QforttEU Inioeraita Siibtatg 3tl{aca. Hew ^atk B'OUGHT WITH THE INCOME OF THE SAGE ENDOWMENT FUND THE GIFT OF HENRY W. SAGE 1891 Cornell University Library TJ 795.P98 Diesel engine design, 3 1924 004 406 116 The original of tliis book is in tlie Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924004406116 DIESEL ENGINE DESIGN DIESEL ENGINE DESIGN BY H. F. P. PURDAY, B.Sc, A.C.G.I. NEW YORK D. VAN NOSTRAND COMPANY Printed in Great Britain PREFACE This book is based, on about twelve years' experience of Diesel Engines, mainly from the drawing-office point of view, and is intended to present an account of the main considera- tions which control the design of these engines. The author ventures to hope that, in addition to designers and draughtsmen, to whom such a book as this is most naturally addressed, there may be other classes of readers— for example, Diesel Engine users and technical students — ^to whom the following pages may be of interest. The text deals mainly with general principles as exemplified by examples of good modern practice, and it has not been possible to notice every constructional novelty. Apology is perhaps called for on account of the omission of any special treatment of the stepped piston and the opposed piston types of engine. These, however, are the specialities of a compara- tively limited number of manufacturers, and have been very fully described and illustrated in the technical press. The existence in its fourth edition of Chalkley's well-known book on The Diesel Engine for Land and Marine Purposes has enabled the present writer to proceed to details with a minimum of preliminary discussion. A number of references to other books and papers have been inserted in order to avoid, so far as possible, overlapping with other sources of information. The author has pleasure in acknowledging his indebtedness to : Mr. P. H. Smith (who has at all times placed his unique experience of Diesel Engines at the disposal of the author) for several corrections and suggestions ; to Mr. L. Johnson, M.A. (Cantab.), for his very careful and patient revision of the proofs ; to the author's wife for assistance with the manuscript and for compiling the index. H. F. P. P. CONTENTS CHAPTER I FIBST PBINCIPLBS The Diesel principle — Compression pressure and temperature — The four stroke cycle — The two stroke cycle — Types of Diesel Engines ......... Page 1 CHAPTER II THEBMAL EFPICIENCY Calorific value of fuel — Laws of gases— An ideal Diesel Engine- — Fuel consumption of ideal and real Diesel Engines — Mechanical efficiency — ^Mechanical losses — Efficient combustion — Entropy diagrams — Specific heat of gaseous mixtures . . .16 CHAPTER III EXHAtrST, SUCTION AND SCAVENGE Renewal of the charge — Flow of gases through orifices — Suction stroke of four stroke engine — Scavenging of two stroke engine — Exhaust of two stroke engine ...... 37 CHAPTER IV THE PBINCIPLB OF SIMILITUDE Properties of similar engines — Weights of Diesel Engines — Determination of bore and stroke — Piston speed- — Mean indicated pressure . . . . . . . .53 CHAPTER V OBANK-SHAFTS Materials — General construction- — Order of firing — Details of con- struction — Strength calculations — Twisting moment diagrams 65 viii DIESEL ENGINE DESIGN CHAPTER VI rLY-WHBELS The functions of a fly-wheel — Fly-wheel effeet — Twisting moment diagrams — Degree of uniformity — Momentary governing — Alternators in parallel — Torsional oscillations and critical speeds — Moment of inertia — Types of fly-wheels — Strength calculations ........ Page 97 CHAPTER VII FRAMBWOEK "A" frame type of framework — Crank-case t3^e — Trestle type — Staybolt type — Main bearings — Lubrication — Strength of bed- plate — Cam-shaft drive — Bedplate sections— Strength of "A" frames — Design of crank-cases — Machining the framework . 119 CHAPTER VIII CYLINDBBS AND COVBBS Types of cylinders — Liners — Jackets — Cylinder lubrication — Types of cylinder cover — Strength of cylinder covers . . 141 CHAPTER IX ETJNNING GEAR Trunk pistons — Shape of piston crown — Proportions of trunk pistons — Gudgeon-pins — Water-cooled trunk pistons — Pistons for crosshead engines — Systems of water cooling — Piston rods — Crossheads and guides — Guide pressure diagrams — Connect- ing rods — Big ends — Small ends — Strength of connecting rods Big end bolts — Indicating gears ..... 161 CHAPTER X FUBL OIL SYSTBM External fuel system — Fuel system on engine — Fuel pumps — Governors — Governor diagram — Fuel injection valves — Open type — Augsburg type — Swedish type — Burmeister type — Flame plates and pulverisers — ^Design of fuel valve casing and details .... ^ .... . 190 CHAPTER XI AIR AND EXHAUST SYSTEM Air suction pipes — Suction valves — Exhaust valves — Exhaust lifting devices — Proportions of air and exhaust valves and casings — Exhaust valve springs — Inertia of valves and valve gear — Spring formulae — Exhaust piping — Silencers — Scav- engers — Scavenge valves — Controlled scavenge ports — Ex- haust systems for two stroke engines ..... 220 CONTENTS ix CHAPTER XII COMPEESSED AIB SYSTEM Functions of the air-blast — Capacities of blast-air compressors — Number of stages — Compressor drives — Construction of com- pressor cylinders — ^Volumetric efficiency — Stage ratios — Trans- mission of heat through plates — Air reservoirs — Air-bottle valves — Blast pipe system — Starting pipe system — Starting valves — Air motors ....... Page 243 CHAPTER XIII VALVE GEAB, Cams — Cam rollers — Valve levers — Fulcrum shafts for valve levers — Push rods — Cam-shafts — Cam-shaft bearings and brackets — Cam-shaft drives — Reversing gears — Sliding cam- shaft type — Twin cam-shaft type — Twin roller type — Selective wedge type — Two stroke reversing gears^Moving roller type — ^Manoeuvring gears — Interlocking gears — Hand controls . 263 Index 297 LIST OF ILLUSTRATIONS No. PAGE 1. Curve connecting compression pressure and temperature for various values of " n " 4 2. Indicator card from four stroke Diesel Engine .... 6 3. Light spring card from four stroke Diesel Engine .... 7 4. Valve-setting diagram for four stroke engine 8 5. Valve-setting diagram for two stroke engine 9 6. Diagram showing simple port scavenge 10 7. Diagram showing controlled port scavenge 12 8. Ideal indicator diagram 20 9. Fuel consumption curves for ideal and actual engines ... 24 10. Entropy temperature diagram 30 11. Calibrated indicator diagram 32 12. Diagram illustrating flow of gases through orifices ... 38 13. Curve relating suction-pressure and throat velocity ... 40 14. Curve relating scavenge pressure and throat velocity . . 41 15. Valve area diagram 42 16. Diagram showing position of piston at various crank angles . . 43 17. Valve and port area diagram for two stroke engine ... 44 18. Light spring indicator card from two stroke engine ... 52 19. Three cylinder slow speed land engine 56 20. Weight per B.H.P. of slow speed engines 57 21. Built up crank-shaft 65 22. Solid shaft for two-cylinder engine .66 23. Orders of firing for fotir stroke engines 68 24. Orders of firing for two stroke engines 68 25. Section through ring lubricator 70 26. Crank fitted with centrifugal ring lubricator 70 27. Longitudinal and transverse oil holes in crank-shaft . . . 71 28. Diagonal oil holes in crank webs . 71 29,30,31. Various shapes of crank webs 71 32, 33, 34. Methods of securing balance weights 72 35. Cast iron flanged coupling ........ 73 36. Solid forged coupling 73 37. 38. Compressor cranks .74 39. Diagram showing loads on four-crank shaft 80 40. Diagram showing deflections of imiform cantilever . . . 82 41. Construction for deflected shape of beam ... .82 42. Diagram for obtaining coefficients c™, e^, g„ etc. . 83 43. Loads on four-crank shaft — ^No. 1 firing 84 44. Loads on four-crank shaft — ^No. 2 firing 86 45. Diagram showing positions of connecting rod 90 46. Calibrated indicator card 91 47. Cylinder pressure curve on crank-angle base 93 48. Twisting moment ctirves for four cylinder engine .... 93 49. Portion of twisting moment curve 98 xi xii DIESEL ENGINE DESIGN No. PAOB 50. Resultant twisting moment curve for one cylinder, four stroke engine 101 51. Resultant twisting moment curve for two cylinder, four stroke engine 102 52. Resultant twisting moment curve for four cylinder, four stroke engine 102 53. Resultant twisting moment curve for eight cylinder, four stroke engine 102 54. Resultant twisting moment curve for three cylinder, four stroke engine 102 55. Resultant twisting moment curve for six cylinder, four stroke engine 102 56. Resultant twisting moment cxirve for one cylinder, two stroke engine 103 57. Resultant twisting moment curve for two cylinder, two stroke engine 103 58. Resultant twisting moment curve for four cylinder, two stroke engine 103 59. Resultant twisting moment curve for eight cylinder, two stroke engine 103 60. Resultant twisting moment curve for three cylinder, two stroke engine 103 61. Resultant twisting moment curve for six cylinder, two stroke engine 103 62. Part of resultant twisting moment curve for three cylinder, four stroke engine, 20" bore X 32" stroke 106 63. 64. Diagrams for obtaining angular variation 106 65. Uniform shaft and fly-wheel 109 66. Diagram showing disposition of masses in a shaft system . . Ill 67. Construction for finding radius of gyration of a soUd of revolution 113 68. Solid disc fly-wheel 114 69. Disc fly-wheel with loose centre 114 70. Large fly-wheel in two pieces 114 71. Fly-wheel 115 72. Fly-wheel 117 73. " A " type of framework for four stroke engine . . . .119 74. " A " type of framework for two stroke engine .... 120 75. Four stroke " A " column with tie-rod 121 76. Four stroke " A " column with tie-rod and crosshead guide . 121 77. " A " column for foiu: stroke marine engine 120 78. " A " column for two stroke engine, with crosshead . . . 120 79. Crank-case framework for four stroke engine 122 80. Steel staybolts for crank-case 122 81. Crank-case framework for two stroke engine 122 82. Crank-case framework for crosshead engine (low tjrpe) . . . 123 83. Crank-case framework for crosshead engine (high tsrpe) . . 123 84. Trestle type of framework .% 124 85. Staybolt framework for trunk engine "... . 125 86. Staybolt framework for crosshead engine 125 87. Oil-catcher for main bearings 128 88. Forced lubricated bearings 128 89. Main bearing girder . . 129 90. Lower spiral drive . . . .131 91. Bedplate sections . . 132 92. " A " column 135 93. Crank-case ....... .... 137 94. Types of crank-case construction 138 LIST OF ILLUSTRATIONS xiii No. FACE 95. Cylinder liner jacket and cover 141 96. Integral liner and jacket 141 97. Integral liner jacket and cover . . ■ 142 98. Liner for four stroke trunk engine 143 99. Liner for four stroke crosshead engine 143 100. Upper flange of cylinder jacket 146 101. Cylinder jacket for four stroke engine 147 102. Cylinder jacket for two stroke engine 149 103. Cylinder jacket for two stroke engine 149 104. Cylinder jacket for two stroke engine . . ... 149 105. Cylinder lubricating fittings 150-2 106. Old type of cylinder cover 153 107. Modem type of cylinder cover 154 108. Cylinder cover with separate top plate 155 109. Water tube fitting 156 110. Water inlet in side of cover 156 111. Section of cylinder cover 157 112. Cylinder cover for two stroke engine 158 113. Cylinder cover for two stroke engine 159 114. Cylinder cover for two stroke engine 160 115. Types of piston crown 162 116. Diagram showing temperature gradient in piston . . . .163 , 117. Water-cooled pistons 163 118. Loose piston top ...... . . . 164 119. Device to obviate cracking of the piston crown . . . .164 120. Diagram showing proportions of trunk pistons . . .164 121. Means of locating piston rings - 165 122. Tapering of trunk pistons 166 123. Reduction of piston surface round gudgeon-pin bosses . . .166 124. Alternative forms of gudgeon-pins 167 125. SUt system of gudgeon-pin lubrication 168 126. Piston crown 168 127. Piston cooling systems 169 128. Pistons for four stroke crosshead engines ..._.. 171 129. Piston cooling systems 171 130. Piston for two stroke crosshead engine 172 131. Piston for two stroke crosshead engine 172 132. Piston for two stroke crosshead engine 172 133. Lower end of piston rod 174 134. Lower end of piston rod 174 135. Crosshead guide block 175 136. Diagram showing relation between guide pressure and twisting moment 175 137. Forms of connecting rod big end 177 138. Cast iron big end " brasses " 178 139. Spigoted big end 178 140. Ring for relieving big end bolts of shear stress .... 178 141. Types of connecting rod small ends 179 142. Top end of marine connecting rod 180 143. Small end with drag links 181 144. Curves showing variation of transverse stress in connecting rods, due to inertia ... 181 145. Diagram showing deviation of connecting rod thrust, due to journal friction 182 146. Buckled connecting rod 182 147. Resultant thrust when friction at journal =0 183 148. Resultant thrust when friction at crank-pin =0 . . . .183 xiv DIESEL ENGINE DESIGN No. ^AGE 149. Outline of conneoting rod 185 150. Diagram showing unequal pull in big end bolts .... 188 151. Indicator gears 189 152. External fuel system for land engine 190 153. Diagrammatic arrangement of fuel pumps and governor . . 193 154. Diagrammatic arrangement of fuel pump, governor, and fuel distributors 193 165. Fuel distributor 194 156. High pressure pipe union 194 157. Fuel pimip for markie engine 195 158. Fuel pump for one cylinder 196 159. Fuel pump ~ ... 197 160. Fuelpump 198 161. Horizontal fuel pump 199 162. Horizontal fuel pump 201 163. Diagram showing hand control of fuel pump valves . . . 201 164. Control handle for fuel pump 201 165. Plunger connections 202 166. Plunger packing 202 167. Fuel pump valves 202 168. Hand plunger 203 169. Diagram of fuel pump delivery 204 170. Governor 205 171. Diagram of governor controlling forces 207 172. Speed varying device 208 173. Speed varying device 208 174. Arrangement of governor and fuel pump gear 209 175. Arrangement of governor and fuel pump gear 209 176. Arrangement of governor and fuel piunp gear 210 177. Open type of fuel valve 211 178. Augsburg type of fuel valve 212 179. End of fuel valve 213 180. Sleeve type of pulveriser 214 181. Swedish type of fuel valve 215 182. Burmeister fuel valve 216 183-86. Lever arrangements for fuel valve operation .... 217 187. Air suction strainer 220 188. Suction pipe common to a number of cylinders .... 221 189. Suction valve casing 222 190. Cast iron exhaust valve heads 224 191. Water-cooled exhaust valve casing 224 192. Valve spindle guides 225 193. Valve casing with renewable seat 226 194. Exhaust lifting devices (hand) 227 195. Exhaust lifting devices (mechanical) 228 196. Exhaust valve proportions 229 197. Valve directly operated by caA 230 198. Velocity and acceleration curves for tangent cam .... 232 199. Diagram of valve and levers 233 200. Exhaust system for land engine 235 201. Exhaust system for land engine 236 202. Cast iron silencer 237 203. Scavenge pump driven off crank-shaft 238 204. Lever drive for scavenge pump 238 205. Combined scavenge and L.P. compressor cylinders . . . 238 206. Scavenge valve 240 207. Valves for controlling scavenge porta 241 LIST OF ILLUSTRATIONS xv No. PAQE 208. Silencer 242 209. H.P. piston-rings for air compressor 246 210. Heat flow diagram 250 211. Diagrammatic arrangement of H.P. air-bottles .... 252 212. Valve for bottle-head 253 213. Air-bottle with head screwed to neck 254 214. Blast air pipe for land engine 255 215. Blast shut-off valve 255 216. Starting pipe for fom: cylinder land engine 256 217. Flange for starting pipe 256 218. Tee-piece for starting pipe 256 219. Starting valve 257 220. Starting valve 259 221. Starting valve 259 222. Diagrammatic arrangement of Burmeister starting valve . 259 223. Pneumatic cylinder with oil bufier cylinder 260 224. Rotary air motor 261 225. Exhaust cams 263 226. Double cam for reversing engine 265 227. Fuel cam-piece 265 228. Tangent exhaust cam profile 265 229. Smooth cam profile 265 230. Construction for cam profile 265 231. Profile based on clearance circle 265 232. Starting cam 266 233. Valve roller 266 234. Cast iron valve roller 266 235. Fulcrum shaft and valve levers 267 236. Diagram of starting, neutral and running positions . . . 268 237. Sections of valve levers 269 238. Forked ends of levers 269 239. Tappet screws 270 240. Swivel tappet 270 241. Dimensions of levers and fulcrum spindle 271 242. Outline of exhaust lever 272 243. Section of lever 272 244. Plan of lever 272 245. Push-rod ends 273 246. Stepped cam-shaft and cam-trough 274 247. Cam-shaft with separate bearings 274 248. Cam-trough 275 249. Spiral drive for cam-shaft 276 250. Vertical shaft couplings 277 251. Spiral and bevel drive for cam-shaft 278 262. Spin- drive for cam-shaft 278 253. Spiu' and coupling rod drive for cam-shaft 278 254. Suspended gear-box 279 255. Diagram for spiral gears 280 256. Reversing gear for four stroke engine 281 257. Reversing gear for four stroke engine 282 258. Mechanism for sliding cam-shaft 282 259. Double cam-shaft reversing gear 283 260. Twin roller reversing gear 284 261. Reversing gear for starting valves 285 262. Selecting wedge type of reversing gear 285 263. Valve settings for two stroke angina 286 264. Reversing gear for two stroke engine 287 xvi DIESEL ENGINE DESIGN No. PAGE 265. Reversing gear for two stroke engine 287 266. Reversing gear for two stroke engine 289 267. Reversing gear for two stroke engine 290 268. Reversing gear for two stroke engine 291 269. Diagrammatic arrangement of manoeuvring gear .... 293 270. Interlocking gear for parallel shafts . . .... 295 271. Interlocking gear for shafts at right angles 295 DIESEL ENGINE DESIGN CHAPTER I FIRST PRINCIPLES The Diesel Principle. — ^The characteristic feature of the Diesel Engine is the injection of oil fuel into air which has been previously compressed by the rising of a piston to a pressure corresponding to a temperature sufficiently high to ensure immediate ignition of the fuel. In the course of the pioneer experiments by which the commercial practicability of this engine was demonstrated, it was found advantageous to effect the injection of the fuel by a blast of air, and this feature was retained in all Diesel Engines until the lapse of the original patents. At the present date there exists a class of high-compression oil engines operating on the Diesel principle, in which the injection of oil is effected by mechanical means without the assistance of an air blast. The design of these engines presents a variety of problems differing materially from those which arise in the design of Diesel Engines as defined below. Further- more, the use for war purposes of one of the most conspicuous members of this class of engine prohibits anything like a satis- factory discussion of these so-called " solid injection " engines. The well-known " surface ignition," " hot bulb," or " hot plate " engines form a very numerous class by themselves and have in the past been misnamed " semi-Diesel " engines. The cycles on which they operate and the principles underlying their design differ so widely from those relating to Diesel Engines proper that they also fall outside the scope of this work. The features which characterise the true Diesel Engine, in the correct use of the term, are now understood to be the following : — DIESEL ENGINE DESIGN (1) Compression sufficient to produce the temperature requisite for spontaneous combustion of the fuel. (2) Injection of fuel by a blast of compressed air. (3) A maximum cycle pressure (attained during combustion) not greatly exceeding the compression pressure, i.e. absence of pronounced explosive effect. Item 3 is deliberately worded somewhat broadly as the shape of a Diesel indicator card is subject to considerable variation under difEerent conditions of load, blast air pressure, fuel valve adjustment, etc. In the earlier days of Diesel Engine construction the square top indicator card, showing a period of combustion at constant pressure, was considered the ideal to aim at. It has since been found that a card having a more peaked top is usually associ- ated with better fuel consumptions. When tar oil is used as fuel the square top card appears to be almost out of the question. It should further be remembered that the existence of a period of combustion at constant pressure is no guarantee that all the combustion takes place at that pressure. This ideal is never realised. Combustion probably proceeds slowly well after half stroke, even under the most favourable conditions. Compression Pressure. — The height to which compression is carried is governed by the following considerations : — (1) The attainment of the requisite temperature. (2) The attainment of a desirable degree of efficiency. (3) Mechanical considerations. Considerations of temperature for ignition fix the lower limit of compression at somewhere in the neighbourhood of 400 lb. per sq. in. The temperature actually attained depends on the initial temperature of the intaken air and the heat lost to the jacket during compression, so it is clear that the temper- ature attained on the first few strokes of the engine will be considerably lower than the'^alue it assumes after the .engine has been firing consecutively for some time. As regards efficiency, it is well known that increasing the degree of compression beyond certain limits does not very materially increase even the theoretical efficiency. In practice the compression most usually adopted is about 500-550 lb. per sq. in. for four stroke engines. For two stroke engines the compression is frequently in the neighbourhood of FIRST PRINCIPLES 600 lb. per sq. in. or over, owing to the fact that the charge of air delivered by the scavenge pump may itself be at a pressure slightly above atmospheric. At first sight it might be thought that this initial compression might be considered as a first stage in the temperature rise of the charge of air ; but apparently compression in the scavenge pump is not so effective in raising the temperature as compression in the main cylinder. The mechanical considerations which limit the compression are numerous, and some are mentioned below. Higher compression involves : — (1) Heavier load per sq. in. of the piston and necessitates massive construction of all the main parts. (2) More highly compressed air for injection and consequently increased trouble with the air compressor, and its valves particularly. (3) Increased wear of cylinder liners due to increased pressure behind the piston rings. Compression Temperature. — ^With a compression of 500 lb. per sq. in. in a fair-sized four cycle cylinder working under full load conditions the compression temperature is about 1200° F. On starting the engine from a cold state the com- pression pressure and temperature are considerably lower owing to the cold state of the cylinder walls and the piston crown. In addition to this the injection of cold blast air with the fuel in the proportion of about 1 lb. of blast air to 12 lb. of suction air still further reduces the temperature apart from the probability that the blast air has momentarily a local cooling effect in the zone of combustion. The middle curve (Fig. 1) shows graphically the connection between the compression temperature and compression pressure on the assumptions that : — (1) The initial temperature of the intaken air is 212° F. (2) That the exponent in the equation PV''=const. is 1-35. These assumptions correspond approximately to the condi- tions obtaining with a fully loaded engine of fair size — say an 18" cylinder. The noteworthy point about this curve is the slowing down of the rate of increase of temperature with pressure as the latter increases. Expressed mathematically dT diminishes as P increases. dP DIESEL ENGINE DESIGN The Four Stroke Cycle. — The well-known four stroke cycle consists briefly of : — (1) The Suction Stroke. (2) The Compression Stroke. (3) The Combustion and Expansion Stroke. (4) The Exhaust Stroke. These are considered in detail below. ■noo 100 200 300 too 500 600 100 Compression-Lbs per Sq. In Gauj«Pressure , Fig. 1. Suction Stroke. — If the engine crank is considered to be at its inner dead centre and just about to begin the suction stroke, the suction valve is already slightly open. In steam engine parlance it has a slight lead. At the same time the exhaust valve, which has been previously closing on the exhaust stroke, has not yet come on its seat. The result of this state of afEairs is that the rapidly moving exhaust gases create a partial vacuum in the combustion space and induce a flow of air through the suction valve, thus tending to scavenge out exhaust gases which would otherwise remain in the cylinder. As the piston descends its velocity increases and reaches a maximum in the neighbourhood of half stroke. At the same time the suction valve is being lifted further off its seat and FIRST PRINCIPLES attains its maximum opening also in the neighbourhood of half stroke. The lower half of the suction stroke is accom- panied by a more or less gradual closing of the suction valve, which, however, is not allowed to come on its seat until the crank has passed the lower dead centre by about 20°. At the moment when the crank is passing the lower dead centre the induced air is passing through the restricted opening of the rapidly closing suction valve with considerable velocity and an appreciable duration of time must elapse before the upward movement of the piston can effect a reversal of the direction of flow through the suction valve. It will be clear from the above that owing to the effect of inertia more air will be taken into the cylinder in the manner described than by allowing the suction valve to come on its seat exactly at the bottom dead centre. The exact point at which the suction valve should close is doubtless capable of approximate calculation, but is usually fixed in accordance with current practice or test-bed experiments. Compression Stroke. — The piston now rises on its up stroke and compresses the air to about 500 lb. per sq. in. The clearance volume necessary for this compression being about 8% of the stroke volume. During the compression the ternper- ature rises and a certain amount of heat is lost to the cylinder walls and cylinder cover. The final compression temperature is in the neighbourhood of 1200° F. Combustion and Expansion Stroke.