio is-^ 1 -- Data and calculated results obtained 24-3 by this method may then be tabulated as follows: Conditions of Test. Required Conditions. Correction Ratio. Percentage Correction. Vacuum inches . 27 ? 28 25.0 2 34%* Superheat, degrees F SO. o 25-6 25.0 + 4 17% Admission pressure, pounds absolute . . Net correction 175- i65 24.0 24.8 24-3 + 2.06% + -? 80% * Steps in the calculation are omitted in the table, thus ' = .0766 or 07.66 per cent., making the 25.6 correction 100 97.66, or 2.34 per cent. It may seem unreasonable to the reader that these percent- ages are calculated to three figures when the third figure of the values of steam consumption is doubtful . In practice, however, the ruling of the curve sheets must be much finer and to larger scale so that the curves can be read more accurately. The signs + and are used in the percentage column to indicate whether the correction will increase or decrease the steam consumption. " Net correction " is the algebraic sum of the quantities in the last column. The following table gives the results of applying the above "net correction" to fractional loads. Load \ Load | Load |Load f Load 31-3 kw. 62.5 kw. 93.8 kw. i 25 kw. 156.3 kw. Steam consumption from test (Fig. 73) 31.2 26.9 25.2 24-5 23-6 Net correction + 3 . 89 %. . . + 1.2 + 1.1 + 1.0 + 1.0 + .9 Corrected steam consump- tion 32 A. 28 o 26 2 2 C ? 24. ^ Curve B in Fig. 74 shows the corrected curve of steam con- sumption for the steam turbine as plotted from the above table. By thus combining, on the same curve sheet, curves A and B as in this figure, the points of better economy of the turbine are readily understood. 132 THE STEAM TURBINE Results of economy tests of the various turbines given on the preceding pages are of very little value for comparison when the steam consumptions or " water rates " are given for all sorts of con- ditions. With the assistance, however, of curves like those shown in Figs. 75, 76, and 77, if they are representative of the type and size of turbine tested, it is possible to make valuable compari- sons between two or more different turbines. Some very recent data of Curtis and Westinghouse-Parsons turbines are given below, together with suitable corrections adopted by the manu- facturers for similar machines. Steam Consumption Lbs.per Kw.Hour t-L t-i |_L |_k M M o to *- os oo o x x X x^ ^^ X X X. N, \ \ \ 20212223342526272829 Vacuum inches of Mercury FIG. 78. Typical Vacuum Correction Curve of a 5ooo-Kilowatt Impulse Turbine. The following test of a Westinghouse-Parsons turbine, rated at 7500 kilowatts, was taken at Waterside Station No. 2 of the New York Edison Co., and a comparison is made with a test of a five-stage gooo-kilowatt Curtis turbine at the Fisk Street Station of the Commonwealth Electric Company of Chicago. As no pressure correction is given for the Curtis machine, the New York Edison test is corrected to the pressure at which the other machine was operated (179 pounds per square inch gauge). Approximately an average vacuum for the two tests is taken for the standard, and 100 F. superheat is used for comparing the superheats. These assumed standard conditions make the METHOD FOR CORRECTING STEAM TURBINE TESTS 133 corrections for each turbine Comparatively small. When two tests are to be compared, by far the more intelligent results are obtained if each is corrected to the average conditions of the two tests, rather than correcting one test to the conditions of the other. There is always a chance for various errors when large corrections must be made. 7500-KILOWATT WESTINGHOUSE-P ARSONS TURBINE, WATERSIDE STATION NO. 2, NEW YORK EDISON COMPANY. Tested Sept. i, 1907. Corrected to Correction, per Cent.* 8 7CQ Average steam pressure, pounds gauge Average vacuum, inches (referred to 30 in. barom.) Average superheat degrees F 177-5 27-3 nc 7 179 28-5 IOO ~ -15 -3.36 2O Average load on generator kilowatts Vo / 08 ?o ? Steam consumption pounds per kilowatt-hour 1C . ICT Net correction per cent i.'io Corrected steam consumption, pounds per kilo- watt-hour 14.57 * The following corrections were given by the manufacturers and accepted by the purchaser as representative of this type and size of turbine: Pressure correction ., per cent, for x pound. Vacuum correction 3 .8 per cent for i inch Superheat correction 7.0 per cent, for 100 degrees F. t This is 7$ per cent, better than the manufacturer's guarantee. November, 1907, p. 658. 9000-KILOWATT CURTIS TURBINE, FISK STREET STATION, COMMON- WEALTH ELECTRIC COMPANY, CHICAGO. Tested in 1907. Corrected to Correction, per Cent.* Speed revolutions per minute 7?o Average steam pressure, pounds gauge Average vacuum, inches (referred to 30 in. barom.) Average superheat degrees F 179 29-55 116 I 79 28.5 IOO .O + 8.40 + 1.28 Average load on generator kilowatts 8070. Steam consumption pounds per kilowatt-hour . I^.O + 9 68 Corrected steam consumption, pounds per kilo- watt-hour 14.26 * The following percentage corrections were used : Superheat correction 8 per cent, for 100 F. Vacuum correction 8 per cent, for i inch from curve in Fig. Pressure correction not given . G. E. Bulletin, 7O' f -VT No. 4531. 134 THE STEAM TURBINE These results show a difference of only .20 pound in the cor- rected steam comsumption, so that for exactly the same con- ditions these two machines would probably give approximately the same economy. Each turbine is doubtless best for the special conditions for which it was designed. These results are equivalent to respectively 9.58 pounds and 9.72 pounds per indicated horsepower, assuming 97 per cent as the effi- ciency of the generator and 91 per cent as the mechanical efficiency of a large Corliss engine according to figures given by Stott.* From experience with other similar turbines it seems as if the vacuum corrections given are too low for each turbine. The correction for the Curtis turbine was obtained from the curve in Fig. 78 as given between 27 and 28 inches, while it was used between 28.5 and 29.5 inches, where the curve of steam consump- tion most likely slopes somewhat as shown by the dotted curve in the figure, which was derived from the percentage change of the theoretical steam consumption calculated from the available energy. The correction of 2.7 per cent, per inch of vacuum for the Westinghouse-Parsons turbine is probably too low also, although the percentage correction would not be nearly as large as for the Curtis. If both of these corrections are too low, the effect of increasing them would be to increase the cor- rected steam consumption of the Curtis turbine and reduce that of the Westinghouse-Parsons. Large sizes of steam turbines are also made by the Allis- Chalmers Company, but sufficient data are not given with pub- lished tests to make a comparison here. Tests of a 5ooo-kilowatt Curtis and a 75ookilowatt Westing- house-Parsons turbine are also recorded here for comparison. The two tests are corrected to the assumed standard conditions of 173.7 pounds gauge pressure, 28 inches vacuum, and o F. superheat. For the test of the Curtis machine the same per- centage corrections were used as for the 9Oookilowatt turbine; * Electric Journal, July, 1907. It is stated also in this article that the vacuum correction of a Westinghouse-Parsons turbine is 3.5 per cent, per inch betv/een 28 and 28.5 inches. Jude states that the vacuum correction for Parsons turbines is five to six per cent. METHOD FOR CORRECTING STEAM TURBINE TESTS 135 and for the test of the Westinghouse turbine the vacuum correc- tion is that given in the footnote at the bottom of page 134 (3.5 per cent, per inch), while the other percentage corrections are the same as in the preceding test of a similar machine. The West- inghouse turbine was operated with wet steam. In a test of a reciprocating engine the equivalent economy with dry steam is calculated by merely subtracting the percentage of moisture, but in a turbine test the correction is generally stated as being a little more than twice the percentage of moisture. In other words, in a turbine test the moisture must be subtracted twice. The reason for this difference in the methods of correcting water rates of engines and turbines is the very large increase in the disk and blade rotation losses in wet steam (cf. Fig. 69). 5000-KILOWATT FIVE-STAGE CURTIS TURBINE, L STREET STATION, BOSTON EDISON COMPANY. Tested Jan. 29, 1907. 1 Corrected to Correction, per Cent. Duration of test hours Speed revolutions per minute 720 177. 7 177. 7 O Average vacuum, in. (referred to 30 in. barom.) . . . Average superheat degrees F 28.8 142 28 o + 6.40 + 11.36 Average load on generators kilowatts CIQC Steam consumption pounds per kilowatt-hour . . I?. C2 + 17.76 Corrected steam consumption, pounds per kilo- I ^ 02 7500-KILOWATT WESTINGHOUSE-PARSONS TURBINE (SINGLE FLOW TYPE), INTERBOROUGH RAPID TRANSIT COMPANY, NEW YORK. Tested in 1907. Corrected to Correction, per Cent. 3 Average steam pressure pounds gauge 149-7 27.70 3-o 7135 17. 70 173-7 28 24. Average vacuum, in. (referred to 30 in. barom.) . . -1.05 6.0 Steam consumption, pounds per kilowatt-hour -9-45 Corrected steam consumption, pounds per kilo- 16. 10 136 THE STEAM TURBINE It is stated that the steam consumption of the Interborough Company's turbine is 15.87 pounds at full load and 15.54 pounds at 9000 kilowatts when the overload valve opens. The gen- erator connected to this turbine is rated at only 5500 kilowatts. With a generator more nearly the rating of the turbine it is probable still better results would be secured. Corrected tests of a 20oo-kilowatt Curtis and a looo-kilowatt Westinghouse-Parsons turbine-generator are also given here. Assumed standard conditions and corrections are taken the same as in the two tests preceding, except that the Westinghouse test is corrected to "the steam pressure of the Curtis test. 2000-KILOWATT CURTIS TURBINE, COMMONWEALTH ELECTRIC COMPANY CHICAGO. Tested May, 1905, by Sargent & Lundy. Corrected to Correction, per Cent. Duration of test hours I 2s Speed, revolutions per minute ooo Average steam pressure, pounds gauge Average vacuum, in. (referred to 30 in. barom.) . . Average superheat degrees F 166.3 28.5 2O7 166.3 28 o o + 4.0 4- 16 s? Average load on generators kilowatts 2O24 Steam consumption, pounds per kilowatt-hour I S O2 Net correction 4- 20 ? ~l Corrected steam consumption, pounds per kilo- watt-hour 18.10 1000- KILO WATT WESTINGHOUSE-PARSONS TURBINE. 1907, by S. Gilliard. Tested September, Corrected to Correction, per Cent. Duration of test hours i Speed, revolutions per minute 1800 Average steam pressure, 'pounds gauge 147. 6 166.? -I 87 Average vacuum, in. (referred to 30 in. barom.).. . \verage moisture per cent . ... 27.02 7? 28 o -3-40 i co Average load on water brake, horsepower ICO3. C Equivalent average load in kilowatts (generator efficiency QJ.%)) IOs < Steam consumption, pounds per brake horse- power-hour (wet) 1 3 61 Steam consumption, pounds per equivalent kilo- watt-hour (wet) IO 3? -6.77 Corrected steam consumption, pounds per kilo- watt-hour 18.04 METHOD FOR CORRECTING STEAM TURBINE TESTS 137 Curves in Fig. 79 are given to compare the steam consump- tion of a standard 5ooo-kilowatt turbine-generator and a 4-cyl- inder compound 5ooo-kilowatt reciprocating steam engine of \ O of it I? Is" i.8 51 8- the type used by the Interurban and Metropolitan Companies of New York, assuming both units operating under the same conditions. These curves illustrate the good overload economy THE STEAM TURBINE of the turbine, showing that at 50 per cent, overload the engine designed for equal work in the cylinders requires for the same output 43 per cent, more steam than the turbine. These results are particularly interesting because the peak capacity of a station with a given equipment of boilers and auxiliaries is increased in proportion to the reduction of steam consumption at overloads. For a given investment the turbine gives a much larger range 800 1200 Speed -H.P.M. 2000 2400 FIG. 80. Torque, Speed Output, and Efficiency Curves of a Typical 5oo-Kilowatt Steam Turbine. of load and, moreover, affords the means by which the peak capacity of existing stations can be greatly increased. The speed output curve (Fig. 80) is very useful to engineers to determine if a turbine is running at its best speed. If the cor- responding curves of steam consumption per kilowatt output (usually called water rate per kilowatt) and efficiency are calcu- lated according to the form on page 272, a great deal of informa- tion is obtained about the operation and economy of a turbine. The torque line in Fig. 80 is always drawn straight, just as a METHOD FOR CORRECTING STEAM TURBINE TESTS 139 Willans flow line. A curve of total steam consumption is usually a straight line for the normal operating limits of a turbine, but usually becomes curved when a by-pass valve opens on overload, or when the turbine is over its capacity so that the pressures are not normal in the stages. The torque line shows why a turbine engine is not adaptable to automobiles. The starting torque of a small commercial tur- bine is not large, so that starting would be difficult with a small wheel, and reversing and speed reduction would be as difficult as with a gasoline engine. The reciprocating steam engine as well as the gasoline engine has, therefore, advantages over the steam turbine for this service. CHAPTER VII. COMMERCIAL TYPES. IN some respects the order in which the commercial types of steam turbines are discussed on the following pages is somewhat arbitrary; but, essentially, it is in the order of relative simplicity. De Laval and Parsons, of the modern designers, were first in the field. They were in fact pioneers in the development of commercial steam turbines, and other designers have* followed more or less in their steps. The reasons for giving precedence to the types which they developed are therefore obvious, and no other explanation is needed. Because of its greater simplicity the commercial De Laval is first discussed, and is followed with descriptions of the various forms of the Parsons turbine and the more recent types. DE LAVAL STEAM TURBINE. Rational engineering development is nowhere better exempli- fied than in the successful performance of the De Laval steam turbine. In nearly every respect, even to details, it is still prac- tically the same as the turbines designed under the personal direction of De Laval. The essential elements of this turbine are: (i) the nozzles in which the steam expands; (2) awheel or disk with suitable blades on its periphery; (3 ) a slender shaft on which the wheel is mounted; and (4) a set of reducing gears to change the high speed of the turbine shaft to a lower speed adaptable for driving machinery. Drawings of a small De Laval turbine are shown in Fig. 82. The turbine wheel, W, is supported upon the flexible shaft between the bearing, Z, provided with a spherical seat, and the gland or stuffing-box, P. Teeth are cut into the metal of the turbine shaft to make the pinions on each side of K fit the gear 140 COMMERCIAL TYPES 141 wheels A and B, from which the power is transmitted. The design shown here is intended for driving two electric generators which are direct-connected by means of the couplings shown at the left in the figure. De Laval turbine-generator sets of from 50 horsepower up- wards are supplied with two gear wheels, two power shafts, and two dynamos for each turbine wheel, while the smaller sizes o,veN T COUPLING C / I 'omviNtt > COUPLING FIG. 82. Section of a De Laval Turbine with Two Power Shafts (on Gear Wheels A and B). have gear arrangements for a single generator. Because of the higher speed at which the small sizes operate (see page 144), making the pressure on the gear teeth considerably smaller than with the larger sizes, more power can be transmitted with a single set of gears. The large size of the gear wheels compared with the turbine is a noticeable feature of these turbines. Turbine Wheel. On account of the very high speeds at which these turbines operate, the wheels or disks require very careful designing. In the small and medium sizes, a wheel similar to 142 THE STEAM TURBINE the drawing in Fig. 83 is used. When this design is used, the hub of the wheel is bored out and a thin steel bushing is drawn into it by means of a nut shown in the figure at the right-hand end. Before this bushing is put into the wheel, it is forced on the shaft FIG. 83 De Laval Turbine Wheel with a Hole at the Center and Details of the Blades. and pinned in place as shown. The wheel can be removed from the shaft by taking off the nut and drawing it from the bushing. The strength of a disk, or a wheel of a disk type, in which there is a hole at the center is at best not more than half as strong as one FlG. 84. De Laval Turbine Wheel without a Hole at the Center. without a hole.* On this account in the larger sizes of De Laval turbines it has been found necessary to use the design shown in Fig. 84. In this arrangement a solid disk is permitted. The hub is recessed at each end, and the flexible shaft is made * An explanation of this remarkable phenomenon is given on page 333. COMMERCIAL TYPES 143 with enlarged flanged ends which fit into the recesses and are bolted solidly in place. The recesses and flanges are machined with a four per cent, taper in order that the parts maybe accurately centered and fitted. This form of wheel disk with the section increasing from the rim towards the hub is arrived at by proportioning it to have equal unit stresses throughout. But this condition does not hold true at the rim, where just below the blades annular grooves are turned on each side. Weakening of the wheel at the rim is a very good method of providing for abnormal stresses that result in case of a failure of the governor to control the speed. The purpose in making these grooves is to have the wheel burst at this reduced section where the stresses per unit of area are about 50 per cent, larger than at any other part of the wheel, rather than near the center where the damage from failure would be so much greater. At normal speed the factor of safety, at this smallest section, is about five, and since the unit stresses vary as the square of the speed,* the wheel will fail at this place at a little more than twice the rated speed. As these wheels are constructed, no great damage to the turbine will result, therefore, from the failure of the wheel rim. It has been shown by actual experiments with such wheels that when failure occurs, the rim holding the blades is broken up into very small pieces which will not damage the wheel case. It is stated, however, that wheels without this reduced section, when tested to failure, have been broken up into two or three large pieces by bursting through the center, and these pieces have been driven through an experi- mental wheel casing made of two-inch steel castings. There is also another consideration that is especially interesting to engineers. When a portion of the rim breaks off the wheel becomes unbalanced, and as the clearance between the heavy hub of the wheel and the safety bearings in the surrounding casing is very small, as can be seen in Fig. 82, the flexibility of the IFF 2 * Centrifugal force = - (see page 314) and is therefore proportional to the square of velocity (speed). The factor of safety at other sections of a De Laval wheel is about eight. 144 THE STEAM TURBINE shaft will permit the hub of the wheel to come into contact with the circular openings in the casing into which it extends. The friction of these surfaces will act as a brake and assists in bringing the wheel to rest. And this is easily accomplished, because with the blades removed the steam no longer acts to rotate the wheel. The diameters of the wheels are relatively small, as can be seen from the following table: Horsepower . 5" 30 IOO ?oo Revolutions per minute 30,000 20,000 13,000 IO OOO Diameter to center of blades, inches. . . 3-94 8.86 19.68 29.92 Wheels for De Laval turbines are usually made of a special forged nickel steel said to be rather high in carbon. Nozzles. Fig. 85 is a typical illustration of a 20-kilowatt De Laval turbine-generator and gives a general idea of how the nozzles which direct the steam against the blades are arranged around the periphery of the turbine wheel. They are attached to the turbine mechanically by being fitted into the circumfer- ence of the steel casting which serves as the casing for the wheel. The number of nozzles varies according to the size of the turbine. The nozzles are provided with hand valves, which can be seen in the figure, by which they can be closed when the turbine is running at light loads. In this way some of the nozzles are "cut out" and a relatively high efficiency is obtained at light loads. In this particular case, about half of the openings in the casing for nozzles are closed by plugs; but by removing these plugs and inserting nozzles instead, the capacity of the turbine would be greatly increased. The nozzles are the only parts of a De Laval turbine that are changed to make it suitable for any particular pressure, degree of superheat, or vacuum. The ratio of the admission (usually boiler) pressure to the exhaust pressure is the most im- portant factor influencing the design of a nozzle. Briefly stated this ratio of pressures determines the areas of the cross-section of the nozzle at the throat and at the mouth, and therefore its divergence or taper. COMMERCIAL TYPES 145 For the same output more steam is required at a low pressure than at a higher pressure. De Laval turbines are readily adjusted for a change of boiler pressure by adding more nozzles if they are needed. Sometimes turbines are fitted with two sets of nozzles, one suitable for condensing and the other for non-condensing operation. 146 THE STEAM TURBINE Reamers are used to produce the required taper on the inside of these nozzles. In the works at Trenton over 600 reamers are kept in the tool room. The taper of the nozzle ranges from six to twelve degrees, and the clearance between the mouth of the nozzle and the blades (axial clearance) is about an eighth of an inch. Blades. De Laval blades are made of drop-forged steel and have bulb shanks which are fitted into suitable slots in the wheel, shown in Fig. 83, which are milled across the rim and then drilled. The blades are lightly calked to secure them in place. At the upper ends of the blades they are provided with "extensions" which are designed to make adjoining blades fit closely and thus form a continuous ring over the blades at the periphery of the wheel. Details of these blades are shown more clearly in Fig. 64. Shaft. De Laval steam turbines have two important features distinguishing them from all other types. The first is the long diverging nozzle with the hand wheel control already mentioned; and the second is the slender flexible shaft * of the turbine. A wheel revolving at a very high speed tends to rotate about its center of gravity. If it is mounted on a stiff, unyielding shaft, of which the axis does not pass through the center of gravity of the wheel, this tendency causes violent vibrations of the wheel and shaft due to the very large centrifugal forces. It is stated that a weight of one ounce attached at the circumference of the wheel of a 3oo-horsepower De Laval turbine will produce an unbalanced centrifugal force of nearly 2000 pounds. It is mechanically difficult and almost impossible to construct a wheel so perfectly balanced that its center of gravity will exactly coincide with the geometric center of the shaft on which it is mounted. De Laval, therefore, devised a long, slender shaft which, as the speed of the wheel increases, yields somewhat and allows the latter to assume its own position of rotation about its center of gravity. * The diameter of the shaft of a loo-horsepower De Laval turbine is i inch and of a 30ohorsepower turbine is about i^ inches. COMMERCIAL TYPES 147 The wheel is not mounted midway between the bearings but considerably nearer the spherical seated bearing Z, Fig. 82, at the governor end. Now when the wheel is started up from rest, if its center of gravity is not precisely in the axis of the shaft, it will bend, and the plane of revolution of the wheel is then no longer perpendicular to the axis of rotation. When, however, a sufficiently high speed is reached, so that gyroscopic action is great enough to pull this plane back to a position perpendicular to the axis of rotation, a "node" is formed at the center of the hub and rotation will then take place about the center of gravity of the system. The speed at which the amplitude of vibration is greatest is called critical.* Bearings. Typical bearings of De Laval turbines are illus- trated in the section drawings in Fig. 82. At the right-hand or "governor" end there is a spherical seated bearing (Z). A design of this kind is used for the purpose, primarily, of giving greater flexibility to the shaft and to take the small end thrust exerted on the wheel by the steam issuing from the nozzles at a very high velocity. In single wheel turbines of the De Laval type this pressure or thrust is, however, very slight, as the steam is expanded to the exhaust pressure before it leaves the nozzles. It is obvious, therefore, that the wheel rotates in steam of very nearly the same pressure on both of its sides. Such a design has also the advantage of being self-aligning. A helical spring shown in the same figure holds the spherical bearing against its seat in the turbine casing. On the other side of the turbine wheel the shaft passes through a loose-fitting bearing, P, serving primarily as a gland or stuffing-box to prevent the leakage of steam from the casing. The shaft does not pass through the casing on the right-hand side, so that no precautions are necessary to prevent leakage of steam on that side. At each side of the pinions of the reduction gearing, the turbine shaft is supported on plain white-metal (Babbitt) bearings C and CC. The sur- * "Critical speed" is the name given to that speed of a wheel at which it tends to rotate about its own center of gravity. In the De Laval turbines it occurs at about ^ to | of the normal running speed. 148 THE STEAM TURBINE face speed in these bearings is usually designed to be about 70 feet per second. bO 3 ^ 'O l-l (U c o O ?H ^ I I I ^a i i o Speed-reduction Gears. On account of the high speed of the turbine shaft, reduction gears are required to bring the speed within practicable limits for utilizing the power. The reduction COMMERCIAL TYPES 149 is usually about ten to one, anji is accomplished by means of small pinions on the turbine shaft meshing with steel helical gear wheels. The teeth of the pinions are very small and are cut directly into an enlarged section of the flexible shaft.* The teeth for this gearing are cut spirally at an angle of 45 degrees. As indicated in Fig. 86 the teeth on one side are cut on a right- hand and on the other side on a left-hand spiral. This method effectually prevents any movement of the shaft in the direction of the axis and balances the thrust of the gears. Previous to the time when De Laval demonstrated that gears could be oper- ated at a linear velocity of more than 100 feet per second, the high speeds which he introduced were not considered practically possible. His success. at these high speeds was due largely to the fine pitch f and spiral angle of the teeth. It is thus possible to bring a large number of teeth into mesh at the same time, so that the working pressure on each tooth is made very small and abrasion is reduced to a minimum. The reduction gears are enclosed in a casing entirely separate from that around the turbine wheel. This casing prevents dust and grit from getting into the gears and avoids accidents from persons or objects falling upon them. With careful attention these gears sometimes run for several years without visible wear. Formerly the gear wheels were made of bronze, but experience showed that the teeth became crystallized after a few years of operation, and pieces of the teeth which were sometimes broken off, were liable to injure other teeth. Such gears should always be supplied with a little oil for lubrication. This speed-reduction gearing introduces two important dis- advantages: first, the friction loss is considerable; and second, the construction is necessarily expensive. The friction loss, obviously, will depend largely on the quality of workman- ship. It is stated that this loss in the gears is about 5 per * The pinions are said to be made of .60 to .70 carbon steel, and the teeth of the larger gear wheels are cut in .20 carbon steel of a grade similar to that used for locomotive wheel tires. t The pitch of the gears varies from .15 inch in the smallest, to .26 inch in the largest sizes. ISO THE STEAM TURBINE cent.* of the power transmitted when they are in good con- dition, and sometimes as much as 10 per cent, in moderately worn gears. After a few years of service it is usually found that the steam consumption of a De Laval turbine is slightly greater than when it was new. This poorer economy is probably due to the increased loss in the gears from wear as well as to the wearing away of the .30 20 40 80 100 120 140 Superheat-Degs. Fahr. 160 180 200 FIG. 87. Percentage Curve for Correcting De Lava! Turbine Tests for Superheat. blades on the turbine wheel, which by changing the shape of the blades causes a loss of efficiency. Governor. The De Laval governor is shown in Figs. 162 and 163, page 220, where methods of governing are discussed. The valve arrangement controlled by the governor is a plain throttling type. * Regarding these losses the results of experimenters differ a great deal. Lewicki found the gearing and bearing loss in a 3o-horsepower De Laval turbine-generator to be 7.5 per cent, of the full load output. Delaporte states that the gearing losses of a 200-horsepower De Laval turbine are about i per cent, when new; and he states also that in his opinion the combined gearing and bearing friction losses of a 300- horsepower De Laval turbine should be taken roughly at about 3 per cent. COMMERCIAL TYPES !$! Superheat, Vacuum, and Economy Curves. Fig. 87 shows by percentages the effect of superheat on the steam consumption. For low values of superheat the gain for a De Laval turbine is much greater than for larger amounts of superheat. Such curves on a percentage basis are sometimes very serviceable to show strik- ing variations clearly. Fig. 88 is a similar percentage curve to show how the vacuum influences the steam consumption. With a high vacuum the improvement in economy is much more 10 12 14 16 18 Vacuum in Inches of Mercury 24 26 FIG. 88. Percentage Curve for Correcting De Laval Turbine Tests for Vacuum. marked than at low values. Fig. 89 shows approximately the steam consumption for any size of De Laval turbine-generator operating non-condensing or with 28 inches vacuum at 165 pounds per square inch absolute pressure, and o degrees F. superheat. It is stated that the half load steam consumption of a De Laval turbine is 12 percent, greater than the full load value, and that at quarter load it is 25 per cent, more than that at full load. For such good performance at light loads it is necessary to operate the turbine with no more valves open than are needed. THE STEAM-TURBINE Because the valves must' be operated by hand such good economy could probably not be obtained with a rapidly fluctuating load. 30 10 GO 80 100 120 140 160 180 300 230 240 Hated Full Load Output Kw. FIG. 89. Approximate Steam Consumption of any Size of De Laval Turbine-generators. Dry Saturated Steam at 165 Pounds pej Square Inch Absolute Pressure. Turbine Losses. The following table shows how the losses in a De Laval 2oo-kilowatt turbine-generator have been divided up by Stevens and Hobart: Nozzle losses 12 per cent. Radiation losses and leakage i " " Rotation losses due to the turbine wheel revolving in steam Losses due to the steam traveling over the blades Bearing friction losses Losses in speed-reduction gearing Generator losses Losses due to residual kinetic energy in the steam passing to the condenser 8 Electrical output 59 Total.. ioo COMMERCIAL TYPES 153 PARSONS' TURBINE The Parsons type of steam turbine differs from that commonly known as De Laval's principally in the substitution of stationary blades in the place of nozzles. These stationary blades are so shaped as to direct the steam upon the moving blades just as nozzles would. In turbines of this type a large number of rows of moving blades are employed, which are attached to the cylin- drical surface of a revolving drum, called a rotor. There is also another difference which, from a theoretical view- point, makes a Parsons turbine entirely different from other types. All the impulse turbines, of which the De Laval is a good example, make very little, if any, provision for the expansion of the steam in the moving blades, while the Parsons type is designed to give approximately as much expansion of the steam in the moving as in the stationary or " guide" blades. In turbines of this type each set of one row of moving and one row of stationary blades is called, technically, a stage. Compared with the De Laval turbine in which the blades of a single wheel revolve in a medium of uniformly low density with the pressure very nearly the same on both sides of the wheel, most of the blades of a Parsons turbine revolve in steam of high density. Blades at the admission end revolve in steam at very nearly the boiler pressure, and only those at the low-pressure end' are in steam of low density. In the Parsons turbine, because of the large number of blades, many of which revolve in steam of very high density, the disk and blade rotation losses are very much larger for a given periphe- ral speed than in a De Laval turbine. Also because there is a considerable drop in pressure in every row of blades, and con- sequently a difference in pressure between the two sides of every row, there is always a leakage of steam over the edges of the blades, increasing with the amount of radial clearance between the stationary and moving parts. It is a matter of the greatest importance, therefore, in designing turbines of the Parsons type to make radial clearances as small as possible, consistent with 154 THE STEAM TURBINE proper allowances for the expansion due to unequal heating of the parts,* which in a turbine with a large number of stages is a very important consideration. Fig. 91 is a section of a typical Parsons rotor and casing showing by arrows the leakage spaces for steam through the radial blade clearances a and b. FIG. 91. Section of a Typical Parsons Rotor and Casing Showing the Radial Blade Clearances. \ A section of one of the simplest Parsons turbines is illustrated 1 in Fig. 92. The turbine rotor consists of a long drum of three different sections supported on the two bearings one at each end. The moving blades are mounted on the circumference of * Aside from the question of radial clearance, all other points affecting the design are of minor importance as regards economical and satisfactory operation. The most successful design of a Parsons type is the one which operates successfully with the smallest radial clearances. Unequal expansion of the different parts of the casing and drum introduces factors which are very difficult to estimate. If the blades are made of different materials from the drum, at some temperatures they are likely to be loose. COMMERCIAL TYPES 155 156 THE STEAM TURBINE this drum and the stationary blades are fitted in similar rings to the inside of the turbine casing. The annular space A is the steam chest which receives high- pressure steam. The steam passes through the alternate rows of moving and stationary blades of the first section of the rotor, through a second annular space to the blades of the second sec- tion which discharge into a still larger annular space, from which it passes through the blades of the last section of the rotor to the exhaust B. At the second and third annular spaces, where the diameter of the drum is increased, an unbalanced pressure or thrust toward the right is produced by the pressure of the steam; and this thrust is increased by the expansion of the steam in the moving blades (see Fig. 34). To balance this axial pres- sure, three balance pistons are provided at the left-hand end of the casing one for each section of the rotor. The smallest balance piston is made just large enough to equilibrate the thrust due to the blades of the first section; the intermediate piston balances the thrust on the second annular area and that due to the blades of the second section; and the largest piston equili- brates the pressure on the third annular area and the thrust in the third section. Steam passages are cored out in the casing, as shown in the figure, to make each balance piston communi- cate with its corresponding section of the rotor, so that the pres- sure in the section is always the same as that acting on the corresponding balance piston. In some designs these cored-out passages are replaced by pipes on the outside of the casing. Small annular grooves are usually cut in the balance pistons to join with similar annular projections in the casing. This con- struction, called a labyrinth packing, makes the steam path so devious as to effectually prevent undue leakage of steam around the balance pistons. Oil is supplied to the bearings under pressure by the small pump shown in the figure. The position of the moving blades with respect to the stationary blades (axial clearance) is usually adjusted by means of a thrust or adjustment bearing at the extreme left-hand end of the tur- COMMERCIAL TYPES 157 bine. It consists of a number of rings or collars turned in the steel shaft into which corresponding brass rings in the adjust- ment bearing are fitted. The upper and lower halves of this bearing are adjustable and are moved by the screws shown in the figure. If the lower half of the bearing is set so that the collars on the shaft are in contact on their left side, the upper FIG. 93. Propeiter of a Water-packed Gland of a Westinghouse-Parsons Turbine. half would have the collars in contact on the right side. By this means, when the bearing is once set, the rotor cannot move an appreciable distance either to the right or to the left. A typical adjustment bearing is shown more clearly at the right in Fig. 108. In this design the upper and lower halves are moved by micrometer screws, so that the axial position of the rotor is indicated at all times by the dials on these adjusting screws. In Fig. 92 a very common method for operating the governor 158 THE STEAM TURBINE of steam turbines is illustrated. A worm gear on the main tur- bine shaft engages with a gear wheel which by means of other gears rotates the governor shaft. Packing Glands. In every turbine, glands or stuffing-boxes must be provided where the shaft passes through the ends of the casing to prevent the escape of steam at the high-pressure end and the entrance of air at the low-pressure end of condensing turbines. Steam-packed glands of various types are often provided ; but in the Westinghouse-Parsons turbine water-packed glands are now generally used. This arrangement consists of the propeller of a centrifugal pump (Fig. 93) which rotates in the water supplied to an annular groove in the casing. When 1 Stationary Blades 2 Moving Blades f Stationary Blades Moving Blades FIG. 94. Typical Blading of a Parsons Turbine. the turbine is operating the water is thrown outward by the vanes and completely fills the space around the periphery of the pro- peller. By this means the leakage of steam or air is effectually prevented. As there are no rubbing surfaces in these glands and no oil is used, there is no contamination of the exhaust steam. Blades. The shape and relative position of the moving and stationary blades in a Parsons turbine are shown clearly in Fig. 94. Stationary blades are shown by cross-hatched sections, and moving blades by shaded sections. The blades of Westinghouse turbines are secured to the rotor by means of slots turned on its periphery, which are slightly COMMERCIAL TYPES 159 narrower at the top than at the 'bottom. Into these slots the blades which have been cut at the roots to fit, are put singly. Soft metal spacing pieces of the required shape to fill the space in the slot between the blades are calked to hold the blades firmly by a dovetail construction. This construction is required for the attachment of the moving blades to give the necessary sup- port against centrifugal forces; but as the stationary blades, which are fixed to the inside of the casing, are not subjected to centrifugal forces, the slots for these blades are not usually dovetailed. FIG. 95. Blades on the Rotor of a Westinghouse Turbine. Blade Lashing and Shroud Rings. It has been found necessary to bind the blades together at their ends to make a stronger con- struction. In the earlier designs of Parsons turbines the blades were usually bound together with wires soldered to their ends. Some- times, however, the blades were turned over at their outer ends to form flanges which were soldered together into a solid shroud. Fig. 95 shows several rows of blades of a Westinghouse turbine. i6o THE STEAM TURBINE All blades more than two inches long are reenforced by lashing with a wire of special section threaded through punched holes in the ends of the blades. This method of lashing is illustrated Sfl/Y!t*LCS OF THE. LR5HWG FIG. 96. Method of Lashing Westinghouse Blades. by Fig. 96. The lashing wire, which is drawn to have a cross- section resembling a comma, binds the blades together firmly enough to give adequate strength for normal service, yet, unlike a very rigid blade construction, it will yield in emergencies without COMMERCIAL TYPES l6l seriously damaging other parts of the turbine. The blades are lashed in sections three feet long. Because of the peculiar shape of the section of the lashing wire, it can be calked at the end so that a "key" remains in the punched hole to prevent the blade x \ ALLIS- CHALMERS CO PATENTED xx FIG. 97. Sankey's Bidding for Parsons Turbines. from getting out of line. In many respects it is practically as effective as a shroud ring. A type of blading for Parsons turbines, patented by H. R. Sankey in 1903, has been applied with certain modifications in the Allis-Chalmers and the Willans turbines. A typical illus- tration of this blading is shown in Fig. 97. It is distinguished 1 62 THE STEAM TURBINE principally from the usual Parsons blading by the attachment of a U-shaped shroud ring, B, around both the moving and the stationary blades. The blades are cut to the required length from bars of copper alloy drawn, like wire, to a suitable shape. After the blades are cut from the bar, they are formed in machine tools of special design, so that at the root they have an angular "dove- tail" shape as illustrated in the figure, where the blades are shown inserted in a suitable foundation ring, A. After this foundation ring is turned to the proper diameter, " dovetail" slots for the blades (see Fig. 98) are cut by a special milling machine intended for very accurate spacing and inclination so as to give the required pitch and angle to the blades. FIG. 98. Spacing for Sankey's Blading. After the roots of the blades have been inserted in the founda- tion rings, which, in cross-section, are also of a dovetail shape, the rings are inserted into corresponding grooves in the drums of the rotor and in the inside of the casing where they are held in place by "key pieces." Each of these "key pieces" after being driven into place is upset in an undercut groove which serves as a locking device. The dovetail shapes used in this construction make the attachment of the blades at their roots very secure. The channel-shaped shroud rings are purposely made thin at the flanges so that in case of contact between the revolving and stationary parts these flanges will be worn off at their edges without tearing out or bending the blades. By this method, as well as with all other types of shroud ring construction, the strength of the blading depends, not on the strength of a single blade, but on the total strength of as many blades as are bound together. In the Allis-Chalmers turbines all the blades in a semi-circum- COMMERCIAL TYPES 163 ference are joined by a shroud ring. The blading is thus made up in half rings, which are made almost entirely by machinery. FIG. 99. Interior of an Allis-Chalmers Turbine Casing, Showing Blades Pro- tected by Shroud Rings. Each ring can be thoroughly inspected before being placed in the turbine and the possible inaccuracies of hand work are likely to 164 THE STEAM TURBINE be eliminated. Fig. 99 shows the interior of the casing of a turbine fitted with shroud rings on the blades. If small radial clearances are desired, exceptional precautions in designing must be taken to avoid unequal expansions of the parts of the rotor, the casing, and the blades, because shroud rings in reaction turbines are liable to produce disastrous results by "stripping" the blades. Usually in case of accident, however, damaged or worn rings can be removed and the turbine continued in operation until they can be replaced. FIG. 100. A Westinghouse High Speed Flexible Bearing. Bearings. In turbines of the Parsons type operating at above 1800 revolutions per minute, a design of flexible bearing (Fig. 100) is used to reduce the vibrations of the shaft by permitting the rotor, when passing its critical speed, to revolve about its center of gravity instead of its geometric axis. This flexible bearing consists of a nest (usually four) of loosely fitting cylindrical bronze sleeves between which oil films are maintained by capil- COMMERCIAL TYPES 165 lary attraction.