The Una-Flow Steam-Engine By Prof. Dr.-Ing. h. c. J. Stumpf Technische Hochschule, Berlin. Translated by the Stumpf Una-Flow Engine Company, Inc., 401 S. A. & K. Building. Syracuse N.Y. Second Edition 1922 SI Copyright 1922 by Prof. J. Stumpf, Berlin THE BA TTLE OF THE ELEMENTS By J. A. STUM PP. (Chaos 1831, Nr. saw what he had made and found it good, wrote a man of noblest mind and mood. No longer with this doctrine is the world content, The doubter does in bitter words lament: One need but cast a fleeting glance at life, What sees one there? True happiness? No, strife, Death, need and misery far and wide, The elements in constant war abide, And storms of passion breeding endless hate Rob life of peace, till many curse at fate. We ask ourselves why we are so surrounded By raw materials and forces of all kinds, On what the longing in our breasts is founded, United with the curious impulse of our minds? As master of the earth, man shall create! All labor does the builder's hand await. To carry out His plan, God made him wise, That he might force and matter utilize. Such is the man whose work success has crowned, Who truth and light by his research has found, Who tested fire's flame and water's might, And thus their deepest secrets brought to light: Who through their elemental strife conceived, Instead of ruin, mankind's gain achieved. Nature's forces he has sought, And under his control has brought. The foes who storm with rage and hate, He keeps by thin walls separate. Around the boiler roars the flame, The seething waves within to tame, Who, in revenge, their enemy to reach, Strive through the prison's walls to force a breach. A polished rod ascends, by magic trained, Propelled by steam within a pipe contained. But lo! into the angry steam so bold, Now pours a rage-appeasing flood of cold; Down slides the rod, but in an instant back Pursued again by live steam in its track. The shining steel glides to and fro And, driving other parts, all show A striving to one goal. The great machine Obeys the master's mind, it may be seen. How many nature's wondrous course deride, And what they do not grasp, they claim unproved, The man of science does regard with pride How parts and whole in best accord are moved. 520223 ERRATA Page 70-71 124 125 125 206 234 234-235 303 Line or Fig. Fig. 1 Line 17 Heading Curve Last line Line 11 Fig. 13 Title Change from Max. Continuous L.H.P. Omit "& T. Hall" Omit "& Hall" Corliss Lead Fig. 14 German State Rys. Omit "Triple speed" To Max. Cont. I.H.P. Counterflow Load Fig. 13 Russian State Rys. TABLE OF CONVERSION FACTORS 1 mm. 1 cm. 1 m. 1 km. 1 sq. cm. 1 sq. m. 1 kg. I metric ton 1 metric ton-km. 1 m. kg. 1 kg. per sq. cm. 1 atmosphere 1 metric H.P. 1 C. Temp, in F. 1 calorie 1 cal. per kg. 1 cal. per I.H.P. hr. (metric) 1 kg. per I.H.P. hr. (metric) 0.039 in. 0.394 in. 3.28 ft. 1.609 miles 0.155 sq. in. 10.76 sq. ft. 2.205 Ibs. 2204.6 Ibs. 0.685 ton-mile 7.233 ft. Ibs. 14.22 Ibs. per sq. in. 0.9863 H.P. 1.8 F. 1.8 X temp, in C. + 32 3.968 B.T.U. 1.80 B.T.U. per Ib. 4.024 B.T.U. per I.H.P. hr. 2.236 Ibs. per I.H.P. hr. GENERAL INDEX Acceleration, valve, 88, 308 Admission, 14 to 16 A. E. Cr., 69 Area, inlet valve, 50 to 58 Area, exhaust port, 60, 69 Auniliary exhaust valves, 46, 47, 167, 177, 208, 225, 264 B Balancing, 72, 297 Bearings, proportions of, 73 Belt, exhaust, 12, 69 Blast, 112, 248 Bleeding, 187 automatic control for, 190 Bonnet, valve, Lentz, 87 Stumpf, 128 Cage, valve, 2, 80, 170 Cam oscillating, 89, 143, 144, 164, 167, 170, 172, 206, 246, 264, 270 reciprocating, 128, 160, 208, 217, 231, 257, 277, 279, 287, 291, 292 revolving tapered, 208, 218, 297, 300 revolving stepped, 271, 293, 294 Compounding, 2, 5, 190, 194 Compression, best length of, 17 to 34 Compressor, 218 combined steam and air cylinder, 213 Condensation, initial, 1 Condenser, 68 jet, 69 surface, 276, 294 Westinghouse-Leblanc, 69 Condensing engine, 68, 126 Connecting rod, 173, 264 Consumption, steam, lowest for various compressions, 17 to 26 lowest for various clearances, 27 to 29 Convection, losses due to, 105 Corliss una-flow engine, 182 Crank, 74, 75 center, 71, 132 Crank pin, proportions of, 73 Crosshead, 171, 294 pin, proportions of, 73 Cut-off, range of, 126 Cylinder, boring of, 93, 155 material of, 92 D Diagram, indicator, 166, 175, 203, 269, 286 Diffusor, 109, 110, 249 Doerfel, 172 Draft, smoke box, 112, 113, 248 Drop, pressure,. 51 Dual clearance, 177 E Eccentric gear, single, 150 Efficiency, mechanical, 71 Engines Ames Iron Works, 160, 161 automotive, 272 blowing, 213, 223 Borsig, A., 245 Burmeister & Wain, 132, 137, 278 compressor, 213, 218 Corliss una-flow, 182 Dehne, A. L. G., 218 Ehrhardt & Sehmer, 137, 139, 143, 195, 196, 202 Elsaessische Maschinenfabrik, 129 Erste Bruenner Maschinenfabrik, 129, 262 Filer & Stowell Co., 178, 181 Frerichs & Co., J., 273 Goerlitzer Maschinenbauanstalt, 144 Gutehoffnungshuette, 209 Harrisburg Fdy. & Mch. Works, 177, 178 hoisting, 207 Hungarian State Rys., 264 Kingsford Fdy. & Mch. Works, 292 Kolomna Engine Works, 225, 234, 237, 238, 262 Linke Hoffmann Works, 218 List, Gustav, 213 Locomotive, 225 Marine, 272 Maschinenbauanstalt Breslau, 232 Maschinenfabrik Augsburg-Nuern- berg (M. A. N.), 144 Maschinenfabrik Badenia, 252, 255 Maschinenfabrik Esslingen, 155 Mesta Machine Company, 206, 223 Musgrave & Sons, Ltd., 143 Neuruppin-Kremmen-Wittstock Ry., 240 Nordberg Mfg. Co., 172, 177 Northeastern Ry. of England, 240 Northern Ry. of France, 234 portable, 252 pumping, 213, 218 Robey & Co., 262 Rolling mill, 139, 195 Schmid, Karl, 294 Schweizerische Lokomotivfabrik, 232 Skinner Engine Co., 168 to 171 Soumy Machine Works, 159 stationary, 127 Stork & Co., 144 Sulzer Bros., 10, 76, 150, 151, 155 Vulkan Engine Works, 225, 227, 240, 274 Worthington Pump & Mchy. Corp., 218 Ejector effect, 110, 112 to 117, 245, 249, 250, 269 saving due to, 117 Expansion, cylinder, 128 Experiments, Prof. Naegel's, 118 to 124 Friction of driving parts, 77 H.P., 75-77 loss due to, 71 single eccentric, 150 valve, cam (locomotive), 230, 245 double-speed, 295, 303 Gooch, 207 Klug, 275, 278, 280, 290 link, disadvantage of, 16 Marshall, 237 Saeuberlich, 274 Skinner auxiliary exhaust, 171 Stumpf, 128, 129 Walsnhaert, 22, 230, 247, 285 Zvonirek, 143, 195 Governor, flywheel, 252 Gueldner, 110 I Inertia, curves of, 72 Insulation, 105 J Jacketing, 27 effect of, 2, 3, 6, 10, 11, 12, 33 Sulzer's tests on, 11 Jackets, proportions of, 12 Joints, piston ring, 94, 95 Lagging, 105 Lap, 128, 202, 207, 240 Lead, exhaust, 5, 61, 108, 239, 269 Leakage losses due to, 79 valve, effect of, 85 Lentz packing, 100 Liner, 160 Locomobile, 252 Locomotive, 225 Borsig, 245 Kolomna Engine Works, 225, 234, 237, 238 Maschinenbauanstalt Breslau, 232 Neuruppin-Kremmen-Wittstock Ry., 240 Northeastern Ry. of England, 240 Northern 'Ry. of France, 234 Schweizerische Lokomotivfabrik, 232 Vulkan Engine Works, 225, 227, 240 pistons for, 99 , three cylinder, 116, 240 Losses friction, 71 incomplete expansion, 108 leakage, '.^ radiation and convection, 105 surface, 1, 4 throttling, 49 volume, 13 Lubrication cylinder, 99, 155, 227, 248, 273 driving parts, 78 valve gear, 87, 128 M Machining of clearance surfaces, 2 of cylinders, 93, 155 of pistons, 93 Marine engines Burmeister & Wain, 278 Frerichs, & Co, J, 273 Kingsford Fdy. & Mch. Works, 292 Schmid, Karl, 294 Vulkan Engine Works, 274 N Nagel's experiments, 118 to 124 Nozzle, 109, 116, 249, 269 Parts driving, proportions of, 71, 73, 75, 76 reciprocating, 71, 72 Packing, piston rod, 100 to 104 Pin crank, proportions of, 73 crosshead, proportions of, 73 Pipe blast, 111, 112, 115, 248, 249 exhaust, 109, 110 steam, 105 Piston cast steel, 78, 227 expansion of, 93 floating, 77, 92 to 94, 99, 155 machining of, 93, 155, 227 overrunning of, 94 radial clearance of, 93, 155 self-supporting, 77, 92, 155 shoes, 77, 92, 155, 240 three-piece, 227 two-piece, 77, 240, 264 Portable engines Erste Bruenner Maschinenfabrik, 262 Hungarian State Rys., 264 Kolomna Engine Works, 262 Maschinenfabrik Badenia, 252, 255 Ports, exhaust, 5, 69 area of, 60, 113 Pressure, back, 29 to 33, 42 ' critical, 34 to 43 Pump, tube, 223 Pumping engines List, Gustav, 213 Worthington Pump & Mchy. Corp., 218 R Radiation, losses due to, 105 Rating, load, 71 Relief, compression, for starting, 203, 207, 229, 230, 240, 247, 275 Resilient valve, calculation of, 81 Rings piston, 94 hammered, 96 overrunning of, 97, 98 proportions of, 96 wear of, 98 Rod, connecting, 72, 173, 263 tail, 77, 93, 206 Rolling mill engines Ehrhardt & Sehmer, 195, 196, 202 Mesta Machine Co., 206 Schlick, 297 Seats, valve, 80, 85, 86 Separators, oil, 69 Series arrangement of parts, 85 Shaft, crank, proportions of, 73 Shoes, piston, 77, 92, 155, 240 Slap, piston, 94 Speed, piston, 71 Stack, smoke, 112 Stationary engines Ames Iron Works, 160, 161 Burmeister & Wain, 132, 137 . Ehrhardt & Sehmer, 137, 143 Elsassische Mascheninfabrik, 129 Erste Brunner Maschinenfabrik, 129 Filer & Stowell Co., 178, 181 Gorlitzer Maschinenbauanstalt, 144 Harrisburg Fdy. & Mch. Works, 177, 178 Maschinenfabrik Augsburg- Nurn berg, 144 Maschinenfabrik Esslingen, 155 Musgrave & Sons, Ltd., 143 Nordberg Mfg. Co., 172, 177 Skinner Engine Co., 168 to 171 Soumy Machine Works, 159 Stork & Co., 144 Sulzer Bros., 10, 76, 150, 151, 155 Stepped cams, 271, 293, 294 Stodola, 110 Strahl, 112 Superheater, 258 Superheating, effect of, 2, 3, 5, 8, 273 Surface clearance, 1 to 3, 7 Surface clearance, minimum, 1 T Temperature steam, experiments on, 118 to 124 high, 8 Tests, engine, 129, 137, 144, 157, 161, 234, 236, 265 Throttling, losses due to, 49 Tube pump, 203 V Vacuum, high, 68 Valves clearance, 45, 46 Corliss, 79, 183 to 186 double beat, leakage, of, 79 locomotive, 230 resilient, calculation of, 81 tightness of, 79, 80 two-piece, 168 exhaust, auxiliary, 46, 47, 166, 167, 177, 208, 225, 264 automatic, 166, 264 inlet, area of, 50 to 58 leakage of, 85 piston, 79, 177, 178, 196, 210, 239, 240, 275 single beat, 86, 245, 247, 262, 272, 295, 303 slide, 79 locomotive, 237 Valve spring, calculation of, 88, 308 Volume clearance, additional, 44 to 46, 126 clearance, % of, 48, 126, 127, 228 w Walschaert gear, 22, 230, 247, 28.5 Westinghouse-Leblanc, 69 z Zeuner, 111 Zvonicek gear, 143, 195 Preface. The second edition of this book represents a complete revision of the first one, little of which remains. The first edition contained a good many opinions in addition to facts and was intended rather for the purpose of defending the una- flow engine against antiquated theories and attacks. In the meantime the una- Fie. 1. flow principle has been widely tried out and scientifically investigated. It has become an accomplished fact and is in common use. This book therefore contains scientific proof from an objective point of view, as well as a description of the development of the una-flow engine. In the opening chapters of the book the different losses of the steam engine are investigated. The causes and effects are defined, as well as the relations be- tween them, and the manner is pointed out in which the minimum value of each loss is obtainable. After considering all the seven different losses occurring in a steam engine, the question is asked as to how a steam engine must be designed in order to have a minimum total of all the seven losses. In answer to this query two different designs are presented, one being a stationary una-flow engine with single-beat valves for condensing operation, and the other a una-flow locomotive also equipped with single-beat valves. On the former, the use of single-beat valves was made possible by the use of a double speed valve gear (see Chapter VI). Both types of engines were developed during the year 1920. Lower steam consumption figures than those given by the best multi-stage engines are Fig. 2. obtainable with these types for both saturated and superheated steam. The expe- rience gained with una-flow engines in widely different fields and under the most varying conditions was utilized in the design of these engines to the fullest pos- sible extent. A number of Chapters is devoted to the description of this development in all its phases. The novelty in this case is the single-beat valve, which has so far been used only in internal combustion engine practice. All previous attempts to apply it to steam engines have miscarried. The application of this type of valve to the una- flow engine represents figuratively the keystone in the development of the latter. The fundamental conformity between the new una-flow engine and the two-stroke internal combustion engine is surprising. This refers to the uni-directinal flow, the single stage expansion, the piston-controlled exhaust and the single-beat inlet valve. ?! Surprising also is the close agreement in the essential parts of the cylinder between the latest design shown in Fig. 3 (see Chapter VI) and the first original sketch of a una-flow engine made in the year 1900 and reproduced in Fig. 2, which also incorporates single-beat valves. Since, as Descartes says, doubt may be considered the origin of every philo- sophy, the question regarding the doubt which originated the una-flow philosophy may well be asked. This doubt arose in the year 1896 during the starting up of two pumping engines designed by myself for the Pope Mfg. Co., of Hartford, Conn. (Fig. 1). These were vertical triple expansion engines with Corliss valves and a central condensing system, in which everything then considered good practice was carried to the extreme limit. This resulted in a very complicated construction which appeared to me to be a sign of weakness. The doubts which then arose in my mind eventually led to the sketch shown in Fig. 2 during the year 1900. The construction of steam turbines of several stages, which began at that time, Fig. 3. was developed along the lines of pure uni-directional flow, arid this brought up the question whether it would not be possible to raise the reciprocating steam engine to the same thermal plane as the turbine by the use of the una-flow prin- ciple. The application of the una-flow action of the turbine to the steam engine, although in a somewhat imperfect manner, by properly designing the cylinder, valve gear, steam jackets and condenser connection, etc., finally led to the una- flow engine with single-beat valves as shown in Fig. 3. The object in view was the attainment of the minimum total of all the seven different losses of the steam engine, as well as the utmost simplicity and reliability of operation. This goal now seems to have been fully reached. The fact that the una-flow engine pos- sesses the uni-directional flow in common with the steam turbine and has a con- structional basis similar to that of the two-stroke internal combustion engine, may be cited in support of this. The design and adaptation of the una-flow engine to different requirements and conditions of service represents an immense amount of work, in which I received the full support of my assistents as well as that of Mr. Rosier of Miihlhausen, Alsace, Mr. Arendt of Saarbriicken, Prof. Bonin of Aachen, Mr. Dutta of London, and Drs. Mrongovius and Meineke of Berlin. To the splendid support of the last four gentlemen may be attributed the positive developments of the chapters on volume loss, throttling loss, exhaust ejector action, the una-flow locomotive, and the valve gear with double speed lay shaft. I am particularly indebted to those gentlemen who took up my proposals at a time when no one would yet believe in the una-flow engine, namely, Prof. Nol- tein of the Technical Hochschule at Riga, Messrs. Hnevkovsky and Smetana of Brtinn, Mr. Lamey of Miihlhausen, Alsace, Mr. Mtiller of Berlin, and Mr. Schiller of Grevenbroich. Berlin, January 1921. J. Stumpf. Index. Page Preface V I. Steam Engine Losses 1 1 a. Losses due to Cylinder Condensation 1 1 b. The Una-Flow Arrangement as a Means for reducing Surface Losses. Jacketing of the Cylinder 4 2 a. The Influence of the Clearance Volume upon the theoretical Steam Con- sumption (Volume Loss). Critical back Pressure 13 2b. Additional Clearance Space 44 3 a. Losses due to Throttling. Determination of Inlet and Exhaust Port Areas 49 3b. Relation between the Una-Flow Engine and the Condenser * . . . . 68 4. Lossesdueto Friction (mechanical Efficiency). Dimensioning of drivingParts 71 5. Losses due to Leakage. Valves, Pistons, Piston Rod Packings .... 79 6. Losses due to Radiation and Convection 105 7. Losses due to incomplete Expansion. The Exhaust Ejector Effect . . 108 8. Prof. Nagel's Experiments 118 II. 1. The Una-Flow Stationary Engine 126 2. The Una-Flow Corliss Engine 182 3. . The Una-Flow Engine arranged for Bleeding 187 4. The Una-Flow Rolling Mill Engine 195 5. The Un^-Flow Hoisting Engine 207 6. Una-Flow Engines for driving Air Compressors, Pumps, etc 213 III. The Una-Flow Locomotive 226 IV. The Una-Flow Locomobile and Portable Engine 252 V. The Una-Flow Marine Engine 273 VI. The Una-Flow Engine with single-beat Valves and double-speed Lay Shaft 303 Summary 316 I. Steam Engine Losses. The losses in a steam engine may b.e classified as follows: 1. Losses due to cylinder condensation (surface loss). 2. Losses due to the volume of the clearance space (clearance volume loss). 3. Loss due to throttling or wire drawing. 4. Friction loss. 5. Loss due to leakage. 6. Loss due to heat radiation and convection. 7. Loss due to incomplete expansion. > la. Losses due to Cylinder Condensation (Surface Loss). The amount of initial (or cylinder) condensation (termed surface loss in the following) is determined by the size, kind and arrangement of the harmful sur- faces, by the steam jacket, by the quality of steam passing these surfaces, by the temperature gradient and the number of stages used, by the amount and period of the steam flow and the path of the steam through the cylinder (counterflow or una-flow). Initial condensation, which is usually over by the end of admission, is caused by the clearance surfaces and increased by any moisture carried over with the steam. The ability of these surfaces to receive and give off heat forms a kind of heat bypass, with a corresponding loss to the cycle. Part of the steam condenses during admission and re-evaporates during exhaust and the last part of expansion. The harmful surfaces comprise the inner surfaces of the cylinder and* the inwardly exposed surfaces of piston, piston rod and steam distributing parts. Surfaces which are continually exposed even in the dead center position of the piston may be termed harmful surfaces of the first order, and those which are progressively uncovered by the piston during its motion, harmful surfaces of the second order. The former usually cause the essentially greater part of the surface loss. Since the amount of surface loss is determined by the extent of the harmful surfaces, the latter should be kept as small as possible and should also be machined. A good many designers pay attention only to the amount of clearance volume without considering its surface. The minimum harmful surface of the first order comprises an area equal to twice the cylinder cross-section (cylinder head and piston), and it is convenient to express the additional surface of the first order in percent of this minimum surface. In actual engines these additional surfaces, which are mostly not machined, are found to be from 150 to 200% of this minimum harmful Slump/, The una-Flovv steam engine. surface, although it is possible by careful design to reduce tiiis figure to 3 or 5%. Piston valves with snap rings working in separate bushings, as well as slide valves with long curved ports, which latter are in most cases left rough and serve for both steam admission and exhaust, have large surfaces which are especially harmful on account of their very nature and arrangement. Engines with separate admission and exhaust ports are far better in this respect, because the latter are usually very short and the hot admission and cold exhaust steam enter and leave the cylinder through separate passages, thereby avoiding the alternate heating and cooling of these surfaces and the corresponding surface loss which takes place in engines having common inlet and exhaust ports. Slide and piston valve engines with their perpetual reversal of flow are sub- ject to extensive turbulence and heat exchanges, although careful design can generally improve conditions. The Corliss engine may be considered an improve- ment by reason of .the smallness and different arrangement of its additional sur- faces; and, with the valves in the heads, the ports are straight and short with inlet and exhaust separated. Conditions can be improved still further if the exhaust valve, which forms the major part of the additional surfaces, is made to fill its bore completely and has a straight port through its center only. Poppet valve engines in general have rather large additional surfaces, espe- cially where valve cages are used ; and notwithstanding separate inlet and exhaust ports the frequent flow reversal has a deleterious effect. Valve cages considerably increase the additional surfaces. Machining of the latter may be provided for in many cases by clever design, thereby reducing their extent and the corre- sponding turbulence and surface loss. Further means of reducing the surface loss are:' 1. Jacketing, 2. Compounding, 3. Superheating, 4. The una-flow system. Jacketing of the harmful surfaces is a further step in reducing surface losses. The heating medium is usually steam, seldom flue gases. Engines with great sur- face losses will be largely benefited by jacketing, and single cylinder condensing engines working with saturated steam will show the greatest gain since they offer the largest scope for improvement. The effect of the jacket is diminished if the expansion takes place in two, three, or four stages, and if the steam is superheated in addition. These means improve the thermal condition to such an extent that there is little to be gained from jacketing. This applies to superheating, espe- cially if the whole working cycle takes place in the range of superheat. Locomo- tives, in which the steam, when leaving the cylinder is still superheated, will derive no benefit from jacketing. Superheating is such a far reaching remedy that the number of stages in counterflow engines working with superheated steam has been generally reduced from three to two for condensing operation, and from two to one when operating non-condensing. Increased speed, later cut-offs, and larger units tend to reduce the surface losses and hence the effects of jacketing. High speed damps the temperature fluctuations of the walls; the late cut-off raises the mean wall temperature; and the larger size gives a more favorable ratio of volume to surface. For these reasons a large, fast running, and heavily loaded engine will show the least gain from jacketing, especially if working with superheated steam or a small temperature gradient. All this applies to superheated steam locomotives as an example. Jacketing has the greatest effect in low pressure cylinders, since surface losses, temperature gradient, and jacket surfaces are large and the weight ratio of jacket to working steam is the most favorable. The gain from jacketing is accordingly smaller in intermediate and high pressure cylinders, and in many cases there is hardly any in the latter. Similar conditions prevail in two cylinder compound engines. Head jacketing is usually more effective than cylinder jacketing because the cylinder surfaces are temporarily covered by the piston, and the oil film acts as a heat insulator. These surfaces of the second order cause small sur- face losses and consequently show a smaller gain from jacketing. Saturated steam is very bad in this respect because the water particles act as heat conductors and increase the surface losses. Dry steam is better, and best of all is superheated steam. Saturated steam is an excellent, and superheated steam a very poor heat conductor. The action of the cylinder becomes the more adiabatic the more the superheated region extends through the cycle. Superheat, furthermore, by increasing the specific volume, makes the steam lighter and reduces both the weight per cycle and the surface loss. The commonly prevailing amount of superheat allows non-condensing engines to work in the superheated region throughout the cycle; but this is not the case with condensing engines, assuming proper ratios of expansion in both cases. In condensing engines the low pressure part of the cycle always extends beyond the saturation point. Superheating is of such far reaching effect that the reduction of harmful surfaces, their arrangement, and in many cases even jacketing, lose their impor- tance. Generally speaking, among the different ways of reducing surface losses, one or the other may be so effective that there is nothing left for the remaining ones. The amount, extent and kind of steam flow, especially of the exhaust, may have considerable influence in engines working with saturated steam. Wet exhaust steam flowing with high velocity through long unfinished ports having large sur- faces may cause great surface losses. Summing up, it may be stated that the application of the different means for reducing initial condensation resulted in the common use of the two-cylinder compound engine for condensing service, for the reason that the low pressure part of the cycle takes place in the saturated region. The una-flow principle, however, permits the use of the single cylinder single-stage expansion engine for this ser- vice, since the uni-directional flow of steam eliminates initial condensation despite the fact that a part of the cycle takes place in the saturated region. The una- flow principle is also of great advantage for non-condensing and multi-stage engines. 1* Ib. The Una-Flow Arrangement as a Means for Reducing Surface Losses. The una-flow engine, as its name indicates, utilizes the steam energy by a um-directional flow, i. e. the steam passes through the cylinder always in the same direction. As shown in Fig. 1, the steam enters the cylinder head from below, ; heats the surface of the latter, and then enters the cylinder through the inlet valves located in the top portion of the head. Doing useful work, the steam follows the piston and after having expanded, leaves through ports at the opposite end of the stroke, i. e. in the middle of the cylinder; the opening and closing of these ports being accomplished by the piston during its motion. This is in marked contrast to the ordinary or counterflow engine, where the steam enters at the end of the cylinder, follows the piston during the working stroke, and, returning with the piston, leaves at the cylinder end. The result of this kind of flow is an intensive cooling action upon the clearance surface, the exhaust steam being usually wet and thus an excellent heat conductor. The consequence is increased initial or cylinder condensation during the following admission. The una-flow principle avoids the cooling of the clearance surfaces, thereby eliminating initial condensation to such an extent that compounding becomes superfluous. Una- flow engines may, therefore, be built with a single cylinder and single-stage expan- sion, and yet show the economy of compound or triple-expansion engines. The exhaust ports of the una-flow cylinder have an area about three times as large as can be realized with slide or poppet valves, with a consequent complete pressure equalization between cylinder and condenser if long and restricted pas- sages between them are avoided. In other words, if the condenser is placed close to the cylinder and the connection is of ample area, then complete equalization of pressures is assured. In order to get a clear conception of the magnitude of this port area one has to consider that the engine piston acts as a piston valve and the crank as eccentric, while the cylinder constitutes the valve bushing. The exhaust lead is usually taken at 10%, which fixes the compression at 90%. The una-flow exhaust ports do away with separate exhaust valves and their leakage loss, their additional clearance volume and surface, as well as the neces- sary valve gear. The elimination of exhaust valves is therefore an additional advantage of the una-flow construction. Indicator cards of una-flow engines show adiabatic lines for expansion and compression. Adiabatic expansion results in considerable moisture even with highly super- heated steam. The entropy chart shows at a glance that with an initial pressure of 12 at. and a temperature of 300 C an expansion to 0.8 at. abs. produces 7% moisture. During the exhaust period this expansion continues until at a terminal 'pressure of 0.1 at. abs. the moisture amounts to 17%. On account of the un- dvoidable heat losses, the temperature at the end of admission will be somewhat less than the above, with a consequent increase in the final moisture. The extension of the working cycle into the wet region rendered compound engines necessary, the high pressure cylinder working with superheated, and the low pressure cylinder with saturated steam. The una-flow system made a return to the single stage engine possible for both superheated and saturated steam. Superheated steam is a very effective means of combating initial condensa- tion. The combined use of superheat and the una-flow construction will still better conserve the heat during admission. Expansion will therefore start at a higher temperature and terminate with less moisture; or in other words, better economy will result. The head jackets have the effect of a partial regeneration of the steam during expansion and exhaust. On account of the large difference of temperature, the large heating surface, and the great difference in the specific weights of the live steam and the exhaust or compression steam, an intensive heating action takes place during expansion, exhaust, and compression. This affects particularly that part of the steam located close to the cylinder head, while in consequence of the adiabatic expansion that part following the piston will sustain both a drop in tem- perature and an increase in moisture. The greater part of the moisture produced will accordingly be found close to the piston head with a progressive dryness and 6 an increase in temperature towards the cylinder head. This moist steam, being close to the exhaust ports, will escape first when they are uncovered, while that part which received heat from the head jacket is trapped by the returning piston and compressed along adiabatic lines, partly as saturated, and partly as super- heated steam, but mostly the latter, because during the early part of compression