— At the upper dead centre, or slightly previous thereto, the injection valve opens and fuel oil is driven into the cylinder and starts burning im- mediately. The actual point at which the fuel enters the cylinder is not qtiite certain, as there is inevitably some lag between the opening of the injection valve and the entrance of fuel. The point at which the fuel valve starts to open, as determined by a method described below, varies from about 3° (slow speed engines) to 14° (high speed engines). The method of determining the point of opening of the fuel valve is as follows : — With the engine at rest, air at about 100 lb. pressure is turned on to the injection valve and then communication with the blast air-bottle is cut off to prevent unnecessary waste of air and the possibility of the engine turning under the impulse of the air which is subsequently admitted to the cylinder. The indicator cook is now opened and the engine slowly barred DIESEL ENGINE DESIGN round by hand until the air is heard to enter the cylinder by placing the ear to the indicator cock. The position of the engine when this occurs is the nearest possible approximation to the true point of opening, assuming the operation has been oarefuUy done. The duration of the fuel valve opening is usually about 48°, and in the majority of engines is fixed for aU loads. It is evident that at light load the opening is longer than necessary, and in some designs the duration of opening is regulated by the governor in accordance with the load. The combustion is by no means complete when the fuel valve closes, and usually continues in some measure well past the half stroke of the engine. This is known as " after burning," Fio. 2. and takes place with the very best engines in the best state of adjustment. Exaggerated after burning is the surest sign of misadjustment, and makes itself apparent by abnormally high terminal pressure at the point at which the exhaust valve opens, and is readily detected on an indicator card by com- parison with that taken from an engine in good adjustment. As will be shown later, the presence of " after burning " is most clearly seen on an Entropy Diagram. Expansion continues acc(mipanied by loss of heat to the cylinder walls until the exhaust valve opens. Exhaust Stroke. — The exhaust valve opens about 50° before the bottom dead centre, in order that the exhaust gases may efEect a rapid escape and reduce the back pressure on the exhaust stroke. The pressure in the cylinder when the exhaust valve starts to open is about 40 lb. per sq. in. with an engine working with a mean indicated pressure of 100 lb. per sq. in. The temperature of the exhaust gases at this point is some- FIRST PRINCIPLES where in the neighbourhood of about 1600° F. and the velocity is consequently very high. The pressure falls nearly to atmospheric shortly after the bottom dead centre has been passed and the back pressure during the remainder of the exhaust stroke should not be more than about 1 lb. per sq. in., or less. Excessive back pressure may arise from : — (1) Insufficient diameter or lift or late opening of exhaust valve. (2) Exhaust pipe too small in diameter. (3) Obstructions or sharp bends in the exhaust pipe or silencer. (4) Interference by another cylinder exhausting into the same pipe. It is interesting to note that owing to the higher velocity of air at high temperature per unit pressure difference the back pressure is more at light load than at full load. Indicator Cards. — Figs. 2 and 3 show typical indicator cards taken with a heavy and a light spring respectively. The latter is particularly useful for investigating the pro- cesses of suction and exhaust. It is to be observed that in Fig. 3 the compression is seen to start at a point which is indistinguishable from the bottom dead centre, thus indicating a volumetric efficiency of practi- cally 100%. This is to be regarded as a normal state of affairs, obtainable with both high speed and low speed engines. The volumetric efficiencies of internal combustion engines are frequently quoted at figures varying between about 95% for slow speed engines to 80% for high speed engines. The former DIESEL ENGINE DESIGN figure is reasonable, but the latter can only be due either to imperfect design (or adjustment) of the engine or to erroneous indicator cards. The use of too weak a spring in the indicator may lead to a diagram showing not more than 60% volumetric eflftciency, owing to the inertia of the indicator piston, etc. Consequently, fairly stifE springs are to be preferred. Valve Setting Diagram. — Fig. 4 is a typical valve setting diagram for a four stroke engine, and shows the points relative to the dead centres at which the various valves open and close. Fuel Valve Open§_ Suction Valve Opens, ^^ 'xhaust Valve Closes fuel Valve Closes Suctfan Valve Closes Exhaust Valve Opens Fig. 4. The Two Stroke Cycle. — As its name implies, the two stroke cycle is completed in one revolution of the engine. The revolution may roughly be divided into three nearly equal parts : — (1) Combustion and Expansion. (2) Exhaust and Scavenged (3) Compression. The exhaust and scavenge take place when the piston is near the bottom dead centre, and consequently only very small portions of the expansion and compression strokes are lost, in spite of the fact that nearly 120° of the crank revolution are occupied with exhaust and scavenge. This point is clearly seen on reference to Fig. 5. FIRST PRINCIPLES 9 Exhaust Period. — The exhaust starts when the piston un- covers slots in the cylinder wall. The point at which this happens is different in different designs of engine, an average being about 15% of the stroke before bottom dead centre. The exhaust ports are usually of large area, and consequently the pressure falls to atmospheric very rapidly. The period required for this process naturally depends on the port area and the piston speed, and average figures are about 20° to 30°. It is well to dwell carefully on the state of affairs at this point. Fuel Yslve Closes Admission of Scairenge Air Ceases Exhaust Ports Close Exhaust Port s Open Scavenge Air is Admitted Valve Setting 2 Stroke Diag ram for Engine . Fic. 6. During exhaust the cylinder pressure has fallen from about 55 to about 15 lb. per sq. in. absolute, and there is no reason to suppose that the remaining exhaust gases have fallen greatly in temperature. (Given adiabatic expansion, the fall in absolute temperature is less than 20%.) The conclusions are, therefore : — (1) Something like 50% by weight of the gases have effected their escape. (2) The remaining gases are rarefied compared with atmo- spheric air. Scavenge Period. — The scavenge air is admitted by ports or valves (or both), and the instant at which admission starts is timed to coincide with that at which the cylinder contents 10 DIESEL ENGINE DESIGN attain appreciably the same presstire as the scavenge air, or a trifle less. The incoming scavenge air is supposed to sweep the remaining exhaust gas before it and so fill the cylinder vfdth a charge of pure air by the time the piston has covered the exhaust slots on the up stroke. Actually certain processes take place which do not enter into the ideal programme. Some of these are : — (1) A certain amount of mixing between the incoming scavenge air and the retreating exhaust gases. ( 2) Short circuiting of scavenge air to the exhaust pipe before all the exhaust gas has been expelled. The effects of both these processes are minimised by provid- ing a large excess of scavenge air. The figure adopted for the ratio of scavenger volume to cylinder volume was about 1 -4 in the earlier designs, but later experience points to the advis- ability of increasing this figure to about 1-8. There are a number of different systems in use for admit- ting scavenge air, and some of these are discussed below. Simple Port Scavenge. — In this system the scavenge air is admitted by means of ports in the cylinder liner opposite a Exhaust Ports Scavenge Ports Diagrammatic Arrangement of Simpleton Scavenge Fit Fio. 6. row of similar ports for the exhaust (see Fig. 6), the piston top being provided with a projection to deflect the scavenge air to the top of the cylinder. This system is simple but possesses some disadvantages which are enumerated below. (1) The scavenge air slots have to be made shorter than the exhaust slots in order that the cylinder pressure may fall to FIRST PRINCIPLES 11 the same value as the scavenge air pressure before the piston begins to uncover the scavenge slots. This entails the latter being covered by the piston on its upward stroke before the exhaust ports are covered, and consequently the pressure at the beginning of compression can barely exceed the pressure in the exhaust pipe. There is also a possibility of exhaust gases working back into the cylinder. (2) The projection on the top of the piston necessitates an irregularly shaped cylinder cover in order to provide a suitable shape for the combustion space. Engines provided with this system of scavenge are only suited for a relatively low mean indicated pressure of about 80 lb. per sq. in. Cylinder Cover Valve Scavenge. — In this system the scavenge air is admitted by means of one to four valves located in the cylinder cover, and avoids some of the disadvantage of the simple port scavenge. By allowing the scavenge valves to close after the exhaust ports have been covered by the piston, the cylinder may become filled with air at scavenge pressure before compression starts, and consequently such a cylinder is capable of develop- ing a higher mean effective pressure. The greatest drawback to this system is the complication of the cylinder cover, and this appears to be rather serious. Strenuous efforts are being made to design suitable cylinder covers for this type of two stroke engine, but with the exception of one or two designs, which are now in the experimental stage, most of them appear to have a comparatively short life. Valve Controlled Port Scavenge. — This system (usually associated with the name of Messrs. Sulzer Bros.) appears to combine most of the advantages of both with the disadvantages of neither of the above systems. The air ports, or a certain number of the air ports, are bo situated that they are uncovered before the exhaust ports, but are controlled by a valve of the double beat, piston, or other type, in such a way that communication does not exist between the cylinder and the scavenge pipe until the exhaust ports have been uncovered for a sufficient period to allow the cylinder pressure to fall to or below the scavenge pressme. On the up- ward stroke the controlling valve remains open, so that the cylinder is in communication with the scavenge pipe until the scavenge slots are covered by the piston. (See Fig. 7.) 12 DIESEL ENGINE DESIGN Engines controlled on this principle are at present the most successful of the large two stroke Diesel Engines. Amongst small or medium powered two stroke engines the simple port scavenge principle is the favourite. Sca\/eng^ Ports Diagrammatic Arrangement of 2 Stroke Cylinder with Controlled Port Scavenge, Fig. 7. Types of Diesel Engines. — The existing types of Diesel Engines can be divided into groups in various ways, according as they are : — (1) Stationary or Marine. (2) Four Cycle or Two Cycle. (3) Slow Speed or High Speed. (4) Vertical or Horizontal. (5) Single Acting or Double Acting. It suffices here to describe shortly the outstanding features of the commonest types in commercial use. Four Stroke Slow Speed Stationary Engines. — ^There are probably more Diesel Engines falling under this heading than any other. The usual Mrangement is a vertical engine having one to six cylinders provided with trunk pistons. The main features are (1) a massive cast-iron bedplate made in one or two pieces, the bottom being formed in the shape of a tray to catch any lubricating oil which is thrown out of the bearings or drips from the cylinders. (2) Massive columns forming the cylinder jacket into which the liners are pressed. (3) A heavy fly-wheel of large diameter, necessitated by the low speed of revolution. FIRST PRINCIPLES 13 These are perhaps the most rehable Diesel Engines hitherto constructed, and differ but little from the machines made by the Maschinenfabrik Augsburg in the very earliest days of Diesel Engine manufacture. For economy in fuel oil, and lubricating oil consumption these engines are unsurpassed, and the massiveness of their construction secures for them a long, useful life. Somewhat similar engines are made in horizontal form, in accordance with popular gas engine practice. The horizontal engine is obviously suitable where head room is limited, and it is a trifle cheaper to manufacture than the vertical design. The latter, however, is still the general favourite, and is made in very large numbers in standard sizes, varying from 8 to 200 B.H.P. per cylinder. Four Stroke High Speed Stationary Engines. — ^The high speed Diesel Engine was introduced in order to cheapen manu- facture and to provide a more compact prime mover for use in limited spaces. Forced lubrication is usually adopted, and an enclosed crank-case is therefore provided. The fly-wheel is of course much smaller and lighter than that of a slow speed engine of the same power. In other respects the design follows fairly closely that of the slow speed engine only that additional care has to be exercised in the design of those details in which inertia plays an important part. The fuel consumption is practically the same as that of a slow speed engine, but the lubricating oil consumption is usually higher and the useful life of the engine somewhat shorter. Two Stroke Slow Speed Stationary Engines. — ^These are built in sizes varying from about 200 to 700 B.H.P. per cylinder. Their chief advantage over the four cycle equivalent is reduced cost and size. The fuel consumption is slightly higher than that of the four stroke engines of the same power and the cost of upkeep would also appear to be slightly greater. These engines are usually fitted with a cross-head, and this enables forced lubrication to be used without an unduly high con- sumption of lubricating oil. The use of cross-heads is also becoming increasingly common with four cycle engines, and this appears to be a move in the direction of increased reliability. Other Types of Stationary Engines. — Horizontal two stroke Diesel Engines have been made in small numbers. Examples of double-acting horizontal engines are also to be 14 DIESEL ENGINE DESIGN found, but the three types described above appear to fulfil most requirements for land service. Four Stroke Marine Engines. — Most of the successful marine Diesel Engines are of this type. Cross-head and guides are usually fitted and forced lubrication is preferred. The piston speed is usually moderate, being about 700 to 850 per minute. In Messrs. Burmeister, Wain's well-known design the bedplate and columns form an enclosed box construction of great rigidity and strength. The total enclosure of the crank- pase is also conducive to economy of lubricating oil. In other respects the engines follow land practice very closely in almost every essential point of design. It has been frequently con- tended in the past that the failure of certain marine Diesel Engines has been almost entirely due to lack of knowledge of the requirements of marine service. The actual facts of the case appear to be that almost every complete failure of a marine Diesel Engine has been due to causes which would have led to the same result on land. Under the category of failures must be placed a number of large two stroke Diesel Engines, which have been installed on board ship before they had proved their reliability on land. The actual modifications required to convert a successful land design into a successful marine design are comparatively trivial. When once the essential difficulties of Diesel Engine construction have been successfully sur- mounted the adaptation to the requirements of marine service is a small matter in comparison. Two Stroke Marine Engines. — ^As indicated in the pre- vious paragraph, there have been many failures with large two stroke engines. In cases where the manufacturers have had no experience of marine work, they appear to have been un- duly influenced in favour of accepted marine traditions. Where the engine has been constructed by marine steam engineers, under licence from a Diesel Engine manufacturing firm, the result has been similax. The machines resulting from this fusion of ideas have generally resembled marine steam engines with Diesel Engine cylinders. The columns, guides and bedplate are scarcely to be distinguished from those of a marine steam engine. The open crank-pit, adopted out of deference to steam engine practice, appears to lead to an excessive consumption of lubricating oil, and renders the economical use of forced lubrication impossible, besides facilitating the ingress of dirt. Another feature borrowed from FIRST PRINCIPLES 15 steam practice consists of the rocking levers driven off the cross-head for working the scavenge air pumps. These have not always been quite successful. These and other circum- stances all seem to point to the advisability of developing the marine Diesel on lines a little more independent of steam engine precedent. At present the two stroke marine Diesel Engine, of large power, must still be considered to be in the experi- mental stage (without prejudice to two or three isolated cases of large two stroke Diesel Engines at present operating success- fully). For moderate powers, up to about 500 H.P., two stroke engines have been fairly widely and successfully used. High Speed Marine Engines. — ^A large number of designs of both four stroke and two stroke high speed marine Diesel Engines have been described in the technical press of recent years. Many of these differ but little (except in the provision of a reversing gear and other details) from high speed land engines. Others, particularly of the two stroke type, have been developed along strongly individual lines for some special purpose, such as the propulsion of submarines. To this latter class belong the stepped piston engines, in which an extension of the main piston serves at once as a cross-head guide and a single acting scavenge pump piston. With one or two exceptions, these engines appear difficult to dismantle, and are not likely to become popular until this objection is removed. Literature.— Chalkley, A. P., "The Diesel Engine."— Con- stable. WeUs and Taylor, " Diesel or Slow Combustion Oil Engines." — Crosby Lockwood. Supine, Bremner and Richardson, "Diesel Engines." — Griffin. Milton, J. T., " Diesel Engines for Sea-going Vessels." — Inst. N. A., April 6th, 1911. Milton, J. T., " The Marine Diesel Engine."— Inst. N. A., April 2nd, 1914. Knudson, I., " Performance on Service of Motor Ship 8uecia."—ln6t. N. A., June 27th, 1913. CHAPTER II THERMAL EFFICIENCY The overall thermal efficiency of a heat engine is the ratio of the useful work performed to the mechanical equivalent of the heat supplied during a given period of working. Problem : What is the thermal efficiency of a Diesel Engine which consumes 0-4 lb. of fuel per brake horse-power hour ? To solve this problem two things require to be known : — (1) How much heat 1 lb. of fuel gives out on combustion, i.e. the calorific value of the fuel. (2) How much mechanical work is equivalent to a given quantity of heat. The calorific value of different qualities of liquid fuel varies from about 15,300 (Mexican crude) to about 19,300 (Galician crude) British Thermal Units per lb. For calculations and comparison of test results fuel consumptions are usually reduced to their equivalents at a calorific value of 18,000 B.T.U. per lb. One B.T.U. (the amount of heat required to raise the temperature of 1 lb. of water 1° F.) is equivalent to 778 ft. lb. This is Joules' equivalent. Therefore, since 1 H.P. hour = 1,980,000 ft. lb., the required thermal efficiency is equal to:— 1,980,000 ^ 0-4x18,000x778 ' From this it is seen that roughly one -third of the heat supplied has been converted into useful work. It is within the province of thermodynamics to determine what proportion of the heat loss is theoretically unavoidable, and to what extent the performance of an actual engine approximates to that of an ideal engine working on the same cycle of operations. It is not proposed to give here more than a brief summary of the physical laws relating to the behaviour of air under the in- 16 THERMAL EFFICIENCY 17 fluence of pressure and temperature, which form the basis of thermodynamic investigations. Pressure, Volume and Temperature of Air. — The relation between these three quantities is expressed by the formula : — P.V.=53-2 xw xT — (1) where P= Pressure in lb. per sq. ft. abs. V=Volume in cubic ft. w=Weight in lb. of the quantity of air under con- sideration. T=: Temperature in degrees abs. P. This relation holds good for any condition of temperature and pressure, and for a specified weight of air, given the values of any two of the quantities represented by capital letters, the third can be calculated. Example : Find the volume of 1 lb. of air at atmospheric pressure and 60° F. In this case P=14-7 lb. per sq. in. abs., T = 60 + 461=521° abs. F., and w = l. Hence: V=53-2x621 ,^ , . .^ -^-T-f, — =-- = 131 cub. ft. 147 xl44 Isothermal Expansion and Compression. — If the tempera- ture remains constant during compression or expansion the process is said to be isothermal, and the value of T in equation (I) becomes a constant quantity. Thence for isothermal pro- cesses equation (1) becomes: P.V.=constant (2). If 1 lb. of air is under consideration the value of the constant is equal to 63-2 times the absolute temperature. Work done during Isothermal Compression. — If P^ and Vi represent the pressure and volume before compression, and Pj and Vg the same quantities after expansion, then : — V Work done=const. log^ ^' (3), the constant being that of equation (2). It should be borne in mind (though the bare fact can only be stated here) that the internal energy of a gas depends on its temperature only, regardless of the pressure. It therefore follows that aU the work done in isothermal compression must pass away as heat through the walls of the containing vessel, and for this reason 18 DIESEL ENGINE DESIGN ^_ — - isothermal processes are not attainable in practice, though they may be approximated to by slow compression in cylinders arranged for rapid conduction of heat. Specific Heat at Constant Volume. — ^If 1 lb. of air is heated in a confined space, i.e. at constant volume, 0-169 B.T.U. are required to raise the temperatxu-e by 1° F. This then is the specific heat of air at constant volume. For many purposes it is near enough to consider the specific heat as constant, though actually its value increases sUghtly as the temperature increases. The amount by which the internal energy of 1 lb. of air increases as the temperature rises is there- fore : — 0-169 (Ta— Tj) where (T2—Ti)=the increase of tempera- ture. Specific Heat at Constant Pressure. — In this case, on the other hand, work is done by expansion if heat is being added, and by compression if heat is being discharged ; consequently the specific heat at constant pressure exceeds that at constant volume by the equivalent of the work done. The specific heat at constant pressure is 0-238 B.T.U. per lb. per degree Fahrenheit. Adiabatic Expansion and Compression. — Expansion or compression unaccompanied by the transfer of heat to or from the air is termed Adiabatic. It should be noted that the air may lose heat by doing external work or gain heat by having work done on it by the application of external force during an adiabatic process ; but this heat comes into existence or passes out of existence within the air itself and does not pass through the walls of the containing vessel. The following relation holds good between the pressure and the volume during an adiabatic process : — P V° =constant (4) where n =ratio of the two specific heats, \ With the information supplied by the above six equations it is possible to construct the Indicator Diagram corresponding to an Ideal Diesel Engine, in which all defects of combustion and heat losses to the cylinder walls are supposed to be eliminated. An Ideal Diesel Engine. — Before proceeding further, it is necessary to define what is to be understood by the term ideal engine for the purposes of this investigation. In the first place, frictional losses and all leakage are eliminated and compression and expansion are supposed to take place adiabat- ically (n= 1 -41). The specific heats are supposed to be constant and the same whether for air or a mixture of air or exhaust gases. On the other hand, the provision of compressed air for the purpose of fuel injection will be recognised, and conse- quently the machine will have a mechanical efficiency less than unity. The compression in the air compressor will be regarded as isothermal and the compressor itself free of all mechanical or other losses. It will also be supposed that all the exhaust gas, including that contained in the clearance space, is ejected on the exhaust stroke and that every suction stroke fills the cyHnder with air at 14-7 lb. per sq. in. abs. and 521° abs. F. The stroke volume is taken for convenience of calculation at 100 cub. ft. 20 DIESEL ENGINE DESIGN The Ideal Indicator Card. — ^The indicator card correspond- ing to this ideal engine is now readily constructed, and is shown in Fig. 8. Point A denotes a volume of air equal to 100 cub. ft. plus the clearance space, which has to be calculated. Line AB represents adiabatic compression from 14 -7 lb. per sq. in. to 514-7 per sq. in. abs. n.'