* The clearance between these sleeves is about .004 inch. These films of oil^ have also a cushioning effect in absorbing vibrations that occur when bringing the turbine up to speed. This flexible bearing accomplishes the same purpose for which De Laval used a flexible shaft. In the figure the outer casing of the bearing is at the right-hand side and the holder for the Babbitt metal lining and the cylindrical sleeves around it are shown at the left. In larger machines which run at lower speeds, balancing is less difficult and single spherical-seated bearings lined with Babbitt metal are used. Quadrant liners are provided for either type of bearings to accurately adjust the rotor to a central position. Stages. In this type of turbine low blade speeds are secured by using a larger number of stages. Thus in a 4oo-kilowatt Westinghouse-Parsons turbine there are 58 stages or 116 rows of blades. In such a turbine there are about 30,000 blades. It is important to notice why the pressure difference for each row of blades gradually decreases from the admission to the exhaust in such a turbine. Since there are 58 stages, if the pressure dif- ferences were made equal for a total drop in pressure of say from 175 pounds per square inch to i pound per square inch, the drop in pressure in each stage would be 3 pounds per square inch. But because the steam velocity for a given difference in pressure is very many times as great at i pound as at 175 pounds, such a division is not desirable, and instead the pressure drop is made to give nearly constant velocity in the different rows of blades. A good rule for designers is to set 150 and 450 feet per second respectively for the minimum and maximum velocities at the high-pressure end, and 500 to 600 feet per second at the low- pressure end. A large Westinghouse-Parsons turbine is shown in Fig. 101, with the upper half of the casing removed to show the rotor, * Bearing pressure in pounds per square inch times peripheral velocity of the shaft in feet per second is generally about 2500. Proc. Inst. Elec. Engrs., June, 1905. i66 THE STEAM TURBINE & J COMMERCIAL TYPES 167 blades, and balance pistons. The collars on the balance pistons which form the labyrinth packing are plainly visible. The increasing length of the blades of the third (exhaust) section is also very apparent. Besides the Westinghouse Machine Company of Pittsburg, Pa., other important manufacturers of Parsons turbines are the following : Allis-Chalmers Company, Milwaukee, Wis. C. A. Parsons & Co., Newcastle, England. Willans-Robinson Company, Rugby, England. Brown-Boveri & Co., Baden, Switzerland, and Mannheim, Germany.* British Westinghouse Company, Manchester, England. The Allis-Chalmers steam turbine is a reaction type which differs from the original Parsons machines principally in manu- facturing details intended to remove some of the operating diffi- culties of the older designs. An innovation in the design of these turbines is in the arrangement and construction of the balance pistons. In the older types of reaction turbines the three balance pistons were put at the high-pressure end of the turbine. Some- times, however, there was difficulty with this construction, as the largest or low-pressure piston in large turbines was of com- paratively large diameter, so that an inner web was required in its construction. This web sometimes tended to warp so as to bring the "dummy" or baffle rings of the labyrinth construction on these pistons into contact with those attached to the casing. To overcome this difficulty the largest balance piston has been placed at the low-pressure end of the rotor behind the last row of blades. In this location its effective area starts from a smaller inner diameter, so that the required area can be obtained with a smaller outer diameter. Fig. 102 represents diagrammatically an Allis-Chalmers * A 24,000-horsepower steam turbine has been constructed at the Mannheim works of Brown-Boveri & Co. for the Krupp steel works and blast furnace plant at Rheinhausen. It is probably the largest turbine yet ordered for stationary service. The governing and overload valve designs of Brown-Boveri & Co.'s turbines are described and discussed on pages 232 and 241. 1 68 THE STEAM TURBINE COMMERCIAL TYPES 169 o' I/O THE STEAM TURBINE design of Parsons turbine. There are three sections of the rotor H, J, and K and three corresponding balance pistons, L, M, and Z. The construction of the rotor of one of these turbines is shown in Fig. 209. Steam admission valves are shown as in the usual Parsons designs. The valve D admits steam to the high-pressure end of the turbine and is always under the direct control of the governor. The second valve, V, called the overload valve, is opened only when the turbine must be operated at overload or non-condensing when the condenser equipment is out of service (see page 243). At C the main steam pipe enters the steam-chest and the exhaust is at G. Main bear- ings are at A and B. A 55oo-kilowatt Allis-Chalmers turbine-generator is illustrated in Fig. 103. Governors and Low-Pressure Turbines. The various methods for governing Parsons turbines and the designs of low-pressure steam turbines of the Parsons type are discussed in Chapters VIII and IX. r s & 5 10 * * - -^ i __ 1000 2000 3000 4000 5000 COOO 70 Rated Full Load Output Kw. FIG. 104. Approximate Steam Consumption of Any Size of Parsons Turbine. Economy Curves. Fig. 104 shows fair average values of the steam consumption of good designs of Parsons turbines for 165 pounds per square inch absolute steam pressure, 28 inches vacuum, and o F. superheat. American Parsons turbines, until recently, were not made in smaller sizes than 400 kilowatts. Typical tests and load curves of 300, 500, and 1000 kilowatt Westinghouse-Parsons turbines are given on pages 268 and 269. The curves in Fig. 105 are based upon the results of tests of a Westinghouse-Parsons steam turbine of standard construction. COMMERCIAL TYPES I/I It is stated by the manufacturers that the performance as shown by these curves is typical of machines of this type. The diagonal lines or "Willans lines" in the figure show the total water weighed or steam condensed per hour at various loads. The curves or " water rate curves" show the variation in water, or more correctly, in steam consumption per horsepower-hour at 1000 1600 LOAD IN KILOWATTS FIG. 105. Typical Economy Curves of a looo-Kilowatt Westinghouse-Parsons Steam Turbine. various loads, that is, the "water or steam rate" of the turbine. Each " water rate curve" corresponds to a " Willans line" --the upper curve to the upper line, the lower curve to the lower line, etc. Operating conditions of these tests are: (1) Condensing saturated and superheated steam (iooF.) (2) Non-condensing saturated and superheated steam (100 p.). (3) One-quarter rated load to 100 per cent, overload. 172 THE STEAM TURBINE In the two overload tests the operation of the automatic secondary or overload valve may be observed. As before noted, it comes into action at a definite predetermined load as indicated by a bend in the water line. With the aid of this valve the best economy of the turbine is secured throughout the range of normal loading, while large overload capacity is available when desired, although at slightly decreased efficiency. When the secondary valve, however, has come fairly into action, the efficiency under- goes gradual improvement, as shown by the reversal of curvature of the curves of steam consumption. V \ Vo * Effect of Vacuum and Superheat on Steam Consumption 1500 K.W. Turbine Full Load x^ I ^N. \ 150 Ll 28 >s. Stea Inches m Pres Vacuui sure l N S^ 150Lbs.Stei Dry Satur m Pres ited St -sure am V ^ tj ^ \^ ^ 16 tt 15.5 H w 15 - 13.5 13 12.5 12 Vacuum - Inches 20 FIG. 106. 40 60 80 100 120 140 Superheat Deg. F. Curves of Steam Consumption of a i5oo-Kilowatt Westinghouse Turbine with Varying Vacuum and Superheat. A turbine designed for condensing work will not operate non- condensing with quite as good economy as if designed to exhaust against atmospheric pressure. That this economy is, however, excellent is shown by the upper pair of curves. The water rate is somewhat less than double the condensing water rate. Fig. 1 06 illustrates graphically the effect of vacuum and super- heat on the steam consumption of a i5oo-kilowatt Westinghouse turbine. The percentage change in the steam consumption is said to be about the same for all sizes. COMMERCIAL TYPES 173 THE WESTINGHOUSE "IMPULSE AND REACTION" DOUBLE- FLOW TURBINES. The double-flow principle has been adopted recently for the design of large sizes of Westinghouse turbines largely for mechan- ical reasons primarily to avoid the end thrust which is an impor- tant factor in all reaction types. In small machines, however, the double-flow principle does not have the same advantages as in the large machines. It is very obvious that the economy of two small machines is not nearly as good as one of twice the capacity. With large machines, however, the change in econ- omy is not nearly so great when the capacity is doubled. This fact is well illustrated by the curve in Fig. 104. Only for large Westinghouse turbines above $000 kilowatts capacity, therefore, are the double-flow designs used; and for the smaller sizes the regular Westinghouse-Parsons single-flow type is used as previ- ously. Fig. 107 illustrates a Westinghouse double flow turbine with an impulse element. In its essential parts this turbine consists of a set of nozzles, an impulse wheel with two velocity stages, one intermediate section, and two low-pressure sections of Parsons blading. Steam enters the turbine through an opening in the lower half of the casing, from which* it is piped directly to the nozzle block shown at the top of the figure. Steam escapes from these nozzles* at a high velocity to impinge on the impulse blades. The casing around the impulse wheel is made of suf- ficient size to permit a good distribution of the steam, so that it will enter the intermediate Parsons section evenly around the entire circumference of the rotor. After the steam has passed through the intermediate section it divides along two separate paths. One half enters the left-hand section of the low-pressure Parsons blading and the other half passes through the interior of * These nozzles are made non-expanding. It has been shown that non-expand- ing nozzles give higher efficiencies than expanding nozzles with steam at less than about 70 pounds gauge pressure. (See footnote, page 47.) The designers of these turbines have recognized that there are nozzle losses due to under-expansion in a diverging or expanding nozzle when the steam is throttled at light loads. 174 THE STEAM TURBINE the rotor shell which forms the connecting passage to the right- hand low-pressure section. Arrows indicate in the figure the passage of the steam through the shell. When the steam is dis- charged from the last rows of low-pressure blades, it passes into the exhaust pipes of which there is one at each end and then to the condenser. As there is practically no expansion in the impulse blades, these blade areas are made to increase only in proportion to the reduc- tion in steam velocity in each row of moving blades. COMMERCIAL TYPES 175 As the same pressure exists on both sides of the impulse wheel disk, this is not subjected to any end thrust, and requires no balancing. The small thrust due to the difference of pressure between the inlet and outlet of the Parsons intermediate section is accurately equilibrated by a "dummy" or balance piston, of moderate dimensions, located between the impulse wheel and the right-hand low-pressure section. The thrusts in the low- pressure sections are in opposite directions, and are therefore balanced. With these arrangements it is possible for the entire turbine to run in perfect equilibrium under all conditions of vacuum, pressure, and load. It is, of course, necessary to provide means for accurately fixing the axial position of the rotor, and for this purpose an adjustment bearing, shown at the right-hand end of the shaft in Fig. 107, is provided. It consists of a number of collars turned in the steel shaft, into which fit corresponding brass rings fixed in the adjustment blocks. The upper and lower halves of the adjustment bearing may be moved by means of micrometer screws, thus permitting the axial position of the rotor to be accurately known at all times. All double-flow cylinders are made in two parts, the upper and lower halves each being a one-piece casting. The design is symmetrical throughout, devoid of longitudinal flanges except those at the center required for bolting the two parts together. The castings are first rough-bored, after the flanges have been planed and drilled, and are then " seasoned" with high-pressure steam for a number of hours to remove any local casting stresses in the metal. They are then given the finishing cut and assembled with the boring bar running in the bearing housing so as to insure a concentric bore. Manholes are provided at each end of the cylinder to permit access for interior examination, and auxiliary relief valves are fitted in each of the manhole covers to prevent the pressure in the exhaust passages from rising to a dangerous point in case of failure of the condensing apparatus or the sticking of the atmospheric relief valve in the exhaust piping. A Y-connection, fitted with two corrugated copper expansion THE STEAM TURBINE joints located below the base of the turbine, connects the sepa- rate exhausts to the main exhaust pipe. These expansion joints provide for the desired freedom of movement of the turbine casing due to expansion and contraction. An atmospheric exhaust valve at the side of the exhaust " Y" can.be opened to permit non- condensing operation. The rotating element of the turbine is built up of five cast-steel parts, in addition to the shaft. As may be seen in Fig. 107, FIG. 108. 7500- Kilowatt Westinghouse Turbine. these are the three Parsons blading supports, the impulse section, and a dished plate. It is stated that the shaft carries its load at one-third the distance from the points of support, so that this design allows a lighter shaft than would be required for dis- tributed loading, and the consideration of deflection is practically eliminated. This built-up part of the rotor is rigidly attached COMMERCIAL TYPES 177 to the shaft only at the right-hand support, and the opposite end is fitted with a bronze bushjng surrounding the shaft, so as to permit the rotor to move axially, without appreciable resistance, under any differential expansion of shaft and rotor. The impulse section consists of a flanged cast-steel disk forced on the body carrying the intermediate Parsons blading. The flange of this disk is grooved at the base and forms the dummy piston for balancing the thrust of the intermediate Parsons section. Fig. 1 08 is a half-tone illustration of a 75oo-kilowatt Westinghouse double-flow turbine. The rotor of a 6ooo-kilowatt double-flow turbine is shown in Fig. 109. Details of the arrangement of nozzles and blades are FIG. 109. Rotor of a Westinghouse Double-Flow Turbine. shown in Fig. no. It is seen that the nozzle block is a casting quite separate from the turbine casing. As it receives steam from the governor valve this restricts the high pressure and high temperature to a comparatively small casting which is free to expand and contract with changes of temperature. A new type of shaft coupling for Westinghouse turbines is illustrated in Fig. in. Westinghouse Emergency Speed Limit. A very interesting mechanism is provided with Westinghouse turbines for shutting off the steam supply in case the governor fails to act and a dangerous speed might be attained. Details of this mechanism are shown in Figs. ii2a and nib. In its essential elements it THE STEAM TURBINE FIG. no. Westinghouse Nozzle Block, Showing Arrangement of Nozzles and Blades. FIG. in. Westinghouse Shaft Coupling. COMMERCIAL TYPES 179 consists of a " weight pin " P, placed diametrically at right angles to the axis of the shaft, in a cylindrical "body" screwed on the main turbine shaft at the high-pressure end. Centrifugal force tends to drive this pin away from the center and through the loosely fitting collar N. This force is resisted, however, by the " weight spring " shown around the pin in the figures. The strength of this spring can be adjusted by means of the COVER - TRIGGER CAM BUSHING \>ji5i} TRIGGER CAM TRIGGER CAM BUSHING) TRIGGER WEIGHT SPRING RETAINER LOCK BALANCING BLOCK PIN BALANCING BLOCK BODY WEIGHT WEIGHT SPRING RETAINER BODY LOCKING SCREW WEIGHT SPRING RETAINER LOCK SPRING VALVE LEVER VALVE LEVER PIN VALVE LEVER PLATE ., TRIGGER PLATE TRIGGER STOP TRIGGER 'ALVE BODY COVER - VALVE BODY (UPPER) - VALVE SPRING ADJUSTING SCREW VALVE SPRING RETAINER(UPPER) VALVE SPRING VALVE SPRING RETAINER (LOWER) - VALVE BODY (LOWER) .VE FIG. ii2a. Phantom View of Westinghouse Emergency Speed Limit. collar N, which is provided with a screw thread. Such adjust- ment determines the speed at which the centrifugal force over- comes the spring and forces the pin outward to engage with a trigger cam L. This cam is rigidly attached to one end of a short shaft S, which carries at its other end a trigger H. A small plate at the bottom of the valve lever C is supported normally at one end on the trigger H and at the other end on a screw provided for adjusting the spring on the auxiliary steam valve E. i8o THE STEAM TURBINE If the speed of the turbine should become higher than the limit for which the "weight spring" is set, the pin P is forced out to engage with the cam L, which in turn moves the trigger H away from the valve lever plate which it supports. In this way the valve E is opened because the tension in the spring on its spindle FIG. ii2b. Drawings of Westinghouse Emergency Speed Limit. is released. There is always high-pressure steam on the upper side of the valve E, and when it is removed from its seat this steam rushes through a pipe connecting the lower side of the valve to a small cylinder at the side of the main steam pipe. A short rod attached to a piston in this cylinder is moved by the steam pressure to strike a trigger which releases and closes the emergency valve on the main steam pipe. COMMERCIAL TYPES l8l Advantages of the Westinghouse Double-flow Type. In large capacities the following advantages are claimed for the double- flow type over the usual Parsons designs: (1) Reduction in size and weight due to higher permissible speed. (2) Almost negligible end thrust. (3 ) Blades and casing are not exposed to steam at high temper- atures. (4) Large volume per pound of steam at the admission to the first Parsons section avoids the use of very short blades. (5 ) Only one balance piston is required and this is of relatively small diameter. 1 (6) Exhaust connections are considerably reduced in size, due to divided flow. (7) The impulse element is well suited to high pressure and superheat, and by this modification the shaft length is reduced nearly 50 per cent. An exact reproduction of a section drawing of a Westinghouse double-flow low-pressure turbine rated at 1000 kilowatts is shown in Fig. 184 in Chapter IX. The following figures 1120, d, e, and f show designs used by the Westinghouse Machine Company for 25,ooo-kilowatt turbine- generators to operate at 200 pounds per square inch absolute steam pressure, 29 inches vacuum and 200 F. of superheat. Fig. ii2c shows the double-flow turbine designed for these con- ditions and operating at 1500 r.p.m. Fig. ii2d shows a similar design operating at 750 r.p.m. A tandem compound arrange- ment operating at 750 r.p.m. is shown in Fig. ii2e. In this last design it will be observed that the high-pressure portion is of the ordinary single-flow arrangement while the low-pressure end is made double-flow. The combined unit is connected to a single 25,ooo-kilowatt generator. A cross-compound turbine arrangement, Fig. ii2f, with the high-pressure portion operating at 1500 r.p.m. and connected to a i2,5oo-kilowatt generator is shown here together with a low-pressure portion which is of the double-flow arrangement and operating at 750 r.p.m. The low- i8ia THE STEAM TURBINE pressure portion is also connected to a i2,5oo-kilowatt generator. These figures show the comparison to scale of the arrangements described above. All of these designs would give very excellent economy, and the choice of the unit would depend primarily on the two factors of first cost and economy, assuming that in each case the relia- bility for continuous operations is the same. FIG. ii2c. Double-flow Reaction Turbine Designed for 1500 r.p.m. A close study of the four arrangements indicates that the double-flow turbine at 1500 r.p.m., direct connected to a single generator (Fig. H2g), is the cheapest construction. The large FIG. ii2d. Double-flow Reaction Turbine Designed for 750 r.p.m. areas required in the low-pressure stages of this turbine make high velocity and long length of blades essential, with the necessity of careful designing to properly take care of the stresses due to centrifugal force in the low-pressure end. COMMERCIAL TYPES l8lb The most economical combination of the four is the cross- compound reaction turbine with the high-pressure portion run- ning at 1500 revolutions and the low-pressure portion at 750 FIG. 1 1 2e. " Tandem-compound " Reaction Steam Turbine. revolutions. With this arrangement the highest efficiency is obtained, because the method of combining the unit into highl- and low-pressure cylinders, running at 1500 and 750 r.p.m., FIG. ii2f. " Cross-compound " Reaction Steam Turbine. gives the condition for best blading proportions throughout the turbine without departing from established standards of prac- i8ic THE STEAM TURBINE tice. The construction, however, is considerably heavier than the single unit of the double-flow type (Fig. ii2g) and is more costly to construct and install. At powers and speeds attain- able with single alternating-current units of, say, below 15,000 kilowatts' capacity, the double-flow turbine will be nearly, if not equal, to the cross-compound " straight reaction" turbine under the same operating conditions, and any difference in efficiency would probably be offset by the lower first cost of the double-flow machine. Taking into consideration both first cost and effi- ciency, between 10,000 and 40,000 kilowatts' capacity, the double- flow machine is undoubtedly the proper type of construction. FIG. ii2g. Section of Combined Impulse and Double-flow Reaction Turbine. The demand for turbine-generators is greatest between 4000 to 15,000 kilowatts maximum rated capacity, within which range the double-flow machines of the combined impulse and reaction types of blading satisfactorily meet commercial conditions both with respect to cost and efficiency. There is, however, practi- cally the same to be said regarding turbines of the Curtis type having four to five pressure stages. Below 4000 kilowatts' capacity the turbines consisting of an impulse wheel followed by single-flow reaction blading, and a " straight " single-flow reaction turbine (Fig. 102, page 168, and Fig. 1121), represent the machines best suited for average operating conditions. In most cases the former would be pre- COMMERCIAL TYPES l8ld f erred; but when the speed is to be made particularly low the preference goes to the latter. Thus with a speed of 3600 r.p.m. driving a 6o-cycle generator at, say, 500 kilowatts, the best design would be a combined impulse and reaction machine for best efficiency and lowest cost. If, however, the generator to be driven was for 25-cycle service, with an allowable maximum speed of 1500 r.p.m. and the same capacity the single-flow reaction turbine would be selected, providing reaction blading was to be used at all. In general, however, the application of the combined impulse and reaction turbine, consisting of an impulse element for the high-pressure portion and reaction blading for the low-pressure portion, is well adapted for complete expansion turbines over wide ranges of power and speed; and since the introduction of FIG. ii2h. Relative Lengths of Rotors in Two Common Types. this type, a large proportion of the firms building steam turbines have utilized this construction, either with Parsons or Rateau blading following the Curtis impulse element in the high- pressure end. The principal advantage of this type of con- struction is the shortening of the machine without very much loss in efficiency, the elimination of balancing pistons with the avoidance of the very considerable leakage of steam through them, and the securing of high economies at light loads by the application of the method of governing by "cutting out nozzles" (see page 221), now applied to Westinghouse turbines of the combined impulse and reaction type. i8ie THE STEAM TURBINE Fig. H2h shows the relative lengths of the rotors of the latest design of Westinghouse turbine with combined impulse blading and single-flow reaction blading compared with the conventional single-flow Parsons type with " balance pistons " (see page 156). Nearly 50 per cent, in length is saved, making the difficulties due to springing and expansion of the casing and rotor relatively small. Another important consideration in choosing between the " straight " single-flow reaction and the combined impulse and reaction types is that the former is generally preferred for FIG. ii2i. Single-flow Reaction Turbine with Kingsbury's Thrust Bearing. moderate superheats and pressures, while the latter is selected when the superheats and pressures are high. Another important improvement in the construction of reac- tion turbines is shown in Fig. 1121 where the Kingsbury type of thrust bearing is applied. By this means the usual " balance pistons " required for the single-flow type are eliminated. A recent design of a 2O,ooo-kilowatt " tandem " type of reac- tion turbine is shown in Fig. H2J. At the right-hand side is the COMMERCIAL TYPES l8lf FIG. H2J. Tandem Reaction Turbine. high-pressure turbine and at the left the low-pressure. This figure shows more clearly the type shown also in Fig. ii2e. There are very few applications as yet for " land " service of this arrangement, although it is common for marine service. FIG. 1 1 2k. Relative Sizes of Steam Turbine-generators, from One Kilowatt to 35,000 Kilowatts. Fig. 1 1 2k illustrates the relative sizes of steam turbine-gener- ators for capacities from one kilowatt for the smallest to 35,000 kilowatts for the largest. The turbine is shown on the right- hand side and the generator on the left. 1 82 THE STEAM TURBINE THE CURTIS TURBINE The Curtis steam turbine, of which the original patents were issued to C. G. Curtis about 1895, is manufactured by the General Electric Company at Schenectady, N.Y., and Lynn, Mass., the British Thomson-Houston Company at Rugby, England, and the Allgemeine Elektrizitats Gesellschaft at Berlin, Germany. As in the De Laval turbine, the steam is expanded in nozzles before reaching the moving blades,- but the complete expansion from the boiler to the exhaust pressure occurs in this type usually in a series of stages or steps, as the steam passes through a suc- cession of chambers, separated from each other by diaphragms. The diaphragms and blade wheels of a four-stage Curtis turbine are shown by a section drawing in Fig. 113. Each chamber or stage contains usually one disk or blade * wheel. Steam at the admission pressure enters the first set of nozzles through the port A, where it expands to the pressure in the first stage and delivers a portion of its energy to the blades in the wheel F. The steam then expands again through a second set of nozzles in the dia- phragm C leading to a still lower pressure in the second stage, where it gives up a portion of the energy remaining to a second set of blades, and so on. In the very small units but one pressure stage is usually employed, but in the larger sizes from two to five are used. The general arrangement of the nozzles and blades in a single-stage Curtis turbine was shown diagrammatically in Fig. 39. It is typical of these turbines that there are always three or more rows of blades following each set of nozzles, and at least one row is stationary. These stationary blades are technically called intermediates. There is practically no expan- sion in the stationary blades; the object of the several rows of blades is only to reduce the velocity, and for a given blade speed * The terms vane, blade, and bucket are often used interchangeably. Common practice, however, seems to apply blade to the Parsons turbine, and bucket to the Curtis, De Laval, and those of the Pelton type. In order, however, that the notation may not be confused, the term blade will be used in connection with Curtis as well as other types. COMMERCIAL TYPES I8 5 rim. A dovetailing method similar to Fig. 63 is now generally preferred to the method of inserting the blades by casting. The fixed blades, or intermediates, are also either cut or cast in seg- FIG. 115. Curtis Moving Blade Segments. ments (Fig. 116), and are fastened by bolts to the interior of the casing as shown in Fig. 57. These intermediates cover only the portion of the circumference upon which the belt of steam delivered by the nozzles can impinge. To make the blades more rigid, thin bands or shroud rings are riveted in segments to pro- jections on their ends. FIG. 116. Curtis Intermediate Blade Segments. The wheels of a four-stage Curtis turbine are shown in Fig. 117. There are two rows of blades on each wheel, so that in this design there are two velocity stages in each pressure stage. The shroud rings on each row of blades are plainly visible. Shafts and Bearings. The smaller sizes of Curtis turbines have horizontal shafts with standard bearings, as devices for flexibility are unnecessary at the speeds employed. The larger 1 86 THE STEAM TURBINE COMMERCIAL TYPES I8 7 sizes, however, are built with a vertical shaft supported on an ingenious step bearing, shown at the bottom of Fig. 118, which is supplied with oil or water under pressure, the shaft thus revolving on a film of liquid. The small disk D is attached by dowels E to the main shaft. The bearing is between the stationary plate C and the disk D. This vertical shaft arrangement is one of the special characteristics of the large sizes of Curtis turbines, and O/J Dra/n O/V dupp/y FIG. 118. Step Bearing for a Vertical Curtis Turbine. produces a very compact design. The direct-connected electric generator is mounted immediately above the turbine, as shown in Fig. 119, which is a section of a pooo-kilowatt Curtis turbine - generator. Fig. 120 is a "phantom" view of a 3oo-kilowatt Curtis turbine- generator, showing the wheels, armature, and couplings as if the turbine casing and generator frame were transparent. Curtis units are manufactured from 15 kilowatts (about 20 188 THE STEAM TURBINE FIG. 119. Section of pooo-Kilowatt Vertical Curtis Turbine-Generator. COMMERCIAL TYPES 189 FIG. 120. Phantom View of a Curtis Turbine Showing Wheels, Armature, and Couplings. FIG. 121. Ring Type of Emergency Stop. 190 THE STEAM TURBINE horsepower) at 3600 to 4000 revolutions per minute to as high as 9000 kilowatts (nearly 12,000 horsepower) at about 750 revolutions per minute, the general application being to direct- connected electric generators for power or lighting purposes. Emergency Valve. Since a steam turbine can accelerate at a rapid rate and this increase in speed is not easily perceptible, am. FIG. 122. Emergency Stop Valve. it is important that all these machines be equipped with simple speed limiting devices which are operated automatically in emer- gencies. The device shown in Fig. 121 consists of a steel ring (13) placed around the shaft between the turbine and the gener- ator. This ring, which is held in place by stud bolts (4), is placed in a slightly eccentric position, and the centrifugal force due to this unbalancing is counteracted by a helical spring (u). COMMERCIAL TYPES 191 When the speed increases, the centrifugal effort overcomes the spring and the ring moves into a still more eccentric position as indicated by the dotted lines. In this position the ring strikes a bell-crank lever, which trips, by means of a simple auxiliary mechanism and the tension rod L (Fig. 122), the throttle valve on the main steam supply pipe. The rod L is connected to the crank D, which operates to release the spring S, pulling up the gear and throwing out the hook G, which holds the valve open. When released by this emergency ring mechanism, the valve descends upon its seat with a very positive motion due to its FIG. 123. Details of "Spring Type" of Emergency Stop. own weight and the unbalanced pressure on the area of the valve stem. Fig. 123 shows a little different arrangement for tripping the valve. The free end of a spiral spring is thrown out by centrifu- gal force and strikes a bell-crank lever in very much the same way as the ring does. The emergency valve is opened by means of the hand wheel shown at the bottom of the figure. No turbine should be kept in operation unless it is known that this speed limiting device is in reliable condition. Governor. Curtis turbines are governed by a method com- monly known as "cutting out nozzles." By this method the THE STEAM TURBINE number of nozzles which are open for the discharge of steam is regulated according to the requirements of the load. This method is described and typical Curtis governors and valve gears are illustrated in Chapter VIII. FIG. 124. 25-Kilowat* Curtis Turbine-Generator. Small Turbines. Fig. 124 shows a 25-kilowatt Curtis turbine and generator suitable for lighting a factory. The whole set occupies very little space compared with that required for a reciprocating engine. The shaft, armature, and turbine wheel of this set are shown separately in. Fig. 125. One of the latest and FIG. 125. Wheel, Shaft, Armature, and Commutator of a Small Curtis Turbine. most efficient designs of blade or bucket wheels for Curtis tur- bines with two pressure stages is shown in Fig. i25a. The disks or wheels in the two stages are of the same diameter but the much greater blade length toward the low-pressure end COMMERCIAL TYPES 193 makes the actual over-all diameter at that end considerably larger than at the high-pressure end. FlG. I25a. Latest Construction of Blade Wheels of Curtis Two-stage Turbines. The most recent improvement in valve gears on Curtis tur- bines is shown on the turbine illustrated in Fig. 125!). The FIG. 1250. Horizontal Curtis Steam Turbine with Latest Steam-operated Valve Gear. 194 THE STEAM TURBINE centrifugal governor is placed at the upper end of a vertical shaft between the turbine and generator, which is driven by worm gearing from the main shaft of the turbine. The motion of the main governor is transmitted to the valve gear by means of levers and rods, which operate a small pilot valve, controlling the ad- mission of steam to a steam cylinder at the upper end of the valve mechanism. That is, the pilot valve serves to admit steam either above or below the piston in the steam cylinder. The piston rod extends into the steam chest and on this rod are mounted a series of spiders, which engage a corresponding series of annular double-seated admission valves. The spiders on the valve rod are arranged so that the valves are lifted from their seats in sequence as the rod is raised by the steam cylinder under control of the pilot valve. As each of the valves is lifted from its seat, steam is admitted from the central space within the annular valves to cored passages leading to the turbine nozzles. Each one of these passages in general communicates with two sections of the turbine nozzle. Correction Curves. Typical curves showing the variation in steam consumption of a 5oo-kilowatt Curtis turbine, due to in- creasing superheat and vacuum, are shown in Figs. 126 and 127. r;- 17 i* o II" 20 10 60 80 100 120 140 160 180 200 220 Superheat Degs. Eahr. 22 L 1" 18 17 2 N \ S \ \ \ 2 v "24 26 28 3( Vacuum in Exhaust Base FIG. 126. FIG. 127. Curves Showing the Effect of Superheat and Vacuum on the Steam Consumption of a 5oo-Kilowatt Curtis Turbine. Such curves become most useful, however, when they are re- duced to equivalent percentages like those for De Laval turbines COMMERCIAL TYPES shown on pages 150 and 151. In Chapter VI the correct method for making this transposition was explained. Steam Consumption. Fig. 128 is a curve to show approxi- ao 2000 3000 4000 RatedJFull Load Output * 5000 6000 7000 FIG. 128. Approximate Steam Consumption of Any Size of Curtis Turbine with 165 Pounds per Square Inch Absolute Pressure, 28 Inches Vacuum, and no Superheat. mately the steam consumption of any size of Curtis turbine at the rated full load. All the data for this curve were corrected by using percentage curves like those referred to above, which served to reduce the conditions of the various tests to assumed conditions of 165 pounds per square inch absolute steam pressure, 28 inches vacuum, and no superheat. To get sufficient data for this curve it was necessary to include some tests made with com- mercial loads, making its values probably a little higher than they would be if all the tests had been run with a constant load. Analysis of Losses in a Curtis Turbine. Steinmetz has calcu- lated the energy distribution in a typical two-stage Curtis tur- bine and has given the results in the form of the diagram in Fig. 129. WESTINGHOUSE IMPULSE TURBINES. Still another type of steam turbine intended particularly for small capacities has been developed by the Westinghouse Ma- chine Company, a$ illustrated in Figs. 130 and 131.* Machines of this type are suitable for a capacity as low as one kilowatt. By this construction it is possible to secure with the use of only * Turbines of this type are known abroad as " Electra " designs. THE STEAM TURBINE FIG. 129. Analysis of the Losses in a Turbine with Three Velocity Stages in Each of Two Pressure Stages. FIG. 130. Westinghouse Impulse Turbine (with one reversal). COMMERCIAL TYPES 1940 one row of moving blades an effect similar to the velocity stages in a Curtis turbine with only one row of moving blades or buckets as illustrated in Fig/ 130. This design is suitable for the pressure drop in non-condensing operation. The arrange- FiG. 131. Westinghouse Impulse Turbine (with two reversals). ment shown in Fig. 131 has two reversals of the steam and is suitable for condensing operation. FIG. 13 la. Three Sizes of Small Westinghouse Impulse Turbines. The advantages of this construction are that it is essentially simpler than the De Laval in the elimination of speed-reduction gears, and requires a very much smaller number of blades than the Curtis type. THE STEAM TURBINE COMMERCIAL TYPES 195 RATEAU TURBINES. Professor Rateau of Paris is' also a pioneer in the development of a well-known type of steam turbine. His first experiments were made with a turbine having a single impulse wheel; but he soon abandoned this type in favor of a multiple wheel construe- .-'-:.. y, FIG. 132. Diagrammatic Representation of Four Stages of a Rateau Turbine. tion. The Rateau turbine is often called " multicellular," meaning that it consists of a large number of " cells " or pressure stages of which the separating walls are diaphragms similar to those in a Curtis turbine. The principle of the Rateau turbine is illustrated by the section drawing in Fig. 132, which shows diagrammatically four stages. Essentially the 196 THE STEAM TURBINE Rateau type differs from that of Curtis in that it has a much larger number of pressure stages or "cells" but no velocity stages. There is therefore only one row of blades in each stage. Except for the fact that turbines with simple disk wheels can be operated at higher blade speeds than reaction turbines of the drum type (Parsons), making the efficient utilization of steam at higher velocities possible, the Rateau and the Parsons types would require the same number of stages. Rateau turbines have from 20 to 40 stages respectively, depending on whether they are for non-condensing or for condensing service. For given blade-speed, steam pressure, and superheat, the number of stages increases, although not proportionally, as the exhaust pressure is reduced. Nozzles and Diaphragms. Annular nozzles are set in each of the diaphragms between the stages. Because of the large num- ber of stages, the pressure drops are very small, so that the nozzles -are made with a uniform cross-section along their length; that is, they are non-expanding. To allow for the increased volume of the steam as it expands, in almost all the other types of impulse .turbines the nozzles are made with at least somewhat larger radial width for the lower pressures. In Rateau turbines, however, the same increased nozzle area is secured by increas- ing only the arc or part of the circumference occupied by the nozzles. In the last stages, then, where the entire circumference of the diaphragm is made use of, a complete annular jet results. Rateau nozzles are arranged in groups very much like the Curtis nozzle plate shown in Fig. 114. Diaphragms of several sizes of these turbines are shown in Fig. 133. Several groups of nozzles can be seen in each diaphragm. At the high-pressure end of .the turbine there are only a few groups (usually about three), but in each succeeding stage there is a greater number. Because the steam discharged from the blades is carried along a short distance by the rotation of the wheel, a portion of each group of nozzles is located a little in advance of the preceding set. COMMERCIAL TYPES 197 One of the advantages claimed by Rateau for multi-stage types over those in which the steam is admitted around the whole periphery in all the stages, is' that since the volume of the steam FIG. 133. Diaphragms of a Rateau Turbine Showing Nozzles and a Shaft Packing with "Water Grooves." at the admission end is small, the blades, in the Parsons type for example, have necessarily a small radial height, so that there is more friction due to the passage of steam than where the steam spaces are larger and the volume of the steam is large in pro- portion to the surface of the blades. 