LIBRARY UNIVERSITY OF CALIFORNIA. Class STEAM PIPES STEAM PIPES THEIR DESIGN AND CONSTRUCTION A Treatise of the Principles of Steam Convey- ance and Means and Materials Employed in Practice, to Secure Economy Efficiency and Safety By WM. H. BOOTH AUTHOR OF " LIQUID FUEL AND ITS COMBUSTION*' Member of the American Society of Civil Engineers ; late of the Manchester Steam Users' Association, The Neiv South Wales Government Railway Def>t. t etc. SECOND IMPRESSION LONDON ARCHIBALD CONSTABLE & CO LTD 1906 1 BUTLER & TANNER, THE SELWOOD PRINTING WORKS. FROME, AND LONDON. OF THE UNIVERSITY OF PREFACE IF any engineer will refer to his various text-books for information on steam piping, he will find very little to assist him. The earliest steam pipes served to convey steam at atmospheric pressure into cylinders of large size. The object of the steam was simply to fill the cylinder with a condensible vapour, so that upon cooling it there was a vacuum formed and the air pressure on the opposite face of the piston was made to do work. So long as the piston could be drawn up by the weight of the pump spear rods hung to the opposite end of the beam, it mattered little that the pipes were small, for there was no object in filling the cylinder with steam at full boiler pressure, for the vacuum ob- tained on condensing would be better as the steam was the less in quantity. When the steam engine was made rotative its piston speed became great and more regular, and pipes of a definite size began to be found necessary. Most of the early books on the steam engine gave engine proportions of a very empirical order, deducing them from the cylinder diameter by formulae that were sound only so far as they were applied to engines constructed 201593 PREFACE to standard practice. Standard practice in so far as beam engines were concerned was very much as Boulton and Watt had made it, and no doubt the proportions given by these curious old formulae were properly deduced from what experience had fixed upon as good practice. Our modern practice in steam pipes is thus the outgrowth of Boulton and Watts' early experience. It is a well-known fact that when steam escapes to atmosphere through an opening, the weight of flow is proportionate approximately to the absolute pressure. Thus steam at ninety pounds absolute pressure will escape three times as fast as regards weight as will steam of thirty pounds absolute pressure. Yet as steam pressures have advanced the area of pipes has not diminished accordingly. Any steam boiler, no matter at what pressure it may be worked, will con- vert a steady weight of water into steam. It follows therefore that for boilers of high pressure the steam pipes may be proportionately less than for low pressures. But as all practice has been towards higher pressures, it has always been the case that steam pipe mounting blocks, steam entrance pipes to engines, steam valves for any stated size of boiler, 'have been gauged by last year's practice rather than this year's, and, as is always the case in en- gineering practice, there has been a delay in cutting down pipe dimensions to the equivalent of the rise in pressure. While steam pipes may be too small they are probably more often too large, and not the less so that they have been called frequently vi PREFACE to supply many engines of high rotative speed which make a demand on the steam supply that is prac- tically continuous. Apart from mere sufficiency of piping, as such, there are the numerous details involved in flange diameters, bolts, centre lines, sockets, joints, and, not less important, valves, which all demand investigation. It is the object of this book to bring together such information as may be useful in connection with steam pipes as will be of assistance to the engineer who has to face problems of this sort. Steam piping to-day is so costly an item, especially when large valves are employed, that every effort should be made to mini- mize such cost without sacrifice of efficiency. No attempt has been made to put together everything that is published on steam piping. Selections only have been possible, and the author is indebted among others to the Babcock & Wilcox Co. for permission to reproduce from their book Steam and other of their pamphlets ; to Mr. A. J. Lawson, of the British Electric Traction Co. ; the Mannesmann Tube Co. ; the Cruse Superheater Co. ; and to Mr. Arthur Venning ; also to Messrs. Holden & Brooke ; Messrs. Templer & Ranoe, whose productions the author has employed to illustrate the book ; Messrs. Yates & Thorn, for numerous illustrations of Lancashire practice ; Mr. Stromeyer, of the Manchester Steam Users' Association ; Messrs. J. E. & S. Spencer, Ltd. ; Mr. Thos. Walker, of Tewkesbury, and others. In the book will be found various tables of dimen- sions of junction pieces, valves and flanges. In vii PREFACE printing these the author has merely been influenced by a desire to put before readers a few examples as a guide, and not as a fixed and determined standard. There are many so-called standards which differ from one another in but little, and per- haps the most important detail to standardize is the flange as regards diameter, bolt circle and bolt numbers, and their relation to centrelines. A com- mittee is now sitting on this subject, and doubtless will arrive at a result which engineers can accept. It will probably be useful as a guide to have general dimensions of even a single make of valve, but the author would suggest that overall dimensions of valve bodies ought also to be made standard, for such would make it possible to get out a whole system of pipes before deciding on where the valves should be obtained. There are so few and so small differences between one maker's products and those of another that a universal standard should be quite practicable. 2, QUEEN ANNE'S GATE, WESTMINSTER. vm TABLE OF CONTENTS CHAP. PAGE I THE DUTY AND OBJECT OF PIPES FAULTS OF DUPLICATE PIPE SYSTEMS i II FLOW OF STEAM Loss OF HEAD VELOCITY FORMULA TABLES EQUATION OF PIPES RESISTANCE OF ELBOWS, ETC. ... 4 III MATERIALS COPPER, CAST IRON, STEEL THICKNESS OF PIPES JUNCTION PIECES-- DIMENSIONS FLEXIBLE PIPES RIVETED PIPE FLANGES JOINTS SOCKETED PIPES WHITWORTH PIPE THREADS AMERICAN PIPE THREADS 23 IV EXPANSION COEFFICIENTS SPRING BENDS GENERAL ARRANGEMENT SLIDING JOINTS SWIVEL JOINT ANCHORING TEMPERATURE AND PRESSURE OF STEAM 55 ix CONTENTS CHAP. PAGE V STRENGTH OF PIPES THREADS BOARD OF TRADE RULES . . . .-.'.. .72 VI ANTI-PRIMING PIPESOUTLET VALVES DRAIN PIPES INCLINATION ISOLATING VALVES WATER HAMMER BRANCHES . .. .76 VII PIPE JOINTS SPIGOT SOCKET SCREW FLANGES JOINTING RINGS, ETC. SUPERHEATED STEAM 83 VIII PIPE SUPPORTS BRACKET SUSPENSION PILLARS PLAIN BRACKETS VIBRATION ANCHOR BRACKET TABLES OF DIMENSIONS . . 88 IX ERECTION OF PIPES TEMPLETS FOR PIPES TAPER JOINT RINGS EXTENSIONS TO EXISTING PIPES PIPE BENDING .... 97 X GENERAL ARRANGEMENTS RELATIVE PostriON OF BOILERS, ENGINES AND SUPERHEATERS- SIZE OF VALVES BOILER OUTPUT MODERN BOILER CAPACITY 105 x CONTENTS CHAP. PAGE XI VALVES, GLOBE, ANGLE, FULLWAY OR GATE BYE-PASS RELIEF REVERSED FLEXIBLE SEATS ISOLATION MATERIALS .... 112 XII DRAINAGE STEAM TRAPS .... 128 XIII J UNCTION PIECES AND FLANGES WEIGHTS CON- STRUCTION MATERIALS FLANGE DIMENSIONS BOLT PITCH STANDARDS . . . .133 XIV SEPARATORS EXHAUST HEADS ATMOSPHERIC VALVES . 145 XV SUPERHEATED STEAM, RELATIVE VOLUMES PIPE COVERINGS NORTON'S EXPERIMENTS ATKINSON'S REPORTS ..... 152 XVI WEIGHTS OF PIPE SPECIFIC GRAVITY OF MATERIALS 173 XVII THE KINETIC THEORY OF GASES IN RELATION TO THE FLOW OF STEAM . r .178 XI CHAPTER I Steam Pipes : Their Duty and Object HE object of a steam pipe is to convey steam T from point to point. A steam pipe as ordinarily understood is for the conveyance of steam from the boiler to the engine,, or to other apparatus, such as a dye vat or brewing copper, etc. In the early days of steam engineering there were no steam pipes ; the working cylinder of the engine was wholly or partially connected to the boiler or joined by a narrow neck, which, becoming gradually longer, developed into a pipe. Sometimes of rect- angular section for convenience under special cir- cumstances, the natural and usual cross-section of a pipe is the circle, that being a figure which con- tains a maximum of area within a minimum cir- cumscribing boundary, and also being the only figure of maximum strength to resist bursting, and there- fore requiring no internal or other stays. Since a steam engine gives the best results at the highest pressures, the duty of a steam pipe is to con- vey steam to the engine with a minimum of loss of pressure. Steam being hotter than the air surround- ing the pipes must lose some of its heat in its passage I B STEAM PIPES through pipes. Obviously, therefore, pipes must be of minimum size. This is inconsistent with a minimum pressure loss, and a compromise must therefore be come to between loss of heat and loss of pressure. If that compromise could be worked out, it would be found by equating the loss of coal due to loss of economy consequent on loss of a given amount of pressure, and the loss by radiation of heat that would be incurred by making the pipe large enough to prevent said loss of pressure. Beyond a certain small loss of pressure, any further increase of pipe diameter affords so little further reduction of friction and adds so much to the heat radiation losses from the pipe surface that very large pipes must not be used, for they involve also increased capital expenditure in pipe sizes, coverings, flanges and valves out of all proportion to the small gain of pressure. The broad principles to be observed to enable a pipe to perform a maximum duty are that it shall take as direct a course as practicable between any two points, shall be as smooth internally as it can reasonably be made, and shall have bends of large radius. All these points are compelled to be neg- lected by circumstances, but they afford a basis for design. As a steam pipe is meant to convey steam and will be called on to convey as much water as may be formed in it by condensation due to cooling, this must be provided against by suitably covering the pipes with a non-heat conducting substance. Other- 2 THEIR DUTY AND OBJECT f wise not only is heat lost, but, water being formed, must be impelled along the pipe at the expense of the steam, which will lose pressure as a result. As a steam pipe failure will cause the stoppage of a whole power station, the practice has grown up among electrical engineers of duplicating the steam pipes. Hence arose that nuisance the ring main, with its myriad of costly valves, its maximum of condensation and its minimum of safety. The ring main ought only to be employed where other conditions render it obligatory, and these only occur as a rule with initially bad designs. The ring main is not a steam engineer's device. Steam engineers have always arrived at directness of pipes knowing the losses of heat in long pipes, and they insist on the use of the best materials so as to minimize the chance of failure rather than countenance the dupli- cation of bad work which has arisen from want of knowledge of steam engineering conditions and an apparently foregone conclusion that break-downs are necessary, and must be encouraged to occur by provision of a maximum of parts to fail. CHAPTER II The Flow of Steam RANKINE, in his work on the Steam Engine, gives the following formula for the velocity of flow of steam where V = velocity in feet per second. g = gravity = 32-2. 7 = 1-3- po = ideal pressure at 32 F. To = absolute temp, at 32 F. Ti = at pressure p t . T* = p2. pi = pressure in boiler. p 2 = at steam chest. v = volume ideal at 32. po v = 42141. k = coefficient of contraction. Vi = volume of i pound of steam at pi. Substituting values this becomes = /J6 4 >4 x i> 3 x 42141 X V \ 493 . x 0-3 = ^71066 = 265 feet per second, where pi = 200 Ib. absolute and 4 THE PLOW OF STEAM p 2 = 197 lb. That is to say, with a drop of pres- sure of 3 pounds the velocity in a short straight pipe may be 265 feet per second. For small differences of pressure Rankine gives an approximate formula /2gx 42141- rifri-fr) (6) ' V Substituting values gives F = ^72260 = 270 feet per second, which is not a serious difference from the complicated formula. In his Rules and Tables, Rankine gives a rough approximation of the weight of steam flow per second where the initial and final pressures are p and p 2 respectively, and q = pounds per second per unit of area. (1) Where p 2 = or < 5 p q = p t 4. 70 nearly. o This formula is only useful when the external or final pressure is low. (2) Where p, >3. #1 , q = Applying this formula (2) to the case of steam of 200 lb. (28,800 lb. per square foot) flowing with a loss of 3 lb., or to a final pressure of 197 lb. (28,368 lb. per square foot), we have by (2) ' or 675-4 x 0-1511 = 102-06 pounds. As at 197 pounds pressure there are 2-26 cubic feet per pound, the velocity of flow per second will 5 STEAM PIPES be 2-26 x 102 = 230-6 feet, which again is not very seriously different from the velocities found by the more strict formulae. Ordinarily gases flow by virtue of the same rules as apply with liquids. The rule for the flow of a liquid is v = x/2p or v = 8 the Page 8, line 14 for Then >= -36 jr read ThenZ)= .36 A/('* g 38 LlTD)^"^ ,,__-_^ __^ . 5Q per second, and Va = 278 x 0-196 = 54-5 cubic feet of flow per second. Hutton gives a rule for the outflow of steam into an external pressure not more than 58 per cent, of the internal pressure, as follows : W = weight of steam discharged per minute. Then W= ^- y where P = absolute pressure in pounds per square inch. A = area of pipe in square inches. C 1-38 for pipes up to 10 ft. long. 9 STEAM PIPES velocity is reduced to one half, showing that friction is very considerable. It is usual in practice to require to know the dia- meter where the quantity and the fall of pressure are given as in fixing the pipe diameter for a boiler of a given evaporative capacity at a certain dis- tance from the boiler. Calling Q the volume in cubic metres per second. L = pipe length in metres. D = pipe diameter in metres. P = difference of pressure available in assumed value is then used under the root sign and the formula worked out. If the calculated and the assumed values of D are found to differ, the new value of D as calculated can now be used under the root sign and a fresh value calculated which will be very close to the first calculated value where pipes are of usual lengths or several diameters long. For convenience in using these metrical formulae the following equivalents will be useful in reducing British data to metrical i metre = 39-37 inches = 3-28 feet. i foot = 0*305 metre. 8 THE FLOW OF STEAM i inch = 0-0254 metre i cubic foot = 0-0283 cubic metre, i metre = 35-316 cubic feet, i pound per square inch difference of pressure = 0-052 metre of mercury column. A formula sometimes employed for the velocity of flow in a pipe is V = 50 V ; the value of JL/ H being v.p. 144, where V = feet per second, v the volume in cubic feet of i pound of steam at the initial pressure, L and D the length and diameter of the pipe in feet, and p the difference of pressure between the two ends of the pipe. The number of cubic feet per second is then Va, where a is the pipe area in square feet. Thus for a 6-inch pipe 50 feet long, carrying steam of 100 pounds absolute pres- sure, with a drop of 5 pounds, we have v =4-29 from any steam table, whence H = 4-29 x 5 x 144 = 3088-8. Hence, V = 5oA/3 88>8 x '5 = 278 feet 50 per second, and Va = 278 x 0-196 = 54-5 cubic feet of flow per second. Hutton gives a rule for the outflow of steam into an external pressure not more than 58 per cent, of the internal pressure, as follows : W = weight of steam discharged per minute. Then W= ^4, where O P = absolute pressure in pounds per square inch. A = area of pipe in square inches. C 1-38 for pipes up to 10 ft. long. 9 STEAM PIPES = i'39 for pipes up to 40 feet long. = *'4 2 >> >> 7 = 1 '45 y, )> IO >y P of course must not be less than 12 pounds above the atmosphere. The velocity of flow through a valve or short pipe is V = 32 J T + 460, when T is the temperature and V = feet per second. Neither of these rules bears upon boiler steam pipes, because a drop of 42 per cent, or more could not be tolerated. The rules are useful for special cases. PRACTICAL RULES. Hutton gives as a good practical rule that in ordin- ary cases steam velocity should not exceed 85 feet per second = 5,ioo feet per minute. If very short and straight the velocity may be no feet. He supposes steam to follow the piston for full stroke, and gives the following rule for steam pipe area Cylinder Area x Piston Speed per minute ~5i55~ =Pipearea. This rule is purely empirical, for it supposes con- tinuous steaming and neglects the higher piston velocity at middle stroke. Referred to evaporation the steam pipe area is given as Pounds of Steam per minute x volume of steam relative to water A t Velocity in feet per minute x 62-42 -f- 144. Mr. Stromeyer, of the Manchester Steam Users' Association, gives as a rough rule for sectional area of steam pipes in square inches = A. 10 A = THE FLOW OF STEAM 180 x Sum of widths of furnaces in inches Absolute Pressure. This rule corresponds with a velocity of 8,000 feet per minute and a fuel consumption at the rate of 25 pounds per square foot per hour on grates 6 feet long. Obviously the rule has been empiricised from ordinary rates of evaporation and combustion. In ordinary land practice the constant becomes 240, but where there is an ample excess of boiler pressure, as in case of Belleville boilers, for example, which carry pressures much in excess of what is permitted to reach