^ Volume in Cubic Feet. ' / Fig. 8. Line BC represents increases of volume at constant pressure due to addition of heat by combustion of fuel. Line CD represents adiabatic expansion of the products of combustion and DA the fall in pressure due to exhaust release. It is not necessary to deal with the exhaust and suction strokes, as these are supposed to take place at atmospheric pressure. Determination of Clearance Value. — 41 From equation (4). P« 514-7 m- 100 This determines Points B and A. Construction of Line AB. — ^This is effected by calculating the pressure at various points of the stroke in accordance with the following schedule : — Pa ~ 14-7 "" \35/f4i 12-45 = 11-45 .-. Vb = 8-75 cub. ft THERMAL EFFICIENCY 21 1 2 3 4 s e 7 Per cent of stroke. V- cubic feet. Va V 1..;=' («X1.«. AntUoK (5) P=(6)X14-7 lb. per sa. in. 108-75 1-00 0-000 0-000 1-00 14-7 20 88-75 1-23 0-099 0-127 1-34 19-7 40 68-75 1-58 0-199 0-280 1-90 27-9 60 48-75 2-23 0-348 0-490 3-09 45-5 80 28-75 3-78 0-579 0-815 6-53 95-5 90 18-75 5-79 0-763 1-075 11-89 174-5 95 13-75 7-90 0-898 1-265 18-41 270-6 100 8-75 12-45 1-095 1-545 35-08 514-7 This simple calculation is given in detail, as it is typical. Determination of Point c. — ^The value of V^ depends of course on the amount of heat added. The efEect of the oil itself in increasing the weight of the working fluid will be ignored. The blast air, however, will be taken into consideration and assumed to be equivalent to 8 cub. ft. of free air.. The weight of air concej^ned has now to be calculated. (1) Suction air. PxV From equation (1) w=::;^^-;^ — — = 14:4:X (2) Blast air. 14-7x108-75 53-2xT 53-2x521 = 8-3lb. Weight of blast air= 8-3x8 =0-61 lb. 108-75 „ „ suctionair+blast air=8-3+0-61=8-91 lb. The efEect of the heat released by the combustion of the fuel is to increase the temperature of : — (1) The suction air (8-3 lb.), (2) The blast air (0-61 lb.), at a constant pressure of 514-7 lb. per sq. in. abs. The temperature of the blast air is taken to be 521° F. abs., and that of the adiabatically compressed air is found from equation (6) as follows : — Ta \Pa/ V 14-V 2-805 Therefore Tb=521 x 2-805 = 1460' F. abs. = 3,100° F. abs. 22 DIESEL ENGINE DESIGN Now let H=heat added in B.T.U. Then since the pressure remains constant, H==-238 [8-3 (To-1460) + -61 (T„-521)] = 2-12T„-2961 from which T,=^4^^ (7) " 2-12 and from equation (1) Now suppose that 0-2 lb. of fuel is added, having a calorific value of 18,000 B.T.U. per lb., the resulting temperature will be:— 0-2x18,000 + 2961 212 and V„ = -0064x 3100 = 19-8 cub. ft. The expansion line CD can now be constructed in the same way as the compression line, both being adiabatics. The value of the terminal pressure P^ is required for calculating the work done, and is found by means of equation (4) as follows : — p^ [yj V108-75/ -"^^ .-. Pa = -091 X 514-7 =46-8 lb. per sq. in. abs. Calculation of Work Done. — Referring to Fig. 8, the work done is clearly equal to the areas BCNM plus CDRN less BARM. Using equation (5) for the two latter, we have : — Area BCNM = 144 x 514-7 x (19-8 -8-75) = 819,000 ft. lb. „ CDRN = 144(514-7xl9-8-46-8xl08-74)^^g^^^^^^^^^ -41 By addition 2,619,000 „ „ BARM = 144(514-7x 8-7^14-7 xl08-75)_, ^„^ ^„^ ,, ^jr 9 — 1,UZU,UU0 „ ,, Work done by difference = 1,599,000 „ ,, Since 1 H.P. hour = 1,980,000 ft. lb. the fuel consumption is : — 0-2X1,980,000 „.oM. TTT-DT- 1,599.000 =-248lb.perI.H,P.hour. ,J [> THERMAL EFFICIENCY 23 From equation (3) the work done in compressing the blast air is given by 914-7 8 X 14-7 X 144xloge (assuming blast pressure of ^*"' 900 lb. per sq. in.). = 70,000 ft. lb. so the nett work done is 1,599,000-70,000=1,529,000 ft. lb. and the consumption of fuel per B.H.P. hour is : — 0-2x1,980,000 OKOIU T}TTT)U =-259 lb. per B.H.P. hour. 1,529,000 ^ The mechanical efficiency being ,'„„'„_ =95-6% 1,599,000 The M.I.P. of the diagram is equal to the indicated work divided by the stroke volume. 1,599,000 icnn/^iu i^ — ^ = 15,990 lb. per sq. ft. 100 f ^ = 111 lb. per sq. in. This is a trifle higher than the M.I.P. usually indicated at full load in an actual commercial engine. The brake thermal efficiency is given by : — ^'^SQ'OQQ =0-548 0-259x18,000x778 A splendid figure for the brake thermal efficiency of an actual engine of large size is 0-35 ; comparing this with the above figure, it is seen that the actual engine attains 64% of the efficiency attributed to the ideal. The indicated thermal efficiency of the ideal engine is : — 1,980,000 =0-573 0-247x18,000x778 That of an actual engine at full power is about 0-472, the ratio of actual to ideal being about 82%. From the above it will be seen that so far as the thermal actions within the cylinder are concerned, an actual Diesel Engine in good order leaves comparatively little room for improvement as long as the accepted cycle is adhered to. As a matter of fact, slight deviations from the constant pressure cycle are frequently made, and improved efficiencies obtained thereby. Most high speed Diesels, for example, are arranged to give an indicator 24 DIESEL ENGINE DESIGN card, showing a certain amount of explosive effect, i.e. combus- tion at constant volume, causing the pressure at the dead centre to rise to a figure which may be anj^hing up to about 100 lb. above the compression pressure. This is found to have a beneficial effect on the fuel consumption, which is readily explained on theoretical grounds ; but it is obvious that considerations of strength must limit the extent to which this principle is used. Fuel Consumption at Various Loads. — If the foregoing calculations for the fuel consumption of the ideal engine are 09 0-8 I .0-7 :i:-a (u m Q. u S;-oo-4 — ™ -a -> 0-1 n 1 \ \ , \j ^ 1 ^ ^Z ^ A " TT p Mechanical efficiency = ^ ^ _. — ".''., i.H.r. of mam cyunders. Continental engineers, on the other hand, sometimes subtract the indicated power of the compressor from that of the main cylinders for the purposes of the above equation. The usual British practice will be adhered to throughout this book. Variation of Mechanical Efficiency with Load. — Examina- tion of a large number of Diesel Engine test results reveals the fact that the difference between the I.H.P. and the B.H.P. remains nearly constant as the load is varied. This fact enables one to calculate the mechanical efficiency at any load if the full load efficiency is known. Example: What is the mechanical efficiency at three-quarter, half and quarter load if that at full load is 72% ? Let the full load I.H.P. be represented by 100, then the following tabulated figures hold good :- Const. Meoh. B.H.P. I.H.P. Diff. Effioy. Full load . 72 100 28 •72 Three-quarter load 54 82 28 •66 Half load . 36 64 28 •56 Quarter load . 18 46 28 •39 This method yields sufficiently accurate results for most estimating purposes. Mechanical Losses. — ^The work corresponding to the differ- ence between the I.H.P. and the B.H.P. is approximately accounted for in the two following tables, which apply to 26 DIESEL ENGINE DESIGN medium-sized engines of good design, and of the four stroke and two stroke types respectively : — Four Stroke Engine — Full Load Mech. Efficy., 75%. per cent Brake-work 75-0 Work done on suction and exhaust strokes . . 3-0 Indicated compressor work 5-8 Compressor friction 1'2 Engine friction, opening valves, etc. . . . 15-0 Work indicated in main cyUnders , . . . 100-0 Two Stroke Engine— Full Load Mech. Efficy., 70%. per cent Brake- work 70-0 Indicated compressor work 6-5 Compressor friction 1-4 Indicated scavenger work 7-5 Scavenger friction 1-6 Engine friction, opening valves, etc. . . . 13-0 Work indicated in main cylinders . . . 100-0 Improvements in bearings and guides on the principle of the well-known Michell bearing or the use of roller bearings for the main journals and big ends suggest possibilities for reducing engine friction which will possibly materialise in the future. The adoption of some form of limit piston ring to prevent e^essive pressure on the liners would also help matters in the 'same direction, besides increasing the life of the liners. Influence of Size on Mechanical Efficiency and Fuel Con- sumption. — Piston speed, within the range of present practice, appears to have very little influence on either mechanical efficiency or fuel consumption- This remark does not apply to the abnormally low speeds obtaining, for example, with a marine engine turning at reduced speed. The cylinder bore is the principal factor in economy, always assuming a reasonable ratio of bore to stroke and good design generally. The following table is a rough guide to the mechanical efficiency and fuel consumptions to be expected from cylinders of various sizes working at full load. THERMAL EFFICIENCY 27 FoTTR Stroke Two Stroke Cylinder £>iam., in. Meoh. Effcy. % Fuel per B.H.P. hr. lb. Fuel per I.H.P. hr. lb. Cylinder Diam., in. Meoh. Effcy. % Fuel per B.H.P. hr. lb. Fuel per I.H.P. hr. lb. 10 •70 •46 •320 10 •65 •51 •332 15 •73 •43 •315 15 •68 •48 •326 20 •75 •41 •308 20 •70 •46 •322 25 •76 •40 •305 25 •71 •45 •320 30 •76 •40 ■305 30 •71 ■45 •320 The fuel consumption per B.H.P. at loads other than full load is readily found by first calculating the probable mechanical efficiency, as described in the previous article, and then allow- ing for a fall in the consumption per I.H.P. proportional to that shown on Fig. 9 for a typical engine for the same M.I. P. Heat Balance Sheet. — An elaborate trial of a Diesel Engine includes the measurement of the quantity of cooling water used, the inlet and outlet temperatures of the water and the temperature of the exhaust gases. Apart from very slight losses, such as radiation, etc., these data usually enable the heat supplies by the fuel to be accounted for. A typical heat balance is given below : — Accounted for by indicated work .... 42% Rejected to cooling water 34% Rejected to exhaust 24% Total heat supplied . . .100 A striking feature of this balance is the large amount of heat appearing on the cooling water account, which at first sight would appear to indicate very poor utilisation of heat within the cylinder. It has been shown that so far from this being the case, an actual engine in good order indicates about 80% of the work attributable to an ideal engine. The explanation of this lies in the fact that a large propor- tion of the heat received by the cooling water is given out by the exhaust gases after combustion is complete, particularly on their passage through the cylinder cover in the case of a four stroke engine and through the ports in the case of a two cycle. Most of the friction work done by the piston and all the compressor work appear on the cooling water account. 28 DIESEL ENGINE DESIGN Efficiency of Combustion. — In all actual oil engines there is a considerable amount of " after burning," i.e. gradual burning during the expansion stroke. In a well-tuned Diesel Engine this effect is not sufficiently pronounced to cause smoke even at considerable overload, 120 lb. per sq. in. M.I.P. for example. Exaggerated " after burning " is to be avoided as, in addition to increasing the fuel consumption, it increases the mean temperature of the cycle and of the exhaust stroke particularly, and gives rise to accentuation of all the troubles which arise from the effects of high temperature. The most prominent of these troubles are enumerated below : — (1) Cracking of piston crown and cylinder covers. (2) Pitting of exhaust valves. (3) Increased difficulty of lubricating the cylinders, resulting in — (a) Increased liner wear. (b) Sticking of piston rings. (c) Liability to seizure of piston. (4) Increase of temperature of gudgeon pin bearing. The above formidable list is probably not exhaustive, but is sufficient to show the desirability of securing the best possible conditions, apart altogether from the question of economy in fuel and lubricating oil consumption. The attainment of good combustion, assuming a good volumetric efficiency of cylinder and good compression, depends more than anything upon small points in connection with the fuel valve, which are easily adjusted on the test bed, provided the design of the fuel valve is satisfactory. Entropy Diagrams. — Entropy is a convenient mathematical concept which it is difficult, and perhaps impossible, to define in non-mathematical terms. It is sometimes described as that function of the state of the working fluid which remains constant during an adiabatic precess. Entropy increases when heat is taken in by the working fluid and decreases when it is rejected. If heat is supplied to the working fluid at constant temperature, then the increase of entropy is equal to the amount of heat so supplied divided by the absolute tempera- ture. If the temperature is variable during the process of heat absorption, then the increase of entropy is determined by the integration of the equation : — THERMAL EFFICIENCY 29 , _dH where 0=Eiitropy. ^"""t" H= Heat taken in or given out. T =Absolute temperature. The zero of entropy may be located at any convenient level of temperature except absolute zero. It is convenient for our purposes to consider the entropy to be zero when P = 14-7 lb. per sq. in. abs. T=521°F. The value of a diagram connecting T and (j> during the working cycle of an internal combustion engine depends upon the following properties : — (1) Increasing and diminishing values of ^ denote heat supplied and heat rejected respectively. (2) For a complete cycle the diagram is a closed figure, the area of which is proportional to the quantity of heat which has been converted into work during the cycle. The area of the diagram is given by : — y"T d(p==/ — — — =Hi— Hj the difference between the heat supplied and that rejected. Construction of an Entropy Chart for Air (see Fig. 10). — The axis of T is drawn vertically, and includes temperatures from to 4000° F. abs. The axis of is horizontal, and is graduated from to 0-36. It will be clear that the axis of T is an adiabatic line of zero entropy and that any selected temperature on this line corresponds to a definite pressure, which can be calculated by means of equation (6). Actually it is more convenient to tabulate a series of pressures from to, say, 7001b. per sq. in. abs., and tabulate the corresponding temperatures. Points on the axis of T found in this manner are the starting points of constant pressure lines. Constant Pressure Line. P = 14'7 lb. per sq. in. — A constant pressure line is a curve connecting T and

are calculated and plotted on the chart, a fair curve passing through the points being the required constant pressure line. One such line having been constructed, lines correspond- ing to other pressures are readily drawn, since their ordin4tes Pounds per Sq In. Absolute. 600X0400300 200150 ■05 -10 -IS -20 26 -30 -35 Entropy per Lb of Cycle Air Fig. 10. are proportional to the temperatures indicated by their starting points. Constant Volume Lines. — These can be constructed in a similar manner, using the specific heat at constant volume instead of that of constant pressure. On Fig. 10 only one constant volume line is shown, ^z. that corresponding to P = 14-7 lb. per sq. in. T=673°r. abs. as this is required to complete the diagram by representing the rejection of heat at constant volume at the end of the stroke. This process is of course a scientific fiction, as the pressure in an actual cylinder is reduced to approximately atmospheric pressure by the discharge of exhaust gases and not by cooling THERMAL EFFICIENCY 31 the latter. The reason for using the value 0-25 for the specific heat must now be explained. Although the specific heat of pure dry air is 0'238, that of the gases present in the cylinder of a Diesel Engine is a variable quantity for the following reasons : — (1) The composition of the working fluid is altered by the addition of the fuel. (2) The specific heat of the exhaust gases increases slightly with increase of temperature. If these variations were rigidly taken into account the con- struction of the entropy diagram would be a very laborious business, and the work is greatly curtailed by adopting certain approximations which will be described. In the first place the specific heat is assumed to be constant and equal to the calculated specific heat of the exhaust gases at 60° F. The variation in the weight of the working fluid is dealt with by first treating the diagrams as though the weight were constant and then correcting the diagram (entropy) by increasing the entropy and decreasing the absolute temperature of points on the expansion line in the same proportion by which the fuel and blast air increase the weight of the charge. Use of the Entropy Chart. — A method of constructing the entropy diagram, corresponding to an indicator card, will now be described. The clearance volume of the engine must first be ascertained and the card accurately cahbrated. Points on the indicator diagram are then marked, corresponding to every 15 or 30° or other convenient division of revolution of the crank past the top dead centre. Each of the 12 or more points so marked is given a reference number and the absolute pressure in lb. per sq. in., and also the volume in any arbitrary units (hundredths of an inch on the diagram, for instance) corresponding to each point is read off and tabulated. The ■ apparent temperature corresponding to each point is then calculated by means of equation (1) on the assumption that the initial temperature just before compression is, say, 673° P. abs. Points on the entropy diagram corresponding to the selected points on the indicator diagram are now found by following the appropriate constant pressure lines until the calculated temperatures are reached, thus obtaining the apparent entropy diagram, which requires to be corrected in accordance with the preceding article. 32 DIESEL ENGINE DESIGN The entropy diagram shown in Fig. 10 has been constructed in the manner described, and the data are given below : — M.I.P. shown by indicator diagram, 82 lb. per sq. in. (see Pig. 11). ISO 135 ISO M.I P = 105 '90 75 82 lb s. per GO S(|. n. 4! 30 3 u ';3 500 300 fe- Deqrees anerT.D.C. 200 -| 100.1 Atmospheric Line Fig. 11. Initial temperature of suction air before compression assumed to be 673° F. abs. Absolute temperature (apparent) calculated from PV=kT where P=Pressure in lb. per sq. in. abs. (scaled off diagram) V= Volume measured in linear inches off diagram T=Temperature in degrees F. abs. Initial conditions are P=14-7, V=6-38, T=673. Whence k = -1395. Figures are given in the table below : — Ref. No. Degree after firing centre. V. in in. 0-48 P. in lb. per sq. in. Apparent Temperature. 1 490 1690 2 15 0-63 498 2260 3 30 0-95 421 2880 4 45 1-53 251 2760 5 60 2-17 163 2540 6 75 2-94 114 2400 7 90 3-Z5 4*6 84 2260 8 105 65 2130 13 180 6-38 14-7 673 19 265 2-94 44 930 20 300 2-17 61 950 21 315 1-53 99 1090 22 330 0-95 184 1250 23 345 0-63 331 1500 THERMAL EFFICIENCY 33 The apparent entropy diagram was plotted from the above values of P and T, and the corrected diagram, constructed in accordance with the preceding article, is shown in Fig. 10. Area of corrected entropy diagram 9-90 sq. in. Temperature scale . . . 1 in. =500°. Entropy scale .... lin. = -05°. Therefore work done per lb. of suction air is given by : — 9-90 X 500 x0-05 = 248 B.T.U. 1 lb. of suction air @ 673° F. abs. occupies 16-9 cub. ft. Since clearance volume = 8% of stroke volume, therefore corre- sponding stroke volume=;16-9-^l-08 = 15'65 cub. ft. There- fore work done by suction air, according to the indicator card, @ 82 lb. per sq. in. M.I.P. is given by : — 82xl44xl5-65 ^^33^^^ 778 This figure is about 4% less than that shown by the entropy chart, and suggests that perhaps the assumed temperature of the suction air before compression (viz. 673° F. abs.) is too high. Note. — ^The low pressures at the beginning of the compression stroke and the end of the expansion stroke have not been used in the construction of the entropy chart for the following reasons : — (1) Low pressures are very difficult to scale off the indicator card with any degree of accuracy. (2) The low pressures indicated on the card are invariably erroneous, unless the greatest care has been exercised in taking the card, and that with a suitable indicator in perfect order. Many of the reputable makes of indicator, though quite suitable for steam engine work, give very inaccurate results when applied to internal combustion engines. The following points should be noticed : — (1) The compression line deviates from the true adiabatic in a manner which indicates that heat has been lost to the cylinder walls. (2) The expansion does not become adiabatic until well after half stroke, showing that there is a certain amount of " after burning." 34 DIESEL ENGINE DESIGN Area of Entropy Diagram. — If the diagram has been care- fully done the area of the diagram in heat units should corre- spond fairly closely to the work shown on the indicator diagram. Deviations from equality may be due to : — (1) Variation in the specific heat at high temperatures. (2) Incorrect assumption of the initial temperature of the induced air. Owing to the approximations referred to above there is generally a discrepancy of a small percentage between the amounts of work shown by the indicator card and the entropy diagram. Also the total area under the upper boundary of the entropy diagram should correspond to the total heat supplied per lb. of suction air less the heat discarded to the water-jacket on the expansion stroke. Investigation of the sort described seems to indicate that not more than about 10% to 15% of the total heat supplied by the fuel is lost in this way at 100 lb. M.I.P. Specific Heat of Exhaust Gases. — The specific heat of a mixture of gases is the sum of the products of the specific heats of the constituents into the proportion by weight in which the constituents are present. The specific heats (at constant pressure) of the constituent gases present in the exhaust of an oil engine are given below : — Water vapour . . '. 0-480 (Varies considerably with Nitrogen Oxygen CO2 . . CO 0'247 temperature) 0-217 0-210 (Varies considerably with 0-240 temperature) The average composition of air and the specific heat derived from that of its constituents are given below : — N2 . . 75-7%x0-247 = -187 O2 . . |2-7 x0-217 = -049 H2O . . J^ X 0-480:=^ -007 Total . _100 -2^ This is about 3% higher than the accepted value for pure dry air. The approximate composition of exhaust gases assuming complete combustion is readily calculated as foUows : — Data M.I.P. - 100 lb. per sq. in. THERMAL EFFICIENCY 35 Fuel per I.H.P. hour -0-31 lb. r Carbon, 86% | Composition of fuel . I Hydrogen, 13% r by weight, (assumed). I Oxygen, 1% J One H.P. hour = 1,980,000 ft. lb. .". Volume swept by piston per I.H.P. hour=-^— — ^— - ^ -^ ^ ^ 100x144 = 137-5 cub. ft. Clearance volume, say, 8%= 11 -0 ,, ,, Total weight of suction air =148-5 (assuming perfect scavenge of clearance space). w - i-j. t 4.- ■ 148-5 X 14-7x144 onoiu Weight of suction air=^ -— - — — -r =8-78 lb. @ 212° F. Free volume of blast air @ 8% of stroke volume @ 60° F. = •08x137-5 = 11 cub. Weight of blast air=i^^^^^^^=0-83 lb. .: Total air = 8-78+0-83=9-611b.= 97% Weight of fuel . =0-31 „ = 3% Total mixture . 9-92 „ 100% Composition of mixture before combustion is given by : — (Na. . . -76x97=73-7% Air O2(22-7x97) + (-01x3)=22-0% (H^O . . -015x97= 1-5% *"^Hh2. . . -13X3= 0-4% 100-1% Combustion of carbon and hydrogen takes place in the pro- portions given by : — C+0j=C02 12+32=44 by weight And H2+0=H20 2 + 16 = 18 by weight 36 DIESEL ENGINE DESIGN The following table shows the composition after combustion and the specific heat (constant pressure) of the mixture : — Per cent. Specific heat. Product. Na 73-7 X -247 = -1820 CO2 2-5x44^12 . . 9-2 X -217 = -0191 O2 22-(32x2-5-i-12) . 12-1 x -210 = -0254 -(16x0-4^2) H2O l-5-(18x0-4-2) . _5J[ X -480 = - 0245 1004 Specific heat of exhaust gases = '2510 Repeating the above calculation for different values of the M.I.P., the following figures are obtained : — M.I.P. lbs. per sq. in. 30 45 75 105 130 155 Specific heat . . -245 -247 -248 -260 -252 -254 -256 Literature. — For thermodynamics of the internal combus- tion engine consult : — Wimperis, H. E., " The Internal Combustion Engine." Judge, A. W., " High Speed Internal Combustion Engines." Clerk, Sir Dugald, "The Thermodynamics of Gas, Petrol, and Oil Engines." For test results see : — Chalkley, A. P., " The Diesel Engine." Dalby, W. E., "Trials of a Small Diesel Engine."— Inst. N. A., April 2nd, 1914. WUkins, F. T., "Diesel Engine Trials."— Inst. M. E., October 20th, 1916. — " Description and Tests of 600 B.H.P. 2 stroke cycle direct reversible Diesel Marine Engine." — Engineering, Dec. 22nd, 1916, CHAPTER III EXHAUST, SUCTION AND SCAVENGE Renewal of the Charge. — In the theoretical study of heat engines the charge of working fluid is supposed to remain enclosed in a working cylinder and to undergo a cycle of physical changes due to the introduction of heat from, and discharge of heat to, external sources by conduction through the walls. In actual engines of the internal combustion type one constituent of the charge, viz. the oxygen, takes an active part in the chemical processes which constitute the source of energy. In such engines, therefore, the air charge must be renewed periodically. In all existing types of oil engine the charge is renewed as completely as possible for each cycle of thermal changes. The way in which this is done in four stroke and two stroke Diesel Engines has been described in general terms in Chapter I. In the present chapter it is proposed to discuss the questions of the discharge of exhaust gases and the introduction of a new air charge from the quantitative point of view, apart from the consideration of the mechanical details, such as valves, cams, etc., of which these processes involve the use. At the outset it will be necessary to state in a form con- venient for application, the laws governing the flow of gases through orifices. Flow of Gases through Orifices. — Fig. 12 is intended to represent a chamber containing gas at a pressure and tempera- ture maintained constant at the values P^ and Tj, and from which a gaseous stream is issuing through an orifice into the surrounding space, which is filled with gas at a constant pressure Pa- In the first instance, it will be supposed that P^ is very much greater than Pj. Then, according to the elementary theory of the flow of gases, it can be shewn that if the gas composing the stream expands adiabatically and all the work done is expended 37 38 DIESEL ENGINE DESIGN in increasing the kinetic energy of the stream, then the velocity of the latter will increase as expansion proceeds and attain a maximum value given by : — V = V2g J K„ (Ti-Ta) = 109-6V(T7=T;) (1) where : — V=maximum velocity of stream in ft. /sec. g =32-2 ft. /sec. 2 (acceleration due to gravity). J = Joules' equivalent (778 ft. lb. /B.T.U.). Kp=Specific heat at constant pressure (for air Kp = -238 B.T.U. /lb. deg. F. T2=Temperature attained by the stream after adiabatic expansion from P^ to Pj, and is given (for air) by : — (p \ 0.285 In equation (1) the velocity in the chamber is supposed to be negligibly small compared with V. For values of Pi up to P, ,T, Ps Fig. 12. 