198 THE STEAM TURBINE COMMERCIAL TYPES 199 Description. Fig. 134 is a section of a 5oo-horsepower Rateau turbine of 24 stages. It will 'be observed that this turbine is divided into three sections, the high-pressure and intermediate sections being separated from the low-pressure section by a " middle" turbine bearing H. This turbine is designed to FIG. 135. A Rateau Disk. operate at 2400 revolutions per minute. Dimensions can be calculated from the scale given at the right-hand side. In this figure B is the main steam pipe, C is the throttle valve controlled by the governor, D is a blade of one of the moving disks, E is a nozzle in a diaphragm, F is a steam pipe connecting the inter- mediate and low-pressure sections, L is an auxiliary valve to 200 THE STEAM TURBINE admit high -pressure steam to " carry an overload," and G is the exhaust pipe leading to the condenser. Bushings of anti-friction metal are fitted in the diaphragms where the shaft passes through them. Wheel Disks. A typical Rateau disk is shown in Fig., 135, Details of construction are shown better, however, in Fig. 132. The disks are dished to add to their lateral strength, or to make them stiffen* In the recent designs a shroud ring is FIG, 136. Rateau Disks Assembled on the Turbine Shaft. fitted around the blades as illustrated in the figures. The blades resemble those used in De Laval and Curtis turbines except that they have a flat projection at the root which is provided to fasten them to the flange of the disk by riveting. The holes shown in the disk in Fig. 135 were drilled for balancing. Fig. 136 shows a group of Rateau disks assembled on the turbine shaft. * Diaphragms of large sizes of Curtis turbines are dished in the same way to give increased strength. See Figs. 57 and 119. COMMERCIAL TYPES 201 ''I / i /i r PL, i "3 < W) S o 1 1 202 THE STEAM TURBINE Manufacturers. Rateau turbines are constructed by the pioneers, Sautter, Harle & Co., at Paris, by the Maschinenfabrik Oerlikon in Switzerland and by many other companies in Europe. American rights are controlled by the Southwark Foundry & Machine Company of Philadelphia. Rateau designs are fre- quently used in combination with other types, as for example when Curtis blading is used for the first stage and Rateau blad- ing for the remaining stages (see example, pages 86-95) . Licenses to manufacture Rateau turbines have been issued to Goldie & McColloch Company, Gait, Ontario, Canada. A typical Rateau turbine connected to a high-lift centrifugal pump is illustrated in Fig. 137. Low-Pressure Rateau Turbines are extensively used in Europe to operate with the exhaust steam from rolling mill and mine engines. Professor Rateau has designed a steam accumulator (Fig. 184) for application in such cases where the steam supply is intermittent. It is described in Chapter IX in the discussion of low-pressure steam turbines. WILKINSON TURBINES. Rateau turbines are governed by throttling the steam pressure by means of valves controlled by the governor. Mr. James Wilkinson has invented a system of governing steam turbines (see page 238) which is intended to be equivalent to the Corliss "cut-off" governing of reciprocating engines. He has applied this method of governing, together with some other unique features, to steam turbines of the Rateau type which are made at the Corliss Engine Works, Providence, R.I. A Wilkinson turbine-generator rated at 100 kilowatts for non-condensing service (six stages) is shown by side and end sections in Figs. 140 and 141. It will be observed that in this design the diaphragms are "dished" as in Curtis turbines, while the disks are flat. The disks are made of forged steel, but the blades are bronze castings which are filed to a sharp edge on the side where the steam enters. p. 203 2O4 THE STEAM TURBINE Stage Packing. To prevent the leakage of steam between the diaphragms and the shaft (stage leakage)^ which in some impulse turbines is a considerable loss often 10 to 20 per cent. a very ingenious system of steam packing has been devised. A drawing illustrating this system is shown in Fig. 142. By this device, steam containing a large amount of condensation is dis- charged into grooved packings between the diaphragms and the shaft through ducts drilled into the hubs of the disks. This wet steam is taken from a part of the labyrinth packing at the high- pressure end of the turbine through which there is always FIG. 142. Wilkinson Labyrinth Stage Packing. some leakage of steam and is conducted in the ducts shown in the figure, which are arranged so that the steam discharged into a diaphragm packing is at a slightly higher pressure than that on either side of the diaphragm. It is probably possible in this way to practically eliminate the loss due to stage leakage. HAMILTON-HOLZWARTH TURBINES. A steam turbine called the Hamilton-Holzwarth is being developed by the Hooven, Owens, Rentschler Company of Hamilton, Ohio, which is a slight modification of the Rateau type. According to designs which have been published, this turbine is divided into two sections (high- and low-pressure) which are separated by a bearing. The principal difference COMMERCIAL TYPES 205 between this and the Rateau type is that the nozzles are arranged in complete rings around the circumference of the diaphragms in all the stages, instead of being grouped at the high-pressure end. As in the Curtis turbine, the blades and nozzles increase in radial height gradually toward the low-pressure end. The number of stages is about the same as in a Rateau turbine for the same conditions of pressure, superheat, and. vacuum, so that the nozzles are always designed to be non-expanding. This turbine has not been developed commercially,* so that it is not necessary to give other details. THE ZOELLY TURBINE. The Zoelly turbine is a modified form of the multi-stage, impulse type. It has fewer stages (about 5 to 10), and is gen- erally a much simpler design than a Rateau turbine. It repre- sents a noteworthy attempt at increasing the steam velocities in the blades; but with it results the great disadvantage that the surfaces of the numerous large wheels and blades, many of which move at high speeds in steam of high pressure, produce excessive losses due to fluid friction. This fluid friction of disks and blades increases, of course, enormously as the speed and the pressure of the steam are increased. In a Zoelly steam turbine there are a number of single impulse wheels, each rotating in a separate chamber, the walls of which are formed by stationary flat disks to which the nozzles are attached. At the high-pressure end the nozzles occupy only a portion of the periphery; but the area covered gradually in- creases till at the low-pressure end practically the whole circum- ference is covered. When there are about ten stages the pressure in no stage drops to less than .58 of that in the preceding stage, so that non-expanding nozzles are used. The blades are dove- tailed as represented by Fig. 61, and there is no shroud ring. The * One of these turbines was operated for a few days at the St. Louis exposition in 1904, but the author has heard of no other important installation. 206 THE STEAM TURBINE tops of the blades are cut off parallel to the shaft, but at the roots they are made with a considerably greater height on the dis- charge side than on the entrance side. .This is of course desir- able to allow for the loss of velocity in the blades, but it is stated that the height is made unusually large to cause the steam to flow smoothly through them without producing eddies.* In order to accommodate, in the different stages, the size of the nozzles to the expansion of the steam, the radial widths of the nozzle parts are gradually increased toward the low-pressure end. The most interesting part of the design of this turbine is, however, in the construction of the blade wheels to resist the stresses due to extraordinarily high peripheral speeds. As the blades for this turbine are made at present, they -are much longer in comparison with the size of the wheel than in any other turbine; in fact, the length of the blade is sometimes nearly one-half the radius of the wheel. These long blades are tapered off toward the outer ends in order to make them of uniform strength. The disks are made of forged steel and the blades of nickel steel which resists erosive action very effectually. This simple construction of the wheels and blades makes a great saving in weight. The large radial divergence of these long blades makes possible the use of very small angles on the discharge side of the wheel. Turbines of this type intended for condensing service are usually made in two sections each about 5 stages placed far enough apart to permit a bearing to be located to support the turbine shaft at the middle. Zoelly deserves the distinction of being the first to adopt, in ' impulse turbines, the use of blades with unequal angles at the entrance and exit sides. Simplicity in the design of the working parts is the most striking feature of these turbines. There are a number of manufacturers of Zoelly turbines in Germany and France. It is stated that a Zoelly turbine has been constructed at the Providence Engineering Works, Providence, R.I. * It is probable that there is considerable expansion of the steam in such blades. COMMERCIAL TYPES 207 PELTON AND SIMILAR BUCKET WHEEL TURBINES. Impulse turbines with bucket wheels of the Pelton type have recently received a great deal of exploitation from inventors in America. This type has probably received so much attention because the Pelton water wheel, commonly known as the " hurdy- gurdy " wheel, has proved so efficient in American water power plants where a high head is available. Professors Rateau in Paris and Stumpf and Riedler in Berlin have done a great deal of experimental work on such turbines, but they have practically abandoned them for those with blade wheels of the common axial flow type. Rateau has now adopted his famous " multicellular " type, and Riedler is engaged in developing the Curtis turbine in Germany. Sturtevant Turbine. A steam turbine has been developed by the B. F. Sturtevant Company, Boston, Mass., from designs prepared by Mr. W. E. Snow, which in the general bucket arrangement is similar to the old Riedler-Stumpf type.* This turbine was developed primarily for driving blowers, but it is, of course, equally applicable for other purposes. It is notable particularly for its extreme simplicity and strength. Fig. 143 is a good illustration of this turbine, showing the buckets on the wheel and the segments on the inside of the casing including the nozzles and the stationary " reversing " buckets. Three, four, or five of the latter are cut into the seg- ment, following each nozzle, depending on the velocity of the steam. Fig. 144 shows more clearly the arrangement of the * The unique feature of the Riedler-Stumpf turbine was in the bucket wheels, of which the Sturtevant wheel in Fig. 143 is a good illustration except that there was usually a double row of buckets on the rim of each wheel. These wheels were patented by Prof. Stumpf and developed with the assistance of Prof. Riedler. The buckets were cut into the rim of the wheel by a milling machine, and were arranged to overlap each other like the shingles of a roof, instead of being placed one in front of another as in a Pelton water wheel. Unusual attention was given to balancing the wheels, which were in the form of flat disks. Stumpf states that these disks were balanced so accurately that the center of gravity came within .004 of the diameter from the geometric center. These disks were similar to the design in Fig. 216. It is stated that such wheels were designed for a factor of safety of 5 at a rim speed of 1200 feet per second. 208 THE STEAM TURBINE nozzle with respect to the stationary buckets. The nozzle i? the nearly square opening shown next to the first bucket, count- ing from the left. As shown here the steam flow will then be FIG. 143. Sturtevant Turbine with the Wheel Removed to the Side to Show the Arrangement of the Buckets. toward the right from the nozzle into the bucket opposite it on the wheel. From this moving bucket the steam will be diverted back into the stationary bucket next to the nozzle and the steam FIG. 144. Sturtevant Nozzle and Stationary Buckets, showing Flanged Connec- tion to the Steam Chest. path continues alternately through moving and stationary buckets until the last stationary bucket has been passed, when it will escape into the casing and into the exhaust pipe. The stationary bucket shown to the left of the nozzle is called a COMMERCIAL TYPES 209 " supplementary " bucket intended to utilize the velocity of the steam escaping over the top of the first moving bucket oppo- site the nozzle. Its function is to divert this steam leakage into the moving buckets. FIG. 145. Section of Sturtevant Turbine. Fig. 145 is a sectional view of the turbine. Hand wheels are shown on valves by which the flow of steam into the nozzles can be controlled. It is thus possible to close some of the nozzles on light loads and obtain nearly as good efficiency and steam consumption as at full load. The method is the same as ex- plained for De Laval turbines on page 144. The governor of the centrifugal throttling type is shown at the extreme right-hand end of the turbine shaft. It is one of the type with weights acting on knife edges, in principle somewhat like the De Laval governor (Figs. 1 60 and 161). Like other parts of this turbine 210 THE STEAM TURBINE it is made as simple as possible, consisting of very few parts as shown in Fig. 147. The main bearings have solid linings of phosphor bronze. They are of the self-aligning, ring-oiling type. The weight on these bearings never exceeds 14 pounds per square inch of bear- ing surface. The speed of these turbines is from 1600 to 3000 revolutions per minute. These low speed limits compared with the speeds of single-stage De Laval turbines are made possible by the FIG. 147. The Parts of a Sturtevant Governor. application of the velocity stage principle in the use of the reversing buckets. Fig. 148 is an illustration of a Sturtevant turbine direct-con- nected to a ventilating fan or blower. The governor mechanism is at the left-hand end. Valves for closing nozzles to adjust the steam supply to the load, to get the best efficiency of the nozzles and blades, are shown clearly outside the casing. The deep base which the small diameter of this turbine neces- sitates, is utilized for steam chambers, to which the main admis- sion and exhaust steam piping is connected. Overhead pipes axe in this way eliminated. COMMERCIAL TYPES 211 The bucket wheel is a single forging of open hearth steel, and as the buckets are cut out of /the solid metal, a wheel of great strength is secured. Blade breakage and "striking" are elimi- nated, because if the bucket wheel should get out of line and touch the casing on its sides, the result would be merely like the rubbing together of two steel plates, which would produce no serious injury. FIG. 148. Sturtevant Turbine Direct-connected to a Blower. This turbine was designed to require the minimum amount of attention and repairs. It is stated that it can be operated continuously under ordinary conditions with no more attention then the weekly filling of the oil wells in the main bearings. It is therefore particularly well suited for driving any type of auxiliary machinery, especially such as may be located in inac- cessible places. Such turbines make operating expense and depreciation low, and it is stated by some engineers that they have operated turbines of this type for five years at a time without any expense for repairs. Kerr Turbine. An impulse turbine of the Pelton type has been patented by Mr. C. V. Kerr and is manufactured by the Kerr Turbine Company, Wellsville, N. Y. In this turbine typical Pelton double cup-shaped buckets are used into which jets of steam at high velocity are discharged from nozzles, located >\ 212 THE STEAM TURBINE as in Fig. 149, around the periphery of the wheel. The inside surface of each bucket is formed of two intersecting surfaces of FIG. 149. Kerr Bucket Wheel and Nozzles. FIG. 150. Sectional View of the Kerr Turbine. revolution, approximately ellipsoidal, somewhat like the reflector of a locomotive headlight. A section of a Kerr turbine is shown in Fig. 150. In this COMMERCIAL TYPES 213 design there are five compartments or stages, each with a single bucket wheel. In the design of this turbine provision was made for its manufacture in standarcf " unit parts." In this sense the turbine casing shown here in section, consists of steam and exhaust end castings and a number of nozzle diaphragms between the ends. In the chambers thus formed steel disks revolve, each having a row of double buckets dovetailed in the rim. By this simple arrangement it is possible to build up turbines of any size or for any pressure and vacuum by adding sections or nozzles as may be required. Steam flowing from one stage into the next is discharged into these buckets in tangential jets by the nozzles screwed into the steel nozzle bodies, which are accurately set in place and riveted into the diaphragm castings. The governor is of the centrifugal type, consisting of weights moving on knife-edges. A section of the governor weights and &g^ \v 4/^0. G. Open Position y*^C.G. Closed Position Ctr. Line of Shaft ->" Section A-A FIG. 151. Section of Kerr Governor Weights and Mounting. their mounting is illustrated in Fig. 151. The weights are sup- ported at three points. The hardened knife-edge at B is straight, and of sufficient width for the stresses on it. At 90 degrees on each side is a rolling contact at C. The curve at this point is such that the bearing between the weight and the cam collar is always on the line of centers. Pure rolling contact is thus secured, and the weight, without being fastened in position, is firmly driven by its triangular support. The outward movement of the weights compresses the governor spring and operates, through lever connections, a balanced piston valve controlling the flow oi steam. 214 THE STEAM TURBINE Fig. 152 shows a typical Kerr turbine direct-connected to an electric generator. Because of the simplicity of the design these turbines are particularly suitable for isolated lighting plants and for driving centrifugal pumps and blowers. Terry Steam Turbine. Like the Sturtevant turbine, the one invented by Mr. Edward C. Terry belongs to the Pelton impulse type in which there are two or more velocity stages. Stationary reversing buckets are arranged in groups one for each nozzle around the interior of the casing. These bucket groups are shown in Fig. 153, where a Terry turbine is shown with the upper half of the casing raised for inspection. In this illustration there are four stationary buckets for each nozzle. Obviously the steam I FIG. 152. A loo-Kilowatt Kerr Turbine-Generator. is returned to the moving buckets as many times as there are stationary buckets in each group. These stationary buckets are made of gun metal, and each has a crescent-shaped hole at the center through which the steam partially exhausts. There is, therefore, apparently considerable expansion in the moving COMMERCIAL TYPES 215 blades. A valve is provided for each nozzle, so that when it is desired some of them can be closed. Speeds of these turbines FIG. 153. Terry Turbine with the Casing Raised. vary from 2500 for a ic-horsepower size to 1600 for 3oo-horse- power. Fig. 154 is an illustration of a Terry turbine direct-connected to a five-stage high-pressure turbine pump. 216 THE STEAM TURBINE COMMERCIAL TYPES 217 Dake Steam Turbine. This turbine has stepped buckets, but the nozzles do not discharge radially. Stationary and moving buckets, with a section of the bucket wheel, are shown in Fig. 155. Steam Port Expansion Wedge Nozzle Bucket Guide Passage FIG. 156. Diagram of Nozzles and FIG. 155. Steam Passages in a Dake Buckets of a Dake Turbine Showing Turbine. Expansion Wedges. This arrangement is unique in that the steam passes through the stationary buckets b in an axial direction, and is deflected radi- ally by the two sets of buckets a and c on the wheel. Steam enters the nozzles from the steam chest S. Relative positions of nozzles and buckets are illustrated diagrammatically in Fig. 156. Actually in the turbine the horizontal lines in this figure are, of course, arcs of circles. The "steps" shown here in the wall of the bucket ring are intended to bring the surfaces upon which the steam impinges nearer to the nozzles and to present always approximately the same angle to the flow of steam. The nozzles are designed with the object of delivering the steam to the buckets in parallel jets. Throughout their lengths the nozzle walls are the same distance apart, and expansion is secured by the use of " expansion wedges," shown plainly in both figures, which are set centrally in the nozzles. These wedges can be readily removed and replaced, so that it is not difficult to insert a wedge properly proportioned to give the best expansion for a given steam pressure. THE STEAM TURBINE p a COMMERCIAL TYPES SPIRO TURBINES. The Spiro turbine (Fig. 157) consists simply of two herringbone gear wheels which mesh together and revolve in a close-fitting casing. Steam enters at the inlet pipe at the bottom and passes around the gears in its expansion to the exhaust pipe at the top. Steam discharges from the inlet pipe through the small holes, equivalent to non-expanding nozzles shown in Figs. 158 and 159, FIG. 158. The Spiro Casing or Cylinder. The two holes near the central rib inside the cylinder are the steam nozzles. into the " pockets " or spaces between adjacent gear teeth. As the rotors revolve the " tooth-space " occupied by the steam increases in length as the steam expands. Finally the steam escapes when the outer ends of the teeth pass the line of contact between the two rotors. The increased length of this " tooth- space " from the time the steam is admitted until it is exhausted is shown in Fig. 160, by the comparison of the length of the tooth-grooves at "A" with the length of the outer white lines. By having the steam inlet at the bottom the weight of the rotors is partly carried by the steam pressure and friction is much reduced below what it would be if the inlet were at the top. THE STEAM TURBINE Spiro turbines are suitable only for non-condensing operation and find application usually for driving the auxiliaries like blowers and pumps in a power plant or in office buildings where the exhaust steam is passed through feed- water heaters or is used EXHAUST INLET- FIG. 159. Section of Spiro Cylinder and Rotors at Mid-length. FIG. 160. The Rotors. for heating buildings. Under these conditions low steam con- sumption, as would be obtainable with condensing operation, is unimportant. Sufficient expansion for condensing operation is impracticable with this type of turbine. Compactness is also an important feature. The casing of such a turbine is alone no COMMERCIAL TYPES larger than the cylinder of a good high-speed engine of the same capacity and is even smaller than comparable commercial sizes of electric motors. Governing is accomplished by throttling the steam pressure by a method similar to that used for nearly all steam turbines of relatively small size (less than 100 horsepower). CHAPTER VIII. GOVERNING STEAM TURBINES. METHODS of governing steam turbines, or, in other words, of regulating the supply of steam to suit the load on the machine, may be classified as follows: 1. Throttling or partly closing the steam admission valve. 2. Varying the cross-section of the steam passages by " cutting out nozzles." 3. Varying the time of admission, or "blast" governing. 4. Admitting steam at boiler pressure at various points along the direction of steam flow, or " by-pass" governing. When the steam admission of a turbine is partly closed, the amount of steam passing through the valve ports is, of course, reduced; but at the same time the steam is throttled, meaning that the pressure is reduced without changing the heat contents in a unit weight. Although in this throttling process the total heat in a pound of steam remains unchanged, the energy avail- able from expansion is considerably reduced. If steam at 165 pounds per square inch absolute pressure which contained ini- tially 2 per cent, moisture is throttled without loss of heat, that is, without doing work, to 25 pounds per square inch absolute, the steam at this lower pressure will have 43 degrees F. of super- heat. Now if the available energy is calculated for adiabatic expansion from this lower pressure and 43 degrees F. super- heat to i pound absolute, it is found to be 207 B.T.U. The available energy for adiabatic expansion of steam at the initial condition before throttling (containing 2 per cent, moisture) to the same final pressure is, on the other hand, 316 B.T.U.* * Although the moisture is removed and the steam is superheated there is no gain to offset the loss in available energy except that the disk and blade rotation losses are reduced; but the gain from this cause could not probably exceed 10 per cent. " Drying " action then is not very important, and it will have very little influence in remedying large losses due to throttling. 218 GOVERNING STEAM TURBINES 2I 9 In this extreme case of throttling, the available energy of the steam is reduced about 35 per cent, and consequently, for the same work, approximately 35 percent, more steam is needed with throttling valves than if the steam could be used at light loads without throttling. It is not unusual for turbines governed by throttling to take steam at full load at 135 pounds pressure when the steam supplied is at 165 pounds. Then the maximum pressure becomes available only on overload just before the stage valves open. Efficiency of an expanding nozzle is considerably reduced when it is used with pressures very much different from that for which it was designed, as shown by the curve in Fig. 28. Blade efficiencies are similarly reduced when the available energies and consequently the velocities are not those for which the blades were designed. Fig. 161 shows very FIG. 161. Effect of Throttling on Steam Consumption. plainly the effect of throttling on the economy of steam turbines. The two curves in this figure show the steam consumption per electrical kilowatt for a 6oo-kilowatt turbine when operating (i) with a throttling governor, and (2) with a governor varying the steam supply by changing the area of the steam passages, that is, governing without appreciable throttling. In spite of these defects, however, governing by throttling has been fairly satis- factory. In the De Laval, Rateau, and Zoelly turbines governing is effected by throttling devices. For most turbines, the governor itself is similar to centrifugal governors used in reciprocating engine practice. 220 THE STEAM TURBINE FIG. 162. De Laval Governor and Vacuum Valve. FPG. 163. Section of the Main Admission Valve of a De Laval Turbine. GOVERNING STEAM TURBINES 221 Fig. 162 shows cross-sections of a typical De Laval governor. It consists of two half cylinders B, B which are pivoted in' a short outer casing by the knife-edge A. Inside the casing these cylin- ders are fitted with pins C, C which press on a collar D when the other ends of these cylinders (at B) are thrown out, or tend to separate, by centrifugal force. The pressure on the collar D transmitted by the pins compresses the springs and forces a central spindle G toward the right, which moves with it the bell-crank L. This bell-crank moves a short shaft which passes through the steam pipe and has attached to its other end by means of a set-screw a lever (shown in the section of the valve, Fig. 163) operating the main admission valve. The weight of the valve and levers is balanced by the small spring N. The bell-crank L has a certain "play" in M which is adjusted to make the governor not too sensitive to momentary changes in speed. The valve travel is only about one-eighth of an inch from the closed to the wide open position. The governor frame is supported on a tapering rod E which is fitted into the end of the main turbine shaft K. With condensing De Laval turbines a vacuum valve T is arranged in connection with the governor to act as an emergency stop valve. In case the turbine exceeds the allowable speed limit due to the failure of the main admission valve to operate properly, the vacuum valve admits air to the turbine exhaust pipe through the passage P. The steam consumption when operating non-condensing is so much greater than when condens- ing that it is said to be impossible to exceed the rated speed when exhausting into the atmosphere, with all the nozzles and the main admission valve open. FIG. 164. A Slide Valve Arrangement for a Turbine. Governing by " Cutting Out Nozzles." One of the simplest forms of governing is represented in Fig. 164 showing a plain 222 THE STEAM TURBINE slide valve arrangement for regulating the flow of steam through a series of nozzles. This is one of the best systems that can be employed for an impulse turbine if an elaborate valve gear is to be avoided. As the full initial pressure is always maintained in all the nozzles that are open, there can be very little throttling except when the valve is in a position so that one of the nozzles is partly covered. The loss, however, due to this amount of throttling is practically negligible for other than very light loads. Valve gears have been designed to improve on this slide valve method by providing a separate valve (usually of the poppet type) for each nozzle or, at most, for a small group of nozzles. These valves are opened and closed suddenly by the governing apparatus by the use of either springs or dash-pots, very much as with our modern Cor- liss valve gears for reciprocating engines. The difficulty with this last method is, however, that there will be abrupt, although perhaps small, variations in speed every time a valve opens or closes, unless special precautions are taken in the design. If the service is for electric lighting, speed irregularities due to such governing may be sufficient to produce a flicker in the lights. In turbines with more than one pressure stage, as, for example, in the Curtis and Rateau types, it has often been proposed to control the admission to each stage. Apparently the only objec- tion to such a scheme would be in the very complicated valve gear that would be needed; but, contrary to what one might expect, it can be shown by tests and demonstrated mathemati- cally that such an arrangement would not give as good economy as if only the first stage is controlled. The only advantage resulting from this method of controlling the steam supply is that by making the light load pressures more nearly in the same proportion to each other as for full load and overload, the stresses in the diaphragms separating the stages are more nearly the same as calculated in the original design. There is probably no commercial type of turbine using such a compli- cated method of governing except for large overloads, when econ- omy is not of importance and the conditions are more of emergency GOVERNING STEAM TURBINES 223 than of continuous operation. Governors for the larger sizes of the Curtis turbines show the merits of this method to the best advantage. By governing in this way it is possible to vary the number of valves supplying steam to the turbine in proportion to the size of the load, thus maintaining a constant initial pres- sure and therefore constant velocity in the nozzles and blades. In a single stage turbine there is no difficulty in applying this method, and consequently the energy and velocity are always those suited for the best nozzle and blade efficiency. Usually in a turbine of several stages no attempt is made to regulate the number of nozzles after the first stage, on account of the mechanical difficulties inseparable from a complicated valve gear. In order to secure a correct energy and velocity distri- bution throughout the turbine, the nozzles in all the different stages should, of course, be changed in the same ratio. This scheme is not impossible and has been attempted in some German designs. With turbines like the Rateau and Parsons, where the drop of pressure is very small in each stage, and where there are, therefore, a great many stages, any method of cutting out some of the steam passages to reduce the area at Jight loads is impracticable. Types of governors to be used depend a great deal on the capacity and the kind of service. The smaller sizes have usually simple forms, while the larger ones are necessarily more com- plicated. On the small turbines, where an elaborate valve gear is not desirable, the valves are moved by the direct action of the centrifugal force of the governor weights. This is called direct governing, to distinguish it from the " relay" system used by most turbine manufacturers for large machines. By the direct method a comparatively large centrifugal force is necessary to move the valves; and unless they are carefully balanced it is difficult to make the governor sensitive to fluctuations in the load. Besides, if for any reason a valve sticks, there may be wide variations in speed. By the indirect or what is commonly called the "relay" method the centrifugal force of the governor is needed only to 22 4 THE STEAM TURBINE 226 THE STEAM TURBINE "give the signal," as we may say, which sets in motion an auxil- iary mechanism by which the valves are moved by gearing con- nected to the main shaft or by steam or hydraulic pressure. In Curtis turbines of all sizes up to 500 kilowatts the valves are operated mechanically, and for larger sizes a hydraulic apparatus* is used. Electromagnetic Control of Valves. Formerly, in large Curtis turbines, the valves opening the nozzles were operated by the pressure of steam admitted through a port opened and closed by a " pilot" valve controlled by electromagnets. The governor was connected to a very simple mechanism for the purpose of making and breaking the current through the electromagnets, which, in turn, moved the "pilot" valves operating the main valves on the turbine. Mechanical Valve Control. One of the recent developments in the valve gears for large turbines governed by cutting out nozzles is the successful replacing of the electromagnetic "relay" outfit formerly used on Curtis turbines by a positive mechanical valve gear, due to Mr. Richard H. Rice. This valve mechanism is well illustrated in Figs. 165 and 166, where it is shown applied for regulating the steam admission to the first stage of an impulse turbine. Steam in the steam chest C is maintained constant at the pressure for which the turbine was designed, and the valves are operated so that they are always wide open or else tightly closed. When the valve rod t, Fig. 166, is raised steam is admitted through the port A, from which it passes into a nozzle plate (like Fig. 114) at B to be discharged at high velocity into the blades of the first stage. The valve gear consists essentially, besides the worm gears shown at the right-hand side of the figures, of a connecting rod moving a bell-crank 1, to which two dogs or " catches," w, w, are attached by pins. The extreme ends of these dogs, marked i will engage with the teeth on the steel plates u and v. An eccentric, h (Fig. 165), gives the connecting rod k a reciprocat- ing motion which, being transmitted to 1, moves the dogs w, w up and down. In Fig. 166 the lower dog is shown sliding on GOVERNING STEAM TURBINES 227 the plate u, and in its lowest position it touches the tooth on this plate. The upper dog is kept out of contact with the tooth on the plate v by the lever x, which by engaging with the lower end of this dog, marked 2 in the figure, raises the end i out of reach of the tooth. The letters x and s are at opposite ends of the same lever supported on the shaft m. In the top view of this valve gear shown in Fig. 165 there are five valves operated by the connecting rod k. On the same eccentric, h, there is also another similar connecting rod, j, operating five valves on the opposite side of the turbine. The steam supply of the tur- bine is therefore regulated by ten valves. The position of the end of the lever at x is regulated by means of the rod q, which is connected to the Curtis governor illustrated in Fig. 167. Speed regulation by means of this governor is accomplished by the balance maintained between the centrifugal effort of moving weights and the static forces exerted by springs. The governor is keyed to the main turbine shaft at S and, of course, rotates with it. It is protected on two sides by a stationary looped casing, of which a section is shown at the top of the figure. In the order of action of this governor the weights A fly out on account of centrifugal force, moving on knife-edges near their largest diameter, and pull down the governor rod C by the pres- sure exerted on other smaller knife-edges B. The governor rod is pulled down against the action of the heavy spring D. At E a ball-bearing gimbel joint, thoroughly lubricated, forms a junction point between the revolving shaft of the turbine and the stationary lever of the governor (shown in the figure extending toward the right, nearly horizontally).* This stationary lever is connected by means of a bell-crank to the rod q (Fig. 166) and thus determines the position of the lever x. To illustrate the action of this valve gear and the governor, assume the load on the turbine has been increased and the speed * Connected to the stationary lever of the governor is an auxiliary spring F fer varying the speed when synchronizing. By means of a small motor G the tension of this spring can be adjusted from the switchboard. 228 THE STEAM TURBINE has dropped a little, indicating that more steam is needed and that the valves have been so arranged* that the one shown in Fig. 1 66 is the next to be opened. With a reduced speed the governor weights A (Fig. 167) will move in slightly toward the center, reducing the tension on the governor spring D, so that the rod C and the left-hand end of the stationary governor lever are raised. By means of an auxiliary lever and a bell-crank the rod q is raised and the end x of the lever attached to it is lowered FIG. 167. Sectional View of Curtis Governor. to engage the catch 2 of the lower dog w, releasing at the same time the upper dog, which now comes into contact with the tooth in the plate v, raises the valve rod t, and admits steam through * Each of the levers s controlling the dogs is set at a little different angle to the horizontal. The lever which has its end (x) lowest will open its valve first. GOVERNING STEAM TURBINES 229 the port A. When again the speed becomes too high the rod q is lowered, x is raised, and the lower dog closes the valve. The dogs are held in position when not in contact with the lever x by the flat springs o. The eccentric h is moved by means of gears connected to the main turbine shaft S. A ring on this shaft has a single tooth a, which engages with a gear wheel b, on the shaft c, which by means of worm gearing is connected with another horizontal shaft f, and thus moves the eccentric h. The hub of a turbine wheel is shown at H (Fig. 166), and carbon packing rings to prevent the leakage of steam from the first stage are illustrated by D and E. The speed reduction is designed sometimes for one worm gear instead of two as shown here. Hydraulic Motor Control of Valves. The hydraulic governing device used in the designs of Curtis turbines of the 500 to 9000 kilowatts sizes is illustrated in the following figures. The movement of the horizontal governor lever shown in Fig. 167 is transmitted through the rod D (Fig. 168) to a second lever arm C operating the pilot valve of the oil cylinder B. The piston A operates the main power arm of the mechanism, which trans- mits the motion either by a rack connecting with a pinion, or by means of cranks, to the "side" rod, shown in Fig. 169, carrying the cams for operating the valves. These cams act directly on the valves, opening and closing them according to the demands of the load. Because this device has a very slow motion it has the advantage of being practically independent of lubrication for its successful operation. Governing by Varying the Time of Admission. Governing by periodic admission or by "blasts" was invented by Parsons and has been applied to practically all types of steam turbines using his name. In its ideal form steam is admitted to the turbine by a poppet valve in puffs or blasts in periods of long or short duration depending^ on the demands of the load. The method is explained usually by saying that there are alternate periods when the turbine casing is either filled with steam or there is no 230 THE STEAM TURBINE FIG. 1 68. The Hydraulic Operating Mechanism for Valves of a Curtis Turbine. steam at all. At light loads the valve opens for short periods, remaining closed the greater part of the time. When the load increases the valve remains open longer, and at about full load GOVERNING STEAM TURBINES 231 there is full pressure in the high-pressure blades, the valve merely vibrating without sensibly affecting the pressure of the steam in the passages. It is thought that in this way the full benefit of high-pressure steam can be secured at all loads. This fS 1? C/3 U II H 8 u i! is the ideal condition, but practical considerations greatly modify it. Brown-Boveri-Parsons Governing Device. The method of reg- ulating the steam supply by intermittent admissions or "blasts" is typical of nearly all the governing devices fitted to Parsons 232 THE STEAM TURBINE turbines. The design used for the Brown-Boveri-Parsons tur- bines is illustrated by Fig. 171. Steam enters the turbine through a main admission valve N which is given a vertically oscillating motion. A small piston mounted above this valve and on the same spindle, has steam at the pressure in the main steam pipe on its lower face acting against the pressure of a strong spring on its upper face. An auxiliary valve fitted on the spindle L is given an oscillating motion by an eccentric on the governor shaft at M, which causes, at every stroke, the small passage at the FIG. 171. Brown-Boveri-Parsons Governing Device. lower face of the piston to communicate with the exhaust, making the main valve N fall upon its seat. The spindle L is linked up to a collar sliding on the governor shaft. The height of the governor balls determines the position of this collar. Thus the height of the governor augments or diminishes the amplitude of the oscillations of the auxiliary valve on L, and in consequence causes the main valve N to open a longer or a shorter time at each admission of steam. The frequency of the steam admis- sions is about 150 to 250 per minute according to the speed of the turbine. GOVERNING STEAM TURBINES 233 Westinghouse-Parsons Governor and Valve Gear. Diagram- matically, the governor and valve gear of a Westinghouse- Parsons turbine are shown in Fig. 172. A small pilot valve, marked A in the figure, is actuated directly by the governor by means of levers and links. This pilot valve controls the steam supply of the turbine by regulating the operation of the main poppet admission valve which opens and closes at . uniform intervals when the turbine is in operation. Speed variations FIG. 172. Diagrammatic Arrangement of the Governing Mechanism of a Westinghouse-Parsons Turbine. change the height of the governor balls which, in turn, change the position of the collar F of the lever on the governor spindle. By means of a system of links this lever varies the throw of the pilot valve relatively to the valve port. This pilot valve controls the main admission valve by means of the auxiliary piston valve B in the same way as in the Brown-Boveri design which has already been explained. Reciprocating motion for operating the valve mechanism originates in an eccentric driven by the 234 THE STEAM TURBINE turbine from a worm on the main shaft. This eccentric gives an oscillating motion to the levers supported at D, F, and E. The governor is of the fly-ball type, the ball levers being mounted on knife-edges instead of pins, to secure sensitiveness. The speed of the turbine may be varied, while running, within the limits of the governor spring by grasping a knurled hand wheel at the top of the governor and bringing the spring and tension nuts to rest. Adjustment of the tension of the spring can then be made. This device is particularly useful for syn- chronizing the speed of small turbine-alternators operating in parallel, or for distributing the load between them. For syn- chronizing large Westinghouse turbine-alternator units a small motor controlled from the switchboard is used to adjust the governor spring. Allis-Chalmers Governor Mechanism. The governing device of the Allis-Chalmers steam turbine is of the Parsons type, using hydraulic instead of steam pressure. The governor is required to operate a small balanced oil relay valve only, while the two steam valves, main and by-pass, are controlled by oil pressures of about 20 pounds per square inch, acting upon a piston of suitable size.. The by-pass valve opens when the turbine is required to develop overload or the vacuum fails. The oil supply to the bearings and to the governor can be interconnected so that the governor will shut off the steam if the oil supply fails. " Blast " Governing Compared with Throttling. When the main steam admission valve of a Parsons turbine closes there is still some steam in the turbine casing, and this steam expands, of course, to fill the space. The same effect occurs also when the valves are first opened, and the steam rushes into a region of very low pressure. In these two ways low pressures are pro- duced just as with throttling valves, although, for the same average pressure, the loss is not nearly so great. The pressure variation for a i5oo-kilowatt Parsons turbine at one-quarter, three-quarter, and full load is shown in Fig. 173. At a little overload there is practically no variation because the GOVERNING STEAM TURBINES 235 steam valve is then closed for shorter periods. Probably the greatest disadvantage from this method of governing results from " initial condensation at light loads." There is usually one blast or puff of steam in every thirty revolutions. Steam admis- sions are, therefore, far enough apart to allow the interior of the turbine to be cooled by the falling temperature between the blasts. Now when there is a fresh admission the steam comes into contact with the relatively cooler walls of the interior of the 2570 2240 1270 Atmospheric =0 Absolute Zero Ilodgkinson, F. FIG. 173. Indicator Cards Showing Initial Pressures in a Parsons Steam Turbine. turbine, and condensation must take place just as in a recipro- cating engine. If it were possible, practically, the number of "periods" would be made so small that free expansion would be reduced to a minimum; but for a satisfactory speed regulation long periods are not permissible. It appears, therefore, that unless the periodicity can be made low, the economy at light loads is no great improvement on the method of plain throttling. A very important feature of this method, however, should not be over- looked. This is the advantage of having a valve mechanism which is constantly moving, precluding the possibility of "sticky" valves. 