the engines, the constant need not be more than 120. The combined formula for flow of steam due to differ- ence of pressure and length of pipe and diameter is where (i) W = the weight in pounds per minute. D = the weight per cubic foot of steam at the pressure. pi and p 2 = the initial and final pressures. L = the pipe length in feet. d = the pipe diameter in inches. Table i has been calculated for Steam (B. & W. Boiler Co.) from this formula for pipes having a length of 240 diameters, straight and smooth. The results are in pounds per minute. In using this formula it must be noted that actual pipe diameters are employed as per Table II., which gives the pipe diameters on which the Table I. is calculated for all sizes below six inches. IT STEAM PIPES a WM .3 5 txoo co O (SJ I^IOH row H o HO *O co H m cr\ co m >. c^oo Tf uo t>. OO C^H -"a" Tf^o O ao O O C>t^H OO C\H O r> Tf o H M H oo x iSs O lO CO OO CN, COTfiOO Cx-OO CN O H CO H H H H M M M iOfOH O lOtxlOHOO H CO tovO HOO C^ CTNC^O t^OO OO O M iO ON M O ON o H 9 oo oo M tx o o >sin O CO CT> H lOrOOO O ONVO H rfvD C^O M HOOiOCQOTfH CO rfiOO-OO C^H O Tf CO Tf COOOO ONlpOO H ^ H 6 CNOO CM M 10 10 O Tf C OO 9 M OO Tf COO CO t^ l>s Tf oooobo Tf H j>scooo coob l^OO oo CO co x o H Tf OO O oo tx a\ H OO O 1OHVO H M COCOTf H Tf OO IO TfH TfCpOO H H HOOlOO H H M C o c* Tfio^c^H m COCOCOCOCOCOTfTf CSJ Tf Tf H HOO tO, O> H M Tf ip CNOO O^ H Tf HHHMMMMMMMCOCO 4! Jj3 ^ HOOOOOOOOOOOO H M COTfioO C^CO C^O CvJiO H H H 12 THE FLOW OF STEAM In order to render the table suitable for pipes with bends and valves, the values of these are as follows. The resistance of the first opening from the boiler and that of a globe valve are each equal to a length 114 of pipe = for various pipes This length works out as follows in. in. j in. in. in. in. in. in. m. in. in. in. in. in. d = I i ij 2 at 3 A 5 6 8 IO 12 15 18 L = 20 25 34 41 47 52 60 66 7i 79 8 4 88 92 95 An elbow is equivalent to f of a globe valve These equivalent lengths are all added to the straight pipe length, and the total is the equivalent straight pipe. Thus a 4-inch pipe 40 feet long (120 diameters) with a globe valve and three elbows, and the open- ing from the boiler is equivalent to a length of 120 + 60 + 60 + (3 x 40) = 360 diameters, or ij times the tabular length. The flow through this pipe will be that given in the table multiplied by i + >/i5, or 81-6 per cent, of the tabular number. That is, for any length of pipe other than 240 diameters, divide 240 by the equivalent length, and take the square root of the quotient, which divide into the tabular weight. The result is the weight of flow for the new length. Again, for any loss of pressure other than i pound, multiply the tabular figure by the square root of 13 STEAM PIPES the pressure drop. Thus a drop of four pounds instead of one pound should double the output. The formula (i) is sometimes written W = 303-25 d* \/ D & ~ where L is the number of times the length is of the diameter. Ob- viously this brings the term d into the denominator, and enables the d 5 to come from under the root sign, for d 2 = \/~. it The number 303-25 is the 87 of the previous for- mula multiplied by v'lz, which is necessary where it is changed to feet instead of being a multiple of a diameter in inches. When steam flows from one pressure to any other pressure less than three-fifths of the initial pressure its velocity has the constant value 888 feet per second, so that the weight discharged varies with the density. Hence the rule for weight of outflow per minute W pounds. W = area of opening a x 370 x weight of a cubic foot of steam. Rankine's formula is W = - -, where a is the area in square inches and p is the absolute pressure. A coefficient of reduction k =0-93 is employed for a short pipe and k= 0-63 for an opening in a thin plate, as a safety valve for example. When the steam flows into a pressure more than two-thirds the initial the formula becomes THE FLOW OF STEAM W= 1-9 a k ^ (p d)d, where d is the difference of pressure. The result is substantially what all other formulae give. In a system of pipes in order that a correct balance may be found the proper proportion of any size of pipe to allot as an equivalent of any other size must be found. Pipes deliver according to the square of their diameters, but the same head will not pro- duce the same velocity of flow in four 5-inch pipes as in their equivalent one lo-inch pipe. The relative flow W in different pipes varies as where W = weight of fluid and d = diameter in inches. In Table II., from Steam (Babcock & Wilcox Co.) the true or standard diameters of pipes are given, and in Table III. are given the equivalents of pipes in terms of other pipes. That part of the table above the diagonal line refers to standard pipes of the nominal diameter only. Below the diagonal the pipes are actually of the given diameter. Thus below the top line 7, along the line 2, we find 29, or the number of nominal 2-inch pipes equal to a single 7-inch nominal, or again, 6' 21 pipes of standard 7-inch size = i pipe of actual 14-inch size, but it requires 6-45 pipes of 7-inch standard to equal one standard 14-inch. The table is useful, but it is calculated for American pipes and must be used with discretion with English pipes, though no very serious discrepancy will arise. 15 STEAM PIPES TABLE II. OF STANDARD SIZES, STEAM AND GAS PIPES. Diameter. Diameter. Diameter. Size, Ins. Inter- Exter- Size, Ins. Inter- Exter- Size, Ins. Inter- Exter- nal. nal. nal. nal. nal. nal. 4 27 40 2* 2-47 2-8 7 9 8-94 9-62 i 36 54 3 3-07 3-5 10 IO-02 1075 1 49 6 7 3* 3-55 4 ii II n-75 J 62 8 4 4 4-03 4-5 12 12 12-75 1 -82 1-05 i 4i 4-5i 5 13 I3'25 14 I i'05 I'3I 5 5-04 5-56 14 I4-25 15 ij 1-38 1-66 6 6-06 6-62 15 15-43 16 i* 1-61 1-90 7 7-02 7-62 16 16-4 17 2 2-07 2-37 8 7-98 8-62 17 I7-32 18 Mr. Geipel gives the following rules : d = diameter in inches. L = length in feet. p = loss of pressure due to friction. D = weight of steam in pounds per cubic foot. Q = pounds of steam per hour. v = velocity of flow in feet per minute. 9000-000 d* D v = 9170 V jj^y whence the Tables IV., V. are JLt LJ deduced. 16 PO > HI OO O\ HI O 00 O -rf-OO VO O vo vo ON N uo ONVO PO HI HI vo Tf POOO HI IN txvNOvOPOHiHH^^^^^,;^,; M ;< ^ 4j-vo 6 - ON ^ N rj- N HI M HIM vo M tx ON ON HI POVO PO POVO vo txoo vo IN HI HI Q CO PO POVO Tj- rf tx txVO OO ^t PO HI O O vp vp op POVO op Tj-op IN tx rj- M HI PO txop O O ONVO ""* OO HI N PO ONVO 4f PO N IN HI M M M M M 6NVO^pV>VOMPOMVOMVoO l-H| -'ONHivoN IN vr>OO M ^OO tx O tXVO^OO IN M ^ ^^ QVo4j-vNvNHIHI MMM^^^ CO^O tx O\ HI ON PO OOOOTfOO VQ vN M PQ M M M op ON HI op op vp ON O ""> Tj- CN POVO rj- txOO Vp VO O\ M M tx. O vN vooo M vo ONOO CN tx ^X IN PO tx Tj-OO vo PO M ON PO PO T*" ? 9 T 1 " ? 7 ? ^V^T*" ? ^ 9 ? 9 N *? 9 NV 9 O\ >-" O "- 1 txvo 6 rf POVO PO tN M N rj-vo ob M vooo POOO 4j- ONVO 4j- 6 M n co ^^ PO o\ tx O vo N ON^ vo PO HI 4l-vo 6 H: 6 ^ tx VO tx O PO voVO HIOO OVOVO IN ^rj - Os PO Tt- povp vp vp tx ONVO opvoONHi MVOO IN rftx ONVO M ro tx PO PO N PO M PO M M POVO 6vOPOvNvNvo6NTl-'~''~ l '~' hH '~ lN '^" lXM . ^"^t VON M M M N PO Tj- VOVO OO M M IN IXOOOvN OOMOpvOTt-vN VOOO tx""> VO - 1 ^ M Os PO PO txoo VO POOO rJ- -^ PO -fr o ONOO ^ ^" ^" <* PO ONVO vo HI oo oo POVO O txoo oo O ON tx O O -rf PO ONVO vo M oo oo PO IN ON PO O - ~' (S ^ ^ ^^^ M ^^O ONtx txOO O ONVO O PO N oo H. TttxON ^tOO HI oo N PO ON HI N oo POVO PO TJ- Tf Qvo NOVOVOVOtxvovoPOvNtxtxQ ^OO tx IXVO "^ PO O ^ HI VO vo N OO O tx txOO M O PO PO ON Tj-OO voO IN txPON txtNvoO IN vo A L 4*. 4*. M M POVO M tx ^ PO POVO O vo vo oo ON vo POVO vo rt M t\ PO vNvo HIMCv>POTfvotxOOO'vN~TftxO'vo"vN'voQ'txON M HI c* PO rf "^VO txoo ON O -i N PO rf 17 STEAM PIPES TABLE IV. Absolute Pressure. Pounds per Square Inch. Diameter of Pipe. 2 5 7.0 100 2OO Inches. Velocity in Ft. per Minute with i Ib. Loss of Pressure per 100 Ft. I I2OOO 7790 4070 1900 1380 2 I7OOO IIOO 5760 2690 1950 3 20850 13520 7070 3290 2390 6 2950O I9IOO 990O 4660 3380 9 36IOO 23400 I2O5O 5700 4140 12 4I60O 26900 I4IOO 6580 4770 15 46600 3O2OO 15600 7400 18 5IIOO 33200 I7IOO 24 58900 38150 TABLE V. Diam. Absolute Pressure. Pounds per Square Inch. of Pipe. Inches. 2 | 5 10 20 50 100 150 2OO Pounds of Steam per Hour with i Ib. Drop of Pressure per 100 Ft. I 22'8 35*2 487 67'6 104 144 174 200 2 129 199 275 382 590 815 984 1130 3 356 549 760 1054 1620 2245 2715 3I2O 6 2015 3100 4290 5960 9170 12700 15350 17640 9 5550 8550 11820 16400 25300 35000 42300 48600 12 II380 17500 24300 33700 51800 71700 86600 99600 15 19900 30610 42400 58500 90500 125500 I 18 31400 48400 67000 93000 143000 24 64400 99200 137000 190500 The accompanying diagrams are drawn from these formulae. 18 THE FLOW OF STEAM 50,000 ft. o> 40,000' 30,000' o 20,000' 10,000' I ' > 100,000* & 80,000* IH 3 GO ft "o 60,000* 4O,OOO* 2O,OOO* 8" 12 Pipe diam. 1 6* 20 dia. 4O* 80" 1 20" 100" 20O* Pressure, absolute, STEAM PIPES For other losses of pressure, multiply the tabular numbers by the square root of the new loss. For other lengths of pipe = L feet, multiply the tabular numbers by -/=. The values of /& are best obtained by means of logarithms : they are given here up to 40 inches in Table VI. TABLE VI. I i ii 401-3 21 2020-9 2 5-66 12 498-8 22 227O-I 3 15-6 13 609-3 23 2537-0 4 32-0 14 733-4 24 282I-8 5 55-9 15 871-4 30 4929-5 6 88-2 16 1024-0 4 IOII9-3 7 129-6 17 1191-6 8 181-0 18 1374-6 9 243-0 19 1573-6 10 316-2 20 1788-9 The length of pipe equal to a globe valve or to an opening into a pipe is given as L^ = 8-66 ^ The length equivalent of an elbow is d L. = 57 6 i + d 20 THE FLOW OF STEAM f For different diameters the equivalent lengths thus figure out in Table VII. TABLE VII. Equivalent Length to Diameter in Inches. Globe Valve or Pipe Opening. Elbow. I 1-9 i-3 2 6-2 4-2 3 7'9 5-2 6 32-5 21-6 9 55-6 37*0 12 79-9 53-2 15 100-5 69-6 18 129-9 85-9 24 180-6 123-8 Properly speaking, the opening to a pipe should be by a short converging piece, the wider end of which has an area about 10 per cent, greater than the pipe in order to allow for the vena contracta effect. Boiler mounting blocks do approximate to this shape, but their good effect is spoiled by the usually clumsy anti-priming pipe, which is not led up to the mouthpiece in an easy curve, and is usually plugged into the mouthpiece in such a way as to destroy the effect of this. The resistance of openings and elbows by the 21 STEAM PIPES above rule is given in the Table VII. on page 21, and it will be observed that the results, in the smaller sizes, are much below the figures given on page 31. CHAPTER III Materials STEAM PIPES are made of one of the four following materials : Cast Iron, Wrought Iron, Steel, Copper. CAST IRON. No material is so convenient or has been so largely employed as cast iron. Though a material of no flexibility, cast iron is strong and cheap, and with care can be cast sound and free from blemishes. The flanges are readily faced in the lathe, for which purpose stout bars carrying the centreings are com- monly employed. Bolt holes are easily drilled.. Pipes can be cast to any convenient length, and in brief, cast iron is without a serious rival for general purposes up to pressures of 100 pounds per square inch gauge pressure. Above that pressure the safety of cast iron admits of doubts ; above 120 pounds very serious doubts are to be entertained. The stresses in steam piping are not so much those of pressure as those which arise from expansion due to temperature changes and from water hammer, and, perhaps even more seriously, from forcing pipes to fill places which they do not fit properly. Some of these stresses are increased by pressure, viz., those due to expansive movements, and the high temperatures of superheat also have the same effect. Above 100 pounds, therefore, cast iron 23 STEAM PIPES should not be employed. True, junction pieces, valve bodies, etc., are still made by reputable firms, of cast iron up to 200 pounds pressure, and, while the author would condemn this practice, it is perhaps but fair to state that in such cases the choice of metal, the care in casting, and the rej ection of faulty bodies, combine with the abnormal stoutness of parts to render such castings very much less unsafe than ordinary pipe castings from a jobbing foundry, with more or less uncertain coring and no special selection of the iron. For exhaust pipes, however, cast iron holds the field. Exhaust pipes are usually larger, much larger than the steam pipe to the same engine, for it is their duty to carry the same steam in a wet condition and at much smaller pressures. To enable exhaust pipes to be tightly jointed the flanges require to be stout and to be faced. They should be cast from metal of good quality and tough, and not too hard to tool easily. Pipes should be cast vertically if they are to be reasonably safe against floating of the cores. The common fault of cast- iron pipes is the chaplet, which does not become melted fast in the pipe and causes blow holes, which admit air and vitiate the vacuum. D P A usual rule for pipe thickness is - - = 0*5, 4000 when D = diameter in inches, and P = pressure, but this rule will give too small a thickness for ex- haust pipes, and no pressure should be assumed less than, say, 4 pounds per square inch for each inch 24 MATERIALS of diameter of pipe. Flanges are made one-third thicker than the pipe body. Their duty is greater than the mere withstanding of pressure stress. In practice they are subject to very severe stresses of error which arise when pipes do not fit their places and joints are screwed up much too severely for good workmanship. TABLE VIII. THICKNESS OF CAST-IRON PIPES. in. in. in. in. in. in. Diameter . 4 5 6 7 8 9 Thickness . 1 A i i & * Diameter . 10 12 14 16 18 20 Thickness . f 1 1 1 I i Neither of the above rules is suited for a large range of diameters. A more adaptable rule is to make the thickness of the pipe "]"= -^~. This rule serves for pipes from 2 to 12 inches, up to 100 pounds pressure. Above 100 pounds the rule is T = P.P. 4000 + as given above, but these last rules give a pipe un- necessarily heavy for exhaust purposes, for which the author's rule is T= d D 2 i - + 0-5 -g, the result being taken only to the nearest sixteenth of an inch. Thus a 20-inch pipe has a thickness ? 22_)_ 0-5 4000 = -875". The nearest sixteenth to this is J inch, and this rule gives close results to ordinary practice, 25 STEAM PIPES as per Table VIII., which is that given by the Babcock & Wilcox Co. for exhaust pipes. Cast-iron pipes can be obtained in the form of tees, small radius bends and large radius bends. Straight pipes are made in g-feet lengths, or, for sizes below 3-inch bore, in lengths of 6 feet. Making- up lengths are, of course, made to any f templet' length (see Templet). Crosses, pockets, and Y" pieces are also made, and the Table XXIlA., later and figures 1-7, herewith, give the sizes of such pieces as made by the Babcock Company. It will be noted that all pieces of the same main size must always have equal overall dimensions. Thus the projection of the branch on a lo-inch J_ will always be made 13 inches, whether the diameter of the branch piece be 2 inches, 7 inches, or 10 inches, and, simi- larly, the dimensions A Ai of a cross will always be the same as the dimension A of a T, while of course the dimension B will always be half of A. Unless these precautions are taken, piping systems are liable to prove very inconvenient to put together. Some engineers do not hesitate to employ cast- iron junction pieces for the highest pressures, taking care to relieve the cast pieces of the stresses of expansion, but the author does not recommend this. When so used they are made specially stout, while stout pieces are used for lower pressure and lighter castings for exhaust steam. There is always some risk of a light casting getting into a high pressure line. Engineers differ as to the mode of facing flanges. A flange faced right across makes an excellent joint 26 MATERIALS MHn ~ T T 27 STEAM PIPES with woodite for steam pipes, or with a plain ring for exhaust pipes. Some engineers recess their flanges in the form of a shallow spigot and socket, as shown in Fig. 50, by Yates & Thorn. This certainly is a safeguard against blowing out the joint ring. Such pipes are often difficult to take down or to fit with a new ring. The Babcock Co. have a narrow facing strip only round the pipe, and make the joint with a light corrugated ring of brass or copper, all the bolt pressure being concentrated on the narrow face. See Figs, u, 12. In Table IX. are given the dimensions of the cast- iron exhaust tees of the standard of the British Electric Traction Co., kindly supplied me by Mr. A. J. Lawson, of that Company, and shown in Fig. 8. The only remark that might be made on these is that bolts of J-inch diameter, as needed for the f holes, are smaller than is perhaps desirable, noth- ing less than fth bolt diameter being very satis- factory in practice, though the smaller size was put in to be proportional to the rule of bolt numbers in multiples of four. This Company face pipe flanges straight across, with no projecting ring and no recess. Flanges faced flat across are often scored with two or three circular grooves put in to the depth of -^ or ^ with a single V-point tool, with the object of better hold- ing the joint rings. These grooves should only be employed where soft rings can be used. That is, they must not be used where joints are made with simple copper wire, for steam will leak at the joint where a wire may run across a groove. 28 MATERIALS FIG. 8. CAST-IRON EXHAUST TEES OF BRITISH ELECTRIC TRACTION CO. TABLE IX. CAST-IRON EXHAUST TEES. Size. A B C D E F G Bolt Holes. 