70 X 144 lb. per sq. ft., Pa being atmospheric, equation (1) has been f oimd to agree with experiment within two or three per cent, which is ample accm'acy for our purpose. It may, however, be mentioned in passing that experiments at high pressures, and particularly with steam, indicate that the elementary theory on whreh equation (1) is based stands in need of correction. It is to be observed that equation (1) is equally valid for any stage of the expansion of the stream, so that if any pressure value P^ be selected lying between Pj and Pj, then the velocity at this stage wiU be given by equation (1) with P^ substituted for Pg. If values of the velo6ity V^ be calculated for various values of P^ and the specific volumes v^ corre- EXHAUST, SUCTION AND SCAVENGE 39 spending to these values, then the ratios ==^ will be proportional to the areas of the stream at the several stages of expansion. On making such a calculation it wiU be found that the area of the stream at first contracts and afterwards converges to a final value corresponding to the maximum velocity. In other words, the stream has a neck or throat, as shown in Fig. 12. The pressure Pq at the throat is known as the critical pressure, and for air is equal to O-SS P^. Furthermore, it is evident that if Pa is equal to Pc (a con- tingency which was ruled out at the beginning of the discussion), then the stream will converge to its throat area and remain parallel instead of diverging. Supposing again that P2£00 u ^ o ^400 y / / rhroa / / ( lb 16 17 18 19 20 21 22 23 Scavenge Pressure Lbs/In? Abs- Fig. 14. 2a forced into the cylinder after the exhaust ports have been covered. During this latter stage the cyhnder pressure con- tinually increases until either the scavenge valves (or ports) close and compression begins, or until the cyhnder contents and the scavenge air supply are in equilibrium. The first stage of the process may be dealt with in a similar manner to that indicated in the preceding article, and Fig. 14 shows the calculated throat velocity of the scavenge air plotted against the absolute scavenge pressure, on the assumption that the back pressure is 15 lb. per sq. in. abs. and the temperature of the scavenge air supply is 130° F. (591° F. abs.), this being a good average figure found in practice. Before applying the above to a concrete example some observations will be made on port and valve areas. If a fluid flows for a time " t " through an orifice of constant 42 DIESEL ENGINE DESIGN area " A," with a velocity " V," then the volume " v " discharged is evidently given by : — v=V(Axt) If on the other hand A is variable, then v=V/A-dt The quantity yA-dt known as the time integral of the area is readily found by plotting A as ordinate against " t " as abscissae, as in Fig. 15, which exhibits the opening area of a valve from the time it lifts (t=o) to the time it seats. The time integral of the valve area for the whole interval is the area under the curve. For any other interval from, say, t=ti to t=t2the time integral of the area is the area bounded by curve, the axis of t and the two ordinates which define Fig. 15. tj and tg. This area is shown shaded on the diagram for particular values of t^ and tg. It is often convenient to take as the unit for " t " i^th sec. or xrnyrtli sec, or even 1 degree of revolution of the crank-shaft. Also it is very convenient to plot on a " t " base a curve whose ordinates representyA-dt from t=o. This has been done on Fig. 15. It will be noticed that between the instants t=ti sAd t=t2 the value ofyA-dt has increased by an amount X. The volume discharged between these instants would therefore be (X-V). Example of Scavenge Calculation .-j-Data for two stroke engine : — Bore of cylinder 10" Stroke 15" Revolution per minute . . . 300 EXHAUST, SUCTION AND SCAVENGE 43 Two scavenge valves 3 J" diam., maximum lift 1 ", opening 25° before bottom dead centre and closing 60° after. Lift curve harmonic, i.e. the lift plotted on a time base is a sine curve. Exhaust ports occupy 60% of the circumference of the cylinder bore and become uncovered by the piston 15% before the end of the stroke. The connecting rod is 5 cranks long. Fig. 16 has been drawn to show the relation between percentage of Fig. 16. stroke and the number of degrees between the crank position and the bottom dead centre. From this diagram it will be seen that the ports are uncovered 50° before B.D.C. and covered again 50° after B.D.C. Compression space 7% of the stroke volume. Free air capacity of scavenger 50% in excess of stroke volume of impulse cylinders. „ , , -785x102x15 Stroke volume =- 1728 Compression space =0-07 x 0-68 Stroke volume + compression space Volume of scavenge air (at atmospheric pressure and temperature) deUvered per cylinder per revolution=0'68 x 1'5 Maximum exhaust port area _ 0-6XTrxl0x0-15xl5 " 144 Maximum scavenge valve area _ 2X7rX3-5xl ~ 144 =0-68 ft.3 =0-05 „ =0-73 = 1-02 =0-295 = 0-153 ft. 2 44 DIESEL ENGINE DESIGN Intermediate values of the exhaust port and scavenge valve areas have been plotted on a crank angle base on Fig. 17. Number of seconds corresponding to 1 degree of revolution of the crank-shaft In Fig. 17 values of/A-dt for both exhaust and scavenge areas have been plotted in accordance with the preceding article. In particular, since the scavenge valve lift is harmonic, Degrees Before B.D.C. Degrees AFter B.D.C. Fig. 17. the mean area is one-half the maximum, and the final value of fA-dt is given by : — ^^x0-000556x85° = 10-3x3-61 ft.-'secs. And since the port area curve is nearly parabolic the maximum value ofy A-dt for the ports is given by : — 0-295 Xfx0-000556xl00° = 10-3x 10-94 ft.3 sees. The intermediate values are readily found by planimetering the areas under the port and^alve area curves. Scavenge Air Pressure. — ^A first approximation to the value of the pressure in the scavenger air pipe required by the condi- tions of the problem is readily obtained, on the assumption that the scavenge air is delivered against a constant pressure of 15 lb. per sq. in. abs. Volume discharged . . .=l-02ft.^ /A-dt =3-61x10-3 EXHAUST, SUCTION AND SCAVENGE 45 1-02 X 10"* Apparent velocity . . .=■ — — — =283 ft. /sec. Dividing by a discharge coefficient of 0-7 the " calculated " velocity is 283 ^0-7 =405 ft. /sec. and the corresponding scavenge air pressure from Fig. 14 is : — = 16-35 lb. perin.^abs. This figure cannot, however, be accepted as final, since the effect of discharging the excess of air through the exhaust ports has been ignored. Assume for trial a scavenge pressure of 17 lb. per in.^ abs. The calculated velocity from Fig. 14 is 498 ft. /sec. and the apparent velocity =498x0-7 = 349 ft. /sec. The value oty^A-dt which must be attained to fill the stroke volume and clearance space (0-73 ft.*) is therefore : — ^3=2.09X10- This corresponds to point A on Fig. 17, and shows that with the assumed scavenge pressure the cylinder would be filled with scavenge air at about 21° after the bottom dead centre. From this point until the point where the exhaust ports are covered three processes are occurring simultaneously, viz. : — (1) Scavenge air is escaping out of the exhaust ports. (2) Scavenge air is entering the cylinder and raising the pressure of its contents. (3) The piston is rising and reducing the volume of the cylinder contents, at the same time tending to raise the pressure. The pressure which exists in the cyUnder at the instant at which the piston covers the ports may conveniently be termed the " initial charge pressure," and will be denoted by Pj. Now if the value of Pj were equal to the scavenge pressure, which is its upper limit, the amount of air introduced into the cylinder in the interval under consideration would be about equal to : — f/A-dt (from 21° to 50° after B.D.C.) X 349 (for valves) 46 DIESEL ENGINE DESIGN From Fig. IT/A-dt for this interval is l•4:XlO"•^ so the vdlume introduced is f X 1-4x10-3 X 349=0-326 ft. 3 Adding to this the volume already introduced, viz. 0-73 tt.^, we obtain 1-056 ft. 3, which is a trifle in excess of the required quantity. The value assumed for the scavenge pressure, viz. 17 lb. /in.^ abs., is, therefore, probably not far out. The precise value of P, is a matter of considerable im- portance, as on it depends the value of the charge weight, and, consequently, the power capacity of the cylinder. Its pre- determination, however, appears to be a difficult matter, and in practice it is usually adjusted experimentally by advancing or retarding the scavenge cam, according as the value of Pj found by light spring indicator diagram is too low or too high. In the former case an adjustment is easily effected by putting a baffling diaphragm in- the exhaust pipe. The above calculations, however, are a sufficient check on the design to ensure the provision of suitable valve or port areas. The volume (@ 15 lb. /in. ^ abs.) of scavenge air lost through the exhaust ports is roughly equal to ^/A-dt (from 21° to 50° after B.D.C.) X 498 x 0-65 (for ports) From Fig. 17/A-dt for this interval is 2-15 x 10" ^ ft.« sec, so the volume required is ^X 2-15x10-3x498x0-65=0-23 ft. 3 So that the air charge is equivalent to (1-02— 0-23) =0-79 ft.* @ 15 lb. per sq. in. abs. Its actual volume at the point when the piston covers the ports is 0-05-)-(0-85x0-68)=0-63 ffc.» and its pressure is therefore about 15x0-» ,^ ^,, ,. „ -^;g3-=18.8lb./in.^ This figure being 1-8 lb. /in.^ above the assumed scavenge pressure of 17 lb. /in.^ indicates that the latter figure is in- sufficient for the supply of the requisite quantity of air under the conditions specified. The true value probably lies some- where between the two. Exhaust of Two Stroke Engines. — ^As already mentioned, EXHAUST, SUCTION AND SCAVENGE 47 it is essential that the scavenge receiver should not be put into communication with the cylinder until the contents of the latter have fallen to a pressure almost equal to that of the scavenge air, by the release of the products of combustion, through the exhaust slots. This process usually takes a period of time equivalent to 20 to 30 degrees of revolution of the crank-shaft. The calculation of this period is much facilitated by the fact that in the interval considered the port area is increasing very approximately in proportion to the time (see Fig. 17). It can easily be shown that if during an interval from t=o to t=ti an orifice area increases uniformly with respect to time from O to A^ and the velocity of efHux also varies uni- formly with respect to time from V, to Vj, then the discharge is given by : — Q=^Mv,+iV„) — (1) If A represents ft. 2, t sees., and V ft. /sees, of an incom- pressible fluid, then Q represents cub. ft. If, as in the case we are about to consider, V represents lb. per sec. per unit of area, then Q represents lb. We proceed to apply equation (1) to the exhaust period of the two stroke engine specified in the previous article. Exhaust Calculation. — Data : Cyhnder pressure at the point at which the ports become uncovered, 60 lb. /in. * abs. Temper- ature at the same point is equal to initial charge temperature x 60_ 670x60 — say r-; ,, ,, pressure. 18 =2240° F. abs. Pressure in exhaust pipe, 15 lb. /in.* abs. Problem : To find the position of the crank when the cylinder pressure has fallen to 18 lb. /in.* abs./ Assume that the charge has the thermal properties of pure air and that the expansion is adiabatic. At the instant when the ports first become open the volume of the charge is : — 0-05+0-85x0-68=0-63ft.3 and the weight of the charge is : — (60 x 144 x 0-63) -^(63-2 x 2240) =0-0456 lb. When the cylinder pressure has fallen to 18 lb. /in.* abs. the 48 DIESEL ENGINE DESIGN volume is not certainly known, but wiU not differ much from 0-05+0-92 x0-68=0-675 ft.^ and its temperature will be :— 2240(4^1 =1590°F. abs. /18^Y28B_ \60-0/ ■" and the charge weight is reduced to (18 X 144 xO-675) ^(53-2 x 1590)=0-0207 lb. The weight discharged in the interval is therefore : — 0-0456 -0-0207 =0-0249 lb. The next step is to calculate the rate of discharge per unit of port area. At the higher pressure of 60 Ib./in.^ the throat pressure is 0-53 x 60=31-8 lb. /in. 2 abs., and the throat temperature is : — (Ql .C\-288 ^j =1870° F. abs. The calculated throat velocity is therefore : — 109-5 V2240-1870=2110 ft. /sec. and the apparent velocity =2100 x 0-65 = 1365 ft. /sec. Now the specific volume at the throat is : — 53-2x1870 „,„,,, 1, • ^j-:g-^=21-8ft.3perlb. The rate of discharge is therefore : — — — =62*7 lb. per sec. per ft.* of port area. 21-8 At the lower pressure of 18 lb. /in.'' the throat pressure is 15 lb. /in. 2 abs., and the throat temperature is : — 2240(^y285 = 1510° F. abs. The calculated throat velocity is therefore : — 109-5^1590-1510 = 978 ft. /sec. and the apparent velocity =§78 xO-65=635 ft. /sec. Now the specific volume at the throat is : — 53-2X1510 or, OX^ lU — - — — — =37-2 ft. 3 per lb. 15x144 ^ and the rate of discharge is therefore : — 635 — — =17-1 lb. /sec. per ft.* of port area. EXHAUST, SUCTION AND SCAVENGE 49 We now make the assumption that the rate of discharge per unit area changes its value from the higher to the lower value uniformly with respect to time, so that equation (1) can be applied. We then have : — 0-0249= 3 from which Aj t^ = 10" ^ X 1 -54 and7'A-dt = i Aj ti (approx.) = 10-3x0-77 ft.^ sec. Reference to Fig. 17 shows that this value occurs about 18 degrees after the ports begin to open. Alternative Method of Calculation. — Let :— t=time in seconds, counting from the instant when the ports begin to be uncovered by the piston. P=pressure of cylinder contents in Ib./ft.^ during the period considered, so that P varies from 60 x 144 to 18x144. T == temperature of cylinder contents deg. P. abs. (variable). w= weight in lb. of cylinder contents (variable). A=port area in ft.^ (variable). v= volume in ft.^ of cylinder contents (actually varies during the period considered, but treated as constant) =0-65 ft. 3 Vg= specific volume in ft.* /lb. at the throat of the stream issuing from the ports. V=throat velocity in ft. /sec. dP f (P) =--T — hA=the rate of pressure drop per unit of port area, dtj then^=f(P)xA dt and/A- dt=/j^ f (P) has now to be calculated. KQ.O P.v = 53-2wT .-. P- w.T. V , dP 53-2 /^dw , dT\ and -r--= ( T.-rr+W.- * dt V \ "dt 'dt/ 53-2 /^ dw , dT dP\ T-dt+^-dp-d^) 50 DIESEL ENGINE DESIGN so that dP 53;2 / dw\ 53-2 dw V V 'dt / V dt /, 53-2 dT\ / p dT\ (1-— •^•dp) (l-fdP/' AT dw . V Now -r- =A.— dt V. /|J\U. ZIJ5 and since T=To f.^) , where T^ and P^, are the initial ^^"^ values of T and P, dT 0-285 T„P«'- 285-1) dP P^o-^ss P dT and =--^=0-285 (i.e. constant), dP 53-2 V so that f(P)=^-^ A = V" V ^ ' dt V v„ 0-715 Values of f (P) are easily calculated for evenly spaced values f dP of P and the integration of \-TTf^, can be effected by Simpson's rule. J *(^) The table opposite shows the application of this method to the example of the preceding article. Po=601b./in.2abs. To=2240° F. abs. V=0-65ft.3 For convenience P has been expressed in lb. /in.^, instead of lb. /ft. 2, so that a factor of 144 is required. Since the interval between successive values of P is 10-5 lb. /in.", we have : — r dP /A-dt=^ =10-5x144x96-18x10-8 f(P) 3 =0-485x10-3 Dividing by the discharge coefficient 0-65 the required value of^'A-dt is :— • 0-485x10-3 „ ,. -^ „ — — =0-74 ft.^ sec. H- 1000 0-65 which agrees with the value obtained previously within about 4%- The value of the discharge coefficient (0-65) has been found by comparing the results of calculations similar to the above, with the information afforded by light spring indicator cards. EXHAUST, SUCTION AND SCAVENGE 51 00 GO 00 <=P 00 00 ^_^ i 6 6 tb O 1^ r-l 1— ( 1— 1 l-H i-t I-H pH X X X X X X o 1— 1 o o o I-H O I-H a H > X X X X X CO > 00 00 o to 14* CO lO l~i r-H ■"* o 6 CO t^ ■* r~i CO 00 >\> CD 1— t to pH to a 00 to lO IN t» 00 1—1 00 IN IN > i-H (N 05 IN CO CO ■" ^ „• S.-S o o o lO o o 00 05 to Thrc veloc V ft. /SI l—t o IN o 1— ( Temp. drop. deg. F. O o w o i la t^ lO IN o CO CO CO CO ■| df-i ■ o o lO o o ■* Thro Tern deg. abs 00 r-l l> 1— 1 l—t •o I-H •g .". O m P OQ t^ per^n. ' of bore', ^ X No. of cylind's.) 9-5 to 14 5-2 4-5 4-0 4-8 3-5 For preUminary estimating it is convenient to have an approximate idea of the weights of the piston, connecting rod, etc., of a proposed engine, and the following figures are a rough guide to average practice. A designer wiU find it convenient to make notes of similar figures for actual engines with which he has had experience. THE PRINCIPLE OF SIMILITUDE 59 Table IV Name of Part. Weight per in. ' of bore' lb. Trunk piston (land engines) .... 0-18 Connecting rod „ „ .... 0-17 Crank-pin „ „ .... 0-036 One web „ ,, .... 0-036 Reciprocating weight (land engines) . 0-24 Revolving weight „ ... 0-18 Piston and piston rod (marine engine) 0-13 Crosshead „ ,, 0-13 Connecting rod „ „ 0-24 Crank-pin „ ,, 0-046 One web „ ,, . . 0-046 Reciprocating weight ,, „ 0-356 Revolving weight ,, „ 0-236 The above figures are necessarily approximate only, as they depend not only on the materials used and the views as to design stresses, etc., held by individual designers, but also on the bore to stroke ratio and the rules of insurance societies. Determination of Bore and Stroke. — Nowadays Diesel Engines are generally built up of standardised units in com- binations of 2, 3, 4, 6 and 8 cylinders, and by this means a large range of sizes may be covered with a minimum of stock patterns, jigs, etc. The choice of sizes of cylinders and the numbers of different sizes stocked will of course depend on the capacity of the factory and the estimated demand. Marine engines are usually provided with 6 or 8 cylinders in the case of four stroke engines and 4 or 6 cyhnders in the case of two stroke engines, to facilitate starting and manoeuvring and also to keep the size of cylinder and the aggregate weight within reasonable limits. Having decided on the brake-power to be developed by a certain standard cylinder the indicated power" must be inferred, either from previous experience of engines similar to the one proposed, by direct estimate of the various losses, or with reference to published data of comparable engines. Typical figures for mechanical efficiency have been given on page 27. 60 DIESEL ENGINE DESIGN A very close approximation of the mechanical losses is obtain- able as follows : — 1. Air compressor and scavenger losses may be reckoned to be equal to the indicated power of these accessories divided by a mechanical efficiency of 0-8. 2. Losses due to friction of piston rings about 5% of the indicated power of the engine. 3. Lubricated friction loss is given very approximately by the formula : Horse-power lost in lubricated friction _ 0-3 [D.LiVi^-s+n(A-KBL,)V,i-^] 550 Where D =Diameter of crank-shaft in in. Li=Aggregate length of journals and crank -pins in in. Vi=Peripheral speed of journals in ft. per sec. n =Number of cylinders. A = Area of slipper guide in sq. in. B =Bore of cylinder in in. L2=Length of piston in in. Vj^Piston speed in ft. per sec. Assuming that the desired indicated power per cylinder is now known, the former is related to the cylinder bore by the formula : — TXTT> V A 0-785.B2P.V B^P.V ... I.H.P. per cylmder=^^^33^^^ = ^^^ (1) for two cycle engines. And half this amount for four cycle engines, both assumed single acting. I Where P=Mean indicated pressure in lb. per sq. in. V= Piston speed in ft. per minute. The piston speed and mcOTi indicated pressure both vary considerably in different cases, and existing practice in this respect will be discussed later. Table V gives values of the I.H.P. per sq. in. of bore^ for various values of the piston speed and M.I.P. commonly adapted. The I.H.P. of a proposed cylinder is found by multiplying the square of the bore by the appropriate coefficient from Table V in the case of a two stroke and J)y half that figure in the case of a four stroke cylinder. THE PRINCIPLE OF SIMILITUDE 61 Hi n H > I f^ CO 12; o H SQ I— I 8 O i-H »0 i-l » JI5 05 lO rH 03 O O i-H "-I N CO 6 -^ p^ -^ -^ ^ '^ 05 CO n 05 OS O O i-H r-( (N 6 O rH rH rH ,^ ,A t- (N CO lo 1— 1 CO (M t- ro CO 00 05 OJ O O ■— 1 i— 1 © O O 1— 1 r-( 1— 1 F-l o 00 O O !M T)( rH CO rH CO i—i CO 1—1 00 OO OS OS O O r-l O O O O rH i-H ,i 8 00 fO CO rt m o lo o i« I> 00 OO OS 03 o o O O O © © I-H I-H o lO © lO © ■«*( © CO ^ CO © »0 05 ■* 00 I> t- 00 00 CO 03 OS o © o © © o © o o l> 00 © -^ Tjl CO t- CO © lO OS CO t- rH CO 1> l> C-- OO CO 05 © O O O © © C5 © 00 t- CO lO ■* !M (N m 05 CO i> t-H lo CD CO CO I> t^ 00 00 © © © © © © 6 o # 00 I-H lO 00 lO CO CO CD l> l> 1> © © © 6 © © © •I M © lO © lO © lO © 00 OO 05 05 © © rt I-H I-H r-H g fc< 6(1 a « o M o OJ o & (N "fll >) <2. ,i2 o S s 1> -a a e i>>^ o o Tl >■ .c ■V o <^ = 51001b. g Inertia effect of reciprocating parts at top dead centre : — 236x615 ,, K.nAiu X l+=5400 lb. g Inertia effect of reciprocating parts at bottom dead centre : — ?5^^^X(l-i) = 3600lb. g Combined centrifugal and inertia effect at top dead centre =5100-1-5400 = 10,500 lb. upwards. Combined centrifugal and inertia effect at bottom dead centre = 51004-3600 = 8700 lb. downwards. 80 DIESEL ENGINE DESIGN Maximum load due to cylinder pressure =0-785 X 102 X 500 = 39,000 lb. Resultajit downward effect of pressure, centrifugal force and inertia force at firing dead centre = 39,000-10,500 = 28,500 lb. Method of Calculating the Reactions at the Bearings. — Let A 1 (Fig. 39) represent the centre of the crank-shaft, A C E G I being the centres of the journals and B D F H those of the crank-pins. Rg R4 Re Rs a-re the applied forces due to cylinders 1, 2, 3, 4 respectively. RjRgRjR^Rg are the reactions at the bearings unknown in magnitude and direction. Under the influence of these forces the centre line of the shaft assumes some deflected shape, Cyin. CyirZ. Cyl':3. CylTA. /?? f{4 /?5 Rg I Fig. 39. Rg and the deflection at any point above or below the straight line joining A 1 is equal to the sum of the deflections at the same point which would be produced by each of the forces R2R3R4R5R6R7R8 acting alone, supposing the shaft supported freely at A and 1 . At C E and G the sum of these deflections must be zero if the bearings are level (ignoring the effect of running clearance). Let Cg Cg and g^ be the deflections at C E and G due to unit load applied at B (the position of Rj), assuming the shaft support||d freely at A and 1. Cgesandga' 0464 and g4 Cgegandgs etc., etc. Caegandgg are the corresponding deflections at C E and G due to unit load applied at C D E, etc. The values of Cj Ca and gg, etc., are readily found by the usual formulae for the deflection of beams. CRANK-SHAFTS 81 Now since the total deflection at C E and G is zero, the following equations hold good : — R2 C2+R3C3 + R4C4 + R5C5+ReC6 + R,C,+R8C8 = (1) R2e2+R3e3+R4e4+R6e5+R6e6+R,e,+R8e8=0 (2) R2 g2+R3 g3+R4 g4 + R6 gs+Re g6+R7 gT+Rg g8=0 (3) In these three equations the only unknown quantities are R3 Rj R7, which can therefore be determined. The remaining unknown reactions R^ and Rg are found by equating moments about 1 and A respectively. In solving the above equations downward forces and deflections will be considered positive and upward forces and deflections negative. Determination of C2e2g2, etc. — In determining these constants it will be assumed that the shaft deflects under load as though it were a cylindrical beam of the same diameter as the crank-pins and journals. Considering that the webs of a Diesel Engine crank are short, the presumption is probably not inaccurate, but it would be interesting to see this point investigated, as it could readily be, by means of models. So long as bearings at constant level are assumed, the actual value of the deflection is of no importance, as the method of calculation depends only on the relative deflection at the different points considered. The problem therefore resolves itself into finding the deflected form of a beam freely supported at each end under the influence of a concentrated load placed anywhere between the supports. This may be done by treat- ing each end of the beam as a cantilever. The deflection of the end of a cantilever carrying a load at the end is given by : — W 1' Deflection at the end of cantilever in in. = „' — 3 EI Where W=Loadinlb. E = 30,000,000 lb. per sq. in. (for steel). I=Moment of inertia (transverse) of the section of beam in. For the shaft under consideration : — I=Zx5-25* = 37-3in.* 64 Fig. 40 shews the values of the deflection at various fractional points in the length of the cantilever, the deflection at the end being unity. These results are applied to the case of a beam as follows : — Let AB be a beam .