heat is still being transmitted to it from the head jacket. This elimination of liquid condensate avoids its deleterious effect of heat exchange as well as damage due to water hammer. Tests conducted on triple-expansion engines in regard to the action of steam jackets have shown that there is no gain in high pressure cylinders, very little in intermediate cylinders, but a large gain in low pressure cylinders, despite the large heat losses prevailing in cylinders of the usual type. Owing to the counter- flow action, a great part of the jacket heat is necessarily carried by the exhaust steam into the condenser. One has only to consider that at the time of exhaust valve opening a considerable amount of pressure energy is available to exhaust the steam with velocities as high as 350 to 400 m. per sec., that the steam with this velocity impinges on the clearance surfaces, depositing water, and that on account of the sudden drop in pressure the heat absorbed by them during admis- sion starts an intensive re-evaporation which extracts considerable quantities of heat from these surfaces. The fact that the latter are in many cases jacketed presents a most unfavorable picture of the very inefficient way in which admission and jacket heat are utilized. From the point of exhaust valve opening up to the beginning of compression the jacket heat is carried into the condenser in the most wasteful fashion. During the remaining part of the cycle, heat exchange occurs under more unfavorable conditions and at lower velocities, but notwithstanding the immense losses of jacket heat the low pressure cylinder derives the greatest benefit from steam jackets. This may be explained by the fact that the low pres- sure cylinder has the greatest temperature differences, the greatest heating sur- faces, the greatest surface losses, and a very favorable density ratio of jacket to working steam. It follows that una-flow cylinders necessarily have a particularly energetic heating action, since, as in the case of low pressure counterflow cylinders, heating takes place under the influence of the full temperature difference, the large surfaces, and the great difference in density of jacket and working steam. The counterflow of steam, with its great losses, is replaced by the uni-directional flow where no jacket heat is lost to the exhaust. As shown in Fig. 2, exhaust steam in a una-flow cylinder never passes heated surfaces. The layer next to the cylinder head, at the very worst, approaches the exhaust ports without leavipg the cylinder, and therefore hardly any jacket heat can be lost. The beneficial effect of steam jackets proved to e-xist in low pressure counterflow cylinders must therefore be in evidence in a high degree in una-flow cylinders, since a loss of jacket heat is avoided. It is assumed, of course, that jacketing is limited to the cylinder head and that the cylinder is unjacketed (Fig. 1 and 2). The head jacket extends to the point where cut-off normally occurs so that the clearance surfaces of the first order are effectively heated from the outside, while the highly superheated com- pression steam prevents cooling from the inside. A further reduction of surface losses can be accomplished by enlarging the jacket surfaces and by the utmost reduction and machining of clearance surfaces. Even though the clearance surfaces in a una-flow engine are exposed during the whole cycle, they are not subject to the rush of exhaust steam passing them, nor to re-evaporation; and they therefore suffer little cooling on account of the com- parative tranquility of the steam molecules adjacent thereto, and furthermore have the benefit of both the jacket and compression heat. All these causes combined with the una-flow principle and una-flow construction produce an almost adia- batic action of the clearance surfaces. Fig. 2. The una-flow engine fundamentally avoids the thermal mixup characterizing the counterflow engine. The cylinder is composed of two single acting cylin- ders with the exhaust ends in common. On account of the long piston the stroke volumes of both cylinder ends are relatively displaced for a distance equal to the length of the piston. The two inlet ends are hot and remain hot, while the common exhaust is cold and remains cold, and the temperature changes gradually from the hot inlet to the cold exhaust. The jacketing, comprising the heating cham- bers at the inlet ends and the cold exhaust belt around the common exhaust, is in perfect accord with this. In a counterflow cylinder the two stroke volumes overlap. The exhaust end of one side more or less reaches to the admission end of the other, according to the length of the piston. From a thermal point of view the arrangement is obscure. The central exhaust port and belt in the middle of the cylinder effectively cool that part of it at which the piston attains its highest velocity. This favorable 8 action is further augmented by the omission of jackets on the adjacent part of the cylinder. Furthermore, the piston on account of its large area exerts a very low unit pressure. The cylinder is of very simple form and can be kept free from badly distributed material, thus avoiding local heating and warping. The large bearing surface of the piston, the cooling action of the exhaust belt, as well as the simplicity of the cylinder, which precludes the possibility of warping even at the highest steam temperatures, render the use of a tail rod unnecessary, provided the material used is suitable, the design correct, and proper lubrication supplied. (See una-flow locomotives and engines built by Sulzer Bros.) The piston has two sets of rings, comprising four or six altogether. During the period of high pressures both sets are in action, and the pressure has dropped to about 3 at. when one set has overrun the exhaust ports. The many una-flow engines in operation are proof that the piston is not the cause of any difficulties even with the highest degrees of superheat, and that with good workmanship a piston can be made perfectly tight. If, however, a cylinder should become scored, its simplicity allows it to be easily replaced without great expense. There can be no doubt about the possibility of employing, in una-flow engines of the kind described, steam temperatures far in excess of anything used at present. Even with the highest initial temperatures the cycle extends far into the moist region, thus insuring moderate working temperatures for cylinder and piston, the highly superheated steam being limited to the cylinder head. The una-flow engine, therefore, opens up further possibilities of development in the utilization of higher' degrees of superheat. It is all the more suitable for this because the superheat benefits the whole cycle, while in counterflow compound engines the high pressure cylinder receives too much and the low pressure cylinder too little superheat. This feature of the una-flow engine does not, however, contradict the fact that it is also suitable for saturated steam, excellent results being actually obtained with both kinds. The una-flow engine has the uni-directional flow, the hot inlet and the cold exhaust in common with the steam turbine. From a thermal point of view it forms the missing link between the reciprocating steam engine and the steam turbine. The opinion is frequently advanced that, in regard to their thermal action, the cylinder head and piston surfaces of a una-flow engine are merely interchanged. This leaves out of consideration the fact that the piston surface is protected against the action of the exhaust by a cone of stagnating steam. The existence of this phenomenon has been frequently proved beyond dispute in the case of air and water. The surface corresponding to the face of the slide valve of a counterflow engine is located in the una-flow engine at the circumference of the cylinder. The port necessary with a slide valve, together with the clearance and clearance sur- faces involved, are completely avoided in the una-flow engine. The surfaces of the una-flow exhaust ports are outside of the cylinder and therefore have no bearing upon the thermal conditions. Piston and steam have a different velocity only after the former has opened the exhaust ports, when the available pressure energy is transformed into kinetic energy. The steam, however, attains its full velocity only after it has passed the edge of the piston and therefore the cooling action of this steam upon the piston can only be small. The cooling action caused 9 by the low velocity of approach in the cylinder now remains to be examined. Considering the exhaust ports as nozzles bounded on one side by the piston edge, the steam at this narrowest point attains a mean critical velocity of about 410 m/sec. On examining cross-sections ahead of this smallest section or throat, very moderate velocities are found. Fig. 3 makes these conditions clear for a cylinder of 600 mm bore and 800 mm stroke, the piston being assumed to have overrun the ports for a distance of 20 mm. Calculation of the velocity at this narrowest point shows it to be 410 m/sec., while at a distance of 15 mm ahead of this point it is only 71 m/sec.; at 40 m it is 36 m/sec.; at 85 m it is 20 m/sec., and at 130 m it is 15 m/sec. It should be noted that in considering the various cross-sections a reduction of area due to the bridges has been assumed at the narrowest section (throat), and this does not apply to the cross-sections of approach. As the piston progressively uncovers the exhaust ports the proportions of the cross-sections are changed in such a way that the velocity of approach is increased. At the same time a reduction of pressure occurs, the effect of which is to reduce the velocity of approach, beginning at the point at which the critical pressure is reached. Most indicator cards of una-flow engines show clearly that when the piston reaches the dead center the pressure inside the cylinder has dropped to the back pressure; or in other words, the greater part of the steam has in this position been exhausted. Therefore, owing to the short duration and low intensity of the flow of exhaust steam along the piston surface and the small harmful exhaust surfaces, the resulting cooling action is small. The whole cylinder section acts as an approach to the exhaust nozzles. A further protection is given by the layer of stagnating steam, and there is finally a very intense heating action during compression and admission. The 10 heating of the piston surface by the hot live steam is so effective that this surface acts almost adiabatically during the following exhaust period. The very favorable steam consumption figures obtained with this type of engine are further proof that a simple interchange of the cylinder head and piston surfaces, in regard to their thermal behaviour, is out of the question. Such favorable economy can only result if the piston and its cooling action is negligible. This is further confirmed by tests of Prof. Nagel. A test with saturated steam of 184 C, and a cut-off at 12%, showed that the temperature of the piston surface at a point near its circumference was about 164.5 G. The total fluctuation at this point was only 1.3 G. This surprisingly high temperature, and moreover the small fluctuation, justifies a very favorable conclusion. These figures should be still better towards the center of the piston surface. The thermal, constructional, and operative advantages of this type of prime mover are such that in continuous operation the economies of compound and triple-expansion engines can be obtained with both saturated and superheated steam. Jacketing of the Cylinder. The firm of Sulzer Brothers, of Winterthur, Switzerland, constructed an ex- perimental engine of the una-flow type after such engines had been put on the market by the Erste B runner Maschinenfabrik Gesellschaft, who were the first to take up the manufacture of una-flow engines on the author's recommendation. Fig. 5. Sulzer Bros, entrusted the author with the design of the first una-flow cylinder of 600 mm bore, 800 mm stroke, and 155 r. p. m. All later Sulzer engines are built with only slight changes from this design, which is shown in Figs. 4 and 5. According to Sulzer Brothers' usual practice the engine was put on the testing floor and a great number of tests were carried out with the object of observing 11 its performance and determining its economy under the most varying conditions. An important part in the program was the study of the effect of the jackets. For this reason the author incorporated in this design not only head jackets, through which the live steam had to pass before entering the cylinder, but also jackets at the ends of the cylinder barrel, which were separated by a neutral zone from the exhaust belt. These cylinder jackets could be shut off separately. During the tests the head jackets were necessarily always in operation, but the cylinder jackets were either in service or shut off, as indicated in Fig. 6 by the words ,,with jacket" and ,, without jacket". The tests proved that the effect of the cylinder jacket decreases with increasing steam temperature. With saturated steam the difference was almost 1 kg/I HP-hour in favor of cylinder jackets, while it was barely 0.5 with steam of 265 C and only 0.2 kg with steam of 325 C. All figures refer- red to the most economical M. E. P. The steam pressure was 9.2 at. gage and the vacuum 66 cm. For the point of best operating economy, i. e., an M.E.P. of about 3 kg/sqcm, these differences change in such a way that a small increase results when running with cylinder jackets and with a steam tempe- rature of 325 G, while steam of 265 G shows a small decrease in steam consumption. For a steam temperature of kg/cm* 325 C the point of equality of steam consumption, when operating both with and without cylinder jackets, is found to correspond to an M.E.P. of 2.5 kg/sqcm. The corresponding point for steam of 265 G occurs at an M.E.P. of 3.4 kg/sqcm. For saturated steam this point moves towards a still higher M.E.P. This explains the customary omission of cylinder jackets for superheated steam. It is also noteworthy that the best economy with steam of 325 G very clo- sely approaches the value of 4 kg/I HP-hour. The results for saturated steam, especially with cylinder jackets, are extremely favorable. It must be considered, however, that the saturated steam had a very slight degree of superheat in order to make sure that it was actually dry. It should further be noted that this engine was well designed and well built. The clearance volume and clearance surface (the latter being machined) were moderate, and the whole engine was built with the high precision usual to Sulzer Brothers' shop practice. The measurements were made by means of a surface* condenser, which, as is well known, gives slightly lower but more accurate results than boiler feed water measurements. Comparing these curves with those of compound or triple-expansion engines, it will be found that the steam consumption of the una-flow engine is influenced 12 Fig. 7. by the load to a much smaller degree. This is shown particularly by the curves marked ,, without jackets'"', where there is hardly any change in the steam con- sumption between mean effective pressures of 1 and 3 kg/sqcm, especially with high superheat. Even with saturated steam little change is noticeable between mean effective pressures of 1 and 2,4 kg/sqcm. The curves of Fig. 6 justify the following conclusions in regard to cylinder jacketing: For highly superheated steam (300 C and more) and high mean effective pressures cylinder jacketing is useless (Fig. 7), and has a deleterious effect even at as low an M.E.P. as 3 kg/sqcm. For low mean effective pressures cylinder jackets may yet be expected to yield a small gain. For moderate steam temperatures of about 250 G a short cylinder jacket as shown in Fig. 8 is advisable. It will slightly im- prove the economy even for a high M.E.P. and produce considerable gain for a low M.E.P. With saturated steam or low de- grees of superheat a cylinder jacket separa- ted only by a narrow zone from the exhaust belt (Fig. 9) should be used under all circum- stances. Head jackets are essential in all cases. The separation of cylinder jacket and exhaust belt is advisable in order to avoid unnecessary loss of jacket heat and to pro- vide more favorable operating conditions for the piston, especially when using super- heated steam. The center part of the cy- linder, where the piston attains its highest speed, has the lowest temperature. Ex- cellent results can be secured with self- supporting pistons, even with very high temperatures of superheat, if the designer pays attention to these thermal conditions by providing the piston with bearing sur- faces at its center only, leaving its extremities to project in the manner of a plunger towards both ends of the cylinder without actually touching the walls. The head jackets do not impair the operation of the piston, since no rubbing surfaces are in contact with the former. Their heating action is very effective because they are continuously in contact with the working steam ; they heat harmful surfaces of the first order, and with proper lubrication the transmission of heat from them is not impeded by an oil film (Fig. 7). There is practically no loss of jacket heat to the exhaust. Conditions are not so good in Fig. 8 and still worse in Fig. 9. In these constructions the amount of jacket heat lost to the exhaust increases more and more because the exhaust steam to a certain extent flows past heated surfaces, although these may be partly protected by an oil film. Fig. 8. 13 2 a. Influence of the Clearance Volume upon the theoretical Steam Consumption (Volume Loss). (The Una- Flow System as a Means for Reducing the Volume Loss). A certain amount of steam admitted per stroke into a cylinder with clearance will produce an indicator card of less area than in an ideal cylinder without clea- rance. The difference in area may be termed volume loss. This volume loss is represented in Figs. 1 to 4 for different conditions. The diagrams corresponding to the cylinder with clearance are drawn in heavy lines, while those for the ideal cylinder are shown dashed. Volume losses can be expressed absolutely or relatively. In most cases it is convenient to figure the volume loss in per cent of the engine output. u-^ H Fig. 1. A comparison of the two areas AO PG and ABPG in Fig. 1 shows area BOP to be a loss. The parts of the diagrams lying below the line G P show a loss of the area GES and a gain of area PCV = GFQ T. By subtracting the latter, the resultant loss is represented by the area TQFES. In Fig. 2 the admission has been lengthened until the points F and .P of Fig. 1 coincide with the beginning and ending of the diagram. Consequently the shaded areas BOP and GES represent the loss for the diagram with clearance. In Fig. 3, with still longer admission, a comparison of the diagrams shows the loss to be equal to the shaded areas BOVC and GHS. 14 The reduction in diagram area produces a corresponding increase of the loss due to incomplete expansion. A special case is shown in Fig. 4, which represents a diagram without volume loss. This may occur for instance in high pressure cylinders of compound engines when expansion is carried to the back pressure and compression to the initial pressure. Although the diagram has no direct volume loss, the stroke volume of the cy- linder with clearance must be increased from F! to F 2 , with a consequent increase in the surface loss. \d O \r Fig. 3. Fig. 4. The diagrams clearly disclose the fact that the compression is a means of reducing the volume losses, since they would be essentially larger without com- pression. It is also clear that for a constant admitted steam volume cp the point of cut-off will vary with different lengths of compressions. There will therefore be a certain position of 9? for which the diagram area produced will be a maximum and the corresponding volume loss a minimum. \ X^ Fig. 5. a) Determination of the best position \ \ of 9?, if pv = constant, and /? 1? /? 2 , cp and ^ S Q are known (see Fig. 5). Diagram areae F = ABEDRF = = h + h-h-h=ABHG + BEKH FRJG RDKJ. \\ 15 The admitted steam volume
)pilog e = Pi(V E S ) + V E /?! loge -jr~ (VE
i + Pi lo ge ~y -- V E -p 1 - v - - p l log e -- *- y E V E V H. *->0 Pz B 99 Dj A H
The quantities F c , p e , Zl i e , ^ i c and p fc can be obtained directly from the Mollier chart. The values plotted in the following diagrams were obtained by this method. In Figs. 10, 11, 12, 13, 14 and 15 the steam consumption for an initial pressure of 14 at. abs., 1 at. abs. back pressure, superheated steam of 300 G and clearance volumes of 5, 8 and 11%, has been plotted against the length of compression A;, for various values of constant values e and constant M. E.P. The construction of auxiliary diagrams (Fig. 9) is recommended for determining the curves of M. E.P., the abscissae representing M.E.P., and the ordi- nates the steam consumption. These curves are plotted for constant compression, and a vertical line intersecting them determines the values of the steam consumption for a given constant mean effective pressure. Meow e pir Fig. 9. 21 The great influence exerted by the compression upon the steam consumption is shown in Figs. 10, 11, 12, 13, 14 and 15. As indicated by previous results, the Fig. 10. Saturated steam 14 at. abs. Noncondensing, Clearance space 5/ . Fig. 11. Saturated steam 14 at abs. Noncondensing, Clearance space 8/ . best economy is obtained with long compressions for early cut-offs e and small mean effective pressures p t and vice versa. It will also be noticed that link valve gears and valve gears controlled by shaft governors acting upon both inlet and 22 exhaust have approximately correct variation of compression (Fig. 14). Valve gears operating with fixed compression can rightly be used in connection with small clearance volumes. Fig. 14 is especially interesting, since it also con- tains the curve of compression obtained with the standard Walschaert gear of the German State Railways for an ex- haust lap of 3 % mm. It is surprising to. find that this curve agrees closely with the calculated minimum values of the compression for constant M. E.P. The shortest cut-off of the link valve gear necessitates a large clearance vo- lume, the bad effect of which is only partly neutralized by a correct variation of compression. It would be more im- portant, however, considerably to reduce the clearance volume required by this type of gear, since the bad effect of large clearance volume far outweighs the correcting influence of the com- pression. The best steam consumption for constant mean effective pressure p t on the one hand, and constant cut-off e on the other, occur at different lengths of compression. If, as it must be, p t is considered the governing variable, then shorter compressions are arrived at. The smaller the clearance volume is, the shorter will be the compressions at which both these minima occur. Figs. 16 and 17 explain the different sets of curves more clearly. In Fig. 17 are repeated two curves from Fig. 15, one being a constant M. E.P. curve for Pi = 10 at., and the other a constant cut-off curve for e = 50%. The diagrams in Fig. 16 marked A (9% compression), and B (76% compression), have the same theoretical steam consumption of 8 kg/HP-hour, the cut-off being in both cases at 50% of the stroke. Dia- gram C, also with a cut-off at 50% but with 47% length of compression, has a steam consumption of only 7.85 kg, while still another diagram D, having the same M. E.P. (10 at.) as diagram C, but with a cut-off at 44% and a Fig. 12. Saturated steam 14 at. abs. Noncondensing, Clearance space 11 / . 23 compression of 19%, only has a steam consumption of 7.6 kg. This last diagram is represented by the lowest point of the curve for constant M.E.P. (Fig. 17). At this point D, however, the constant M.E.P. curve is intersected Fig. 13. Superheated steam 300 C, 14 at. abs. Fig. 14. Superheated steam 300 C, 14 at. abs. Noncondensing, Clearance space 5/ - Noncondensing, Clearance space 8/ . by a curve of constant cut-off at 44%. The lowest point of the latter in turn is intersected by another M.E.P. curve which also has a minimum. By following from minimum to minimum along the M.E.P. and cut-off curves a point is 24 10 finally reached at which the .two minima coincide. This point represents a dia- gram with complete expansion and with compression to the initial pressure, a dia- gram which, as previously proved (Fig. 4), has no volume loss and gives the lowest theoretically possible steam consumption for the assumed range of pressures. Similar curves are shown in Fig. 18 for condensing operation. They refer to superheated steam of 13 at. abs., a steam temperature of 300 C, a back pressure of 0.08 at., and a clearance volume of 2%. In this diagram also the M.E.P. curves essentially determine the best com- pression. For the M.E.P. of 2 to 3 at. ordinarily used it is evident that the best compression approximates 90% at this low back pressure, but even for considerably higher mean effec- tive pressures the difference in steam consumption between 90% and the best compression is negligible. The, difference gradually disappears as the back pressure approaches the absolute vacuum, since in this case com- pression naturally would have no in- fluence whatever upon the steam consumption. Nevertheless, for 2% clearance, superheated steam of 13 at. abs. having a temperature of 300 G, an M.E.P. of 2,8 at., and a back pressure of 0.044 at. abs., the best compression is 90%. These may be considered average conditions for con- densing una-flow engines. This proves conclusively that the long compression of the una-flow en- gine is in no way a necessary evil accompanying the use of piston-con- trolled exhaust ports. The flatness of the M. E. P. curves also indicates that it is permissible to keep the compression constant in the 20 W 60 <30 100 Compress/on //? %* Fig. 15. Superheated steam 300 C, 14 at. abs. Noncondensing, Clearance space 11/ . above case, which is a further argument for the correctness of the una-flow system. The, long and constant compression of 90% of the condensing una-flow engine is therefore correct and admissible. 25 Many authors discuss the ''high compression" of the una-flow engine in the sense of its being unavoidable or undesirable. "High compression" is evidently confused with "long compression". A compression line may be long with low terminal pressure, or short with high ter- minal pressure. Generally speaking, ter- minal compression pressures are too low in the majority f of condensing una-flow Fig. 16. 5,0 X I 3,5 5 'eat ^ 20 ffff 00 Fig. 17. Superheated steam 300 C, 14 at. abs. Noncondensing, Clearance space 11 / . Fig. 18. Superheated steam 300 C, 14 at abs. Condensing (0.08 at. abs.) Clearance space 2/o- engines, and should be considerably higher according to paragraph 9 of the summary. Steam consumption and compression curves Jfor saturated steam, correspond- ing to those for superheated steam shown in Fig. 18, would show only a slight deviation from the latter, with the effect that the best compressions are slightly 26 shorter throughout. The assumption should, however, be remembered that ex- pansion and compression are adiabatic and that the compression is thought to cf 18 Iff 3Md//C/?er flat/m in Fig. 19. Saturated steam 14 at. abs. Noncondensing. C/ect ranee s/icrce Fig. 20. Superheated steam 300 C, 14 at. abs. Noncondensing. begin with the quality of team resulting from continued adiabatic expansion during exhaust. Jacketing of the cylinder, especially in una-flow engines, requires shorter com- pressions, since the effect of the jacket is to increase the temperature of the resi- dual steam, thereby superheating it at an earlier stage and thus raising the com- pression line. The heavy, full line curves in Fig. 19 give steam consumptions plotted against clearance volumes, for different mean effective pressures, a constant length of Fig. 21. Saturated steam 13 at. abs. Noncondensing. Most favourable compression. d 10 Fig. 22. Saturated steam 13 at. abs. Condensing (0.1 at. abs.). Most favourable compression. compression of 90%, saturated steam at a pressure of 14 at. abs. and atmospheric exhaust. The heavy dashed line curves also give the theoretical steam consump- tions plotted against clearance volumes for various mean effective pressures, but for the best compression in each case. The dashed-and-dotted lines are lines of constant best compression. The points of their intersection with the dashed lines give the best compression for that particular combination of M.E.P. and clearance volume. For example, an M.E.P. of 10 at. requires a best compression of 20% for a clearance of 12%, the steam consumption being 9.4 kg. Also for 6% clearance, 28 an M. E.P. of 10 at., and a best compression of 10%, the steam consumption would be 8,8 kg. Fig. 20 shows similar curves for superheated steam of 14 at. abs., a tem- perature of 300 C and atmospheric exhaust. It should be observed that in Figs. 19 and 20 the dashed curves for an M. E.P. of 2 at. show a distinct minimum. At this point the expansion reaches the back pressure. Reduction of the clearance volume beyond the point in- dicated by the minimum of the M. E.P. curve, for a constant M.E.P. of 2 at., results in a loop on the indicator card with a corresponding increase in steam consumption. Fig. 23. Superheated steam 300 C, 13 at. abs. Noncondensing. Most favourable compression. Fig. 24. Superheated steam 300 C, 13 at. abs. Condensing (0.1 at. abs.) Most favourable compression. The dashed curves of Fig. 19 are repeated in Fig. 21. They give steam con- sumptions for different mean effective pressures plotted against clearance volumes, for saturated steam of 13 at. abs., atmospheric exhaust and best compression in each case. Fig. 22 shows similar curves for a back pressure of 0.1 at. abs. Figs. 23 and 24 show corresponding curves for superheated steam of 14 at. abs. and a temperature of 300 G, Fig. 23 being for atmospheric exhaust and Fig. 24 for condensing operation (back pressure 0.1 at. abs.). Apart from the curves for low mean effective pressures it is interesting to note that for atmospheric exhaust the steam consumption shows an almost linear 29 dependence upon the clearance, the best compression being assumed in each case. It is therefore possible to calculate the mean specific volume loss per 1 % clearance and per HP-hour. For dry saturated steam of 14 at. abs., this is found to be 0.0918 kg/HP-hour and 1% clearance for atmospheric exhaust, and 0.1715 kg. HP-hour and 1% clearance, for condensing operation. The corresponding figures for the mean specific volume loss for superheated steam of 14 at. abs. and a tem- perature of 300 G are 0.072 kg for atmospheric exhaust, and 0.12 kg for con- densing operation. For instance in the case of a single cylinder condensing engine running on saturated steam, an increase in clearance volume of 6% will raise the steam consumption by 1 kg/HP-hour. The curves of Figs. 21, 22, 23, 24 also contain the steam consumption of the ideal engine without clearance, which is given by the intersection of the M. E.P. curves with the zero clearance line. The distance of a horizontal line drawn through such points from the corresponding M. E.P. curves gives the volume loss for any particular clearance. It may therefore be stated that for the same initial and back pressure, the same M.E.P. and best compression, the. theoretical steam consumption and the volume loss increase almost linearly with the clearance volume. Excluding small clearances and mean effective pressures, this relation is strictly true for condensing operation and approximately so for atmospheric exhaust. A mathematical expression of the volume loss R can be based on the diffe- rence in area of the diagrams with and without clearance, for