24 2j 9 4* 7 5i I f f 1 3 3 9 5 74 6 A i 1 4 34 3* 10 5 8 6J * 1 i J 4 4 ii 54 9 7i 4 i * i 44 4i ii 6 94 8 i 1 f 5 5 13 64 104 8| i 1 1 6 6 14 74 12 IO i 1 1 o 7 7 15 8 13 I0| * f 1 8 8 16 84 14 12 * f 1 J 9 9 18 9 15 13 * I 1 } 10 10 19 10 i6i 141 i I 1 ii ii 20 10 17 15 I I t 12 12 22 ii X9i *7 f il J .12 13 13 23 12 20j 18 i ii 1 14 14 24 13 2li 19 1 ij 1 16 16 27 14 24 2li 1 if I 29 STEAM PIPES In case of reducing tees, no difference to be made in the dimensions B or C. Bolts smaller than holes. In Table X. and Fig. 9 the standards of the same Company are given for cast-iron steam tees, and in # f /SS////s/s////////777)(/// M T FIG< 9. CAST-IRON STEAM TEES OF BRITISH ELECTRIC TRACTION CO. TABLE X. CAST-IRON STEAM TEES TESTED TO 200 LB. Bolt Holes. Size. A B c D E F G H I N S 3 3 9ft 5* 7J 6J 4 i ii I 8 1 3i 31 IO )> 7l 6| >J 5J J) ?> i 3* 3i IO 6 8 61 * 3f 3l II 8ft 7 A } J> j j j> 4 4 II* 6J 9 7t JJ I Ii 5) ?> 5> 5 5 13 7i IOJ 8| i J) If i 5> J 6 6 15 8 12 IO i I* Ii )J )5 I 7 7 16 8ft 13 II Ii J J J) 12 I 8 8 18 9 14 12 1 Ii If 1 j) I 30 MATERIALS Table XL and Fig. lothe same Company's standard for mild steel tees with standard flanges. It will be noticed that these standards appear to FIG. IO. MILD STEEL STEAM TEES WITH STANDARD FLANGES OF BRITISH ELECTRIC TRACTION CO. TABLE XI. MILD STEEL STEAM TEES WITH STANDARD FLANGES. Size. A B c D E N s 1 1 5i 3 4i 2| 4 f I I 6 55 41 31 j) 55 ij II 7 3i 4l 3i 55 5 it It 55 4 5i 4 5 5 1 if If 8 4i 6 41 5 5 55 2 2 ,, 4i 61 5 8 f 2j 2j 81 55 j j j ,, J5 2j 21 9 5 7 51 55 55 2} 2j 91 55 7i 51 55 55 have been designed somewhat on the idea of each size standing by itself of the best proportion accord- ing to the designer's views for the particular piece. STEAM PIPES Thus the dimension C is not made one-half of B in every case, as it ought to be in the author's opinion. C is, however, always made the same for every tee of a given size A , no matter what the diameter of the branch or A. In event of a + being required it would not measure equally over each rim of flanges, or if it did do so it would not work evenly with the tees of the same size. Though good in themselves and useful as a guide in flange proportion and generally, these standards should be changed in such respects in order to secure uniformity, so that the set of a T 7 the half -breadth of a +, and the set of a quarter- bend may all measure alike. Mr. Venning advises that as regards bends in pipes which do not need to be proportioned to suit other junction pieces, the radius should not be less than five diameters of the pipe. An easy bend is better for the flow of steam. Small quarter-bends, however, when in particular situations, have to accommodate themselves to the size of other junc- tion pieces, hence the dimension \A in Fig. 4 corre- sponds with the dimensions of Figs, i and 2, which may be looked on as the leading junction pieces. COPPER. As a material for steam pipes, copper has long held a place it can no longer claim with high tem- perature steam. Copper pipes are flexible because weak, and have been much used for lengths or bends intended to give way under stress. 32 MATERIALS Copper for pipes must not contain more than 0-7 of i per cent, of impurity, and the pipes should be solid drawn. A rule for brazed pipes is D x P |-o- 1 25". where D = pipe diameter in inches 10000 and P = pressure by gauge per square inch. Brazing must be looked on with great suspicion, though it may be employed in attaching flanges, which should be four times as thick as the pipes. At a temperature of 360 F. the strength of copper is reduced 15 per cent. Copper is therefore to be employed cautiously for high pressures, and it is not a suitable material for conveying superheated steam. The diminution of tenacity is shown in the an- nexed table. DIMINUTION OF STRENGTH OF COPPER AT TEMPERA- TURES ABOVE 32. Loss of Loss of Temperature. Tenacity. Temperature. Tenacity. Per cent. Per cent. 68 2 638 35 138 5 748 45 248 IO 788 50 328 15 838 55 418 20 938 66 438 22 968 68 488 25 1168 88 The above figures must always be allowed for in calculating pipe strengths, the bursting pressure of which is found by the following rule : 33 D STEAM PIPES 4- V> /^ \^ Q , where t = thickness in inches. a s = tenacity in pounds per square inch. d = pipe diameter in inches. Good ordinary metal has the following values for s : Cast Iron = 15,000. Copper (cold) = 30,000. Wrought Iron = 49,000. Mild Steel = 60,000. In modern practice with superheated steam copper must not be assumed to have a tenacity above 16,500 pounds, and brazing at superheat temperatures becomes rotten. In Admiralty practice copper pipes are wound with steel wire close laid as a precaution against ripping. Mr. Ferranti, to avoid danger from large copper pipes, built up large pipes of a number of small pipes closely spaced in flange plates, which when bolted together gave a large number of pipes in cluster, but the system was expensive. Solid-drawn copper pipes are said to be liable to longitudinal splits. The ductility of copper is not great. When pulled apart by tension its reduction of area at fracture is small, and copper has lost any superior value it once possessed, perhaps very properly as compared with cast iron, in comparison with which copper first gained its character for elasticity and safety, 34 MATERIALS Mr. W. E. Storey states that a common cause of deterioration of copper is its contact when hot with reducing gases, such as coal gas, which makes the metal brittle. The same result is produced in the brazing hearth, when the air supply is insufficient. Such copper is properly to be termed gassed, rather than burnt, and this would tend to explain a fact and avoid a danger. He attributes failures in steam pipe to improper design, and urges solid-drawn tubes for bends with a radius at least three diameters of the pipe. He deprecates severe hydraulic tests. Though he is a maker of copper pipes, his advocacy is far from urgent, and engineers would be well advised to avoid copper for steam pipes, and especially superheated steam pipes, but copper may still be well employed for pipes containing water, such as feed pipes, the spring piece of a boiler blow-off not on the boiler side of the blow- off cock, however. Electro-deposited copper pipes .are said to be 50 per cent, stronger than ordinary copper. The author has used such copper only in water work, and cannot speak to its use in steam work. FLEXIBLE METALLIC PIPES. Flexible pipes are made by coiling into a closely interlocked helix a peculiarly folded strip of metal. This may be steel, zinc, brass, copper or a bronze alloy. These pipes are suitable even up to 300 pounds steam pressure. They are very flexible, 35 STEAM PIPES and are maintained steam tight by means of a pack- ing of asbestos, which is the more tightly held in the folds of the helix the greater the pressure inside the pipe. Bronze is considered best for steam pipes and these pipes are particularly suited for rapidly connecting boilers and engines on contractor's or temporary work. Flanges are attached by means of a screwed gland and collar, with asbestos packing, which holds firmly on the ridges of the pipe. Any gap in a length of pipe can readily be made good by a piece of flexible pipe slightly long. It will accommodate itself to any flange angle. The author has no knowledge to go upon as to long con- tinued durability, but there can be no doubt that it would form a perfect connection between a boiler and the main steam pipe. If used where it is not supported at each end it might be advisable to restrain end movement, in order to keep the flange connections free from tension. Where a bad foundation causes settlement and undue strains, a short length of flexible pipe would prevent all trouble It is well engineers should bear this flexible tubing in mind, for at times it may prove useful and of marked convenience. WROUGHT IRON AND STEEL PIPES. Pipes of steel up to 10 inches diameter can be had in weldless steel. Above that size, as well as below 36 MATERIALS TABLE X!A. WROUGHT STEEL PIPES, WITH WROUGHT STEEL FLANGES. Internal diameter } 1 1 2 9* 3 3i A of Pipe in inches J J . 2 ! i Thickness of Pipe \ in inches . . J 9W.G. 8W.G. 6W.G. * * i i i Weight per ( Ib. foot of I 1-48 1-89 3-31 5-0 5-5 8-5 IO-O II-O Pipe 1 (approx.). Ikilos. 673 86 1-51 2-3 2-5 3-86 4-55 5-o Weight per f Ib. 2-8 5-8 6-9 17-0 20-0 25-0 26-0 38-0 pair of J Flanges | (approx.). [kilos. 1-27 2-63 3-09 773 9-IO n-35 n-8 17-3 Internal diameter "1 of Pipe in inches J 5 6 7 8 9 10 12 14 Thickness of Pipe \ l 1 1 a 3. in inches . . J 4 32 8 8 Weight per f Ib. I4-0 21 24-5 27-5 35-5 41-0 49-0 67-0 foot of 1 Pipe | (approx.). Ikilos. 6-35 7-5 n-i 12-5 16-1 18-6 22-3 30-4 Weight per C Ib. 42-0 53-o 84-0 80-0 96-0 130-0 162-0 154-0 pair of I Flanges j (approx.). Ikilos. 19-1 24-1 38-2 36-4 43-6 59'2 73-6 70-0 i 37 STEAM PIPES it, pipes can be had lapwelded. There is always some little doubt remaining as to the absolute sound- ness of a weld, and where such doubts are felt the riveted pipe may be relied on. The riveted pipe, when reasonable care has been taken in choosing good material, can be made to show a strength 70 per cent, of a solid pipe, and it can be relied on. Steel pipes and steel flanges are made of all forms, including straights, quarter-bends, quarter-bends with a length of straight, double quarter-bends joined by a bit of straight, and set-off or cranked lengths. Table XlA. (see previous page) will be useful in getting out approximate weights of pipe. It is copied from a list of the Babcock Co. Steel for pipes should be of strictly mild quality, similar to boiler plate, with a tenacity of 24 to 27 tons per square inch and an elongation on tenacity test of not less than 20 per cent, in a length of eight inches. The flanges of steel pipes are attached by three usual methods, viz., riveting, welding and screwing. In the practice of the Babcock Co. pipes below 6 inches are screwed into their flanges (Fig. n) and expanded by a tube expander. The Whitworth thread should be used. It has n threads per inch, in all sizes above \" . The coarser American thread does not produce such good work or so tight as the English or Whitworth thread. Pipes are faced and drilled after fixing the flanges. Above 6 inches the Babcock Co. rivet on the 38 MATERIALS flanges (Fig. 12), and they frequently also rivet on branches, as in Fig. 13. FIG. II. FLANGE SCREWED ON. FIG. 12. FLANGE RIVETED ON. BABCOCK & WILCOX CO. The flanges of steel pipes are stamped out of solid forged pieces, and weigh as in the annexed Table 12 for high pressure pipes, Table 13 for FIG. 13. RIVETED BRANCHES. BABCOCK AND WILCOX CO, lighter purposes, and Table 14 for heavy cast-iron flanges. Riveted pipes are usually double-riveted lap- jointed scarfed down and tucked into the flange, 39 STEAM PIPES as in boiler-making practice ; the longitudinal seams are thinned at the overlap and tucked into the ring seams. The Mannesmann Co. of London make solid rolled steel tubes up to 12-inch external diameter, from 0*104 inch thick in 2-inch tubes up to 0-312 for sizes above loj. These pipes can be had with flanges FIG. 14. SOLID WELDED FLANGE (YATES & THOM). attached by any approved method. They also manufacture pipes, the flanges of which are loose and are slipped on to the pipes, which are afterwards flanged or lipped up in various ways. These flanges, or lips, are drawn together by bolts through the loose heavy flanges and are offered for use with high pressures and superheated steam. Steel pipes are also made with their flanges welded solid with the pipe, as shown at Fig. 14. The steel pipe is par excellence the proper pipe for high pressure and for superheated steam. The tenacity of steel at 60,000 pounds per square inch 40 MATERIALS gives steel pipes of ordinary thickness a very large margin of strength, a 6-inch pipe at 200 pounds pressure, and J inch thick, only carrying a unit stress of 2,400 pounds per square inch, or a 25-fold margin when solid drawn, and perhaps 15 to 2O-fold if lap welded. The best practice for junction pieces is to make these of wrought steel, but this becomes FTG. 15. HEADER OF CRUSE CONTROLLABLE SUPERHEATER, SHOWING METHOD OF MAKING JOINTS. x" expensive in the large sizes, and they are often made of cast steel, of similar pattern to cast iron, but they need not be so stout. The medium dimensions of cast-iron pieces should be ample for cast steel. Some engineers consider it quite good practice to employ cast-iron junction pieces, which, if stout 41 STEAM PIPES and carefully cast of strong metal, they look on as safe for the highest pressures. Perhaps they are safe when due provision is made to relieve expan- sion stresses, but their chief danger is perhaps from the sudden shock of water hammer. For super- heated steam steel pipes are most desirable, especi- ally in the superheater itself. As an example of the highest class of pipe work the steel pipes of the Cruse Controllable Superheater (Fig. 15) may be cited. These pipes are usually 6 inches external diameter, and T Vinch thick. They are of solid rolled weldless steel. Their extremities are staved or thickened up for threading, so that the diameter at the base of the thread is a little in excess of the external diameter of the pipe body. The staved portion is threaded, and the pipes are simulta- neously screwed at both ends into headers of i J-inch rolled steel plate. They are then expanded. The cover box which encloses the ends of two pipes with the coupling box of the internal 2 -inch water- control pipe is of pressed mild steel, and the joint between cover and header plate is made by means of a solid ring of -nrinch round copper wire. These joints have never been known to fail, and similar joints may be made between ordinary faced flanges by means of copper wire. The writer has made them with f -inch wire only looped into a circle and the ends simply crossed over each other, the bolts tightening the wire sufficiently to flatten the crossing to a steam-tight condition. 42 MATERIALS TABLE XII. HIGH PRESSURE STEAM FLANGES, PATTERN A. Internal Diameter of Pipe. Inches. Outside Diameter of Pipe. Inches. Diameter of Flange. Inches. Approximate Weight. Lb. Kilos. ,O 2f 7 7i 3-49 2i 2f 71 9 4 3 3i 8} I2i 5'5 3i 4 9 I2i 574 4 41 10 i6J 7'43 5 54 ii 22 9'9 6 6f 12 25 11-25 7 7f 14 43 19-35 8 8f 14 35i 16 9 9H 15 46f 21 10 IOJ 17 58 26 12 I2f xgi 72| 3274 14 I4J 2li 87i 39-38 TABLE XIII. LIGHT WEIGHT STEEL FLANGE, PATTERN B. Internal Outside Diameter Approximate Weight. diameter of Diameter of of Pipe. Inches. Pipe. Inches. Flange. Inches. Lb. Kilos. I i* 3J I 45 I I& 4i 2j I-I3 1} 1$ 4i 2 i 1-02 ij ij 5 2| 1-24 2 2J 6 41 2-14 2\ 2 f 7 6J 2-81 3 3i 8J 91 4-28 4 4i 91 14! 6-64 43 STEAM PIPES HEAVY PATTERN CAST-IRON FLANGES, FACED AND DRILLED, FOR AUXILIARY, STEAM, FEED AND BLOW- OFF PIPES. TABLE XIV. CAST-IRON FLANGE. Internal diameter of Outside Diameter of Diameter of Approximate Weight. Pipe. Pipe. Flange. Inches. Inches. Inches. Lb. Kilos. t I* 31 2 9 I I* 4i 3i 1-6 it if 4i 31 i-5 ij U 5 31 i-5 2 2| 6 61 3'i 2j 2j 7 8i 4 3 3i 81 14 6.4 4 4i 91 18 8-2 STEEL ALLOY. Some boiler makers will provide what they call steel pipes, which are really malleable iron, or so- called steel alloy. They are often tough, but are difficult to face, and are apt to suffer from bad spots, which leak out steam. No doubt such pipes are much superior to cast iron, but they are not equal to steel pipes with either screwed or welded flanges. 44 MATERIALS Though Tables XII., XIII., XIV. are given as the Babcock standards, it may be doubted if it is worth while having flanges of different diameters for auxiliary pipes, etc. At Fig. 16 is shown the riveted flange as made by Yates & Thorn, with the recessed face and projecting shallow spigot, which, while it is FIG. 1 6. RIVETED FLANGE ( YATES & THOM). apt to make it difficult to pull pipes apart, is very efficacious in preventing joint rings from blow- ing out. Pipes put together with indiarubber rings in these recesses are sometimes very diffi- cult to part. Plain flanges with sheet joints can be sawn apart with an old saw, which will clean off the flange faces ready for a new sheet to be inserted. 45 STEAM PIPES SOCKETED JOINTS. A form of joint of extreme neatness not much employed is the socketed joint. Sockets are usually screwed upon the pipe as tightly as possible, coming to a stop where the thread dies away into the barrel. The next length of pipe is screwed upon the pre- viously erected length, and it is obvious that the last length must be flanged at one end, for it cannot be screwed two ways at once. But the flange is not always even possible, and socketed pipes are joined finally by means of a " long thread " or " connector/' which consists of a piece of pipe of any length, one end of which is threaded a long way so that the threaded portion will hold the full length of a socket as well as a back nut. In making a final joint by this long thread the socket is fully screwed back, the ends of the pipes to be joined are brought together, and the socket is then screwed forward upon the end to be joined up. For each thread it advances on one pipe it leaves the other, and of course is somewhat slack upon the long thread. The back nut is then screwed against the socket, a thread of asbestos with cement being wrapped round the thread and forced tightly against the socket end by the back nut. There should be a back nut at each end of the socket for steam work. Artesian pipes are faced off to dead lengths of 10 feet, or other desired lengths, and have special sockets which screw exactly half length on each 46 * MATERIALS pipe and tighten up on the last of the thread just as the faced-off ends meet at the middle of the socket. Such pipes could be made to form a steam- tight joint by means of a ring of copper between the ends. It is possible that socketed pipes will come more into use for special work, as they are very neat, offer the minimum of surface for loss of heat, and can be covered with sectional or other covering to look very neat, having no flanges. The long thread affords every facility for making up lengths exactly. It is, however, certain that sock- eted pipes are troublesome to take apart. The sockets become very fast. They are best put together with Dixon's smear-grease, a compound of mineral oil and graphite, which is said not to become hard. As regards wrought-iron pipes these are lap- welded for steam purposes, are to be treated as described for steel, and it is the author's belief are often supplied of steel to fill wrought iron orders. FLEXIBLE METALLIC TUBING. For temporary work, flexible metallic tubing coiled up from pecular doubled or folded steel strip, interlocked and flexibly packed with a thread of asbestos or other fibrous matter, may be employed. It may be obtained attached to flanges, and for rapid connection is easily put in to occupy the place of making-up lengths not yet arrived from the makers. 