236 THE STEAM TURBINE The time required for the steam entrapped in the casing when the valves are closed to drop in pressure by a given amount can be calculated very simply as follows. Let W l = weight of steam (pounds) entrapped when the valves are closed, W = weight of steam (pounds) in the turbine casing after expanding a time /, W = weight of steam (pounds) flowing per second when the valves are wide open, that is, when the pressures in the casing are those for which the blading was designed, P l = initial absolute pressure of the steam delivered to the turbine, P 2 = final absolute pressure after a time /, and to avoid complex mathematical terms assume that in the expansion in the casing, in general terms, Pv = K, where v is the volume of a pound of steam and K is a constant. Then since v and W are reciprocals, W = P X constant; then also W = P, X C, where C is another constant. In a time dt we have thus dW = CPdt, also K v K. GOVERNING STEAM TURBINES 237 therefore , KC ji/! = X" and If we take for convenience 1/F' = 2 J pounds of steam per second, JF t = f pound, and P l = 165 pounds per square inch absolute pressure, the time required for the average pressure in the casing to fall to 100 pounds is 3 . 165 4 X 2.25 100 159 seconds. The time required for steam at 165 pounds absolute to fall to various other pressures is shown in Fig. 174. 200 .5 Time-Seconds FIG. 174. Time Required for Pressure Variations in the Casing of a Parsons Turbine. With 165 pounds per square inch absolute initial pressure, usually the no load pressure varies from 25 to 50 pounds absolute, 2 3 8 THE STEAM TURBINE when, according to the curve, the time required to reach this pressure without throttling is about .4 to .8 second; and as the load is increased correspondingly shorter times. Wilkinson Governing Device. An important type of valve gear has been invented by Mr. James Wilkinson and is being applied to the Wilkinson turbines. The general arrangement of his governing device is illustrated in Fig. 175 showing governor, FIG. 175. Wilkinson Valve Gear. eccentrics, and a series of valve casings. One of these casings contains what is called a governor nozzle which is connected mechanically with the eccentrics. This. is a form of auxiliary valve of which the function is not primarily to discharge steam into the turbine blades but to admit steam into or eject it from the other valve casings for the purpose of opening or closing them. This governor nozzle is illustrated in Fig. 176. The important feature of the governor nozzle is a cone-shaped piston at the lower end of the valve rod passing through the stuffing-box and connected to the eccentrics as shown in Fig. 175. A cone-shaped jet flows continuously over this cone. The central chamber of the governor nozzle as well as the spaces around it except a narrow annular passage communicating with a similar passage shown by a circular section in Fig. 176 at the left-hand side of the GOVERNING STEAM TURBINES 239 central chamber contains, in normal operation, steam at the "admission" or initial pressure. The valves in the other casings are operated by the force produced in this annular chamber by the injector or ejector action of the cone-shaped jet. Steam to be admitted to the annular passage must pass around the cone-shaped piston, and the position of this piston with respect to the annular passage deter- mines the effective pressure of the steam operating the admission valves. When the cone-shaped piston is in its lowest position the steam in passing around it to enter the turbine nozzles opposite, produces an ejector effect in the annular passage; but when the cone-shaped piston is at the other end of its stroke the steam produces an injector effect in the annular passage. When the injector effect predominates the pressure in the annular passage is greater than that in the steam chest, while with the ejector effect predominating this pressure is considerably less. FIG. 176. Governor Nozzle. FIG. 177. Admission Valve. One of the admission valves is shown in Fig. 177. A small passage of circular section shown here in the wall of the steam chest communicates with the governing valve or governor nozzle. This passage communicates with one side of the piston valve illustrated here. A spring is provided to keep the valve closed when the pressure in the passages communicating with 240 THE STEAM TURBINE the governor piston is the same as in the steam chest. When the pressure in the passages is, however, less, corresponding to the ejector effect, by an amount greater than the tension in the spring (about 25 pounds) the valve is opened. With the injector effect, on the other hand, the valve will be closed. All the admission valves operate together, as the pressures are approximately the same in each of them. The governor valve oscillates 150 times per minute. The position of its cone-shaped valve with respect to the annular passage communicating with the admission valves is determined by the height of the governor weights and is adjusted by means of the levers shown in Fig. 175. By this arrangement the duration of the ejector effect opening the valves is controlled by the speed. All the valves open a fixed number of times in a minute, but the duration of the period they are open varies with the load. Because the governor nozzle is always open to the steam chest, steam is never cut off from this nozzle, but w\ih careful designing in proportioning the sizes of the nozzles this is no particular dis- advantage. In its action this valve gear is not unlike the usual Parsons governing device which has already been explained. It is likely that both are affected by quasi throttling at very light loads. To some extent the magnitude of this effect would probably be in proportion to the length of the casing. By-pass Governors. In all turbines the area of the steam passages increases in going from the high-pressure end to the exhaust. Consequently it is possible to pass a larger quantity of steam through a turbine for an overload, by admitting high- pressure steam into the middle stages in addition to the steam coming through the high-pressure nozzles. This is accomplished usually by the use of an auxiliary valve which opens slowly when an overload comes on the turbine and admits high-pressure steam directly into the low-pressure stages. As the steam entering through the by-pass valve acts on fewer rows of blades than the steam admitted under normal conditions, obviously, of course, the method is uneconomical and should, therefore, be used only for GOVERNING STEAM TURBINES 241 emergency loads. When a by-pass valve is used, the turbine is designed to be large enough to carry a little more than the normal full load, at which, of course, 'it is most economical; and for overloads it is expected that the efficiency will be considerably reduced. All the makers of Parsons and Rateau turbines use by-pass valves. The Westinghouse turbine has by-pass overload valves under the control of the governor, so that they open automatically when an overload comes on, but on most turbines by-pass valves are opened by hand. Overload economy is not usually of great importance, so that, practically, it is considered more feasible to use overload valves than to install additional turbines. In turbines of the Curtis type, which can be made to take a large overload with the addition of only a few extra nozzles without increasing the other dimensions, by-pass valves for overload have no advantages. In Parsons turbines, as anticipated in the design, there is usually a falling off in speed when the overload valves open. FIG. 178. By-pass Valve Designed by Brown-Boveri & Co. A by-pass governor is shown in Fig. 178, which is a diagram- matic sketch showing the method of admitting high-pressure steam to the low-pressure stages of a Parsons turbine. This particular device is due to Brown-Boveri & Co. This design shows the by-pass method applied to an exceptionally well-made 2 4 2 THE STEAM TURBINE turbine much used in Europe. The by-pass ports open only at overload, and the speed is regulated for small fluctuations by the throttling method. In the figure the centrifugal governor is marked 9, and operates by means of levers a balanced throttle valve. The by-pass valve 7, on the other hand, is operated by the pressure on the piston 10. Since this piston and the by-pass valve are on the same valve stem, they are raised or lowered together according as the pressure in the steam chest is high or low. With a high pressure the piston rises, lifting with it the valve 7, thus uncovering the ports, shown on one side of the turbine at 6, which admit steam through the pipes 3, 4, and 5 to different parts of the turbine casing. Obviously there is a considerable change in the power developed immediately after the steam is admitted to one of the pipes 3, 4, or 5; and the consequent fluctuation in speed is taken care of by the throttling governor 9, or by an electrical solenoid governor indicated at 12. FIG. 17*9. By-pass Valve Arrangement for a Parsons Turbine. A simpler type of by-pass valve and governor arrangement is illustrated in Fig. 179. The by-pass valve is here directly under the control of the governor. The governor is marked 9 in the figure and operates a by-pass piston valve 7. The steam enters the turbine through the steam chest over the by-pass valve. When there is no overload on the turbine, the steam passes through the side port, which is shown open in the figure, to the GOVERNING STEAM TURBINES 243 steam space 2 below, and from here it passes through the turbine. When, however, there is an overload, the by-pass valve 7 is raised by the governor and high-pressure steam is admitted through the pipe 3 to some lower pressure stages. With still more over- load the ports for the pipes 4 and 5 are also opened and high- pressure steam is .admitted to stages intended for still lower pressures. When in order to make repairs to the condenser equipment, or for other reasons, it is necessary to run a Parsons turbine non-condensing the by-pass valve is opened and high-pressure steam is admitted to the intermediate stage of the turbine. By this method, because of the larger area of the passages, more steam can be used and the turbine is able to carry full load without a vacuum, although, of course, at a sacrifice of economy. If turbines are designed to take a large overload without a by-pass, the turbine must be of correspondingly greater capacity than the full rating indicates. The best economy of steam will then be at the highest output, and not quite so good at three- quarter load and full load. This is the usual practice in designing impulse turbines; but in the large sizes of Curtis turbines special overload valves are provided. Curtis Overload Valves. In Curtis turbines of four or five stages, especially in the larger sizes, automatic valves are pro- vided to open additional nozzles in the diaphragm between the first and second stages at times of overload. The usual designs of such valves are similar to the one shown in Fig. 180, which is arranged to operate when the pressure in the first stage due to a large flow of steam is larger than the normal. This design consists essentially of a piston valve fitted with a spring of sufficient strength to balance the unequal pressures on its faces for normal operation of the turbine. In the position shown in the figure the valve is closed; that is, no steam passes through it from the first stage to the nozzles discharging into the second stage. The pressure on its upper face is that in the first stage, while the pressure on the lower face is approximately that in the third stage. 244 THE STEAM TURBINE Stage Second 6&3ffe Th/rcf Stage FIG. 180. Curtis Overload "Stage" Valve. GOVERNING STEAM TURBINES 245 As the flow of steam increases due to increasing the load with a constant nozzle area the pressure will become greater in each of the stages; but obviously the pressure will be increased much more in the first stage than in the other stages, and at an overload the difference in pressure between the first and third stages becomes great enough to overcome the resistance of the spring on the overload or stage valve, so that it is forced from its seat. Steam then passes through a port communicating with an extra set of nozzles discharging into the second stage. These valves should be adjusted by varying the tension of the spring, so that they tend to open and close within a compara- tively small range of first-stage pressure. If the adjustment of such a valve has not been properly made and the valve remains open with a load fluctuating in a wide range between overload and considerably less than normal, the economy may be seriously impaired, or if one of these valves remains in a partly closed position so as to throttle the steam, the economy will be affected at all loads. Such valves to be efficient should open and close abruptly. Experimental Data Concerning Governing. Something should be said about the experimental results at hand concerning the different methods of governing. Curves illustrating the effects of throttling have been shown in Fig. 161, but a more satisfac- tory comparison can be made from the following table.* Rated Full Load. Fraction of Load. Kw. i 1 Full. Curtis $ 600 20 2OO 500 500 400 1250 350 5*-7 52-4 68.0 52-3 56.0 S5-o 57-o 57-3 58.8 76.2 76.5 82.0 76.2 78.0 77.2 78.5 78.5 80.3 TOO 100 IOO IOO IOO IOO IOO IOO IOO Curtis De Laval. De Laval C. A. Parsons Company Rateau Westinghouse-Parsons Westinghouse-Parsons Zoelly * Mechanical Engineer, Jan. 20, 1906. 246 THE STEAM TURBINE Fractions of load given at the top of each column refer to fractions of the most efficient load. Steam consumption at the different loads is expressed as a percentage compared with the steam consumption at the most economical load for each partic- ular machine. In other words, if the economy of any of these turbines were as good 'at half load as at full load we should have in the table under the column for one-half load 50 per cent, etc. Results in this table must be used guardedly and not confused with steam consumption. For example, the De Laval 20okilo- watt turbine appears to such good advantage here because the system of governing used for these tests was nearly ideal. Full load steam consumption, on the other hand, was high com- pared with any other make of turbine in the list. The Zoelly, Rateau, and the 2O-kilowatt De Laval turbines used simple throttling governors. The Curtis and the 2oo-kilowatt De Laval were governed by varying the number of nozzles to suit the load. The original Parsons and Westinghouse-Parsons turbines used the " blast'' governor. These data lead to the conclusion that the Curtis and the experimental De Laval (200 kilo watt) give the best results as regards the method of governing. Sufficient data are not available of the performance of Wil- kinson and Allis-Chalmers turbines to be included in this com- parison. CHAPTER IX. LOW-PRESSURE STEAM TURBINES. EARLY in the period of steam turbine development it became apparent that these new types of prime movers were capable of operating with ratios of expansion far beyond those economically possible with reciprocating engines. In the discussion of the effect of vacuum on steam consump- tion the good results obtained with turbines running at a high vacuum were clearly shown. With a high vacuum the heat efficiency of a reciprocating engine is not nearly so good as that of a turbine, because it is not desirable to make the engine cyl- inders large enough to handle economically the great volume of steam we have to deal with when the exhaust pressure is very low. For pressures slightly above atmospheric, however, a first- class, slow-speed reciprocating engine has a slight advantage over the turbine. We can see, then, that a combination of a non-condensing reciprocating engine with a condensing turbine, the latter taking exhaust from the former, might well be sug- gested. Fig. 181 shows graphically the volumes of the steam in each of the five stages of a Curtis turbine and illustrates how rapidly the volume increases at very low pressures. Reciprocating engines may be designed to operate with improved economy up to 25 or 26 inches vacuum; but this is about the limit. Steam turbines, on the other hand, will operate economically* with steam at the highest vacuum practically obtainable. The initial pressure of low-pressure steam turbines is usually that of the exhaust steam from non-condensing engines. With * It is not always commercially profitable to design a plant for operation at an extremely high vacuum, as the first cost of condensers and auxiliaries is usually a deciding factor. 247 248 THE STEAM TURBINE steam admitted to the engine at 200 pounds and exhausted from the turbine at 28 inches vacuum, theoretically there is no differ- ence in the total economy of a unit consisting of a reciprocating 11 U C > Is. -3 3 .s s a ^ - '' engine operating with an exhaust turbine taking steam within the Umits from 7 pounds to 15 pounds per square inch absolute.* The curve in Fig. i8ia shows the energy (B.T.U.) made available * Proc. Inst. of Naval Architects, 1908. LOW-PRESSURE STEAM TURBINES 249 by expansion from 26 inches vacuum to higher. The rapid in- crease above 27 inches is an important consideration and makes 29 28 27 28 10 20 30 40 50 60 70 Energy in B.T.U. available between 26" and 29" vacuum. FIG. i8ia. Curve Showing Enormous Energy Made Available for Work by Expansions to High Vacuums. it very essential to get a very high vacuum in the exhaust on account of the enormous energy obtainable. Flattening out of the curve of the specific volumes of steam, Fig. i8ib, is also interesting in this connection. The increase in volume as the expansion is carried beyond 28 inches of vacuum is not generally realized. The following table prepared by the General Electric Company shows the large amount of work that is made available by the use of turbines in connection with existing non-condensing and condensing plants, the steam being delivered to the steam en- gine at 165 pounds per square inch absolute pressure. Owing to the rapid development of the turbine industry for high speed work and the close attention to this branch of turbine applications required of designers, the " combination " system of reciprocating engines and turbines was comparatively neglected. Only recently the advantages of this system have come to be 2 5 THE STEAM TURBINE generally recognized, and particularly in connection with marine propulsion. Parsons has never advised the installation of an " all turbine " arrangement for ships designed for a speed of Atmos- pheric Inches of Vacuum. Pressure. Pressure of steam at tur-j bine admission valve [ O 4" 8 12 16 20 24 in inches of vacuum. J Per cent, gained over output of engine when worked with high vac- uum, the turbine ex- 26.1 26.5 26.8 26.3 25-3 23-6 20 hausting to a vacuum of 28^ inches. less than 1 5 knots, and for moderate or slow speeds his designers have recommended the " combination " system. According to one of his designs for a cargo vessel intended for a speed of nj 29 28 27 26 100 200 500 300 400 Cubic Feet per Pound. FIG. i8ib. Curve of Specific Volumes between 26 and 29 Inches of Vacuum. knots, if provided with a reciprocating engine discharging steam into the turbine at 7 pounds per square inch absolute pressure, the steam consumption was estimated to be 15 to 20 per cent, less LOW-PRESSURE STEAM TURBINES than that of an "all turbine" arrangement, or of triple expansion engines of the type usually fitted to this class of vessel. The " combination " system gives a vessel also greater maneuvering power than if driven only by turbines. On land low-pressure turbines have been installed principally in connection with rolling mill engines in steel works and wind- ing engines in mines. In both cases the engines are stopped or are running practically idle a large part of the time. These engines are usually reversing and are operated non-condensing. When a low-pressure steam turbine is installed to take the exhaust from such engines an equal amount of power can be obtained from the turbine (at 28 inches vacuum) as from the engine, thereby doubling the power of the plant without increas- ing the consumption of coal or the size of the boiler plant. Throughout the country there are a great many reciprocating engines exhausting into the atmosphere, and the exhaust from these engines is often wasted. There is no doubt that when these plants need increased capacity an installation of exhaust turbines will be profitable, even in most cases where there is no supply of water for condensing, and cooling towers must be erected. There are also many power plants equipped with high grade compound reciprocating engines operating condensing which have a highei efficiency (not steam consumption) when operating non- condensing, or at a comparatively small vacuum, than when operating condensing. In such cases the installation of low- pressure turbines is probably always profitable. As an instance of the uses of exhaust steam turbines the following paragraph is quoted from a report prepared by a company manufactur- ing large sizes of both reciprocating steam engines and steam turbines. " A compound reciprocating engine with cylinder ratios of 3.5 : i, say of diameters 28 inches and 52 inches, with 150 pounds initial pressure, may be assumed to have 1000 kilowatts econom- ical capacity when running condensing and having a steam con- sumption of about 22 pounds per kilowatt-hour. This engine if operated non-condensing should have valve gears adjusted to 25ob THE STEAM TURBINE develop 1700 I.H.P., when it would consume about 20 pounds of steam per I.H.P. per hour. This gives 30,600 pounds steam available for the turbine, allowing 10 per cent, of moisture in the exhaust of the reciprocating engine. The total amount of steam passing the reciprocating engine, however, being 34,000 pounds, 30,600 pounds would develop not less than 1073 brake horsepower in the turbine. Allowing 94 per cent, for the mechanical effi- ciency of the reciprocating engine, the combined horsepower developed would be 2673 brake horsepower and the steam con- sumption of the two units 12.7 pounds per brake horsepower, or 1 8 pounds per kilowatt-hour, which is a remarkable performance for engines of such capacities operating without superheat. Compared with the performance of the reciprocating engine running condensing, this gives 75 per cent, increase of power and 1 8 per cent, saving of steam." Fig. i8ic shows a very important installation of 5Ooo-kilo- watt steam turbine-generators in combination with reciprocating engines of the same power rating. A low-pressure steam tur- bine has been installed to take the exhaust from each engine. Condensing engines when changed to non-condensing operation do not necessarily have their capacity in horsepower reduced because of the great increase of back pressure against which they must then operate. Such reduction would appear, however, on first thought to be the natural result; but, contrarily, the capacity of such an engine when changed to non-condensing operation may be unaltered or even in exceptional cases may be actually increased; particularly is this the case if the engine is one designed for a high expansion ratio. Under these conditions the high-pressure cylinder must have enough volume to pass the required amount of steam, without having the cut-off come so late as to sacrifice all opportunity to use the steam with a reasonably good expansion. There is an interesting reason for the capacity of many compound engines not being reduced when this change is made. In this adjustment the cut-off of the high- pressure cylinder has been shifted to make it late enough so that expansion in the low-pressure cylinder will not cause a loop in LOW-PRESSURE STEAM TURBINES 2500 FIG. i8ic. Steam Turbines taking the Exhaust from Large Reciprocating Engines. 25od THE STEAM TURBINE the indicator diagram or to a final pressure in the low-pressure cylinder when its exhaust valve opens, which is lower than the average pressure in the exhaust line supplying the turbine, which is also the engine exhaust pipe. As the result of this adjustment of cut-off in the high-pressure cylinder the top part of the indicator cards taken from it will be observed to be much enlarged and of greater area than before, the increase being in some cases greater even than the area which is lost at the bottom of the low-pressure diagram by raising its exhaust pressure to about atmospheric. But there are also many compound engines operating condensing in which the release in the low-pressure cylinder occurs when the pressure is relatively high, possibly as high as atmospheric. Now the application of a low-pressure turbine to take the exhaust from this engine would have the effect of very materially reducing the capacity of the engine, as the benefits to be obtained to avoid the loop in the low-pressure diagram have been sacrificed in the original design of the engine and there is no chance to increase the area of the top of the diagram. Engines designed to operate both condensing and non-con- densing have generally the valves adjusted so that normally there will be a rather high back pressure at the point of release in the low-pressure cylinder, so that this " looping" of its indi- cator diagram will be avoided when running non-condensing or on light loads. When carrying full loads or an overload, on the other hand, the expansion will not be complete and a serious loss results in such engines designed for operation with loads varying considerably. In this case when the low-pressure tur- bine is applied there will be no occasion for the high back pressure at release, and the valve setting can be changed to exhaust the steam from the low-pressure cylinder when running at full load so that this pressure will be about atmospheric or possibly a half pound or a pound above to assist in getting the steam readily through the exhaust ports of the engine. For this condition the blades of the low-pressure turbine taking this exhaust should be designed for initial pressure approximately atmospheric when LOW-PRESSURE STEAM TURBINES 2506 getting the amount of steam used by the engine at full load. Now when the cut-off is shifted back by the governor to keep the inlet valve of the engine open only half as long as at full load, only half as much steam will be delivered to the turbine and the absolute inlet pressure to it will also be reduced to half its former value. The final result is that the expansion of the steam in the low-pressure cylinder of the engine will be about to the inlet pressure of the turbine and there will be no appreciable " looping " of its indicator diagram, indicating that the condi- tions as regards effective expansion of the steam are excellent. In large power plant installations, it is the best practice to adhere to the " unit system " throughout; that is, in providing a separate low-pressure turbine for each engine and also a sep- arate condenser for each turbine. This method, although con- siderably more expensive than that of passing the exhaust steam from several engines to a single receiver supplying a relatively larger low-pressure turbine is, however, much to be desired as it gives so much greater flexibility in the operation of the plant and reduces very much the liability to an enforced shut-down of the plant due to condenser troubles. Maximum load on a low-pressure turbine connected to the engine as here described, that is, without having a governor on its steam supply pipe, is usually reached in conventional designs when the pressure at the inlet to the turbine is about twenty pounds per square inch absolute. A relief valve must always be provided hi this low- pressure supply line which should open to let steam out to atmospheric exhaust when this pressure is exceeded. Steam discharged through the relief valve is an excess above what can be used in the turbine and is obviously wasted. All possible precautions should be taken to prevent the leak- age of air into all that part of the system operating at less than atmospheric pressure. This air is very detrimental to the proper action of a good condenser as it reduces the attainable vacuum and consequently renders impossible the full gains to be ex- pected from the installation of low-pressure turbines. Such leaks occur most generally in the joints of the exhaust piping of 25of THE STEAM TURBINE both engine and turbine, also in imperfectly tight relief valves, as well as through the stuffing-boxes on the piston rods of the low-pressure cylinder of compound engines. To eliminate these difficulties the piping should be examined and tested frequently by applying a lighted taper to all questionable joints to observe whether the vacuum inside the piping tends to draw the flame toward it, as would occur if there were a leak. Another pre- caution often necessary is to put a special type of stuffing box on the piston rods of the engines, these boxes being supplied with steam at a pressure slightly above atmospheric so that air leakage inward is prevented. When several engines are connected up to supply exhaust to a single low-pressure turbine, it is always most desirable to turn over each engine for several revolutions with the piping connec- tions arranged so that the engine just starting will exhaust into the" atmosphere. This should be done in order to avoid dis- charging the air in the engine cylinders into the low-pressure turbine and into the condenser. If this precaution is not taken when several engines are operating and another is started the effect in vacuum reduction will be observed. In matters of design, the low-pressure steam turbine presents no new problems. In fact its construction is in many respects simpler because of requiring fewer complicated details than the usual types of high-pressure or " complete expansion " turbines that have been studied. The relatively short length of the shaft or drum required for low-pressure turbines makes for rigidity and freedom from vibration stresses. The skill of the engineer comes into play in this new field almost entirely in the methods of application to conditions that a few years ago were not thought of. The primary consideration in practically all these applications is to utilize as much as possible of the avail- able exhaust steam about the plant, either in the low-pressure turbine or in some still more efficient method. While it is the object to show here the great advantages of this type of prune mover, yet it should be pointed out that there are often condi- tions arising in power plant practice when exhaust steam can LOW-PRESSURE STEAM TURBINES 25Og be used much more efficiently than in any known type of prime mover. The maximum thermal efficiency of a low-pressure tur- bine cannot well be made to greatly exceed 10 per cent., and even with this low efficiency when exhaust steam is a by-product, with no other available use, the addition of such a turbine to the plant will in such cases produce a great saving in coal bills. In other cases, however, where exhaust steam can be used advanta- geously for heating water, as for example in feed-water heaters, hot-water vats in manufacturing processes, or for heating build- ings, it would certainly be false economy to use the exhaust steam in a turbine and install low-pressure boilers for heating water. In the cases cited the thermal efficiency of the process of heating water with low-pressure steam is usually about 80 per cent., which is to be compared with the 10 per cent, efficiency of the turbine. In this analysis it must not be overlooked, how- ever, that a steam power plant operating non-condensing can use only a very small percentage of the total amount of its exhaust steam for heating the feed water. In fact in the aver- age steam engine plant operating non-condensing only about one-sixth of the exhaust steam can be used for heating the feed water, the other five-sixths being discharged into the atmos- phere through the exhaust head. Of first importance among the general considerations affecting low-pressure turbine applications is the providing of adequate facilities for the removal of water and oil from the steam before it enters the turbine. Very wet steam can have no considerable deleterious effect on the turbine compared with the disastrous results often experienced in steam engine practice. It has the effect, however, of increasing enormously the fluid friction in the turbine blades and therefore of reducing the output and raising the steam consumption. Oil, when clean and pure, is not nec- essarily objectionable in the turbine and will pass through with- out accumulating, but in cases where boilers sometimes foam and discharge sulphates and carbonates with the steam, these will mix with the oil and form a gummy deposit on the blades. This deposit is not ordinarily removed by erosion, particularly THE STEAM TURBINE when steam velocities are not over 400 feet per second, and will often choke the steam passages between the blades. Until recently in America exhaust steam turbines were usually arranged to take steam directly from the exhaust pipe of the engines without intervening valves or governing mechanisms. A generator direct connected to the turbine will operate very satisfactorily with generators adapted for connection in parallel to engine-driven generators, and the turbine set thus " floats on the system." As it receives only steam exhausted from the engine its output will therefore vary as the load on the engine. When the load becomes light the steam supply will be reduced by the governor on the reciprocating engine. In the case of direct current units the generators may have shunt windings, and as the voltage will vary nearly as the speed, the load will be automati- cally proportioned between the reciprocating and turbine units. FIG. i8id. Simplest Combination of Low-pressure Steam Turbine and Recip- rocating Steam Engine. The most common and probably also the simplest application of the low-pressure turbine is shown in Fig. i8id. As shown the LOW-PRESSURE STEAM TURBINES 250! generators connected to both the reciprocating engine and the turbine are connected to the same three-phase alternating-current circuit and here also no governor is used on the turbine. By this arrangement the turbine will take automatically its share of the total electrical load in proportion to the amount of steam supplied to it. If it tends to forge ahead of the reciprocating unit it will take more of the load, leaving less for the engine whose speed will immediately increase until its governor re- duces the flow of steam to both the reciprocating engine and to the turbine, thus controlling with one governor the amount of steam supplied to the complete system. In case the generators driven by the engine and turbine are of the direct current type, as the turbine forges ahead, and takes more load, the increase in its speed raises the voltage slightly, which puts more current through the fields of the generators and tends to reduce the speed. Self -regulation is thus admirably accomplished. Obvi- ously for the same reason it is possible to vary the speed of the turbine slightly by adjusting the field rheostat. The quantity of steam used determines obviously the relative amount of load carried by the low-pressure turbine; the greater the amount of steam the greater the proportion of load taken by the turbine, which is due, of course, to the variation of pres- sure in the receiver. As this pressure increases the total range of pressure available for the turbine increases, the heat available per pound of steam increases, and consequently more work is done, assuming, of course, a constant vacuum in the turbine exhaust. In the method of low-pressure turbine installation described in the preceding paragraphs, where the turbine operated with- out a governor of its own, the electrical machines driven (gen- erators) were of similar types; that is, all the current generated was supplied to a single line. It is not infrequent, however, for low-pressure steam turbines to be installed to operate with re- ciprocating engines in power houses where the generators on the engines are to supply direct current lines and the generators on the turbines are to supply alternating current for transmission THE STEAM TURBINE to a distance. Such an arrangement is shown in Fig. i8je. Obviously the engine and the turbine must each have its own governor. If the low-pressure turbine were arranged to take only the load from the alternating current line there would be much steam wasted when the direct current load happened to be heavy and the other light; and conversely' it might be neces- sary to sometimes supply the low-pressure turbine with high- FIG. i8ie. Application of a Rotary Converter in the Combination of Low- pressure Turbine with Reciprocating Engine. pressure steam when the load on the engine was light. A most satisfactory method to avoid these difficulties is to install a rotary converter or motor-generator set as shown. Any in- equality of the two loads will then be taken care of and the load coming on the engine and on the turbine will be divided auto- matically to give the best results. This sort of arrangement might not be very satisfactory for taking care of electric lighting loads if there were likely to be exceptionally frequent reversals of the operation of the converter from alternating to direct current and vice versa, as there might be a voltage change of several per cent, which would perceptibly affect illumination until again adjusted at the switchboard. In most cases, how- LOW-PRESSURE STEAM TURBINES ever, the demand for one kind of current will always predomi- nate, so that this sort of reversal is not likely to be troublesome. Another application of a low-pressure turbine is illustrated in Fig. 1 8 if, where the steam engine drives a line of shafting through a belt drive and the low-pressure steam exhausted from the engine goes to a turbine generator unit supplying an electrical transmission line. In this case the engine and the turbine must each have a governor, as the loads on the two machines are en- A.C.Line FIG. i8if. Application of Condenser in Combination System. tirely unrelated. The device adopted in this case to make the plant as economical of steam as possible is to operate the engine with the excess of steam, above that required for the low-pressure turbine, to be discharged directly into the condenser. By this method the engine will operate at times at a fairly good vacuum as determined by the relative amounts of steam and absolute pressures in lines A and B. The governor on the turbine oper- ates only the by-pass valve V, regulating the flow of steam from. 2 THE STEAM TURBINE the engine exhaust into the condenser. When the load on the turbine is light this valve will be nearly wide open, deflecting only a small amount of steam into the turbine; but when the turbine load gets near its full capacity, the condenser by-pass valve will be nearly closed. It becomes thus possible for the engine to obtain the advantages of nearly full vacuum when the turbine is running light. It is a good practice to put the usual type of valve on the inlet pipe to the turbine which will also be controlled by the governor to prevent the turbine run- ning away on very light load. A very interesting type of installation is shown in Fig. i8ig. FIG. i8ig. Application of Synchronous Motor in Combination System. The method illustrated here consists in the installation of an electric motor of the synchronous type supplied with current from the generator driven by the turbine and having the pulley on its shaft belted to the line shafting driven by the engine. In case the electrical load on the turbine-generator set becomes too large for it to handle it will slow down slightly with the result that the synchronous motor will be driven from the line shafting LOW-PRESSURE STEAM TURBINES 251 as a generator to supply more current to the electrical supply lines. The additional load coming on the engine as the result of driving the motor will cause the governor to open the inlet valve wider on the engine and admit a larger amount of steam to the system. Conversely, when the line shafting is overloaded the governor on the engine admits more steam to the system, in greater amount, however, than the turbine requires. This re- sults in a speeding up of the turbine and a forging ahead of the synchronous motor so that it acts now purely as a motor to assist in driving the shafting. Provisions for Intermittent Supply of Steam. An ingenious development, largely due to Professor Rateau, has been applied to cases where the supply of exhaust steam is intermittent, as in the case of rolling mill and winding engines. Rateau's device, called an accumulator, is used to bridge over the "dead periods," and by providing sufficient capacity it can be made to provide a practically constant supply for an exhaust turbine. Rateau's Accumulator. This regenerator or accumulator is shown in Fig. 182, illustrating longitudinal and transverse sec- tions. This regenerator consists of a large cylindrical shell partly filled with water. When the engine exhausting into it is running the steam is delivered as a spray through the small holes in a num- ber of pipes immersed in the water. By this method some of the steam is condensed and gives up heat to the most of the water. As these accumulators operate usually with steam at atmos- pheric pressure, the entering steam will have a temperature of 212 degrees F. and will tend to heat the water to that temperature. If, now, the engine stops, the supply of exhaust steam is discontinued, and the flow of steam to the turbine will tend to make the pres- sure fall off slightly so that 212 degrees F. will then be slightly above the temperature of boiling water at this lower pressure. In this way the water will be evaporated to supply steam as a boiler would. If, now/the engine starts again, steam will be delivered to the accumulator at a temperature slightly above that to which the water has fallen, due to the cooling effect of the evaporation for supplying the turbine, and the mass of water will again absorb 252 THE STEAM TURBINE i-^CW ] C 3 C "h jf ] C 3 C 3 t 3 " 'C J 3 411 LOW-PRESSURE STEAM TURBINES 253 heat from the exhaust steam. Water has a higher specific heat than any other substance except hydrogen, so that it is a most suitable and convenient substance for heat accumulation. In actual practice it is more convenient to run the regenerator at a pound or two pressure above the atmosphere, as in this case the piping is not under vacuum, so that so much care does not have to be exercised to avoid air leaks. In certain cases, how- ever, it is desirable to run below atmospheric pressure. In this way the power of the primary engine may be augmented by letting it operate at a partial vacuum. Plants are actually run- ning with a delivery pressure to the turbines as low as six pounds below atmospheric pressure. On account of heat radiation from the accumulator, water gradually accumulates in it; but by means of a float trap shown at the right-hand side of the longitudinal section this excess of water is removed. If for any reason the engine shuts down for a considerable period, the supply of heat stored in the accumulator will become exhausted and the pressure will fall below the practical limit for operation of the turbine. To provide for such an emergency an automatic reducing valve is inserted in the piping to deliver live steam to the accumulator. There is also a relief valve on the accumulator through which excess steam will pass off into the air when the pressure becomes 3 or 4 pounds above atmospheric. The pressure in the accumulator should be always about five to ten pounds per square inch, gage pressure, or at least a few pounds above the atmospheric to avoid the possibility of air leaking into the system. The important consideration in the selection and designing of an accumulator for a low-pressure turbine is the length of time the regenerator will be expected to carry full load on the turbine without receiving any low-pressure steam from the engine or engines. Obviously the longer this time is, the greater the capacity required of the accumulator. Quite generally the mis- take has been made, according to Hodgkinson,* of supplying * Tlie Electric Journal, April, 1913, page 335. 