ii /2/o 55?? 55 "KO/ " " " " " " 1 A / 3 /O5 ** /O ?5 ?? 55 55 HO/ 5? 55 55 55 55 55 O \ Q/ ' /05 Zi /O 55 55 55 55 Q O/ " " " " " " 07 O/ These figures also show an approximately linear relation between volume loss and clearance volume, except for very small clearances. As a final result of the foregoing discussions, Fig. 25 shows a simple rule which allows the best compression to be determined for any given case. For a given amount of steam (p the compression must evidently be correct if a displace- ment of the line (p by an amount dq> produces equal changes of area, shown shaded in Fig. 25, on both the expansion and compression sides of the diagram, in such a way that the following equation is satisfied: 427 (ij i e ) d
' w-1 " Pi ^ The bracketed part of the fifth term can be expressed as a series by the bino- mial theorem; considering the first three terms only, the expansion gives: Substituting this value in the equation for <9, , o ~ r /IT/ OP n 1 ft *-\V HI \ / ^/ ^ r- 1 /PW . /^\ 2 /
* * /y6 lilt/lclUlt; I */ **/ ** JU JL/ ,// w \ft/ \^v Hence equation (XV) can be expressed in the form: ax 2 c = x or x 3 -I a; = 0. a a This cubic equation may be written x 3 -}-px-{-q = Q where p = and Cv = , and may be solved by Cardani's formula as follows: \ With the help of equations XV, XVI and XVII, the critical back pressures p 2 have been figured for various initial pressures, clearance volumes, lengths of compression, and amounts of admitted steam. The results are plotted in Figs. 32, 33, 34 and 35, and show the influence of these different variables upon the critical back pressure. Fig. 32 gives the variation of the critical back pressure for different clearance volumes and initial pressures, for a constant amount of steam admitted (p = 10% and a constant length of compression of 100% (una-flow steam engine). The ordi- nates of the individual curves indicate that for a given clearance volume the cri- tical back pressure p z is proportional to the initial pressure p^ which is also evi- dent from the previous equation, 1__ 1 33 =-, or p z = p- n =p i x--*=p l * . (XVIII) Therefore, for the same clearance volume, the same output and the same length of compression, the critical back pressure is proportional to the initial pressure. For the same initial pressure, output and length of compression, the critical back pressure varies with the clearance volume at a steadily increasing rate, at first slowly, then more rapidly, until the rate of increase attains a linear maximum. The critical back pressure is zero for zero clearance volume, and is small for very small clearances. This is self-evident and also proved by the curves, but is frequently left out of consideration in the design of steam engines. The clea- rance volume is therefore the cause of all evil. If the clearance volume is made zero the critical back pressure as well as the volume loss are zero, and the reci- procating engine in this respect is put on the same level as the steam turbine. Noteworthy is the slow initial increase of the critical back pressure, which again calls attention to the necessity of small clearance volumes. 38 Fig. 33 gives the relation of the critical back pressure to the amount of steam admitted, for different clearance volumes, for a constant initial pressure of 13 at. abs. and a constant length of compression of 100%. The critical back pressure grows with decreasing amount of steam admitted. In this case also the necessity for small clearance volumes is evident, especially for early cut-offs, as for instance in condensing una-flow engines. Therefore: for a constant clearance volume, for the same initial pressure and length of compression, the critical back pressure increases with decreasing output. Fig. 32. Critical back pressure plotted against length of compression is shown in Fig. 34. Each curve refers to a different amount of steam admitted, the initial pressure being 13 at. abs. and the clearance volume 2% in all cases. The curves show first a rapid increase of the critical back pressure which attains a maximum at about 30 or 40% length of compression, followed by a steady decrease. The effect of the compression is the less, the higher the MEP or the greater the amount of steam admitted. The curves confirm the lengths of compression ordinarily used in counterflow engines for normal cut-offs of 30 to 40% and more, and the com- pression of una-flow engines (100%) for the usual admissions of 15 to 10% and less. On the other hand Figs. 32 and 33 seem to lend support to the long admis- sions and subdivided pressure ranges of counterflow engines. In the case of multi- stage engines the low pressure cylinder and receiver pressure are the determining factors for the critical back pressure. The compression has the least influence, especially for late cut-offs. It should be noted here that the scale of ordinates in Fig. 34 is twice that of Figs. 32, 33 and 35. 39 The interrelation of critical back pressure, length of compression and clea- rance volume is given in Fig. 35, the initial pressure being 13 at. abs. and the amount of steam admitted 10%. The curves show the immense influence of the clearance upon the critical back pressure, which latter seems to follow a geometrical progression with increasing clearance. The critical back pressure increases rapidly up to about 40% length of compression, the rate of increase being progressively larger for larger clearances, after which it shows a steady decrease with increasing length of compression. In the order of the influence exerted, the clearance volume ranks first, then initial pres- sure, then admission and finally length of com- pression. The need for small clearance volume 9.0 too Fig. 34. loo is imperative. This is fulfilled in the best pos- sible manner by the una-flow engine, especially if fitted with high lift single-beat poppet valves which allow a clearance of 1% to be realized. The length of compression (90%) combined with the short cut-offs as used in una-flow engines is also favorable, since Fig. 32 gives a critical back pressure of only 0.004 at. abs., for P! = 10 at. abs., 9 = 10% 7*= 100%, and S Q = 1 % ; a back pressure which is beyond even the most modern condensing equipment. At the same time the above combination also gives a very small volume loss. The question is not to produce an engine which has the smallest critical back pressure, but one which combines the latter with a small volume loss. It is possible to have a large volume loss and yet a critical back pressure equal to zero, for instance in the case of large clearance volume and compres- sion equal to zero. 40 On account of the adverse influence of high initial pressure combined with short admission, the single stage una-flow engine has to rely upon small clearance volumes which, however, can be realized without difficulty. Compound or triple expansion counterflow engines can be run with more liberal clearance volumes by reason of the long admission and the low initial or receiver pressure and still have a critical back pressure beyond that attainable with the best condensers. The case is very different for single stage counterflow engines which usually have very large clearance volumes. For instance, according to Fig. 35, a single stage engine with iS" =10%, p l = 13 at. abs., 99 = 10%, and F fc = 40%, has a cri- tical back pressure of about 0.5 at. abs. All the bad influences are cumulative; large clearance volume, short admission and the most unfavorable length of com- pression of 40%. It might not be impossible to find an engine combining all these points, in actual operation. For 20% clearance, 15 at. abs. initial pressure, 10% admission and 40% length of compression, the critical back pressure becomes 1 at. abs. To run such an engine condensing would be utterly wasteful. The use of several stages not only reduces the volume loss, but also lowers the critical back pressure, and to such a degree that other defects, such as large clearance volumes, lose their significance, to a certain extent. As can be seen from Figs. 32, 33, 34 and 35, the critical back pressure becomes zero for S = 0, ^ = 0, V k = and attains a maximum for zero admission. In all the above considerations, admission, mean effective pressure and output have the same meaning and may be used indiscriminately. The possibility of causing an increase in steam consumption by going beyond the critical back pressure, as well as the useless generation of too high a vacuum are out of the question in case of well designed una-flow engines. These condi- tions, however, sometimes occur in counterflow engines, even to such an extent that the engineer and fireman are able to notice the bad effect of too high a vacuum. Prof. Josse reports such a case in the Zeitschrift des Vereines deutscher Ingenieure, 1909, page 324. He states that "the economy of the engine improved until the back pressure fell to 0.2 at. abs. From this point onwards a further reduction in steam consumption due to increased vacuum was not noticeable". Such a result in this particular engine was caused not only by the critical pressure being exceeded, but also by increased losses of initial condensation and leakage due to the higher vacuum. The initial condensation was considerable in this case. Further more, the pressure difference between the engine cylinder and condenser increases with a high vacuum by reason of the usual deficiency in exhaust area. In una-flow engines the exhaust port areas can always be made sufficiently large; leakage and initial condensation are reduced because the engine has no exhaust valves and may be fitted with single-beat inlet valves, and furthermore has the benefit of the una-flow action. In well designed and well built una-flow engines the results of the above calculations, which implicitly contain the rule of equal heat changes, do not, therefore, require any appreciable corrections. For given initial and mean effective pressures, the lowest steam consumption is obtained when the clearance volume is zero and the back pressure is zero; the length of compression has then no influence. The critical back pressure, i. e. the most economical back pressure, and the length of compression become of impor- 41 tance as soon as the clearance volume has a definite value. According to Fig. 31 the length of compression can then be increased to 100% and the back pressure reduced until the changes of total heat during compression and expansion become equal, thus offering the best basis for low steam consumption. In other words: for a given initial pressure, given mean effective pressure and given clearance volume, the minimum steam consumption will be obtained when the length of compression is 100% and the back pressure is such that the change of total heat during expansion is equal to the change of total heat during compression. (Una-flow steam engine.) This rule may also be arrived at if the second wording of the fundamental law given on page 31, dealing with back pressure, is applied to una-flow engines. The minimum steam consumption requires the shortest possible cut-off and the longest possible compression of 100%, these being related to each other by the rule of equal heat changes. An examination of the common types of steam engines will reveal the fact that incorrectly designed engines are the rule and correctly designed engines the ex- ception. There is hardly a steam engine designer who is not guilty of some viola- tion in this respect. To begin with, the average designer is not aware of the harm- fulness of the clearance volume, which explains the carelessness with which un- necessarily large clearances are used. The latter are rendered necessary for short cut-offs, for instance in locomotives, marine engines, or non-condensing engines with shaft governor controlling both inlet and exhaust. In marine engines this is the case with the high and intermediate cylinders, while the low pressure cylin- ders usually have unnecessarily large clearances. The lengths of compression are frequently incorrect. These "necessarily" and "unnecessarily" large clearances can be avoided. A knowledge of the above rules is indispensable, as well as re- cognition of the fact that changes in load as well as change of rotation may be accomplished merely by the steam admission organs without change in the exhaust timing. The proper choice of steam distributing organs as well as their arrange- ment and mechanism are also important factors. For instance, a single-stage condensing una-flow marine engine with single-beat valves arranged in the cylinder heads fulfills all of the above conditions. (See Fig. 31, Chapter V.) This engine has the great advantage of a small clearance volume of less than 1 % ; the exhaust timing is independent of the inlet gear and the constant length of compression is correct and permissible. The same reasoning holds true for the stationary una- flow engine having single-beat valves. (See Fig. 6, Chapter VI.) Both types of engine therefore have very small volume losses despite the large pressure ranges. In the same way a considerable reduction of clearance volume in non-condensing una-flow engines can be accomplished by shortening the compression (large exhaust lead). This is shown in Fig. 32, Chapter III, including also the effect of an exhaust ejector, which produces a proper change of compression with the cut-off, thereby further reducing the volume losses. Summary. 1. The volume loss is determined by the clearance volume, initial pressure, back pressure, mean effective pressure and length of compression. It increases with increasing initial pressure and clearance volume, decreases with increasing 42 back pressure and mean effective pressure, and becomes a minimum for a certain length of compression. 2. Correct compression tends to reduce the volume loss; compression may be kept constant for small clearance volumes, but should be varied inversely with the cut-off in case of large clearances (Single cylinder engines). 3. Change of compression ,in case of high vacuum has no material effect upon the steam consumption. 4. The clearance volume loss is zero if expansion reaches the back pressure and compression rises to the initial pressure. 5. The theoretical steam consumption, for the same initial pressure, back pressure, mean effective pressure and best compression in each case, increases nearly linearly with the clearance volume. Apart from very small values of the clearance and mean effective pressure, this linear dependence is almost exact for condensing operation, and approximate for atmospheric exhaust. 6. For given initial pressure, back pressure, mean effective pressure and clea- rance volume, the length of compression must be such that the change of total heat during expansion is equal to the change of total heat during compression. 7. For given initial pressure, mean effective pressure, clearance volume and length of compression, the back pressure must be such that the change of total heat during expansion is equal to the change of total heat during compression. 8. For given back pressure, mean effective pressure, clearance volume and length of compression, the initial pressure must be such that the change of total heat during expansion is equal to the change of total heat during compression. 9. For given initial pressure, back pressure, mean effective pressure and length of compression, the clearance volume must be such that the change of pressure during expansion is equal to the change of pressure during compression. 10. For the same initial pressure, back pressure, the same terminal compres- sion pressure and terminal expansion pressure, and for equality of total heat changes, the lengths of the best compressions have the same ratio as the clearance volumes. (Different mean effective pressures.) 11. With proper proportioning of the length of compression, the clearance volume has to be kept as small as possible; this applies especially to single cylinder condensing engines and the low pressure cylinders of compound and triple expan- sion engines. 12. Subdivision into stages results in reduction of volume losses, the high pressure cylinder having the smallest and the low pressure cylinder the largest volume loss. Intermediate cylinders have a loss between both according to their relative size. 13. For given initial pressure, mean effective pressure and clearance volume, the lowest steam consumption is obtained if the length of compression is made 100% and the back pressure chosen so as to make the change of total heat during expansion equal to the change of total heat during compression (una-flow engine). 14. The critical back pressure is determined by the initial pressure, the clea- rance volume, the mean effective pressure, and the length of compression. It increases in proportion to the initial pressure and faster tlmn proportionally to 43 the clearance volume; it first increases with increasing length of compression, then decreases with increasing length of compression and increasing mean effec- tive pressure. It is zero for initial pressure = zero, clearance volume = zero, length of compression = zero, and attains a maximum for mean effective pres- sure = zero. The clearance volume has by far the greatest influence, the pressure range has less, and the length of compression and mean effective pressure have the least. 15. In compound engines the low pressure cylinder determines the critical back pressure. Under the same conditions compounding reduces the critical back pressure corresponding to the lower initial pressure of the low pressure cylinder. 16. It is not so important merely to achieve low critical back pressure alone as it is to obtain simultaneously small volume losses and low critical back pres- sure. The volume loss may be very large and the critical back pressure may still be zero. Of all single cylinder condensing engines, the single-beat poppet valve una-flow condensing engine has at the same time the smallest volume loss and a critical back pressure which is far below anything that can be reached even with the most modern condensing equipment, mainly on account of its small clea- rance of less than 1% and its favorable length of compression. 44 2b. Additional Clearance Space. Practically all condensing una-flow engines must be able to run non-condensing. In case of breakdown of the condenser, lack of cooling water, or during the winter months when the exhaust steam is used for heating purposes, the engine must be capable of operation with either atmospheric or higher back pressures. The Fig. 1. Fig. 2. simplest way of accomplishing this purpose is the provision of an additional clea- rance space. (See Figs. 1 and 2.) The amount of additional clearance depends upon the initial and back pressure. If the latter is for instance 1 at. abs., the initial pressure 13 at. abs., and the clearance for condensing operation 1.5%, then the additional clearance should be 14.75%, according to the tables to be given later. At the same time this increased clearance will cause a lengthening 45 of the cut-off from 8 to 12% for the same output with non-condensing operation (Fig. 3). The drop in pressure at the end of expansion amounts to 0.8 at. for con- densing and 1.0 at. for non-condensing operation. It is found that for other initial pressures, with approximately the same drops of pressure (0.8 or 1 at. abs.) at the end of expansion, the mean effective pressures produced are about equal. In the previous chapter it was demonstrated that the mean specific volume loss for saturated steam of 13 at. abs. and non-condensing operation was 0.0918 kg/HP- hour and per 1% clearance, and 0.072 kg/HP-hour, and per 1% clearance for superheated steam of 300 G, the other data being the same. For 14.75% clea- rance the total losses amount to 1 .35 and 1.06 kg respectively in the two cases. The general adoption of the additional clearance space despite this considerable increase in steam consumption is due firstly to its simplicity, and secondly to qualities which tend partly to coun- teract this heavy loss. The effect on the overall economy is negligible if an engine operates with additional clear- ance only for several days or hours during the course of a year. Fig. 3 also indicates that although the drop in pressure at the end of expansion is higher when using the additional clearance space, the loss due to incomplete expansion is less; and the condensing cylinder being rather large for non- condensing operation becomes more or less adapted to this condition. The additional clearance also preserves the una-flow principle, including the series arrangement of live steam space, inlet valve, piston and exhaust, which is such a valuable feature of the una-flow engine. Although it is possible to use auxiliary exhaust valves instead of additional clearance, and these valves being relieved of pressure at the time of opening can be of single beat or annular con- struction, no joint at all being preferable to even a tight joint or seat. Care must be taken that the clearance valves which control the additional clearance pocket do not materially add to the cylinder clearance for condensing operation (Figs. 4 and 5). In this respect it is advantageous to provide the clea- rance valves with projections which fill up the space between valve seat and cylin- der surface. The valve area of the clearance valves must be large enough to avoid throttling during expansion and compression. It is also advisable to arrange the additional clearance so that it will act as a kind of heat insulator when the engine is running condensing, which is especially of .value for the crank end of the cylinder. The clearance valve may also be designed in the form of a spring loaded safety valve, but then the above mentioned projections cannot be used. The safety valve action of the clearance valve is unnecessary when the main steam valve is not, or only partly balanced so that it can act as a safety valve. The "safety" inlet valve is preferable to the "safety" clearance valve since its weight and spring load are less. The inlet valve is designed for high speed and held closed by steam pressure, its spring being only strong enough to overcome inertia. The clearance valve on the other hand is heavy and its spring has to overcome the total steam pressure. In case of sudden failure of the vacuum this heavy spring load combined with the great weight of the valve cause an objectionable hammering, which can only be stopped by screwing the valves back. In Fig. 6 is reproduced a diagram such as is obtained from a una-flow engine running non-condensing and fitted with auxiliary exhaust valves, the clearance being the same (1%%) as for condensing operation. It is evident that the ratio Fig. 4. Fig. 5. of expansion is too high. A construction of this kind is shown in Fig. 7, having the auxiliary exhaust valves arranged in the cylinder heads. The increase in clea- rance volume due to these valves was not taken into consideration in the diagram of Fig. 6. The diagram indicates that, assuming the same mean effective pressure, the expansion line reaches the back pressure while the piston uncovers the exhaust ports. The loss due to incomplete expansion is zero. The shape of the diagram indicates the counter-flow action and proves that the cylinder is too large for non-condensing operation, especially for smaller loads, when the toe will change into a loop. This produces a backflow of exhaust steam into the cylinder and a corresponding increase in condensation losses. For loads higher than normal the exhaust action will be partly una-flow and partly counterflow. The loop at the end of expansion cannot occur in engines fitted with additional clearance spaces. 47 The worst feature of auxiliary exhaust valves, however, is their detrimental effect upon the condensing operation of the engine. They increase the clearance space and the harmful surfaces as well as the possibility of leakage, and sacrifice the very valuable series arrangement of live steam space, inlet valve, piston and exhaust. For 1% increase in clearance volume, an additional steam consumption of 0.12 kg/IHP-hour may be expected, superheated steam being assumed. This figure does not include the effect of leak- age and the surface losses caused by the valves and their pockets, nor additional losses due to the operation of these valves while the engine is running condensing. It is advisable to keep these valves in operation even while running condensing, Fi 6 since they are liable to stick after re- maining out of use for some time. If the auxiliary exhaust valves shorten the length of compression also for condensing operation, a larger volume loss results, because 48 an increase in clearance necessitates a corresponding lengthening of the compres- sion. The bad effect upon condensing operation appears all the more objectionable since it occurs during the whole working period; while on the other hand, for short periods of non-condensing service even a considerable increase in steam consump- tion due to additional clearance could be tolerated. For long periods of non-con- densing operation, as for instance during the winter, single-beat auxiliary exhaust valves are preferable to additional clearance. When auxiliary exhaust valves are used, they are usually placed below the engine room floor level, which renders their attendance difficult and the arrange- ment of the piping awkward. The amount of the necessary clearance is determined by the following rule: for a given length of compression, mean effective pressure, initial pressure and back pressure, in order to keep the volume loss as small as possible, the clearance volume should be made large enough to produce equal variation of pressure during expansion and compression (see chapter on volume loss). On an average, the ter- minal expansion pressure for non-condensing operation may be taken as 1 at. gage, and this implies a terminal compression pressure of 1 at. below initial pressure. The following table gives the total amount of clearance volume required, when operating non-condensing, for 90% length of compression, starting with a pressure of 1.03 at. abs. and ending 1 at. below initial pressure, with adiabatic compression and saturated dry steam. The figures are based on the latest Mollier chart. Initial Pressure 8 9 10 11 12 13 14 15 16 at. abs. Total Clearance 27.9 23.8 21.3 19.2 17.6 1625 15.15 142 13.4 O/ /o 49 3 a. Losses due to Throttling. Throttling is a change of state in which the total heat remains constant, the effect of which is to diminish the amount of heat and the pressure difference avail- able for utilization between boiler and condenser. Losses due to throttling may occur in the superheater, steam main from superheater to the engine, stop valves, inlet valves, piston-controlled exhaust ports or exhaust valves, and in the exhaust pipe between engine and condenser (Fig. 1). The losses due to throttling occuring in the superheater, steam pipe, stop valves, and inlet valves may be partly regained in connection with the subsequent expansion, although the greater part is lost. The percentage of this regain depends W Fig. 1. Fig. 2. upon the extent of the expansion. The temperature-entropy diagram in Fig. 2 shows the conversion of a certain quantity of heat at high temperature and small entropy into an equal quantity at lower temperature and larger entropy. The change is represented by the area GCDQKHG and is equal in area to the strip QKLWQ which results from increasing the entropy. The part KNVJK, falling within the area of expansion will be regained, while the part NLWVN below the line of terminal expansion pressure is definitely lost. Throttling losses occuring in the exhaust valves or exhaust pipe are irretrievable, for which reason they must be restricted to the smallest possible amount. They are especially harmful because S/ump/, The una-flow steam engine. 4 50 their effect extends through the whole compression stroke. (See chapter on the relation of the una-flow engine and the condenser.) The heat losses and throttling losses in the steam pipe cannot be separated and are usually combined in one figure; a loss of 0.5 to 1.0% is considered an average and corresponds to a steam velocity of about 40 to 50 m/sec, calculated on the total amount of steam flowing through. The throttling losses occuring between the stages of multistage engines are eli- minated in the una-flow engine. In order to estimate the throttling losses in the inlet and exhaust valves it is necessary to know the relation between effect (losses) and cause (valve area), to FA which the following calculations refer. f ^5 A P: Fig. 3. Determination of inlet valve areas. In Fig. 3, p t and v t represent the pressure and volume of the steam at the dead center position of the piston; piston travel x 1 = and corresponding crank angle 6^ = 0. p 2 , v 2 are the pressure and volume at the point of valve closing, for a piston travel = x 2 and crank angle = (5 2 . p and v are the pressure and volume at any intermediate point where the piston travel = x and crank angle <5; w represents the velocity of the entering steam corresponding to the pre- vailing pressure difference. F is the valve area in sqm,
w-F'dt
n
D"- H (x + s) :-D z -H(x
dx 0,2125 y w-F'dd
x-{- s A (x + s)
(2)
For a small drop in pressure it may be assumed with sufficient accuracy that
w = i * g (Pi P) !