47 STEAM PIPES 1 1 is best made of bronze for steam purposes, as to which a further note is made under the head of " Copper/ American pipe of ij and up to 2-inch sizes is screwed nj threads per inch. Above that size it is screwed 8 threads, which seems coarse to English engineers, and is not so good as our Whitworth ii threads for light work. The following is the American pipe list abridged for lap-welded wrought-iron pipe : Inside Outside Weight Threads Diameter. Diameter. per Foot. per Inch. ii I-900 2-68 "J 2 2-375 3-6o Hi 2l 2-875 5-73 8 3 3-500 7-54 3i 4-000 9-00 4 4-500 10-66 4i 5-000 12-34 5 5-563 14-50 6 6-625 18-76 7 7-625 23-27 8 8-625 28-18 9 9-625 3370 10 10750 40-06 ii 12-000 45-95 12 12-750 49-00 13 14-000 54-00 14 15-000 58-00 16-000 61-77 18-000 70-00 20-000 77*57 22-000 85-47 ~ 24-000 93-37 , 48 MATERIALS American pipes are screwed with a taper per inch of length of screw of & up to 8 inches diameter and -gr above that size. As with English pipes, the inside diameter of American w.i. or steel pipe is not the nominal diameter, but varies as the thickness of the pipes varies, the outside diameter being constant for any nominal size. WHITWORTH PIPE THREADS. Internal Diameter. External Diameter. Diameter at bottom of thread. Threads per Inch. 4 826 734 r 4 1 1-04 949 *4 I I-309 1-192 ii rj 1-650 1-533 ii ii 1-882 1765 n 2 2-347 2-23 ii 2j 3-00 2-882 ii 3 3'485 3-368 ii 4 4-340 4-223 ii Pipes are not made exactly to their nominal inside diameters. All pipes, whatever their strength, have equal outside diameters for the same nominal in- ternal diameter. Any change of thickness adds to or subtracts from the inside dimensions. The outside diameter of English and American pipes differ very slightly. Messrs. John Spencer, Ltd., say that for general 49 E STEAM PIPES work there is nothing to beat lapwelded steel pipes, with solid welded flanges, and branches riveted on, for sizes from 2-in. bore to 12-in. inclusive. For larger pipes riveted flanges are preferred, and for low pressure and small pipes, screwed flanges. This firm's list of standard flange diameters, drilling, etc., and also thickness of pipes for both high and low pressure steam main work is annexed. The thick- ness of pipes is given for straights ; bends are always made somewhat thicker : TO 120 LB. PRESSURE. Bore. Diameter. Thickness. No. of Holes. Diameter of Pitch Circle. Size of Bolts. iin. 3iin. i in. 4 2} in. -ft in. 1 3l i ,. 4 2| TV i 4i 4 .. 4 9t:N i Iin 5 i ., 4 3f i Ii,, 5i >, 4 4 4 i 2 6 1 4 41 i 2 6| 1 4 5 i 2i,, 6J,, 1 .. 4 5 ,> 4 2j 7 1 6 5i >, 4 3 8 i ,. 6 6J,, 1 3l 8i,, i,, 6 6| i ,, 4 ,, 9 i 6 7i i 5 ,, ioj J ,. 6 8i,, i -. 6 12 i 8 10 1 ,, 7 134 i 8 nj i,, 8 15 i*,, 8 I2f J 9 ,, 16 tt i*,, 10 I3l t 10 17 i*,, 10 I4J J ii 18 ii,, 12 I5l I ., 12 19 if,, 12 i6| t ,, 50 MATERIALS TO 200 LB. PRESSURE. Bore. Thickness of Pipe. Thickness of Flange. No. of Holes. Diameter of Pitch Circle. Size of Bolts. i in. 10 g. J in. 4 2j in. w in. 1 9 *> ^ 5> 4 2| ,, i 8 ,, ii si ,, 2^ > Ij > 7 , M 31 ,, 1 2 6 ,, 4 ,. ,, 2 1 II 41 I M 2 ii 1 II 5 2 j > 5 i> ,, ,, 11 M 2g Jin. II 6 5i ., II 3 >, ,, 1 ,, 6J ,, 1 ># 31 ,, ,, 6 i II 4 >* ,, ,, ,, 7i > ,, 5 ii ~5 " 8 8i,, II 6 ii * . ii 10 5) 7 ,, 112 ,, 8 "fo >f IB 12} I 9 . II M 10 X 3i || 10 ,, ,, 14! ,, ,, ii 8 i> *$ > 12 15} ,. II 12 II If',, 12 16} " Flange diameters as on previous list. For making steam joints, Taylor's corrugated brass rings give the best result, and they also strongly advise the adoption of a facing strip -gi-in. deep on the flanges of all high pressure pipes. The following empirical formula is given for get- ting out quickly and accurately the length of tube in a right-angle bend ; it is found to give excellent results : Take the sum of the two arms and deduct 5 inches for every foot of radius, plus f inch for STEAM PIPES stretching. Thus, for example : a bend setting 6 feet with a radius of 3 feet 6' o" + 6' o" = 12', less i' 3f" = 10' 8J". The lengths of tube in the pipe will be 10' 8J". In getting out steam mains, the chief point to watch is free expansion, which is obtained by using a sufficient number of lapwelded steel bends of large radii, or by the insertion of special expansion pieces. The rate of expansion of steam pipes is taken by Messrs. Spencer as follows : Copper. Steel. C. Iron. W. Iron. 012 in. -00822 in. -0077 in. -0082 in. per 10 feet for a rise of 10 degrees Fahr. The number of joints is to be minimized by using as long pipes as possible. Appended is a list giving what may be considered stock lengths for various sizes of pipes. Short bends and cast steel elbows should not, of course, be used when they can be avoided, nor should cast iron be used for bends at all ; in fact one should always endeavour to eliminate it from steam main work, especially where there is high pressure and superheated steam, as cast iron de- teriorates very much when carrying superheated steam. Two other very important points are the proper supporting and draining of ranges ; if the former is not well done vibration will ensue, and if the latter is insufficient water hammer is set up ; excessive 52 MATERIALS vibration will cause leaky joints, and may lead to very serious consequences. The following formulae for thickness of pipes are given : t = thickness of pipe in inches. p = pressure per square inch. d = diameter (internal) in inches. For lap welded wrought iron, t = g d_ 3500 j) d Copper (brazed) t = cast-iron, t = f + i> d (solid drawn) t = g These are as used by the Board of Trade, and hold good, generally speaking, for bores up to 12 -in., and water Pipes of 200 Ib. per sq. in. Another very fair formulae for cast-iron up to 200 Ib. per sq. in. water-power, is t = ^ - + \ ] this gives somewhat heavier castings. A good formulae for thickness of flanges on cast- iron pipes is : T = 1-4 x t x -15, where T = thickness of flange, For bolts : d = -83^ + -3 d = diam. of bolt, n = *6D + 2 n = number t = thickness of pipe. D = diameter 53 STEAM PIPES The following bursting pressure of steel pipes is given : Diameter, Pipes T 1_ 6 in. 7 in. 8 in. 9 in. 10 in. ii in. 12 in. in Inches. Thickness. i i n - 1600 1372 1:00 1066 960 8;3 800 * 2400 2058 1800 J 599 1440 1309 I20O 1 3200 2744 2400 2132 I92O 17 5 1600 "& > 4000 3430 3000 2665 2400 2181 200O ! " 4800 4116 3600 3198 2880 2617 240O iS > 5600 4802 4200 373i 3360 3053 2800 i ,, 6400 5488 4800 4264 3040 3489 3200 13 in. 14 in. 15 in. 16 in. 17 in. 18 in. 19 in. Jin. 738 685 640 6OO 56i 533 505 f 2583 2401 2240 2100 J 975 1866 i? 6 5 i 2952 2744 2560 240O 2257 2133 2017 Stock Lengths, 16 to 17 ft. up to 10" diam. 15 to 16 n" to i2 /r diam. 54 CHAPTER IV Expansion 'TVHE necessity of cooling-off boilers for cleaning * and repair, and the fact that some boilers are spare and cold, causes the connecting pipes of the boilers to the main to vary in length, not only as between their hot and cold conditions, but as between one boiler and others. In a length of 100 feet a steam pipe may expand 2 inches or more. The coefficient of expansion of cast iron is 0-00000618 per degree Fahrenheit =0-0000111 per i C. Wrought iron expands 0-00000656 per iF. = 0-0000118 per i C. Between the temperature at which the pipes were fixed, say, 59 F. = 15 C., and the temperature of steam at 190 pounds gauge pressure per square inch, say, 383 F.= 195 C., the expansion will be about 4^ inches per 100 feet. Mr. Yenning says a good practical rule is to allow i inch for each 50 feet. Beyond the superheater the expansion effects will be even greater, for the temperature may be 653 F. = 345 C., a rise of nearly 600 F. = 330 C., or nearly 6| inches per 100 feet above the cold length, for at high tempera- tures the expansion coefficients become greater (see table, p. 68). 55 STEAM PIPES Obviously, therefore, such variations must be provided for. In the case of a long battery of boilers with a straight main steam pipe athwart them, if the middle of the pipe was anchored fast the two ends would extend considerably. This would push outwards the branch pipes of the boilers to an amcunt gradually increasing as the distance from the central point was increased. FIG. I/. FIG. 18. FIG. 19. FIG. 20. FORMS OF EXPANSION BENDS. FIG. 21, The branch pipes would give way by their elas- ticity, but they would of course exert a twisting stress upon the mounting block, the vertical pipe above this, and the valve, if this was at the top of the vertical branch. In long lengths of main, there- fore, bends of the form of Figs. 17 to 21 are employed. 56 EXPANSION If circumstances are such that one of these expan- sion lengths is to be fixed, its place should be as nearly central as possible in the main, which would be an- chored, if anchored at all, at one-fourth its length from each end, thus dividing it up into four dis- tinct lengths, and limiting expansion in any one section to one-fourth the total. The anchoring of a pipe in this way prevents the steam pressure on the extreme ends of the pipe from acting to pull open the expansion bend, but if anchored in such a way as to prevent lateral movement there would be introduced a stress in the nearest boiler branch pipes. Anchoring, therefore, should be longitu- dinal only. In arranging for expansion bands the loop should, if possible, be horizontal, so as to ob- viate water pockets. If vertical, there must be a small connecting pipe looped down from the straight main to carry water across the gap of horizontal continuity. The expansion bend cannot be allowed to hang downwards from the pipe unless the bottom of the loop is drained by a trap, and this position must not be used, if possible to be avoided. Figs. 17, 18, 19 are usual types of bends. Fig. 20 is convenient where there is a change of level greater than the pipe diameter, for the lateral displacement of the loop may be little or considerable. In Fig. 15 the stress, says Mr. Stromeyer, is tor- sional. 1 1 The Manchester Steam Users' Association. Memorandum by Chief Engineer, June, 1901. 57 STEAM PIPES A bend is more elastic than a straight of eqnal length from crown to crown. Thus, if in Fig. 17 the two straight arms were connected by a rigid casting in place of by a bend, this form would be the most rigid of all the arrangements shown. Mr. Stromeyer represents the elasticity of the two straight arms by 2. Then each of the forms, Figs. 18, 19, 20, will spring an amount 2 x 6 = 12. Fig. 17, con- sisting of two-thirds straight pipe and one-third bend, will be represented by 2f . In Fig. 21 the elasticity is 21. The permissible stretch of any form varies with the mean height of the loop, is inversely as the square of the diameter and independent of the thickness (in all practical thicknesses). The force to stretch a bend is proportional, how- ever, to the thickness and to the diameter squared. Bends are therefore made thin and weak if they have to relieve stress on a weak point, such as a cast-iron valve or junction piece, but for expansion of long pipes the bends may be of the same material and thickness as the pipes. Copper, once so much used for bends, is not so very suitable, though it may be made thin. Its elastic limit is low, and it has less spring than mild steel or wrought iron. It is a metal that grows brittle with age, and it is dangerous at high tempera- tures. With bends of 4 feet crown to crown, and a dia- meter of pipe of 6 inches, Mr. Stromeyer gives the following (Table XV.) of safe extensions of bends : 58 EXPANSION TABLE XV. Material. Two Straight Pipes. Fig. 17. Figs. 18, 19. Fig. 20. Fig. 21. Steel . Copper . 0-21 O-I2 0-42 0-23 074 0-4I 0-74 0-41 2-60 i*45 The values in the table, except for Fig. 21, may be doubled where the pipes they relieve have free- dom for lateral movement, and, again, this double value may be again doubled if the bends are ini- tially stretched by the same amount they will be compressed when hot, so that a copper bend of Figs. 18, 19, 20 type would, if erected cold and stretched 0-82, allow of a total difference of length between cold and hot of 1*64 inches, or enough for a length of 50 feet of main. In a battery of four boilers, as very commonly arranged in cotton mill work, see Fig. 22, as ar- ranged by Yates & Thorn. The branch pipes of the boilers are about 12 feet long from crown of steam mounting block to crown of main steampipes. The only relief necessary here is given by the two bends and long straight piece of pipe between the boiler main and the engine, the starting valve of which is not in line with the boiler main. A similar double bend connects the high-pressure cylinder with the first initial-pressure cylinder. Expansion joints are sometimes made of the form 59 FIG. 22 GENERAL ARRANGEMENT OF COTTON MILL PLANT (YATES & THOM). 60 EXPANSION of Fig. 23. These are small and compact, but are apt to become choked with deposit and to split. Larger expansion discs are made by riveting to- gether two flatly dished plates. These are also liable to choke with deposit, and they also offer a large surface to the steam pressure, which exerts FIGS. 23, 24. EXPANSION JOINTS (YATES & THOM). a very heavy thrust and helps to nullify the move- ment of pipes when these are wanted to contract. They resist the very movement they are designed to accommodate, and they are therefore only to be recommended for exhaust pipes, in which case the outside pressure exceeding the inside pressure tends to move the pipes in the same direction as the ex- pansion will move them 61 FIG. 25 END VIEW FIG. 25. TELESCOPIC EXPANSION JOINT (YATES & THOM). 62 EXPANSION Expansion joints of the form of Fig. 25 are some- times used. This particular one, as made by Messrs. Yates & Thorn, consists of a sliding pipe and stuffing box to provide movement on each side of the joint the pipe has legs upon it which are joined across the joint by two long bolts that must be strong enough to resist the steam pressure on the area of the pipe. In a length of pipe served by such a joint as this the bolts would be of a length FIG. 26. BARTER'S SWIVELLING EXPANSION COUPLING. equal to half the length of the pipe. This class of expansion joint is only advised where spring bends cannot conveniently be introduced. Shorter ex- pansion joints are made complete in overall lengths of from 27 to 32 inches, according to size, i.e. about 20 + 1-5 d inches, and weighing from 75 to 80 pounds per inch of pipe diameter, according to size, the larger ones being of the heavier proportion. Barter's joint is shown in Fig. 26. This explains 63 STEAM PIPES itself. The joint provides a universal movement of a pipe in the way of bending, but this joint is not an expansion joint to be placed in the length of a pipe. It is rather to be placed in the branch pipes of each boiler, so as to allow free movement of the main pipe without straining of the branch pipes. Where economy is specially desirable these joints would not be necessary on the first (or perhaps also the second) boiler, on each side of the middle point of the main. If this joint is so arranged that when cold it lies in a straight line, or nearly so, with the boiler branch pipe, it is obvious that when the main steam pipe lengthens, these swivelling joints will be displaced laterally, and will be no longer straight and in line. The flanges they connect will therefore need to approximate each other by the amount of the versed sine of the arc of swivelling. As, however, the boiler branch pipe becomes longer also by heat expansion, this angular movement of the swivel piece will provide to some extent for this expan- sion. Thus, in a swivelling length of 30 inches a movement of i inch in the main steam pipe would imply an angle of 2 degrees, the versed sine of which is -0006 or 0*018 inch. This is only about one- eighth of what a branch pipe of 5 feet in length would expand at usual pressures. But if the swivel piece be already placed at an angle of 4 degrees, when cold, its movement only 2 degrees further would increase the versed sine movement by -0048, or eight-fold. Obviously, therefore, a boiler branch 64 EXPANSION of a length of 5 feet, with a 3O-inch swivel-piece placed when cold 2 inches out of line, would allow for a movement of the steam main of i inch when hot, and would take up the expansion of the boiler branch. If boilers and engines are carefully fixed to drawn positions, as they may be, and pipes are made to the same measures so as to come right without final make-up lengths, as is also not merely possible but practicable, then it would be possible so to arrange the whole scheme of pipes as, by plac- ing the swivel at greater initial angles towards the end of a range