254 THE STEAM TURBINE these accumulators in much too large sizes for the require- ments. In many cases the time interval has been assumed to be six to seven minutes, during which the accumulator must supply the steam, while more careful study shows that five to six seconds would have been a much better estimate. The case of a steel mill is cited. If the exhaust is to be taken from a blooming mill the time element should bear some relation to the period between the passes of an ingot, as well as to the maximum tune from the last pass of one ingot to the first pass of the next ingot. The accumulator should not be designed, therefore, to cover such delays as would arise from the clogging of the mills or because a new ingot might not be ready to be bloomed. For these cases of unusual delays another method is recommended. When the demand for steam at the engines is interrupted there will be a sudden rise of pressure in the boilers and the safety- valves will blow off. This steam should be piped to the accu- mulator inlet, instead of being allowed to escape to the atmos- phere. This steam from the safety valves will assist materially in helping the low-pressure turbine in carrying its load. A very good arrangement for the accomplishment of this idea is to place -a " cross-connection " of piping between the steam main sup- plying the engine and the engine exhaust line, and to put into this line a globe type of spring loaded valve set to permit steam to pass through the cross-connection when the pressure is a few pounds lower than that at which the boiler safety valves will blow. For engine power plants, where the supply of exhaust steam is often stopped for long periods, the accumulator installation is usually dispensed with, and the low-pressure turbines are provided with piping to take steam directly from the boilers, in addition to the exhaust steam piping. (See pages 258a to An exhaust steam turbine has, of course, relatively few rows of blades compared with ordinary high-pressure turbines. From several tests made with 5oo-kilowatt exhaust turbines in England, a steam consumption of 34 pounds per kilowatt-hour LOW-PRESSURE STEAM TURBINES 255 was obtained with 15 pounds per square inch admission pressure and 28 inches vacuum. The curve in Fig. 183 shows the steam consumption in pounds per horsepower-hour at the switchboard of a 5oo-kilowatt exhaust steam turbine of the Rateau type. -f--+ -H-- Illllllllll -- CURVE OF STEAM CONSUMPTION H.P.-Hour 487 27.9 37-i II. 8 7-7 5-2 26.97 27.03 26.98 592 321 IO2 17,720 11,980 6 >57 29.9 37-3 64-4 ii. 6 8.7 2 7 .8 28.00 586 458 16,400 13,920 28.0 3-4 6.1 4-5 27.90 27.99 234 114 9.036 6,248 38.6 54-8 2 5 6 THE STEAM TURBINE Fig. 184 is a copy of a shop drawing of a looo-kilowatt Westing- house double-flow low-pressure turbine. The exhaust steam from the engines enters through the annular space H and is distributed FIG. 184. looo-kilowatt-Westinghouse Low-Pressure Turbine. to the right and left sections of Parsons blading. The upper half of the drawing is a section of the rotor and shows the method of construction. The exhaust is discharged through the base as indicated by arrows. The openings I, I are provided for con- venient inspection of the blading. They are covered with suitable covers in which automatic relief valves are fitted. Economy curves of this turbine are shown in Fig. 185. The pressure of the steam delivered to the turbine was approximately atmospheric. The vacuum, as shown by the curves, was 27^ inches for one test and 28 inches for the other. Another Westinghouse turbine built to operate in connection with high-pressure reciprocating engines gave the following results in a shop test: Initial steam pressure, 15 pounds per square inch absolute. Superheat, 40 degrees F. Vacuum referred to 30 inch barometer, 23 inches. Load, 1500 brake horsepower. Steam per brake horsepower hour, 35.5 pounds. LOW-PRESSURE STEAM TURBINES 257 In all these tests the exhaust was condensed in a surface con- denser, which assures accuracy in measuring the steam con- sumption. A Curtis exhaust steam turbine installed in Philadelphia receives the exhaust steam from reciprocating engines at a pres- sure of 15 to 1 6 pounds per square inch absolute, and exhausts into a condenser with an average vacuum of 28 inches. The turbine has no governor, but takes all the steam the engines will supply. The output over and above that obtained from 1-50 -S43 .5 KO K3 3CO 4< O 51 6( *. "" I?P1 1830 IK o ifl o uwnwaxjfcoo o ia JOB x> K 10200 FlG. 185. Curves of Steam Consumption of a looo-kilowatt Westinghouj Low-Pressure Turbine. the reciprocating engines is increased about 66 per cent, for the same steam consumption. With dry, saturated steam at atmos- pheric pressure delivered to the turbine, the guaranteed steam consumption is 36 pounds per kilowatt at full load, and 40 pounds per kilowatt at half load. It is stated that the actual test results are probably at least 10 per cent, better. Since the exhaust from reciprocating engines is often very wet, it is good practice to insert a steam separator in the steam pipe leading to the low-pressure turbine. Most applications of low-pressure turbines have been made in collieries and steel mills, where non-condensing engines are the 'rule. Results are, however, so satisfactory that the design of 258 THE STEAM TURBINE new plants having a compound engine with a smaller low-pressure cylinder than is usually provided, which .is to discharge its exhaust into a steam turbine is likely to become common, as giving better steam economy than can be obtained from either reciprocating engines alone or turbines alone. Low-Pressure Steam Turbines Combined with Gas Engines. There is also another field open to the low-pressure steam turbine. The hot cooling water from the jackets of large gas engines could be heated by the exhaust gases, and the low-pressure steam thus formed would drive a steam turbine. CHAPTER X. MIXED-PRESSURE TURBINES. IT sometimes happens that there is an available source of low- pressure steam which it is desired to utilize for the development of power, but where unfortunately there is not always at hand a sufficiently large amount of this low-pressure steam to take care of the power requirement. To suit this condition it is not infrequent to provide in a steam turbine a high-pressure section to take steam at boiler pressure and thus help out the low- pressure section when its normal supply of steam is low. In this type of construction when, however, the supply of low-pressure steam is sufficient for the power requirements, the supply of high-pressure steam is cut off entirely by the governor; and on the other hand when the supply of low-pressure steam becomes again sufficient in quantity, no more high-pressure steam is used. The generous use of live steam in low-pressure steam turbines is not by any means as poor engineering practice as at first thought it appears. The obvious reason for admitting live steam to the turbine is that the supply of low-pressure steam from the engines is insufficient for the turbine requirements, and that consequently some of the engines have been relieved of their load more or less suddenly. The boiler plant continues, however, to make steam at the former rate, and the safety valves will soon blow off unless the excess steam can be used in the power plant. By taking this excess of steam to the turbine to help in carrying its load, which we shall assume has not been reduced, will serve to use this excess of steam to the best possible advantage. It is not an unusual practice even to pipe the discharge from the safety valves on the boilers into the receiver in the low-pressure piping supplying the turbine. These conditions are met most 258b THE STEAM TURBINE frequently in the suddenly variable loads in rolling mills and in hoisting operations where the reciprocating engine drives the rolls or hoists and the low-pressure turbine supplies a more or less constant electrical load for both lighting and comparatively light power requirements. If the intervals requiring the use of high-pressure steam are relatively long, as for example five to ten hours on the average, then a so-called " mixed turbine " type should be used. The accumulator method is also adap- table for the longer period but is expensive as regards first cost. The so-called " mixed" steam turbine has been a development of the applications of steam turbines to suit two important con- ditions of operation which are as follows: (i) the case where a low-pressure turbine is to be used to develop an amount of power for which there is not constantly available a sufficient amount of low-pressure steam to carry the average load; and (2) when there are large enough quantities of low-pressure steam at certain times to carry the load but at more or less long inter- vals there is no exhaust steam supplied at all. Both of these cases require the supplying of large quantities of steam from sources independent on the exhaust lines and live steam direct from the boilers is invariably the substitute. For this sort of service with widely varying steam pressures the mixed-pressure turbine has found acceptable application. In speaking of a mixed-pressure turbine in this chapter we shall think of one hav- ing separate high- and low-pressure portions in a single casing. A good example is shown in Fig. 1853, where the high-pressure portion is provided with an impulse wheel which is made easily removable so that when for long periods high-pressure steam is* not needed it can be taken off. High-pressure steam enters at the steam chest opposite the nozzles in the "impulse" section. Low-pressure steam enters only the reaction blading through the vertical pipe coming up behind the " reaction" section. Such a turbine differs essentially from the ordinary low-pressure tur- bine which is provided with its own governor only in having under the control of the governor a special set of valves arranged to supply live steam to nozzles directing steam into a section of MIXED-PRESSURE TURBINES 2 5 8C high-pressure blades before discharging through the low-pressure sections along with a supply of low-pressure steam with which it mixes. A very common type of mixed-pressure turbine con- FIG. i85a. Mixed-pressure Turbine. sists of an impulse wheel in which the energy drop from boiler to about atmospheric pressure is absorbed and the remainder of the energy is taken out by expansion in low-pressure reaction blading. An installation of this kind is illustrated by Fig. i8sb. The valves under the control of the governor are adjusted so that no high-pressure steam is admitted until the valves on the low-pressure line are wide open. There are often excellent oppor- tunities for the installation of mixed-pressure turbines in con- junction with accumulators; but in every case a check valve must be provided between the accumulator and the turbine in the low-pressure line to prevent live steam getting back into the shell of the accumulator, which will probably not be strong enough to withstand the excessive stresses that might be pro- duced. In many mixed-pressure turbines the high-pressure section, if of a simple disk construction, is frequently made removable 2 5 8d THE STEAM TURBINE as in Fig. i8sa, so that the " windage" loss due to its revolution when not in use can be eliminated. A good estimate is that about 2 per cent, of the power of the turbine is lost in the air resistance of such a high-pressure section. FIG. iSsb. Low-pressure Turbine with Live Steam Valve Installed at Peace Dale, R. I. Mixed-pressure turbines are not often used in sizes larger than about 2000 kilowatts. To meet the requirements of the larger capacities it is best to use a regular or "complete expan- sion" turbine in combination with a smaller simple low-pressure turbine. CHAPTER XI. BLEEDER OR EXTRACTION TURBINES. THE name " bleeder" or ''extraction" turbine is given to one specially designed to take steam at boiler pressure and to exhaust part of this steam at a normally low vacuum while another part is " extracted " or taken out from one of the stages at a pressure of five to ten pounds per square inch gage pressure; that is, just a little above atmospheric. In many cases, as for example in cotton, woolen, and paper mills, this steam is " extracted" for manufacturing purposes, usually heat- ing water in vats. More commonly, however, such turbines find their application for supplying the low-pressure steam re- quired in a heating system for houses, factories, office buildings, etc. Because this latter supply is needed only a part of the year and otherwise is variable with the seasons, there will be times when the turbine operates by complete expansion of all the steam supplied to it by the boilers. Obviously it is necessary to provide in such turbines a means whereby the " bleeder " steam can be taken out at any time, and with sufficient back pressure even at light loads to maintain a pressure in the section from which the steam is to be withdrawn to overcome the resistances of pipes and valves, so that steam can flow freely as required. It is desirable also that the pressure in this section of the turbine should be fairly constant. To accomplish this result in Wes- tinghouse-Parsons turbines a partition diaphragm has been used to separate completely the high-pressure from the low-pressure portion, as shown in Fig. 1850. The steam from the high-pres- sure section passes normally out through the bleeder passage and into the mains to be supplied with steam. When, how- ever, the turbine uses more steam than is needed for the service supplied by these mains then the pressure in these passages 2586 THE STEAM TURBINE " backs up," and when above that pressure for which the auto- matic * valve shown at the top of the figure has been set, this steam flows out into the low-pressure sections of the turbine, where it does work, and thence into the condenser. It is the function of this valve to maintain sufficient pressure in the passages to create the desired flow into the mains. There is therefore some throttling action in this valve which causes a slight loss in the available energy of the steam supplied to the low-pressure sections. FIG. 1850. Westinghouse-Parsons " Bleeder " Turbine. To avoid as much as possible the throttling action referred to above, Curtis turbines are designed without any additional partitions or diaphragms. The method adopted is to place a ring-shaped valve over the nozzles leading from the stage from which the steam is to be " extracted. 1 ' This valve is operated automatically by a mechanism responsive to the pressure in the stage so that the effective area of the nozzles is changed as required to maintain a constant pressure. By this method the closing of the nozzles occurs only in groups, so that any slight throttling action that might occur due to partial opening would create its loss only in a small group rather than in all the nozzles * Similar to a relief or safety valve in its action. BLEEDER OR EXTRACTION TURBINES in the stage; those not throttled would be either closed off entirely or else wide open. FIG. i85d. Ring Valve of Curtis Bleeder Turbine. FIG. 1856. Side of Diaphragm of Bleeder Turbine. Parts of this device are illustrated by the following figures: Fig. i8sd shows the ring valve used for covering the nozzles THE STEAM TURBINE which are bolted in the usual construction to the side of the diaphragm (Fig. 1856). This valve is operated by the piston in an oil (or steam) cylinder which is in turn moved by being subjected to oil or (steam) under pressure admitted from a high- pressure supply by a small pilot valve actuated by leverage con- nections to the diaphragm (30) in communication by small piping with the stage in which the constant pressure is to be main- tained. Fig. i8sf shows a cross-sectional view of the mechan- / I FIG. 185!. Valve Gear of Curtis Bleeder Turbine. ism which actuates the valve. By means of a flexible joint at (i) the piston rod (2) moves the valve plate back and forth over the face of the nozzles. Pressure on the piston (14) in the oil (or steam) cylinder (13) gives the movement to the piston rod. Movements of the piston are effected by means of the pilot valve (22) which is in turn actuated by the diaphragm (30) by means of the rods (34) and (39) . This diaphragm with its " corrugated " or " accordion " sides forms a cylindrical chamber which is in communication by means of small piping with the stage to be con- trolled, and from which the " bleeder " steam is to be taken. BLEEDER OR EXTRACTION TURBINES 2581 Movements of the diaphragm are opposed by the spiral spring (36) which can be set to maintain any desired steam pressure in the stage. Careful inspection of Fig. i8sd shows that the ports in the ring valve are not all of the same size but are of progressively increasing width around the circumference from the narrowest to the largest. The narrow parts begin closing up on the first movement of the valve. There are four groups. The second group begins closing only after the first or narrowest set is fully closed. A balance-plate (Fig. i8sg) is put on top of the ring valve (Fig. i8sd) for the purpose of assisting in equalizing the pres- sure on the two sides of the valve, and thus reducing the force required to move it. FIG. i8sg. Nozzle -plate for Curtis Ring Valve. It is comparatively a very easy matter to remove some of the steam which has been partly expanded in the turbine by the use of suitable automatic or hand-controlled valves even when the quantity of steam required at a constant pressure in such a bleeder line is quite variable. By this method it is possible to 258J THE STEAM TURBINE extract the greatest amount normally possible as required for generating power and at the same time supplying at a reason- ably constant pressure "usually about atmospheric," or about 5 pounds above, the requirements for heating or industrial purposes. Fig. i8sh shows the satisfactory filling of the blades of an impulse turbine of the " bleeder " type. FIG. i8sh. Flow Lines in a " Bleeder " Impulse Turbine. CHAPTER XII. MARINE TURBINES, ONE of the most important fields for the steam turbine is the propulsion of ships. In the mercantile marine the progress of the turbine had been extremely rapid, the first mercantile vessel propelled by turbines having been built only a very few years ago. That vessel had about 700 tons displacement, and developed 3500 indicated horsepower, comparing with a tonnage of 45,000 and 70,000 horsepower in the Lusitania and Mauretania. Care- ful trials had shown that at all speeds above 14 knots the turbine was more economical than the reciprocating engine, being 15 per cent, better at 18 knots, 31 per cent, better at 20 J knots, and 36 per cent, better at 20. i knots. In the Dover-Calais service it had been found that the turbine boats carried passengers at two knots greater speed with 25 per cento less coal per passenger than boats propelled with reciprocating engines. A saving in coaj of about 9 per cent, was computed for the turbine steamers belonging to the Midland Railway (England), as compared with similar steamers of the same company equipped with reciprocating engines. The difference in initial cost and in weight of machinery was found to favor the turbine driven ships by ij and 6 per cent, respectively. When used for marine service, doubtless the greatest defect of practical steam turbines is that they cannot be reversed. Many attempts have been made to devise a turbine to reverse in a simple way comparable with a reversing reciprocat- ing engine. It is the present practice to provide turbine driven ships with two turbines used only for reversing, and as they are not intended for high speed, they may be of small power compared with the main turbines. These two reversing turbines are usually fitted to the same shafts as the low-pressure turbines, and when the ship is running ahead their rotors revolve idly in a vacuum. 259 260 THE STEAM TURBINE When the ship is to be run backward the steam is shut off from the "ahead" turbines and is admitted to the auxiliary reversing turbines. There is, of course, a disadvantage from not having at times the full normal motive power of the ship available for backing. Besides, conditions are not ideal when a large portion of the plant is idle for a greater part of the time. These reversing turbines will occupy a great deal of longitudinal space, so that the floor space required for an installation of marine steam turbines is larger than that required for reciprocating engines for the same conditions of service. The White Star Company (International Mercantile Marine Company) has decided to operate ocean steamers with a com- bined reciprocating and turbine engine plant. The two outer shafts will be driven by quadruple expansion reciprocating en- gines and the central shaft by a low-pressure turbine operated by the exhaust steam from the low-pressure cylinder of the reciprocating engines. For going backward, the reciprocating engines will be used, as they are readily reversed, and in the ordinary service the turbine and reciprocating engines will be operated together. By this combination the advantages of reciprocating engines for reversing are secured, together with the great range of expansion which is possible with the steam turbine. It is difficult to say what developments the future will bring in the applications of steam and gas turbines to the marine service. Practically all the new battleships and cruisers for the British navy are now turbine driven. If we consider that the steam turbine in its practical form commenced its real develop- ment only in 1885, the future certainly may have rich possibilities. Fig. 185! represents the results of tests made at variable speeds and powers on a standard combined impulse and reaction type of turbine. In explaining the results of these tests Mr. H. T. Herr * states that investigations now under way by the " Westinghouse interests " will insure the elimination in the near future of the reciprocating engine in the field of marine propul- * Journal of the Franklin Institute, March, 1913. MARINE TURBINES 26oa sion as the turbine generator has practically eliminated it for electric power plant service. On account of the difficulty of adjusting the inherent re- quirements of the steam turbine for operation at relatively high rotative speed and the corresponding difficulty, opposite how- ever, in effect, of the efficient operation of the propellers of steamships at anything but relatively low speed, the applica- tions of steam turbines to the propulsion of ships has been very much limited. If it were not for these difficulties there is no reason why the steam turbine should not displace the recipro- -s ^ a rf 45 40 30 - o S g 8 2 Efficiency in per cent. O i i ~- ~. s. ~ t *- -. j Pounds of steam per brake horsepower hour. \. 1 \ i \ / s 4 s ^ 7 S ^ r l -- s ^ ^^< ** / ?> . > / 10 < / ^ *^ / e v / ^f V 2 * L 5 - / / x ** / ^ ^^ / 2.5 / ^-- , ^ ^ 2 f ' 800 1600 2400 3200 4000 Revolutions per minute. FIG. 1851. Curves Showing Variation of Steam Consumption, Horsepower Efficiency of Latest Designs of Steam Turbines with Speed. , and eating steam engine almost entirely for this service. On this account, however, the application has been confined almost entirely to merchant and naval vessels designed for high-speed service. Experience has shown that in the applications of steam turbines to slow-speed ships there has been no appreciable saving in weight, in space, or in the cost of operation, over what it would have been with reciprocating engines. The nearest approach to the solution of this problem is to be secured prob- ably by the application of gearing essentially similar to that designed by De Laval. This sort of gearing cannot, however, 26ob THE STEAM TURBINE be applied without modification to turbines developing more than possibly 1000 horsepower. A design much better suited to high-speed conditions and also adaptable for large power has been developed by the Westinghouse Machine Company with the cooperation of Mr. George Westinghouse, Admiral George Mellville, and Mr. John H. Macalpine. The essential principle embodied in their improvements consists in the application of a so-called " floating frame " designed to carry the pinion on the main turbine shaft. The experimental gear developed in the early stages had a floating frame supported on pivots, permitting flexibility as regards horizontal movement of the pinion, but was rigid as regards vertical movements. Very recently Mr. Westinghouse developed a very important improvement con- sisting in the substitution of a flexible support by means of hydraulic pistons taking the place of the rigid vertical supports, and in this way improving very much the efficiency and the wearing properties of the gear. This improvement in wearing properties had also the effect of reducing to a minimum all noise and vibration which in the original design were consid- erable. In this later design (Figs. 1853* and k), the main frame supporting the pinion is held up by the pistons in the hydraulic cylinders filled with oil under pressure. This construction permits vertical movements of the pinion along with its flexi- bility in its floating frame and is therefore a great improve- ment in that the earlier design permitted only lateral move- ment. The vertical movement is permitted by the supporting of the floating frame on the piston connected to the support- ing rods. Similarly the lateral movement is permitted by the flexibility of the horizontal pistons on the two sides of the pinion. For a more complete description and discussion of the " float- ing frame " type of reduction gear see Engineering, vol. 95 (1913) > pages 169 and 609, and The Electric Journal, January, 1912. Water rate curves drawn from the data of acceptance tests of the battleships North Dakota and Delaware are shown in Fig. 185!. Curves A and BI show the steam consumption per MARINE TURBINES 2600 FIG. 185]. "Floating Frame" Reduction Gear, Showing Gears when Side of Casing is Removed. 26od THE STEAM TURBINE FIG. iSsk. Reduction Gear, Showing Flexible Support of Pinion. I 18 16 N 12 04 2000 6000 10000 14000 18000 22000 26000 30000 Horsepower. FIG. 185!. Water Rate Curves of U. S. Battleships and Computed Curves if Geared Steam Turbines were used. MARINE TURBINES 2 6oe shaft horsepower per hour of the engines in the two battleships, while C and Ci show the corresponding results if geared turbines had been used instead of reciprocating engines. It is estimated that with a geared turbine combination of the Westinghouse " floating frame " type the economy of the prime movers in vessels of the Delaware class could be improved 30 per cent, at full speed and 25 per cent, at cruising speed. Fig. i8sm FIG. 185111. Low-pressure Turbine Casing for German Steamship. shows the enormous size of the casings for the low-pressure sections of the steam turbines installed in modern battle- ships. Electrical Transmission for Ships. Another method differ- ent from the use of reduction gears has been frequently sug- gested for making the steam turbine more adaptable for marine service. This method consists in using on the vessel steam turbines direct connected to high-speed electric generators, which can operate then under practically identical conditions as in " land " service. These turbines, obviously, can then be designed to operate at a speed best suited to obtain high effi- ciency. The electric current from the generators is used to drive slow-speed electric motors on the shafts of the propellers, the speed here being that giving best efficiency for the pro- pellers. This method offers great flexibility in the handling of a vessel, as the motors can be very quickly reversed, and changes of speed are readily obtainable. 26of THE STEAM TURBINE Although this method for marine propulsion has been ad- vocated by engineers for many years it has not as yet re- ceived very favorable acceptance; but, doubtless, it is con- stantly receiving more favorable attention from well-known designers. CHAPTER XIII. TESTS OF STEAM TURBINES. Testing Steam Turbines.* In every power plant the means should always be available for making tests of the steam equip- ment to determine the steam consumption. Usually tests are made to determine how nearly the performance of a turbine approaches the conditions for which it was designed. The results obtained from tests of a turbine are to show usually the steam consumption required to develop a unit of power in a unit of time, as, for example, a horsepower-hour or a kilowatt-hour. In such tests a number of observations must be made regarding the condition of the steam in its passage through the turbine and of the performance of the turbine as a machine. To get a good idea of what these observations mean, it may be profitable to follow the steam as it passes through the turbine. The steam comes from the boilers through the main steam pipe and the valves of the turbine to the nozzles or stationary blades as the case may be. It then passes through the blades and finally escapes through the exhaust pipe to the condenser. It is pref- erable to have a surface condenser for tests so that the exhaust steam can be weighed. The weighing is done usually in large tanks mounted on platform scales. Methods for Testing. The important observations to be made in steam turbine tests are: 1. Pressure of the steam supplied to the turbine. 2. Speed of rotation of the turbine shaft, usually taken in revolutions per minute. 3. Measurement of power with a Prony or a water brake, if the power at the turbine shaft is desired ; or with electrical instru- ments (ammeters, voltmeters, and wattmeters), if the power is measured by the output of an electric generator. * For complete and detailed information regarding the testing of steam tur- bines and other prime movers, as well as the revised codes of testing adopted by the American Society of Mechanical Engineers, . see Power Plant Testing, pages 294-363, by the author (McGraw-Hill Book Co., N.Y.). 261 262 THE STEAM TURBINE 4. Weight, or measurement by volume, of the condensed steam discharged from the condenser. Unless a surface con- denser is used it is very difficult to obtain the amount of steam used by the turbine. All leakages from pipes, pumps, and valves, which is part of the steam which has gone through the turbine, must be added to the weight of the condensed steam. The accuracy of a test often depends a great deal on how accurately leaks have been provided against, or measured when they occur. 5. Temperature of the steam as it enters the turbine. If the temperature is higher than that due to the pressure of the saturated steam given in steam tables, the steam is superheated; if, how- ever, the temperature is not higher the steam may be wet, and a calorimeter must be attached as near the turbine steam chest as possible.* All gauges, electrical instruments, and thermometers should be carefully calibrated before and after each test so that observations can be corrected for any errors. The zero readings of Prony and water brakes for measuring power should be carefully observed and corrected to eliminate the friction of the apparatus with no load. Unless all these precautions are taken the dif- ficulties in getting reliable tests of turbines are greatly increased. In all cases tests should be continued for several hours with absolutely constant conditions if the tests are to be of value. The most valuable test of a steam turbine is made when varying only the load; that is, with pressures, superheat, and speed con- stant. When the steam consumption is then plotted against fractions of full load, a water-rate curve is obtained. For such a curve a series of tests are needed, each for some fraction of full load ; and in each separate test the power as well as all the other conditions must be held constant. * The most satisfactory tests of turbines are made with steam slightly super- heated rather than wet. When steam is very wet (more than about 4 per cent, moisture for ordinary pressures) the determination of the quality is difficult. There is also a danger that steam showing only a few degrees of superheat by the reading of the thermometer is actually wet. The high temperature is due in such cases to heating from eddies around the thermometer case or in steam pockets near it. TESTS OF STEAM TURBINES 263 Another important test of the performance of steam turbines is made by varying both the speed and the power and keeping the other conditions constant. The observations of speed and power from such a test give a power parabola as illustrated in Fig. 80. This curve shows at what speed the turbine gives the greatest output. For complete tests of a steam turbine the steam consumption should be determined at full load (i) with varying initial steam pressure; (2) with varying vacuum; and (3) with varying superheat. A complete set of tests as outlined will give sufficient data to determine all the corrections usually required. Commercial Testing. The methods used by the New York Edison Company in commercial tests of steam turbine-generator units may well be explained briefly. During a test the load on the turbine unit is maintained as con- stant as possible by "remote control" of the turbine governor by the switchboard operator. The maximum variation in load is to be held within 4 per cent, above and below the mean. For some time previous to the test the turbine is run a little below the load required for the test, but at least ten minutes before the starting signal is given the test load must be on the machine. Three-phase electrical load is measured by the two-wattmeter method,* using Weston indicating wattmeters of the standard laboratory type. These instruments are calibrated by a well- known testing laboratory immediately before and after the test. Power factor is maintained substantially at unity and all electrical readings are taken at one-minute intervals. When the turbine is provided with a surface condenser, the steam consumption, or water rate, is determined by weighing in a large tank supported on platform scales the condensed steam delivered from the condenser hot well. Above the weighing tank a reservoir is provided which is large enough to hold the condensation accumulating between the weighings which are made at intervals of five minutes. By using a loop connection * Cf. Kent's Mechanical Engineer's Pocket-Book, 7th ed., page 1069, 8th ed., page 1396, or Foster's Electrical Engineer's Pocket-Book. 264 THE STEAM TURBINE for the gland water supply (of Westinghouse turbines) or the water from the step bearing (of Curtis turbines using water for this bearing) the necessity for correcting the weighings for these amounts is avoided. Because the circulating water at the stations of this company is usually quite salt, any condenser leakage is detected by testing the condensed steam by the silver-nitrate method with a suitable color indicator. This color method is said to be a decided advantage over the usual method of weighing the leakage accu- mulating during a definite period when the condenser is idle and is tested for only one particular vacuum. By taking samples of circulating water and condensed steam at the same time, it is possible to detect any change in the rate of condenser leakage. The water level in the hot well is maintained at practically a constant point by means of a float valve in the well automatically controlling the speed and, therefore, the amount of the delivery of the hot-well pump. This device avoids the necessity for the difficult correction to be made in a test when the levels in the hot well are not the same at the beginning and end of a test. Tem- peratures and pressures of the admission steam are determined by mercury thermometers and pressure gauges located near the main throttle valve of the turbine; the amount of superheat is determined by subtracting from the actual steam temperature after making thermometer corrections the temperature of saturated steam corresponding to the pressure at the point where the temperature is measured. All gauges and thermometers are calibrated before and after the test. Vacuum is measured directly at the turbine exhaust by means of a mercury column with a barometer alongside for reducing the vacuum to standard barometer conditions (30 inches). By this latter arrangement the necessity for temperature corrections which are necessary when the two mercury columns are not at the same place is avoided. Fig. 1 88 shows a 55oo-kilowatt Westinghouse-Parsons turbine set up for testing in the shops before shipment to the customer. TESTS OF STEAM TURBINES 265 266 THE STEAM TURBINE The power is measured by means of a large water brake shown in the figure at the left of the turbine. Reports of Tests. The tables given below have been prepared to show the steam consumption, together with the most impor- tant other data, of what are believed to be reliable tests of standard makes of steam turbines. The vacuum given in the tables is the equivalent referred to 30 inches barometer. Curtis Turbines. The following results were obtained in 1905 by Messrs. Sargent and Lundy with a 2ooo-kilowatt Curtis turbine-generator. Kilowatts. Steam Pressure (Gauge) . Superheat, Deg. F. Vacuum, Inches. Pounds per Kilowatt-hour. 555 1067 2024 155-5 170.2 166.3 204 120 207 28-5 28.4 28-5 18.09 16.31 15.02 Also the following results are reported in 1907 with a 9000- kilowatt turbine-generator in Chicago : Kilowatts. Steam Pressure (Gauge). Superheat, Deg. F. Vacuum, Inches. Pounds per Kilowatt-hour. 5.374 182 133 2 9-43 I3-I5 8,070 179 116 29-35 13.00 10,186 176 147 29.47 12.90 13,900 198 140 29.31 13.60 Parsons Turbines. A i5oo-kilowatt Parsons turbine was tested at Sheffield, England, with the following results: Kilowatts. Steam Pressure (Gauge). Superheat, Deg. F. Vacuum, Inches. Pounds per Kilowatt-hour. 53 1071 1585 145.0 131.0 128.5 no 124 125 28.9 28.3 27-5 21.58 18.24 17.60 TESTS OF STEAM TURBINES 267 The results of two tests of a 3oo-kilowatt Parsons turbine installed at the Hulton colliery are also given to show the change of economy from running condensing 26.58 inches vacuum and non-condensing. Kilowatts. Steam Pressure (Gauge). Superheat, Deg. F. Vacuum, Inches. Pounds per Kilowatt-hour. 303 297 158.0 i6i .0 o 26.6 O. 2 3-l5 34-20 These last tests show well the increased steam consumption (about 50 per cent.) when running non-condensing. Westinghouse-Parsons Turbines. The table below gives the results of tests in 1904 by F. P. Sheldon & Co., Providence, R.I., of a 4oo-kilowatt Westinghouse-Parsons turbine with about 100 degrees F. superheat. Brake Steam Pressure Superheat, Vacuum, Pounds per Horsepower. (Gauge). Deg. F. Inches. B.H.P. Hour.* 279.4 I53-I 92-5 28.0 14-34 410.7 657-3 153-2 152.7 102.9 100.3 28.0 28.0 .13-45 12.48 967.5 149.6 IOO. 2 2 7 .6 12.79 1207.5 152.0 99-9 27-3 13-55 * Observe the steam consumption is in pounds per brake horsepower hour, instead of pounds per kilowatt-hour as for some of the other results given here. The curves given in Fig. 189 were plotted to show graphi- cally the steam consumption of 300, 500, and 1000 kilowatt Westinghouse-Parsons turbines with varying loads.* Data of the tests from which these curves were drawn, as well as of a test of a 3ooo-kilowatt turbine are given in the following tables. These tests were reported by J. R. Bibbins in 1906 and 1907. * The numbers marked on the curves to indicate the vacuum represent the actual readings taken in the test and are not referred to a standard (30 inches) barometer. 268 THE STEAM TURBINE m 300-KILOWATT TURBINE (3600 R.P.M.). 1 o u II U I . *. %> %<% Illal 00 PO fO OvOO ro N ei 4- 10 rOO t N Ov O t- M 00000 E E i2 i2 2 2(2{g{g Illll ooooo iiid CQCQCQ " Hill *-0 N fO M O t-00 Ov O I .2 .9 II s 4 se a 272 THE STEAM TURBINE at the initial conditions and the heat ("of the liquid") in the condensed steam at the temperature of the exhaust. By this method the full load test of a Westinghouse-Parsons turbine reported by F. P. Sheldon & Co. will be calculated from the data given in an official report. In order to make the results of such calculations of steam turbine tests comparable with the usual heat unit computations of reciprocating steam engine tests the results are generally expressed in terms of indicated or " internal " horsepower. F. P. Sheldon & Co. assumed the mechanical efficiency of a reciprocating engine of about the same capacity at full load to be 93.3 per cent. THERMAL EFFICIENCY OF A 400-KILOWATT STEAM TURBINE. Brake horsepower 660 Corresponding indicated or "internal" horsepower of a recip- 660 rocatmg engine = 708 '^OO Total steam used per hour, pounds 9169 Steam used per " internal" horsepower per hour, pounds 12.96 Steam pressure, pounds per square inch absolute 166.9 Superheat, degrees F 2.9 Vacuum, referred to 30 inches barometer, inches 28.04 Temperature of condensed steam, degrees F. (at ,96 pound per square inch absolute pressure) 100. 6 Total heat contents of one pound of dry saturated steam at the initial pressure, B.T.U 1193 . 