In order to facilitate the calculations, the admission line or rather the curve
representing the change in pressure plotted against crank angle or time will be
replaced by a parabola. It will be shown later that this assumption is admissible
if the object of the calculation is not the shape of the admission line but the final
drop of pressure at the end of admission. This final pressure drop, however, is to
form the basis of the determination of the inlet valve areas. Therefore we may
write
p l p = a d- and p 1 p z = a d z 2
or 2
and therefore
combining equations (3) and (2)
dp dx 0.2125-
+
Jdp . C dx
P h j s-h
2 .9)
p 1 (p p t ) F-6-dd
0.2125 . 001 3 56sqm 2 :=:1 3 i 56 sqcm 2 >
If the maximum valve lift for the assumed valve gear cut-off of 12.5% is
equal to Vio tne valve diameter, then.F max = n - d 0.1 d = 13.56 sqcm, d = 6.6 cm,
and A max = 0.66cm. The common empirical formula, based on mean piston velo-
c m
city would give a steam velocity w m = -- m = 92 m/sec. These calculated values
of F' max and ft max can be realized without difficulty if the single-beat valve is ope-
rated by a lay shaft gear running at twice the engine speed (See final chapter).
Permissibility of the use of the parabola.
It is still to be proved that the actual diagram admission line plotted against
crank angle or time may rightly be replaced by a parabola. For this purpose a
diagram is laid out with the crank angles d as abscissae and the valve openings
as ordinates. This results in a curve such as that shown in Fig. 16. The total
crank angle is now divided into a large number of parts or intervals and for each
59
part the mean valve area is determined (F^ F 2 , F s etc.). It is assumed that p v
= const. In the above example it was assumed that p 1 = 130000 kg/sqin, ^ = 300,
i>! = 0.2 cbm/kg and /v ^ = 26000.
y
Volume of clearance space Fj = s\ weight of steam Q l = - ; stroke volume
-f w 2 w 2 ^ *
~~2~ v 2
= H\ '.=
= O.S- dt =
for i intervals. The
n 360 i
piston is first considered to be moved forward a distance corresponding to the
V 1 26000
first interval without admission of steam, so that p 2 = p r y ', v 2 = ;
;> 2 and y m = -^ 2 >
('o
These values represent the state of the steam at the end of the first interval,
produced by expansion only, without the admission of live steam. w m and y m
being also known, it is now possible to calculate the additional weight of steam
Fig. 16.
admitted, which is dQ = = coefficient of velocity, w = the velocity of the exhaust steam in
m/sec, F = instantaneous value of the exhaust port area in sqm, y = specific
weight and t = duration of exhaust.
Even with highly superheated live steam, the exhaust steam of condensing
engines is always, and that of non-condensing engines in most cases, saturated.
The change of state within the cylinder can therefore be assumed to follow
Mariotte's law:
/?! t>j = pv = p 2 v z = const.
Pii y i5 7i represents the state of the steam at beginning of exhaust,
p, y, y the state of the steam at any intermediate point,
Pzi v zi 72 the state of the steam in the exhaust pipe (condenser, atmo-
sphere, etc.),
Pe, v e, 7e the state of steam at the narrowest place of exhaust port.
x represents the piston travel in % of the stroke measured from the
admission end.
As long as - < 0.577 is p e = 0.577 p. The value 0.577 remains about
P
the same for any steam wetness. If p < 1.735 p 2 then p e = p 2 . The change
of the cylinder volume during exhaust is neglected. Since the weight of
steam present in the cylinder is proportional to the absolute pressure,
p dp Q dQ _ T . , n ,
J - = - PC ; Q =V -y; dQ = w-(o- F y e -dt
p Q
*
dp d Q y (Q-F y e d t ..
~p"~ = ~Q~ ~^ T r~
t = - ; dt = pr (<5 = crank angle corresponding
*- 360 ' 6 to time t)
dp max
Pz
= 129 sqcm. With a port diameter of 12.5 cm,
the area of one port would be 122.7 sqcm
and . a single port would be almost suffi-
cient.
4. For the same engine running non-con-
densing, the conditions may be p t = 3 at. abs.
and p 2 = 1 at. abs., hence = 3, a =20%.
Pz
Then according to Fig. 20, -F max = 48.2 sqcm,
and one port of 10cm diameter having an area
of 78.5 sqcm is therefore already too large.
The formula commonly used, based Fig. 22.
upon mean piston speed, would give the
velocities w =. 13.8 m/sec in case 1, w = 9.4 in case 2, w 24.3 in case 3, and
w = 65 in case 4. This proves the inadequacy of this formula.
In Fig. 22 is shown a diagram in which the admission and exhaust lines were
calculated point by point after the areas for inlet and exhaust had been found
by the above method, a drop of pressure at the inlet from 13 to 11 at. abs. and
at the exhaust from 1.2 to 0.05 at. abs., 13% valve gear cut-off and 10% exhaust
lead being assumed.
68
3b. The Relation of the Una-Flow Engine
to the Condenser.
A high vacuum is of great advantage to the operation of una-flow engines.
Fig. 1 shows compression curves for different back pressures and the same ter-
minal pressure, the clearance volumes being correspondingly changed. These
curves indicate how appreciably the diagram area increases with better vacuum.
At the same time it is possible to keep the compression up to the desired value
by properly proportioning the clearance volume. For a high vacuum the clearance
Fig. i.
volume used may be very small. The limit is usually determined by the design,
2% being considered an average figure.
The duration of the exhaust of a una-flow engine with 10% exhaust lead
and 90% length of compression is only about one half of the time available for
the exhaust of a corresponding counterflow engine. The working steam of the
una-flow cylinder must therefore be exhausted into the condenser in one-half the
time. It is a fact that in the usual design of counterflow engines there exists a
considerable pressure difference between the interior of the cylinder and the con-
denser, which is used to overcome the resistances in the usually too narrow ex-
haust passages. The shortening of the duration of the exhaust in the una-flow
engine is all the more a reason for diminishing to the utmost the resistance between
condenser and engine cylinder, and this can be accomplished by short passages
69
of large area. Furthermore, the exhaust port area of the una-flow cylinder can
easily be made three times as large as the exhaust valve area of the ordinary
counterflow engine. If now the remaining cross-sections have sufficient area to
harmonize with these large exhaust port areas, and the length of the passages is
kept down to the minimum, then complete pressure equalization will result. This
is proved by experience as well as theory (See end of this chapter).
In Figs. 2 and 3 are shown a longitudinal and a cross section of a una-flow
cylinder where the exhaust belt connects over its full width to the jet condenser
placed immediately below it. The injection water enters by means of a perforated
tube placed horizontally across the condenser. As may be seen from these illustra-
tions, the exhaust passages are extremely short and wide so that there is practically
no resistance.
Fig. 2.
Fig. 3.
Figs. 4 and 5 show the application of a jet condenser of the Westinghouse-
Leblanc type to a una-flow cylinder. This condenser and a similar one developed
by the A.E. G. are based upon a principle which formed the substance of a patent
issued to the author. The condenser body in this case forms the support for the
engine cylinder. As in Figs. 2 and 3, this gives a very short connection and large
transfer area, thus insuring equalization of pressure between the cylinder and
condenser.
On account of this complete equalization, the compression begins at the lowest
possible pressure with the result of a considerable gain of diagram area, a corre-
sponding reduction of clearance volume and clearance surfaces, as well as increased
thermal efficiency (Fig. 1). The short duration of the exhaust period due to the
piston-controlled exhaust correspondingly reduces the cooling action of the con-
denser upon the interior of the cylinder. As soon as the exhaust ports are covered
on the return stroke, the connection with the condenser is cut off, any further
cooling is prevented and the heating effect of the steam jacket at the cylinder
end comes into full play without any adverse influence due to the exhaust.
It is fundamentally wrong to interpose oil separators, change-over valves,
feed water heaters or elbows in the connection between engine cylinder and con-
70
denser. Such accessories cause very large resistances to the flow of steam and should
be avoided unless their use is rendered necessary by other important considera-
tions.
The atmospheric exhaust pipe should be connected to the condenser body.
If the connection between the condenser and air pump is shut off, the former
Fig. 4.
Fig. 5.
then acts as a kind of exhaust muffler or silencer (See Fig. 2). This silencer action
should be assisted not only by the volume of the condenser but also by a change
of direction of the steam flow. If no such provision is made, the loud exhaust
will be very objectionable, as is shown by experience.
<. <^ ^ CQ PdJP uoj sy/^i/o
cr-p
71
4. Losses due to Friction. (Mechanical Efficiency.)
Dimensioning of Driving Parts.
Very complete data are available for the dimensioning of driving parts of
stationary una-flow engines. From these data have been compiled the curves
shown in Fig. 1, which apply to steam pressures of from 10 to 12 at. gage, and
condensing operation.
In Fig. 1 may be seen the weight of the reciprocating parts including two-
thirds of the weight of the connecting rod, plotted against cylinder diameter.
The average values may be represented by a curve according to the equation
r __ (Cylinder dia. in cm) 2 - 5
&VT -26-
The weight of the piston is about 3 / 10 , that of the connecting rod 2 /7 f the total
weight of the reciprocating parts.
Since the ratio of stroke to cylinder bore is the determining factor for the
reciprocating weights, the ratio of stroke to cylinder bore is also shown in this
chart. It will be observed that small engines have a proportionally long stroke,
while large engines have a proportionally shorter stroke. Since the average buyer
of engines generally has a prejudice against what may be called high speed in
the sense of high number of revolutions per minute, regardless of piston speed,
small engines are therefore built with a comparatively long stroke. For large
engines a high number of revolutions is usually demanded, and since the majority
of builders have a similar dislike for high piston speeds, an engine of short stroke
is the result. It seems strange, however, that the type of frame used does not
appear to have any bearing whatever upon the bore and stroke ratio although
some designers are inclined to make side crank engines with long, and center
crank engines with short strokes (See Fig. 1).
The piston speed based on cylinder diameter shows a more rapid increase
for the smaller sizes than for the larger ones.
The maximum continuous load rating of una-flow engines usually corre-
sponds to a mean effective pressure of about 4.5 kg/sqcm. If the mechanical effi-
ciency for this load is assumed to be 0.94, then the figures for the HP output
obtained agree closely with those given by the makers. The normal rating usually
corresponds to a mean effective pressure of 3 kg/sqcm based on brake HP. This
figure is evidently a compromise between high economy and low initial cost.
The lowest curve in Fig. 1 represents weight of reciprocating parts divided
by rated brake HP. The values show small variation (3.45 to 4.15 kg/BHP) but
increase gradually with the cylinder bore.
The curve between 3.6 and 5.6 kg/sqcm represents inertia of reciprocating
parts for an infinite length of connecting rod.
72
Fig. 2 gives first an indicator card having an MEP of 4.3 kg/sqcm corre-
sponding to the maximum continuous load. From this card are developed three
net pressure diagrams containing inertia curves plotted for a length of connecting
rod equal to five times the crank radius, for three different cylinder bores of 500 >
Net pressure and Inertia force Curves.
MEP = 4,3 kg/ cm
10-
Cylinder dia. 500 mm
Crank pin
Main bearing, for
50% balancing
Cylinder dia. 900 mm
Piston rod
Crosshead pin
Fig. 2.
900 and 1300 mm. For any piston position the vertical distance between inertia
curve and net pressure line represents the load upon that driving part to which
the inertia curve refers. The dashed and dotted lines apply to the load on the
piston rod, the dashed lines to the crosshead pin, and the full lines to the crank
pin. The dotted curves in the same way give the load on the main bearings, 50%
of the weight of the reciprocating parts being balanced. For center crank shafts
two equally loaded main bearings of equal size are assumed, while for side crank
73
shafts the main bearing loads have been increased by about 20% on account of
the overhang. In regard to bearing load, apart from impact, it would be more
advantageous if the inertia forces for the smaller engines would correspond to
those of the larger size in the last diagram of Fig. 2; and this could be accom-
plished by increasing the speed, for an engine of 500 mm cylinder bore, from
162 r. p. m. to say 188 r. p. m.
The proportions of piston rod and tail rod as well as diameters of the diffe-
rent bearings are plotted as functions of the cylinder bore. It will be noted that
the ratio of piston rod diameter to cylinder bore is slightly less for engines of
large size, by reason of the proportionally shorter stroke of the latter. The dis-
tance from rear end of piston to center of crosshead pin is usually about 3.33 times
the stroke. The factor of safety against buckling of the piston rod, based on the
loads represented in the diagram of Fig. 2, is 10 for small engines and 9 for larger
engines.
Average ratios.
Crosshead pin diameter to cylinder bore 0.265
Side crank, crank pin diameter to cylinder bore 0.33
Side crank, main bearing diameter to cylinder bore . . . 0.5
Center crank, crank pin diameter to cylinder bore . . . 0.425
Center crank, main bearing diameter to cylinder bore . . 0.425
Length of crosshead pin to its diameter 1.3 1.6
Side crank, crank pin length to its diameter . ... . .1.0- 1.2
Center crank, crank pin length to its diameter 0.9 1.0
Side crank, main bearing length to its diameter 1.3 1.8
The crank pin diameter of center crank shafts will be found only in rare cases
to be larger than the diameter of the main bearing, and the main bearing at the
flywheel side longer than the opposite main bearing.
F
Fig. 1 further shows the ratios of piston area to bearing areas which
(I ' a)
give the following averages: Crosshead pin 8, side crank, crank pin 7. 7, center crank,
crank pin 4.25, side crank, main bearing 1.95, and each main bearing of center
crank shafts 1.5.
Combining these data with the specific loading taken from the diagrams in
Fig. 2, the resultant bearing pressures were calculated and are shown at the top
of Fig. 1. These values refer to maximum continuous load and smallest dead
center inertia, horizontal forces only being considered. Strictly speaking, the
additional forces due to flywheel weight, belt pull, etc. should be combined with
the horizontal forces, but this would not materially alter the results. It will be seen
that crank pin and main bearings of side crank shafts sustain about 50% higher
loading than the corresponding bearings of center crank shafts. The highest bearing
pressures given in Fig. 1 are undoubtedly permissible in case of force feed lubri-
cation.
On side crank shafts, excessively large leverages, i. e. distances from the con-
necting rod center to the main bearing center, cause increased bending, higher
74
bearing pressure and on account of the deflection of the crank shaft, increased
pressure at the inside edge of the main bearing. This tendency can be reduced
by shortening the leverage or the use of self aligning bearings, or both (Fig. 3).
Since, there are no secondary forces acting on the connecting rod in the horizontal
plane, the factor of safety against buckling in this plane need not be more than 5,
as against a factor of 9 or 10 in the vertical plane. This condition can be easily
met by flattening the otherwise circular rod section. The crank hub should be
placed as close as possible
to the connecting rod and
should not be wider than
0,65 times the shaft dia-
meter for large engines
and 0.75 for small engines.
On some large Belgian
engines this figure is even
cut down to 0.6. The
projecting part of the
crank pin bearing cor-
responds to the projecting
part of the crank hub.
The rear side of the crank
in this case becomes flat.
The crank is frequently
pressed in place on the
shaft, although a shrink
fit is preferable by reason
of the lesser chance of
damage to the structure
of the material. A key,
although frequently used,
is unnecessary. The same
applies to the connection
of the crank pin to the
crank arm, and the former
should be ground after
assembly.
Fi 8- 3 - UsssL^J A still shorter over-
hang may be obtained by
casting crank, crank pin
and crank shaft in one
piece of cast steel. < By designing the crank in the form of a disc, as shown in
Fig. 4, an especially ,large reduction in the overhang may be obtained, with a
corresponding decrease in the shaft diameter. The present state of foundry prac-
tice allows such a construction to be used without anxiety.
Constructional data of a stationary una-flow engine built by Sulzer Bros,
and installed in a cotton spinning mill at Grefeld are as follows: Cylinder bore
| 23000 Kc)
75
1100 mm, stroke 1200 mm, speed 110 r. p. m., steam pressure 12 at. gage, steam
temperature 320. The engine is of the center crank type.
Main bearings, 475 mm dia. by 650 mm long
Crank pin, 475 mm dia. by 380 mm long
Crosshead pin, 300 mm dia. by 430 mm long
Piston rod 220 mm diameter, tail rod 170 mm diameter
Weight of piston 2125 kg, weight of piston rod 1500 kg
Weight of crosshead 1532 kg, 2 / 3 connecting rod 1780 kg
50% of the reciprocating parts are balanced
by counterweights fastened to the crank
cheeks.j \
The friction HP of this engine at 110 r. p. m.
with a smooth flywheel was 113,6. The correspon-
ding mechanical efficiency for a rated load of 1700 1 HP
is therefore 0.933, a figure which disproves the
opinion frequently advanced that the mechanical
efficiency of una-flow engines is low. The assumption
that the engine friction must be nearly independent
of the HP output is based on the fact that the load
on the driving parts is practically
the same for idling as it is for K
the rated HP Starting with the '-
-j
^
^
^
s s/ s
1
\
^
^r
1:
^
1
1
1
^
ti
OOx
SS>
^s
xxX
^
^
^N:
^t
mi
-
**
engine idling, and gradually in-
creasing the output, the gross \
^
load on the driving parts first A
decreases slightly, at rated output ^
reaches the same value as for \^
idling, and then becomes some-
what greater for larger output.
The engine at Crefeld, for instance, has
center inertia load of approximately 6 ]
a corresponding mean pressure of the ine
gram of 3 kg/sqcm, and a useful rated me;
tive pressure of 3 kg/sqcm. It follows 1
engine friction for idling and rated outp
be the same. Further, the weight of the
flywheel, crank shaft and the rest of the
parts as well as the centrifugal force oftheco
rod end, crank and counterweights are res
for a constant portion of the total friction
The above will be further illustrated
following test results obtained by Sulzer B
\
^ / ^^ / /'s
a dead
cg/sqcm,
rtia dia-
m effec-
,hat the
ut must
piston,
driving
rmecting
ponsible
by the
ros.
Fi
/s
I
->v
1
j
g-
4.
A una-flow engine of 700 mm cylinder bore and 900 mm stroke, having a rope
flywheel of 4000 mm diameter, gave the following friction at different speeds (without
ropes). 133 112 100 85 68 r. p. m.
66 53 46 38 28 I HP friction.
76
The corresponding rated load would be
820 690 615 520 420 I HP,
so that the friction HP in % would be
8 7.7 7.5 7.3 6.7%.
Another engine of 600 mm cylinder bore and 725 mm stroke, fitted
with a rope flywheel of 2400 mm diameter, with 14 grooves, gave the following
results: 150 100 50 34 r. p. m.
41 22 8.5 4.5 I HP friction.
The corresponding rated load in this case would be
475 317 158 108 I HP,
so that the friction HP in % is
8.65 7 5.4 4.15%.
In reducing the engine speed from 150 to 50 r. p. m. the friction HP should
diminish from 41 HP to 1 / 9 of the same, or 4.5 HP. It actually was 8.5 HP, which
reflects the influence of the weight of the driving parts. The weight and inertia
of the latter have an equalizing effect, so that for constant speed the friction HP
remains nearly constant independently of the instantaneous output.
Dimensions of Driving Parts of Other Una-Flow Engines-
(Sulzer Bros, design.)
Steam pressure at admission valves 12 at. gage. All bearings have force feed
lubrication.
1. 550 mm cylinder bore, 650 mm stroke, 158 r. p. m.
Two main bearings 250 mm dia. by 360 mm long
Crank pin . . , . . 250 200
Grosshead pin 150 220
Grosshead shoes 520 long 300 wide
Piston rod 110 dia.
Connecting rod length .... 5.5 times the crank radius.
2. 500 mm cylinder bore, 600 mm stroke, 165 r. p. m.
Two main bearings ..... 230 mm dia. by 330 mm long
Crank pin 230 180
Crosshead pin 140 200
Crosshead shoes 480 long 275 wide
Piston rod 100 dia.
Connecting rod length .... 5.5 times the crank radius.
3. 850 mm cylinder bore, 1000 mm stroke, 125 r. p. m.
Two main bearings 375 mm dia. by 550 mm long
Crank pin 375 310
Crosshead pin 240 ,, 350 ,,
Crosshead shoes 800 long 500 wide
Piston rod 150 dia.
Connecting rod length .... 5.5 times the crank radius.
77
Sulzer Bros, also mention the fact that they have found the friction HP of
their una-flow engines equal or slightly less than the friction HP of their tandem
compound engines of equal power.
The driving parts of a una-flow engine should naturally cause more friction
than the parts of a tandem compound engine since the size, or rather diameter,
of the bearings of a una-flow engine is larger on account of the higher piston load.
In a una-flow engine the single piston carries live steam pressure, while the low
pressure piston of a tandem compound engine only carries receiver pressure and
the much smaller high pressure piston sustains the difference between the live
steam and receiver pressures. On the other hand, the single una-flow cylinder
with its one piston and one or two piston rod packings will cause less friction than
the two cylinders with two pistons and three or four rod packings of the tandem
counterflow engine. Steam cylinders arranged in tandem are furthermore subject
to misalignment with accompanying binding of the moving parts, and the friction
Fig. 5.
caused by such misalignment may be considerable. The piston system in una-
flow engines can always be supported at two points only, for instance by means
of a crosshead and self-supporting piston, or on crosshead and tail rod support
with floating piston. It is assumed of course that the metallic packing used is of
such design as to permit of lateral movement of the piston rod. Furthermore,
the tandem counterflow engine requires four times the number of steam distri-
buting elements as the una-flow engine (8 valves against 2 valves), with a corre-
sponding increase in friction. A comparison between a una-flow and a cross-com-
pound engine will still further emphasize the advantages of the una-flow system.