(Fig. 41) supported at A and B and 82 DIESEL ENGINE DESIGN carrying a load W^ at any point C. The reaction at A is equal OB to Wj j^. Let this be denoted by Wj. At A erect a perpendi- cular AD equal to some convenient scale to the deflection of the cantilever AC, due to the load W 2 at its end. Draw the deflected shape of this cantilever (DC) by means of the proportions given re 'A fa k % % ^e -^ -^ r» •s^ «5 c\ 1 2 1 10 CM Fig. 40. in Fig. 40. Proceed similarly with cantilever CB, obtaining the deflected shape C F. Join D and F, then the deflection of the beam at any point X is the vertical intercept " x " shewn in the figure. Application to the Case in Hand. — The constants c^e^gi being the deflections at various points on a beam due to unit load applied at other various points, are independent of the system of loads and wiU therefore be dealt with before special A 1 X D < c X ""--.^ F W3 Fig. 41. cases of loading are consiJfered. In order to obtain manage- able figures, the deflection will be reckoned in thousandths of an inch and the unit load will be taken as 10,000 lb. Deflection of cantilevered portion of shaft in thousandths of an inch is given by:- W.(A)« _ W.(3^)3 3x30x37-3 Where W=Loadinlb. Z=Length in in. 3360 CRANK-SHAFTS 83 The diagram Fig. 42 shews the process of determining ^2 ©2 §2! etc., in particular : — Unit load at B = 10,000 lb., AB = 10", BI = 70" -D ,. ^ . 10,000x70 „„^^,, Reaction at A = — '—— =8750 lb. Hi) T 10,000x10 ,„.„,, „ 1= g„ =1250 lb. Deflection of cantilever AB= — -— - — =2-6^-^°^ odOU ^-^ ^ f \ ^ ^ y /// / N ■v ITj ^ -^Z ^ /] / ^'^ ^ 4 i / ^ V / ABCOeFGHI Fig. 42. Deflected shapes of cantilevers drawn by plotting ordinates from the proportion given in Fig. 40. By scaling the diagram : — C2 = 30-5 e2=35 g2=20-5 Since the deflection at G, due to a load at H, is the same as that at C, due to the same load at B, and so on, therefore : — g8 = 30-5 68 = 36 C8=20-5 By the same methods (see Fig. 42) : — C3=51 e3=62 g3=39 g, = 51 e, = 62 c, = 39 c^ = 65 64=88 g4=55 g6 = 65 66 = 88 C6 = 55 C6 = 65 65 = 95 g6 = 65 84 DIESEL ENGINE DESIGN These values could be computed with rather less trouble and with greater accuracy by the formulae for the deflection at any point in a beam due to a load at any other point (see Morley's " Strength of Materials "). In more complicated cases, with a larger number of unknown reactions, the constants c, e, g require to be ascertained very accurately, otherwise large errors arise. Z-85 0-87 087 -105 Fio. 43. PJ Rgi The conditions of loading will now be considered. Case I. Crank 1 on Firing Dead Centre. — ^The magnitudes and directions of the applied forces are shewn in Fig. 43. Hence : — R2C2 = 2-85x30-5 = 87-0 R Ra 62 = 2-85x35 =99-5 R R2g2 = 2-85x20-5 = 58-5 R4g4=0-87 x 55=48-0 04=0-87x65=56-5 4 64=0-87x88 = 76-5 Rj 06=0-87x55 = 48-0 Rg C8=-l-05 x20-5= -21-5 Rg 66=0-87x88 = 76-5 Rg e8=-l-05 x35 =-37-0 Rgg5=0-87x65 = 56-5 R8g8=-l-05x30-5=-32-0 From which Ra C2+R4 C4+R6 C6+R8 08=170-0 R2 e2+R4 e4+R6 ee+Rs 68=215-5 R2g2+R4g4 + R6g6+R8g8 = 131-0 Substituting these values in equations (1), (2), and (3), we obtain : — * 51 R3 + 65 R6 + 39 R7= -170-0 (4) 62 R3+95 R5 + 62 R,= -215-5- 39 R3+65R5 + 5IR, = -131-0- -(5) -(6) From which R3= -2-295 R5= -1-402 R,= -0-965 Equating moments about A : — 8 Rs + 7 Re-F6 R7 + 5 R6+4 R6 + 3 R4-K2 R3-hR2=0 CRANK-SHAFTS 85 .-. 8 R9 = 7x 1-05-6 x0-965-5x0-87+4x 1-402-3 xO-87 + 2x2-295-2-850 Whence Rg=0-118. Equating moments about I : — 8Ri + 7R2 + 6R3+5R4+4R5 + 3R6 + 2R,+R8=0 .-. 8 Ri= -7 X 2-85 + 6 X 2-29-5x0-87 + 4x1-402-3x0-87 -2x0-965 + 1-05 Whence Ri= -1-079. Knowing all the forces, the bending moment at A B C, etc., can now be tabulated thus : — Point. Moments. B.M. in in. lb. A B C D E F G H I 10,510x10 10,510x20-28,500x10 .... 10,510x30-28,500x20+22,950x10 . 10,510 X 40-28,500 X 30+22,950 X 20-8700 XlO -2440x30 + 10,500x20-9,650x10 . -2440x20 + 10,500x10. -2440X10 105,100 -74,800 -25,200 -62,600 30,300 56,200 -24,400 Since the bending modulus of the shaft is 32 x5-253 = 14-2in.3 Therefore, Maximum stress due to bending (occurring in No. 1 crank- pin at top firing centre) is equal to : — 1^0^7,400 lb. per in.^ Bending stress, according to fixed journal method : — W.L 28,500x20 __ ,, rz^ 8x14-2 =50201b.sq.m. According to free journal method : — „, 28,500x20 ,„„^„,, Stress = A^iAo =10.040 lb. sq. m. 86 DIESEL ENGINE DESIGN Case II. No. 2 Crank on Firing Dead Centre.— The magnitudes and directions of the applied forces are shewn in Fig. 44. R2=0-870 R4=2-850 R6=-l-05 R8=0-870 R2C2=0-87x30-5 = 26-5 R^ C4=2-85x65-0 = 185-0 Ra 62=0-87 x36-0 = 30-5 R^ 04=2-86 x88-0 = 251-0 R2g2=0-87x20-5 = 18-0 R4g4=2-85x55-0 = 156-5 Rg 06= -1-05 X 55-0= -58-0 Rg 66= -1-05x88-0= -92-5 R5gg= -1.05x65-0 = -68-0 Rg 08=0-87x20-5 = 16-0 Rg 68=0-87x35-0 = 30-5 R3gg=0-87x30-5 = 26-5 From which Rg C2+R4 C4+R6 C6+R8 C8 = 171-5 Rg Og+Ri e4+R6 eg+Rg 68 = 219-5 0-87 Z-85 -l-OB 0S7 R, A?3 ffg /?7 Rg Fig. 44. The three equations for determining R3R5R, are there- fore : — 51 R3+65 R6i39 R7= -171-5 I -219-5 62Rj-|-95Rs-f62R, 39R3-f65RB-|-5lR7=-133 from which R, = -2-005 Rs=-l-788 R, = 1-204 Taking moments about A and I, Ri=-0-161 Rj = -0-790 CRANK-SHAFTS 87 from which the following figures for the bending moments are obtained : — Point. Moments. B.M. in in. lb. A B 1610x10 16,100 C 1610x20- -8700x10. , -54,800 D 1610x30- -8700x20-1-20,050x10 . 74,800 E 1610x40- 10 . -8700x30-1-20,050x20- -28,500 X -80,600 F 7900X30- -8700x20-12,040x10 , -57,400 G 7900x20- -8700x10. 71,000 H 7900 X 10 . 79,000 I • ■ Maximum bending stress ,' „ =5670 lb. sq. in. ° 14-2 Considerably less than in Case I. Case III. Loading as in Case I, but bearings C and G supposed worn down 20 thousandths and bearing E 25 thousandths of an inch below the level of the line joining A I. The difference in height of the bearings is very simply allowed for by putting 20, 25, and 20 respectively on the right- hand side of equations (1), (2), and (3) instead of zero. The equations for determining RgRjand Rjthen become :— 51 R3-t-65 R5-h39 R7 = 20- 62 R34-95 Rs-l-62 R7 = 25- 39 R3-I-65 R5-f51 R,=22- -170-0=-150-0 -215-5=-190-5 -131 =-109 The following values are obtained for the reactions at the bearings : — Ri=-1-412 R3=-l-610 R5=-2-135 R7=-l-820 R, = -0-202 88 DIESEL ENGINE DESIGN The bending moments are as follows : — Point. Moments. B.M. in in. lb. A B c D E F G H I 14,120x10 14,120x20-28,500x10 14,120x30-28,500x20 + 16,100x10 . 14,120x40-28,500x30 + 16,100x20-8700 XlO 2020x30 + 10,500x20-18,200x10 2020x20 + 10,500x10 2020x10 141,200 -2,600 146,000 -55,200 88,600 145,400 20,200 Maximum bending stress 146,000 14-2 = 10,250 lb. sq. in. Comparing the above figures with those obtained in Case I, it will be seen that the maximum stresses have been increased to the extent of about 35% by the difference of level of the bearings. In view of the uncertainty which exists with regard to the deflection of cranked shafts the above figures are not strictly reliable, but the writer is of opinion that they under- rather than over-estimate the stresses. Conclusions. — 1. The value of the bending moments at a crank-pin on the top firing centre is greater for a crank-pin situate at one end of the shaft than that at one nearer the centre of the shaft. 2. The bending moments at certain crank-pins and journals may be as great or greater than the bending moment at the crank-pin which is receiving the greatest applied load. 3. No general rule for the be)|ding moment at a Diesel Engine crank-pin or journal can represent the true state of affairs, but every different arrangement of cranks and number of cylinders requires to be investigated individually. 4. Difference of level of the bearings, due to wear or other- wise, gives rise to greatly increased bending moments, which can be calculated approximately in the manner described. The methods of calculation which have been illustrated in CRANK-SHAFTS 89 this chapter can be applied to cases involving any number of cranks, and the effects of fly-wheels, the rotors of electric generators, outboard bearings, etc., can be included. When the number of cylinders is three or less the weight of the fly- wheel is considerable, and cannot therefore be ignored. In these cases allowance should be also made for the practice of packing the outward bearing above the level of the engine main bearings. For engines of four cylinders and over, the weight of the fly-wheel and the presence of outboard bearings can probably be neglected with safety. An extended treat- ment of this subject will have to be reserved for a future occasion. The labour involved in solving simultaneous linear equations increases as the square of the number of unknowns. For a description of a machine devised to do this work mechanically, see the " Treatise on Natural Philosophy," vol. i., Kelvin and Tait. Graphical Determination of the Twisting Moments. — ^In the processes described below the following approximations have been made : — 1 . The negative twisting moments due to the air compressor at the forward end of the engine have been neglected. These moments are small in comparison with the moments due to the working cylinders, and being opposite in direction to the maximum moments tend to reduce the latter by a small amount. 2. Moments due to the dead weight of the revolving and reciprocating parts have also been neglected. In a very large engine it would be advisable to take these into consideration, as other things being equal the dead weight of the running gear per square inch of piston area increases as the scale of the engine. 3. The twisting moments due to mechanical friction have been neglected, as (so far as the present writer is aware) the distribution and variation of the friction forces are not known with any exactitude, and in any case one is a little on the safe side in neglecting them. These friction moments of course accumulate as one passes from the forward to the aft end of the engine, where they amount in aggregate to about 15% of the mean indicated twisting moment, so their effect on the forward end of the shaft is quite negligible. 90 DIESEL ENGINE DESIGN 4. The moment of inertia of the fly-wheel has been assumed to be large in comparison with the fly-wheel effect of the revolving and reciprocating parts of the running gear. In cases where the fly-wheel is very small, or omitted altogether, as in some two stroke marine engines, the irregularities of turning effort are mainly absorbed by the angular acceleration of the crank masses. Allowance is readily made for this effect in the following manner. The combined twisting moment curve for all the cylinders is first found without allowance for fly-wheel effect, and a new zero line is taken at the height corresponding to the mean twisting moment. Or- dinates measured to this new zero line represent fluctuations of the twisting moment from its mean value. These ordinates are now divided into segments proportional to the fly-wheel effects of the fly- wheel and crank masses. For ex- ample, if there are four cylinders and the moment of inertia of the fly-wheel is three times that of one set of crank masses, then the or- dinates will be divided into seven parts, one part being appUed in opposite sense to corresponding points on the twisting moment curve of each cylinder and the re- maining three parts of each ordinate form the ordinates of a curve of the twisting moments absorbed by the angular inertia of the fly-wheel. In the example worked out below itgsdll be assumed that the fly-wheel is large compared with the crank masses, so that the process described briefly above is not necessary. Fig. 45 is a skeleton diagram of the connecting rod positions for every 20 degrees of revolution of the crank-shaft for the determina- tion of piston displacements. CRANK-SHAFTS 91 Fig. 46 is a typical full load indicator card calibrated for pressures vertically and percenta;ges of stroke horizontally. Points corresponding to each 20 degrees of revolution are marked on the diagram by scaling the piston displacements ofE Fig. 45. On Fig. 47 cylinder pressures are plotted on a crank angle base from to 720 degrees (four stroke engine). The pressure izo° m'miscf during the suction and exhaust strokes is assumed atmospheric. The inertia effect of the reciprocating parts per square inch of piston area is plotted from the following figures : — 236 X 615 Inertia effect at top dead centre (1 +J) =5400 lb. bottom „ 236x615 236x615 g (1-1) = 3600 90 g X^= 9001b. 5 ^^.^236X615^ l = 3200lb. g ^2 in.) Corresponding figures per sq. in. of piston area {78-5 sq. are 69, 46, 11-5, and 41 lb. per sq. in. respectively. The centrifugal forces of the revolving masses being radial produce no twisting effect on the shaft. The dotted line (Fig. 47) is the resultant of the pressure and inertia curves. The twisting moments are now computed as in the following table : — 92 DIESEL ENGINE DESIGN toS 14,100 20,400 14,100 -11,200 -17,900 -15,100 -9,750 ©©©©©©©© ■i^a ©©©OO©©© ® ■*" o" in" so t-" a> rf r-" Twis mom in. Iffl so rH 1—1 r-H i-H i-H | mini' 43 d iiltan ce in /in." of n are 05(— ,5ClO 00»t-05Ci5 00r-lO ©©CO(N®i— 100©!— 1© 1 1 1 1 1 1 1 1 SClfflI>l>t-10Tt<(N 1 M 1 1 1 1 1 CO o,g ID O & © H ® oooooooooo ©©©©©©©OO© i»T)H{MO00ffl-*(NOa0 ©(Mt-COlOTHMi— 1 i-H CTi<»ft«l i-H'^OOCO©01050000t^ 5O«O-^(M©o M >ffl Qoap © © CD (N 1© rH C30 ©^ © ©so©»ffleo»«oo© ©O®CSlC0rH00©f-l© MlOI>I>t^lO'<*I>I>»0-<*M Degrees from top ead centre. ©©©©©OO©©© ©©OOOOOOOO ^■^^(^©(NTtlJOOO 5000© f R S 2 S •Baj\/ uoqsjj jo-u| •bg'jad spunoj' 94 DIESEL ENGINE DESIGN piston area in lb. per sq. in. and by the piston area in sq. in. (in this case 78-5 sq. in.). Forces acting towards the crank are considered positive, whether they are expansion forces or otherwise, and those acting away from the crank negative. Assuming rotation clockwise, leverages to the right hand of the centre line are positive and those to the left negative. The signs of the moments then look after themselves according to the signs of their factors. It is not unusual to see the leverage in a case of this sort treated as though it were always positive. The dis- advantage of this proceeding is that in order to get the signs of the moments correct those of the resultant forces have to be reversed at every dead centre, which besides being incorrect from a mathematical standpoint is inconvenient for the draughtsman and confusing to others. The tabulated values of the twisting moment are plotted in Fig. 48 (full line curve). Identical curves for cylinders 3, 4, and 2 could be plotted in their respective places at 180 degrees apart, in the order named, but are omitted, for the sake of clearness. Dotted curve numbered 2 is the resultant of the curves belonging to cylinders 1 and 2. Dotted curve 3 is the resultant of the curves belong- ing to cylinders 1, 2, and 3, and so on. The simplest way of obtaining these resultants is to trace the primary curve on a piece of transparent paper and move it sideways into its required position for the next cylinder, and then for every required ordinate move the paper vertically (guided by the vertical degree lines) until the zero line coincides with the top of the ordinate of the curve to which it is required to add the effect of another cylinder. In this position prick through the top of the ordinate of the curve on the tracing-paper to the diagram underneath. These resultant curves enable the twisting moment at any crank-pin or journal at any angular position to be read off the diagram. Combined Effect of Ben^ng and Twisting. — It will be seen that the peaks of the twisting moment curves occur about 30 degrees after the dead centres, and that the results previously obtained for the bending moments with the cranks on dead centre apply very closely to this position also, so that the tabulated values of the bending moments at the various journals and crank-pins combined with the twisting moments existing at these points 30 degrees after the corresponding firing dead centres have been passed, represent the maximum CRANK-SHAFTS 95 conditions of stress at the points in question. The conditions of bending when cranks 3 and 4 are on firing centre are of course the same as those obtaining when cranks 2 and 1 respectively are in that position, the order in which the bending moments occur being reversed. For example, the bending moment at No. 2 crank -pin when No. 4 cylinder is firing is the same as the bending moment at No. 3 crank -pin when No. 1 cylinder is firing, and so on. A comparison of the following table with the twisting moment curves and the bending moments tabulated in the previous articles will make the matter clear. The equivalent twisting moment equals VT^+B^, and is that twisting moment which would give the same shear stress as the maximum shear stress due to the combined action of twisting and bending moments actually obtaining. Position. Which crank 30° past firing dead centre. Bending moment in. lb. (see pre- vious tables). Number of twist- ing moment curve. Twisting moment in. lb. from curves. Equivalent twisting moment in. lb. Maxi- mum shear stress lb. per sq. in. A. Journal ■ — — — — — — B. Crank- pin No. 1 No. 1 No. 2 No. 3 No. 4 105,100 16,100 79,000 24,400 — — 105,100 3700 C. Journal No. 1 No. 2 No. 3 No. 4 74,800 54,800 71,000 56,200 (1) 127,000 11,000 13,000 17,000 147,500 5200 D. Crank- pin No. 2 No. 1 No. 2 No. 3 No. 4 25,200 74,800 57,400 30,300 (1) 127,000 11,000 13,000 17,000 137,000 4820 E. Journal No. 1 No. 2 No. 3 No. 4 62,600 80,600 80,600 62,600 (2) 112,000 115,000 31,000 31,000 140,000 4930 96 DIESEL ENGINE DESIGN Which crank 30° past Bending moment in. lb. Number of twist- Twisting moment in. lb. from Equivalent twisting Maxi- mum shear Position. firing dead (see pre- vious ing moment moment in. lb. stress lb. per centre. tables). curve. curves. sq. in. F. Crank- No. 1 30,300 112,000 pm No. 2 57,400 (2) 115,000 128,500 4530 No. 3 No. 3 No. 4 74,800 25,200 31,000 31,000 No. 1 56,200 100,000 G. Journal No. 2 No. 3 No. 4 71,000 54,800 74,800 (3) 93,000 93,000 45,000 117,000 4120 H. Crank- No. 1 24,400 100,000 pm No. 2 79,000 (3) 93,000 122,000 4300 No. 4 No. 3 No. 4 16,100 105,100 93,000 45,000 No. 1 82,000 2890 I. Journal No. 2 No. 3 No. 4 (4) 82,000 82,000 82,000 Conclusions. — Maximum Shear Stress, 5200 lb. sq. in. — Taking the fatigue stress in shear for mild steel, subject to combined bending and twisting at 15,000 lb. per sq. in., the factor of safety for a shaft newly lined up is about 3, and diminishes very considerably as the bearings become worn out of level. The high values of the ggresses at the centre of the shaft point to the advisability of making all couplings between Sections of the crank-shaft of the full torsional strength of the shaft, i.e. the aggregate shearing area of the coupling bolts multiplied by the radius of their pitch circle should be equal to the twisting modulus of the shaft. Thickness of coupling flanges ^|^ diameter of the shaft. CHAPTER VI FLY-WHEELS The Functions of a Fly-wheel are : — 1. To keep the degree of uniformity within specified limits. 2. Where alternators running in parallel are in question to limit the angular advance or retardation of rotation, to a specified fraction of a degree ahead of or behind an imaginary engine rotating with perfectly uniform angular speed. 3. To limit the momentary rise or faU in speed when full load is suddenly thrown off or on. 4. To facilitate starting under compressed air. In addition to the above the fly-wheel usually serves as a barritig or turning wheel and a valve setting disc ; also the inertia of the fly-wheel has great influence in determining the critical speed at which torsional oscillations of the crank-shaft are set up. Fly-wheel Effect. — ^The fly-wheel effect of a rotating body is its polar moment of inertia (mass X radius of gyration squared) about its axis of rotation. For a fly-wheel or pulley it is found approximately by multipljdng the weight of the rim in pounds by the square of the distance in inches from the axis to the centre of gravity of the section of the rim, the result being in in.^ lb. units. This underestitnates the moment of inertia slightly, and a more accurate method will be described later. The fly-wheel effect of the running gear of one cylinder is found with sufficient accuracy for most purposes by adding the weight of the revolving parts (crank-pin plus unbalanced part of two crank webs plus 0-65 of the connecting rod) to half the weight of the reciprocating parts (0-35 of the connecting rod plus cross-head plus piston-rod plus piston, etc.), and multiplying the sum by the square of the crank radius. For a screw propeller the radius of gjrration may be taken as 0-35 of the extreme radius if details are not available. H 97 98 DIESEL ENGINE DESIGN Degree of Uniformity. — -p. r -i- -x Max. speed— Min. speed Degree oi umformity = fj ^—^ ° •' Mean speed. Let d=Degree of uniformity. Wi=Max. angular speed in radians per second. W2=Min. „ „ „ „ „ Thend=2i^^i=:^ (1) For a specified value of " d " the necessary fly-wheel effect is calculated by means of the resultant twisting moment curve of the engine. Let Fig. 49 represent the twisting moment curve and the line A C the mean twisting moment. Let ABC be the loop of largest area (with multi- cylinder engines there are in general as - 'r.^ , - ^^ ^" ,x V - =^^ ID ^^ \ 7 O t ■ : 7 . y" -^ it -^t ' L -Z- _,Z _ [~ ■ , > o o dj H 'A O •A o A H CO & O OOC\J — OCiOO'^^ »0<:jcn 0\J**O-- C\jC*3^uliO two revolutions in the case of a four stroke engine. The brake energy developed during this period is entirely devoted to accelerating the fly-wheel and other rotating masses. Owing to the fact that the governor does not act immediately the load is thrown off, the wheel should be capable of absorbing the 102 DIESEL ENGINE DESIGN /\ ""■ 1 1 L L 2 Cyli iders lO'^ /5 " _i I 4 A 7 i , I — ^^5. t ^ t V i t _. 5 ^ -- -V -- ■- N ^ Z5 ^^ V t-. S Z^ iJ 0° m'' ak Fig. 51. 3f0» ^0 - 1 ~ 1 1 4 (.yiinders Iff' 15 ^ 1 \ 1 \ / \ 1 \ / \ /-- ... -4 — /- -- — -- J J ISO' Fig. 52. ^ [- A - - 4 - - I ., . f ._ . t- = .-. ^ , . r— 1 S Cwlinder. ; k^lf 0° 30' FlQ. 63. /4 /3 /2 /; w 9 8 7 e 6 A 3 2 / -/ -? -3 -4 -5 -6 1 1 / \3 ly inder ? \ "^ Px' 1^' i \ ^ \ \ , l_ L \ """\ ,-_. " ■ V \ 4 i\ ' \ s \ \ \ \) It iO' i 4rf elCvlilid'er s 1 lb";'■ Fig. 58. 17 16 15 Il- ls /2 // 10 9 8 7 6 S 4 3 ? / -<-J W? 45' Fig. 69. Cylinders 1 V'lE r \ \ \ ', I \ I ^-• — V — — //2Q Fig. 60. /5| I I I W — 76ML 6 5 4 3 2 / o\ I I 0" 60° Fig. 61. TWO STROKE ENGINES. 104 DIESEL ENGINE DESIGN whole power of the engine for about three revolutions in the case of a four stroke engine, about 1 -5 of a revolution in the case of a two stroke engine. Example : B.H.P. of engine (four stroke). 180 Revolutions per minute . . 375 Momentary rise in speed when full load is suddenly thrown off 12% Radius of gyration of wheel . 18 in. It is required to find the weight of the fly-wheel, neglecting the fly-wheel effect of the running gear. „- , , , ^. ^, ,,, ,180x33,000x12. „ Work done per revolution at full load r^^ in. lb. Energy corresponding to three revolutions 180x33,000x36 375 =570,000 in. lb. Angular speed at full load — — — =39-3 radians per second. Momentary angular speed when load is suddenly thrown off 39-3 X 1 •12=44-0 radians per second. If W= weight of wheel, then : — Wx 182 •; „ ° {442-39-32) = 570,000 in.^ lb. 2x386 ^ ' and W = 3470 lb. Alternators in Parallel. — ^When two or more alternator sets are being run in parallel it is a necessary condition for working that they keep almost exactly in phase. Due to inequalities of twisting moment, slight differences of phase inevitably occur, and these give rise to synchronising currents between the various machines, the tendency of these currents being to accelerate the lagging machinfp and retard the leading ones. This effect keeps the whole system in a state of stability, but cannot be relied on to correct any large fluctuations, and on this account it is usual to specify that the maximum deviation from uniform rotation shall not exceed three electrical degrees on either side of the mean. If the alternator under considera- tion has a field of two poles only, then the electrical degrees correspond to crank-shaft degrees. In general, if the number of pole pairs is " p," then one crank-shaft degree corresponds FLY-WHEELS 105 to " p " electrical degrees. So far as the engine designer is concerned, then, the problem consists in ascertaining the fly- wheel effect required to keep the cycUc fluctuations on the engine fly-wheel within a certain number of degrees, or more commonly within a certain fraction of a degree of revolution, oil either side of the mean. The method of calculation may be described briefly thus :- — (1) Assume any convenient figure for the fly-wheel effect, e.g. 100,000 in. 2 lb. (2) Plot twisting moment curve for complete period, taking the zero of ordinates at the mean twisting moment. (3) Reduce crank angles to time in seconds, assuming uniform rotation. (4) Reduce twisting moments to angular acceleration in degrees per second^ by dividing by the assumed fly- wheel effect and by the acceleration due to gravity (386 in. per sec. 2) and multiplying by the number of degrees in a radian (57-3). (5) Plot angular acceleration to time or crank angle base. (6) Integrate by planimeter, or otherwise, obtaining angular speed curve. (7) Integrate again, obtaining angular displacement curve. (8) Measure maximum deviation from the mean position in degrees. (9) Increase or decrease the assumed fly-wheel effect in pro- portion as the angular deviation so found is more or less than the deviation specified. This gives the fly- wheel effect required. Example : Three cylinder, four stroke engine : — Bore 20 in. Stroke 32 „ Revolutions per minute . .150 Number of pole pairs ... 20 Angular deviation. 3 electrical deg. Twisting moment curve as in Fig. 62. The mean twisting moment being taken as the basis. The whole calculation is contained in the table below, in con- junction with Figs. 62, 63 and 64. Since the engine makes 150 revolutions per minute, therefore 20° = -|^=0-0222sec. 150x360 106 DIESEL ENGINE DESIGN Assume fly-wheel effect of 10® in.* lb. for purposes of calcula- tion. Then : — Twisting moment x 386 . , , ^..,^. ^-. „^„^o T.M. m degrees per sec.