of boilers, to eliminate ail stresses of expansion. Even if such stresses were reduced to half or a third, it would be a desirable thing to accomplish. In many power stations the boiler branch pipes enter the main steam pipe, and the engine branch pipes are taken out from points very near to them. Where possible, expansion is provided by bending the boiler branches so as to enter the steam main at the top. If the engine pipes leave the main from its upper side also, the main acts as a water separator, and must be drained. If the engine branches leave from the bottom of the pipe, all water must then be dealt with at the engine separators. Both the boiler and engine branches may enter at the opposite sides of the main without bends. In this case the engine branches are usually bent down to the engine some feet further on, and the separator is placed in the horizontal part of the engine branch. With water-tube boilers, where there is a clear gangway 65 F STEAM PIPES behind the boiler seating, the boiler branches have been brought down by bends to the steam main placed in the gangway, and the engine branches have been carried up from the main, bent through the engine-room wall, thence carried a few feet, and bent down to the engine stop-valve. This arrangement is very elastic, because the various vertical pipes are several feet long. The main must be carried above the passage ways be- tween pairs of boilers, at least 6 feet above floor. It must also be drained. In case of very long mains the expansion, if not otherwise provided for, may be allowed for by an expansion T-piece. The one pipe has a closed end and passes right through the head of the T, being made steam-tight by glands at each end. That part of the pipe inside the T is perforated by slots to permit steam to pass to the T and to the pipe, which is rigidly bolted to the single end of the T. The expansion of the long pipe can take the place of sliding, and there is no end- thrust. The disadvantage is, that the steam has to take a sharp square bend and pressure is lost. Sub- stantially the provision of suitable bends and suffi- ciently long branches is alone necessary for general work, and a scheme of pipe work must be carefully thought out so that expansion shall not be con- centrated at one point, but shall be well distributed throughout the system, bearing in mind always that any one boiler may be at rest between two working boilers, or vice versa ; and due consideration 66 EXPANSION must be given to each movement that will occur under the extremes of conditions. The expansion of any other bends than those of 4 feet x 6 inches given in the table, page 59, can be calculated by the rule given, or E = e x 4 **j or E = - -i * 2 '-, where d is the diameter of pipe in inches, H l the loop height in feet, and e is the tabular expansion for 6-inch pipe, E being the per- missible expansion of the pipe sought. Then ,as found, may be doubled or quadrupled, according as the pipe is free for lateral movement, or is extended when cold. The calculation of the expansion of any length of pipe is made by the following formulae : S = L t /, where S is the movement sought, L is the length in feet, t the range of temperature over which the expansion occurs, and / is the coefficient of expansion per foot of pipe per degree of temperature. In a long main taking branches from several boilers, the middle of the main may, as stated, be anchored fast. This is easily effected when the pipe is carried on a bracket, as a cap may be bolted over the pipe, but not so as to prevent lateral movement. At other points the pipe may be suspended by rods from brackets above, and a short spring may be placed between the bracket and the nut of the suspender. Anchoring of one point is desirable for the pur- pose of checking the vibration which is often set 67 STEAM PIPES up by the connection of the pipes to the engines, or by the intermittent impulses of the moving steam. A pipe which swings in this way may usually be steadied without locking it fast, if a stop is placed against a point of chief movement to limit the ampli- tude of the vibration and destroy the natural rhythm of the movement. EXPANSION COEFFICIENTS. The expansion of cast iron between 32 and 42, or a range of 180 F., is given in Molesworth as o-ooii, wrought iron, 0-0012, and copper, 0-0018. Kempe gives the coefficient of linear expansion per degree Fahr. as follows : Metal. Coefficient. Tested Between. Cast Iron . 0-OOOOo6l8 32-2I2 Steel . . . 0-OOO006OO 32-2I2 Wrought Iron . O-OOOO0895 32-572 0-OOO00656 32-2I2 Copper , . j 0-00000955 32-2I2 j 0-OOOOIO92 32-572 Firebrick 0-00000275 32-2I2 Good Red Brick 0-00000305 32-2I2 From this it wculd appear that at temperatures of superheated steam the coefficient of expansion per degree Fahrenheit for steam pipes may be taken as o 000008, which is nearly 2 inches in 50 feet for a temperature rise of 400 F. While undoubtedly stresses are often very severe and manifest themselves by failures of cast-iron 63 EXPANSION junction pieces, and by weeping joints and even rivets, it must not be necessarily inferred that all the movement calculated does actually occur to produce stress. Much may be done in the way of giving counter-stresses initially, and cold to reduce the working stresses when hot. Nor can we assume that the boiler does not move from its cold position. The expansion of firebrick is about half that of iron, and under boiler conditions it is much hotter and its actual expansion is as much or more. The seating of a Lancashire boiler lengthens as much as the boiler, the movement one way of the steam outlet blcck on a boiler may be as much as the ex- pansion the other way of the steam pipe. An ultra precision and refinement is therefore not called for, but it is easy to see that the expansion of the boiler branch pipe may be very fairly balanced by com- pelling the boiler steam-drum to expand from a determined and anchored point. Practice has taught what may and what may not be done, but special cases may require special consideration, and for these the methods indicated may be followed. The use of superheated steam not only increases the pipe temperatures but also increases the co- efficient of expansion, which (as per table p. 68) becomes 0-000009 nearly between cold and 572 F. Such an expansion as this figure implies is very considerable. At the same time, probably, the pipes and bends are more yielding and take up the stresses by further movement. Where super- heaters are placed behind the boilers, as in case of 69 STEAM PIPES Lancashire type boilers, the pipes to the super- heater have only the ordinary expansion. But n .ct M Temperature Q 70 EXPANSION f the superheater is connected with the main, and this is to be considered in the design. The annexed diagram will be of use in ascertaining the temperature of saturated steam from the known pressure. With the prospect of superheat being added, the temperatures found by the diagram should be increased by about 150 F. when calcula- ting probable expansions to be provided for. CHAPTER V Strength of Pipes THE thickness of a steel or wrought-iron pipe necessary for screwing is more than sufficient for all ordinary sizes at high pressures. The stress on the material of a pipe per inch of length is the product of the diameter and the pres- sure per square inch. This product, divided by twice the thickness of the pipe, gives the unit stress. Good wrought iron may be assumed to have an ultimate strength of 20 tons per square inch, and steel of 28 tons. On this basis, with a marginal factor of 5, the stress permissible will be about 9,000 pounds for iron and 15,000 pounds for steel. Double riveted joints have a 70 per cent, efficiency, and if lapwelding be allowed to have the same, the unit working tenacity will be 6,300 pounds and 10,500 pounds respectively. Thus a 12-inch pipe at 200 pounds, if only J-inch thick, only carries a unit stress of 4,800 pounds. For very large work, even to 24-inch pipes, the stress on pipes J-inch thick would only be 9,600 pounds at 200 pounds pressure, or within the working stress of mild steel. Steam pipes of ordinary manufac- 72 STRENGTH OF PIPES turer's thickness of tube walls are thus of ample and excessive strength when only double riveted. Bad welds should be provided against by hydraulic test up to 12,000 pounds in iron and 21,000 pounds in steel, as calculated on the actual thickness, which will usually exceed J-inch in pipes of even 10 inches diameter. Very large pipes may be worth riveting with butt strips. Solid rolled pipes can be calculated to stand a unit stress of 15,000 pounds, and need not exceed the thickness proper to this stress so long as the threading at the flanges does not unduly reduce them and render them liable to crack off at the flange. Probably steam pipes are made too heavy, and being so they throw undue stresses on cast-iron junction pieces, which are therefore made unduly clumsy to stand the stresses. Solid rolled pipes with flanges double-riveted appear to offer the maximum strength per unit of weight. Flanges screwed on to large thin pipes involve the weakening due to the threading. The Whit worth thread is the best for pipe threads, as it is finer than the American thread and cuts less of the pipe away. Its main dimensions are given here within Table XVI., up to 4 inches. All larger sizes have the same thread of n threads per inch. Pipes of all kinds are of the same diameter outside. The nominal inside diameter becomes less as the pipe is made stronger. Threads may thus be all standard, but some pipe makers do not make to the Whit- worth standard even to-day. 73 STEAM PIPES TABLE XVI. WHITWORTH PIPE THREADS. Size. No. of Threads per Inch. External Diain. Diameter at Bottom of Thread. 4 14 0-8257 0-7342 1 14 I-04I 0-9495 I II I-309 1-1925 ii II I-650 1-5335 ii II I-8825 1-765 2 II 2-347 2-2305 A II 3-0013 2-8848 3 II 3-485 3-3685 34 II 3-912 3-7955 4 II 4-339 4-223 According to the Board of Trade, the strength of copper pipe, well made, with brazed joints, is, work- 1. T- J T\ , where T and D are 6OOO (T - mg pressure = - ^ the thickness and diameter in inches. If solid drawn and not over 8 inches diameter, the iV mc h is re- placed by sV-inch. Wrought iron lapwelded pipes of good material 6000 x T are allowed D = working pressure. It is well when pipes are screwed into their flanges to taper the hole on the face side of the flange and roll over the pipe end. This provides an extra longitudinal strength. Pipes too thin in relation to their diameters, if exposed to heavy end pull, will jump clean out of their sockets without showing any injury to threads. They cannot do this so well 74 STRENGTH OF PIPES with rigid flanges, but too much reliance must not be placed on mere screwing. Rivets, again, must be proportioned for shear to stand the maximum possible end stress in the pipes, which is not likely to be greater than the steam pres- sure multiplied by the " area of pipe." This " area of pipe " may be greater than the nominal area, for it is the area enclosed within the joint ring. Rivets may be allowed a working stress in shear of 10,000 pounds per square inch, and will always be found to have a large excess over the stress on even the largest pipes. Rivets, therefore, are pro- portioned for steam tightness. Longitudinal rivet seams, of course, are propor- tioned as in boiler work. CHAPTER VI Anti-Priming Pipes and Outlet Valves IT was formerly the practice to lead off the steam pipe from a boiler by way of a steam dome. But these being found not to abolish priming have been discarded, and the steam pipe is attached directly to a mounting block, which, as elsewhere stated, ought to taper, so as to allow steam to enter it without loss by vena contracta effect. Inside the boiler is fixed the anti-priming pipe. This is a length of pipe which usually extends each way from the steam outlet block, into which it is fitted by a short neck The. sides and top of the pipe are slotted with holes, the joint area of which should be 25 per cent, greater than the area of the steam pipe sup- plied. Each branch of the anti-priming pipe should have a diameter of about three-fourths at least of the steam pipe. The anti-priming pipe is held to place by hangers attached to riveted lugs on the boiler crown. Illustrations for ordinary practice are given in Figs. 27 and 28. It would be good practice to enlarge the central part of pipe so that it could be brought down at the bottom, and an easy leading 76 ANTI-PRIMING PIPES curve made to the outlet, so as not to oppose too sudden a bend at this point. The holes in the pipe should be confined to the upper quarter or third of the circumference, and it is usual to drill a drain hole at the lowest point to let out any water. There ought properly to be a drain pipe carried to below water level in order to free the drainage water from the rush of steam. FIG. 27. ANTI-PRIMING PIPE (YATES & THOM). I FIG. 28. ANTI-PRIMING PIPE (YATES & THOM). Anti-priming pipes are made about 6 feet long. Their object, of course, is to collect steam from a considerable length of a boiler, so as to avoid local rushes and formation of vortex whirls which would pick up water. They are sometimes made of copper of great length. The author has seen them ex- tended to a double or ring main of over 20 feet of thin copper, perforated or slit. This seems need- lessly long, for a sufficient spread of the intake can 77 STEAM PIPES usually be obtained in a length of 6 feet, and the area of holes can be got in to the requisite extent of 1-25 times the steam-pipe area. If this is exceeded, the steam will enter over too limited a length of the anti-priming pipe and defeat the object of the pipe. It is probable that excellent anti-priming pipes could be made from slitted brass sheet similar to the slit brass used for covering driven wells, but experiment is wanting to determine the amount of steam that will pass by a given length of slit. In the water sheets the slits are merely cuts without removal of material, and are very effective to keep back sand, while passing water much more freely than it could pass small holes of many times the area of opening. THE STEAM OUTLET VALVE. These have always been of the mushroom variety, and have necessarily been opened with or against the pressure in the boiler. When the valve opens against the pressure it can of course be easily shut, and the pressure keeps it shut. It possesses the fatal objection that when shut it depends on the strength of its spindle to withstand the steam pressure in the main steam pipe, where other boilers are at work. This is a fatal objection, because it endangers the safety of the boiler cleaner or inspector in the idle boiler. The shut-down type has the objection that in order to prevent it being opened when the boiler is at rest it must be loose on its spindle, so that it only opens 78 ANTI-PRIMING PIPES by steam pressure, and these loose valves float on the outgoing steam and keep up a constant ringing sound, which, however, is objectionable only as showing wear. The author would prefer the full-way double- faced slide-valve type to either of the other two forms, though this variety has the fault that it can be opened during the presence in the boiler of work- men, and ought to be specially safeguarded by a lock. The position of the outlet valve is important. If when placed close to the boiler the steam pipe extends vertically above the valve for any distance before bending away to the main, or as in some cases the main is immediately on the top of the vertical pipe, then the vertical length of pipe becomes a water pocket should the boiler be at rest. This necessitates the employment of a drain pipe taken off the valve body at the lowest point and fitted with an automatic steam trap, for it is too dangerous to trust to the opening of a drain cock, when perhaps a labourer is sent to open up the valve. The water in the vertical pipe would all be blown forward and produce a most violent con- cussion at the first square end, or at the blank end of the main, or at any interposed resistance. Drain pipes and steam traps are a source of loss at any time. Good practice eliminates both if possible. To do this at the stop valve the vertical pipe is placed directly on the boiler block, and the stop valve, if of mushroom type, is made an angle valve, 79 STEAM PIPES and placed at the top of the vertical pipe, or if it is a slide valve it is placed past the quarter-bend which follows the vertical pipe. Sometimes there is no vertical pipe, but simply a large radius quarter-bend instead. In that case the stop valve comes next after the bend. No matter how placed, the broad principle to be ob- tained is that the valve shall be dry, there being a fall each way to the boiler and to the engine or steam main. This ensures freedom from water disasters, and avoids the loss and annoyance of drains and traps. In the course of the nearly hori- zontal pipe between the stop valve and steam main, it is a fairly common practice to fix an automatic non-return valve, which is intended to safeguard a boiler should it be out of work. This valve pre- vents the boiler from absorbing steam from the other boilers should its pressure fall when cleaning, etc. It also prevents escape of steam should any failure of a boiler tube take place, and confines any escape of steam to the one boiler. The idea of this valve is excellent, but as usually made with a large rolling ball it is probable that if called on to act suddenly the ball would shatter the valve box, and the last disaster might be worse than the first. These check or non-return valves must be applied with caution. In the improved form made by Templer & Ranoe and described in the Chapter on " Valves/' the moving of a heavy weight through a long distance is avoided. 