According to what is said above, the lesser friction of the una-flow cylinder must
outweigh the increased friction of the una-flow driving parts, if the experience
of Sulzer Bros, is accepted as having general application.
That part of the engine friction caused by the piston, especially if self-sup-
porting, may be considerable. This friction may be reduced by fitting the piston
with shoes of bronze, babbitt or Allan Metal. The friction is least with a floating
piston, i. e. a piston supported by its rod, despite the additional friction of the
second stuffing box and tail rod support.
78
The una-flow piston should be made as light as possible, and this may be
accomplished by constructing it in two parts of cast steel (See Fig. 5), thereby
reducing friction, inertia and impact.
Next to the piston, the main bearing, crank pin and crosshead pin contribute
the largest share to the total friction. The friction of high grade metallic packings
is extremely small, as is also the friction of the poppet valves and their gear. This
applies especially to una-flow engines with only two valves and one packing.
The friction of the driving parts can be considerably reduced by a proper
oiling system, especially by means of force feed lubrication. The latter type
also reduces impact. The provision of force feed lubrication for the condenser
pump driving parts and the use of a housing around the flywheel are further mea-
sures in the right direction.
Short stroke engines with bearings of large diameter naturally have a higher
friction loss than long stroke engines.
In engines having steam jackets on the cylinder barrel the friction is usually
greater if the jackets are shut off. In the same way a new engine while being run
in will have more friction than later, and the friction of an engine immediately
after starting will be larger than when in regular operation, especially if it has
not been warmed up previously.
Taking it altogether one may say that the una-flow engine has a slightly better
mechanical efficiency than the ordinary tandem compound engine.
79
5. Losses due to Leakage.
Valves, Pistons, Piston Rod Packings.
Tight steam distributing organs are a rare exception. Slide valves are gene-
rally considered to be tighter than piston valves, this being the reason for the
practice of many concerns to use piston valves for the high pressure and slide
valves for the low pressure cylinders, a practice which is also supported by pres-
sure and temperature considerations.
Corliss valves are fairly tight, but they are far from being absolutely tight.
Well made piston valves fitted with snap rings may be considered fairly tight.
Piston valves without rings should only be used in small sizes for saturated steam
and must be made a good fit. Larger piston valves for use with superheated steam
should always be equipped with snap rings on
account of the necessary clearance required for
expansion. Even then a certain amount of
leakage .must be expected in those positions in
which none or one ring only is active, in
addition to the constant amount of leakage
past the ring joints. In case of highly
superheated steam, carbonized oil may be
the cause of increasing leakage.
Double beat valves are usually leaky. The
leakiness increases with the amount of balance,
the pressure and the temperature. With all
types of valves leakiness will be enhanced with
increasing superheat, on account of warping
and the increasing fluidity of the steam.
The body of double-beat valves as shown
in Figs, 1, 2 and 3 will sustain a heavy load in
the direction of the axis during the expansion
and exhaust periods. The corresponding
deflection will cause the lower face of the valve
to leave its seat and leak. The radial forces
can be neglected if the seats are made flat.
In the same way, if the temperature of the valve is higher than the tempe-
rature of the material forming its housing and seat, then the valve body will
expand more than the latter, and the upper valve face will lose contact and start
to leak. The above temperature difference may have several causes. In a valve
design as shown in Fig. 1, in which valve and seat have the same height and the
same thickness of material, equal expansion, in the most favorable case, will
Fig. 1.
Fig. 2.
80
occur only if the material of both parts has the same coefficient of expansion.
This can be realized by due attention to the work of the foundry. Both parts are
exposed to live steam temperature on one side and to the varying temperature
of the cylinder steam on the other.
Conditions are not so good in the design shown in Fig. 2, in which a valve
cage of the ordinary type is used. Valve and seat are of different thicknesses and
therefore expand unequally, especially during the first period after starting. Fur-
thermore, the two parts will have different temperatures during operation, since
the valve is exposed on one side to live steam, while the cage is entirely surrounded
by cylinder steam.
The most unfavorable design is the one shown in Fig. 1, page 4, which
has the valve seats cast in one piece with the cylinder head. The difference in
expansion between valve and seat, especially just after starting up, must
be considerable. The coefficients of expansion and the mean tempera-
ture of both parts will certainly be different and the corresponding
leakage will therefore be considerable.
Fig. 3.
Fig. 4.
The valve will leak the more, the higher it is. The first rule in poppet valve
design is therefore to make the valve as low as possible. For this reason, valve
gears should be avoided which at late cut-offs give unnecessarily large valve lifts
and therefore require high valves, unless a restriction of the upper passage is not
objectionable. It is further advisable in cases where valve cages or similar con-
structions are objected to on the ground of the number of tight joints required, to
use a kind of saucer to form the lower valve seat, thus at the same time reducing
the height of the valve. (See Fig. 3.) In this design the vertical forces as well
as the deflection of the valve body are reduced on account of its small height
and the small radial width of the valve ring. By grinding in at the operating
temperature, a close approximation to complete tightness for one particular pres-
sure and temperature will be obtained. During the first period after starting and
at any other pressure and temperature the valve will leak. Perfect tightness under
81
all conditions can be accomplished with a resilient valve such as is shown in Fig. 4.
The lower valve face is smaller in diameter than the upper one. This difference
produces a force in addition to the spring load, which tends to press the lower
rigid face as well as the resilient upper face against the corresponding seats. The
upper resilient face also takes care of unequal expansion. In order to make the
valve low and to reduce its deflection, a saucer forming the lower valve seat may
be used to advantage. Dirt in the steam, however, may cause leakage even with
this type of valve.
Theory of the Resilient Valve.
The following forces act on the valve:
Downwards:
1. The spring pressure minus the steam pressure on the stem,
where p a is the absolute pressure in the valve chest.
2. The pressure on the upper side of the resilient annular ring,
where p t is the absolute pressure in the cylinder.
Upwards :
3. The pressure on the lower side of the lower annulus
(r* Q *)n(p a ~ Pi ).
4. The upward reaction of the upper valve seat W v
5. The upward reaction of the lower valve seat W z .
The horizontal forces balance each other.
The sum of the vertical forces must be zero, so that
P + (& Q z )n( Pa Pi )-(r* Q*)n(p a - Pi ) W, W 2 = . (1)
If the radii, the steam pressures and the spring pressure are known, only
the bearing reactions W 1 and W z remain unknown.
W 1 may be found from the condition that the deflection f of the resilient
ring due to the steam pressure (measured at the middle of its face), must be equal
to the deflection / 2 caused by the seat reaction W^ provided the lower valve face
is to remain on its seat. Therefore f l = / 2 .
Imagining a piece cut out radially from the valve with the angle dcp at the
center, then the steam pressure acting on this element of the resilient surface is
and the corresponding deflection is
*>(*-<>) (p. -p0 .... (2)
where E = the modulus of elasticity and J = the moment of inertia (approxi-
mately constant) of the cross-section at right angles to the plane of bending.
Stumpf, The una-flow steam engine. 6
82
The seat reaction on the radial element under consideration is W 1 H^-. and
2n
the corresponding deflection is
^ = Wl ' In' 3E-J ' ' ' (3)
Then on equating / x = / 2
W l = -Q- n (R -\- Q) (R Q) (p a p^ (4)
From equation (1)
Substituting in (5) the value of W as found in (4), then
In the limit, when the valve just rests on the lower seat, W 2 = 0. Neglecting
P, the excess of the spring pressure over the steam pressure on the valve stem,
the highest permissible value for Q may be obtained from the following equation:
"8" '~ r + ~s^'- < 7 )
Denoting R by Q -f- a, and r by Q -f- 6, equation (7) becomes
10
or since a-\-1q and 6 -f- 2p are approximately equal,
If the valve expands by the amount A I in excess of the casing, in consequence
of unequal temperatures and coefficients of expansion, the resilient ring must
deflect by the same amount in order to remain steam tight. In this case /, / 2
is not zero but is equal to 41 _ ./ / /cn
//t /i /a ......... (v)
The seat reaction W-^ for any given A I is obtained by inserting the values
for /j (from equation 2) and / 2 (from equation 3) in equation 9; hence
For the valve to be tight, W 1 must be positive, for which purpose a pressure
difference (p a Pi ) is necessary, as may be seen from equation (10). The mini-
mum pressure difference required for tightness may be obtained from equation
(10) by making W^ 0, i. e.
d
83
From this equation it will be seen that the pressure difference is proportional
to Al and consequently to 1. In order to keep (p a />) small, the valve must
be low. The dimension d is determined by the strength of the material, and the
values of R and Q by the desired steam velocity.
To find the bending stresses of the resilient ring during operating condi-
tions; let
^ = the temperature of the valve,
Q! = the coefficient of expansion of the valve material,
t 2 = the temperature of the surrounding casing,
a 2 = the coefficient of expansion of the casing,
I = the distance between valve seats at normal temperature t .
Then Al = l\a (t t) a (t Ml (12)
Substituting this value in equation (11), the pressure difference (p a p^
will be found at which the valve will commence to be tight. At all greater pressure
differences the valve will be perfectly tight.
For example, a valve is taken in which / = 30 mm, R = 125 mm, Q = 104 mm,
d 3 mm, p a = 12 at, and p t = at.
Assume also that t =15
*! - 300
% = 0.000012
a 2 = 0.000011.
The following table gives the relative expansions Al of this valve for certain
temperature differences (t Z 2 ), calculated according to equation 12. In this
table are also found the stresses K^ and the pressure differences (p a p^ neces-
sary for tightness calculated by means of equation 11.
t l t z
50
100
150
200
Jl =
in mm
0.009
0.0255
0.042
0.0585
0.075
Pa ~ Pi =
in at abs.
1.35
3.8
6.3
8.75
11.25
K =
230
645
1070
1490
1920
From these figures it will be seen that many prevalent valve constructions
are far from being steam tight. The excessive height of many valves, neglecting
other constructional errors, makes steam tightness absolutely impossible. Even
in a valve only 30 mm high, with a temperature difference between valve and
seat of 200 G., a steam pressure of 11.25 kg/sqcm is necessary in order to obtain
steam tightness; i. e. at all lower pressures there will be leakage.
6*
84
Calculations for a Resilient Valve of 160 mm mean diameter, for a
condensing engine working with steam of 9 at. abs. (Fig. 5).
The following two condi-
tions are assumed for the cal-
culations of W l and W 2 :
1. The actual closure oc-
curs at the mean circum-
ference in both the top
and bottom seats (Radii
R m and r m ).
2. In the most unfavorable
case closure occurs at
the inner edge of the
upper seat (Ri) and the
outer edge of the lower
seat (r a ).
In the latter case W 2 on
the lower valve seat is to be
zero.
In Fig. 5,
Fig. 5.
= 81.5 mm
= 76 mm
Ri = 80 mm
r a 77.5 mm
6*71
Q = 67.5 mm
d = 17 mm.
The steam pressure on the valve stem is (p a 1) =18 kg.
Wj, according to equation (4) above, is
221 kg, for R m and r m
195.5 kg, for R t and r a .
W 2 = in case 2. The spring pressure F is then calculated from equa-
tion (5),
(Pa-1)
( Pa - Pi ) (R? n-r*. n]
W l
= 18 + 195.5 107 = 165.5 kg.
With this spring pressure in case (1)
W 2 = P +(p a Pi) (R m * n r m 2 n
= 88.5 -f 245 221 = 112.5 kg.
The thickness d of the resilient ring is calculated for the case in which the
valve is opened at the greatest pressure difference p a p,, the valve and seat
being assumed to have the same temperature.
Again taking a radial element d = 10.5, and the time for 1 of crank angle t 1 / 600 sec. For any other speed n
the valve velocity v must be multiplied by TT-T- and the acceleration by
Time or crank angle is taken as the abscissa for all calculations. The investi-
gation will extend only to the point at which the cam lever reverses, since the
j-i
Fig. 11.
subsequent closure of the valve is an exact repetition of the phases during lifting.
The necessary dimensions of the valve gear parts are assumed to be known or are
determined according to Figs. 12, 13 and 14. From Figs. 13 and 14 the travel
of the cam lever (a = KK'} can be found corresponding to the crank angle a.
90
\
Fig. 12.
This corresponds to the part MM' of the valve lift curve (equidistant from the
cam profile through the center of the cam roller) and the valve lift h. The cam
lever therefore changes a
I into h. The line / K in
Figs. 12 and 13 corresponds
to the z-axis in Fig. 14. The
distance a = KK' corresponds
to the angular displacement
/? of the cam lever. Imagine
the roller to move along the
cam (the latter supposed
held stationary) from M to
M', corresponding to the
angle /5. ML is the center
line of the valve stem.
Drawing M' L' tangentially
to an arc struck from
center N with radius N L,
the base circle MM' will be
intersected at a point whose
distance from M' =h= valve
lift, corresponding to a and p. The valve lift curve M M'Q is obtained from
the cam profile by increasing or decreasing the radius of the latter by the roller
radius. O^M is usually from 3 to 10 mm, and O^M'
= roller radius -f- 3 to 10 mm. The cam profile as well
as the valve lift curve consist of two circular arcs with
a common tangent between them. As long as the roller
remains on the arc of the cam with center O^ it is
advisable for the sake of accuracy to take the in-
crements of the crank angle a equal to 2, while,
later on, increments of 5 are sufficient. Fig. 15 shows
the valve lifts thus determined plotted against the
crank angle a. ^ max cor-
responds to the dead center
position of the eccentric.
This curve is also important
for the determination of the
admission line (see chapter
on losses due to throttling).
Instead of now differentiating
this valve lift curve to find
the valve velocities, the latter
may be determined in a more
exact manner from the in-
stantaneous angular velocity
and the corresponding lever
91
arms. The velocity of the eccentric rod H K in Fig. 12 is found from Fig. 14
to be
>! = V - 1 =
r
2 n r n
bO
= co
co r-,
The angular velocity of the cam lever is o^ = - (Fig. 13). The "roller
r 2
contact line" M'0 2 P always intersects the roller center and either of the centers
O : or <9 2 , or i g at right angles to the straight middle piece of the valve lift curve.
The velocity along the roller contact line is v' = o^ r 3 '. For the velocity along
the relative valve stem axis M' L' (valve velocity y), 7*3 has to be considered instead
of r 3 ', since the right-angled triangles NSP and M' TU are similar, and v : v'
= r 3 : r 3 '. Therefore the valve velocity v
r 3 = co r 3 . The distances
r-,, r 9 and r, are to be measured in meters. The roller contact line M'0 Z P coin-
cides with MN for the point of valve opening
and therefore r s = 0. The same occurs when the
roller reaches the upper land and M' falls on Q,
in which case also r 3 = 0. The valve velocities
thus determined are plotted in Fig. 15 on a
crank angle basis, giving the curve y, which
indicates a rapid increase of the velocity up
IT
Fig. 14.
Fig. 15.
to the point 7\, corresponding to the point at which the lifting curve and
straight portions merge, after which it decreases until it reaches zero for the
dead center position of the eccentric, or for the point at which the upper curve
and upper land run together. The velocity u =. corresponds to the dead center
position of the eccentric after which the whole procedure is repeated with the
signs reversed. In case the roller reaches the upper land or runs a certain distance
on it, a corresponding part of the velocity curve coincides with the #-axis.
The maximum valve acceleration p = corresponds to the steepest tangent
Wv
7\ T 2 of the velocity curve. Receding from point T z a distance equal to 6 crank
angle or 1 / 100 sec, an ordinate at this point will represent the change of velocity
in Vioo sec - In this case the latter amounts to v x = 0.205 m/sec. The accelera-
92
tion is consequently p =- - ==20.5 m/sec. If the scale of velocity is chosen
/ioo
so that 100 mm = 1 m/sec, then the ordinates in mm give directly the accelera-
tions in m/sec. The accelerating force during the period of increasing velocity is
exerted by the eccentric through the cams, and for decreasing velocity the retarding
force must be produced by the valve spring. The opposite holds true for the
closing period of the valve.
The force to be exerted by the valve spring depends upon:
1. The maximum acceleration p max to be produced by the spring.
2. The weight G of the valve and the parts connected to the same.
3. The friction of the valve stem, valve and roller head.
4. The steam pressure upon the valve stem area.
The weight G is generally assumed to be balanced by the friction and there-
fore neglected.
g
The accelerating force P = jz-rrr p max . An additional 10 to 20% are neces-
y.oj.
sary to take care of inaccuracies in the construction of the valve gear and the
increased friction during the first period of operation. The steam pressure on
the valve stem area may either relieve or oppose the spring and must in every
case be considered. Torsional stress of the spring material k d = about 3500 kg/sqcm.
Pistons.
Pistons may be classified as:
1. Self-supporting pistons.
2. Floating pistons.
3. A form intermediate between the other two.
All questions concerning the design, construction and operation of the piston
should receive thorough consideration. The self-supporting piston is undoubtedly
the most difficult, and the floating piston the easiest to deal with. The self-
supporting piston together with its cylinder form a bearing. The first condition
for satisfactory operation is therefore sufficient difference in the properties of
the materials of which the two parts are made. For instance, a steel shaft runs
quite satisfactorily in a babbitt, phosphor bronze, brass or cast iron bearing. All
these combinations present sufficient difference in the properties of the materials
employed. An exception exists in the combination of hardened steel on hardened
steel which may frequently be found in valve gear joints. Babbit on babbitt, bronze
on bronze, or mild steel on mild steel never work together satisfactorily. Cast
iron pistons, however, can be made to work well in cast iron cylinders, as is shown
by the una-flow engines built by Sulzer Bros. Cast iron is a collective designation
which includes materials of very heterogeneous composition. Sulzer Bros, after
extensive experiments have found the proper mixtures for the cylinder and piston,
which possess sufficient difference in properties to work together safely and satis-
factorily. If a designer lacks sufficient faith in his foundry he will do well to equip
his piston with a babbitt or bronze mounting in order to provide materials with
93
sufficient difference in properties. A cast steel piston should always be equipped
with such a bearing surface. There are still, however, concerns who try to make
piston and cylinder from the same mixture, and in addition to this cardinal error
commit others of equal consequence which result in certain failure. There are also
materials which do not work well together despite sufficient difference in their
properties, such as for instance, cast steel on cast iron. Even with a bronze mounting
the difference in expansion between these two materials must be taken into con-
sideration.
A journal and bearing must ordinarily have sufficient clearance to allow room
for the oil film. This condition applies all the more to piston and cylinder since
the dimensions are larger and the temperature difference is greater. A clearance
of 3,5 to 4 thousandths of the diameter between piston and cylinder bore have
been found satisfactory. Machining the piston by first turning it to the exact
cylinder diameter and afterwards turning it off eccentrically so as to produce a
bearing surface for about 90 to 120 has not proved satisfactory for superheated
steam. This method gives enough clearance at the top but on account of the higher
temperature of the piston and its greater expansion, the weight concentrates at
the edges of the bearing surface and the piston is likely to seize. A proper way
of finishing the cylinder is to bore it barrel-shaped, or machine it while heating
the ends, for instance by passing steam through the jackets and cooling the center
by blowing air through the exhaust belt. The heads of a una-flow piston expand
more than the center by reason of their higher temperature, and should therefore
be of smaller diameter than the center. They should be out of contact with the
cylinder and only act as plungers. The part of the piston forming the bearing
surface should be rounded off liberally at its ends to prevent it from scraping the
oil off the cylinder wall. Briefly, the endeavor must be to produce a piston and
cylinder with exact cylindrical surfaces and sufficient clearance, and to maintain
this condition at high temperatures. If this can be accomplished for the long una-
flow piston with its large bearing surface and temperature difference, the most
difficult part of the problem is solved.
The case of the floating piston which is carried by the piston rod is much
simpler. A radial clearance of 2 to 3 mm, according to the size of the cylinders,
may be used, so that no consideration of bearing action is necessary. Only the
piston rings project beyond the piston surface and are in contact with the cylinder
wall. The tail rod can either have a stationary bearing behind the stuffing box or
be carried on a crosshead. The great length of the piston rod between the bearings
caused by the long piston, makes a light cast steel construction and a rod of large
diameter a necessity. From a thermal point of view the floating piston is to be
preferred, since only the piston rings transmit heat from the hot to the cold por-
tions of the cylinder, while in the case of the self-supporting piston the large bearing
surface also takes part in this action. When using a stationary bearing for the
tail rod behind the stuffing box there will be a rise and fall of the piston at every
stroke, which must be considered. If the cast steel used is very soft, then the
cast iron rings are liable to seize.
The long and heavy piston together with the long span of the rod usually
bring about a condition which is a mean between the floating and self-supporting
94
constructions. Part of the weight is then carried by the cylinder wall and part
by the piston rod. The wear of the cylinder tends to alter the weight distribution
in such a way as to increase the part carried by the piston rod and thus relieves
the piston and cylinder bearing surfaces. The piston rod also resists possible
forces acting on the outside of the piston. If for instance, the piston rests in the
cylinder and by reason of leaky rings the steam obtains access to the clearance
space above the piston, then a heavy downward load will result. Floating pistons
offer greater safety against this possibility, this safety being imparted by the
piston rod. This applies especially to vertical engines where the piston, if not guided
by a tail rod, is in a condition of unstable equilibrium and is liable to slap. This
may even be occasioned by the piston overrunning the inlet ports, which should
be fundamentally avoided. Such overrunning can only be permitted in the case
of floating pistons, although even then the possibility of vibration should be reckoned
with. The greater number of mistakes which give rise to lateral forces are incurred
in the arrangement and construction of the piston rings, which latter should pro-
tect the piston surface against such forces. For this reason they should be placed
as far as possible towards the ends of the piston, in order to leave as little area
as possible for the formation of lateral forces. Even if this is done there still
remains some possibility of an unbalanced load, especially with the large surface
of a una-flow piston. If, for instance, the rings of a horizontal engine are not
secured against creeping, their center of gravity, being located eccentrically oppo-
site the joints, will move to the lowest possible position and all the joints will
fall in line at the top. The steam leaking through them will then undoubtedly exert
a heavy pressure upon the large surface, thus forcing the piston downwards and
causing rapid wear. This cannot happen with a floating piston since the pressures
will equalize in the annular space and temporary lateral forces will- be resisted by
the piston rod. The ring joints of floating pistons and of pistons having the rings
on the plunger heads should therefore be equally spaced over the circumference;
for instance, where three rings are used at each end, the joints should be at 120.
In self-supporting pistons without plunger heads, the ring joints should be kept
within the bearing area; thus for three rings one joint may be arranged in the
center and one each at 30 to the right and left. The bearing surface then pro-
tects the ring joints against the steam. With such an arrangement complete tight-
ness may be attained if the workmanship is good and the rings are sufficient in
number. In floating pistons the action of the rings is similar to a labyrinth packing
which always passes a certain amount of steam, since the ring joints can never be
made absolutely tight. The pressure ratio, in case of una-flow engines always being
above the critical value, the weight of steam flowing past the ring joints may be
calculated by means of the formula
in which / denotes the free area of the joint, z the number of joints in series, p^
and v l the absolute pressure and the specific volume of the steam inside the
cylinder. In Fig. 16 are illustrated five different types of ring joint fastenings
which can only partly render the joints tight, but perform the important addi-
95
to
96
tional function of securely locking the rings against creeping. In a una-flow piston
where the rings are usually mounted on the piston heads which expand conside-
rably, the locking elements must in no case project to the cylinder wall.
The slot at the ring joints must be from 2 mm to 5 mm wide, according to
the diameter of the rings, in order to allow for expansion. Friction will cause
the rings to assume a higher temperature, especially with superheated steam and
poor lubrication. If the clearance provided is insufficient, the joints will close
and the rings expand against the cylinder wall, so that increased heating, greater
expansion and a heavier pressure result, which may lead to a complete destruction
of the cylinder surface and rings.
Dimensions of Concentric Cast Iron Piston Rings in mm.
Cylinder Bore]
Piston Ring
Length of
piece cut out
Thickness
Width
300
12
12
24 '
400
15
15
35
600
20
20
GO
800
22
22
84
1000
25
25
108
1200
28
28
132
1400
30
30
155
1600
32
32
180
1800
34
34
205
2000
36
36
230
The different phases in the manufacture of ordinary piston rings are pre-
sented in Fig. 17. First is shown the rough casting of sufficient width to hold it
in the lathe, rough turning and cutting off follow; a piece is then cut out and the
ring closed up for finish turning to the correct diameter. It is not possible to
obtain a uniformly distributed pressure with concentric rings, and only a very
rough approximation to this condition can be reached with excentric rings whose
thickness increases from the joint to the opposite side. The concentric type,
however, is preferable in order to avoid a one-sided center of gravity and a large
clearance in the groove behind the ring.
Attention may also be drawn to what are known as hammered rings. These
are made of Swedish iron, turned to correct diameter and width in one operation,
are then split, and afterwards given the required tension by hammering. Approxi-
mately uniform pressure distribution, greater strength and reliability in operation
are obtainable with this form. Even in small sizes they may be sprung sufficiently
to be slipped into the piston.