*=^ — Acceleration in radians per sec.* — T.M. X 57-3x386 10« 45-3 Referring to the table below : — Values given in column 2 are scaled off Pig. 62. „ „ 3 are obtained by dividing those in column 2 by 45-3. „ „ 4 are obtained by multiplying the values in column 3 by 0-0222 sec. Column 5 is obtained by successive addition of speed increments. Column 6 contains corrections necessitated by the fact that the resultant of column 5 is not zero, owing to errors. Column 7 gives corrected speeds which are plotted in Fig. 63. 8 r "" \ ,--8« \ £=4 \ |R? s «<=.. \ s \ s / s K 1 — 1 V\ ^ '^ I — 1 1 — 1 1 — 1 1 — 1 / s. / s f fe « 1 V ' ■2^4 s c^ ? \, ■=c ' J V z \ 20 to SO 80 100 120 m 160 180 200 2X2400^ Fig. 62. 20 u a m mnaim miso 200220 zionej- Via. 63. 30 1 1 1 — — __- 1 — ■*"■ S 20 /* s, 5^'" / \ tSo / \ ■3?i/fl \ d &,* \ M f-20 ^ ^ — 1 — 1 — 1 _ ■30 ~ _J — —J ' 1 ~ 20 40 60 SO 100120 UOI{0ISOl00 220»)00ig- Fig. 64. Columns 8, 9 and 10 are obtained similarly to columns 2, 4 and 6. Total swing in phase (see Fig. 64) 24+28 = 52° or 26° each side of the mean. FLY-WHEELS 107 ■"13 6< ftT3 CO 00 C<1 CD © © o -^ M SP in lO 05 6 CD 6 cs !> CO o 00 g ffl (B O 3 O •?. g Cl,g gg OiMl O o o CO o o o CO o 00 o CO lO o o CO CO o m 03 o CO o OO! o o CO 03 CO GO CD O CO o I— t CO 05 CO o era oo 05 CO ■* ■* CO CD o 00 00 00 05 O o CD I— I CO o o o CO CO CO 00 CO CO CO ss-^ ii 9 o CO I— ( CD O CO CO CO I— I CO 00 00 CO CO i-H CO CO I-H a S © K 03 o o o o © CO OS CD CD O 00 CO © © © 00 03 © o as © © 05 © fclD-43 O W-3 ©M •I- © © 00 © CO r- © o © CO © 00 © CO © CO © © CD © © © © © © © © © © © © (N, ■* CD 00 © r— ( IN I— t 3 CD 00 © (N © Tin 108 DIESEL ENGINE DESIGN Q Allowable swing, 3 electrical deg. = —= crank-shaft deg. JO Fly-wheel effect assumed for calculation, 1,000,000 in.^ lb. Therefore fly-wheel effect required 1,000,000x20x26 ,„„ ,„. . .„ =— — = 173xl0*in.2 lb. If radius of gyration of wheel is 65 in. Weight of wheel=gg^^^^ = 18-3 tons. Allowance for the fly-wheel effect of the alternator rotor would reduce this figure a little. Torsional Oscillations and Critical Speeds. — In the great majority of Diesel Engines the critical speed at which torsional oscillations of the crank-shaft would occur lies far above the practical range of the engine. Serious cases do arise occasion- ally, however, and will arise more frequently when high speed two stroke engines of six cylinders and upwards become more common. The trouble when it arises is generally disposed of very simply by altering the fly-wheel effect. Supposing for the moment that a marine engine is under consideration. Then the fly-wheel, the crank masses, the propeller and the shafting, etc., constitute an elastic system having a natural frequency of torsional oscillations which depends on the amounts and positions of the fly-wheel effects of its component parts and on the stiffness of the shafting. If this natural frequency happens to be the same as the frequency of the torsional impulses due to the working strokes of the engine, then the oscillations tend to become accumulative, vibrations are felt in the shafting, and the crank-shaft may hammer in its bearings. Then the engine is running at a critical speed. In general a two stroke engine gives as many torsional impulses per revolution as there are cylinders and a four stroke engine half this number, so that a two stroke engine attains i^s critical speed at one-half the number of revolutions required by a four stroke engine. For example : Suppose the crank-shaft, etc., of a six cylinder four stroke engine has a natural frequency of 2400 complete oscilla- tions per minute, then the critical speed will be 800 revolutions per minute. A similar engine working on the two stroke cycle would have a critical speed of 400 revolutions per minute. It will be obvious that all the revolving masses in connection with the crank-shaft (apart from trifling items) must be taken into FLY-WHEELS 109 consideration, so that the critical speed of a marine engine coupled to a dynamo for testing purposes will be different to that obtained when the cBigine is installed in the ship. Natural Frequency of Torsional Oscillation. — Consider the simple system shewn in Fig. 65, consisting of a shaft fixed at one end and carrying a fly- wheel at the other. If the fly- wheel be turned through an angle against the torsional re- sistance of the shaft and then released suddenly the system will oscillate until the energy has been dissipated in friction, the angle through which any section of the shaft oscillates distance from the fixed end rm^mai .^ L 1 u^^w Mk^ Fig. 65. being proportional to the The latter is called the Node. Let f =Frequency in complete oscillations per second. F=Frequency „ ., „ minute. d=Diameter of the shaft in inches. I=Polar moment of inertia of shaft section in inches*. Z= Length of shaft in inches. W = Weight of the wheel in lb. K=Radius of gyration of the wheel in inches, g = Acceleration due to gravity in inches per second^ (386) G=Modulus of rigidity of the shaft material (about 12,000,000 lb. per sq. in.). Then :— f =^Vj^>-"=-V,^--'" If the shaft is not of the same diameter throughout its length, but consists of sections of length ?i, h, etc., of diameter dj, dj, etc., then the equivalent length ^o of shift of standard diameter do is given by ''<^Mf)'^- ■ ■ *■ For example : 1 ft. of 6 in. shafting is equivalent to 16 ft. of 12 in. shafting, 16 ft. being ( -g ) • For this reason, the lengths occupied by coupling flanges, etc., are negligible. 110 DIESEL ENGINE DESIGN If there are a number of fly-wheels or other rotating masses at different distances from the node, then the frequency is given very closely by the following : — ^-'•''^m^ etc. -(2) In the case of an engine and shafting no point on the latter is fixed, and the position of the node is to be inferred from considerations of dynamic equilibrium. It will be obvious that a uniform shaft with an equal wheel at each end will have its node at the centre, the oscillations of the two wheels being equal in magnitude and opposite in sense at every instant. It is almost equally apparent that in a similar case, with un- equal wheels the position of the node will divide the shaft into two lengths in inverse ratio to the fly-wheel effects at their ends. In general, the position of the node is determined with sufficient accuracy to enable the critical speed to be predicted within a few revolutions per minute by treating the fiy-wheel effects (moments of inertia) as though they were weights and locating the node at their centre of gravity. A crank-shaft may be treated as a uniform shaft of the diameter of the journals. Example : Six cylinder, two stroke marine engine. Bore 24 in. Stroke .... Revolutions per minute Diameter of shaft Lengths of crank-shaft Weight of fly-wheel Radius of gyration of fly-wheel . Weight of revolving parts for one crank ,, reciprocating* arts . ,, propeller .... Radius of gyration of propeller . Shafting . 35 in. 120 . 15 in. as in Fig. 66 12,000 lb. . 40 in. 4,500 lb. 51,000 lb. 10,000 lb. . 20 in. as in Fig. 66 Reduction of Shafting to Standard Diameter of 15 Inches. Length of 14" shafting 620 in. Equivalent length of 15" shafting, 62o(ijy =815 in. FLY-WHEELS 111 Equivalent length of 15* shafting between the pro- peller and fly-wheel = 150 + 81 5 + 60 = 1025 in. Moment of inertia (polar) of l&"shaft=^x 154=49,600 in.* [_^:^ ■eipgi-^ - fe 4e o- o— f op O^ ■0-^ o c 10 O^ CM fH -^eip 5/ ■oj 112 DIESEL ENGINE DESIGN Fly-wheel Effects.— W.K2 for propeller = 10,000 X 20^=4,000,000 in. lb. W.K2 for crank masses =(4500 +2500) 6 x 17-52 = 12,900,000 in.Mb. W.K2 for fly-wheel = 12,000x402 = 19,200,000 in." lb. Position of Node. — Take moments of fly-wheel effects about A : — 19-20x1025 = 19,700 12-9 (1025+37 + 48 + 74+24) =15,600 Total 35,300 The distance of the node from A=— ^-p=978 in. Dealing with the part of the system to the left of the node and applying equation (1) : — t:, „ -- 749,600 X 12 X10«X 386 „„„. .,, ,. P=9-55 . / — '—- i =2320 oscillations V 978x4xl0« per minute. 2320 And the critical speed — -- =386 R.P.M. which is far above the working range of the engine. Critical speeds of the second and third order, and so on, are possible at 193, 128, 96 R.P.M., etc. ; but these are not, as a rule, important. It is sometimes useful to repeat the calculation for different weights of fly-wheel and plot a curve connecting moment of inertia of fly-wheel and critical speed, so that the possible variation of critical speed obtainable by altering the fly-wheel can be seen at a glance. Torsional oscillations about two or more nodes are also possible, but as these involve higher speeds, and are therefore more subject to damping, it is doubtful if they are of much practical importance. « To find the Moment of Inertia of a Fly-wheel. — In the first instance, suppose the wheel in question is a disc wheel, i.e. a solid of revolution. Referring to Fig. 67, the thick, full line represents the section of the wheel. Z Z is the axis and S S is a line through the extreme radius of the wheel parallel to the axis at a distance R from the latter. Rule any line A B parallel to the axis, cutting the outline of the section in A and B. Project A and B on to S S at C and D. Join C and D to any FLY-WHEELS 113 convenient point on the axis, cut- ting A B in Aj and Bj. Proceed similarly with different positions of the line A B and join up the various positions of A^ and B^, thus obtain- ing a new figure — the First Derived Figure. Treat this figure as though it were the original figure, and ob- tain the Second Derived Figure. Similarly with this figure obtaining the Third Derived Figure. 9 C D S .■< T^ V ' 1 J J c X ■o '*[ T 1 Z \ *s > \ z Fig. 67. Let A = Area of original section in sq. in. A 1== Area of First Derived Figure in sq. in. A2=Area of Second Derived Figure in sq. in. A3=Area of Third Derived Figure in sq. in. w= Weight in lb. of one cubic inch of the material. (■E 1 P 1 p ThenA= y.dx, Ai=g^j x.y.dx, Aa^^J x^.y.dx, R3J0 ^.y.dx. Weight of wheel=2w.w 1 x.y.dx=27r.w.R.Ai Moment of inertia of wheel=27rw I x'y.dx=27rw R^A3 Radius of gyration ''=R^.-^ Ai The above hold good for any position of the line S S, which may therefore be taken where most convenient. In cases where the section tapers towards the extreme radius (a screw propeller, for instance) the line S S is best located at a distance of about one-half or one-third of the extreme radius from the axis. In the case of a screw propeller or a fly-wheel with arms, the rotating body must first be reduced to an equivalent disc wheel. This is readily done as follows : Describe a radius R which cuts through the arms or blades, as the case may be. Divide the total area of section at this radius by 27r.R and the result is the thickness of the equivalent disc at this radius. Repeat for a number of different radii covering the whole range. Types of Fly-wheels. — ^Fig. 68 shews a disc wheel cast in one 114 DIESEL ENGINE DESIGN piece and provided, with a number of drilled holes in the rim for turning the engine by means of a bar. Degree marks are cut on the edge of the rim to facilitate valve setting. Fig. 69 shews a disc wheel with a separate centre, an arrangement which Fm. 68. Via. 69. makes it easier to obtain a sound casting. Large wheels are usually cast in two pieces, and Fig. 70 shews a design which is suitable for weights up to at least 20 tons. It should be noted that no keys are provided for securing the wheel to the shaft. If the boss of the wheel is bored one-thousandth per inch of Fig. 70. diameter less than the shaft and the bolts are drawn up at about the temperature of boiling water the frictional grip is quite sufficient for the largest wheels and the danger of split- ting the boss of the wheel involved in the use of keys is avoided. The same applies to pulleys for belt or rope drives. A rather more elaborate wheel, in which greater precautions have been FLY-WHEELS 115 taken, is shewn in Fig. 71. In this case it is advisable to make the bore of the wheel the same as the shaft diameter and to give a shrinking allowance of about one -thousandth per inch of diameter to the bore of the shrunk ring. For large stationary engines some form of barring gear is necessary, and where electric power is available a motor-driven gear is a great con- venience. For marine engines a worm, or other self -locking gear, is essential, and where the auxiliaries are electrically driven an electric turning gear should be fitted, as the use of the latter greatly expedites adjustments to the valve gear. Fig. 71. Strength of Fly-wheels. — Continental practice favours a peripheral speed of about 100 feet per second for cast iron fly- wheels, and this corresponds to a stress of about 1000 lb. per sq. in. An investigation by Mr. P. H. Smith into the case of a split fly-wheel which burst at Maidenhead in 1912 shewed that the engine (the fly-wheel of which had a normal working peri- pheral speed of about 100 feet per second) was running about double its normal speed, and as the stress varies as the square of the speed, it follows that the factor of safety under normal conditions was about 4. Destruction tests of wheels and models of wheels shew that the bursting speed is about 200 feet 116 DIESEL ENGINE DESIGN per second for split wheels and 400 feet per second for solid wheels. The discrepancy seems very large and difficult to account for. Average British practice is in favour of a slightly lower peripheral speed (about 90 feet per second). With marine engines considerations of space generally necessitate a still lower figure. The strength calculations for a fly-wheel will be illustrated by an example. Example : Required to find the approximate dimensions of a fly-wheel suitable for 1 80 revolutions per minute given that W.K2=40,000,000 in.2 lb. Peripheral speed, 100 ft. per sec. Maximum twisting moment due to engine, 450,000 in. lb. Outside radius of wheel = — — — — =63-7 in, say 64 in. 27rXl80 •' Take inside radius of rim =52 in. Then radius of C.G. of rim section=58 in. Let B= width of rim. Then: — Weight of rim = I2xBx2irx 58x0-26 And approximate moment of inertia = 12 xB X 1-64 X 583=40,000,000 in.^ /lb. From which B = 10-5 in. Since the stress due to a peripheral speed of 100 ft. per sec. is about 1000 Ibs./in.^, the total tension at each joint of the rim is equal to 12x10-5x1000 = 126,000 lb., for which pull the dowel and cotter section must be designed. Allowable stress in dowel, say 6000 lb. per sq. in. Effective area of dowel section ^ ' ^ =21 sq. in. 6000 ^ Since about one-third of the section of the dowel is cut away by the cotter hole (see Pig. 72), the gross sectional area of the 21 X 3 dowel must be — - — =31-5 sq. in., say 4 in. X 8 in. 8 • Thickness of cotter = — "about 2| in. Bearing pressure of cotter on dowel 126,000 „.-„,, 2-75x4 ^ ' P®^ ^^' ^^' which is allowable. If the hole for the dowel is made i in. wider than the dowel itself, the bearing length for the cotter on the rim will be 10-5 — 4-5 = 6 in., and bearing pressure of cotter on rim FLY-WHEELS 117 126,000 =7650 lb. per sq. in., also allowable. 2-75x6 Allowable shear stress for cotter (which is in double shear), say 5000 lb. per sq. in. Depth of cotter ' =9-2 in., say 9| in. over the rounded ends. 5000x^-75 The distance " 1 " between the inside edge of the cotter hole and the rim joint must be sufficient to obviate risk of the Fig. 72. intervening metal being torn out in double shear. (This point is sometimes overlooked in otherwise well-proportioned rim joints.) Allowing a shear stress of 1000 lb. per sq. in. : — , 126,000 ,. _ . 1= ,^^^ ^ — - = 10'5 in. 1000 X 2 X 6 118 DIESEL ENGINE DESIGN Owing to the difficulty of analysing the straining actions on the arms it is well to give the latter ample proportions. An approximate method of calculation is given below. Assume that the maximum twisting moment due to the engine (450,000 in. lb.) is transmitted to the rim by means of a constant shear force across the arms, and that the be'nding moment is a maximum at each end of an arm and zero at the centre. The length of each arm from boss to rims is about 40 in., and the distance of its centre from the centre of the wheel about 32 in. Then shear force in each arm= ' =2340 lb. (assuming six arms). ^^^^ And maximum bending moment at end of each arm = 2340x20=46,800 in. lb. Taking a low stress of 500 per sq. in., to allow for direct tension in the arms, bending modulus- of arm section 500 This is satisfied by a rectangle section 6 in. x 10 in., which could be replaced by an oval section about 7 in. x 12 in. to reduce wind resistance. The bolts at the hub of the wheel are sometimes made as strong as the rim joint, in which case the core area of two bolts will be the same as the net efEective area of one dowel, viz., 21 sq. in. This gives a bolt of about 4 in. diameter. If shrunk rings are employed these will have a square section about 10-5=(3iin.)^- The above calculations must be regarded as preliminary only, and give the draughtsman a basis on which to start designing. The next step wiU be to check over the weight and radius of gyration of the complete wheel in the manner already describe r* . The dimensions and stress calculations will then be amende accordingly. ^ Literature. — For information on the strength of fly-wheel , see : — Unwin, W. C, and Mellanby, A. L., " The Elements of Machine Design," Part II. For information on critical speeds, see:— Morley, A., "Critical Speeds for Torsional and Longitudinal Vibrations," Engineering, December 9th, 1910. CHAPTER VII FRAMEWOBK A LARGE number of diflEerent types of framework have been employed in Diesel Engine construction, and a complete classification will not be attempted here. The outstanding types in successful practice may, however, be broadly divided into a few well-defined classes, as under : — / «' A " Frame Type. — This is the earliest type of Diesel Engine construction, and on account of its merits is stUl very extensively used. Referring to the diagrammatic drawing Fig. 73, it wiU be seen that a stiff bedplate of box section is provided, and that each cylinder stands on its own legs without support from its neighbours. The legs of the column are cast integrally with the cylinder jacket, into which a liner is fitted. The breech end of the cylinder is closed by means of a deep cylinder cover of box section. 119 120 DIESEL ENGINE DESIGN The main tensile load due to the cylinder pressure is trans- mitted from the cover through the jacket and legs to the bedplate. The reaction corresponding to this load occurs, of course, at the main bearings, and con- sequently that part of the bedplate be- tween the column feet and the main bearing housings must be designed to deal with the bending moment occa- sioned by the fact that the tensile load in the columns and the reaction at the bearings are not in the same plane. Casting the cylinder jacket and column in one piece reduces fitting and machin- ing operations to a minimum and the independence of the individual cylinders would appear to have no disadvantages so far as land engines are concerned. Fig. 74 shews the same type of con- struction applied to a two stroke land engine and Fig. 77 to a four stroke marine cylinder. Fig. 74. Fi8. 77. Fio, 78. FRAMEWORK 121 Slightly lengthening the column legs enables crosshead and guides to be fitted (see Fig. 78, which represents a large two cycle land engine). Occasionally one of the column legs takes the form of a steel tie rod, with a view to giving greater accessi- bility to the running gear and to enable the crank-shaft to be replaced, if necessary, without dismantling the whole engine. Unfortunately this arrangement nullifies many advantages of the " A " frame construction, as special splash guards must now be fitted to retain the lubricating oil, which office they do not always perform very efficiently, and also additional Fig. 75. Fio. 76. machining and fitting operations are introduced which add to the cost of production without increasing the efficiency of working. Figs. 75 and 76 shew this construction applied to trunk and crosshead engines respectively. Crank-case Type. — The crank-case type of Diesel Engine was introduced when a desire was felt for higher speeds, necessitating forced lubrication. The crank-case bears external resemblance to that of a high speed steam engine (see Fig. 79), On the other hand, the high pressures dealt with in the cylinder of a Diesel Engine necessitate the crank-case being strengthened internally to an extent which is not found necessary in steam practice. Sometimes the box or girder construction of the crank-case is relied upon to transmit the tensile stresses from 122 DIESEL ENGINE DESIGN the cover to the bedplate ; more frequently, however, steel staybolts are provided for this purpose (see Fig. 80). The latter procedure, however, does not justify flimsy construction Fig. 79. or a careless distribution of metal in the crank-case, as the guide pressure has still to be reckoned with and the pressure caused by tightening up the staybolts may be relied upon to * cause serious distortion of a poorly ribbed case. Fig. 80. Fig. 81. The cylinders being separate are secured either by a round, studded flange or by passing the staybolts through each corner of a deep, square flange of hollow section cast at the lower end FRAMEWORK 123 of the cylinder jacket for this purpose. The latter arrangement requires four staybolts for each cylinder, whereas with the former it is usual to arrange a pair of staybolts only at each main bearing girder. With the crank-case construction it is not necessary to make the side girders of the bedplate so strong as for an " A " frame type of engine, as the bending action referred to above is avoided and the bedplate and crank-case when bolted together form a girder construction of great rigidity. On the other hand, the upper part of the crank-case is clearly subject to bending actions similar to those which ^t ^d '^ Fig. 82. Fis. 83. occur in the bedplate of an " A " frame engine and must be designed with this fact in view. Fig. 81 shews a section through a two stroke trunk engine of the crank-case type. Figs. 82 and 83 shew the crank-case construction applied to crosshead engines. The suitability of this type of framework for marine service has been amply proved in practice. In some cases the crank-case is common to two or more cylinders, and in others the case for each cylinder is a separate casting, the individual cases being bolted together to form a virtually continuous box of great strength and rigidity. Trestle Type. — With this construction, illustrated in Fig. 84, 124 DIESEL ENGINE DESIGN the cylinders are bolted to a base plate or entablature resting on trestle-shaped columns, the feet of the latter being secured to the main bearing girders. If the guide casting itself and its attachment to the trestles are sufficiently strong, this construc- tion can compete with the crank-case type of frame in the matter of rigidity. The same effect could doubtless be achieved by some form of bracing. In Fig. 84 an alternative form of crosshead and guide is shewn, to which the trestle arrangement M 7^ E^l r Fio. 84. y I H particularly lends itself, viz.,lfche fore and aft double guide usually found on paddle steamers. The advantages of this form of guide for Diesel Engines are : — 1. Accessibility of running gear, the piston cooling gear in particular. 2. The guide blocks being free to adjust themselves to the guides are capable of sustaining a greater specific pressure, and consequently require less bearing surface than the usual type of guide shoe. FRAMEWORK 125 It is to be noted, that the trestle construction involves some little extra care to retain lubricating oil when forced lubrica- tion is used. Staybolt Construction. — ^With this construction the cylinders are connected to the bedplate by means of turned bolts only, and the saving in weight on this account amounts to the considerable figure of about 25% of the complete weight of the engine. The reduction in cost of manufacture must also be considerable when the staybars are made of ordinary bright shafting screwed at the ends. Examples are shewn in Figs. 85 Fig. 85. Fio. 86. and 86 for trunk and crosshead engines respectively. In the latter the cast columns are relatively light, being designed for the guide pressure only, and their top ends are made free to slide, in order to avoid the subjection of the column to tensile strains on the compression and firing strokes. The difficulty of enclosing such engines adequately renders them unsuitable for forced lubrication under circumstances where economy of lubricating oil is a consideration ; on the other hand, large slow-running engines, fitted with drip or ring lubrication, appear to be quite satisfactory when built on this principle. Design of Bedplates. — ^The design of a suitable bedplate involves consideration of the following points, which will be dealt with in order, viz. : — 126 DIESEL ENGINE DESIGN (1) The provision of a suitable main bearing. (2) A girder construction under each main bearing, capable of supporting the full bearing load without central support. (3) A sufficiently strong and stifE connection between the main bearing girders, forming at the same time an oil- tight tray. (4) Suitable studding or staybolt arrangements for carrying the tensUe pull of the columns. (5) Arrangements for supporting the cam-shaft driving gear. (6) Means for collecting drainage of lubricating oil to some convenient sump, whence it can readily be drawn off With a view to filtration and repeated use. (7) Facings for barring gear, auxiliary pumps, etc. Main Bearings. — Examination of badly worn crank-shafts indicates that the high bearing pressure obtaining for a short time when the piston is at its firing centre gives rise to far less abrasive action on the bearings than the less intense but longer sustained pressures due to inertia and centrifugal force in a four cycle engine. It appears that a film of oil is capable of sustaining a heavy pressure for a short time, but once the film has broken down, relatively feeble pressure is sufficient to cause abrasion, and there is small chance of the surfaces receiving a new film of lubricant until the pressure is removed. The result is that very little trouble is experienced with the lubrication of the main bearings of four stroke engines (the pressure on the journals being frequently reversed) unless the peripheral speed is so high as to reduce seriously the viscosity of the oil film by means of the heat generated by friction. In this case forced lubrication improves matters up to a certain point, but if the speed is still further augmented there comes a point when the bearings requite to be water cooled. With two cycle engines the direction of pressure is probably not reversed at all in most cases, and lubrication is con- sequently more difficult. When the peripheral speed is low, and the oil film in consequence as stable as possible, satis- factory results are obtainable even with ring lubrication of good design, if the maximum bearing pressure is kept about 30% lower than would be considered good ordinary practice with four stroke engines. If, on the other hand, high peripheral FRAMEWORK 127 speeds, or moderate bearing surfaces, or both, are required, then a system of high pressure forced lubrication would appear to be necessary, preferably in conjunction with a system of water cooling in extreme cases. The following table gives a rough guide to the limitations of the various systems of main bearing lubrication : — System of Lubrication, Peripheral Speed of Journal, feet per minute. Projected Area of One Journal expressed as percentage of the Area of Piston. FoxjE Stroke Engines — Ring lubrication Forced lubrication . 