9 Heat equivalent of superheat in one pound of steam, B.T.U. (Cp. from Fig. 30) 1.9 Total heat contents of one pound of superheated steam, B.T.U. 1195.8 Heat of liquid in condensed steam, B.T.U 68.6 Heat used in turbine per pound steam, B.T.U 1127.2 Heat used in turbine per "internal" horsepower per minute, B.T.U. = 1127.2 X -^- = 243.5 Heat equivalent of one horsepower per minute, B.T.U. = 3 42 .42 770 Thermal efficiency, per cent. (42.42 -r- 243.5) *7 4 Standard forms for data sheets and for tabulating results of steam turbine tests are given in Power Plant Testing by the author. (See pages 315-340.) Full explanations of methods and of necessary precautions are given. CHAPTER XIV. STEAM TURBINE ECONOMICS. The Best Conditions of Vacuum, Superheat, and Steam Pressure. For normal operating conditions, a great deal can be learned about the most profitable and satisfactory vacuum, superheat, and initial steam pressure for steam turbines from a comparison and study of existing modern power plants. For this purpose a table * is given on the following page in which data are given regarding the vacuum, superheat, and steam pressure of a large number of steam power plants. This table is compiled from fifty-eight turbine plants in America and in England. The figures represent the number of plants working under the conditions stated at the head of each column. There is no doubt that such comparative data of operating conditions are, from a practical viewpoint, of considerable impor- tance. Although these figures were collected in 1904 and 1905, they may be taken to represent very well the average practice of the last few years as well, except that in America there has been a tendency to operate a larger percentage of the plants with from 100 to 150 degrees F. superheat and at about 28 inches vacuum. The Question of the Most Profitable Vacuum. Steam turbine manufacturers are inclined, naturally, because of the obvious advantages of turbines over reciprocating engines for operation at high vacuums, to draw attention to the reduction in the steam consumption when a plant is operated at a high vacuum. Then the question is often raised as to the actual economy considering the increased first cost of the condensers, pumps, and piping, * J. R. Bibbins, in the Report of American Street Railway Association, October, 1904, page 201, and from data collected in 1905 by Messrs. Stevens and Hobart. 273 2/4 THE STEAM TURBINE together with probably larger operating expenses. A great deal depends on the local conditions, particularly on the average temperature of the condenser cooling water. At places only slightly elevated above the sea-level and where the temperature of the water supply for the condensers is very low near the Limits of Capacity in Rated Kilo- watts of Plant. Character of Service. Limits of Vacuum, Inches of Mercury. Limits of Superheat, Degrees F. Limits of Steam Press., Pounds Gauge. 00 IN 2 oo (N 2 ^ ) . = ! = < ' t ! ^^ ~ ^ 10 ^^* ~1 2 ~ T~ ~ ~ ~T r~ ~ ~ ~ 2~ ~ ~ ~ T" ~ ~ - Vacuum in Inches of Mercury FlG. 190. Percentage Curve of the Effect of Vacuum on the Steam Consumption of a Single-Stage Impulse Turbine. turbines of the single-stage type is probably shown very accurately by Fig. 190 (reproduced from Fig. 88).* * In the catalogs of the General Electric Company it is stated that a curve like Fig. 127 is typical for most Curtis turbines. Actually, however, a curve like Fig. 78 is more accurate. Emmett has stated recently in a published communi- cation that "around 27 inches the change in economy per inch is 6.6 per cent.; 28 inches 7.8 per cent.; and 29 inches 9.5 per cent." Parsons states in a paper read before the Institution of Electrical Engineers in 1904 that in a turbine "the benefit derived from a good vacuum is much more than in a reciprocating engine. Every inch of vacuum between 23 and 28 inches affects the steam consumption on an average about 3 per cent, in a loo-kilowatt; 4 per cent, in a 5oo-kilowatt; and 5 per cent, in a i5oo-kilowatt turbine, the effect being more at high vacuum and less at low." It seems very doubtful to the author whether, in general, vacuum corrections can be classified according to the size of the turbine. There are some very large turbines of the Parsons type of which the vacuum correction is less than 4 per cent, per inch of vacuum. STEAM TURBINE ECONOMICS 279 The variation of the steam consumption of a 5oo-kilowatt Westinghouse-Parsons turbine for vacuums from 25 to 29 inches FIG. 191. Curves of Steam Consumption of a 5oo-Kilowatt Westinghouse-Parsons Turbine for 25, 26, 27, 28, and 29 Inches Vacuum. from light loads to overloads is illustrated by the curves in Fig. 191. What might be called a curve of normal vacuum correction 280 THE STEAM TURBINE factors for comparing those of 26, 27 and 29 inches with 28 inches in Westinghouse turbines is given in Fig. 192. Chilton,* after stating that the impression is no longer so common that a high vacuum is necessary to secure good results with steam turbines, says that the difference in economy of Allis- Chalmers-Parsons turbines between 24 and 27 inches vacuum is 5 per cent, per inch. Between 27 and 28 inches the saving is 6 per cent., and between 28 and 29 inches is 7 per cent. 300> ill FIG. 192. Vacuum Correction Factors for Westinghouse Single-Flow Turbines. An idea of the relative quantity of condensing water required for different vacuums may be gained by comparing that required for the usual operating vacuums. For example, with injection water of 70 degrees F., the usual temperature upon which con- denser guarantees are based, it is customary to estimate that to obtain a vacuum of 27 inches about 36 pounds of water will be used for each pound of steam condensed, and about 1.4 times this quantity is required for a vacuum of 28 inches. With injection water at 60 degrees F., which may be considered the winter tem- perature, the quantities required for the foregoing vacuums are * Street Railway Journal, Oct. 19, 1907. STEAM TURBINE ECONOMICS 28l approximately 28 and 34 pounds respectively. Having the quan- tity of condensing water required, the cost of fuel, and cost of water delivered to the condenser, the vacuum best suited to the conditions under consideration may be readily determined. Theo- retically, the effect upon the turbine of reducing the vacuum below that for which it is designed, is to reduce the capacity and to lower the rating at which maximum economy is obtained. The following table * illustrates the percentage gain in economy per inch of vacuum for various vacuums. The close agreement between the actual results and the theoretical values should be observed. The table applies, however, only to turbines using very high steam pressures and superheats. For " land " turbine practice it is serviceable only for comparison. Inches of Vacuum. Gain in Per Cent. 28 27 26 25 Curtis 1 5-i 5- 3-J4 5-2 4.8 4.0 3-05 4.4 4.6 3-5 2.Q5 3-7 4.2 3-o 2.87 3-0 Parsons \Vestinghouse-Parsons Theoretical Effect of Superheating on Economy. The effect of superheat on the economy of De Laval and Parsons turbines is usually stated to be 10 per cent, per 100 degrees F. superheat. This statement is probably very nearly correct for the usual ranges of superheat in practice and is the usual* correction employed by most consulting engineers for correcting steam turbine tests f up to about 150 degrees F. superheat. Some investigations made by Professor Hobart show that the mean superheat correction for Parsons turbines is almost exactly * Mechanical Engineer, Feb. 24, 1906. f The superheat corrections used by the engineers of the Westinghouse Com- panies, by Dean & Main, and by Parsons, are all approximately 10 per cent, per 100 degrees F. superheat. 282 THE STEAM TURBINE 10 per cent, per 100 degrees superheat for all superheats from o to 100 degrees F. Between 100 and 150 degrees superheat it is approximately 8 per cent., and between 150 and 250 degrees is about 6 per cent. It is the opinion of the author that the results of this investigation can be considered quite accurate, as a large number of tests were compared. A curve showing approximately the same sort of variation in the superheat correc- tion of De Laval turbines is given in Fig. 87. Chilton states that tests of Allis-Chalmers-Parsons turbines show that the "incre- ment of saving becomes smaller as the superheat is increased"; adding that for 50 degrees F. superheat the steam consumption is reduced 7 per cent, (at the rate of 14 per cent, per 100 degrees) ; for 100 degrees 10 per cent; and for 150 degrees 12.5 per cent, (at the rate of a little more than 8 per cent, per 100 degrees). According to Kruesi of the General Electric Company IOG degrees F. superheat reduces the steam consumption of Curtis turbines 8 per cent., but "the first 50 degrees of superheat is of greater value than the second 50 degrees."* When steam at about 150 pounds per square inch gauge pressure is superheated 100 degrees F. the total heat of the steam is increased about 4.8 per cent, with an additional fuel expenditure of approximately 6 per cent, if the boiler equipment is good. Now since the steam consumption is reduced from 8 to 10 per cent, for 100 degrees F. superheat there is obviously a saving of from 2 to 4 per cent, in the cost of fuel. Experience seems to show that the best economic results will be obtained with from 100 to 150 degrees F. superheat for tur- bines of the Parsons type, and about 50 degrees superheat for Curtis turbines of more than one stage. In all kinds of turbines of the single-stage impulse type there is probably always a saving * Because curves of steam consumption per kilowatt-hour for varying super- heats (like Fig. 126) were apparently straight lines, most turbine engineers, until very recently, believed that at high superheats the percentage correction was increased instead of being reduced as more recent results show. Since it has been fairly well established that the, specific heat of superheated steam has very low and minimum values at from about 200 to 250 degrees F. superheat, the later results seem to be the more reasonable. STEAM TURBINE ECONOMICS 283 of from 4 to 5 per cent, in fuel cost per 100 degrees F. superheat within the practicable limits of superheating. Although there is much yet to be determined concerning superheated steam, it has been shown by experience in turbine plants that a considerable saving in fuel can be secured by super- heating the steam at least a moderate amount. The greater saving in turbines of the Parsons type over multi-stage Curtis turbines is due to the larger "skin-friction" or disk and blade rotation losses of the large number of rows of blades in Parsons turbines. The curves in Fig. 69 show the very large percentage that these losses are reduced when the blades revolve in dry steam instead of wet steam. When the admission steam to a Parsons turbine is dry saturated the steam in the low-pressure stages will probably have nearly 20 per cent, of moisture, while if it is superheated 1 50 to 200 degrees F. the steam in these stages will be nearly dry. Finally, the use of a high degree of superheat must depend not only on the type of turbine, the load factor, and the size of the units but also upon the nature of the service as regards severe and frequent variations in the load, having in mind the difficulties which have been encountered in the practical operation of super- heaters, steam piping, valves, pumps, and auxiliary machinery. Reasons for the Improved Economy in Turbines and Recipro- cating Engines Due to Superheated Steam. A gain in steam and fuel economy results from the use of superheated steam in either turbines or reciprocating engines. In the turbine the gain comes principally from the reduced fluid friction of the steam moving at a high velocity through passages and blades, some of which have also a comparatively high velocity. In a reciprocating engine the gain from superheated steam is due to the reduction of cylinder condensation, resulting in less loss due to the cooling of the cylinder from the reevaporation of moisture at the lower pressures near the end of the stroke. On account of this cooling of the cylinder ends, the loss due to the "initial condensation" of the steam admitted on the return stroke is often 40 to 50 per cent, of the weight of steam admitted. This loss is partly or 284 THE STEAM TURBINE entirely prevented when the steam is superheated, depending upon the degree of superheat. In a steam turbine there is a similar loss due to condensation, but it is due almost entirely to the mere expansion of the steam. The walls of the turbine casing remain, however, at a practically uniform temperature, so that there is no opportunity for loss through reevaporation of condensed steam. Steam Pressure Best Suited to Turbines. It is the general opinion of practical engineers that probably the most economical operating pressure for the usual power-house services is about 150 pounds per square inch gauge pressure (165 absolute) at the throttle valve, and that a greater saving can always be obtained by the use of a moderate amount of superheat than by increasing the pressure beyond this point. Chilton states that there is a gain of 2 per cent, in steam con- sumption from increasing the steam pressure from 150 to 175 pounds per square inch *' and i per cent, for an increase from 175 to 200 pounds per square inch. But against the saving in fuel due to a reduced steam consumption must be charged the increased cost of piping, valves, and boilers, and also the loss due to increased leakage. Increasing the steam pressure will also increase considerably the cost of the turbine. A "rough and ready" correction used a great deal by turbine engineers is one- tenth per cent, per pound. Speed Variation as it Affects Economy. A steam turbine will give its best economy at some particular speed, just as it has been found to give its best economy at some definite load. For this reason the design of a turbine should be worked out very carefully with velocity diagrams to determine whether at the speed required by the operating conditions it will give the best economy. When- ever any changes are made in the design of a turbine, the manufacturers will always make tests to determine the steam con- * The engineers of the Westinghouse and General Electric companies use practically the same correction for initial pressure. It may be added that the correction for exhaust pressure (back pressure) of non-condensing turbines is about ten times as large as the correction for initial pressure. STEAM TURBINE ECONOMICS 285 sumption at various speeds, and curves like those shown in Fig. 80 are calculated and plotted. If it is found that the turbine has a lower steam consumption at a slightly different speed from that for which it is rated, either the angles of the blades or the pressures must be changed. The reasons for such changes are obvious, because the blade speed has a very definite relation to velocity of the steam in the blades. If the designer is not successful in securing this relation for the rated speed, there will be impact of the steam against the blades and a consequent loss of efficiency. The curve of steam consumption in Fig. 80 shows the change in economy at various speeds. At 2000 revolutions per minute the steam consumption is 19.6 pounds per kilowatt-hour; at 1800 revolutions (rated speed) it is 19.45 pounds; at 1600 revolutions, about 19.8 pounds; at 1400 revolutions, about 20.7 pounds; and at 1000 revolutions, about 24.7 pounds. It will be observed in these curves that the ideal conditions have been secured in the design of this turbine; that is, the steam consump- tion is lowest and the output (load) greatest at the rated speed. Within a range of about 50 revolutions above or below the rating (a total variation of about 6 per cent.) the steam consumption is practically constant. These curves are typical for all good designs of steam turbines. When a speed test is. made of an impulse turbine the best results are obtained as regards the accuracy of the design by running the turbine with a number of nozzles wide open to give approximately full load. The test for each speed can then be made of comparatively short duration, as the steam can be weighed continuously between the first and last tests without interruption when the speed is being changed. With a constant number of nozzles discharging steam the rate of flow will be the same at all speeds. Comparative Economy of Steam Turbines and Reciprocating Engines. To summarize the results of tests on a number of large steam turbines and reciprocating engines the following tables have been prepared. Steam consumption of most of the turbine tests was given in the published data in terms of kilowatt- 286 THE STEAM TURBINE hours or electrical horsepower-hours. In order to make com- parisons with the reciprocating engine it was necessary to reduce all to a common standard brake horsepower-hour. To express all the results in this common standard various efficiencies must be assumed. In the calculations the generator efficiencies given on page 362 were used to obtain the following coefficients to change the steam consumptions from the rate per kilowatt-hour to that per brake horsepower-hour: Rating of Turbine, Kilowatts. Coefficient. 300 and 400 500 1,000 to 3,000 5,000 to 10,000 .68 7 1 .72 73 Mechanical efficiency of reciprocating engines of 30x30 to 5000 horsepower is about 91 per cent.; 1000 horsepower, about 90 per cent. ; and 400 to 700 horsepower, about 89 per cent. In the following tables are given the steam consumptions of a large number of steam turbines and some particularly good reciprocating engines. A great many of the steam turbine tests given are approximately the full load data taken from the tests recorded at the end of the preceding chapter, and some others are taken from Chapter VI. The ratings given in the tables are those for what is generally known by engineers as "full load;" meaning that the turbine can carry economically a load at least 50 per cent, larger than this rating. This statement is necessary because some manu- facturers use a rating based on maximum output. Assuming average values of the corrections given above by various authorities, an approximate equivalent steam consumption has been calculated for each engine at o degrees F. superheat, 28 inches vacuum, and 165 pounds per square inch absolute steam pressure. STEAM TURBINE ECONOMICS 287 STEAM CONSUMPTION OF TURBINES. A= American, E= English, F= French, G"= German, S= Swiss, and W-P Parsons. Westinghouse- Turbine. Rated Power. Conditions of Test. Steam per Hour as per Test. Equivalent Steam per b.hp.-hr. at Super- heat, Deg.F. Vacuum Inches. Steam Pressure Lbs. Abs. r.p.m. Pounds per kw. Pounds per b.hp. o Degs. Sup., 28 ins. Vac., 165 Lbs. Abs. Press* 15-81 15-29 14.71 13-99 *3. 77 15.84 13.17 13-20 14.68 13.25 13.04 12.83 15-52 13-48 13-47 12. 00 ".95 ".95 12.00 De Laval (G) . . . De Laval (G) . . . De Laval (A).... Parsons (E) . . . VV-P(A) W-P (A) Zoelly (G) W-P (A) Curtis (A) ... hp. 3 150 300 kw. 300 300 300 hp. 500 kw. 500 5o 500 500 1000 IOOO 1200 1500 2000 300O 5000 7500 0000 o 5 JOO 107 290 104 10 468 125 207 235 142 06 116 non-con 26.4 26.6 26.6 28.0 28.0 28.7 27.8 28.0 26. 9 26.7 27.0 25.0 28.8 27.5 28.5 27.0 28.8 27.3 29. 6 IOO 114 206 158 160 1 68 201 164 165 168 136 163 179 178 144 181 139 189 i93 . 194 39-6 17. 70 15-17 15. 70 13.99 12.48 '3- 37 13.28 10. 71 J 4-55 15-05 13-61 15.80 II. 00 12. 67 10. 82 10. 60 9.87 11.03 9. ii 3000 3600 3600 23- 15 18.82 15- 10 20.5 21. 2 21.98 15.30 17. 60 15.02 14. 74 13- 52 15. is 13. oo 3600 1800 1800 2400 1800 Curtis (E) Rateau (F). . . W-P (A). . . Rateau (S) Parsons (S). . . . Parsons (E).. . . Curtis (A) Parsons (G). . . . Curtis (A) W-P (A) Curtis (A) 000 1350 750 75 750 Additional data, from recent tests of steam turbines are given in tables I and II, page 271. Combined Steam Engine and Low-pressure Steam Turbine. The combination units of a large Allis engine with Curtis ex- haust steam turbines (see Fig. i8ic, page 25oc), as installed in the 59th Street Power Station of the Interborough-Metropolitan System in New York, have a rated capacity of 15,000 horsepower * Correction curves in Figs. 87 and 88 were used to correct the De Laval tests for superheat and vacuum and the usual correction of .1 per cent, improvement in economy per pound increase ot pressure. For Parsons and Westinghouse-Parsons turbines the following corrections were used: Superheat (300-1000 kw.) 10 per cent.; (1200-7500 kw.) 8 per cent, per 100 degrees F. Vacuum (300-1000 kw.) 4 per cent.; (1200-7500 kw.) 3 per cent, per inch. Pressure .1 per cent, per pound, and the following for Curtis, Rateau, and Zoelly turbines: Superheat, 8 per cent, per 100 degrees F. Vacuum (26-28 ins.) 7 per cent.; (28-29.5 ins.) 8 per cent, per inch. Pressure .1 per cent, per pound. t Referred to 30 inches barometer. 288 THE STEAM TURBINE and give a steam consumption of 13.19 pounds per kilowatt- hour (about 8.74 pounds per i.h.p.-hour) with steam supplied to the engine initially dry saturated (no superheat), 194 pounds per square inch absolute pressure and exhausting from the turbine at 28.8 inches vacuum, referred to 30 inches barometer. STEAM CONSUMPTION OF RECIPROCATING ENGINES SHOWING EXCEPTION- ALLY HIGH ECONOMY. Steam per Equiv. & Hour. Steam fe Consump- ^ f J tion per d w cq ja b.h.p. at Engine. C S J jj I r.p.m. i.h.p. b.h.p. oDeg. superheat, References. o3 $ 28 Ins. - Vac. and R 165 Lbs. $2 1 3 Abs. Pres- H > 1 sure. Rockwood-Wheelock . . 595 o 25-4 174 76.4 13.00 14.61 14.62 F. W. Dean, Trans. A.S.M.E., 1895. Mclntosh & Seymour. 1076 20 27.1 138 99-6 12.76 14.19 13-89 F. W. Dean, Trans. A.S.M.E., 1898. lavitt Pumping En- 576 o 27-3 191 Si-6 11.20 12-59 13.03 E. F. Miller, Tech- gine. nology Quarterly, Rice & Sargent (Phila.). 420 297 25.8 157 IO2 9 56 10.75 13-39 D. 'S. Jacobus, Trans. A.S.M.E., 1904. Westinghouse (verti- 5400 27-3 200 76 11-93 13-12 13.76 Eng. Record, May cal). 28, 1904. Kerchove Mclntosh & Seymour 3600308 2000! 92 27.6 25-5 146.5 171 86 IOO II 78 11.05 12.95 13-58 Von der Kerchove. (Boston). Allis-Chalmers (New 7500 25-1 190 80 11.96 York). Moabit (Berlin) Erie-Lentz (simple en- 2500 282 223 141 28.1 203 156 85 208 8.96 15.24* il.45t J. A. Moyer in gine) . Power, 1912. Buckeye (compound 142 2 5 6 27.2 196 195 9 65 I0.82J Power, 1912. engine). 1 Effect of Superheat, Vacuum, and Admission Pressure on the Economy of Reciprocating Engines. According to Professor Schroeter* the steam consumption of reciprocating steam engines is reduced about 6 per cent, for 50 degrees F. and about <) per cent, for 100 degrees F. of superheat. Parsons f has shown that in a triple-expansion engine the steam consumption can be reduced only .4 per cent, per inch with an increase of * Storm Bull., Journal of Western Society of Engineers, December, 1903. t Proc. Insl. of Naval Architects, April, 1908; Mechanical Engineer, May i, 1908, and Die Turbine, July, 1905. J Probably World's records for steam engines. STEAM TURBINE ECONOMICS 289 vacuum between the limits of 25 and 28 inches, and at a still higher vacuum there is practically no gain at all. Increased initial steam pressure reduces the steam consumption of recip- rocating engines .1 to .2 per cent, per pound per square inch. A comparison of the two tables shows that in large capacities steam turbines will give, for the same standard conditions, better economy than reciprocating engines.* It is shown that for sizes from 3000 to 9000 kilowatts the steam consumption of turbines is about 12 pounds per brake horsepower-hour at the assumed standard conditions of o degrees superheat, 28 inches vacuum, and 165 pounds per square inch absolute pressure; and that, operating at the same conditions, the steam consumption of the best designs of reciprocating engines is about 13 pounds. Economy of Small Reciprocating Engines and Turbines. Nearly all small high-speed reciprocating engines rapidly deteri- orate in economy, primarily because the valve leakage becomes excessive. Although an engine of this kind will meet the guar- antees of steam consumption in a shop test, it has been shown that very soon they require a much larger amount of steam.f Tests of seven high-speed engines of various types rated at 100 to 200 horsepower conducted by Dean and Wood in 1907 show that the steam consumption of such engines after a comparatively short duration of service was found to vary from 49.4 to 60.5 pounds per kilowatt-hour at full load. These rates are very high when compared with the economy of small De Laval and Curtis turbines as given in Figs. 89 and 128. Parsons stated in 1904 that the full-load steam consumption of turbine-generators of his design under the conditions of 100 degrees F. superheat, 27 inches vacuum, and 155 pounds per square inch absolute steam pres- sure was approximately 25 pounds per kilowatt-hour for one of 100 kilowatts capacity, while that of the 200 and 500 kilowatt sizes * It must not, however, be overlooked that these standard conditions were selected in the first place for comparing the economy of steam turbines. It happens that the vacuum is taken a little higher than is usual in the operation of reciprocat- ing engines. t "Economy Tests of High Speed Engines," by F. W. Dean and A. C. Wood> Proc. American Soc. Mech. Engineers, June, 1908. 2QO THE STEAM TURBINE was respectively 22 and 20 pounds. He stated that the equiva- lent results with dry saturated steam and 28 inches vacuum would be about ten per cent, larger.* It has been shown by repeated tests that the steam consump- tion of these turbines is not materially increased when operated continuously for long periods. Weithammer f states that he made tests of a De Laval turbine-generator when new and after five years of service, and calculated the deterioration in economy to be not more than two per cent. ; and this lower efficiency was probably largely due to wear of the reduction gears. It would appear that the deterioration of Curtis turbines should be even less because of less erosion from steam at very high velocities and the absence of the reduction gears. It is stated that there are cases where De Laval blades have been so much worn as to Tequire replacing in a year.J Such an experience is, however, unusual. POWER PLANT ECONOMICS. The following table prepared by Mr. H. G. Stott of New York is interesting in many of its items. Actual data were used to determine the values under the heads of "Maintenance" and "Operation." The first column is for a plant with compound condensing reciprocating engines operating without superheat, and in all cases the values have been suitably corrected to make the other columns directly comparable with the first. Mr. Stott advocates the use of an exhaust steam turbine to be operated by the exhaust steam from reciprocating engines. By increasing the pressure of the steam supplied a moderate amount as well as superheating it the output of a power plant of the type represented by the first column in the table can be doubled at a comparatively small cost for turbines and boilers. * Trans. Inst. of Electrical Engineers, May, 1904. f Die Dampfturbinen, page 104. % Lea and Meden, Transactions American Soc. Mechanical Engineers, Vol. 25. STEAM TURBINE ECONOMICS 291 DISTRIBUTION OF MAINTENANCE AND OPERATION. (Charges per Kilowatt- Hour.) Recipro- cating Engines. Steam Turbines. Recipro- cating Engines and Steam Turbines. Gas Engine Plant. Gas Engines and Steam Turbines. Maintenance. i. Engine room, mechanical.. . . 2-57 0.51 1-54 2-57 1-54 2. Boiler room or producer room 4.61 4.30 3-25 I-I5 i-95 3. Coal and ash handling appa- ratus 0.58 o. 54 0.44 o. 29 o. 29 4 Electrical apparatus 1 . 12 I . 12 I . 12 I. 12 I . 12 Operation. 5. Coal and ash handling labor. . 2.26 2. II 1-74 I-I3 I-I3 6 Removal of ashes I 06 O 04. o 80 O. 53 O.53 7 Dock rental O 74 O 74 O. 74 o. 74 O. 74 8 Boiler room labor 7 i? 6 68 5.46 I 70 3.O3 9. Boiler room, oil, waste, etc. . . 0.17 0.17 0.17 0.17 0.17 10. Coal 61 30 IT? ?Q 46.87 26. 31 25. 77 n. Water 7. 14 7 TO >-46 3. 57 2. 14 12. Engine room, "mechanical" labo'- 6 71 I 35 4 03 6 71 4 O3 1 3 Lubrication I 77 O 35 I OI I 77 I. 06 1 4 Waste etc O 3O O 3O o 30 O 3O o. 30 15 "Electric" labor 2 52 2 52 2 52 2 . 52 2 . 52 Relative cost of maintenance and operation IOO OO 86 03 75 72 ^0.67 46. 32 Relative investment in per cent. . . . 100.00 82 . 50 77.00 100.00 91.20 That the steam turbine plant has an inherent economy of 20 per cent, better than the best type of reciprocating engine installa- tion is shown by a comparison of the first and second columns. Prices of Steam Turbines. Fig. 206 shows by means of curves the price per kilowatt of the normal full load rating of turbine- generators operating condensing. The prices given are the averages of those given by a number of manufacturers at a time when the cost of foundry pig iron was about $20 per ton. It is estimated that the values given by the curves will be changed roughly about 2 per cent, for a variation of $i in the price of foundry pig iron. Unless some such standard of values is given such results can be of little value a very short time after the curves are prepared. 292 THE STEAM TURBINE Non-condensing turbines cost about 5 per cent, less than con- densing machines. Prices of 25-cycle and 6o-cycle generators are usually about the same. Prices do not include charges for freight and erection, which in the eastern and middle western states are about $i to $1.50 per kilowatt. 50 100 150 200 250 300 1,000 2,000 3,000 4,000 5,000 6,000 Rated Full Load Kw. FIG. 206. Curves of the Approximate Price of Steam Turbine-Generators per Kilowatt of the Rated Full Load Output of the Generator. Mr. W. C. Gottshall, who has very carefully investigated power plant economics, has collected the data on the following page, published in 1903, regarding the probable maximum and mini- mum costs per rated kilowatt installed of a power plant equip- ment of about 10,000 kilowatts capacity.* High-grade power stations of from 5000 to 10,000 kilowatts capacity with thoroughly modern equipments cost usually from $100 to $125 per kilowatt. In a few very large stations with high-grade equipment the cost has been about $60 per kilowatt installed; but for stations under 10,000 kilowatts' capacity the cost is rarely below $90 per kilowatt. A building of modern, factory type of construction (one story steel and glass) costs about $i per square foot of floor space. * Gottshall, Street Railway Economics. STEAM TURBINE ECONOMICS 293 COST OF A STEAM POWER HOUSE AND EQUIPMENT. . Costs per Rated Kilowatt Installed. Maximum. Minimum. Boilers and settings $l 7 . oo 3.00 4-50 6.00 1.50 12.00 2.00 I. 00 32.00 21.00 IO.OO 4.00 6.00 3.00 3-5 15.00 2.OO $9.00 2.50 2.50 2.OO I. 00 4.00 I .00 I. 00 20.00 18.00 2.OO 1.50 3.00 2.00 50 8.00 I. 00 Stokers "Economizers Coal conveyors and bunkers Ash conveyor Piping and covering Feed-water heater Feed pumps Engines or turbines* Generators Condensers including pumpst Switchboard ..... Power-house cables and conduits Incidentals (as concrete floor and traveling crane) . . . Foundations for machinervj Buildings Chimneys and flues Total cost including 10 per cent, for engineer- ing supervision and contingencies (nearly) . . . $158.00 ' $87.00 * This item is about right for reciprocating engines and turbines in 1903, when these data were pub- lished. Fig. 206 shows, however, that the minimum cost for this item is about the same as the present cost of first-class turbines and generators. t The cost of condensers is not included in GottshalPs data. Prices given here are those given by J. R. Bibbins (Report American Si. Ry. Assn., 1004) for a plant operating at 26 inches vacuum. He estimates that the cost of a plant for 28 inches vacuum is 60 per cent, greater. Bibbins' values may be tabulated as follows: COST OF CONDENSING PLANT PER RATED KILOWATT. Inches Vacuum. 26 28 29 Barometric condenser $6 to $7 .50 $0-SO tO $12 $12 tO $15 Surface condenser (including centrifugal lift pump, air cooler, single-cylinder dry vacuum pump, and centrifugal circulating pump; Surface condenser (including wet vacuum pump and centrifugal circulating pump) $7.50 to $10 $12 tO $l6 $12 to $16 $15 tO $20 $15 to $20 Ejector condenser $2 tO $2.5O $3 to $4 $4 to $5 $ Engine and generator foundations cost from $1.00 to $3.50 per kilowatt capacity. Foundations for turbine-generators cost as a rule about one-fourth as much, usually 30 to 40 cents per kilowatt capacity on fairly good sub-soil. 294 THE STEAM TURBINE Itemized costs of the 85oo-kilowatt power plant of the Fort Wayne and Wabash Valley Railway Company at Fort Wayne, Ind., are given by Bibbins as follows: Substation apparatus and buildings are, of course, not included. Drawings and a photograph of this station are shown in Figs. 199 and 200. The double-deck arrangement and the installation of baro- metric condensers designed for a moderate vacuum make the. first cost of this station very low. Dollars per Kilowatt. Building: Including general concrete and steel work, coal bun- ker, smoke flue, condenser pit, coal storage pit, etc IO -97 Boiler plant: Including boilers, superheaters, stokers, piping, pumps, heaters, settings, breechings, and tank 13 .92 Generating plant: Including turbines, generators, exciters, cables, switchboards, transformers, and ventilating ducts 3-55 Condenser plant: Including condensers, pumps, piping, free exhausts, water tunnels, and intake screen 3-9& Coal-handling plant: Including gauntree crane, crusher motors, and track o . 94 Erection, superintendence, and engineering 5-94 Total, excluding property and siding 66. 25 The costs are based upon the following assumptions: Per Cent. (a) Bond interest and taxes 7 (b) Sinking fund, equivalent to 6.43 per cent, depreciation 4.2 (c) Total fixed charges on capital cost 11.2 Depreciation determined by summing the depreciation on the several parts of the plant as follows: building, 3 per cent.; boiler plant and coal-handling apparatus, 10 per cent.; condens- ing plant, 6 per cent.; generating plant, 7.5 per cent.; general average, 6.43 per cent. In calculating the cost of an electric power plant it is necessary to consider the probable life of the plant, so as to make correct yearly reductions for depreciation. Often this is more or less of guess-work on the part of the constructor or owner, and for STEAM TURBINE ECONOMICS 295 this reason the following table from a recent issue of Zeitschrift des Vereines deutscher Ingcnieure is interesting. The figures given are those used by two English public corporations, two English engineers, and a series of figures taken from German technical publications. Conditions in America are somewhat different from those of Europe and the depreciation is usually somewhat greater. ESTIMATED YEARS OF LIFE. Authority. Local Gov't Board. L. Canby Council. Robert Hammond. J. F. C. Snell. German Publi- cations. Buildings 7Q CQ 00 60 66 Boilers ir 2O 20 20 T P Steam engines ir 2C 2O 2O 2s 2C. 1 J 2O Steam turbines 2O Gas engines I 7 Water turbines 22 Dynamos 2O 2O 2S 2 C. Storage batteries . . c 7 2O I C, *a IO Transformers I c 2O I r 20 Switchboards 1C. 2O 2O 2O 2? T ? Electric cables (conductors) I2-IC, I25O 10 I ^60 Electric meters IO Arc lights 7 IO IO *3 1C. Comparisons have been made by L. G. French of the cost of two sizes of turbine-generators with corresponding recipro- cating engine costs. He states that the cost of a 75o-kilowatt turbine-generator with a surface condenser (operating vacuum not given) and including foundations and installation charges was $37 per kilowatt. A similar reciprocating engine plant cost $40 per kilowatt. A 1 5oo-kilowatt turbine-generator with a similar condenser equipment cost $30.20 per kilowatt, including founda- tions and installation, while a reciprocating engine equipped similarly cost $32.40 per kilowatt. * Orrok of the N. Y. Edison Co. states that the company has had 40 small steam turbines of the impulse type in service for five years with practically no expense for repairs. 296 THE STEAM TURBINE It is generally believed by engineers who have done recent work in the equipment of large new steam power stations that it is not very probable that large reciprocating engines will ever again be installed to develop power for electrical distribution. One reason is that turbine-driven alternators are particularly adapt- able for parallel operation. The low first cost and operating expenses of turbine-generator units as well as the saving in the cost of foundations and floor space are also very important considerations. A manufacturer of very large sizes of both steam turbines and reciprocating engines has stated that a large power station if equipped with reciprocating engines instead of steam turbines would cost at least from 35 to 60 per cent, more than a turbine station of the same capacity. The following interesting tests of the power required to operate the auxiliary machinery * needed for a Curtis turbine were reported by the Turbine Committee of the National Electric Light Association in 1905. The data apply to the auxiliaries of * " The quantity of circulating water required for high- vacuum condensing plants must be increased from the old standard of from 25 to 30 pounds to from 40 to 60 pounds of water per pound of steam condensed for moderate temperatures, and from 60 to 100 pounds of water per pound of steam when used at the higher temperatures common to cooling tower practice. In cases of excessive head or quantity of circulating water the bulk of the power required by auxiliaries is due to the circulating pump. The range of power for this purpose varies so widely that the older method of assuming a given type of plant requiring 5, 10, or 15 per cent, of the total steam consumption to drive auxiliaries is entirely in error without an accompanying statement denning conditions under which circulating water is pumped. The power to drive the air pump is dependent somewhat upon the vacuum, but particularly upon the air leakage into the condensing system. It was for some time assumed that the work of the air pump corresponded to removing the air which entered the boiler in solution in feed-water. As a matter of fact, handling the air in solution is the smallest portion of work done by an air pump, the leakage through piping, pipe joints, pores of castings, stuffing-boxes, etc., imposing the greatest duty, the total quantity of air to be handled ranging from ten or fifteen to thirty or forty times the air dissolved in ordinary water. The actual power to drive the air pump should in good practice be less than .00018 indicated horsepower in the air pump cylinder per pound of exhaust steam per hour. As the amount of power necessary to drive the air pump is a comparatively small portion of the total power for auxiliaries a slight error in this quantity will not largely affect the final result." C. C. Moore, Journal of Electricity, Power, and Gas, March, 1905. STEAM TURBINE ECONOMICS 297 one of the 5ooo-kilowatt turbine-generators of the Boston Edison Company. Test i. Test 2. Test 3. Kilowatts on turbine 2 7I3- 28.4 3410. 28.7 4758. 28.6 Barometer 20 . ^3 20. CK 20.06 Boiler feed pump E 13.0 orsepower Usec 27.7 1. 27.4 Circulating pump 60 I 60. I 60. 1 Dry vacuum pump 24. T> 23.2 23.8 Step bearing pump .... 6.4 5.8 C.6 Wet vacuum pump 8.6 Q. 2 9.8 Totals 122 . 3 1^1 . 135.7 Per cent, power of auxiliaries to power of turbine 7 4 2 Q 2. I Per cent, water used by auxiliaries to that used by turbine 8. 4 7-4 c . 7 PRACTICAL DESIGNING OF POWER STATIONS. Special Fields for Steam Turbines and for Reciprocating Engines. Some space will be given here to the economic considerations entering into the design of a modern power station intended for electric distribution of power. There can be no doubt as to the status of the steam turbine in comparison with reciprocating engines for the generation of electrical energy. Practically all the recently designed power stations for electric services are equipped with steam turbine-generators, and in some of the older stations built originally for engine-driven units it is not unusual to see turbines installed to increase the capacity.* There is a marked contrast between a power station equipped with reciprocating engines and one that is turbine-driven. The large and heavy frames, ponderous moving parts, and the large generators of reciprocating engine plants cannot be made to compete successfully with the smaller, more compact, and cheaper turbine units. But, on the other hand, reciprocating * W. C. L. Eglin, Report National Electric Light Association, 1906. 298 THE STEAM TURBINE engines similar to the Corliss type have a field which for a num- ber of years probably, the steam turbine cannot enter success- fully. For irregular loads, suddenly applied, like those of rolling FIG. 195. Comparative Floor Space Required for Curtis Turbines. mills and mine hoists the reciprocating engine has advantages over the steam turbine, except perhaps when the turbine is used in connection with an electric drive. Because reciprocating STEAM TURBINE ECONOMICS 299 water pumps and air compressors are more efficient, at least up to the present time, than centrifugal pumps and compressors, reciprocating engines are invariably installed in waterworks and compressing plants. In cotton and woolen mills and shops, where the power is transmitted by belts, shafting, and ropes instead of by electrical methods, the reciprocating engine because of its slower speed is generally preferable. Stated in a few words, the steam turbine is unrivaled by steam reciprocating engines for driving apparatus which can be operated efficiently at a high speed so that a direct-connected unit can be made. Parallel operation of alternators is greatly facilitated when they are driven by turbines rather than by reciprocating engines. There are always difficulties when reciprocating motion is to be con- verted into synchronous motion. Besides the advantages of a uniform turning moment which makes possible such close speed regulation that it is possible to operate railway, power, and lighting circuits from one turbine, because of its high speed it produces a more powerful regulating force without the use of a fly-wheel than that of any engine-driven units of the same capacity. Where a steam turbine is installed in a plant with piston engines or water-wheels its inertia or fly-wheel effect has a steadying effect on the whole system. As an example of this inertia effect it is stated * that a 35oo-kilowatt Curtis turbine and generator has at the rated speed (750 r.p.m.) a " storage energy " of 30,600,000 foot-pounds which is sufficient to enable the machine to carry at any load an additional load equal to the full rating for about .75 of a second with a drop in speed of only 3 per cent, and without additional steam. This machine could carry a momentary increase of load of half the rating for 1.5 seconds. Floor Space for Power Plants. " Compactness " expresses well the primary requisite for the economical design of modern power stations. The small space occupied by a Curtis steam turbine compared with that required for a reciprocating (Corliss) engine * A. H. Kruesi, Proc. American Street and Inlerurban Railway Association, 1907. 300 THE STEAM TURBINE of the same capacity is well shown by Fig. 195. Floor space occupied by Westinghouse turbine-generators is given by the curves in Fig. 196, showing the number of square feet occupied per kilowatt or per brake horsepower. Comparisons of the space required for power units are of little value, however, unless the space for the condensing apparatus and auxiliary machinery is also considered. It is probably fair to assume that for the ox * Sp VestL Wt On are/' rtho IOL 'stin 50C >ral Caps et I ise-} 9O-c ?hou W-X. IDi city *toor ^oirso* woe seD WOO men of i Sp* -73 S >KM oubk KM sion '/nit ce p inglt '&> -Fit '<* s ar Tat J"BA ;Ftc >acii w 1 tacit, id / fen. HP.c nv7 y lirJb f ?a/e< rK.l 'jnbir 'ne ^ e ^oia \ $. *o* ' X fe <*. \ / ^ % \ ^ fc >* 3 s \ \ ! ? ^ L ing/f e low Type \ ++^ ^o^. ^ KK QO4 ^~ Doub leFl BH ow 7 o _002 s ~o~ * * *~S *. >L k ^ V: I B.h P t | 1 I I 1 \ 1 I 1 I ! \ I \ | \ FIG. 196. Floor Space Required for Westinghouse Turbines. conditions where a reciprocating engine would be operated at 26 inches vacuum condensers for a turbine plant would be designed for 28 inches vacuum. Now the volume of steam at 26 inches vacuum is very nearly half that at 28 inches. When surface condensers are used, therefore, the very great increase in the size of the condenser equipment for turbine plants is very obvious. For this reason there has been a tendency in recent years to install barometric or the open type of jet condenser for steam turbines. A very recent installation of Westinghouse-Parsons turbines, barometric condensers, and Stirling boilers is illustrated in STEAM TURBINE ECONOMICS 301 -- - iL.