Mention should also be made of the piston rings designed by Schmeck (Fig. 18),
which are made in sections whose joints are secured by spring-loaded plugs which
prevent them from creeping and force them against the cylinder wall. Rings of
this type have the advantage that they are practically tight, especially at the
joints, that their bearing pressure is uniform, they can be easily assembled and
disassembled, and adapt themselves to warped cylinders. The cast iron plugs
97
and ring sections are finished in such a way as to fit the cylinder bore. This type
of ring is much used by the Hannoversche Maschinenfabrik vorm. Egestorff and
is reported to have given complete satisfaction.
n.
V,06S'-
i
ro
V5
-19 60 f-
Fig. 17.
It is advisable not to permit the outer ring to overrun the cylinder bore. Such
overrunning exposes part of the ring surface to the steam pressure, thus causing
the ring to collapse and destroying its function of tightness for at least a certain
distance near the dead center. An ordinary cast iron ring, especially if the de-
Fig. 18.
flections are large, will not withstand the stresses produced thereby for any length
of time. The clearance behind the rings should be made as small as possible, so
as to reduce the deflection and leave as little space as possible for the accumula-
Stump/, The una-flow steam engine. 7
98
tion of steam which would press the ring against the cylinder wall and cause con-
siderable wear both on the ring and cylinder, especially in the middle of the latter.
With superheat and dirt in the steam a ring may under these conditions lose as
much as 5 to 10 mm in thickness in a few weeks, and the cylinder bore several
millimeters, especially in the center. A wide ring will be the most subject to this
destructive action. Small width, high grade
material, no overrunning, small clearance
behind the rings, well secured joints, and a
good fit in the grooves are therefore advisable.
If the cylinder is made of a fairly hard,
close-grained cast iron, then no ridges will
form, thus eliminating any reason for al-
lowing the rings to overrun.
All the foregoing is supported by the
experience which the author gained with
a piston packing of the type shown in
Fig. 19. Both rings overran the cylinder
bore. The rings, in collapsing, had to push
the spring along their sloping surfaces, with the effect that both rings and spring
went to pieces. Fig. 20 shows the pieces which the author found in the cavity of the
piston. The spots where wear occurred on the springs are clearly visible in Fig. 2Q.
The number of
rings used should
vary according to the
pressure range. The
Fig. 19.
the rings at the ends
of a una-flow piston,
with both sets active
during the first part of
the stroke, fulfills this,
requirement in the
best possible manner.
A una-flow piston
should always be ma-
de in two or three
parts, held together
by piston rod and nut.
The castings in this
case are simple, light
and without core plugs, which will be especially appreciated in the case of cast steel.
.It is advisable to test the piston with water pressure if the foundry cannot be relied
upon. A cast steel piston of two-piece construction is shown in Fig. 5, page 77.
Each half carries a bronze shoe fastened to it with copper rivets, covering an angle
of 120. The two halves, fitted with three rings each, have radial clearance over their
whole circumference and are made as light as possible in order to reduce inertia.
Fig. 20.
99
The ring joints have a labyrinth effect. Fig. 40, page 157, shows an older
piston of one-piece design having grooves fitted with Allan metal rings. The
latter are made to project about 1 mm above the surface of the piston when new,
and during the first period of operation part of this metal is transferred to and
fills out the pores of the cylinder surface. Both these rings should be placed in
the middle of the piston.
Una-flow pistons for locomotives are always made in three pieces, i. e., two
heads carrying the piston rings, having clearance all around, and a center sup-
porting piece. It was formerly customary to make the latter of hard steel, but
Swedish iron is now used with better results. The heads which are made of cast
or forged steel, expand considerably under the action of superheated steam. This
expansion is transmitted to the center piece and must be considered in designing
the latter. In order to reduce the weight of the reciprocating parts to a minimum,
Fig. 21.
locomotive pistons are always made without tail rods and have given satisfaction
except for minor troubles. These have been due to errors in the composition of
the material and to disregard of the effect of the expansion of the heads upon
the center supporting piece. The experience obtained with such pistons, as well
as the favorable observations of Sulzer Bros., indicate that the problem may be
solved with self-supporting pistons, provided the important requirement of a
reliable lubrication system is satisfied. One oil feed on top, and one each in or below
the horizontal plane on both sides of the cylinder, every feed being connected to
a separate force feed pump, will give satisfactory results with the proper kind of
oil. The pump plungers should be timed to deliver oil only during the periods
when the corresponding orifices are covered by the piston. If the feeds are con-
nected to the cylinder ends, carbonization of the oil is to be feared; and if the oil
is fed to the center of the cylinder at the time of exhaust, it is likely to be blown
through the ports. Both these conditions lead to a high oil consumption which is
sometimes complained of in connection with una-flow engines. The time during
7*
100
Q
which the piston covers the oil
feeds is longest when the latter
are in the middle of the cylinder,
so that even with pumps having
a continuous feed there is a reaso-
nable certainty of oil being car-
ried between the rubbing surfaces.
In order to avoid losses due to
the exhaust, it is permissible to
arrange the three feeds close to
one side of the exhaust belt in-
stead of in the center. In regard
to lubrication also, the floating
piston offers greater safety, since,
if correctly constructed, the rub-
bing surface is limited to the
rings. If, however, the steam
obtains access to the spaces behind
the rings, heavy friction and large
oil consumption may result. With
self-supporting pistons the sanje
effect may be caused by lateral
forces.
Everything considered, it
may be stated that the self-
supporting piston requires greater
care in regard to design, choice
of material, lubrication and ope-
ration, but has the advantage of
not requiring a tail rod with its
bearing or crosshead. The floating
piston has greater reliability but
is more complicated and increases
the floor space required. Self-
supporting pistons may however
be made to operate satisfactorily.
Piston Rod Packings.
Soft packing is used only in
small and cheap engines, while
metallic packing is the rule with
steam of high pressure and
superheat.
The Lentz packing shown in Fig. 21 consists of a plurality of one-piece cast
iron rings, whose number varies according to the pressure to be carried. These
0, 3
.S o>
02
6X3
s
101
rings are fitted to the rod and work between the ground surfaces of a corresponding
number of housings, thus producing a labyrinth effect. The individual housings
form metal to metal joints and are pressed against the bottom of the packing
space by means of the outside gland, sufficient clearance being provided to allow
the cast iron rings to move laterally. The last chamber collects the water of con-
densation so that it may be drained away. With pure steam, absence of dust,
and good lubrication (the oil being preferably forced into the packing under pres-
sure), satisfactory results should be permanently obtainable. The one-piece rings
may sometimes be found unhandy in assembling.
The American type of packing shown in Fig. 22 is better in this respect,
since it may easily be assembled and disassembled and offers greater freedom
in the design of adjacent parts. Each ring is made in four pieces, the two oppo-
site segments with their babbitt lining being pressed against the piston rod by
means of springs. The second ring of similar construction is set at 90 to the first
one. The two remaining segments of each ring are forced by springs against the
other segments. Botli rings are properly fitted to the housing at their joints and
relatively to each other. The whole packing system is held by means of axial
springs against a spherical seat, thereby accommodating itself to inclined positions
of the piston rod, while any lateral movement of the same is provided for by the
sliding fit of the rings in their housings. This packing is occasionally stated to
be unsatisfactory for vacuum, although this criticism may be unjust. The Duplex
packing shown in Fig. 23 which is equipped with an additional set of conical bab-
bitt rings, is equally satisfactory for both vacuum and high pressure. Different
102
Fig. 24.
f C Ct f7
Fig. 25.
Fig. 26.
103
springs are supplied for various pressures. Good workmanship is claimed for this-
packing. Pure steam, regular and ample lubrication, as well as frequent use of
the drain cocks, especially in running in, are essential for success.
Packings of a similar type of Simplex and Duplex construction are offered
by the firm of Max Dreyer & Co., of Magdeburg.
The Proell packing (Fig. 24) is based
on a similar principle. Each cast iron
ring is cut into six parts which "are held
together by means of a coil spring, the
joints of one ring being staggered in relation
to those of the adjacent one. A pair of
such rings is contained in each housing,
which is easily removable by means of an
internal lip. These housings come together
on metal to metal joints and thus form
chambers in which the rings have suffi-
cient play to enable them to move late-
rally. The whole packing is held together
and against the bottom of the packing
space by the outside gland. The oil is
either fed onto the piston rod or forced
between the middle rings. This packing,
however, does not accommodale itself to
inclined positions of the piston rod. This
packing is furnished in Simplex, Duplex
or Triplex forms, according to the number
of rings employed.
The Kranz packing (Fig. 25) made
by the Elementenwerk Kranz, of Ludwigs-
hafen, also employs cast iron rings in
pairs, each cut into three parts, sur-
rounded by sectional primary housings
held together by coil springs. These pri-
mary housings are radially movable between
the ground surfaces of the secondary hou-
sings, provision being also made for axial
expansion of the former. Oil may be fed
under pressure to the packing, although
it is claimed that a drip feed to the rod
its satisfactory. The use of three pairs of
rings with a large number of cavities
results in a thorough labyrinth effect. The
water of condensation is caught by a further
pair of rings on the outside, so that it
may be drained away.
104
The packing designed by Wilh. Schmidt (Figs. 26 and 27) takes care of in-
clined positions of the piston rod in the best possible manner. The packing as
a whole is inserted between two rings having spherical surfaces with a common
center, these in turn being held between two flat surfaces so that both lateral
and rotative movements are rendered possible. A deep recess insures further
flexibility as well as a cooling effect. The segmental babbitt rings of conical sec-
tion are held together by a powerful spring. A special fitting containing a felt
ring and having an oil connection serves to lubricate the rod as well as to keep
out dirt. This packing has been very successful on locomotives.
105
6. Losses due to Radiation and Convection.
It will be seen from a comparison of the una-flow with the ordinary com-
pound counterflow. engine that the former can have only small radiation and
convection losses. The radiating surface of the counterflow engine with its two
cylinders, receiver and accessories is two or three times as large as that of the
una-flow engine, with correspondingly higher losses. The loss due to radiation
of the una-flow cylinder is very small compared with the radiation losses of the
steam pipe, for which reason the latter will be dealt with first. Assuming a flow
of superheated steam at a very small velocity through a pipe having a length of
100 m, then the steam at the far end will have a lower temperature and a cor-
respondingly larger specific weight, (y x : v 2 = T l : T z ) but practically the same
pressure. The highest point E in the temperature-entropy diagram shown in
Fig. 2, chapter I, 3 a, corresponds to the state of the steam at the entrance of
the pipe, while the lower point Q on the same pressure line represents the state
of the steam at the far end. The narrow vertical strip EFWQE below the part
EQ of the pressure line, extending down to the line of zero temperature ( 273),
represents the total amount of heat lost; but as the heat represented by the area
of the diagram below the back pressure line cannot be utilized, the actual radia-
tion loss is represented by the strip ER VQE. Insulating and lagging the pipe
can therefore only result in a mere reduction of this loss. A further radiation
loss occurs in the cylinder and will show itself in a very slight deviation to the
left of the vertical adiabatic line. Insulating the cylinder can therefore only tend
to reduce this slight loss. Much more important is sufficient insulation around
the cylinder heads, which really form part of the steam pipe. The live steam pipe
in high grade plants is always covered while the cylinder is provided not only with
insulation, but lagging as well. The latter supplements the effect of the insulating
material in an efficient manner and may, if constructed of several casings one
within another, with a bright inner surface, entirely take its place.- All flanges
should also be covered. The materials used for insulating steam pipes and cylin-
ders are Kieselguhr, asbestos, magnesia, cork, or glass or textile waste. The more
porous the material is, and the thicker the layer, the better will be the insulating
effect. Part of the heat is lost by radiation and part by convection. The process
is so complicated that mathematical treatment fails completely and actual tests
have to be relied upon. The following table will help to clear up matters.
106
Maximum drop in temperature per 1 m length of pipe, for 14 at. gage;
Outside
steam pipe insulated, without lagging.
diameter
Thickness of
300 "C
350C
400C
of steam
Insulation
steam temperature
steam temperature
steam temperature
80
60
40
80
60
40
80
60
40
mm
mm
m/sec
m/sec
m/sec
m/sec
m/sec
m/sec
m/sec
m/sec
m/sec
60
0.08
0.11
0.14
0.10
0.14
0.18
0.13
0.19
0.25
108
80
0.06
0.09
0.12
0.08
0.12
0.16
0.11
0.17
0.22
100
0.05
0.08
0.11
0.07
0.11
0.15
0.10
0.16
0.20
60
0.06
0.09
0.11
0.08
0.12
0.15
0.10
0.16
0.20
133
80
0.05
0.08
0.10
0.07
0.10
0.13
0.09
0.13
0.17
100
0.045
0.07
0.09
0.06
0.09
0.11
0.08
0.12
0.15
60
0.045
0.075
0.09
0.06
0.09
0.12
0.08
0.12
0.16
159
80
0.04
0.065
0.08
0.055
0.08
0.11
0.07
0.11
0.14
100
0.035
0.055
0.07
0.05
0.075
0.10
0.065
0.10
0.12
60
0.04
0.06
0.08
0.05
0.075
0.105
0.07
0.105
0.13
191
80
0.035
0.05
0.07
0.045
0.07
0.09
0.065
0.95
0.12
100
0.03
0.045
0.06
0.04
0.06
0.08
0.055
0.85
0.11
60
0.033
0.05
0.065
0.045
0.07
0.09
0.06
0.09
0.12
216
80
0.03
0.045
0.06
0.04
0.06
0.08
0.055
0.08
0.11
100
0.025
0.04
0.05
0.035
0.055
0.07
0.05
0.075
0.10
60
0.03
0.045
0.06
0.04
0.06
0.08
0.055
0.085
0.11
241
80
0.025
0.04
0.05
0.035
0.055
0.07
0.05
0.075
0.10
100
0.023
0.035
0.045
0.03
0.05
0.06
0.045
0.07
0.09
60
0.028
0.042
0.055
0.038
0.057
0.075
0.053
0.08
0.10
267
80
0.023
0.035
0.046
0.033
0.05
0.065
0.045
0.07
0.09
100
0.02
0.03
0.04
0.028
0.042
0.055
0.04
0.06
0.08
60
0.025
0.038
0.05
0.033
0.05
0.065
0.045
0.07
0.09
292
80
0.023
0.034
0.042
0.03
0.045
0.060
0.04
0.06
0.08
100
0.020
0.028
0.036
0.025
0.044
0.050
0.035
0.053
0.07
60
0.022
0.033
0.043
0.03
0.045
0.06
0.042
0.06
O.OS3
318
80
0.020
0.030
0.028
0.026
0.039
0.052
0.038
0.056
0.075
100
0.018
0.027
0.034
0.023
0.034
0.045
0.034
0.05
0.068
60
0.02
0.03
0.04
0.027
0.041
0.055
0.037
0.058
0.073
343
80
0.018
0.027
0.035
0.024
0.036
0.048
0.033
0.05
0.065
100
0.015
0.023
0.03
0.022
0.032
0.041
0.029
0.043
0.057
60
0.018
0.028
0.037
0,025
0.038
0.055
0.035
0.053
0.07
368
80
0.016
0.025
0.032
0.023
0.034
0.045
0.031
0.047
0.061
100
0.014
0.022
0.028
0.021
0.031
0.04
0.027
0.042
0.054
60
0.016
0.024
0.032
0.023
0.034
0,046
0.031
0.047
0.062
394
80
0.014
0.021
0.028
0.02
0.03
0.04
0.027
0.041
0.055
100
0.012
0.018
0.024
0.017
0.026
0.034
0.024
0.036
0.048
60
0.015
0.023
0.031
0.022
0.033
0.044
0.03
0.045
0.060
420
80
0.0135
0.020
0.027
0.010
0.028
0.037
0.026
0.039
0.52
100
0.012
0.018
0.024
0.016
0.024
0.032
0.023
0.035
0.44
107
It will be observed from this table that the heat loss for average thickness
of insulating material is inversely proportional to the steam velocity. Higher velo-
cities of course result in smaller diameter, circumference and surface of the steam
pipe, thus also reducing the heat loss. Higher velocities are therefore advisable
up to the point where the throttling losses become excessive (See chapter I, 3 a,
especially Fig. 2). Heavy insulation (80 to 100 mm thick) is to be recommended.
Assuming an outer temperature of 0., then the heat losses increase faster than
the temperature gradient, according to the law of Stephan Boltzmann.
If the steam cylinder is regarded as a pipe, the steam may be considered to
flow through it with a velocity equal to the mean piston speed. Applying the above
rule of the inverse variation of the heat loss with the steam velocity, it will be
found that in view of the lower temperature the radiation losses of the cylinder
must be extremely small, especially as the oil film and the lagging form part of
the insulation. These losses in fact are so small as to appear negligible in com-
parison with the other losses.
108
7. Losses due to incomplete Expansion.
A loss -of diagram area within the limits of the piston stroke is caused by the
fact that the exhaust begins with a certain exhaust lead or distance / before dead
center (Fig. 1), and this loss increases as the exhaust lead f v and terminal expan-
sion, pressure p e increase. For small exhaust lead and low terminal pressure this
loss is negligible. Even in non-condensing una-flow engines in which a large ex-
haust lead is used in order to soften the exhaust puffs, the loss of diagram area
within the limits of the piston stroke is insignificant when compared with the
lost work represented by the toe of the diagram, shown shaded at D. The pro-
blem of finding a means of utilizing this work without increasing the cylinder
dimensions is well worth while. The solution to be described later has the effect
of reducing the pressure p u at which compression begins, with a consequent lower
Fig. 1.
terminal pressure. A smaller clearance volume may therefore be used, thus dimi-
nishing the volume loss. Without going into calculations, it will be seen that the
gain F at the compression line will be approximately proportional to the shaded
area D, or in other words, the higher the terminal expansion pressure is, or the
longer the cut-off, the lower will be the terminal compression pressure. This implies
an increasing pressure difference during compression for an increasing pressure
difference during expansion, a combination which has been proved desirable in
the chapter on volume loss. This rule would be fulfilled in its entirety if the pres-
sure changes on both sides, i. e. expansion and compression, were equal. Further-
more, the use of a longer exhaust lead / would then become permissible, since the
lost area within the limits of the piston stroke now forms part of the toe of the
diagram, and co-operates in lowering the back pressure at the time compression
begins. There can therefore be no objection to making f v large, since by increasing
the duration of the exhaust, the compression is shortened and the exhaust puffs
are softened. If this is done, the number of the exhaust ports will be so far reduced
that the exhaust belt may eventually be dispensed with and only one port remains
which connects directly to the exhaust pipe. Piston and cylinder also become
109
considerably shorter. The relation between length of cylinder L, length of piston
4, exhaust lead /, stroke /, and exhaust port diameter d, is as follows:
J t = d + I If,
L = 2
For instance, for a stroke of 660 mm, an exhaust lead of 25% and exhaust
ports of 120 mm diameter, the piston length will be 450 mm, and the length of
the cylinder 1100 mm as compared with a piston length of 594 mm and a cylinder
length of 1254 mm for 10% exhaust lead. The distance f v is limited by the maxi-
mum cut-off, since direct exhaust of live steam must be avoided. For locomotives
the value of f v must be limited to 25%, while for locomobiles it may be taken as
large as 30 to 35%.
The utilization of the energy represented in the toe of the diagram is based
upon its complete conversion into kinetic energy by means of conical nozzles
such as are commonly used in steam turbines. Each exhaust puff would therefore
act as a kind of wad or plug moving with a high velocity through the exhaust
ru*
Fig. 2 (Diise = Nozzle).
pipe and finally creating behind itself a partial vacuum whose absolute pressure
is p u . The exhaust pipe must therefore be long enough so that there will always
be at least one such plug moving within it, thus preventing atmospheric pressure
from reaching the nozzle and destroying the vacuum. The end of the exhaust
pipe must form a diffusor to change the kinetic energy into pressure energy at
atmospheric pressure.
Fig. 2 shows such an exhaust pipe diagrammatically. The work corresponding
to the shaded areas D and E (Fig. 1) is determined for various pressures p u . The
weights of steam G e and G u , corresponding to the pressure p e and p u can be found
from the diagram and the dimensions of the engine.
Using the equation of work:
G f G,, w 2
The velocities w are calculated for various values of the pressure p u and plotted
against the latter, as shown in Fig. 3. Each value of p u has associated with it a
certain velocity w d which must exist at the point of entrance into the diffusor in
order that atmospheric pressure may be overcome. This velocity therefore cor-
responds to a pressure difference (1 p u ) and may be easily obtained from the
Mollier chart and also plotted in Fig. 3. Of course the steam when leaving the
diffusor must in practice still have a certain velocity, and the pressure difference
should therefore be reckoned not from the atmosphere, but from a slightly higher
pressure corresponding to this velocity. This shifts the diffusor velocity curve
110
into a somewhat higher position, and the intersection S of this curve with the
nozzle velocity curve, which determines the obtainable pressure p u , is moved
slightly towards the right corre-
sponding to a higher value of p u .
Friction losses in the long exhaust
pipe cause a further loss of velo-
city >!, between the nozzle and
-^>. ~ the diffusor, and this again re-
x ^^; ^i fi t " suits in a shift of the point of
r\^4Jlo7~"" intersection to the right, cor-
responding to a still higher pres-
sure p u . The ejector effect of
the exhaust puffs will eventually
become less and less for higher
steam velocities. This will be the
Fig. 3. (Duse = Nozzle). case especially in single cylinder
engines, because the length of
exhaust pipe required is very great. For instance in an engine running at 180 r. p. m.
which corresponds to 6 exhaust puffs per second, and for a steam velocity o-
540 m/sec, the length of the exhaust pipe must be 90 m, or better 100 m. Calf
culating the loss of pressure required to overcome the resistances by means, of
Eberle's equation: ,
t.o
in which A p represents the loss of pressure in kg/sqm, I is the length of the ex-
haust pipe = 100 m, d its diameter = 0,1 m, w the steam velocity = 540 m/sec,
y the specific weight = 0.58 kg/cbm and $ a constant = 10.5 x 10~ 4 , then the
loss of pressure will be found to be 11.1 at. For d = 0.05 m the loss would be
35.5 at. However, Eberle's tests from which the above formula was obtained,
covered only velocities up to 150 m/sec, as did similar tests by Fritsche, Ombeck,
Lorenz and Ritschel, so that the results are not directly applicable to velocities
higher than the critical value. Such higher velocities will result in still greater
losses. In reality the above results will be smaller since the assumed steam
velocity will not be constant but decreasing.
It would seem from the foregoing that the direct exhaust ejector principle
would not have great prospects if based on complete conversion of pressure energy
into velocity. However, as reported by Giildner 2 ), partial vacua have been observed
in long exhaust pipes of gas engines, although no special provision had been made
to cause and sustain them and a great part of the energy was lost in valves, sharp
edges and elbows. There must therefore still remain a possibility of solving this
problem in another way. It is, however, not easy of solution, since it involves
the calculation of the friction of accelerating and expanding steam, which is very
difficult to express in a mathematical form. Actual tests must therefore be relied
upon.
x ) Stodola, ,,Die Dampfturbinen". 3d. Edition, p. 55.
2 ) Giildner, ,,Entwerfen von Verbrennungskraftmaschinen. 3d. Edition, p. 40.
Ill
Although this problem may seem difficult in connection with a single cylinder,
it is very simple for multi-cylinder engines, of which the locomotive is the chief
representative. Even in an engine having two cranks at right angles, an exhaust
lead of 25% will produce sufficient overlap of the exhaust periods so that the
exhaust of one cylinder begins before the other has ceased (Fig. 5). If now the
exhaust pipes are joined at an acute angle, a jet ejector action is obtained. This
effect is well known and widely used in locomotive practice, where the combina-
tion of blast pipe and stack also form an ejector, thus producing a partial vacuum
in the smoke box which serves to draw off the flue gases. The theory of the blast
pipe was first developed by Zeuner in his classical treatise on the subject 1 ). An
equation in a somewhat simpler form based on
this theory, is found in v. Ihering's book, ,,Die
Geblase". The development of the formula
for the ratio G 2 : G of the quantities of the
ejected to the ejecting steam is very lengthy
and will not be repeated here. The formula is
based upon the principle of the continuity of
flow and includes a number of assumptions
and simplifications, the most important of
which are that friction losses are not con-
sidered, but the unfavorable assumption is
made instead that the velocity w z is entirely
lost and w 3 = 0. (See Fig. 4.) The following
formula then results:
m
m
m ^ n 2
2
I
u.
in which
72
The efficiency of
upon the losses due
as the impact losses
the two streams mix.
regarded as absolutely
ejector
friction
depends
as well
an
to
at the point where
The impact is to be
inelastic. Assuming
*ft
Fig. 4.
F sectional areas w velocities
p pressures / specific weight
These symbols used without index refer to
the junction of the nozzles and the
steam at that point;
Index 2 refers to the variable port opening
at the cylinder;
Index 1, to the throat of the blast pipe and
the mixture;
Index 0, to the final section of the blast pipe;
w 3 = the velocity of the ejected steam at the
junction.