80 ANTI-PRIMING PIPES Danger may arise even when a pipe is being drained of water preparatory to opening the steam valve, especially if the water has become cold. Thus, a horizontal length of pipe below a main steam pipe may become full of cold water, and when partly drained so that the water-level is below the crown of the pipe and steam can enter above the long hori- zontal surface of the water, there wil be sudden condensation of some of the steam ; waves will be set up, and inevitably there will be water hammer and probably burst pipes. 1 The fact that a water hammer takes -effect chiefly at a valve, a bend, or a tee piece, emphasises the badness of the practice which permits such valve bodies or tees to be made from cast iron. In laying out a pipe system, there- fore, the principle to be observed is that of a steady progressive fall from the boiler stop valve to the engine. In addition to this, the question of elasticity must also be fully considered. No absolute fixed plan can be given, because the system must be varied to fit the boilers and their position relative to the engine In a bank of Lancashire boilers, when the steam outlet is central, the valve is sometimes placed directly upon the mounting block, and the steam branch curves out from one side by a quarter- round bend, and thence proceeds to the rear of the boiler at a height but little above the crown of the boiler, avoiding the manhole because of the lateral 1 See Manchester Steam Users' Association. Memorandum by Chief Engineer, June, 1901. Si G STEAM PIPES bend. Without this bend the pipe would be in- conveniently close to the manhole cover, and for this case the vertical branch is employed, the valve being high up and upside down (see p. 120). The horizontal branch connects to the steam main, which is supported by hangers, or by brackets, on the rear wall. The branch pipes from the boilers may enter on the side of the main, or by a downward bend. The length of these branches, often 15 to 20 feet, combined with the length of the vertical pipe, suffice to afford sufficient elasticity to take up stresses of expansion. Ranges of Lancashire boilers have frequently been fitted with steam main closely attached to the side outlets of the stop valves. This is an arrangement which provides too little elasticity and is not at all to be considered, especially for high pressures and temperatures. T CHAPTER VII Pipe Joints HERE are a wide variety of means of connect- ing pipes, though these may be classed under three main headings, as follows : : i (1) Spigot and Socket. (2) Screwed and socketed, or flush- jointed. (3) Flanged. The first named is named only to condemn the spigot- joint as unsafe for steam or hot water. It will draw apart should its supports fail, and should not be used. The second class is only to be used with sockets, and not in the flush-jointed form. It has been sufficiently referred to under the head of " Steel Pipes." The third class is found in many forms. Flanges are (a) Cast with the pipes, whether these are cast iron, cast steel or cast malleable. 83 STEAM PIPES (&) Welded on to mild steel or wrought-iron pipes. (c) Screwed on to mild steel or wrought iron and sometimes partly brazed in addition. (cF) Riveted on to mild steel or wrought iron or copper. (e) Brazed on to copper pipes. (/) Loose upon the pipes, which are gripped to- gether by lips turned up on the ends of the pipes after the flanges are slipped on. Classes (a) to (e) are referred to sufficiently under other headings, except as regards dimensions of flanges. Class (/) are loo numerous fully to describe. A large number will be found illustrated in a paper read by Mr. R. E. Atkinson. 1 Essentially the loose flange joint is formed by slipping a flange over the pipe, which is afterwards turned outwards to form a lip or small flange, or a thick ring flange is welded on to the pipe ends. Sometimes each pipe end, if of copper, is flared out to an approximate quarter-sphere, and the loose flanges draw the two ends upon a solid interior joint ring, the outer face of which is an approximate half-circle and is turned with circumferential V- grooves, into which the copper pipe is forced by the flange pressure. Copper pipes are perhaps more suitable than steel for loose flanges, especially in small sizes. Thus the 2-inch solid-drawn copper water-control 1 Minutes of Proceedings, Inst. M.E., 1901, page 443. PIPE JOINTS pipes of the Cruse Controllable Superheater (Fig. 15) are joined into a continuous spiral by loose thick stamped steel flanges, slipped loose on to each end of the U-pipe. The end is turned over to form a narrow flange, and this flange is drawn up to the face of the connecting link, the copper itself forming the metal to metal joint under the heavy bolt pressure. For various forms of loose flange joints the lists of the Mannesmann Co. may be consulted. To secure steam tightness between flanges various expedients are resorted to. The Manchester Steam Users' Association recommend faced flanges with merely a little red paint. Oiled brown paper is sometimes used on faced joints. Mr. Dewrance recommends scraped surfaces for pressures of 350 pounds, such as he used. Ordinary good practice uses woodite rings with flat-faced flanges. The Babccck Co. use the corrugated copper gasket, the flanges having a projecting face to take the rings. Compound rings of copper and asbestos have considerable elasticity, and make good joints. The author recommends solid rings of copper wire, and has used ordinary r V-inch copper wire with the ends simply crossed to form a joint ring. What is wanted to secure sound work is a truly faced flange free from spongy metal, and a closely pitched circle of 85 STEAM PIPES stout bolts in a strong flange so as to ensure a tight nip on the copper wire. If pipes are pulled apart at any time, the old rings must not be put up again unless they are first heated to a dull red heat and dropped into water to soften them, but old rings become flattened, and it is better not to practise such economies unless the flattened rings can be turned on edge to present their long axes to the flanges. Solid copper rings are made in the form of two flat V's placed back to back with the idea that the sharp edges of the V-grooves will make a good joint ; but as round copper wire is safe and reliable, there seems no good reason to seek further complexity. Superheated steam obviously demands that nothing of an organic nature shall enter into the composition of a joint ring, for the temperature will soon carbonize it. Where flanges are weak a joint covering the whole surface must be employed, or the flange may break when the bolts are tightened up. Flanges as shown in Figs, n, 12, 14, 16 must therefore be strong and stout, to withstand the bolt stress, and when a joint is made with a simple ring of copper wire it should not be attempted to use a light flange. If the flange is light the wire should not be too far inside the bolts, though too large a ring of copper means that more pressure is required to squeeze the ring to a tight joint. For superheated steam, if not too high in tempera- 86 PIPE JOINTS ture, Mr. Vanning has found Jenkins' graphited sheet to stand well. Probably asbestos paper rubbed in graphite would make a good joint over a fully-faced flange. The graphite would prevent the ring from sticking to the metal. Mr. Venning has used Rainbow packing for inser- tion between faced flanges up to 180 pounds in several of the largest power stations in England. CHAPTER VIII Supports (IPES may be variously supported, as follows (i) By pillars from below, as in Fig. 29. FIG. 29. PIPE. PILLAR AND SUSPENDER (YATES & THOM). (2) By hangers from above, as in Figs. 29, 30, 31, 32- 88 PIPE SUPPORTS (3) By brackets on which the pipe rests (Fig. 33), which shows simply the cross-section of a bracket FIG. 30. PIPE SUSPENDER (BRITISH ELECTRIC TRACTION CO.] of the form of Fig. 31 carrying the pipe on its upper , able. 89 STEAM PIPES (4) On brick piers. In the pillar form (Fig. 29), which is convenient for carrying pipes from the crown of a boiler or from the brick walls of the seating, the slightness of the pillar affords play for expansion movements. The head of the pillar is arranged with an adjustable FIG. 31. SUSPENDER FOR ONE PIPE (BABCOCK & WILCOX co.). FIG. 32. SUSPENDER FOR TWO PIPES (BABCOCK & WILCOX co.). screw to enable the weight of the pipe to be carried without either undue upward or downward stress on the connections of branching pipes. The part on which the pipe rests is a bow of about 90 arc, to which the pipe must not be tied down unless the bow is merely fitted into the pillar by a tail free to move up in case of the pipe rising from expansion. Fig. 29 shows also the suspender of Messrs. Yates 90 PIPE SUPPORTS & Thorn, which may be slung as shown from a wall bracket or from an overhead girder. The upper nut may have a helical spring placed between it and the bracket. When adjusting any form of pipe carrier with branches, the latter may be unbolted from the main pipe and their weight carried by a rope and balance weight equal to half the weight of the supported pipe if this is fastened to the rope near one end. The main pipe is then adjusted to height and the branches bolted to it. A little up- ward stress may be given when cold, as this will be relieved when hot by the expansion of the vertical branch on the boiler, if present. Suspended pipes are as free to move as their various attachments will permit. Often they will be set swinging or vibrating by the pulsating action of the draught of steam by the engines. This is usually best at- tended to after setting to work. It may usually be checked by lightly wedging the pipe at a point where one of the branch pipes passes through the wall. If the wedge only stops the full amplitude of the vibration, this may usually be entirely stopped. When pipes are simply supported on brackets they will not vibrate so readily. Brackets are often hollowed to the curve of the pipe, but this is a doubtful advantage, tending to prevent lateral movement under the push of the branch pipes. It is better to provide the table of the bracket quite flat and plain. Riveted to the pipe or fastened 01 STEAM PIPES by a pair of clip-rings encircling the pipe, there should be a rubbing piece of iron interposed between the pipe body and the bracket to take up the wear due to constant movement. These rubbing pieces should be from f to f-inch thick. If a pipe is to be anchored at any bracket the rubbing piece may be as per Fig. 33, with down projecting pieces straddl'ng the bracket table loosely, and with lateral extensions a a to take the clips c c firmly holding the rubbing piece to the pipe. The encircling clips are in halves, the bolted ears being placed at a convenient angle, preferably horizontally. FIG. 33- PIPE BRACKET. Pipe supports are intended only to carry weights and should not be arranged to prevent free move- ment necessary to relieve the expansion stresses, except of course at such point as is intentionally selected as an anchorage. In the type of hanger of Fig. 30, with girder and wr ought-iron rods and clips, the girder is simply a piece of 4^" x 2%" channel built into the wall, and the pipe is carried by a saddle slung by a double-ended nutted suspender, slung over a pin carried on the girder. This is the design of the British Electric Tracjtion Co., and the dimensions are given in the accompanying table. 92 PIPE SUPPORTS TABLE XVII. TABLE OF DIMENSIONS OF PIPE SUSPENDER (Fio. 30) FOR STEAM AND LAGGED PIPES. Dia. of Pipe. External Dia. of r Length of Channel. Size of Strap. Dia. and Length of Sling. B Bi B2 C D E F G H J K L in. in. in. in. in. in. in. in. in. i". in. in. in. in. 2 2i IO 8 4 8 i f Ii 6 2 i 13 3 3i 10 8 4 9i 1 \ Ii 7i 2 i | 14 4 4i 12 9 4 ioi I ft If 18 2* -| I 14 5 5f 12 10 5 ni I g 20 2f- -| I 16 6 6f 12 IO 5 I2i I -| 2 22i 3 J i 18 7 7l 12 12 5 i3i ii f 2 24 3 1 ii 19 8 8| 14 12 5 14 ii f 2 25 3 f J i 20 9 9f 14 13 5 IS* ii * 2 25 3 1 ii 20 FOR EXHAUST AND BARE PIPES. Dia. External Dia. of Length of Channel. Size of Strap. Dia. and Length of Sling. Pipe. Pipe A B Bi B2 C D E F G H J K L in. 3 in. 3i in. IO in. 8 in. 4 in. 5i in. 1 in. i in. Ii in. in. in. i in. i in. 12 4 5 12 9 4 6| I 1 J 4 I 5t 2 i f 12 5 6 12 IO 5 8 I f If 18 3 1 i 14 6 7 12 10 5 9 I f 2 20^- 3 1 x i 16 7 8* 12 12 5 10* ii | 2 21 3 1 ii 16 8 9i 14 12 5 Hi ii f 2 22 3 I i| 17 9 io| 14 12 5 I2i J i * 2 22* 3 1 2 17 10 iii 14 15 6 14 i| 1 2 24 3 I 2 i 18 ii 12* IS 15 6 I4| ii I 2 i 24i 3 I 2i 18 12 18 17 6 16 ii I 2 i 27 3 I 2| 20 13 J 4i 18 18 6 J 7i ii I 2i 30 3i I 3 22 14 is* 18 19 6 19 ii I 2i 30 3i 1 3* 22 93 STEAM PIPES The resting of pipes on rollers carried by brackets is not usually thought so good as the plain rubbing contact which introduces an element of stability against vibration ; but after all, the roller-carried pipe is less liable to vibrate than is the slung pipe. Rollers are, however, liable to become set fast, and they are then liable to wear the pipes which have not perhaps been supplied with rubbing pieces. In the type of bracket of Fig. 29 there should be a projecting lug at the bottom of the wall plate to rest in the wall for the purpose of taking the weight. The two top bolts must be strong enough to carry the load acting with an intensity of pull on the bolts W x A S = ^ , where N is the distance between the top bolt and bottom of the bracket and A is the distance frcm the wall to the pipe centre, W being the weight of pipe. The heads of the bolts must be carried by back plates, which may be either simple double-hole washers, or a full plate, as large as the wall back of the bracket. The Babccck Go. make brackets, as in Figs. 31, 32, the dimensions and weights of which are given in the annexed Table XVIII. Brackets of plain bent angle iron with a riveted jib piece of flat iron, are made with the dimension A reduced to about half the length, and B and C to less than half for small pipes, and to three-fifths for larger pipes, and they weigh less than a third of the cast-iron brackets. 94 ^ tw H'N c 00 C^ ^ 01 tNi t-xOO o^ oo k 01 CO %0 -0 00 V H H 01 01 CO C *co ^ MOO ^ VOO S 01 01 , 01 a V) ^ H 00 ^ VO !>. 01 c v o rH(00 *^rtt Tf OO 10 H 10 V H V CO O Ol 01 01 * *b CO O tO H CO O OJ 01 01 c *b * % Wnt HW 01 H 10 ^ QvJ I/") Ol iO CO v H ^ H 04 01 01 ^ 8,1 n . J !| ^P < PQ ^ c^fi "3" u 7 ,3, , i Sf SJ 6 _JL 3 ,.' .6' * 1 irf * \ 7S - ; '^ ,3r s ,.- y y ,, ~r "TsF 2' -7^ & i rrf 9*' ii ~; ; at - E S -- 1 3J. i -v' ;q" M I3* 1 M3t i 22 -' ' i* * f)'.. i^ 9t 2at^ >**' 1^ sl i Rl FIG. 41. " FULLWAY " VALVE WITH GENERAL DIMENSIONS (TABLE XXI.). 118 VALVES lowers the valve ; E E the two halves of the valves carrying faces of bronze or nickel alloy for superheat, and F F being the wedges which force the valve against their seat when the spindle of the lower wedge F touches the bottom of the body casting. A table of general dimensions of fullway valves (Fig. 41,) by Templer and Ranoe is given on page 121, and will be found fairly approximate for other makes of valves. The hand wheel F may be made into a gear wheel, and turned by a pinion with a long spindle extending downwards within reach, in cases where valves are placed high and out of reach. Such wheels and pinions ought to be made with wide teeth for strength as they do not perhaps afford the same power over the valve as the direct method of Fig/ 42 and Table XXII. ; which, however, demands a re- versed position of the valve with the risk of possible objectionable leakage of water at the gland. The hand wheel of a reversed valve should be attached by a nutted screw, as the wheel may drop off when only keyed. VALVE POSITION. Valves may sometimes be seen upon a ring main with their spindles brought up to a gallery, with a grid floor placed directly over the ring main. Obviously, any accident to the main may envelope the valve gallery in steam and render the hand- wheel inaccessible. This arrangement is on a par with the intelligence which employs ring mains by choice. 119 A B c D E F C K T J K L M Ki P 2 -> / \ - . * 2i V g - 3' '' ? \ ^*. 5" 1 ^ . 3^ '' 34, -. ' j . f'*' f *\ l , , i* 7* ,*' ,^ J .. . j --, 8' ''<' JS : \ & - * i ^\ -x g '-.. N - \ 22 3 | . if! 2* 7* g - ii S IT - . ' -A .- | i? a\ - _L z j FIG. 42. " FULLWAY " VALVE REVERSED, WITH EXTENSION SPINDLE (TABLE xxn.). 120 c X PQ 121 X w C/3 O co 1 X W O. _c O I Zt^a^JS I lO iO VO VO C^ OC i !