- y __._ = - ! ^ss^=n^= > .-=s.?s=sa=irt^llJJ.J FIG. 197. "Double-Deck" Design of Power House Equipped with Horizontal Turbines. Fig. 197.* The important features of this design are the placing of the turbines above the boilers and condensers, the use of * J. R. Bibbins, Ttans. Am. Inst. Elect. Eng., 1908. 302 THE STEAM TURBINE barometric condensers, and the low total cost of power house and equipment. It is probably one of the most compact arrange- ments possible in a steam turbine plant consistent with high economy in operation. In this "double-deck" design the horizontal turbine and barometric condensers are at their best advantage as regards compactness and efficiency.* The connections between the turbines and the condensers are short and direct, which obviates the losses occurring where there are bends in these connections, and the cost of large exhaust piping is saved. The atmospheric relief valve of the turbine is placed between the floor girders, so that it was possible to make the distance between the floor level and the condenser head only 2.5 feet. For stations not at tide- water, turbine plants are usually operated at a moderate vacuum of between 27 and 28 inches. Barometric condensers are now being made to maintain this vacuum without the use of auxiliary dry-air pumps. With surface condensers by far the most compact arrangement is obtained by installing Curtis vertical turbines with a condenser base. By this arrangement a very direct connection between the turbine and the condenser is secured; but in places where there is likely to be trouble with leaky tubes most engineers will prefer a condenser separate from the turbine. Drawings showing the cross-section and plan of a design for horizontal turbines, Babcock & Wilcox boilers, and barometric condensers are shown in Fig. 199. An exterior view of the same station showing the coal-handling equipment is illustrated by Fig. 200. Plan and elevation of a power station with the turbines and boilers on approximately the same floor level are represented by the drawings in Figs. 201 and 202. The first of these figures is particularly interesting because it shows very clearly the arrange- * A similar "double-deck" arrangement has been proposed for power plants operated by horizontal gas engines and producers. In such a design, where pro- ducers, scrubbers, and all auxiliaries are placed on the second floor, it has been shown that the ground-floor area was only 2.25 square feet per kilowatt. STEAM TURBINE ECONOMICS 303 TYPICAL FLOOR STRUCTUSC FT. tV/CTNC STATION LONGITUDINAL FLOOR ELEVATION TRANSVERSE FLOOR FLEVATION FIG. 199. Longitudinal and Transverse Sections of Power Station. 304 THE STEAM TURBINE ment of the auxiliaries in stations equipped with surface con- densers. Fig. 203 is intended to show particularly the piping arrangements for a typical power station having the turbines and auxiliary equipment in a room adjoining the boiler room. FIG. 200. View of the Power House Shown in Fig. 199. OILING SYSTEMS FOR STEAM TURBINES. A perfect oiling system is obviously a necessity for any machinery operating at a high speed. The efficiency of turbines of the Parsons type depends largely on the smallness of the radial clearances between the rotor and the casing. Now if there is any displacement of the rotor with respect to the casing, caused, for example, by the melting of the white metal in one of the main bearings, the blading might be entirely torn or "stripped," and the turbine would probably be out of service for several weeks. In Curtis turbines with vertical shafts, on the other hand, very serious results might occur if the flow of oil to the step bearing should be interrupted. STEAM TURBINE ECONOMICS 305 FIG. 201. Plan and Elevation of Turbine and Condensing Plant. 306 THE STEAM TURBINE STEAM TURBINE ECONOMICS 307 E '3 W I C/3 1 is '3 W I 308 THE STEAM TURBINE There are two usual methods of lubrication for steam turbines: (i) the central system and (2) the single unit system. In the case of the central system an oil tank is placed at a high point in the building and the oil flows through pipes by gravity to the bearings of the turbine. By means of a "parallel" system of piping any number of turbines can be supplied from one oil tank. The oil leaving the bearings flows into a suitable filtering apparatus provided with cooling coils from which it is pumped back to the main supply tank. The chief objection to this system Is the danger of a total shut-down of the oiling system caused by a poor joint or a broken pipe between the supply tank and the turbine bearings. The alternate system, in which each turbine has its own oil supply and pump, has the advantage in that it assists in reducing the risk of a total shut-down of the plant to a minimum, and if the oil is spoiled in one turbine, due to being mixed with water or being overheated, the entire supply of the station is not ruined. Until recently nearly all manufacturers of Parsons turbines supplied their machines with plunger reciprocating oil pumps. In this respect an innovation has been introduced in Westing- house turbines by the use of a rotary oil pump shown in Fig. 204. In the drawings shown here there are two sectional views of the pump. A worm gear on the turbine shaft transmits power to the pump by means of the gear wheel 10. The direction of rotation of the shaft and of the flow of oil is shown by arrows in the sections. The pump cylinder and its rotor are not con- centric, and metal strips, backed by springs, are inserted into slots in the rotor. These strips are forced out by the springs to touch the inside of the pump cylinder in every position, so as to form pockets into which the oil enters on one side and is dis- charged from the other side. Similar rotary pumps are very generally used for all kinds of engineering services. A suitable oiling system for a Curtis turbine (including the step bearing) is well illustrated diagrammatically by Fig. 205. A large storage tank, shown at the right-hand side of the figure, STEAM TURBINE ECONOMICS 309 is fitted with suitable straining devices and a cooling coil. It is usually located low enough to receive oil by gravity from all parts requiring lubrication. Oil from this tank flows to a pump from which it is discharged at a pressure about 25 per cent, greater than that required to sustain the weight of the shaft and wheels on the step bearing. A baffler in the form of an adjustable 12 FIG. 204. Westinghouse Oil Pump. spiral inserted in the pipe leading to the step bearing serves to regulate the oil supply. Another line of piping is provided for oiling the upper parts of the turbine.* This line of piping is provided with a reducing valve and an air chamber partly filled with compressed air to maintain a constant pressure necessary for the hydraulic motor operating the valve mechanism. Drain * Oil pressure on the upper bearings is about 60 pounds per square inch. THE STEAM TURBINE TO 5PRLMG QUALIZP OR ACCUMULATOR FIG. 205. Oiling System for a Curtis Turbine. STEAM TURBINE ECONOMICS 311 pipes from the upper bearings and from the hydraulic motor dis- charge into a common receiver in which the streams are visible, so that the oil distribution can be always observed. At some point in the high-pressure system adjacent to the pump a device is usually installed to equalize the discharge of oil from the pump. Ordinarily Curtis turbines are provided with a small spring accumulator for this purpose, except for cases where weighted storage accumulators are to be installed. A storage accumulator is usually recommended for large power stations. It can be arranged so that it will normally remain full, but will discharge if the pressure fails, and start automatically auxiliary pumps. Piping for Superheated Steam. Much of the trouble resulting from the use of superheated steam is due not so much to want of strength as to the want of elasticity in the parts affected. These troubles are due particularly to the unceasing variations in temperature resulting from fluctuating loads rather than from high temperatures. As it is possible for water to exist in the liquid state in superheated steam, the variations in temperature may produce a spraying of highly heated surfaces, which greatly increases these difficulties. Changes in the design of pipe fittings, valves, boilers, and superheaters should be made to allow for this abnormal condition. It is desirable to use annealed steel castings in place of cast-iron for fittings and valve casings, and the use of copper for internal parts of valves and gaskets should be avoided. Low velocities in steam piping, which have become customary on account of the pulsating flow of reciprocating steam engines, are not suitable for superheated steam. Since flexibility is so important a consideration in piping for super- heated steam, it is necessary to use comparatively small sizes of pipes and fittings. In Curtis turbines, Kruesi states, a velocity of at least 140 feet per second (about 8500 feet per minute) is desirable for dry sat- urated steam. Now if the steam is superheated 100 degrees F. the volume is increased 15 per cent., but "the velocity in the pipes will be substantially the same on account of the reduction 312 THE STEAM TURBINE in the steam consumption of the turbine. 57 Although this statement is not quite accurate because the steam consumption of Curtis turbines is usually reduced only 8 to 10 per cent, per 100 degrees F. superheat, it is an important observation that the size of piping should not be increased in proportion to the increase in volume of the steam due to superheating. CHAPTER XV. STRESSES IN RINGS, DRUMS, AND DISKS. Design of a Bucket Band or Ring. A ring or band is one of the simplest means of fastening together a number of separate pieces attached like the blades of a turbine wheel to the circumference of a cylindrical surface. Such bands are always made a little wider than the blades, especially at the side where the steam enters, so that the edges of the blades may not be easily damaged in transportation and from insufficient axial clearances when the turbine is operated. These bands are very serviceable in taking care of loose buckets which otherwise would be troublesome. The band serves to bind the blades together as a whole, making the blades with weak attachments to the wheel as good as the strongest. The band assists in making a row of blades of uniform strength.* The design of such a ring revolving at high speeds should be determined by careful calculations; but the theory underlying the design of such a ring serves also for the design of turbine drums and disks. Centrifugal forces more than any other considerations deter- mine the design of a blade ring or band for strength. These forces produce, of course, tension and a resulting expansion of the ring both of significant importance. The centrifugal force (CF) in any sector (W pounds) of a * In a Parsons type it cannot be assumed, however, that because the blades can be made stiffer by the use of a band or shroud ring it is possible to reduce radial clearances below the normal amount and at the same time reduce leakage around the blades. There is reason for believing that radial clearances should be increased for satisfactory operation when the "band " construction is used unless the relative expansion of the metals in the ring, blades, drum, and casing is very carefully adjusted. 3*3 THE STEAM TURBINE freely rotating ring of radius r inches, velocity V feet per second, with an angle 6 subtended by the sector, is CF = where g is the acceleration due to gravity. This centrifugal force tending to expand the ring by increasing its circumferential dimensions sets up stresses which, for the purposes of calculation, may be represented by tan- gential forces at the ends of the sector. These forces are necessarily equal for equilib- rium and are shown as T and T in Fig. 208. If the breadth of the sector is represented by m inches and the radial thickness by n inches, then the area of the section over which this stress is distributed Fl0 ' 2 8 ' Fo ' ces ' *? Blade Band or Shroud Ring. is mn square inches; and if S is the unit tensile stress in pounds per square inch, each tan- gential force is expressed by T = mnS. This force T on the section is tangential, and since the radial centrifugal force (CF) must be equilibrated by an equivalent radial force TO * or for equilibrium CF = T0 f, WV 2 =mnS0- * This relation is obvious from the geometry of the figure. It is, of course, not quite accurate, but very nearly correct for small values of 6. f It cannot be assumed that at the moment of rupture the stress will be dis- tributed between the two sections. The assumption made in the equations is, however, very much on the safe side. STRESSES IN RINGS, DRUMS, AND DISKS 3 1 5 Now if z is the weight in pounds of a cubic inch of the material of the ring, the length of the sector (Fig. 208) is rO inches; then (29) This equation shows that the unit stress in a blade ring or band depends only on the weight of the material and on the peripheral velocity. The last equation can also be expressed in another form, remembering that y _ 3.1416 dN 60 X 12 ' where d is the diameter in inches to the central line of the ring and N is the number of revolutions per minute. Then if we make the approximation of TT* = 10, we have S = g x 4320 ^2Tvr2 (30) Equations (29) and (30) are generally used for the design of shroudings and overhanging rims. When such rings are per- forated with small holes for the riveting of blades, bending and shear are produced. The stresses due to this bending and shear are, however, small and do not in practical cases often exceed 400 pounds per square inch. Sometimes rings called "segments" (Fig. 115) are put on the edge of wheel disks and the blades are attached to them. In a construction of this kind the ring must not only restrain the cen- trifugal force due to its own weight, but also part of that from the weight of the blades if they are not tightly fitted. 3l6 THE STEAM TURBINE If the following symbols are assumed: r = radius to center line of blades, in inches, d = diameter to center line of blades, in inches, w = weight of blades in pounds per foot of length of the circumference measured to the center line of the blade ring, rO = length of a short segment of the blade ring to be calcu- lated, in inches, V = peripheral velocity of blades, in feet per second, W = weight of the blades, in pounds, of a segment r # inches wr ft fl long or "- pounds, then the centrifugal force at the blade ring due to the weight of the blades alone is 2 wr 0V 2 w0V 2 Then if T is the tangential force in the blade ring due to the weight of the blades, and S is the corresponding unit tensile stress in pounds per square inch, C = T # = mnS and = mnS 0, g S = ^, gmn or g W7r 2 d 2 N 2 wd W wd W ,, 518,400 gmn 51,840 gmn 1,669,200 mn * then the total stress S t due to the weight of the ring and of the blades is zd 2 N 2 wd 2 N* 139,100 1,669,200 mn STRESSES IN RINGS, DRUMS, AND DISKS 317 If a blade band or shroud is made in a solid ring and is shrunk on the outside of the blades, as is sometimes the case, then the elongation of the ring due to the centrifugal stresses must be allowed for. In other words, the ring must be made small enough so that there will be a tight fit at the highest speed that will ever be attained.* Design of Drums for the Rotors of Reaction Turbines. The blades of steam turbines are, as a rule, fastened to a cylindrical drum or to one or more disks. The drum construction is used where there is a large number of stages with a small drop of pressure between the successive stages and usually comparatively low peripheral speeds. Thus the rotor of a Parsons type is made up of a number of drums of different diameters, increasing in size toward the low-pressure end. The drum diameters are determined by the blade speed which is selected by the designer to give approximately the best efficiency for the velocity of the steam in the stages of each section of the rotor. Calculations to determine the thickness of a section of the drum are the same in principle as for a blade ring as explained in the preceding paragraphs. The thickness of the drum shell is most simply determined by making calculations in the following order: (i) Calculate the stress in the cylindrical shell of the drum due to its own weight by equation (30). This stress can be determined immediately because it is independent of all dimen- sions of the drum except the diameter of the shell at its center line. It is assumed, of course, that before the thickness of the * If 5 = elongation per inch of length, S = the unit stress Ibs. per sq. inch in the ring at the maximum speed attained, E = modulus of elasticity in Ibs. per sq. inch, then s = , and the total elongation of circumference is inches. This means then that the circumference of the ring must be made - inches tL, smaller than if not subjected to centrifugal stress. A very common construction is, however, that of making the ring in segments of about 2 feet in length and riveting the blades to these segments. 3l8 THE STEAM TURBINE metal for the drum is to be determined, the blades have been designed so that their weight can be calculated. (2) Allowable unit tensile stress must be determined. In this connection the factors to be considered are the qualities of the material to be used (see pages 337 and 338) and the grade of work- manship that is available. In some shops in Germany where very expert workmen can be secured and the material is carefully selected and unusually good, a factor of safety as low as three is sometimes used. Manufacturers of De Laval turbine wheels make the limiting factor from four to five; but for average American practice a factor of safety of less than five should not be considered. If nickel steel is to be used of which the ultimate strength is say 120,000 pounds per square inch, with a factor of safety of five, the allowable total stress in the drum shell would be 24,000 pounds per square inch. Now if the stress due to its own weight, of which the calculation has already been indicated, is still represented by the symbol S, and the total stress allowable by S ty then the permissible stress resulting from the weight of the blades S is S = St - S = 24,000 - S. (3) The thickness of the drum shell can now be calculated by equation (31). Since S is now determined and d , N, andw* are given by the dimensions required for the design of the blades, the thickness n can be easily calculated. Equation (31) can be written in the form (33) Since the weight of the blades has been calculated for only one row, the dimension m is the distance between the center lines of successive blade rows on the drum. * Blades made of bronze, zinc, copper, or similar alloys weigh about .30 pound per cubic inch, and steel weighs .28 pound per cubic inch. STRESSES IX RINGS, DRUMS, AND DISKS 319 Example. The following data regarding the shell of a section of a turbine rotor are given by the drawings accompanying the blade design. Diameter at root of blades (approximately = d) 25 inches Diameter at center line of blades (d ) 30 inches Revolutions per minute 2000 Weight of blades in one row, per foot (w) 5 pounds Weight of a cubic inch of material of shell 28 pound Distance between center lines of successive rows of blades of the drum 3 inches The stress in the shell (S) due to its own weight, by equation (30), is .28(25) 2 (2000 ) 2 . , S = - -*- = ^030 pounds per square inch. 139,100 S = 24,000 5030 = 18,970 pounds per square inch. 5 (so) 2 (2000 ) 2 . , mn = ^ Vs7 ' v *-f = .57 square inch. 1,669,200 X 18,970 But m = 3 inches; then n = -57 -*- 3 = ^9 inch - The sections of the rotor are usually supported on disks attached to the shaft. In another paragraph relating to the design of disks, the strength of such forms will be discussed. It should be remembered that, compared with impulse turbines, the peripheral speed is always kept low.* Drums are almost always used for reaction turbines, and separate disks or wheels for impulse turbines. Fig. 209 is an exact copy of the shop drawing of the rotor of an Allis-Chalmers (Parsons type) turbine. It consists of a central cylinder upon which rings are fitted as shown. These rings are made of steel and are forged as a solid ring. The webs -are formed by cutting away the superfluous material in the sides with a lathe. In this type of rotor the central cylinder must be made of sufficient strength to resist the usual torsional stresses * The peripheral velocity of drum types should not exceed 400 feet per second. Impulse wheels, however, are sometimes designed to operate at 1200 feet per second. 320 THE STEAM TURBINE in a " hollow " shaft. The construction of the drums of typical Westinghouse turbines is shown in Figs. 107, 109 and 184. In impulse turbines where all the expansion of the steam takes place in nozzles placed in diaphragms, or partitions between the stages, there is a large drop in pressure between any two stages, and therefore leakage of steam between the stages will be much greater than with the small pressure drop in the reaction type. The fewer number of stages in the impulse turbine necessarily FIG. 209. Section of the Rotor of a Parsons Type of Reaction Turbine. increases the velocity of the steam passing through the blades and at the same time the most economical wheel speed. Within practical limits, wheel speed should always be increased with steam velocity in good designing. Stresses at Right Angles to Each Other. To determine the stress in flat disks a refinement in the calculations is sometimes necessary in order to obtain more accurate values than those secured in the preceding calculations for the stresses in rings and drums. If, for example, two forces R and T act at right angles to each other, theoretical conditions of elasticity show that the maximum stress or elongation is never quite equal to that due to either of the two forces if acting alone. In other words, an elongation in the direction of the line of action of the force R produces a contraction in the direction of the force T.* Thus * This phenomenon is easily observed in a piece of india-rubber. A force in one direction producing an elongation will produce also a contraction in the direction at right angles to the greatest elongation. STRESSES IN INGS, DRUMS, AND DISKS 321 if the elongation due to the force R is s r per unit of length we have the relation s '* c J K Sr ~E where S r is the stress in pounds per square inch and E is the modulus of elasticity of the material. The reduction (s nr ) of the dimension at right angles or normal to the direction of the force producing the elongation is proportional to the force itself and also, of course, to the stress. Then kS r S nr = kS r = , hi where k is a constant and has the value of .3 for metals of a homogeneous structure, such as are usually required for the manufacture of machines. The force T in the same way produces elongation -I ';- and a reduction at right angles (s nt ) (in direction opposite to the elongation due to Sr), _kS, ' E The net elongation in the direction of the force R is Also the net elongation in the direction of the force T is S t kS t 1 " S " r= E" = E (S '~' 3Sr) - When the two stresses at right angles are nearly equal, as in the case of the disk now under consideration, the elongation is, from the results above, only .7 of that resulting from either force * See Greene's Structural Mechanics, pages 7 and 184; Church's Mechanics of Engineering, p. 203. 322 THE STEAM TURBINE acting alone. It follows also that when the stresses are nearly equal the stresses which are, of course, proportional to defor- mations are also only .7 of that calculated from only one of the forces. This effect of forces at right angles to each other will be applied in the discussion of the stresses in disks. Mathematical Treatment of Stresses in Disks. Fig. 210 shows a section of a turbine wheel cut out (i) by two radial planes making the angle with R + d R each other, and (2) by the cylindrical surfaces with radiuses of r and r + dr. The two other bounding surfaces are the sides of the disk. The thicknesses of the disk are t at the radius r, and t + dt at the radius r + dr. If this sector is rotated about the center it develops the centrifugal force (CF). Acting on the surfaces of the sector are also the forces R and R + dR in the radial direction and the forces T, T in tangential directions. The two tangential forces T, T form the angle 180 6 degrees with each other, and their result- ant is approximately TO when 6 is a small angle. We have, then, Forces acting outward = R + dR + CF. Forces acting inward = R + 10. If we call the unit stress in the radial direction S r and in tan- gential direction S<, then at a section at radius r (if all the dimen- sions are in inches) the following relations result: R = r0tS r , R -f dR = (r + di) (t + dt) (S r + } 2 ) is expressed by the following, form if a is the area of the rim section in square inches and E is the modulus of elasticity in pounds per square inch, J * Because the contraction of the cross-section due to stresses at right angles (page 320), has been neglected in the derivation, equation (35) should not be used for values of allowable unit stress less than 15,000 pounds per square inch, as it gives thicknesses at the center, for low stresses, which are sometimes considerably too large. Practical designers who are required to use unusually low stresses for disks will find a suitable discussion in Jude's The Theory of the Steam Turbine, pages 188 to 204. wV 2 f Centrifugal force due to a weight of a cubic inch at r 2 is - , which becomes S r z when multiplied by the area of the rim section. It is easily shown that the tensile stress in a thin cylinder is ?-&-* a where s is the elongation per unit of length and a is the area of the section. Then the total elongation of the circumference (^) is and the radial elongation (>l) is i p ,_ *_* 2?r Ea 326 THE STEAM TURBINE Since the radial and tangential stresses in the disk have been made equal in the original assumptions, the unit elongations in every direction must be equal, so that the linear expansion in the length r is where k is the coefficient of the contraction of the cross-section for stresses at right angles (see page 321). For conditions of equilibrium obviously X 2 = A t , and substi- tuting equation (37) in (38) and equating to (39) we have Usually the percentage error from writing r t for r 2 is very small, so that we have in simpler form, from which either b 2 or t 2 can be solved. In most cases, how- ever, t 2 is determined by the blade dimensions, so that b 2 is expressed thus: (40) Vl 2 - (l-k)S r V t 2 -.78,. In this equation the stresses are in pounds per square inch, V l and g are in inches per second, t t , t 2 , r t , and b, are in inches. Minimum Thickness of the Disk. The thickness of large disks at the smallest section is not determined by the allowable stress but by the requirements for safe transportation and by the liability of thin disks to become distorted and unstable in balance. Disks about 5 feet in diameter should have a minimum thickness of from .4 to .6 inch, depending on the quality of the material and the STRESSES IN RINGS, DRUMS, AND DISKS 327 speed for which they are to be used; and for disks 10 feet in diam- eter the minimum thickness should be from .7 to 1.25 inches.* The breadth of the rim (b 2 ) calculated by equation (40) is the maximum value allowable, but the breadth can be made, of course, less than that calculated. There will be a smaller radial force at the rim of the disk than is necessary to produce the uni- form radial stress S n and the disk will not be one of uniform strength. The stress at the center will be reduced very much less than that at the smallest section. If now equation (40) is used to calculate the minimum thickness tj, with an assumed value for b 2 suitable for the design, negative values may be obtained. In this case a smaller value of S r must be used in the calculation. Limits for S r can be easily determined by putting S 2 = o in (40); then wV 2 ci which is the tangential stress in a freely rotating ring or is the usual "fly-wheel" formula when k= o. Practical Example. Design of the Rim of a Disk Wheel. A disk wheel 50 inches in diameter is to be designed for an im- pulse turbine to operate at 3000 r.p.m. The minimum thickness (t t ) is .4 inch, and nickel steel is to be used with an allowable stress of 28,000 pounds per square inch, which weighs .28 pound per cubic inch (w). Approximately the radius (r x ) at the inner edge of the rim is 25 inches, so that V is 7860 inches per second (about 450 miles per hour). The wheel is to carry two rows of blades, so that the thickness of the rim must be made about 3.5 inches. The weight of these blades is equivalent to a solid ring of steel around the rim .3 inch thick, f The weight of the * Minimum thickness for a wheel 3 feet in diameter is about .25 inch. An approximate rule for the minimum thickness of disks is t m i n = .008 d to .01 d, where d is the diameter in inches. f Centrifugal force of the blades on a wheel is probably most simply determined by this method of calculating from a drawing showing the dimensions of the blades the thickness of a solid band or ring of the same weight. 328 THE STEAM TURBINE blades per square inch of the rim surface is .3 X .28 = .084 pound, and the stress S 2 per square inch due to this weight is (take g = 386 inches per second) = 538 pounds per square inch. 25 Substituting these values in (40), 28,000 X - 538 b2 = .28 X ( 7 86o/' - X 25 - 1.5 inches. ~ - ' 7 ) The thickness of the section at the center (t ) is calculated by (35), using the same allowable stress as before for S': - .28 (786o) 2 t = t o e 772 x 28,000^ t = te'* = 2.2 3 t, t = .4 X 2.23 = .89 inch. The expansion of the radius due to the allowable stresses in the disk can be calculated by (39), taking E = 30,000,000 pounds per square inch and k = .3, , I ; ; ^nr^ 1 " : . :" : : : ';. , .7 X 28,000 X 25 , . , L = -* - - - * = .016 inch, 30,000,0000 and the expansion of the diameter is .032 inch. If reaction turbines are to be operated at higher peripheral speeds than 350 feet per second, the stresses due to the cen- trifugal forces are too large to use a free drum construction, so that the drum must be strengthened with spokes or flat disks. It is considered better practice, however, to divide a drum into short sections, and calculate each section by the method explained here for disk wheels by the use of equations (35) to (39). The Allis-Chalmers Company uses this method for the low-pressure stages of its latest designs as shown in Fig. 209, although the per- ipheral speed of this section of the drum is usually less than 250 feet per second. STRESSES IN RINGS, DRUMS, AND DISKS 329 Practical Example. Design of a Wheel Disk without a Hole. Stresses in disks are difficult because the areas over which the forces are distributed are not readily determined; besides, the forces are not uniformly distributed over any one of the areas to be* considered. The stresses in a disk are calculated usually by determining the force acting on the " boundary" areas of a circular sector imagined cut out of the disk. Such a sector is shown in Fig. 212. The radius is r inches, the elementary radial thickness is dr, t is the thickness of the sector (measured parallel FIG. 212. Forces in a Sector of a Wheel Disk. to the axis of the shaft), and is the angle subtended at the center by this sector. The centrifugal forces cause tangential and radial stresses. If we imagine the disk made up of a series of concentric rings, laid side by side and touching, the tangential forces tend to break the rings in the line of the tangent, and the purely radial forces, on the other hand, will tend, as it were, to break out pieces which would be carried away in a radial direction. In Fig. 212 the tangential and radial forces are shown more simply than in Fig. 210 in the directions to equili- brate the centrifugal force CF. In other words, the tangential and radial forces shown are those balancing the centrifugal forces. 330 THE STEAM TURBINE An actual design of a 5o-inch plain disk of forged steel without a hole at the center for a Riedler-Stumpf turbine is shown in Fig. 213, and the following paragraphs show how the calculations were made. Diameter of the disk (dj is 46 inches (measured inside the rim, which is 2 inches wide). Smallest section of the disk is taken as .5 inch (see page 326). Speed is 4000 revolutions per minute. The allowable unit stress is 20,000 pounds per square inch, and the disk is designed for uniform strength. Weight of the blades is .09 pound per inch of the circumference, producing a centrifugal force of 1840 pounds per square inch at the smallest section of the disk. Now it has been shown* that in a flat disk the stress at the edge due to an external centrifugal load (like blades and shrouds) is superposable by simple addition to the stresses (both radial and tangential) in the disk due to its own rotation. The rim was calculated as in the previous example, and the thickness (t 2 in Fig. 211) was determined by the width required for the blades. From the inner edge of the rim the disk was given a constant thickness of .5 inch till the tangential stress alone as calculated by equation (30) exceeded the allowable limit, f * Jude, The Theory oj the Steam Turbine, page 198. t It will be observed that these approximations are very much on the safe side because of the effect of " forces at right angles " (see page 320). It is probable, however, that whatever the form of the disk (if not abnormally irregular), the stresses at the center are slightly higher than the peripheral stresses. In case of undue racing due to the failure of the governing apparatus or other cause, a disk designed for uniform strength will fly to pieces from the center. A De Laval wheel without the usual "safety groove" near the rim when tested to destruction broke up entirely and projected large pieces through a cast-steel casing two inches thick. When, however, the customary groove was cut just inside the rim, only pieces of the rim were broken off when an excessive speed was reached, and no external damage was done. It is stated by Jude that the metal left between the "safety grooves" of a De Laval wheel is "only sufficient to carry the traction load of the vanes." From this fact the minimum thickness of De Laval disks can be easily calculated, as it is generally stated that the factor of safety at the groove is 5, and the section before the groove is cut is two-fifths larger. Allowable unit stress is probably taken at about 30,000 pounds. per square inch. STRESSES IX RINGS, DRUMS, AND DISKS 331 For some distance from the rim toward the center we have the case of a flat disk. Now in a disk of constant thickness without a hole at the center both the radial and tangential stresses increase from the rim toward the center. At the outer edge of such a disk the radial stress is only that due to the cen- trifugal force of the blades and rim, while the tangential stress is of considerable magni- tude and is always greater than the radial stress, except at the center, where they are equal.* Because both radial and tangential stresses at every section are approximately increased 1840 pounds per square inch by the centrifugal force due to the blades, the net allowable stress is 18,160 pounds per square inch. The point where the "increasing" sec- tion begins is determined then by the following calculations substituting in equation (30): 18,160 X 139,100 .28 X (4000 ) 2 d t = 23.8 inches, or r t = 565, = 11.9 inches. Beyond this point toward the center of the disk the section has been made of uniform strength as calculated by equation (35). The calculation of the thickness where the diameter is 10 inches (r = 5 inches) is given by equa- _ tion (36), FIG. 213. .28X21.2(10)* j 723 X 18,160 Design of a Wheel Disk without a Hole at the Center. in which t^ = .5 inch and e = 2.7183, then, t = V 45 t = 1.56 t t = .78 inch. In the same way the thickness can be calculated for enough points to determine the profile. The section shown in Fig. 213 is a * Jude, The Theory of the Steam Turbine, page 200. 332 THE STEAM TURBINE typical "flat" disk. To facilitate the forging of such disks the profile is not made exactly as calculated but is gradually tapering from the smallest section to the center. The increase in the thickness from the rim to the center is very small compared with many designs, approximating a "concavo-convex" form (Fig. 214). It is argued by some designers that the treatment of the "concavo-convex" forms is entirely wrong and that most prob- ably it is not possible for the stresses all along the central plane to be either equal to or less than those at the rim by merely satis- fying equation (35), and that the metal in the bulging part of such disks has little influence in modifying the stresses in the central plane. Some of the best authorities agree that it seems reason- able that whatever the form of the profile of a disk the "stresses j^r- -] FIG. 214. Typical Solid Disk without a Hole at the Center. in and about the central plane do not differ greatly from those in a flat disk running at the same speed."* A typical solid disk without a hole for bolts or for the passage of a shaft through it is shown in Fig. 214. It was designed for a very much higher speed than the one in Fig. 213, so that it has a bulging from near the center. This design shows an ingenious method for the attachment of the body of the disk to the shaft. It will be observed that the disk is made with a very small sec- tion near the rim, so that the stress there far exceeds that any- where else. If the wheel breaks it will rupture first at this smallest section and the rim and blades will be torn off. When these parts are gone the centrifugal force will be so much reduced on the part of the wheel remaining that there can be still a very * Jude, The Theory of the Steam Turbine, page 204. STRESSES IN RINGS, DRUMS, AND DISKS 333 great increase in speed without further damage. This disk is designed for a factor of safety of about five at the smallest section, and about seven at every other section. In designing a disk for high speeds, obviously a section that gives approximately uniform strength from the rim to the center is desirable. Experience in such calculations has shown that a disk of the shape shown in Fig. 211 fulfills approximately these conditions.* This disk was designed for a speed of 20,000 revolutions per minute. There is a centrifugal force of about .2 pound per inch on the outside of the rim, due to the weight of the blades which are of the irregular shape shown in Fig. 64. Disks with Holes in the Center. Up to this point in the dis- cussion of stresses in disks only designs similar to Figs. 213 and 214, without a hole, have been considered. When, however, a hole is made near the center of a disk the stresses are greatly increased. There are no very reliable methods for determining the stresses in disks of arbitrary shapes with central holes. According to De Laval, any methods for "taking into account the hub influences in the calculation are only rough approxi- mations" to the actual conditions. It can be shown theoretically that a mere pin-hole at the center of a disk makes the tan- gential stress S t at the hole twice that in a disk without a hole. Indeed a small flaw near the center of a disk may seriously affect the magnitude of the stresses. For this reason, steel ingots with any traces of " piping" must not be used for forged disks to be operated at high speeds. For thick disks of the typical De Laval shape when perforated, the exact solution is appar- ently indeterminate. Methods of calculation for such irregular sections have been proposed which depend on the determination of the mean stresses of the whole section. Results from such methods are, however, of no value at all, as it is known that the maximum stresses are often twice the calculated mean stress. * Besides blow-holes and piping in ingots for drop forgings, most makers put holes into the disks for the attachment of tools for removing the disks from the shafts, and for balancing weights. Very few disks are made that do not have some holes. 334 THE STEAM TURBINE The fact remains, however, that disks for turbines are very commonly made with holes in the center for the shaft, and other holes besides are often made for the attachment of tools for forcing the disk from the shaft when the wheel is to be removed. Stress distribution near a central hole of a nearly flat disk can be approx- imately calculated if a disk of comparatively smaller diameter is imagined cut from its center, and this small disk is then assumed to be of constant thickness and subjected to a radial stress at its rim equal to the uniform stress in the large disk if it had no hole. The stresses in this small disk with the g soixx)- 1^ hole can be calculated with some degree of accuracy from equation (34) by put- ting dt = o, since t has been assumed constant in this small disk.* Tangential stresses calculated in this way for a disk 10 inches in diameter with a hole i inch in diameter are shown in Fig. 215. Radial stress is, of course, zero at the center so that it is not important. The large disk (Fig. 213) was designed to make the combined unit stress in the section 20,000 pounds per square inch, and it is assumed, therefore, that the radial stress on the outside of the * Simplified formulas for a disk of constant thickness are given by Eyerman, Die Dampf turbine, pages 88-90 ; Stodola, Die Dampfturbinen, pages 160-161. The algebraic work involved in obtaining equations suitable for calculations is laborious and complicated. Because these equations are not used directly for other calculations they are not given here. This chapter on stresses is not intended to be an exhaustive treatment, mathematically, and the practical designer wishing to use the minimum factors of safety should carefully study the graphical solutions given by Stodola; but he should remember that these methods referred to are only approximations and in a great measure are justified only because they have stood the test when applied in practice. FIG. 215. Variation of Stress in a Disk caused by a Hole at the Center. STRESSES IN RINGS, DRUMS, AND DISKS 335 small disk has this value. The curve shows that the maximum stress at the hole is 40,000 pounds per square inch and that the stress is rapidly reduced as the distance from the edge of the hole increases till it reaches the constant value of 20,000 pounds per square inch, for which the wheel disk was designed. It should be observed, therefore, that the stress at the edge of a hole at the center of a disk is twice that at some distance away from the hole.* It should be carefully noted, however, that this discussion applies only to holes at the center of a disk. Holes near the rim such as are often made for balancing the disk or as a safety device so that the rim will break first in case of excessive speed, would be allowed for in practice merely by the reduction of the section. It is, however, a good practice to make the section at the hub of a disk with a hole at the center of sufficient size to withstand the greatest stress that may come to bear at the normal speed. Fig. 216 shows how the disk in Fig. 214 should be modified that it may be put on a 4- inch shaft. The thickness (z) of the hub will be determined in the usual way as discussed in books on machine design. Only its length (t ) concerns this discussion. Eyermanj and Stodola give elaborate graphic methods for this determination, but they will not be taken up here, as they are of no general interest. For most practical purposes it is satisfactory to make use of the results shown in Fig. 215 and make the length of the boss (t ) twice the thickness at the same section for a disk without a hole. Instead of reducing the section abruptly in proportion to the reduction in the stress the use of a fillet (see curve ab in Fig. 216) of very gentle curvature gives by far the best construction.! Because the distribution of stress is changed when a hole is made in the center of a flat disk, the section where the radial * This has been shown by a mathematical demonstration and the development of suitable formulas by Grubler, Zeit. Verein deutscher Ingenieure, 1897, page 860 ; Kirsch, Zeit. Verein deutscher Ingenieure, 1897, page 798. f W. Eyerman, Die Dampfiurbine, pages 86-98. J Stodola, Die Dampfturbinen, 3rd edition, page 164. 