W
= then the efficiency r] s = i
Since G 2 is small compared with
G+G 2
the ejecting stream loses little of its energy, the impact loss is small and
Tis = 0.75 to 0.80. The efficiency of the blast pipe is considerably lower, being
only about 0.28, because the weight of the ejected air is about 2.6 times the weight
of the ejecting steam. In addition to this there is a further loss equal to the energy
contained in the steam at the final section of the stack. A similar loss will occur
at the end of the blast pipe in regard to the exhaust ejector action, since the
energy contained in the steam at the final section P\ cannot be utilized for steam
ejecting. This finds its expression inequation 1, where A increases with decreasing
*) Zeuner, ; ,Das Lokomotiven-Blasrohr". 1863. Zurich. Meyer-Zeller.
112
F . The area F should therefore be made as large as possible, but is limited by
considerations in regard to the blast action on the flue gases. Strahl 1 ) has deve-
loped a formula for the blast pipe area, which is based on Zeuner's treatise and is
a- R
as follows : F = -y=, where F is the blast pipe area, F^ the smallest stack area,
y xA
R the grate area, A the coefficient of divergence of the stack, x the coefficient of
flue gas friction from ash pan to stack; a = f (m); m = A F : : F - a is nearly con-
stant, and S 0.03 for m = 13 to 19. The weight ratio of the ejected air L to the
ejecting steam D is
L
~D =
A large blast pipe area produces a lower pressure in the cylinder but insuf-
ficient vacuum in the smoke box and therefore unsatisfactory steaming of the
locomotive. It is self-evident that with a given amount of work available in the
toe of the diagram only a certain
total of ejector action for cylinder
and boiler together can be produced,
the distribution of which depends
essentially upon the blast pipe #rea.
In addition to the blast pipe area,
the dimensions of the stack are also
important. Too small a stack area
leads to a large velocity loss at the
outlet, and too large an area causes
a considerable impact in ejecting
the flue gases. It follows from this
that there must be a best stack
area for which, in consequence of
the losses being a minimum, the
blast pipe area is a maximum. Ex-
pressed mathematically, a = f (m)
is a maximum for m = 15.5 approxi-
mately, as is demonstrated in the
above-mentioned paper by Strahl. These best stack dimensions must be
strictly adhered to in a locomotive in which the ejector effect of the exhaust
is utilized, and this leads to considerable difficulties in large locomotives of
the present day. The length of the stack is limited to such an extent by the
loading gage and the high location of the boiler, that the expansion of the jet
leaving the blast pipe may not completely fill out the whole area of the stack,
so that air may enter from above through the remaining area and thus partly
destroy the vacuum. Actual tests, however, have shown that such loss of
vacuum can easily be avoided even with a large stack area.
a ) Strahl, ,,Untersuchung und Berechnung der Blasrohre und Schornsteine an Loko-
motiven, 1912'', Wiesbaden, C. W. Kreidel.
Fig. 5.
113
The further calculations are based on a 010 freight locomotive of the
German State Railways as an example, which is described on page 242. This
engine is a two cylinder superheater locomotive with a cylinder bore of 630 mm,
a stroke of 660 mm, a driving wheel diameter of 1400 mm and a steam pressure of
12 at. gage. Assumed is an evaporation of 7000 kg/hour, a loss of pressure of 1 at.
from boiler to valve, adiabatic expansion and compression at an entropy of 1.7,
and three different speeds of 20, 40 and 60 km/hr, which are referred to below as
cases I, II and III. The exhaust port of the cylinder, having a'diameter of 120 mm,
is placed 7 mm off center to compensate for the angularity of the connecting rod
of 2600 mm length. The diagram shown in Fig. 5 is based on these figures and
the cross-shaded areas represent the periods of overlapping exhaust. The ejector
action is effective only during part of the exhaust
period and therefore only the part A of the shaded
area in Fig. 6 can be utilized for ejector action, the
part B being lost. An increase in the exhaust lead
would of course have the effect of increasing A by
a part or the whole of B; but considerations of
evenness and strength of the draft forbid this. Of
the total work represented by the area A, a part
Fig. 6.
is lost in producing the draft in the smoke box by the rush of the steam at
high velocity from the nozzle (blast nozzle loss) and a part is lost by throttling
and friction in the pipes (pipe loss). Finally, after subtracting the stack loss,
there remains the effective gain of work C at the compression line, corresponding
to an absolute pressure p u .
The investigation begins with the determination of the free exhaust port
areas for various piston positions within the limits of the exhaust lead /. The
secondary nozzle delivering steam to the feed water heater must also be included
in this consideration, since the steam passing through it acts in the same way
as in the main exhaust pipe. Fig. 7 illustrates the profiles of the main and secon-
dary nozzles, as well as the piston positions at crank angle intervals of 5. In
order to find the smallest areas of opening of the nozzles, the outlines of which
are shown shaded in Fig. 7 for 55 before dead center, their plan projections are
first laid out and their areas multiplied by V cos I n calculating the weight of
steam flowing through the nozzles, the velocity loss must be taken into considera-
tion by using a velocity coefficient y = 0.9. On account of the very unfavorable
nozzle profiles when first uncovered, with consequent turbulence, it was consi-
dered necessary to use a smaller value of o
115
and exhaust lines I, II and III shown in Fig. 9 were obtained point by point in
this manner, corresponding to speeds of 20, 40 and 60 km/hour. The part of the
exhaust line where the ejector action comes into effect was also determined point
by point, the weights of steam ejected G 2 being calculated by means of equa-
tion (1).
As long as the pressure in the cylinder is above the critical value p k = 1,73 at.
abs., the pressure energy can only be completely converted into kinetic energy
by the use of conical nozzles; otherwise the jet will still possess some pressure,
and the suction effect will be diminished on account of the lower velocity conse-
quent on the decrease in specific volume. In equation (1) this is taken care of
by an increase in y. The theoretical as well as the actual factors of divergence
i^ and if}' were therefore calculated according to the rules of steam turbine design,
I
I
E
J3mi
6,8*
S.95
2.i
Fig. 9.
and are also tabulated in Fig. 7. According to these figures the actual divergence
is insufficient in case I for crank angles of 50 to 30. A correction was there-
fore made in the calculated values of G 2 . In case II the actual divergence is al-
most, and in Case III exactly correct, so that G 2 need not be corrected. The
compression lines in Fig. 9 were laid out in accordance with these results, and
it is found that the initial compression pressures are respectively 0.7, 0.97 and
1.0 at. abs. for cases I, II and III.
The very small pressure reduction in cases II and III makes it desirable to
analyze the losses during the exhaust period. The blast pipe loss can easily be
calculated since we know the weights of steam exhausted during the time the
piston travels the distances f d and / f , as well as the blast pipe area and the specific
volume corresponding to atmospheric pressure. With these data the mean steam
velocity at the outlet of the blast pipe and the corresponding energy may be
calculated. To be exact, instead of taking the mean velocity w, it would be neces-
8*
116
sary to find the value of the integral w*dd but for a rough approximation this
o
is not necessary. The amount of work represented by the areas A, B and C may
be found with a planimeter, and the efficiency of the ejector is determined by
the ratio of G 2 : G, so that the remainder represents the pipe loss. The major
part of the latter is caused by the throttling of the steam when entering the nozzle.
A high vacuum is formed in the exhaust pipe just before and after beginning of
compression but it only partly reaches the cylinder. This is taken into consideration
in equation (1) by taking a lower value of n. The following table gives the
relative amounts of the different losses.
Case
I
II
III
Area A:
Blast nozzle loss
40
25
14
Pipe Loss
14
30
56
Impact loss during ejection .
Gain of compression
Area B:
Blast nozzle loss
8
27
1
4
15
9
22
Pipe loss
10
17
8
Total . .
100
' 100
100
It will be seen from this table that the blast nozzle loss (part A and B taken
together) is approximately constant and amounts to from 41 to 36%. The pipe
loss increases from 24% at low speeds to 44% at high speeds, and the work re-
presented by area B from 11% to 30% respectively. It therefore follows that
at high speeds the bad effect of area B must be eliminated, and this can be easily
accomplished with a three cylinder locomotive. In such an engine, having cranks
at 120 and an exhaust lead of 25%, there will always be two cylinders exhausting
at the same time. The ejector effect begins at the dead center and proceeds with
greater nozzle areas in the cylinder wall, with the result that the throttling
loss and pipe loss are also reduced. The loss of work may therefore be divided
in all three cases as follows:
Blast nozzle loss 40%
Pipe loss . 24%
Impact loss ....*.... 8%
Useful work. . . ... . . 28%
Assuming this division of losses, and taking the mean effective pressures from
the indicator cards, the following figures were obtained for two and three cylinder
locomotives.
The three cylinder locomotive also shows only a surprisingly slight gain due
to the ejector effect at high speeds or early cut-offs. This is explained by the
fact that the utilization of the toe of the diagram is equivalent to an enlargement
of the cylinder. If the cut-off is early, then a further increase in expansion will
not produce much gain. The steam consumption figures are already so low that
117
I
II
III
Two Cylinder Locomotive.
Mean effective pressure, with ejector effect,
kg/sqcm
Mean effective pressure without ejector
effect kg/5qcm
6.84
6.1
3.93
3.83
2.83
2.83
Gain due to ejector effect /o
12.2
2.6
Indicated HP
940
1080
1170
Steam consumption (7000 kg/hr) divided by IHP.
Three Cylinder Locomotive.
Loss area A -j- B kg/sqcm
7.45
2.7
6.47
0.67
6.2
0.33
Compression gain 0,28 (A-\- B) .... kg/sqcm
Mean effective pressure, with ejector effect kg/sqcm
Gain due to ejector effect /o
0.75
6.85
12.3
0.19
4.02
5.0
0093
2.923
3.0
Indicated HP
940
1110
1200
Steam consumption (7000 kg/hr) divided by IHP.
7.45
6.3
5.8
not much more could be desired. The saving of 12% for late cut-offs is noteworthy,
because it is based on very conservative assumptions. Furthermore, the exhaust ejector
effect allows of a considerable reduction of clearance volume; for instance, in the case
under consideration, from 17 to 11%. Taking into account the saving due to the
una-flow principle a total saving of 15% may be expected with certainty. This saving
is all the more important since it occurs at heavy loads and therefore increases the
hauling power of the locomotive by this amount.
Although in describing the exhaust ejector principle locomotives were con-
sidered exclusively, its field of usefulness is not limited to the latter. It may also
be found advantageous in street railway locomotives, road rollers and stationary
engines, as well as locomobiles which are still frequently built with two cylinders
so that they may be started from any crank position, which is desirable for instance
in peat pressing plants.
118
8. Prof. Dr. Nagel's Experiments.
A series of extremely interesting experiments on the temperature conditions
in a una-flow cylinder were conducted by Prof. Nagel in the engineering labora-
tory of the Technische Hochschule, Dresden, and described by him in the Zeit-
schrift des Vereines deutscher Ingenieure Vol. 1913, No. 27, July 5. Part of his
report is as follows:
"After Prof. Stumpf had published his first communications on the
character and success of the una-flow steam engine a number of years ago,
119
Dr. Mollier and the author approached the Verein deutscher Ingenieure with
the request for an appropriation for investigating the temperature conditions
in a una-flow cylinder. This request was granted in the most whole-hearted
manner. At the same time the Saxon Government provided considerable
sums for the completion of the testing plant. The una-flow cylinder used for
this purpose in the engineering laboratory of the Technische Hochschule,
Dresden, was built by the Ntirnberg Works of the Maschinenfabrik Augsburg-
Ntirnberg, and took the place of the low pressure cylinder of the existing triple-
expansion engine. It was put into operation during September 1911. The
cylinder as shown in Fig. 1, has a bore of 450 mm and a stroke of 650 mm,
and the engine runs at 150 r. p. m.
In order to determine the thermal peculiarities of the Stumpf cycle it
was planned to measure the temperature changes of the working steam at
different points in the cylinder. It was later also found desirable to measure
ffo/ben
OecM
Fig. 2. (Kolben = piston; Deckel = cover).
the temperature variation of the cylinder wall. The determination of the steam
temperatures offered great difficulties. It was at first attempted to use thermo-
couples of copper and constantan wire of 0.2 mm diameter. Tests of a similar
nature on a counterflow engine in the laboratory, using the same elements,
had been started five years ago, but a critical examination showed that their
sensitiveness is by no means sufficient to follow the changes of temperature
with the necessary speed. After long and futile experiments with thermo-
couples composed of thinner wires down to 0.07 mm diameter, it was found,
according to a test report published in an American periodical, that wires of
so small a diameter as 0.01 mm were required for the thermocouple to be suffi-
ciently sensitive. It seemed impossible to produce thermocouples with wires
of this thinness on account of the difficulty in making the junction, and for
this reason the use of electric resistance thermometers was decided upon early
in 1912. The material for the latter was obtained in the form of drawn tungsten
filaments as used in electric lamps. A wire of about 50 mm length was wound
in zigzags upon a glass frame provided with platinum hooks for this purpose,
as shown in Fig. 2. The main difficulty was a satisfactory connection of the
resistance unit to the lead wires in order to enable the termometers so. con-
120
structed to resist the effect of the steam currents inside the cylinder. The
measurement of wall temperatures was rendered difficult by the fact that the
insertion of the measuring unit into the cylinder wall necessitates the drilling
of a hole which more or less disturbs the heat flow. This may be the cause
of an erroneous temperature indication. In order to reduce this possibility to
the utmost, the arrangement illustrated in Fig. 3 to 5 was employed. A hole
of 15 mm diameter was drilled in the cylinder wall, into which was closely fitted
a cast iron plug having a hole of 9 mm diameter bored to within 0.5 mm of
the bottom. Into this hole was fitted a second cast iron plug having two drilled
holes of 2 mm diameter from end to end. These holes contained the copper
and constantan wires of 0.1 mm diameter insulated by small glass tubes, their
Fig. 5.
ends being embedded in grooves at the bottom surface of the cast iron plug.
A thermocouple of similar construction was also fitted to the piston, as indi-
cated at ft, in Fig. 1, and in Figs. 6 and 7. The leads of this thermocouple
were carried through the hollow tail rod, provided with a porcelain lining for
this purpose.
For measuring the change in voltage corresponding to the changes in
temperature, a galvanometer made by Edelmann in Munich was employed.
According to Fig. 8, it consists of a powerful electromagnet which is supplied
with current from a storage battery. In the magnetic field is stretched a fila-
ment of gold or platinum having a diameter of from 0.002 to 0.005 mm, which
carries the current to be measured. The displacement of this wire due to electro-
121
magnetic forces is a measure of the current flowing through the circuit, and
therefore also a measure of the temperature. This displacement, although
amounting to only a fraction of a millimeter, is projected on an enlarged scale
onto the focal plane of a camera by means of a beam of light from a source L
and a system of microscope lenses. The photographic plate is moved pro-
portionately to the piston travel or crank angle behind a slit in the focal plane,
Fig. 6 u. 7.
thus producing a photographic record of the changes of temperature with
stroke or time. Special methods were devised for rapid calibration of the dis-
placement of the filament. In Figs. 9 and 10 are reproduced two records of
steam and wall temperature based on time. The remarkable 'feature about
Bilet-
eberie
Fig. 8. (Bildebene = Focal plane.)
the temperature change of the working steam is the fact that at the end of
compression the latter attains temperatures of such a magnitude as were
hitherto thought impossible. A terminal compression temperature of about
500 was observed when running with saturated steam of 10 at. gage. The
wall temperature was measured at the points a, 6, c, d, e, /, g and /c, indicated
in Fig. 1. The change of temperature at the end of the cylinder barrel at point
b is of considerable significance. A series of tests made with constant cut-off
122
Temperature time diagram of steam close to cylinder head surface. (Point of measurement a.)
tintauchttqfe Q
1 Umdrehung
Fig. 9.
Temperature time diagram of cylinder wall at a depth of 0,5 mm from inside surface.
(Point of measurement /.)
Fig. 10.
of 10% and saturated as well as superheated steam of different temperatures
showed that the highest mean temperature at this point was reached when
operating with saturated steam; even superheated steam of 350 did not produce
the same high wall temperature. The sensitiveness of the thermocouples was
raised to such a degree that the passing of every piston ring over a point of
measurement produced a clearly discernible wave. The moment of passage
of the several rings over the thermocouple is clearly indicated in Fig. 10
by the shading between the various lines; the diagram was taken at
point d"
The above report also includes descriptions of the several instruments which
were used during the tests and for the analysis of their results. Among others,
there are mentioned a harmonic analyser by Mader, an instrument fitted with
a microscope for measuring indicator cards, made by H. Maihak, of Hamburg,
and an apparatus furnished by Steinmuller, of Gummersbach, for automatically
measuring the condensate, which proved to be very exact.
The temperature diagram in the above report by Prof. Nagel merits parti-
cular attention. Instead of a terminal compression temperature of 500 for 3.3%
clearance, a final temperature of 900 should be obtainable with a clearance volume
of 1%. While on the one hand the terminal compression temperature was 500
for saturated steam, it decreased to 480 or 450 for increasing degrees of super-
heat. This may probably be attributed to the more energetic heating action of
the steam jacket in the case of saturated steam. Prof. Nagel further states that
the temperature at the point b of the cylinder wall, for 12% cut-off and saturated
steam of 184 in the jacket, was 128, which fell to 111 for the same cut-off and
superheated steam of 220; and again slowly rose to 118 for a further increase
of the initial steam temperature to 350. The temperature of the inner cylinder
head surface was 177 for an initial or jacket steam temperature of 184, the total
variation during one revolution being only 0,5. The temperatures at the points
*, c, d, (Fig. 1) were found to be 128, 102 and 83, with a total fluctuation of
3, 3 and 2.8 respectively. At the point k on the piston, distant 36 mm from
the cylinder wall, the temperature was found to be 164.5 with a total fluctuation
of 1.3. Attention is especially called to the latter figures, since they prove the
statements previously made concerning the favorable thermal action of the piston
head surface. At the points of measurement the heat had to penetrate a metal
thickness of 0.65 mm. It will also be noticed in Fig. 9 that a very pronounced kink
occurs in the temperature curve where the steam changes from the saturated to
the superheated state, and also that an abrupt drop in temperature takes place
at the moment of admission, from the high terminal compression temperature of
about 530 to that of the live steam.
The contrary effect of heating by the jacket steam and cooling by the cylinder
steam at the point a is also evident in Fig. 9, as well as the corresponding small
temperature fluctuation at this point during the complete cycle, considered apart
from the sudden rise due to the heat of compression. The comparatively high
mean temperature and small fluctuation at the point d are also noticeable in
Fig. 10.
124
In Fig. 11 is shown an especially clear temperature diagram in which the
kinks in the compression and expansion lines corresponding to the change from
saturated to superheated steam are clearly noticeable.
There is a surprisingly high temperature during the last part of expansion,
the first part of compression and especially during exhaust (about 100),
although the engine was operated with a vacuum of 98%. As this card was
taken at the cylinder side of the cover, the explanation is easily found in the
great flow of heat from the cover to the working steam during that time.
3. C. Cooer sieff .
I0f
J).C. Cr
Fig. 11.
The drop of temperature through the cylinder head wall was found to be
7 to 8 for saturated steam, 15 for steam of 250, and 25 for steam of 350 initial
or jacket temperature.
A close study of the temperature diagrams given in Figs. 9 and 10 and 11
has as its final result a confirmation of the thermal advantages of the una-flow
principle and the jacketing of the heads.
This is still more emphasized by the comparison of the una-flow temperature
diagram (Fig. 11) taken by Prof. Nagel with a counter- flow temperature diagram
taken by E. T. Adams & T. Hall from a common slide valve engine of the
Sibley College-Cornell University, as shown in Fig. 12. The comparison elucidates
the striking thermal difference between both engines. Whereas the una-flow
engine shows the highest temperature at the inlet end and the lowest at the
exhaust end, the counter-flow engine shows quite a thermal mixture distributed
over both strokes. Interesting is the postponement of the phases of high metal
temperature caused by the preceding phases of high steam temperature in the
counter-flow engine. -
125
126
II. 1. The Una-Flow Stationary Engine.
The una-flow engine has found a very wide use as a stationary prime mover
mainly by reason of its simplicity, its straight line construction, its high economy
and its adaptability to changing load requirements. The tandem counterflow engine
which still comes occasionally into competition with it, is at a disadvantage on
account of its two cylinders, its two pistons, its piping and the inaccessibility
of its exhaust valves.
The conditions of close regulation required of stationary engines, in which
may be included engines for electric current generation, are satisfied in the una-
flow engine in the best possible manner since the action of the governor is direct
and is not impeded by steam already contained in the engine, as is the case in
multiple expansion engines where the effect of such steam on the regulation makes
itself unpleasantly noticeable.
The range of cut-off in una-flow' engines is usually from to 25%, although
cut-offs are found up to 40%, or even 50 or 60% for instance in rolling mill engines.
The range of the governor must include zero cut-off, in which case the inlet valve
does not open at all. The lead of the steam valves at all cut-offs must be kept
down to the minimum or reduced to nothing if possible, since large lead causes
condensing engines to knock badly, especially if the clearance is large and the
vacuum high. Non-condensing engines with large clearance and long compression
always run quietly and can therefore stand more lead.
It is easily possible to start engines having 25% maximum cut-off even under
load and with a directly driven air pump, more particularly if an additional clearance
space is provided and opened up during the first strokes until sufficient vacuum
is generated. The best location for these additional clearance pockets is in the
cylinder head opposite the end of the cylinder, so that for condensing service they
will act as a very effective insulation between cylinder head and frame, while pro-
viding double the amount of cover jacket surface for non-condensing operation.
Every condensing stationary una-flow engine should be equipped with addi-
tional clearance spaces in order to facilitate starting if the air pump is directly
driven, and to allow of running the engine without the condenser.
The clearance volume averages about 1,5 to 2% for condensing operation
and high vacuum, and 13 to 28.% if the additional clearance spaces are opened up
for non-condensing service (see page 48).
In non-condensing engines the necessary clearance may be arranged in the
cupped ends of the piston. On account of the large work of compression such
engines require comparatively heavy flywheels.
Since the una-flow engine has only two inlet valves, the use of a lay-shaft
is unnecessary. As is shown in Figs. 1 to 3 of this chapter, and in Figs. 4 and 5
(page 10), the inlet valves may be driven from an eccentric on the crankshaft
127
W/////////////////////////////////'/'///"'
Fig. 1.
Fig. 2.
Fig. 3.
128
acted on by a shaft governor, by means of a rocker arm and cam mechanism (Stumpf
gear). A lay-shaft with its bevel gears and bearings is thus dispensed with. Con-
sideration must be given to the expansion of the cylinder. If the latter is pro-
vided with steam jackets receiving their supply from a connection to the steam
pipe ahead of the main stop valve, then the cylinder may be warmed up prior to
starting, and the valve gear, if set correctly for the hot engine, will give proper
distribution from the very start. If the cylinder is unjacketed, then the steam
distribution will be incorrect for a while after starting until the cylinder has reached
its expanded condition. It is always advisable to design the valve gear with out-
side admission, or in other words to arrange cams and rollers so as to make their
action conform to the steam lap of a slide valve, so that while the cylinder is still
insufficiently heated, the head end valve will open late instead of too early. This
negative lead combined with a simultaneous earlier cut-off will do less harm than
an early opening of the valve, which may cause the engine to knock.
Fig. 4.
There remains another possibility of correcting the bad influence of the expan-
sion of the cylinder even if the latter is not provided with steam jackets, by ex-
tending the cylinder lagging around the rod between the valve bonnets (Fig. 4),
thus heating it to approximately the mean cylinder temperature. It is also possible to
drive the head end valve through an equal-armed rocker mounted at the center line
of the exhaust belt. This insures permanent correct motion for the head end valve.
If a lay-shaft is used, the influence of the cylinder expansion is eliminated,
and the steam distribution must always be correct.
Fig. 5 illustrates details of a valve bonnet as used with the Stumpf gear.
The cam is connected to the valve crosshead, and the reciprocating slide is grooved
to accommodate the roller and at the same time form an oil bath. The guide for
the reciprocating slide is long enough so that the groove never runs beyond it,
the loss of oil by splashing and the entrance of dust thus being prevented. The
oil collecting in the groove is transferred by the roller to the cam, so that perfect
lubrication and reliable operation of these important parts is insured.
129
Fig. 6 shows a twin una-flow engine with Stumpf gear, in which a jack shaft
having two crank throws is driven by a pair of eccentrics on the crank shaft set
at 90. This jack shaft carries a shaft governor acting upon an eccentric on each
side of it, which operates the valve mechanism of its corresponding cylinder through
a rocker arm. In this way the valve gears are positively connected, so that both
of them always give the same cut-off. The short vertical eccentric rod also helps
to equalize the cut-offs of both cylinder ends, and the small diameter of the jack
shaft facilitates the design of the governor. This engine possesses great reserve
power since each half is able to carry the whole load.