-S 'co s * " * * M . H ^ H< x H c s 10 M H H 0> TT- O C^ ^9 s , CO 00 ON O^ iO tx CO Tf N o oo Tf H c o oo T|- C-N N Tf 01 ON 1O 00 VO lO M H OO VO ^ ^O > CO H H Tf CO ON Tf H OO CO H CO iO M Tf CO O CO . N 01 H o n- OJ Tt- CO H M IX J>.v> CO H OO CO o ^t- H O ON CO 10 H ^O N 10 H S ft G H co rfoo 00 CO tx ON "t- M 81O Tf H f^ rj- 10 N c CO v* ro COOO OO CO iO ON OO CO tx H Xt- M c IT M CO vo M o o VO CO Tf <^ VO M iOO CO H g a Tt-o CO H CO O "* M ^1- Tj- N iO N W H Diameter. ^ . ni w II II H - ^5 nJ^ II II H S ^ . HJ H II S? 2 - o d H H o & " 3 high pressures. The mild steel and the cast steel tee junction, as arranged by Yates & Thorn, are i35 JUNCTION PIECES AND FLANGES shown in Figs. 50 and 51. The flanges are here shown with the shallow spigot and socket elsewhere described. It is obvious that the recess offers a great safeguard against rupture of the packing, but it is often a great hindrance to the removal of a pipe. It has been suggested that the depth of the recess should be less than the thickness of the joint ring, so that this can be sawn through if necessary to part the pipe. Purely metallic joints of course do not adhere. FLANGES. Since flanged joints are the most usual, it is of importance that their dimensions should be stand- ardized. The lack of a standard has proved an immense inconvenience. There are five points to be standardized. (a) Flange diameter. (&) Bolt circle diameter. (c) Number of bolts. (d) Size of bolts. (e) Thickness of flange. (/) Angle of bolt holes. The dimension (e) is only for convenience in ordering bolts. It would be impossible in the space of this book to publish all the principal flange tables. I have selected a very few only, viz. : Those of the Babcock & Wilcox Co., because so largely employed. The standards of the British Electric Traction Co., i37 STEAM PIPES kindly supplied me by Mr. A. J. Lawson, of that Company, and practically the standard of the Brush Company (Tables XXIV., XXV. ; Figs. 52, 53) The American Standard of 1902. The German standard is omitted because it em- ploys numbers of bolts not divisible by four, and therefore awkward and academic. James Russell & Sons' standard. The cast-iron standard of the Crane Co. of Chicago. TABLE XXIV. B.E.T. Co. STANDARD STEAM, FEED DELIVERY AND SUCTION PIPE FLANGES. Dimensions of Pipes and Flanges. Bolts. A B C D E F G H No. Dia. Length in. i in. 2J in. 3i in. 1 in. i in. 1 in. I& in. Ift 4 in. i in. if 1 2j 4i I J J Ii ii 4 i if i 3t 4i 1 i I I| 2 4 i if ii 3i 4i f 1 I 2i 2| 4 i i| ii 4 5t 1 i ii 2i 2| 4 S 2i ii 4i 6 1 i ii 2| 2i 4 i 2i 2 5 6J 1 1 ii 3i 3i 8 i 2i 2J 5 6J 1 1 ii 31 3 1 8 * 2i 2i 5i 7 1 i il 3i. 4 8 i 2i 21 51 7i i i i| 4, 4i 8 1 2| 3 6* 7i 1 i i* 4f 4i 8 i z| 3i 6f 71 1 f ii 41 4J 8 f 2i 3i 6f 8 1 5 i| 4i 5 8 1 2i 31 7 81 J J i| 5i 51 8 i 2i 4 7i 9 1 I if 5i 51 8 I 2f 5 8| 10} I 1 2 6i 6| 8 i 2| 6 10 12 i I 2 7i 7i 8 i 2i 7 ii 13 i 1 2 8i 8J 12 f 2| 8 12 14 7 8 ii 2i 9i 9i 12 ! 3i 138 JUNCTION PIECES AND FLANGES TABLE XXV. B.E.T. Co. STANDARD EXHAUST AND CIRCULATING PIPE FLANGES. Dimensions of Pipes and Flanges. Bolts. A B c D E F G No. Dia. Length in. ft. in. ft. in. in. in. in. in. in. in. 2j o 54 o 7 f f i 4 4 4 2 3 o 6 o 7* ! 1 * 4 4 1 2| 3i o 6J o 8 1 1 A 4 4 1 2f 4 o 7 i o 9 I f i 1 8 4 2i 44 o 8 o 94 i ! 4 1 8 4 2i 5 o 8} o ioj i i 4 f 8 1 2| 6 IO I O i f j 1 8 1 2| 7 I0| I I i 1 4 1 8 f 2| 8 I I 2 ! 1 ft 4 8 f 2f 9 I I i 3 1 I * f 12 4 3 10 I 2j i 4i I I f i 12 f 3 ii i 3 i 5 i I f i 12 I 3 12 i 5 i 7i 1 ii 1 1 12 1 3l 13 i 6 i 8J I ii i 1 12 1 3l 14 i 7 i 94 i ii i i 12 i 3i 16 i 9 J 2 O I if i I 12 i 3! FIG. 52. FLANGES (B.E.T. CO. STANDARD). 139 STEAM PIPES As regards the number of bolts this should always be divisible by four, and should never be less than eight, if eight bolts can be got in. The size of a bolt should not be less than f-inch, and the holes should be ^--inch larger than the bolts. This rule gives excessive bolt strength in some sizes where the jump is made to an additional four bolts, but the J-inch bolt is not a satisfactory thing in practical engineering unless of manganese steel. FIG. 53. The advantage of the multiple of four is that a piece can always be turned through an angle of go and bolts will still come right. The arrange- ment of bolt holes should always be so that no hole comes on a centre line. The pitch of bolts should not exceed 4^ inches, according to Mr. Atkinson. As soon as with a given number of bolts the pitch becomes 4^ inches, the next size of pipe should have an additional four bolts. Thus with his 7-inch pipe the bolt circle is n|- inches, and the pitch of 8 bolts is 4-51. His 8 -inch pipe, therefore, has 12 bolts. 140 JUNCTION PIECES AND FLANGES TABLE XXVI. BABCOCK & WILCOX STANDARD FLANGES. Bore. High Pressure Steam. Exhaust. Feed. Diam. of Flanges No. of Bolts. Diam. of Pitch Circle. Size of Bolts. Diam. of Flanges No. of Bolts. Diam. of Pitch Circle. Size ! Diam. of of Bolts. ! Flanges No. of Bolts. Diam. of Pitch Circle. Size of Bolts. ins. ins. ins. in. ins. ins. in. ins. ins. in. I 3* 4 2* | - 3* 4 2* I I 4* 4 3* * 41 4 3* * I* 4* 4 3* 1 - 4*. 4 & } Ii 5 4 4 * 5 4 4 * 2 7 6 5* 5 6 4 4* 1 2* 7l 6 6 f 7 4 Si 1 3 8| 6 6| 1 ' 8* 4 6* 1 3* 9 6 7* 1 4 10 8 8* * 9 4 7* * 9i 6 71 1 5 ii 8 9 I 10* 6 8* * 6 12 8 I 10 1 IX* 6 9* 1 - 7 H 12 ii| 1 12* 6 io* * 8 | 14 12 12* 1 13* 8 12 * - 9 15 12 13* 1 14* 8 13 1 10 17 12 14* 1 15* 8 14 1 ii 1 8* 12 16 1 i6f 12 15 1 12 19* 12 17* 1 17* 12 16 | 13 - 19* 12 i7l 1 14 20* 12 18* 1 15 21* 12 19* I 16 22* 12 20* | 17 23i 16 21* 1 18 24i 16 22| | 20 26i 16 24* 1 22 28| 20 2 7 1 24 30* 20 2 9 * The author is inclined rather to limit the pitch to 4 inches, thus giving 12 bolts to the 7-inch pipe as practised by the Crane Co. Ordinary commercial bolts tested by Professor Goodman have shown a tensile strength of 29 to 35 tons per square inch, the smaller bolts coming 141 STEAM PIPES out best, but bolts in practice are exposed to a tor- sional stress, and the smaller bolts are apt to have the biggest stresses put on them, and it is wise to keep to I as a minimum size where possible. TABLE XXVII. DIMENSIONS OF CAST-IRON FLANGED FITTINGS AND CONNEC- TIONS, AS USED BY CRANE COMPANY, CHICAGO. Inside Diameter of Pipe. Diameter of Flange. Diameter of Bolt Circle. Number of Bolts. Set of T- Branch or Quarter Bend, etc. Length of a T. ins. ins. ins. ins. ins. 2 6 41 4 4i 9 2 i 7 Si 4 5 IO 3 7i 6 4 5i ii 3i 8* 6} 4 6 12 4 9 7i 8 6J 13 4i 9i 71 8 7 14 5 IO 8* 8 7i 15 6 II 9t 8 8 16 7 I2| II 12 8J 17 8 13 J 12 12 9i 19 9 15 13 12 io| 2lJ IO 16 I4J 12 nt 23 12 19 17 16 I2| 2 5i 14 21 i8J 16 13^ 26J 16 23i 2lJ 20 i5t 3i 18 25 22j 20 i6J 33 20 27i 24! 2O 18 36 22 294 2 7 J 24 20 40 24 31* 2QJ 24 22 44 For very special work bolts of manganese steel may be procured, which are greatly superior to ordinary bolts. Mr. E. R. Briggs proposes as a suitable stress for bolts, / = 5,000 d, where d is the nominal bolt dia- 142 JUNCTION PIECES AND FLANGES meter. This gives the following stresses per square inch to be allowed in any bolt : Bolt. / J-in. = 2,500 pounds t 3,125 I 3,750 J 4,375 i = 5,000 ,, and over 5,625 He would never allow / to exceed 6,000 pounds. The rule allows for the weakness and liability to overstress of small bolts. TABLE XXVIII. TANDARD FLANGES ADOPTED BY MESSRS- JAMES RUSSELL AND SONS, CROWN TUBE WORKS, WEDNESBURY Inside Diameter of Pipe. Outside Diameter of Flange. Inside Diameter of Pipe. Outside Diameter ot Flange. inches. inches. inches. inches. 4 3* 5 IO i 3* 54 104 i 4i 6 "4 it 5 64 I2J 14 54 7 13 if 54 74 i34 2 6 8 14 2J 61 84 i44 2i 7 9 15 2| 74 94 i54 3 8 10 164 3i 84 104 17 31 84 II i74 3l 9 "J 18 4 9 12 184 44 94 143 STEAM PIPES TABLE XXIX. STANDARD FLANGES ADOPTED IN AMERICA, JANUARY, i, 1902, FOR PRESSURES 101 LB. TO 250 LB. Diameter of Pipe. Diameter of Flange. Thickness of Flange. Diameter of Bolt Circle. Number of Bolts. Size of Bolts. ins. ins. ins. ins. ins. 2 6J 1 5 4 1 2* 7i I 51 4 i 3 8i ii 6f 8 t 3i 9 i* 71 8 i 4 10 il 7J 8 ! 4i ioi i* 8J 8 1 5 ii if 9i 8 i 6 12} i* I0| 12 1 7 14 i* "I 12 i 8 15 i| 13 12 i 9 16 i| 14 12 i 10 i7i ij i6i 16 i 12 20 2 171 16 i 14 22J 2* 2O 20 i 15 23i 2* 21 20 i 16 25 21 22J 2O i 18 27 2| 24* 24 i 20 29i 2i 26J 24 ii 22 3ii 2f 28| 28 ii 24 34 2j 3il 28 ii 144 CHAPTER XIV Separators, Exhaust Heads and Atmo- spheric Valves A WATER separator for removing the surplus water from saturated steam acts always on the principle of the first law of motion, taking into effect the tendency of an inert body such as water to move in a straight line. All separators, there- fore, act by causing the flow of steam to be suddenly reversed in direction. The steam follows the new path and the water continues, and is caught in a suitable receptacle and trapped off. There are legions of separators in the market, but all work on the same principle. Two or three forms only are therefore illustrated, that of Holden & Brooke (Fig. 54), and those of Yates & Thorn (Figs. 55, 56). EXHAUST HEADS, for preventing the escape of oil and water at atmospheric discharges, act on the same prin- ciple, affording a large internal area for the steam, and a quiet part for the collection of oil gathered in the cones and on the outer cylinder. STEAM PIPES Fig. 57 is the exhaust head of Holden & Brooke, who make them so that a 5-inch head will deal with 1,000 pounds of exhaust steam per hour, a lo-inch DRAIN FIG. 54. WHIRLING SEPARATOR (HOLDEN & BROOKE). with 3,060 pounds, a i6-inch with 9,000 pounds, and a 24-inch with 25,000 pounds, and pro rata. 146 SEPARATORS ATMOSPHERIC VALVES. Where an alternate exhaust is desired to a con- denser, or to atmosphere, a valve is fitted in the FIG. 55. FIG. 56. REVERSE FLOW STEAM AND WATER SEPARATORS (YATES & THOM). atmospheric exhaust proper, which will automati- cally open and close when condensation ceases or i47 STEAM PIPES resumes. These valves ought to have a shallow water seal above them so as to obviate any air leak- age. An oil dashpot ought to be fitted outside the body to prevent hammering of the valve, which will occur if no dashpot is present or only an air dash- pot be employed. FIG. 57. EXHAUST HEAD (HOLDEN & BROOKE). A glass water-gauge should show the depth of water-seal, and a drain pipe should prevent its be- coming too deep. A supply pipe should also keep up a supply of water or the seal may leak away and air may leak in. The atmospheric valve is so frequent a cause of bad vacuum that it deserves more attention than it usually obtains. One of these valves by Templer & Ranoe is shown in Fig. 58. It can be placed upside down equally well. There is an outside oil dashpot. An automatic exhaust 148 SEPARATORS FIG. 58. ATMOSPHERIC EXHAUST VALVE. 149 STEAM PIPES valve by Thos. Walker, of Tewkesbury, is shown in Fig. 59. This is shown with the customary air dashpot, but oil can be substituted. In the author's opinion the air dashpot is what is responsible for the clatter of atmospheric valves when opening and closing under the pulsation of the exhaust. He would fill the dashpot with oil on both sides of the piston, 150 SEPARATORS and in place of the snifting valves would connect the top and bottom of the cylinder by a small pipe with a valve. By regulating this valve the proper action of the automatic exhaust valve would be better secured. An air dashpot can be converted into an oil dashpot by means of a small cock and a bit of pipe joining the opposite ends of the cylinder. CHAPTER XV Superheated Steam THE fact is well recognized that moisture in steam is one of the great causes of friction and resistance to flow. The water is inert. When it strikes the pipe surface it is stopped in its progress, and it continually puts a drag on the steam. Steam dried and superheated undoubtedly travels faster, but experiment is wanting to say to what extent. It is possible that more superheated steam will pass through a given pipe in a given time with a given loss of pressure than is the case with saturated steam. The volume of superheated steam varies with its absolute temperature very approximately. Thus steam at 360 F. has an absolute temperature of 360 + 459, or 819. Superheated 100 F., its volume is now increased in the ratio (819 + 100) : 819, or about 12 per cent. Superheated 200 F., the volu- metric increase is nearly 24 per cent. With modern pressures the temperature of superheat will rarely exceed 200 above saturation temperature. Mr. Cruse makes the cross-section of the pipes of his superheater from 25 per cent, for high pressures to 50 per cent, for low pressures in excess of the 152 SUPERHEATED STEAM boiler steam pipe. With this provision the loss of pressure in traversing the long pipe superheater is always under three pounds, and more usually only one pound to two pounds. The superior mobility of .superheated steam is probably such that the size of the steam pipe need not be increased for a given weight of steam. Where the same power is to be developed, the diminution of the weight of steam required will not differ far from the inverse ratio of the increase in volume due to superheat. On the whole, therefore, for a given power the steam pipes may be less in size than usually provided. PIPE COVERINGS. Every manufacturer of pipe coverings will pro- duce figures to show that his special material is the best. It is certain that almost anything sold will pay for itself in steam saved. The best of all material is loose wool. Loose lamp-black, down and hair-felt come next, and generally it may be said that the best heat insu- lators are those which imprison the most air in a finely divided condition. But all organic sub- stances are unsuitable for the modern conditions with superheated steam, and some form of magnesia covering or other similar preparation is probably best. Coverings are sometimes put on in a soft plastic condition and hardened in place by the heat of the STEAM PIPES pipe. Others are built up into sections and fitted to the pipes and held by wire-binding, or clips of hoop-iron, or by hooks and eyes. The neatest cover- ing has an outer case of Russia iron. In all cases the flanges ought to be covered. No covering should be less than one inch in thickness. .This may be exceeded if the value of the heat saved renders it economical, and often it will pay to put two inches of covering upon a pipe. Numerous tests have been made from time to time by various experimenters on different sub- stances, and particulars of these tests may be found in the Proceedings of the American Society of Mechani- cal Engineers. Tables and data may be found in Kempe's Year Book and in Steam y and in various other pocket-books and in the catalogues of makers of coverings. From one of these it appears that the composition prevented five-sixths of the loss with bare pipes. A thickness of one inch of hair-felt also reduced condensation to one-sixth, two inches of felt reduced it to one-eleventh, while the abnormal thickness of six inches reduced it to one-twenty-fourth. Small pipes lose relatively more than large pipes because the area of an equal thickness of covering is greater. It is also more costly to cover small pipes because the same thickness or more is neces- sary, and the area of a small pipe and its steam- carrying capacity is less per unit of superficial area. As a general rule it will be good practice to employ coverings ij inch thick, unless experiment can i54 SUPERHEATED STEAM be made to determine the economy of a different thickness by equating the interest charge of the covering and the fuel value of the heat loss, remem- bering that in a hardly-pressed plant an additional outlay on pipe coverings might render it possible to avoid adding extra boilers. It is possible of course to cover a pipe first with magnesia, and upon this with hair-felt, which would be protected from the most severe heat by the inner mineral layer. Coverings which are liable to loosen or disin- tegrate under vibration should be avoided. Slag-wool is apt to do this, and it is heavy and is liable to cause trouble if it gets into machinery bearings. Lightness is a favourable quality in a covering because it indicates considerable air space, a feature which is sought in fossil meal in the shape of the minute cavities of the diatoms ; in certain asbestos, corrugated millboards, and in wool-felt, both in the fibre itself and the frictional effect by which the myriads of fibres hold the air from cir- culating. The non-circulation of air inside the mass of the covering is one of the more valuable features of the best compositions. The very common practice of leaving pipe flanges and bolt heads and nuts bare of protection is neces- sary with the plastic compounds which are stopped off short of the flanges, but this is no excuse for neglecting to provide loose covers over the flanges. A few tables and deductions abstracted from a report by Mr. Atkinson, of Boston, relative to the i55 STEAM PIPES tests of Mr. C. E. L. Norton, on pipe coverings, will be useful. They have been translated on the basis of i = $5, and they are probably as accurate as any tests made, and will afford a useful guide to the engineer who wishes to have his pipes dealt with in the manner that the importance of the question demands. The tests were made in 1898 by Mr. Norton, on pipe coverings of various types. He employed an electrically-heated apparatus with coils of wire in a bath of oil, and by maintaining the oil at a fixed temperature he was able to measure the heat gener- ated, and therefore lost, by the measure of the current consumed. Particulars of the test need not be detailed ; they may be found in Circular Note of the Mutual Boiler Insurance Co., of 31, Milk Street, Boston, U.S.A., 1898. A few of the tabulated results are here abstracted, and it may be added that Mr. Edward Atkinson, selected A, D, G, E as safe in respect of safety from fire and efficiency in results. Articles containing lime sulphate are not advised because of the danger of corrosion of the covered pipe, and many so-called magnesia coverings con- tain rather lime sulphate than magnesia. Magnesia of course is good if it can be obtained really pure and unadulterated. Mr. Norton also recommended plastic coverings as better than sectional for certain conditions, and especially where vibration is likely to occur. Sectional coverings are looked on usually as better than plastic. Yet at least 20 per cent, of 156 SUPERHEATED STEAM plastic must always be employed for the irregular surfaces. The tables which follow are at least sufficient as a general guide, and prove the undisputed benefit of good coverings. Specimen A, Nonpareil cork standard, consists of granulated cork, pressed in a mould at high tem- perature and then submitted to a fire-proofing process. Specimen B, Nonpareil cork octagonal, is similar in composition, but is made up of several strips of cork, instead of two semi-cylindrical sections. Specimen C, Manville high-pressure sectional cover, is composed of an inner jacket of earthy material and an outer jacket of wool- felt, the whole being one and one-quarter inches thick. Specimen D, magnesia, is a moulded, sectional cover, composed of about 90 per cent, carbonate of magnesia. Specimen E is essentially an air cell cover, being composed of sheets of asbestos paper which has been indented before being laid up, the indentations serving to keep the thin sheets of paper from coming into close contact with one another, thereby causing a considerable amount of air to be held throughout the body of the cover. Specimen F is composed of a wool-felt with a lining of asbestos paper. Specimen G is a cover made up of thin sheets of asbestos paper, fluted or corrugated, and stuck together with silicate of soda. 157 9 PQ Is o ju g id-s ' IH ^*" IS, t^vO"^ininCNin| 1 inooHioO NHmco'^incol l i OTi~'^t'coo QJ C/5 c S'S 8OinN.oo H o co H " W CO CO ^t" "^t" ^O t^ OO OO CO O^ O CO vO o< t 1 at* CQ 5T C*lMc<|fOfOcOf 1 o H ^a ^ 1 rt CXJOflJ * " * *n^J*^ '"^ " "S^S wg< d tl ^ F-I ? 