336 THE STEAM TURBINE stress equals the allowable limit must be calculated by a different method from that used for the disk without a hole at the center. To determine from the general theoretical equations for the stresses in disks the diameter where the radial stress in a flat disk with a hole at the center has a definite value is very laborious and almost impracti- cable; but the following approximate and more or less empirical formula for the radial stress in a disk of uniform thickness with a hole in the center can be used conveniently. It is practically the same as that given by Cree and Jude * except that it has been simplified by grouping constants and changing the units to correspond with those used in the other equa- tions in this chapter. If S r is the radial stress in the disk in pounds per square inch at any diameter d t inches, V is the velocity at the periphery of the disk in feet per second, D is the diameter of the disk in inches, and d is the diameter in inches of a hole at the center, then Now if this equation is to be solved to deter- -s|~ mine d ly it can be written i d, 4 +(^|r S r -0'-d')d 1 '- and putting B = fi~^ Sr - D 2 - d 2 J D 2 d 2 = O, (42) FIG. 216. Design of a Wheel Disk with a Hole at and the Center. C = D 2 d 2 , then (43) * Jude, The Theory of the Steam Turbine, page 204. \ STRESSES IN RINGS, DRUMS, AND DISKS 337 This last equation is easily solved after obtaining the values of B and C from the dimensions of the disk and the allowable unit stress. There are two values of d, because the radial stress increases to a maximum value and then decreases to zero at the edge of the hole. The larger value of d t is always taken to determine the design because between the two values of d l the radial stress has its maximum value. In the design for this example S r = 18,160 pounds per square inch, D = 46 inches, d = 4 inches, and V = 500 feet per second. The value of B is then 1358, C is 33,856, and d t is calculated to be 36.5 inches. The section from the 36.5 inch diameter inward toward the center is made of uniform strength and is calculated by the use of equation (36) in the same way as in the preceding examples. Permissible Stresses and Suitable Materials. It is considered safe generally to use ordinary forged or rolled steel, for velocities not exceeding 600 feet per second; and for lower speeds than this limit wrought iron can even be used if it is of exceptionally good quality. For speeds from 600 to 1000 feet per second crucible cast steel can be used. Nickel steel is recommended for turbine disks by the Krupp Company of Essen, Germany. This nickel steel has an ulti- mate tensile strength, of 125,000 pounds per square inch and 12 per cent, elongation before rupture. The elastic limit is about 95,000 pounds per square inch. It is stated by the Krupp Company that they will produce a nickel steel of still higher tensile strength but only about 6 per cent, elongation. With some small forged pieces of this material an ultimate tensile strength of 285,000 pounds per square inch has been observed, with an elastic limit of nearly 225,000 pounds per square inch. All De Laval turbine wheels used in America are made in Sweden of forged nickel steel, which is rather high in carbon. Allowable working stresses must, of course, be left to the judgment of the designers. An engineer of the Krupp Com- 338 THE STEAM TURBINE pany states that stresses in the same direction may be allowed in turbine disks as high as one-third of the elastic limit. Since the centrifugal force and therefore also the unit stress is proportional to the square of the velocity, if a factor of safety of 4 is allowed, the breaking speed of the wheel will be twice the normal speed, and the elastic limit of the material is only about 1.5 times the normal speed. .Excessive stresses at a hole are " dissipated" very materially if a dangerous stress is reached at the edge of the hole. Before rupture can occur there will be an excessive elongation of the material as soon as the elastic limit is reached at the highly stressed section. CRITICAL SPEEDS OF LOADED SHAFTS. With the high speeds at which steam turbines are operated the centrifugal forces due to even a small eccentricity of the rotating masses produce vibrations, excessive stresses, and "springing" of shafts. As the result of the eccentric forces the shaft is bent farther out of line, so that the centrifugal forces and the amount of the eccentricity are increased until the stress set up in the shaft by the bending produces a force equal to the centrifugal force, and the center of gravity and " center of work" coincide. If W is the weight of the rotating mass in pounds, e is the "original" eccentricity of the shaft in inches, x is the eccen- tricity in inches at N revolutions per minute, P is the force applied to the shaft at the point of attachment of the disk which will bend the shaft i inch, within the -elastic limit, C. F. is the centrifugal force of the rotating mass, and k is a constant, then c F = kWN *. gx The bending of the shaft at this speed is x e, so that kWN 2 (x - e)P = (44) gx and e N STRESSES IX RINGS, DRUMS, AND DISKS 339 The increased eccentricity due to rotation is therefore propor- tional to the original eccentricity of the shaft and increases with increasing values of u and hence also of N. When kWu 2 x becomes oo , that is, the deflection becomes exceedingly large, unless prevented, and would break the shaft. It has been shown by Cree * that the critical speed N c (r.p.m.) of a shaft with some flexibility in the bearings carrying a con- centrated load of W pounds is 6. 94 r 2 . /El -' (45) where E is the modulus of elasticity in pounds per square inch, r is the radius of the shaft in inches, 1 is its length (between two bearings) in feet, and a and b are the distances from the load to the bearings, in feet. This formula f is to be used for only a single concentrated load like the single wheel of a De Laval turbine. When there are a number of wheels with possibly also a revolving field of a generator on the same shaft, the problem becomes very complicated if the loads are considered separately. Experience with such calculations has shown that for the cases occurring in practice $ the critical speed can be determined by the following simple equation derived for the case of uniform loading: N c = i55,ooo r'Y^p' (46) where W is the sum of the several loads on the shaft and the other symbols are used as before. * Proc. Physical Society (London), vol. XIX. t In this formula the weight of the shaft is not taken into account. The influence of the weight of the shaft on the critical speed can be easily cal- culated, but in practical cases it may be neglected without appreciable error. J This applies particularly to the cases of Rateau, Parsons, and Curtis turbines and turbine-driven generators and pumps. CHAPTER XVI. GAS TURBINES. THE development of the gas turbine, which should combine the high thermal efficiency of an internal combustion engine with the mechanical simplicity of the steam turbine, has occupied the attention of a number of able engineers from time to time but without unqualified success. Because of the severe conditions due to the very high temperatures of the gases after combustion, there are many difficulties in construction which in a large measure offset the otherwise simple mechanical construction. It may well be said that the designer of gas turbines is between "the two horns of a dilemma." If he tries to utilize the gases at the temperatures resulting from expansion in a single normal nozzle, the nozzles and blades will deteriorate very rapidly, and for the best efficiency the speed of rotation of the turbine must be made too high for utilization for general power purposes with- out the application of reducing gears; and, if on the other hand, he cools the gases by the injection of water or excess air into the combustion chamber to make the temperature of the gases suitable for the materials available for machine construction, the high thermal efficiency stated by the simplest laws of thermodynam- ics* is, of course, not attained. Since the gas turbine is certainly not yet out of the experi- mental stage, although there are commercial applications, it is not out of place to give some space to its history. Probably the oldest form of gas turbine is the ancient propeller * The thermodynamic efficiency of a heat engine is expressed by where 1 1 TI is the initial and r 2 is the final temperature of the cycle. By lowering the value of T v the efficiency is reduced in much greater proportion than the reduction in the temperature. 340 GAS TURBINES 341 mechanism, known as a "smoke-jack," which was used for operating the turnspit* of large open fireplaces. An illustra- tion of this "smoke-jack" is shown in Fig, 218, which is a copy of an old drawing published in Bishop Wilkin's Mathematical Magic in 1680. A similar apparatus is described by Cardan about 1550. This mechanism was placed in the chimney and was driven around by the ascending cur- rent of hot gases from the fire. Its motion was transmitted by gearing and belting to the spit on which the joint of meat was carried in front of . -. .. . . .. FIG. 218. A Chimney Turnspit the fire. The power of this smoke- or "Smoke-Jack " jack" can only be estimated by the work of the turnspit dog which it replaced. It must, therefore, be rated at least one " dog-power. " The earliest attempt to construct a gas turbine on scientific principles was probably made by Stoltze of Charlottenburg, who received a patent for what he called a " hot-air" turbine in 1873. This apparatus consisted of two turbines on one shaft, one acting as an air compressor and the other as a power turbine. The function of one of these turbines was to draw in and compress the air to about 40 pounds per square inch absolute. Part of this compressed air was then passed through a combustion chamber or furnace, where it supplied the oxygen required for the combustion of the gas or oil fuel. Another part went through a heating chamber and was later mixed with the gases of combustion from the furnace. The mixture of gas and air was then expanded in the second" turbine. The useful power developed by such a turbine is the difference between that developed by the gas turbine and that required to drive * Turnspit is the name usually applied to the dog which was used to turn, by means of a suitable mechanical contrivance, a spit or long iron bar, pointed at one end, used to hang up meat to be roasted. 342 THE STEAM TURBINE the turbine-compressor. A turbine designed to develop 200 horsepower has been constructed on this plan, but it has not been commercially developed. It is very doubtful, if all other difficulties were overcome, whether this method of air injection could give nearly as good economy as water injection. (See page 344.) Some attention has been given to the development of the explosion gas turbine, of which a very simple form is shown in Fig. 219. It consists of a combustion chamber E, of which one end is closed by a large valve A opening inward, admitting air through the parts B, B and fuel through tubes F, F opening into the valve seat. The mixture of gas and air is ignited by electric sparks at I, and the products of combustion are discharged from FIG. 219. A Simple Explosion Gas Turbine. the chamber through a small opening J leading into the nozzle N, where air, as shown by the arrows, is mixed with the gases to reduce their temperature before they reach the blades of the turbine wheel W opposite the nozzle. It is a well-established fact that when a mixture of gas and air is exploded there is first a sudden expansion and then, because of the combination of the hydrogen in the burned gases with the oxygen in the excess air to form water, a vacuum is produced. This phenomenon "is applied in this apparatus to operate the valve A, which by the formation of a vacuum is drawn inward to admit another charge of gas and air. It is stated that in such a turbine the explosions will occur very rapidly from 3500 to 5000 per minute so that there is a practically continuous discharge upon the wheel. The efficiency of an explosion motor of this kind is very low because of the lack of compression; but GAS TURBINES 343 its efficient development does not seem to be impossible. If in some way efficient combustion by explosion can be segured without compression, then a most economical power development could be attained with an explosion combustion chamber with the fuel and air valves operated automatically "by vacuum" and the injection of probably comparatively large quantities of water after combustion. Such an apparatus would be simple Oil Pump FIG. 220. Section of a Zoelly Explosion Gas Turbine. indeed compared, on the one hand, with the complicated com- bination of the steam toiler with external firing and the steam turbine, or, on the other hand, with the complex reciprocating gas engine. Fig. 220 illustrates a Zoelly explosion gas turbine. It consists essentially of an explosion chamber C, a turbine wheel W, water and oil pumps, and an air compressor. The pumps and compressor are of the reciprocating type and are driven by the main shaft by means of the worm gears A x and Bj. The valves regulating water, oil, and air admission and the ignition device TI, are operated by the gases, and steams are 344 THE STEAM TURBINE expanded in the nozzle N and impinge upon a turbine wheel W of the, De Laval type. Some of the heat remaining in the exhaust gases is absorbed by water coils R which serve to heat the injec- tion water. In the operation of this apparatus, air is admitted first into the explosion chamber and then the oil, as the air is supposed to act as a shield against back-firing. After the charge has been exploded and the maximum pressure has been reached, the cooling water is injected. The more successful gas turbines, however, are those operating by combustion at constant pressure. In this type the air and fuel (oil or gas) are delivered under pressure to a suitable com- Water FIG. 221. Diagrammatic Illustration of the Combustion Chamber and Steam Coils of a Modern Gas Turbine. bustion chamber A in Fig. 221 which is maintained at a red heat, so that the combustion is continuous. The products of combustion are usually cooled by water which is injected into the nozzle as in the explosion type. The heat energy in the burned gases is converted into velocity in an expanding nozzle N and are discharged at a high velocity upon the blades R of the turbine wheel. Designers of this type of gas turbines have generally assumed that nozzles and wheels of the De Laval type are most suitable, arid their energies are devoted at present to the pro- duction of a suitable combustion apparatus and a high efficiency rotary compressor. Fig. 222 shows a typical small gas turbine set up for a brake test. GAS TURBINES 345 FIG. 222. A Gas Turbine set up for a Brake Test. 346 THE STEAM TURBINE In practice the combustion chamber is lined with carbo- rundum, and to allow for expansion the carborundum is backed with sheets of asbestos to provide a soft and elastic packing. Exhaust gases are usually discharged over a coil boiler L, and the steam which is produced is also delivered upon the turbine FIG. 223. Arinengaud and Lemale's Gas Turbine. wheel by a separate nozzle M. When the turbine is in operation the lining becomes sufficiently hot to ignite the fuel as it is forced into the chamber. A gas turbine of this latter type designed by Arinengaud and Lemale of Paris is illustrated in Fig. 223. It is a machine developing 300 net horsepower at 4000 revolutions per minute. GAS TURBINES 347 A Rateau turbine-compressor shown direct connected to the gas turbine in Fig. 224 has been specially designed and built by Brown, Boveri & Co. of Baden, Switzerland, for use with this turbine. The compressor gives a mechanical efficiency as high as 65 to 70 per cent, and delivers i cubic foot of air per second FIG. 224. Arinengaud and Lemale's Gas Turbine Direct Connected to a Rateau Turbine-Compressor. at a pressure of from 6 to 7 atmospheres. Compressed air is used for starting, and a simple ignition device is used for firing the charge till the combustion chamber becomes sufficiently heated. M. Barbezat, who has now charge of the development of this turbine, states that the total efficiency is not as high as that of reciprocating gas engines; but no data are given. 34$ THE STEAM TURBINE Gas turbines have been applied practically for the propulsion of submarine torpedoes. Formerly some types of torpedoes received their motive power from a rotary motor like a turbine wheel, driven by compressed air. Recently gas turbines have been installed with an obvious gain in power and saving in weight. These gas turbines develop 120 horsepower at 1000 revolutions per minute. The expansion ratio of the nozzles is 8.4 and the weight per horsepower, without the compressor, is 1.3 pounds. It is obvious, then, that great progress has been made recently in the development of the gas turbine; and when the ratio of progress is compared with the time required to bring the recip- rocating gas engine to its present state of development, there is reason for hoping for greater accomplishments in the near future. The gas turbine question includes, however, a number of un- solved problems; but, on the other hand, the sources available for their solution are numerous. The development of these machines will permit the utilization for power of mixtures of air with coal gas, petroleum, or alcohol; and it will also make possible a combination of the explosion motor and the steam turbine for many purposes. The problem is laid plainly before the physicist, the engineer, and the machinist, and to bring about a satisfactory solution will doubtless require all their combined resources. Questions of Theory. The success of the steam turbine naturally directed the attention of engineers to the possibilities of the gas turbine with the expectation of combining the high thermal efficiency of the gas engine with the constructive advan- tages of the steam turbine. As explained in the preceding pages a gas turbine can be operated by either of two methods: (1 ) By combustion of the fuel in a chamber at constant pressure. (2) By an explosion method. Combus'tion at constant pressure seems to be the more practi- cable method and is the one generally adopted.* In the opera - * Theoretically the same efficiency should be secured with either of these two systems of combustion. Combustion at constant pressure is an adaptation of the GAS TURBINES 349 tion of this method gas and air are compressed in separate chambers or compressor tanks to a suitable pressure, usually about 100 pounds per square inch absolute. The gas and air are admitted through separate valves to the combustion chamber, where the gas is ignited and burned at constant pressure. Just as in a reciprocating gas engine, the air is provided to furnish the oxygen to support combustion. After combustion the burned gases escape through a suitable nozzle to impinge on the blades of the turbine wheel. On account of the extremely high temper- atures resulting from the combustion (about 2500 degrees F.), it is impracticable to design a gas turbine with more than one pressure stage and therefore only "nozzle types" can be used. Comparison of Losses in a Gas Turbine and in a Gas Engine. It is reasonable to assume that the radiation and cooling water losses will be about the same for a gas turbine as for a recipro- cating gas engine; and from a practical viewpoint the work required for the compression of gas and air is about the same for combustion at constant pressure as for explosion. After eliminating, therefore, the radiation, cooling water, and com- pression losses the same energy remains for utilization in each of these two prime movers. In gas engines from 20 to 25 per cent, of this energy is lost in the suction and exhaust resistances, engine friction, and the heat loss in the exhaust. Corresponding to these losses in the gas engine, there are in the gas turbine losses due to nozzle, blade, and disk friction, the heat in the exhaust, and bearing friction. The sum of these latter losses in a steam tur- bine would be about 40 per cent., and they will probably be not much different, in the total, in a gas turbine. It is argued in favor of the gas turbine that it is not impossible "to isolate the well-known Brayton cycle. It has been shown that exactly the same thermal efficiency can be secured by such combustion as in the ordinary explosion process if it is assumed that the specific heat of the gases is practically constant and that the final pressure after compression in the explosion motor is the same as the constant pressure of combustion in the Brayton cycle. It follows v then that the ideal gas tur- bine will theoretically operate with the same fuel consumption per unit of power as the ideal four-cycle gas engine. (Cf. Lorenz, ZeiL Venin deutscher Ingenieure, 1900, page 252.) 350 THE STEAM TURBINE combustion chamber internally" so that no cooling water will be needed. The greatest practical difficulty in the way of the successful operation of gas turbines results from the high temperature of the gases at end of expansion. Nozzles can be cooled by water- jacketing, but the wheel blades are liable to rapid deterioration. The necessity for lowering the temperature of combustion is now generally recognized, and the hot water from the water jackets is sprayed into the compressed air supplied for combustion. By this means the temperature in the combustion chamber can be greatly reduced but at a considerable loss, however, in efficiency. It is making, in other words, the efficiency of the gas turbine approach the lower thermal efficiency of the steam turbine. It seems probable that the most promising field for the gas turbine will be found to be in the utilization of bituminous coals forming tar and asphaltum when used for making gas. Such coals cannot be used in the manufacture of gas for reciprocating gas engines, as the accumulation of tarry matter in the cylinders is particularly objectionable. In a gas turbine, however, the gas is burned under pressure, in an enclosed chamber where accumu- lations of foreign matter cause no serious difficulties. It cannot be expected that gas turbines can be commercially successful for general power purposes if a reciprocating compressor must" be used in connection with them, because a gas turbine with a compressor of this type is quite as complicated as the recip- rocating gas engine. Compressors of the rotary type, on the other hand, have usually a very low efficiency; probably in most cases not more than 50 per cent. Decided progress is being made in the successful designing of compressors of the turbine type which will give from 60 to 70 per cent, efficiency. It is not difficult to understand how the net useful work of a gas turbine may be nil under conditions that are not unusually poor. For if the effi- ciency of the turbine is 60 per cent, and the theoretical work of compression is 40 per cent, of the output (which is not an absurd estimate), then with a compressor efficiency of only 40 per cent, the theoretical power absorbed by the compressor is 60 X .40, or GAS TURBINES 351 24 per cent, of the output, or the actual power delivered to the compressor is 24 + .40, or 60 per cent, of the output. And the compressor takes all the power the turbine can supply. It is obvious then that compressors with the usual low efficiencies of the rotary types are not worth considering. BRAYTON CYCLE CALCULATIONS FOR GAS TURBINES. tf ^ > <* IS fi rt e rt *' yj C 5 If, H| 5 . *0 I*. 8*1 "I ft*:! pf E.s 11 11 li IS i c |! a S c a * 1 1 ^08 o C bo Q jS^o o fe fe ^ 1 -| o 3l 13 w C to g fl *l ^ *4J c H t> in g ft.S TJ 3 , C 3 6 gt: 1^* 111 jl Pi s ffi cu- S h K z. Pi Q * * , r F e Case I: 90 1,000 4,665 2,435 .22 115,300 43 Adiabatic com- pression. Per- fect machine with no losses. 90 195 195 495 495 250 1,000 250 1,000 250 i,55 4,860 1,710 5,190 2,040 652 2,008 546 1,562 434 .87 .27 1.07 34 1-38 71,680 131,200 83,740 147,600 98,300 43 54 54 64 64 Case II: Isothermal com- 90 1,000 9,239 5,039 .08 159,100 93 pression with 90 250 1,965 39 79,57o 72 regenerator. 195 1,000 7,376 3,176 .11 159,100 91 Perfect ma- 195 250 1,498 448 .61 79,570 62 chine with no 495 1,000 6,087 1,887 . 15 159,100 87 losses. 495 250 1,176 126 1.03 79,57o 49 Case III: Isothermal com- 56 333 2,147 1,200 74 75,4oo 27 pression with 82 392 2,398 1,200 .68 84,800 3 regenerator. 61 445 2,867 1,619 50 86,485 Actual ma- chine with as- sumed losses. 98 33 61 545 445 569 o 3,233 1,619 2 ,!39 47 .40 36 99,96o 80,155 96,775 35 28 33 Air excess. Case IV: Isothermal com- pression with 79 625 37 1,275 600 67 75,000 13 regenerator. 47 638 36 i, 680 1,000 43 75,000 15 Assumed losses . Cooling water. 352 THE STEAM TURBINE THERMODYNAMIC THEORY OF THE GAS TURBINE. The elementary thermodynamics of the gas turbine involve apparently no new investigations. The problems are princi- pally mechanical and metallurgical. In the above table the efficiencies of various gas turbine cycles are given as calculated by Sanford A. Moss. In all cases it is assumed that the heat of combustion is developed at constant pressure and that the exhaust gases are discharged at constant pressure. Cases I and II refer to theoretically perfect engines, and cases III and IV to engines with probably normal losses. It is assumed for these latter cases that the turbine efficiency is 70 per cent, and of the compressor is 83 per cent.* The efficiency of the regenerator used for the cases of isothermal compression is taken to be 60 per cent. These figures are certainly above the upper limits of possible results in practice. Thermodynamic efficiencies of gas turbines operating with combustion at constant pressure will now be discussed. The equations given are, in most cases, those relating to a perfect gas. The total heat H of a gas at constant pressure may be expressed by the following equation: H = c v T + RT = (c tf + R) T = c p T + constant, where T is the absolute temperature, c v and c p are respectively the mean specific heats of the gas at constant volume and con- stant pressure between zero temperature and T, and R is a con- stant varying for its value with the kind of gas. Fig. 225 represents the cycle of operations when a pound of a mixture of gas and air is compressed and later expanded in doing work. Adiabatic compression is assumed. One pound of the mixture is taken into the compressor cylinder at the temperature T t and volume v and is compressed as represented by the adia- batic O 3 to the temperature T 3 and volume v 3 . In the passage to the combustion chamber it will be assumed for simplicity in * Efficiencies of 60 per cent, for the turbine and not more than 70 per cent, for the compressor would probably be more reasonable. GAS TURBINES 353 the calculations that the temperature drops to the initial temper- ture T . If the total heat contents at the points o, 3, 3', 4, and 5 are represented by the corresponding symbols H , H 3 , H 3 ', H 4 , Volume FIG. 225. Diagram of the Theoretical Action of a Gas Turbine and Air Compressor. and H 5 , the indicated work of compression is the area 0123, which will be represented in heat units by W II IT c 1 3 1 . It is assumed, however, that immediately after compression the temperature falls to T , so that the volume is reduced from v 3 to v/. Now because the point 3' is on the isothermal 3', it is obvious that H = H 3 '. During combustion a quantity of heat Q t is added to the mixture, increasing the temperature to T 4 and the volume to v 4 ; or, in other words, Q,-H 4 -H,'. When the gaseous mixture is expanded the work W l is per- formed, which may be calculated after determining H 4 in the preceding equation; then W t =H 4 -H 5 . 354 THE STEAM TURBINE The heat lost by the exhaust gases in cooling from T 5 to T is H 5 H ; and since this quantity is called Q 2 , we can write W, - (H 4 - H,') - (H, - H.) = Q, - Q 2 , and placing H for H 3 ' it is apparent that W, =H 4 -H.= 0,-Q,. The theoretical discharge velocity V in feet per second at the mouth of the expansion nozzle is calculated from the usual equation, V 2 2g X 778 = Q l -Q,= H 4 -H 5 =c,(T 4 -T 5 ). Using the same symbols as before for the indicated work of compression, the theoretical effective power of the turbine is W e = W, - W c = Q 1 - Q 2 - W c . If the mechanical efficiency of compression is x and the efficiency of the gas turbine, y, is determined in the same way as for a steam turbine, by constructing velocity triangles and calculating the nozzle, blade, and wheel friction losses for a single stage turbine, then the theoretical net power of the turbine is W.' = (Q, - Q 2 ) y - ^- Since the heat consumed per pound of the mixture is Q 1? the total efficiency z of the gas turbine apparatus is ((Q, - Q 2 )y - Y)-" Efficiency of Gas Turbine with Water Injection. If m pounds of water are injected into the combustion chamber just before the expansion begins, an equal weight of steam is found, which it will be assumed is superheated to the temperature T 4 ', which will now be also at the temperature of the mixture, lower, of course, thai) T 4 . GAS TURBINES 355 The temperature of the mixture of burned gases and steam T 4 ' is calculated by solving the following equation: c, (T, - T/)= m }q/ - q, + r/ + c/ (t/ - t S4 )j, where q/ is the heat of the liquid, r/ is the heat of vaporization, and t S4 is the temperature of saturated steam, all at the corresponding pressure P/.* The other new symbols are q;, which is the heat of the liquid at the injection temperature, and c/, which is the specific heat of superheated steam. In this equation t/ and t s4 are ordinary (not absolute) temperatures. The temperature T 5 ' is calculated for assumed adiabatic expan- sion by using the exponent k' calculated from the equation below : c vi 4- in which the subscript i refers to the specific heats of the mixture and the subscript 2 to specific heats of the steam. The temper- ature T 5 ' is used to determine the value of Q 2? which is the quan- tity of heat abstracted from the mixture to cool it from the condition at 5 to the condition at o. It is calculated from the following equation: Q 2 = m ic p ' (t/ -t S5 ) + r 5 ' + q/ - q l + c p (t 5 ' -t ). In this equation t 5 ' is the " ordinary" temperature corresponding to the absolute temperature T/, t S5 is the temperature of satu- rated steam, r 5 ' is the heat of vaporization, and q/ is the heat of the liquid, all at the pressure p 5 '; q is at the temperature t . It will be assumed also that t is less than t S5 and that the latter is * The "partial" pressure of the steam at the temperature T 4 ' can be calculated approximately by the formula p i _ . 47 * A ' 29.3 + 47 m if we assume the constants for the exhaust gases are the same as for air. In the same way the "partial" pressure of the steam after expansion is calculated thus: r.'-p, 29.3 4- 47 m 356 THE STEAM TURBINE less than t 5 ', as is generally the case, neglecting the small quan- tity of heat in the water vapor remaining in the mixture at the temperature t after the burned gases and steam have been dis- charged from the nozzle. We can write the following equations, applying the same methods as for the case without the use of injection water: Y.'d+m) 2g X 778 Total efficiency, z, increases with the pressure of compression to a certain limiting value and then decreases. But for the practi- cable values of compressor and turbine efficiencies (x and y) the values of the theoretical total efficiency are not particularly good. The equations given here for velocity and efficiency can be used to investigate the best operating conditions by varying the pressure of compressions and the quantity of injection water. Another method for reducing the temperature of the gases is to use a large excess of air above the quantity needed to support combustion. The most economical method is probably that of partly vaporizing the cooling water in the water jackets. The advantage of using water is that likewise it does not need to be compressed, and therefore it can be injected into the combustion chamber without the expenditure of much energy. Many of the troubles in the combustion chamber of a gas turbine are difficult to explain. One of the most serious diffi- culties is the occasional missfire of the incoming charge which is soon followed by a violent explosion. It is also difficult to secure smokeless combustion. Improvements are being made, how- J 37 369 370 INDEX PACK Composition of blades ............ 4 ............................... 111-114 Condensing water, quantity of. ........................................ 280 Condensers and auxiliaries, cost of ................................. 293-296 power for ....................................................... 297 Corrections for economy curves ........ . ....................... 126-136, 287 for rotation losses ........................................... .... 118 for steam turbine tests ........................................ 1 26-136 Cost of engines, turbines, and auxiliaries ........................... 291-296 maintenance and operation ................................... 290-291 Critical speed of rotating shafts ................. . ........... 8, 147, 338-339 Curtis turbines .................................................. 182-194 analysis of losses ................................................ 194 blades .................................................. 113, 182, 185 diaphragms ..................................................... 182 emergency stop valve ......................................... 190-191 governor ........................................ 191, 226-229, 243-245 manufacturers ................................................... 182 nozzles ...................................................... 27, 184 oiling system ................................................ 36~3 1 1 small sizes ...................................................... 192 speeds .......................................................... 19 steam consumption ................................... 133-136, 193, 266 step bearing ..................................................... 187 superheat corrections ......................................... 133, 193 tests on ................................................. 133-136, 266 vacuum corrections ................................. 132-134, 193, 278 valve gear, electric type ........................................... 226 hydraulic ................................................... 229 mechanically operated .................................... 226-229 wheels ...................................................... 184-186 Dake turbine ........................................................ 217 De Laval turbine ............................................ 5~ 6 > I 40-i5 2 analysis of losses ................................................. I 5 2 bearings ........................................................ J 47 blades ...................................................... "3. H 6 disks ........................................................... 142 gears ........................................................... J 4& governor ................................................ *5 220 > 2 45 nozzles .................................................... 3> 2 7> J 44 shaft ........................................................... J 4 speeds of ....................................................... J 44 steam consumption ............................................... J 5 2 superheat corrections ........................ . .................... I 5 tests on ......................................................... 2 7 vacuum corrections II wheels .......................................................... *4 2 Depreciation of power plants ...................................... 294-295 INDEX 371 PACK Design of power stations 297-306 Design of turbines, examples in 86-ioS entropy-heat, diagram applied to 87 of blades 81-83 of steam nozzles 2 9~57 Deterioration of turbines and auxiliaries 294-295 Diaphragms Sr, 182, 196 Disk rotation losses 117 stresses in 322-338 Drums, stresses in 315-320 Economy of engines, best 288 compared with turbines 285-288 with varying superheat, vacuum, and pressure 289 Economy of small engines and turbines 289-290 Economy of standard turbines 287 Efficiency of blades 85 steam engines, mechanical ^ 286 steam nozzles 49, 86 thermal unit basis of 270-271 turbines 85-86, 270 turbine-generators 362 Electric generators for steam turbines 358-362 Entropy 17 of saturated steam 18-23 of superheated steam 18, 54-57 of water 19-20 temperature diagrams J 7~ 2 5 total heat chart 34, 37, Appendix P>osion of blades 114 Experiments with nozzles 2 9~33> 4 2 ~53 on flow of superheated steam 5 2 ~53 with governors 245-246 Flash-over limit of turbine-generators 360 Floor area for engines and turbines 299-304 Flow of steam: experiments on 29, 52 Grashof's law for 2 9~32 Napier's formula for 2 9~3O saturated 2 9~33 superheated 5 2 ~53 weight of steam flowing 2 9~33 Fullagar's blading 161-163 Gas turbines 34~357 compared with gas engines 349 Gauging of Parsons blades 101 Generators, efficiency of 362 Glands, water packed 158, 209 37 2 INDEX PAGB Governing of turbines, methods of 218-246 by-pass ' 240-245 cutting out nozzles 221-229 experimental data 245-246 throttling 218-221 varying time of admission 229-240 Governors, turbine: Allis-Chalmers 234 Brown-Boveri 231-232, 241-243 Curtis 226-229, 2 43 De Laval 220-221 Westinghouse-Parsons 233-234 Wilkinson 238-240 Grashof 's law for flow of steam 20-32 Hamilton-Holzwarth turbine 204, 205 Heat diagram, entropy Appendix applied to design 87 Heat theory 10-28 Heat units. ...- 12 Heat, mechanical equivalent of 13 specific 12, 13, 54-56 Hero's turbine. 4, 8 Horse-power, internal or indicated 271 conversion table for kilowatts to brake 286 Impulse turbines 62-67 and reaction, combined (Westinghouse) 1 73-181 blade design 81-95 blade efficiency 69-77 blade losses 84-85 comparison with reaction turbines 108 distinguished from reaction.. .................................... 5862 shape of blades 62 Initial velocity of steam 16 Injection water (see Condensing water). Intermediate (stationary) blades 64, 182 Jets, impulse and reaction of 58-60 Kerr turbine 211-214 governor 213 nozzles 212 Kilowatts, conversion of, to brake horse-power 286 Knoblauch and Jakob's specific heat of superheated steam 54~j6 Labyrinth packing 156, 167, 204, 209 Lasche's method for rotation losses 121-1 24 Leakage of steam through blades 125 through diaphragms 85, 115, 125, 204 Losses in a turbine 85, 92, 115-125, 152 in a De Laval turbine, analysis 152 INDEX 373 PAGE Low-pressure turbines 247, 258 accumulators for 251-254 combined with gas engines 258, 291 steam consumption of 251, 254-257 Lubrication of turbines 306-3 1 1 cost of oil required 291 of Curtis turbines 311 Maintenance and operation 290-291 Marine turbines 249-250, 259-260 designing constants 97 working with reciprocating engines 249 Materials and permissible stresses in turbine disks 337 Mechanical equivalent of heat 13 losses in turbines 1 15-125 Mixed pressure turbines 258a Moisture correction for rotation losses n8 Motion, absolute and relative 68-69 Multicellular turbines 195-202 Napier's formula for flow of steam 29-30 Non-expanding nozzles 47) 50-51, 173, 196 Nozzles: area of 30-41, 53 design of. 2957, 94 efficiency 49-50 examples of 27 expansion ratio 36-41 expansion wedges in 217 flow of steam^through 2 9~33, 42-45, 52 length of 46-47 non-expanding 47 , 50-51, 173, 196 types of 3)2 7-28 losses in 49-50 Oil pump, Westinghouse rotary 309 Oiling systems 306-3 r i Curtis 3H Over- and under-expansion in nozzles 49-50 Packing glands 158 Parsons turbines 6-9, 153-172 bearings 8, 164 blades and lashing 158 design of 95-106 governors and valve gears 233-234, 245 history 6-9 manufacturers of 167 number of stages 97, 165 pressure diagram (indicator) 235 shroud rings 161 speeds of 97 374 INDEX Parsons turbines Continued PACK steam consumption of 133-136, 170-171, 266-269 superheat corrections 133, 172 vacuum corrections 133-134, 172, 278 Pelton types of turbines 207-2 1 7 Performance, thermal unit basis of 270-271 of engines and turbines 285-288 Piping of turbine stations 305-306 for superheated steam 311-312 Power plant economics 290-291 Pressure (initial steam) effect on economy 284 corrections for economy curves 126-136 in throat of nozzles 2 9~33 stages 65 Pressure-volume diagrams 1 5-1 7 Prices of turbines and auxiliaries 291-294 Quality of steam '. 34~39 Rankine cycle efficiency 28, 93, 272, 364, 368 Rateau accumulator. . 251-254 Rateau multicellular turbines 195-202 diaphragms 196 governor tests 245-246 low pressure 202 manufacturers 202 tests on 255-269 Reaction of jets 58-60 Reaction turbine 67 blade design 95-106 blade efficiency 77-80 blade losses 84-85 distinguished from impulse 62 Relative velocity 68-69 Riedler-Stumpf turbines 207 Rings, stresses in 3 14-316 Rosenhain's tests on flow of steam 47 Rotation losses 115-125 factors to correct for superheat and moisture 118 Rotors, stresses in 315-320 Screw turbine - o/ Shroud rings 161, 185 stresses in 314-316 Space occupied by engines and turbines 299-304 Sparking limit of generators 358 Specific heat 12 of superheated steam 13 Specific volume of superheated .steam 53> 99 Speed of turbines, effect on economy and output 138, 284-285 Stages of turbines, 6465 INDEX 375 PACK Stages of turbines Continued leakage between 85, 115, 125, 204 Steam turbine economics 273-312 Steam turbines: compared with steam engines.. i, 13-17 cost of 291-294 Step-bearing of Curtis turbines 187 Stodola's experiments with nozzles 48-50 Stresses in rotating rings, drums, and disks 3 X 3~338 at right angles 320-322 Sturtevant turbine 207-21 1 bearings 210 buckets 207 nozzles 208 velocity stages 210 wheels. 211 Superheated steam corrections for economy curves 126-136, 287 effect on economy 281-284 flow through nozzles 5 2 ~53 specific heat of 55~56 specific volume of 53, 99, 118 total heat of 271 Taper of De Laval nozzles 46 Temperature-entropy diagrams J 7~57 Temperature 10 n absolute 1 1 Terry turbine 214-216 Tests on turbines 132-139, 261-271 Thermal units, definition of 12 basis of performance 270-271 Thermodynamic efficiency 271-272 Thermometer correction n Thrust, end, in Parsons turbine 156 Torque line 138 Total heat of superheated steam 271 Types of steam turbines 58-62 Vacuum, condensing water required for 275, 280 corrections for economy curves 1 26-136, 278, 280 effect on engine and turbine economy 273-281, 288 most profitable 273 Vanes (see Blades). Velocity, absolute 68-69 diagrams 63, 66-70 relative 68-69 Velocity of steam: calculation of 24, 26 stages 64, 74, 1 73, 210 376 INDEX Volume, specific: PAGE at high vacuums (table) 277 of superheated steam 53, 99 Water-packed glands 157, 158, 209 Water-rate curves 262 Westinghouse oil pump 309 Westinghouse impulse and reaction turbines 1 73-181 adjustment bearing 175 advantages of 181 blades of impulse section 173 emergency speed limit 177-180 nozzles and nozzle block 177-178 reduction gear 26oa velocity stages 173 Westinghouse-Parsons turbines 165-166 auxiliary admission valve (overload) 234 bearings 164 blades and lashing 159 governor and valve gear 233-234, 245 indicator diagram 235 tests on 133-136, 171, 255, 267-269, 271 water-packed glands 157-158 Wilkinson turbines 202, 204 governor and valve gear 202 stage packing 204 Willans and Robinson turbine 161 Willans lines 99, 1 24 Zoelly turbines 205, 206 blades of 206 governor 245-246 nozzles 205 tests on ; . . . 270 wheels. . . 206 v AN IKITIAL FINE OF 25 CENTS - OVERDUE. 291520 UNIVERSITY OF CALIFORNIA LIBRARY VC333% THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL FINE OF 25 CENTS WILL BE ASSESSED FOR FAILURE TO RETURN THIS BOOK ON THE DATE DUE. THE PENALTY WILL INCREASE TO 5O CENTS ON THE FOURTH DAY AND TO $1.OO ON THE SEVENTH DAY OVERDUE. ^ LIBRARY BE '"- -^ 10m-7,'44(1064s)