The elimination of exhaust valves and
their gear will be found very convenient in
horizontal engines, since it leaves the whole
space underneath the cylinder free for
piping and permits of a close arrangement
of the condenser. (See Figs. 2 to 5, chapter
I, 3b, p. 6970.)
The Erste B runner Maschinenfabrik-
gesellschaft was the first concern to take
up the una-flow engine, and decided to re-
build an old 80 HP. single cylinder con-
densing engine with a forked frame by
fitting it with a una-flow cylinder, designed
by the author, having a bore of 400 mm
and a stroke of 420 mm (Fig. 7). On the
free end of the crank shaft was mounted a
shaft governor acting on the eccentric ope-
rating the inlet valves by means of a rocker
arm on the exhaust belt and a pair of
Lentz cam mechanisms. Although this first
design was susceptible of improvement in
many respects, its economy even with
rather low vacuum was 'equal to that of
a compound engine of the same size.
Fig. 8 shows another engine built by
the same Company.
Shortly after the latter had taken up
this work, the Elsassische Maschinenfabrik
decided on a large scale experiment. Their
first una-flow engine, built to the author's
design, had a cylinder bore of 640 mm and a stroke of 1000 mm, with a
rated load of 500 HP. (Figs. 1 and 9). This engine, which was directly con-
nected to an electric generator, was tested by the Elsassische Verein der Dampf-
kessel-Besitzer (Alsatian Association of Steam Boiler Owners), on February 21,
1909. The result of a trial of four hours and eight minutes duration showed a steam
consumption of 4.6 kg/I HP-hour for an initial steam pressure of 12.6 at. gage
and a temperature of 331 G, at a speed of 121 r. p. m. This is a very creditable
Stumpf, The una-flow steam engine. 9
Fig. 5.
130
131
result if it is borne in mind that the engine did not derive the full benefit from
the vacuum on account of too small an exhaust pipe and the use of an oil separator
between cylinder and condenser. (Back pressure 0,145 at. abs.) By correct design
Fig. 7.
of the condensing equipment in the way previously suggested, by jacketing of the
cylinder, and the use of tighter valves, the steam consumption could be conside-
rably diminished, as proved by later engines built by the same makers.
Fig. 8.
Figs. 2 and 3 show a una-flow engine of 900 HP rated load built by the same
firm. Noteworthy is the heavy frame, the engine being of center crank construc-
9*
132
tion which is now used by several concerns. The center crank type is advantageous
where the forces on the moving parts are heavy, especially for large short stroke
engines.
In Figs. 10 and 11 is shown still another engine by the same makers, as well
as its governor and valve gear parts.
Fig. 9.
A una-flow engine built by Burmeister & Wain, of Copenhagen, and designed
by the author, is shown in Figs. 12 and 13. The direct connection of the con-
denser to the cylinder should be noted, as well as the method of supporting the
Fig. 10.
rear end of the latter on two adjustable rods, and the simple air pump drive. The
cylinder is left unjacketed on account of the use of superheated steam (Fig. 14).
The head jackets, however, are carried up to the point of normal cut-off. The
133
Fig. 11.
Fie. 12.
134
piston is turned to a smaller diameter for a corresponding distance to provide for
expansion. The additional clearance pockets are fitted with two clearance valves,
one one each side of the inlet, one of which is sufficient for starting, while the
Fig. 14.
second one has to be opened for non-condensing operation at full speed (Fig. 15).
The area of contact between cylinder head and frame is kept as small as possible
in order to reduce the conduction of heat to a minimum. The cylinder head
casting is perfectly symmetrical so as to increase its range of usefulness.
135
Fig. 15.
Fig. 16.
136
Fig. 17.
Fig. 18.
137
Tests made with this engine by Mr. Bacher, Professor at the Technical
Hochschule in Copenhagen, showed the following results.
Load
Pressure
kg/sqcm
Steam
Tempe-
rature
C
Vacuum
in /. of
760 mm
r. p.m.
Steam
Con-
sumption
kg/hr.
Steam Consumption
KW
BHP
IHP
KW/nr.
BHP/hr.
IHP/hr.
64.50
98.7
116.0
9.90
352
94.0
179.0
479.0
7.44
4.86
4.12
86.46
130.0
149.0
9.87
354
93.8
175.0
632.0
7.32
4.86
4.24
108.66
163.0
184.5
9.84
353
93.5
176.5
798.4
7.36
4.90
4.34
131.24
197.0
222.0
9.80
353
92.6
173.5
976.7
7.45
4.97
4.40
109.00
164.0
186.0
9.75
dry
saturated
93.0
178.0
1150.6
10.55
7.03
6.20
In view of the omission of the cylinder jackets, these results are in close
agreement with the tests of a 300 HP una-flow engine given on p. 11.
A single-acting una-flow engine built by Burmeister & Wain is shown in
Fig. 16. Engines of this type are widely used in Danish dairies. The horizontal
valve is operated by a cam mechanism directly connected to a shifting eccentric
on the crank shaft. (Figs. 17 and 18.)
A stationary engine built by Ehrhardt & Sehmer, of Saarbriicken, for the
power plant of the Saar Valley Railway, is shown in Fig. 19. This engine has
Fi?. 19.
a cylinder bore of 650 mm, a stroke of 1000 mm, and a rated load of 500 HP at
a speed of 130 r. p. m. It has run for long periods at an overload of nearly 100%.
The average cut-off of una-flow engines at rated load being only about 10%, their
138
139
capacity for overload is
far greater than that of
any other type of engine.
If the dimensions of the
driving parts are based
upon the initial pressure
less the inertia, then even
a heavy overload does
not materially increase
the load upon them.
In Fig. 20 is shown
the largest una-flow en-
gine so far built, con-
structed by Ehrhardt &
Sehmer for driving a rol-
ling mill at the steel
works of Gebrtider Roch-
ling, atVolklingen, having
acylinderboreoflVOOmm
and a stroke of 1400 mm,
the speed being 110 to
130 r. p. m. The cylinder
was made with this large
bore on account of the
low steam pressure avail-
able at the time, and is
provided with two inlet
valves at each end for
this reason. Later on,
when new high pressure
boilers have been instal
led, it is intended to
substitute for the present
cylinder a smaller one
with only one valve at
each end.
Another una - flow
engine built by the same
firm and delivered to the
Aplerbeck steel works,
is illustrated in Figs. 21
and 22. The cylinder bore
is 1450 mm, the stroke
1500 mm, the speed 100
r. p. m., and the steam
pressure 8 at. gage.
140
141
142
143
This engine has nearly the same dimensions of driving parts as the one just de-
scribed. The diameter of the piston rod is 250 mm, that of the tail rod 225 mm,
the crosshead pin is 400 mm diameter by 600 mm long, the crank pin 550 mm
diameter by 600 mm long, and the main bearing 730 mm diameter by 1200 mm
long. The use of a side crank in such a large engine of short stroke is noteworthy.
In order to reduce the overhang, the crank and crank pin are of cast steel in one
piece, with a hub length of only 450 mm (hub length : shaft = 0.62). The side
crank construction, together with its corresponding type of frame, makes the engine
simple and inexpensive. The same cannot perhaps be said of the Zvonicek valve
gear employed on this engine, but it has the advantage of giving the late cut-
offs essential for rolling mill engines, and of permitting the use of a standard governor
which is in many cases preferred to a shaft governor on account of its simplicity
and accessibility. The Zvonicek gear consists of a fixed eccentric, the strap of which
Fig. 25.
is provided with a cam profile and held at its armlike extension under the control of
the governor. The combined motion of eccentric and cam is transmitted to the
valve bonnet cam mechanism by a reach rod provided with a roller at its lower end.
Figs. 23 and 24 illustrate clearly the trend of development due to the una-
flow engine and show the replacement of the two cylinders of an old tandem
counterflow engine by a single una-flow cylinder. A number of such reconstruc-
tions have been carried out by Ehrhardt & Sehmer and other firms.
An engine built by Musgrave & Sons, Ltd. of Globe Iron Works, Bolton,
England, is shown in Fig. 25. This firm is credited with the introduction of the
una-flow engine on a large scale in Great Britain and colonies, the Stumpf valve
gear being employed exclusively. A test carried out by Mr. F. Thomas on one
of their engines, having a cylinder bore of 685.8 mm and a stroke of 914.4 mm
gave the following results : Steam pressure 10.67 at. gage at the throttle, superheat
10, speed 129 r. p. m., vacuum at the cylinder 66 cm, load 317 I HP, and steam
consumption 4,98 kg/I HP-hour. The cylinder barrel was unjacketed.
144
Stork & Co., of Hengelo, Holland, also employ only the Stumpf gear on their
engines, one of which is shown in Fig. 26. This firm has been very successful in in-
troducing the una-flow engine in Holland and the Dutch colonies. Stork & Go.
report that a test of one of their engines (650 mm bore by 900 mm stroke, speed
125 r. p. m.) showed a steam consumption of 4.86 kg/I HP-hour, the steam pressure
being 8.24 at. gage, and the temperature 248 G. The cylinder barrel was un jacketed.
The type of una-flow engine built by the Maschinenfabrik Augsburg- Niirnberg
is shown in Figs. 27 and 28. The inlet valves are placed horizontally and are ope-
rated by a rocking shaft and cam mechanism. This cam receives its motion from
a short jack shaft which in turn is driven by the governor eccentric on the lay
shaft. The clearance valve is situated opposite the steam valve and is arranged
to act automatically in case of sudden failure of the vacuum. Partially or entirely
unbalanced inlet and spring-loaded clearance valves may be made to serve the
same purpose.
One engine of this type (903 mm cylinder bore, 1000 mm stroke) was fur-
nished to J. P. Stieber, at Roth near Niirnberg and was tested by the Bayerische
Revisionsverein on February 23, 1912.
Duration of test hours 4.17 4.05 8.06
Boiler pressure at. gage 13.3 13.3 13.2
Steam temp, at engine 255 291 310
Steam pressure in cylinder at. gage 9.3 10.4 11.4
Vacuum in cylinder % 90 90 90
Vacuum in condenser % 93 92 92
Actual cut-off . ~ % 3 6 12
Speed . r. p. m. 125.5 123.3 126.3
Indicated horse power HP 473 793 1109
Steam consumption in kg/I HP-hour including con-
densate from steam pipe 4.69 4.74 4.71
Heat consumption in Gal/I HP-hour based on total
heat of steam entering the engine 3320 3440 3460
Thermal efficiency % 19 18.4 18.3
This engine was designed for a steam temperature of 330 G for which reason the
cylinder barrel was left unjacketed. With jackets the steam consumption would in
this case have been considerably lower on account of the beneficial effect of cylinder
jackets for small cut-offs and low initial temperatures. On the other hand, the
results once more demonstrate the small variation in steam consumption for large
ranges of load (473 to 1109 HP) when no cylinder jackets are employed.
An engine built by the Gorlitzer Maschinenbauanstalt is shown in Fig. 29.
The cylinder has a bore of 1100 mm, a stroke of 1300 mm, and the engine runs
at 91.6 r. p. m. The valve gear comprises a lay shaft with governor and shifting
eccentric acting on short rocking shafts alongside the cylinder. The ends of the
cylinder are jacketed. The air pump is driven from an extension of the tail rod.
Fig. 30 shows one of the latest engines built by this Company. There is only
one governor eccentric, and the valves are operated through Lentz cam mechanisms
from a rocking shaft on the cylinder, having two levers set at 180. The air pump
145
bb
s
Stumpf, The una-flovv steam engine.
10
146
tc
147
10*
149
150
is again driven from the tail rod. In the single-eccentric type of valve gear the
lay-shaft as well as the hole in the latter for the synchronizing device are shorter,
and the governor may be placed close to the rear bearing. The basic idea of this
gear is similar to the one used by the Maschinenfabrik Augsburg- Nurnberg. The
single-eccentric gear is also described in the Z. d. V. d. I., 1914, No. 19, page 729.
The valves are placed on the cylinder and consequently have somewhat larger
clearance volume and surfaces. The jacketing is excellent; the arrangement of
the condenser, however, is not free from objections. The firm reports the following
steam consumption results.
Mean
Engine
Rated
Load
Stroke
Bore
Steam
Pressure
at. gage
Steam
Temp.
r. p.'m.
Steam
Consumption
kg/IHP-hr.
Measurements
based on"
HP
mm
mm
C
1
550
1000
750
11.8
250
127.2
4.8
1
2
810
1300
850
9.5
269.7
91.6 -
4.99
Boiler
3
300
800
650
11.7
253
15.67
488
[ feed water
4
523
800
600
11.5
231
152
4.88
J
5
140 | 600
1
375
9.7
340
200
4.38
Condensate
The steam temperatures in engines 1 and 5 were measured at the entrance
to the cylinder head, and in engines 2, 3 and 4 in the middle of the same.
Fig. 31.
The greatest credit for the commercial introduction of the una-flow engine
is due to Sulzer Bros., of Winterthur and Ludwigshafen. The cylinder of their
first engine, which is^ in operation in the brass rolling mill of Wieland Bros., at
Ulm, was designed by the author (Fig. 4, chap I, Ib, p. 10). The design of later
engines was based on this first one, the only change being the substitution for
the Stumpf gear of a lay-shaft gear, having two eccentrics which operate the
valves by means of cams and roller levers pivoted in the valve bonnets (Fig. 31).
The reciprocating roller slide of the Stumpf gear has therefore been replaced by
the pivoted roller lever. The governor is placed close to the rear lay-shaft bearing
in order to reduce the
deflection of the shaft
and also to shorten
the bore required for
the synchronizing de-
vice. All of the driving
parts, including those
of the air pump in the
basement, are comple-
tely enclosed. A gear
pump on the lay-shaft
supplies oil under a
pressure of about 1 at.
to the bearings of
engine and air pump.
The oil collects in a
reservoir in the base-
ment, where it is fil-
tered and again enters
the circulating system.
Low oil consumption
and smooth running
due to the oil cushion
in the bearings are ad-
vantages of this sy-
stem. The consump-
tion of cylinder oil is
also low since only cne
cylinder and generally
only one piston rod
packing have to be
lubricated, as against
two cylinders and se-
veral packings in an
ordinary multi - stage
engine. The low oil
consumption is also
proved by the high
mechanical efficiency.
An engine furnished
by Sulzer Bros, to the
firm of Junker & Ruh,
of Karlsruhe, having
a cylinder bore of
675 mm, a stroke of
800 mm, and a speed
152
of/150 r. p. m., showed that for a steam pressure of 11 to 12 at. gage and a tem-
perature of 250, 12 to 18 kg of cylinder oil were consumed weekly, and 8 to 10 kg
of bearing oil were added to the circulation when running 10 hours daily and six
CO
tB
co
be
days per week. The whole of the oil in circulation, amounting to about 1 barrel,
is replaced every 6 to 9 months. A hand pump is provided to supply the bearings
with oil before starting. As shown in the sectional drawing Fig. 32, the cylinder
design incorporates all the essentials previously mentioned. It is, however, to be
153
60
154
CO
bo
155
regretted that in many cases cylinder jackets are omitted when their use should
be 'dictated by low steam temperatures.
. Most of the engines built by Sulzer Bros, are fitted with a valve design the
purpose of which could be otherwise accomplished simpler and better.
Extensive experiments have enabled Sulzer Bros, to find the proper mixtures
for cylinder and piston castings, whereby reliable operation of these parts is in-
sured without the use of a tail rod. The cylinders are bored barrel-shaped so that
the cylinder surface becomes almost exactly cylindrical under operating conditions.
To these precautions, in combination with a thoroughly reliable lubricating system,
must be ascribed the fact that Sulzer Bros, have never had piston troubles. All
of their una-flow engines have therefore been built with self-supporting pistons,
except the engine shown in Figs. 33 and 34, supplied to the Crefeld Cotton
Spinning Mill, and a series of engines supplied to the Badische Anilin- and
Soda-Fabrik, where a tail rod was used to meet the purchaser's wishes.
A Sulzer stationary engine of standard design is shown in Fig. 35 (350 BHP
at 150 r. p. m.), while Fig. 36 shows two Sulzer una-flow engines of 450 BHP
each, supplied to the Hafod Copper Works, Swansea, South Wales.
The Maschinenfabrik Esslingen employs a particularly effective method of
boring una-flow cylinders under temperature conditions closely approaching those
of actual operation. The cylinder ends are heated to a high temperature by ad-
mitting live steam to the jackets, and the middle is cooled approximately to con-
denser temperature by a blast of air through the exhaust belt. The cylinder is
then bored cylindrically, and the piston is turned smaller than the cylinder bore
with a correct allowance, a difference of four thousandths of the diameter being
usually sufficient. Since the piston expands more than the cylinder, and is of
great length, ample bearing surface will be obtained. The piston heads should
be turned somewhat smaller in order to allow for their greater expansion.
A piston as shown in Fig. 5, chap. I, 4, p. 77, fitted with bronze shoes, offers
still greater safety against seizing, and this is true to a still greater degree of the
floating piston having clearance all around.
Piston troubles are caused in many cases by a wrong method of supplying
oil to the cylinder. The force pump should be timed in such a manner that deli-
very takes place only while the feed orifice in the cylinder is covered by the piston.
A good distribution of oil to the piston and cylinder wall will then be obtained.
The oil feeds should preferably be placed in the center of or close to the exhaust
belt, where the cylinder has the lowest temperature. One feed should be arranged
on the vertical center line and one each at either side in or below the horizontal
plane, each feed being supplied by a separate plunger. Complaints which are
sometimes made regarding the high oil consumption of una-flow engines frequently
arise from defective methods of introducing the oil. It is fundamentally wrong
to supply two or more feeds from the same pump plunger.
A neat arrangement of piping is obtainable if the steam pipes are placed on
one side of the engine and the exhaust pipe, air pump, and the cooling water and
discharge pipes on the other.
The following is a report of economy tests on the una-flow engine of the Cre-
feld Cotton Spinning Mill, built by Sulzer Bros.
156
Steam was generated by four Lancashire (twin furnace) boilers, having a total
heating surface of 400 sqm. A fifth boiler, the steam and feed lines of which were
blanked off from the others, supplied steam for heating purposes. The feed water
Fig. 38.
was weighed, transferred to a large tank and fed to the boilers by means of a cen-
trifugal pump. Indicator cards were taken every 10 minutes, and the steam pres-
sure, superheat and vacuum were recorded at the same intervals. The guarantees
given for this engine were:
157
Maximum load 2340 I HP:
10% cut-off, 300 temperature 1590 I HP 4.45 kg/IHP-hr
10% 350% 1530 ,, 4.15
13% 300 1920 4.65
13% 350 1860 4.35
These guarantees applied to steam of 11.5 at. gage pressure and condenser
cooling water of 15 C.
The duration of the
test was from 8 : 32 A. M.
to 4 : 06 P. M. a total of
454 minutes.
The load varied from
1477 to 1752 I HP with an
average of 1632.8 I HP at
a speed of 109.5 r. p. m.
Steam pressure at the cylin-
der was 11.1 to 12.1, average
11.6 at. gage; steam tempe-
rature at cylinder was 260
to 300, average 282.6 G;
vacuum was 71.3 to 73, 1cm,
average 72.2 cm. Barometer
reading 76.4 cm.
The temperature of the
cooling water was 11.5
and that of the air pump
discharge 30.1 G.
Since the load was higher than the guaranteed figure of 1590 HP, and the
steam temperature less than 300, a corresponding correction of the test results
was necessary.
According to the guarantees, the steam consumption increases from 4.45 kg
to 4.65 kg or 0.2 kg for an increase of load from 1590 to 1920 HP, or 330 HP;
therefore for an increase in load of 1632.8 1590 =42.8 HP, the permissible
2 42 8
increase may be - - = 0.026 kg and the steam consumption may be 4.45
ooU
-f- 0.026 4.476 kg, and still be within the guarantee.
Tests have shown that a reduction of the initial temperature from 300 to
282.6 produces an increase in steam consumption of 3%.
The permissible steam consumption may therefore be 4.476 x 1.03 = 4.61 kg,
and yet be within the guarantee.
The total steam consumption was 56410 kg, or
56410-60
or
454
7455.06
1632.8
7455.06 kg/hour
== 4.56 kg/I HP-hour.
158
159
The guaranteed figure was therefore satisfied without taking advantage of
the permitted allowance of 5%.
The above report was made by the Association for the Inspection of Steam
Boilers, of Miinchen-Gladbach, Crefeld Branch Office, on October 4, 1913, and
signed by Mr. Rhenius.
In examining thik result it must be borne in mind that the cylinder barrel of
this engine was not jacketed.
A una-flow engine designed by the author for the Soumy Machine Works
is shown in Figs. 37 to 39. It has valve gear of the Stumpf type and is designed
to be used with saturated steam of 7 at. gage. The cylinder has a bore and stroke
of 450 and 600 mm respectively, and the speed is 150 r. p. m. The resilient inlet
Fig. 41.
valves and the clearance valves are designed and arranged in such a manner that
the total clearance volume amounts to only 1.24% for a linear piston clearance
of 3 mm. The nut is flush with the piston so as to avoid the clearance volume
of about 0.5% resulting from a projecting nut. The two-piece cast steel self-sup-
porting piston is fitted with a bronze shoe fastened to it with copper rivets; the
rest of the piston has several millimeters clearance all over. Each half carries three
somewhat narrow rings, none of which overruns the cylinder bore. The harmful
surfaces are small, and are jacketed and machined in addition. The ends of the
cylinder are provided with jackets since saturated steam is used. The suction
of the air pump takes place through ports, and the discharge valves are arranged
in the heads, so that the clearance is small and the suction effect a maximum.
The condenser is placed immediately under the cylinder with a connection of
large area.
The cylinder jackets are supplied with steam through a separate pipe con-
nected to the steam main ahead of the stop valve. This allows the cylinder to
160
be warmed up before starting, and the valve gear therefore gives correct distri-
bution from the very beginning. The eccentric rod is shortened and guided by
a swinging link interposed in the valve gear, thus compensating the angularity
of the connecting rod. This equalization of cut-offs and valve lifts at the two
cylinder ends is very complete for all cut-offs, which range from to 25%.
Fig. 40 shows another similar engine with Stumpf gear designed by the author
for a company in Finland. The lower seats, instead of the valves, are resilient,
and the piston is fitted with Allan
metal rings to prevent seizing. The
clearance volume is 1,25%.
Details of una-flow engines
built by the Ames Iron Works,
of Oswego, N. Y., are given in
Figs. 41 to 44. The head and the
cylinder ends are jacketed and the
additional clearance spaces are
arranged in the cylinder heads.
The upper resilient seat of the
valves is made of steel and shrunk
in place on the cast iron valve
body. For non-condensing service
the engine is fitted with a piston
having cupped ends and the length
of compression is 90%.
The following test results were
verified by Mr. F. R. Low, editor
of "Power" (S. page 160).
In Fig. 45 is shown a small
vertical una-flow engine of 30 HP
at 400 r. p. m. for marine lighting
service.
Fig. 42. A better design is shown in
Fig. 46, illustrating a similar engine
of 30 HP designed by the author for an English firm. The cylinder bore is 220 mm,
the stroke 160 mm, and the speed 400 r. p. m. The governor eccentric oscillates a roller
lever acting on a triangular cam which transmits the motion to the valves. The
whole cam mechanism is enclosed in a separate housing filled with oil. The cylinder
ends are jacketed, and the additional clearance pockets are formed in the cylinder
heads and arranged to be heated by live steam when operating condensing. The
perfect tightness of- the single-beat valves employed, the small clearance space
and 'clearance surfaces, the generous jacketing, and the ample exhaust port area
all combine with the una-flow action to insure high economy. The single-beat
valves provide absolute safety against damage from water which may be trapped
in the cylinder.
The design of the two cylinder vertical single-acting stationary engine shown
in Fig. 47 is worthy of notice. The inlet valves are single-beat and are placed in
161
the center of the cylinder heads. The valve
gear consists of a cam mechanism with
reciprocating slides operated through a bell
crank by an eccentric and shaft governor
located at the free end of the crank shaft.
The cranks are set at 180 in order to obtain
proper balance at the high speed at which
this engine is t<5 run. Single-beat valves are
permissible because the high compression
balances the pressure against which they
open. The valve gear parts may, however,
be easily made strong enough to withstand
the load if the valve should be lifted when
there is no compression to balance the
Fig. 43.
pressure upon it. The cylinder head proper
is a thin dished steel plate, and the cylin-
der is provided with a forged steel liner.
The cylinder head is jacketed and the
upper end of the cylinder is also heated by
live steam admitted to a number of turned
grooves. Each groove communicates with
the adjacent ones at opposite sides so that
a continuous flow may take place through
the grooves. The engine is intended for use
with saturated steam, for which reason the
unjacketed part of the cylinder next to the
exhaust belt is short. The forged steel
cylinder liner should be made of hard ma-
terial in order to insure satisfactory ser-
vice of the cast iron piston rings. The
top surface of the piston is arched to
Slumpf, The una-flow steam engine.
1
r^S - s
K* rri CT> C^ CO
^ 00