55cD -s -gll-i-ll ^ fc "S (Sg^S | M 6 a .f -1 siJlsf.S ^ < <^^S5cofa I-^H ^ yj 'W H-J O P* 158 SUPERHEATED STEAM Specimen H is a plastic covering made of infusorial earth. Specimen I is similar to Specimen F. Specimen J is a plastic cover called magnesia- asbestos. It contains only a slight amount of car- bonate of magnesia. Specimen K is a moulded cover, containing about 45 per cent, of carbonate of magnesia and a con- siderable percentage of carbonate of calcium. Specimen L is composed mainly of sulphate of calcium and some 20 per cent, of MgCO 3 and has upon its outer surface a thick sheet of felt board. Specimen O is similar to Specimen G, except that it has larger cells and contains much more silicate of soda. It is very hard and strong. Specimen P is a sectional, moulded cover, com- posed mainly of sulphate of calcium. It has an outer layer of felt board. Of Specimens C, J, L, and P, the principal in- gredient is stated to be sulphate of lime and not carbonate of magnesia. Prospective purchasers ol pipe covers should not be misled by names. Since the appearance of Professor Ordway's reports it has been recognized that carbonate of magnesia is of great value as a non-conductor of heat, hence the name " magnesia " has been applied to a great many covers. It is to be observed that there is no virtue in a name. Asbestos is merely an incom- bustible material in which air may be entrapped, but when not porous is a good conductor of heat. Magnesia is a most effective non-conductor. This 159 STEAM PIPES name has been applied to many compounds of which the greater part consists of carbonate of lime or of plaster of Paris, materials which are not good as heat retarders. The percentage of magnesia car- bonate and plaster of Paris in several moulded, sectional covers is given in Table B. The Cork, Magnesia, Air Cell, and Imperial covers cause no corrosion. TABLE B. Percentage Composition. Specimen. MgC0 3 CaS0 4 Carbonate of Magnesia. Sulphate of Calcium. D 80 to 90 3 C less than 5 65 to 75 L 20 to 25 50 to 60 P less than 5 75 J 10 to 15 none The conditions of testing were reasonably near the conditions of actual practice. The room tem- perature was kept at 72 F. and the openings into the room were carefully closed. It was found early in the series that variation in the amount of moisture present in the air altered the amount of heat lost from the covers, but no attempt was made to correct this. The error introduced is not greater than i per cent. It was found that the heat loss per square inch of the flat surfaces at the ends of the pipes was less by several per cent, than the loss from the sharply 1 60 SUPERHEATED STEAM curved sides, and as all pipe covers tested were used to cover both sides and ends, the figures given in the table show a loss, less than would be shown were the pipe surface wholly cylindrical, and more than if it were all flat. The pipes were suspended from the ceiling, as described in an early paragraph, and the air cir- culating about them was due only to their own convec- tion currents. The variation in thickness in different places on the same specimen was considerable, but an average of twenty measurements was taken and results given in the table to the nearest one-eighth of an inch. Owing to these variations in thickness, the results of a measurement of the efficiency of any one cover cannot be used to predict the efficiency of a second cover of the same make with an accuracy greater than 2 per cent. Two specimens of each make were tested, and, in some cases, four, the mean value being given in the table. Table C gives the saving, in 's, due to the use of the various covers. Table D shows that at the end of ten years the best of the covers tested will have saved 9-2 more than the poorest. The difference between the several covers of the better grade is exceedingly small. The money saving is computed on the following assumptions : Coal at sixteen shillings a ton evap- orates ten pounds of water per pound of coal ; the pipes are kept hot ten hours a day, three hundred and ten days a year. If computations are made, as is sometimes done, on an assumption that the 161 M O w S {} O rj- Tj-cO O CO O t^ CO tt O ^TfiO [ U"}TJ-Tj-rOCOOJMHMHHOCOVO 1 ^ i D PQ* C/J ^* O O C^ iO d C*s T^ tx O CO ^t" M CO ffc p d S OCOCO l/~>C^M t^O t^CO H O CO H Tf d COCO^r}-vO t^cOCOCO O>O COOCO d-8 *3 MMddddClddC^MCOCOCOCO H . M i .en -fi 5 1 1* ^.1 ^ Qjn S ^ ^5 . e 03 ^ ^^ CrJ fcl J2 ^H rt M HH 3 ^! H +j "-H c/l T3 % w CJ C/5 J-i "*"* /"^ ^^ ^ OJ O <^ , ctf ^S ? i; '8 S (8 ' ' .2 8 o8e o p 2~--> n S^CuO j S . gj '> " " S 2 "g .-5 ,, 1 ' "fill! i MUQWfeOffi^H' .^JO^ I 162 SUPERHEATED STEAM pipes are hot twenty- four hours a day, three hundred and sixty-five days in a year, the saving is nearly three times that shown in Table C. Generally speaking, a cover saves heat enough to pay for itself in a little less than a year at three hundred and ten ten-hour days, and in about four months at three hundred and sixty-five twenty- four hour days. It is evident that the decision as to the choice of cover must come from other considerations, as well as from the conductivity. The question of the ability of a pipe cover to with- stand the action of heat for a prolonged period with- out being destroyed or rendered less efficient is of vital importance. The increasing use of cork as an insulator has led to many questions as to its ability to reaiain ff fire-proof." Exposed to a temperature corresponding to three hundred and fifty pounds of steam for three months, and to a temperature corre- sponding to one hundred pounds for two years, no change was found, and any suspicion of its ability to withstand continued heating is considered ground- less. The magnesia covering is of course unquestion- able on this ground, being almost indestructible by heating. The Imperial asbestos is also perfectly safe from any fire risks, as is the Air-Cell and Fire-Board. The Manville infusorial earth, and also the Man- ville magnesia-asbestos are liable to no accident from fire, nor is the Carey calcite. 163 w 8 1 d o o 00 o o 00 O o m o O N 8 o o H O VO? c^ CO o CO' vO ^ 8 s O "" CO H , O CSJ w &. o 00 o o Tf m 00 tn irx o O O m CM O Tf CM OJ s N N >< N CM CM M C9 CM CM M M rQ % * oj rt < CJ PQ c (LI .1 < PQ W feO S ~ > 3 W-l C PM Qi O5 164 SUPERHEATED STEAM It is not safe to put upon a steam-pipe wool, hair- felt, or woollen felt in any form. The causes of risk are two : First, the wool may become charred by heat from the pipe and finally ignited, though this can hardly happen, even on high-pressure pipes, when the thickness of fire-proof material (asbestos, magnesia, or whatever it may be) is as great as one inch. The second and most serious risk is from the presence in shops or mills of the long tubes of wool, dry as tinder, often connecting one room with another, and ready to flash at the slightest rise in the already too great temperature. Canvas jackets on the covers should be fire-proof. The efficiency of wools is high as non-conductors, but not higher than any other perfectly safe covers. If the wool is separated by about one inch of fire-proof material from the pipe, it is not kept so hot and dry, and the risks from outside ignition is less ; but the practice of many engineers of wrapping hair-felt outside of a sectional cover is not advised. The saving due to this practice is indicated in Table E. The following assumptions have been made in computing the Tables D, E and F. First, that all the covers cost 5 per one hundred square feet, applied. This is a high figure, perhaps too high, yet it is not far from the list price of several makers, and any attempt to get a definite price from them revealed a maze of discounts and double discounts and flexible price-lists too intricate for an uninitiated mind to travel. In case the saving due to a cover, which costs 4 instead of 5, is desired, the simple 165 STEAM PIPES addition to the final saving of the i difference makes the necessary correction. Secondly, by the advice of the makers, the assump- tion is made that the cost is not nearly proportional to the thickness. As the thicker coverings are not now made in great quantities, the actual cost of their manufacture is uncertain. Inspection of Table E shows the saving due to the use of hair-felt outside a standard magnesia cover. In five years one hundred square feet of hair-felt saves 1-4 more than its cost, and in ten years it saves 4 above its cost. The further saving due to a second inch outside the first is 1-60 in ten years. Of course the well- known tendency of hair-felt to deteriorate should be considered. In the case of Nonpareil cork, increasing the thick- ness from one to two inches raises the cost from about 5 to 7 per one hundred square feet, and increases the net saving in five years by 2 and by 6 in ten years. In other words, the second inch of material in use about pays for itself in two years, while the first pays for itself in about one year. The third inch does not increase the saving even in ten years. The second inch, therefore, more than pays for interest and depreciation, while the third fails to do this. In the case of the asbestos fire board, a second inch in thickness causes a saving of 4 in ten years, the third and fourth inches showing a loss. 166 W w .JLS x o gu If o o ooo oooo 9 9 990 9999 vO tx OO lOlfxO tOtfxOcO M3 H H H E T}- Tf O vOvCO Tt'C^OOO 1 O^ CO to O vO C^ CO x vO CO vO Cx x Irx ^* ^x vO vO vO vO ^? k OO N Tf OOOOO OJNOrf a * > .S "^ '> rt H CO CO CM ^J~ (M O^ O OO vO CO CO CO COCOOO C^COCMCM c/: oJ in *X ^ X rt o c^ o o o CM oo o o vo H O vO O tx OO vO ^ CM CO ;* N C^ O^ 00 O O^ vO OO iTx vO CM <^p H 1 tOTj-QvO Tj-iOOCM io o co M^ll 9'?"'! 1 ~ r T > H H 6 CM' HOCMlO s iO O CO vO CO ^t" OO "^"VO O ft* ^x OO OO l^OO OO vO tx tx tx t ij CM OO tX'^TTTt- Tt-OOOcO V-O cO ^x vO OO O^ vO "^^ t^* OO f R t- H CM CM HCMCM QHHH *i o ' o .S -2 * H * N C G c V S rX O '1 w ||,ij ^ .^ , .2 .2 fa .5 fa 'S s c c c c c fi 1 TB" * 384-7 Fahr. 385-0 384-6 384-7 ,, 356 Fahr. 329 302 266 146 pounds 102 7 M 39 Attention being called to the varying loss from bare pipes when their surfaces were in varying con- ditions as regard rust, dirt, paint, etc., a few brief tests to show any large variation which might occur 1 68 SUPERHEATED STEAM from the loss from bare pipe, viz. 13-84 B.Th.U. per square feet per minute, are shown in Table G. TABLE G. Loss OF HEAT AT 200 LB. FROM BARE PIPE. Condition of Specimen. B.Th.U. loss per sq. ft. per minute. New pipe ..... Fair condition .... Rusty and black .... Cleaned with caustic potash inside and out Painted dull white Painted glossy white Cleaned with potash again Coated with cylinder oil . Painted dull black .... Painted glossy black 11-96 13-84 14-20 I3-85 14-30 12-02 13-84 I3-QO 14-40 12 10 The rate of heat loss from a bare pipe is also affected by the air circulation and the temperature of the surrounding bodies. A few tests were made to indicate the magnitude of the errors likely to be caused by variation in these conditions, and a brief examination of some of the results may be inter- esting. They are given in Table H. Table I shows the varying loss from a bare pipe with the change in pressure. A very thorough test was made of the common method of judging a pipe cover by the sensation of warmth given the hand on touching it, and nothing too harsh can be said of this practice. The sensation is dependent to such an extent upon the 169 STEAM PIPES nature of the surface that it fails utterly to give any idea of the actual temperature. TABLE H. EFFECT OF SURROUNDINGS. Condition and Position of Pipe. 1. Standard condition ; hung in centre of room 2. Near brick wall, between windows . 3. Hung horizontally in centre of room inches long 4. Vertical lo-inch pipejjg mch {. lo-inch diameter . 4-inch 6. 4-inch diameter in draft from electric fan . B.Th.U. lost per sq.ft. per minute at 200 pounds. 13-84 I4-26 12-06 13.48 14-42 14-42 15-20 20-IO TABLE I. VARIATION OF HEAT Loss WITH PRESSURE. Pressure. Lb. Bare Pipe Loss B.Th.U. per sq. ft. per Minute. 340 15-97 2OO 13-84 IOO 8-92 80 8-04 60 7-OO 40 574 The ease of removal for repairs or alterations makes the sectional cover better than plastic for some work, but there is much pipe surface which might be covered securely with plastic, where a sectional cover is soon ruined by vibration. Of course, the plastic covers offer no possibility of 170 SUPERHEATED STEAM leaky joints and long cracks. It should be borne in mind that in most cases about 20 per cent, of the entire surface to be covered is irregular, and must be covered by plastic or fittings. It will be well for prospective purchasers of pipe cover to see to it that their contracts call for fittings and plastic of as high an efficiency as the sectional cover shows. TABLE K. MISCELLANEOUS SUBSTANCES. Specimen. B.Th.U. per sq. ft. per minute at 200 pounds. Saving in one year per 100 sq. ft. pipe. Box A : i. With sand 3-l8 6-92 2. With cork, powdered . 175 7-88 3. With cork and infusorial 1-90 778 earth 4. With sawdust . 2-15 7-58 5. With charcoal . 2-OO 770 6. With ashes 2-46 7-38 Brick wall 4 inches thick . 5*17 5-76 Pine wood I inch thick . 3-56 676 Hair-Felt i inch thick . 2-51 7.36 Cabot's seaweed quilt i inch thick. 278 7.18 Spruce i inch thick . 3-40 676 2 inches thick . 2-31 7-30 3 inches thick . 2-02 770 Oak i inch thick .... 3-65 6-62 Hard pine i inch thick . 372 6-58 Eider-down i inch thick loose *r-90 to 270 i inch thick tightly *I7O to 1-80 packed Variable. Table K gives some figures concerning a consider- able number of samples of non-conducting material, 171 STEAM PIPES not, perhaps, classed as pipe covers, but used for heat insulation, which may be of interest. The box A, referred to in the table, is a f-inch pine box, large enough to surround the pipe, leaving a one-inch minimum space at its four sides. In it were tested several materials, which are used in this way for steam and cold storage insulation. 172 CHAPTER XVI Weights of Pipe THE weight of a pipe is usually found by multi- plying the length in feet by the weight of a foot length of pipe, and adding two flanges. Excellent tables of weights of pipes and various junction pieces are published by several firms who supply pipes. For work not using the pipes of any special maker the weight of iron may be taken as 20 pounds per square foot for material ^-inch thick, and pro rota. For steel pipes from | to i inch thick, and from 10 I.W.G. to II.W.G. the tables of the Mannesmann Company are very full. To calculate the weight of a tube multiply together its length in feet and its mean diameter in inches, and the number 10-6 x thickness. Thus a pipe 6J external diameter and J- thick will weigh per i foot long 6x10-6x0-25. Approximately, for ordinary steam pipes, 10 times the external diameter x thickness = weight per foot. Thus for a 6J external diameter pipe J- thick we have 6-25 x 10 x 0-25 = 15-6. The figure in the Mannesmann tables is 15-9. The rule gives results about 2 per cent, light for small thin pipes up to 3 per cent, heavy for larger heavy pipes, but it is very close to truth for pipes of 173 STEAM PIPES ordinary use, and crosses the line of plus and minus error at 8 inches external diameter. Cast iron weighs about 6 per cent, less than steel, or 9-375 pounds per square foot J inch thick. The following weights per superficial foot I inch thick will be useful lb. lb. Cast Iron . . . 37-50 Copper . . . 46-2 Wrought Iron . . 40-42 Brass . . . 43-3 Steel . . . 40-82 Lead . . .59-5 Thus wrought iron is I per cent, heavier than the rule above allows for, and steel is another i per cent, heavier. Brazed copper tubes weigh somewhat more than solid drawn. The specific gravity of tough copper at 6oF. is 8-8917, or 0-3229 lb. per cubic inch, or practically 3 cubic inches to the pound. Specific Gravity. Weight per cubic inch. Aluminium .... 2-56 0926 Brass, from . 8-82 3194 to 7-82 -2828 Gunmetal .... 8-70 3147 Copper .... 8-69 ! -3146 ,, drawn 8-88 3212 pipe . 8-89 3229 Iron, Cast .... 7-21 2607 ,, wrought bar 779 2817 ,, rolled plate 770 -2787 Lead, Cast .... ii-35 4106 ,, rolled .... n-39 -4119 Nickel .... 8-80 3183 Steel plate .... 7-80 2823 Zinc, rolled .... 7-19 26OO Tin .... 7-29 2637 174 WEIGHTS OF PIPE Useful tables for the weight of copper and iron, steel and cast-iron pipes will be found in Kempe's Year-Book, or other pocket-books and manufac- turers' catalogues, and are therefore not given here. JNO. SPENCER'S, LTD., STANDARD DIMENSIONS OF TUBULAR IRON & STEEL FLANGED BENDS (FiG. 60). D A B c Bore of Pipe. Inches. Radius at Centre. Inches. Length Straight. Inches. Centre to Flange Face. Inches. 1 2 2i 41 1 2j 2} 5 I r 3 3 6 ij 3! 3 6| ii 41 3 71 if 5i 31 2 6 31 91 21 . en 71 4 "I 3 JJ, 9 4 13 3i 10} 5 4 12 5 17 41 r 3i 6 5 15 6 21 6 I 18 7 25 7 c 1 24^ 7 8 HSJ | TT 4 1 28 8 3<3 2 9