550 750 55% 40% Forced lubrication and water cooling above 750 40% Two Stroke Engines — Ring lubrication 550 75% Forced lubrication . 700 60% Forced lubrication and water cooling above 700 60% Drip or Syphon Lubricated Bearings are not used on land engines, but are sometimes fitted to Marine Diesel Engines of the open type. Apart from the fact that the caps have only to be proportioned to the inertia and centrifugal loads, these bearings are similar to those provided for steam engines, and need not be described here. Ring Lubricated Main Bearings. — ^These are similar to the bearings fitted to electrical machinery and need not be described in detail. The arrangements for catching the oil squeezed out of the bearings and conveying it back to the oil well merit careful attention, as inefficiency in this direction leads to un- necessary waste of oil. In particular, the oil spaces and holes should be as large as possible, to avoid congestion. Fig. 87 shews a very usual form. Forced Lubricated Bearings. — ^These follow high speed steam engine practice very closely, and the usual form of 128 DIESEL ENGINE DESIGN staggering the circumferential oil groove in the top and bottom brasses is generally adopted. See Fig. 88. Main Bearings Shells. — These are of cast iron in commercial work, and in the best practice the shells are tinned previous to the white metal being poured in. The chief essen- tials are : — (1) Adequate thickness of sheU. (2) Good quality of white metal. (3) Good adhesion between white metal and shell. Common proportions are shewn in Figs. 87 and 88, the unit being the diameter of the journal. The bearing cap should be designed as a beam capable of carrying a central load equivalent to the full inertia and centri- fugal load due to one set of running gear. This is possibly a Cast Iron 0-23 Cast Steel '15 Fig. 88. 0-125- little on the safe side, but reference to Chapter V will shew that the margin is not large in the case investigated there. Main Bearing Girder. — ^Where forced lubrication is used, FRAMEWORK 129 the main bearing girder may conveniently be of I section, the bottom flange being formed by the oil tray ; for ring lubrication a box section lends itself more conveniently to the formation of the oil reservoir. In any case the box section is preferable in the larger sizes. The depth of the girder is determined by that of the oil tray required to give an inch clearance or so to the connecting rod big end at the bottom of its path. Referring to Chapter V, it will be seen that the maximum reaction at a bearing for the case considered is equal to 0-8 of the resultant load due to pressure, inertia and centrifugal force, and this is the load for which the girder must be designed. In other cases the load may be less than this, but it is doubtful if in any case it approximates to the conventional load frequently assumed, viz., one-half the re- sultant cylinder load. A very debatable point is the extent to which the oil tray can legitimately be regarded as a part of the tensile flange of the girder. The author's prac- tice in this respect is to ignore the middle half of that part of the tray lying between two bearing girders (see Fig. 89). The span of the girder is the distance between the two points at which it meets the side girders. If W =Load on girder in lb . 1 =Span in in. M = Bending moment in in. lb. Then M== 0-2 Wl — approxi- mately. The assumption being that the fixing moments at the ends are negligible (which if not correct 130 DIESEL ENGINE DESIGN is on the safe side) and that the load is distributed over the journal. Allowable stress 1500-2500 lb. per sq. in. for cast iron. Side Girders. — ^With " A " frame engines the bending moment on each side girder may be taken as : — Half pressure load x Distance between centres of bearings 6 The usual stress allowance being about 1500 lb. per sq. in. Where the trestle or crank-case type of frame is used the side girders may be of lighter section (proportions will be given later). Arrangements for Carrying Tensile Pull of Columns. — With the " A " frame construction the foot of each column is secured by a row of st^ds, the stress in which when referred to the normal maximum working pressure of 500 lb. per sq. in. in the cylinder amounts to about 5000 to 10,000 lb. per sq. in., according to the size of the stud. It is very convenient to have a list of the loads which studs and bolts of different sizes can conveniently carry, and such a list is given below : — Size of Bolt Stress (Core) ' or Stud allowed, Working Load, (Whitworth). lb./m.=> lbs. i" 2000 240 r 2850 550 i" 3550 1080 i" ' 4250 1800 1" 5000 2750 H" 5250 3650 n" 5500 5000 If" 6000 6300 ir 7100 9300 If" 8500 15,000 2" 9tOO 21,500 2i" 10,000 28,000 2^" 10,000 37,000 2i" 10,000 44,000 3" 10,000 54,500 Care must be taken that none of the studs are at any consider- able distance from adequate supporting ribs. This is best FRAMEWORK 131 obtained by judicious spacing of the studs rather than the provision of special ribs for the purpose. With the crank-case and trestle t3rpes staybolts are usually fitted, and in land work at any rate these should terminate within the bedplate and not penetrate to the under side of the latter for fear of oil leakage, which would- destroy the con- crete. The studs or bolts used to secure the crank-case to the Fig. 90. bedplate may be disposed more with a view to making an oil- tight joint than to carry any definite load. If staybolts are not fitted, then a sufficiency of effective bolt or stud area must be arranged in the neighbourhood of each column foot, and some of the bolts or studs must be inside the crank-case. Cam-shaft Driving Gear. — The motion required by the valve ge£tr is derived from the crank-shaft by spiral or spur gearing in the majority of designs. Fig. 90 shews a very common arrangement of spiral drive, with the driving-wheel between the two sections of a divided main bearing. It is good 132 DIESEL ENGINE DESIGN practice to make the combined length of the two sections about 50% greater than the length of a normal bearing. There would appear to be nothing against having the spiral wheel outside the bearing altogether, provided the gear is at the fly-wheel end. This position for the valve gear drive is preferable to the compressor end as the weight of the fly-wheel tends to keep the journal in contact with its lower bearing shell, whereas the forward journal has freedom of motion (under the influence of forces which vary in direction) to the extent of the running clearance. In six cylinder engines the spiral gear is frequently arranged at the centre of the engine, where it is very easily accommo- dated. There seems to be some feeling that the cam-shaft would whip unduly if driven ^ from the end. This dif&- ■yy culty (if any difficulty can ^3 be said to exist) is easily overcome by making the • cam-shaft about 10% larger in diameter than would be considered sufficient for a four cylinder engine. Where spur gearing is used for the valve gear drive, facings must be pro- vided for the support of the first motion shaft. Oil Drainage. — With land engines of the non- forced lubricated type the oil which drips down from the cylinders and is thrown from the big ends is drained periodically from the for- ward end of the bedplate and holes are cored through the main bearings girders to give the oil free passage. Perhaps the best arrange- ment is a rectangular duct about four inches square FiQ. 91. running down the centre of FRAMEWORK 133 the oil tray. Small holes are useless as they are easily choked. With forced lubricated engines the same arrangements are made with the addition of a collecting sump of good capacity, a pump for forcing the oil into the bearings and filters in duplicate. These features being familiar in steam engine practice need not be described in detail. It must be borne in mind, however, that where trunk engines are being considered the oil is contaminated with carbon, so that the filtering arrange- ments require to be on a more liberal scale than is necessary with engines in which the cylinder is isolated from the crank-case. Proportions of Bedplate Sections. — ^Fig. 91 gives approxi- mate proportions for various types of bedplate sections, the unit being the stroke of the engine. Type "A" is usually associated with the " A " frame construction. Type " B " is a useful one for main or auxiliary marine engines as it enables the engine to be bolted direct to a tank top or to a deck with- out buHding up a special seating. Type " C " is preferable to type " A " for land generating sets as the extra depth of bedplate enables the generator to be flush with the engine room floor without the necessity of building the engine on an unsightly plinth. A deep bedplate is also very desirable with six cylinder engines as the cancella- tion at the centre of the engine of the inertia and centrifugal couples gives rise to vibrations, the amplitudes of which are reduced by increasing the stiffness of the framework. The general thickness of metal may be about 4% of the cylinder bore increased to about 6% or 8% on machined surfaces. These figures are usually exceeded oh small engines on account of the difficulty of obtaining consistent results in the foundry with thin metal. The following figures for different diameters of cylinder represent good practice : — Bore of General thickness of Cylinder, in. Metal for Bedplate, in. 10 f" 12 i" 15 if" 18 W 21 iiV 24 lA" 27 ir 30 If" 134 DIESEL ENGINE DESIGN The above are useful as a guide, but must not be relied upon without check calculations, as special constructions may- require local strengthening to keep the stresses within a safe figure. "A" Frames. — ^An "A" frame for a four stroke trunk engine is shewn in Fig. 92 ; the discussion of those parts which are common to separate cylinders as distinguished from cylinders combined with columns will be reserved for a separate chapter. The liner is a push fit or light shrink fit in the upper flange of the jacket. The fit at M should be a few thousandths slack, to prevent seizure at this point. The jacket is swelled locally, to give adequate water passage, and six or eight strong ribs are provided to transmit the tensile load across this section, which would otherwise be greatly weakened by the discontinuity of the wall. At Q the liner is a push fit, allowing the liner to expand axially relatively to the jacket. Sometimes a stuffing-box is fitted to prevent water leakage. P is the water inlet connection. L L are bosses to accommodate lubricator fittings. N is a cleaning door. Strength of " A " Frames. — Fig. 92 is drawn for a 15 in. cylinder, the stroke being 21 in., and the dimensions given re- present good average practice. The stresses may be checked as follows : — Maximum working pressure . . . 500 lb. per sq. in. „ load 0-784x152x500 = 88,000 lb. Tensile stress in parallel part of jacket (mean dia.=23-5") 88,000 no^,, r , = TT^n^ — ; =1190 lb. /m.^ :rX 23-5x1 ' Next consider section " AA " of one leg. For this section conditions are worst if the nuts at the foot are not tight and the reaction at the foot consists of a simple vertical pull of 44,000 lb. Referring to Fi^ 92, the direct tensile pull in the leg is 42,000 lb., and th* sectional area at "AA" being 57 sq. inches, Direct tensile stress at "AA" = 42,000 -^ 57 = 737 lb. per sq. in. Shear stress „ =13,500-^57 = 237 „ „ „ Bending moment „ =44,000 x 9-5=42,000 in. lb. Section modulus ,, =<^230in.^ Bending stress „ =420,000 ^230 = 1825 lb. per sq. in. Total tensile stress „ = 1825 -|-737 =2562 „ „ „ FRAMEWORK 135 This stress refers to the outside fibres of the column, and is probably in excess of what actually occurs as the fixation of the foot by the holding down studs produces a moment in the reverse direction. Fro. 92. 136 DIESEL ENGINE DESIGN Section " BB " is subject to maximum stressing action if the foot is securely fixed to the bedplate, as it should be. The vertical reaction of 44,000 lb. is again resolved into a direct pull along the axis of the leg amounting to 42,000 lb., and a shear of 13,500 lb. The area " BB " being 64 in.^, therefore the direct tensHe stress at " BB "=42,000-1-64 = 657 Ib./in.^ The shear load of 13,500 lb. at " BB " being opposed by an equal but opposite shear load at " AA " there must be a couple of magnitude 13,500 x 24 in. lb. to keep the part of the leg lying between " AA " and " BB " in equilibrium. Assum- ing that this couple is composed of two equal parts operating at "AA"and"BB," Bending moment at "BB" = 13,500x24^2=162,000in. lb. Section modulus ,, ='^170in. ^ Bending stress „ =162,000-:-170 = 953lb.per sq.in. Total tensile stress . .= 953 -|- 657 = 1610 „ „ „ The stress in the studs depends upon the degree to which the nuts are tightened, and if of sufficient area the stress is prob- ably not affected by the applied load. It therefore only remains to see if the loads induced by tightening them up to their nominal working stress are sufficient to prevent the joint opening. Referring to the table on page 130, the nominal load of five If in. studs = 75,000 lb. Subtracting 44,000 lb. there remain 31,000 lb. to resist tilting about the outside edge of the foot. The distance from to the centre of mean position of the studs is BJin., and the corresponding moment is therefore 31,000 x 6-5 =202,000 in. lb., which is greater than the moment to be resisted. Design of Crank-cases . — ^A simple form of crank-case is shewn on Fig. 93, stroke to bore ratio 1-25. The crank-case is machined on each side and in position is bolted to its neighbours, an arrangement which, though ^nusual, has its advantages, both in the factory and outside, as the small sections are easier to handle than a crank-case in one piece. Provision is made for staybolts, and the thicknesses of metal shewn are about the minimum to give satisfactory results with this design. It must not be thought that these proportional thicknesses are capable of substantial reduction with large sizes of engine without modification to the distribution of metal. On the other hand, if the interior of the case is built up in girder or box formation, FRAMEWORK 137 or generally reinforced by internal ribbing, as shewn dotted in Fig. 93, the general thickness may be reduced by about 25%, and advantage is taken of this fact in designing large engines, the castings of which would be of undesirable thickness if the simpler forms were adopted. In the event of staybolts not being used, it is desirable to check the stresses in the legs in the manner described for an " A " frame. Judging by successful 138 DIESEL ENGINE DESIGN designs, the use or omission of the staybolts has little influence on the strength of crank-case required, and this is readily explained by the fact that whereas the staybolts relieve the crank-case of tensile stresses, the tightening of the former throws a heavy buckling load on the crank-case, perhaps double the tensile load due to the working pressure in the cylinders. These considerations do not apply, however, to those designs in which the staybolts are extended upwards to the cylinder cover. In these cases the crank-case has only the guide pressure to contend with. On the other hand, the use of Fig. 94 FRAMEWORK 139 staybolts passing through lugs on the cylinders enables the latter to be pitched closer together than would be easily possible otherwise, and in any case they add strength and rigidity at very little expense and increase in weight. The above notes on the strength of crank-cases, as well as the figures for the thickness of metal, apply equally to cross- head as to trunk engines. The additional height of the former has little if any influence on the matter, as the guide reaction acts at the same height, above the centre line of the crank-shaft, assuming the same length of connecting rod in each case. Some alternative forms of construction are shewn in Fig. 94. In the case of very small engines the use of the minimum thickness of metal allowable on considerations of strength and rigidity only, would give rise to trouble in the average foundry and additional thickness is usually given. In the following table foundry considerations are neglected, as these must be dealt with by the designer in each individual case, in accordance with his judgment of the capabilities of the foundry in question, and in this matter the foundry manager will be able to give assistance. General thickness of General thickness of Diameter of Whit-worth Bore of Crank-case metal. Crank-case metal. Cylinder, in. m. Plain type,. m. Box or girder formation. Staybolts, in. Fig. 93. Fig. 94. 1 10 i" tV" ir 12 r r If" 15 i" x%" H" 18 i" i" 2i" 21 ItV" I" 2|" 24 ifV" I" H" 27 If" ItV" 3|" 30 H" ifV" 3|" The figures for the diameters of the staybolts are based on the assumption that they carry the whole pressure load. In cases where the cylinders are secured to the crank-case by a studded flange the staybolts if fitted at all may be made considerably lighter, according to judgment or the results of experiment. Other points to be considered in designing a crank-case are : — 140 DIESEL ENGINE DESIGN (1) The provision of oil-tight access doors of ample size for overhauling the bottom ends. (2) End casings provided with oil flingers, stuffing boxes, or other means of preventing the escape of oil. (3) Facings, and other necessary accommodation for valve gear. (4) Bosses to carry lubrication oil connections to the main bearings. (5) Facings for platform brackets. (6) A vent pipe or valve of large area, to relieve pressure in the event of an explosion in the crank-case without loss of lubricating oil during normal working. (7) Steady pins to each section of the case, to fix correct location. Machining the Framework generally. — In designing all parts of an engine the designer will keep in mind the capabilities and limitations of the manufacturing plant and the operatives. This is especially necessary in the case of the framework, on account of the relatively large size of the parts. Where the most modern type of face milling plant is available the element of size offers no difficulties, and bedplates of 60 feet in length may be faced in one operation. Where planing must be resorted to the capacity of the machines must be studied in the early stages of the design. Machined faces should be arranged in as few different planes as possible, and ribs or flanges projecting beyond those planes are to be avoided as much for convenience in machining as for the sake of appearances. The simpler forms of girder or box-girder construction are to be preferred to those designs in which alternate perforation by lightening holes and reinforcement by ribbing mutually defeat each other's object. The lightest, strongest and cheapest forms are to be attained with a minimum of holes and ribs when cast iron is used. Large steel castings, however, are preferably lightened out almost to the extent of lattice-work, in^brder to facilitate rapid stripping of the cores after solidification and to minimise initial stresses. Literature. — The different types of framework used in Marine Diesel Engine construction, and the forces acting on them, are discussed in the following paper : — Richardson, J., " The Development of High Power Marine Diesel Engines," Junior I.E., April 20th, 1914. CHAPTER VIII CYLINDERS AND COVERS General Types. — ^The great majority of Diesel Engines are provided with cyliader liner, jacket and cover as separate pieces, as in Fig. 95, which refers to a four cycle trunk engine. Different arrangements have, however, been used successfully, and deserve mention. With small engines, simplification is achieved by casting the jacket and liner in one piece, as in Fig. 96. Remembering that the bulk of the jacket wall remains Fig. 95. Fig. 96. 141 142 DIESEL ENGINE DESIGN stone cold, it will be appreciated that this construction involves increased tensile stresses on the jacket, due to the tendency of the liner to expand, and jackets of this kind have been known to crack circumferentially. In cases where staybolts have been fitted to carry the tensile stresses from the cover downwards little damage has resulted. On the other hand, when the jacket has been relied on for this function, rupture during working may easily occur, and has sometimes resulted in the cylinder being projected towards the roof. These considerations would appear to indicate that the use of this construction, "without staybolts or other safeguards, is not lightly to be attempted without serious con- sideration of the capabilities of the foundry. In Fig. 97 is shewn a construction in which the cover is incorporated with the cylinder casting in motor-car style. In this case the tensile stresses are mainly carried by the liner, and the jacket is made relatively thin and flexible. This design, though probably safer than that of Fig. 96, also makes some demand on the skill and care of the foundry people. In this connection it is worth while bearing in mind that many failures might possibly have been avoided if it had been realised that certain continental designs, in which lightness has been, a primary consideration, were only practicable if the greatest care were exercised in the selection of material and in making the castings. There are other types of cylinder in successful use, notably .nose in which the liner and cover are cast together apart from the jacket, but the remainder of this chapter will be devoted tcjthe consideration of the details of the more common construction, in which the liner jacket and cover are separate pieces. Unless the contrary is stated, cast iron is understood to be the material in each case. Liners. — Special cast iron is used for liners, but there is little unanimity of opinion as to the most desirable properties be- yond the obvious requirements of soundness and homogeneity. The greatest difficulty to be overcome is abrasion by the piston rings. At present it seems open to question whether the Fig. 97. CYLINDERS AND COVERS 143 problem is most influenced by the material of the liner or the piston rings themselves. It is very seldom that liners crack* either in four or two stroke engines, and this immunity is doubtless due to the rapid conduction of heat from the breech end to those parts which only come in contact with gases at relatively low temperatures, from which it would appear that the best way to cool a relatively inaccessible spot is to connect 03 to 'OS 8 'Atofs^ Fig. 99. the latter to a large, well-cooled area, as near to it as possible. This principle will be referred to again in connection with exhaust valve casings. Typical liners for four stroke engines of the trunk and cross- head types respectively are shewn in Figs. 98 and 99. With the latter the piston is only of sufficient length to carry the rings, and the length of the liner is determined by the position of the bottom ring at the bottom of the stroke . With the trunk engine the liner must be long enough to embrace a sufficient * Except at a piston seizure. 144 DIESEL ENGINE DESIGN length (about equal to the bore) of the parallel part of the piston when at the bottom of its stroke, in order to avoid a piston knock at the bottom dead centre. It is therefore necessary to determine the clearance volume and complete the design of the piston before the length of the liner can be fixed finally. The bore is usually parallel with four stroke engines and slightly barrelled in way of the ports in two stroke engines to allow for the restraints which are inevitably placed at that position against free expansion of the liner. Probably the best bore is produced by finishing with a reamer in a vertical machine. Grinding is frequently adopted, but there is a question if this prpcess does not to some extent destroy those properties of cast iron which facilitate good lubrication. The outside surface is frequently left unmachined in competitive work, and there is probably no serious objection to this practice for four cycle work. For two cycle engines it seems reasonable to take advantage of the increased heat conductivity obtain- able by removing the skin. Strength of Liners. — ^The upper end of the liner is subject to a working pressure of about 500 lb. per sq. in., and the thick- ness at this part measured under the heavy top flange may be found by the following formula, which represents average practice for substantial engines : — Thickness =0-08 bore -|- J" The working stress being about 3000 lb. per sq. in. in the case of a 30 in. cylinder, and less in smaller sizes ; explosions at starting, etc., may nearly double this stress occasionally. Unfortunately the available information on the effects of repeated stress is not sufficiently complete at present to enable one to say definitely whether or not these excesses of stress have any influence on ultimate failure by fatigue, but the writer is inclined to believq|(on the strength of such evidence as has come before his notice) that the elimination of these occasional excess pressures would not enable any substantial reduction of thickness to be effected with the same margin of safety. On account of the diminution of pressure on expansion the liner may be tapered to a thickness of about 0-04 bore at the open end. The breech end of the liner requires to be reinforced by a CYLINDERS AND COVERS 145 heavy flange, to avoid distortion due to the pressure of the cover on the spigot joint. Proportions are given in Fig. 98. Points of Detail. — The difficulty of accommodating the valves in the limited space available in the cover of a four stroke engine usually renders it necessary to make recesses in the top of the liner to clear the air and exhaust valve heads (see Fig. 98). Four or more tapped holes are provided in a circumferential line round the liner to accommodate the lubricating fittings, these being drUled when the liner is in position in the jacket. The holes are located at about the level of the second piston ring (counting from the top), when the piston is at the bottom dead centre. The fittings themselves will be described later. The water-joint between liner and jacket at the lower end may be made by one or more rubber rings. The joint between cover and liner may be of copper or asbestos compositions. Two stroke liners are complicated by exhaust, and some- times air ports (see Fig. 102). In the earlier designs the bars between the latter were always provided with water passages, which introduced difficulties in manufacture, and the value of which seems doubtful, and these are now frequently omitted. The fitting surfaces at this point are preferably ground, to minimise chance of leakage, as the high temperature prohibits the use of rubber packing rings. Cylinder Jackets. — A simple and effective form of jacket for a four cycle engine is shewn in Fig. 95, and in this example the jacket takes the pressure pull without the assistance of stay- bolts. The chief points to be observed are : — (1) A heavy flange at the top to carry the liner and to enable the tensile forces concentrated at the studs to dis- tribute themselves uniformly round the jacket without producing high local stresses. (2) A nearly plain cylindrical barrel, as nearly as possible in line with the pitch circle of the cover studs and provided with sludge doors, bosses for lubricating fittings, and a bracket for supporting the cam-shaft bearings. (3) A circular flange at the bottom for securing to the crank- case. The remarks re tensile forces under heading (1) apply here also, but to a less degree, as the studs are pitched closer to- gether than would be feasible on the cover. On these consider- 146 DIESEL ENGINE DESIGN ations the thickness of the jacket for eoual strength should taper gently towards the middle, and the form shewn in the figure is the practical compromise. Some of these points will be considered in greater detail. Top Flange of Cylinder Jacket. — In small engines this may be soUd, but with larger sizes, say from 15 in. bore and upwards, difficulty is sometimes experienced in obtaining sound metal at this point, and coring of the flange between the studs is resorted to in order to accelerate cooling in the mould. Different constructions are shewn in Fig. 100. Schemes A and B have the additional advantage of increasing the cooling surface. Where four cycle engines are concerned the im- portance of this consideration is probably negligible. Scheme B requires a water outlet connection between each stud if air pockets are to be avoided. On the other hand, the expense of coring is less than with scheme C. Barrel of Cylinder Jacket. — ^This is sometimes conical, instead of cylindrical, and in*his case it is reasonable to provide a vertical internal rib under each stud to discount the addi- tional stress involved. Consideration of manufacturing costs, and of the good appearance of the engine, rule out of court any form of external ribbing. The brackets supporting the valve gear take many forms in different designs. That shewn in Fig. 95 is the modern form, and considered in conjunction with the gear it supports appears to combine most advantages, including that of elegance. CYLINDERS AND COVERS 147 Bottom Flange of Jacket. — If four staybolts are provided for each cylinder, these may conveniently be used to secure the latter to the crank-case. The concentration of the tensile load at four points necessitates a heavy flange, preferably of box form, as shewn in Fig. 101. I . The corners of this flange 8-/^5turfs, ..—"0^ being each subject to a load of one-quarter of the maximum working pres- sure load, deserve atten- tion in the form of a calculation of the bending stress involved. A plain, square shape would appear to be preferable to some of the more elaborate shapes which have occasionally been used, the flat sides lending themselves well to the provision of facings for various purposes. Frequently the flange is spigoted into the top of the crank-case, but as this involves an unneces- sary machining operation on the latter and makes cylinder alignment more difficult, the better prac- tice is to core the aperture in the crank -case suffi- ciently large to allow for adjustment of the position of the cylinder and to locate the latter by means of two steady pins. Strength of Four Stroke Cylinder Jackets. — ^The consider- atidns of strength which enter into the design of a cyhnder jacket are illustrated by the following check calculations relating to the cylinder shewn in Fig. 101. Bursting stress in liner 500 (lb. per sq. in.) X 6-75 „„„„„ = ^ ■, -.2^ — =3000 lb. per sq. m. 1-125 ^ ^ Fig. 101. 148 DIESEL ENGINE DESIGN ■vr • 1 11 • u ^ J 500x0-784x142 „...,, JNominal pull in each cover stua=^ =9650 lb. o Permissible nominal load for 1| in. stud, according to table on page 130, 9300 lb. Maximum working pull in jacket=0-784x 13-52x500 = 71,5001b. 71 500 Tensile stress in jacket= ' — pr-;7^=1670 lb. per sq. in. Owing to the peculiar shape of the bottom flange the calcu- lation of its strength presents a difficulty which is easily evaded by substituting for the actual section a simpler one of obviously inferior strength. Nominal load at each corner, 71,500 lb.-^-4 = ~18,000 lb. Moment from centre of bolt to jacket wall=18,000 x 9 in. lb. Modulus of hypothetical section : — ^^/10-5X53_9-5X3-53N ^,^^3^.^.^3 o. 18,000X9 '. ;- .nAii. .•. Stress< — ^--— — I.e. <5,400 lb. per sq. m. In view of the unfavourable assumptions this is probably not excessive for first-class cast iron. Jackets for Two Stroke Engines. — The necessity for provid- ing exhaust passages or belts, and in some cases passages for scavenge air as well, introduces considerable complication into the design, renders the stresses in certain parts more or less in- determinate, and makes greater demand on the skiU of the manufacturing departments, in comparison with that required by four cycle construction. Referring to Fig. 102, it will be seen that the exhaust belt interrupts the vertical line of the jacket wall, and if the latter has to carry the main tensile stresses internal ribbing becomes a necessity. The arrangement shewn is perhaps as good as any, but the attachment of ribs to*he exhaust belt has a restraining influence on the temperature expansion of the latter which can only result in mutual stresses. It appears, however, that these are not very serious, as cylinders which have failed in other respects have remained intact at this point. Fig. 103 shews a construction in which a good attempt is made to secure continuity of the vertical wall of the jacket. Either of these systems is probably satisfactory for cylinders of medium size. Large cylinders, however (and this applies to Fig. 102. Fie. 103. Fig. 104. 150 DIESEL ENGINE DESIGN other parts as well), are known to be subject to greater temper- ature differences than smaller ones (though not to the extent sometimes suggested,), and the leading designers have had recourse to other expedients when faced with the problem of constructing cylinders of large size. In Fig. 104 the jacket wall may be described as similar to a honey -pot in shape and of abnormal thickness, to allow for the bending stresses caused by the curvature of the walls and the fact that the tensUe supporting forces are localised at two feet. Cylinder Lubrfcating Oil Rinii Main. bl, , . .. „ -" T^Lubncating Oil Inlet Fig. 105 (See page 152). The exhaust belt is of relatively thin metal, with comparatively small support from the walls. It wUl be evident that the strength of the jacket is very slightly influenced by the exhaust belt, and that the latter is free of all but temperature stresses. This construction, therefore, Sttains a good approximation to the correct allocation of the respective duties of jacket and exhaust belt. As disadvantages, may be cited abnormal weight of cylinder and the difficulty of casting a cylinder involving widely different thicknesses of metal. Another and perhaps better way out of the difficulty is to connect the cylinder cover to the bedplate by means of stay- CYLINDERS AND COVERS 151 bolts, thus relieving the jacket of all stresses except those induced by temperature differences. The jacket in this case virtually hangs from the cylinder cover, and only requires to be attached thereto by studs proportioned to a load based on the cyhnder pressure and the annular area lying be- tween the cylinder bore and the spigot at which the cover joint is made. The upper flange is preferably made fairly substantial, but other thicknesses may be made a practical minimum. Cylinder Lubricating Pump. Sti"ofee Hole toprevent AirLock ^ Driven by Eccentric or LinkXvork .t— i Plunders arranged i 7n a Ore le Suction fei by drip Lubricator ^Delireries * Fig. 105 (See page 152) Cylinder Lubrication. — ^The problem of cylinder lubrication in Diesel Engines consists in effecting uniform distribution of minute quantities of oil. The quantity of oil admitted must be the minimum necessary to effect satisfactory lubrication, as the oil " cracks " in service, leaving a gummy deposit, which in course of time causes the piston rings to stick. Under favour- able conditions this may be several months, even a year. Every drop of superfluous oil reduces this period, hence the importance of uniform distribution so that every part may have sufficient, but none a superfluity. These conditions are 152 DIESEL ENGINE DESIGN best secured, by a separate controllable feed to each of about six or eight points round the circumference of the cyhnder. A typical lubricating fitting is shewn in Fig. 105 (pages 150-152), and the point to be observed is that the fitting must adapt itself to slight relative movement between the liner and jacket. The small hole at the end which leads to the surface of the liner reduces to a minimum the chances of the fitting becoming choked with carbon. With forced lubricated engines, in which the cylinder is not isolated from the crankpit, it frequently happens that more than sufficient oil reaches the cyhnder, apart from any arrangements made for the purpose. In this case the problem may be to devise scraper rings, vent holes, or other devices, to remove the superfluous oil. Liner ^Jacket Cylinder Lubricating Fitting. Fig. 105 (See also pages 150, 151). Cylinder Covers. — Owing to a considerable number of failures in service and difficulties experienced in manufacture, cylinder covers for both four and two cycle Diesel Engines have come to be regarded as difficult pieces of design, and it may perhaps be instructive to review the subject in a more or less historical manner. The earlier type of four c^le cover is shewn in Fig. 106, from which it will be seen that the internal coring is compli- cated and that a few core-holes of small diameter only are provided to vent the core in the mould. In spite of these dis- advantages, such covers have given good service when made by the most skilful of continental manufacturers. Dismissing for the moment the question of manufacturing costs, these covers have the following shortcomings : — CYLINDERS AND COVERS 153 Fio. 106. 154 DIESEL ENGINE DESIGN ( 1) The thin walls of uncooled, metal between the recesses for fuel and, exhaust valves are liable to crack on over- loaded engines. (2) The hot exhaust passage is too rigid to permit of much expansion and leads to cracking of the bottom plate. (3) The small core-holes give poor access to the interior for purposes of cleaning away accumulated scale. Assuming first-class foundry work, the two latter considera- tions are perhaps the most important. Modern development is on the following lines : — The provision of large doors, which serve the double purpose of providing good access for cleaning and (1) affording better support and venting for the core when casting. (2) Elimination of all internal ribs, as experience seems to shew that the tubular walls provided to accommodate the valve casings provide all requisite support between the top and bottom plates. (3) Using brass or steel tubes expanded into the recesses for the fuel valve and holding-down studs. (4) The use of square instead of conical seats for the valve casings. Fig. 107. A cover designed on these lines is shewn in Fig. 107. Some makers have simplified the question of casting at the expense of introducing extra machining and fitting operations by making the top plate a separate piece (see Fig. 108.) Another innovation which is becoming increasingly common is to place the fuel valve off centre. This arrangement enables CYLINDERS AND COVERS 155 the cooling space around, the fuel valve to be increased, but too great a displacement of the fuel valve from the centre position necessitates a special shape of combustion space. Points of Detail. — Owing to the large recesses for the valve cages, a four cycle cover is relatively weak, considering the amount of metal in it, and on this account all stud holes should be well bossed under and all inspection openings well reinforced by compensating rings like a boiler. The under face of the cover is machined all over, but on the top face machining is sometimes restricted to those parts which are occupied by valves, etc. This enables the corners to be given a liberal radius, which in addition to improving the Fig. 108. appearance facilitates moulding (see Fig. 107). From all considerations, all internal angles should be well radiused. Water is led to the cover by one of two methods : — (1) By one or more tubular fittings screwed into the top of cylinder jacket and passing through holes in the under face of the cover (see Fig. 109). With the type ot jacket shewn in Fig. 100b it is desirable to fit one such fitting between each pair of cover studs. (2) By means of an opening in the side of the cover (see Fig. 110). Whatever means be adopted, it is advantageous to fit internal pipes or baffles to encourage fiow towards the fuel valve, as accumulation of deposit at this point is to be avoided at all cost. It is usual to arrange the outlet above the exhaust branch, as stagnation at this point is also undesirable. 156 DIESEL ENGINE DESIGN Proportions of Cylinder Covers. — ^The depth of a four stroke cover generally works out to about 0-7 of the cylinder bore, the limiting factors being the size of the exhaust passage and the water space around it. The former should be at least equal in area to the exhaust valve at full lift. The passage starts by being rectangular in shape at the valve end, and gradually becomes circular at the outlet where the diameter is about 0-31 of the cylinder bore. The same applies to the air inlet passage. The thicknesses of metal vary considerably in different designs, and the proportions shewn on the sketches represent average practice. The tendency seems to be to Fig. 109. Fig. 110. rely for safety on a good thickness of metal on the bottom plate. Strength of Four Stroke Cylinder Covers. — The system of loads acting on the cover comprises the tightening stresses of the studs, the reaction at the spigot joint and the gas pressure on the lower plate. The effect of such a system is to produce tensile stress in the top plate and compression on the bottom. Considering the relative weakness of cast iron in tension and the fact that cracks in the td|) plate are of very rare occurrence, it would appear that covers proportioned in accordance with average practice have a good margin of safety so far as pressure stresses are concerned. In view of a few isolated failures, or rather as a matter of principle, the strength should be subject to calculation. Owing to the uncertainty as to actual condi- tions the method of calculation detailed below must be con- sidered comparative rather than absolute. CYLINDERS AND COVERS 157 The assumptions underlying the method are as follows : — (1) That the severest conditions of stress are due to a cylinder pressure of 1000 lb. per sq. in., due to pre-ignition, careless starting or otherwise, and that this pressure causes the cover to lift to such an extent that the reaction at the joint is eliminated. (2) That the stress is uniform across a diametrical section in the case of a cover of constant depth and proportional to the distance from the neutral axis of the section in the case of a cover of varying depth. This is not correct, but probably involves approximately equal percentage of error in different cases. Fig. 111. Example : Referring to Pig. Ill, shewing the weakest section passing through the recesses for the air and exhaust valves, the section modulus is 114 in. ^ Considering the forces to the right or left of this section, we have : — (1) A downward force at the stud circle equivalent to a pressure of 1000 lb. per sq. in. over half the circular area, extending to the joint spigot, viz.: -784 x 13-252 x 1000-^2 = 69,000 lb. This may be considered to act at the centre of gravity of the pitch semicircle, that is at a distance of 9-5x2^7r=6-02 in. from the section under consideration. 158 DIESEL ENGINE DESIGN ( 2) An equal and, opposite force on the under side of the cover acting at the centre of gravity of the semicircular area extending to the joint spigot, i.e. at a distance of 6-625x4 3n =2-8 in. from the centre. The stress is therefore : — 69,000 (6-02-2-8) 114 ' = 1,950 lb. per sq. in. Putting the above into the form of a rule :- f = 1000 R^g (Bz-f Ri) Where 1000= Assumed maximum pressiu-e. f =Stress in lb. per sq. in. E,2=Ba'dius of stud pitch circle in in. Ri=Inside radius of joint ring in in. z=Section modulus in in. * Two Stroke Cylinder Covers. — ^Where port scavenge is adopted the cover has only to accommodate the following fittings : — (1) Fuel valve. (3) Relief valve (if fitted). (2) Starting valve. (4) Indicator tube fitting. Fig. 112. CYLINDERS AND COVERS 159 As all the above are relatively small the construction of the cover is much simpler than that for a four stroke engine, and failures are rare. In addition, the absence of air and exhaust valves enables the lower plate to be dished upwards (see Eig. 112), an arrangement which gives greater freedom of expansion and a shape of combustion space which is probably favourable to combustion. Eig. 112 shews a cover of this type arranged for four staybolts. When valves in the cover are employed for scavenging " ^A^W^^^^^^ Fig. 113. purposes the construction depends on the number of valves. If two scavenge valves are used the cover may be of the four cycle type and interchangeable with those of four cycle engines of the same bore. When fitted to a two cycle engine the average temperature of the lower plate will be higher ; but on the other hand the temperature distribution will be more uniform than that of a similar cover fitted to a four stroke engine where one of the valves is used to conduct exhaust gases. The inequality of temperature over the bottom plate of a four stroke cover is strikingly illustrated by the observed fact that in the event of failure the first crack almost invariably runs 160 DIESEL ENGINE DESIGN between the pockets for the exhaust and fuel valves. A crack between the air valve and, the fuel valve sometimes develops at a later stage. Fig. 113 shews a cover designed to accommodate four scavenge valves. It will be noticed that the interior is divided by a horizontal diaphragm separating the air space and the water-jacket. It appears that this diaphragm and the tubular connections to the bottom plate impose too great restrictions on the expansion of the latter, and fractures have been fre- quent (with both cast iron and cast steel), so that this type of cover, as hitherto designed, must be considered a failure. The writer understands that a modification of this design (patented by Mr. P. H. Smith), shewn in Fig. 114, has proved satisfactory, and at present no failures have been reported. Apparently the additional depth of the water-jacket, and correspondingly increased freedom of expansion, minimise temperature stresses, and the support afforded by the ex- ternal shell keeps the bending stresses to a moderate figure. FiQ. 114. Literature. — "Diesel Engine Cylinder Dimensions." — Article in Engineering, September 5th, 1913, Richardson, J. — Paper on Marine Diesel Engines, loc. cit., p. 64. Hurst, J. E. — " Cast Iron with Special Reference to Engine Cylinders," Manchester Assoc. E., December 9th, 1916. CHAPTEB IX RUNNING GEAR Trunk Pistons. — ^These are so well known in connertion with petrol motors, gas engines and, the like that a general descrip- tion is unnecessary, and we may at once proceed to the con- sideration of their special requirements in Diesel Engine con- struction. The difficulties involved in combining the piston proper with the crosshead arise chiefly from the heat which reaches the gudgeon pin bearing by conduction and radiation, and the high pressures dealt with in Diesel Engines (as com- pared with gas engines) necessitate a high specific pressure at this bearing owing to the limited space available. The most serious troubles to be anticipated are piston seizures, which like all other heat troubles are more pronounced in large than in small engines. For these reasons trunk pistons are not generally used for cylinders exceeding about 22 inches in diameter. For marine engines the inaccessibility of the gudgeon pin bearing is usually considered an insuperable objection for all but the very smallest cylinders. It is possible that prejudice has a little to do with this view, and it is interest- ing to note that Semi-Diesel Marine Engines of fairly large size (500 H.P. for example) apparently give good service with trunk pistons. It is evident that a really efficient system of water cooling would abolish most of the difficulties mentioned. Unfortunately such systems as hitherto fitted to trunk engines have not all been uniformly successful. On the other hand a great measure of success has been achieved in some cases and the ultimate solutions (if not already achieved) are probably near at hand. For crosshead engines water cooled pistons on two or three different systems have already emerged success- fully from the experimental stage. Material. — The use of cast iron for pistons is almost universal on account of its good wearing properties and its cheapness. Owing to the low guide pressure the quality of the metal is M 161 162 DIESEL ENGINE DESIGN probably of minor importance so far as wear is concerned,, as the latter is in any case hardly measurable. On the other hand, so far as that portion of the piston is concerned which is in contact with the working fluid (viz., the piston crown), the quality of the metal is of great importance in determining the liability or otherwise of the crown to crack under the influence of heat. According to Mr. P. H. Smith, the irons whicli give the best results are those of the coarsest possible grain. Experience in other directions seenis to indicate that a large carbon content and low percentage of phosphorus are favourable. At the first glance it might appear strange in view of the ^ ^ ^ 1^ ^ 1 D. ^ .? £ y F. X Fis. 115. fact that heating produces compression at the point of maxi- mum temperature, that the cracks start at the centre of the crown on the side in contact with the gases, but -the facts are easily explained by some such hypothesis as the following : — The local heating causes local compressive stress of high intensity which in course of time causes the particles to re- arrange themselves in such a iaanner that this stress is reduced. On cooling the contraction of the surrounding metal induces tensile stresses in the centre equal in amount to the extent by which the original compressive stress has been reduced. Con- siderable support is afforded to this theory by the fact that the cracks develop in the course of time into fissures, shewing that the material has contracted circumferentiaUy. Also the plastic deformation of cast iron at a red heat is often observed RUNNING GEAR 163 in such familiar articles as kitchen ranges and the like, in which no special provision is made for expansion or growth. Shape of Piston Crown. — ^Fig. 115 shews some alternative shapes. Of these types C is the best for combustion, though types A and B are little inferior in this respect. With existing methods of fuel injection, type E is quite inadmissible on the score of combustion unless the cylinder cover is concave downwards, and even then its efficiency is very doubtful. On the other hand, type F, with the shape of combustion space indicated appears to give good results, due doubtless to the manner in which the charge of air is concentrated. A little reflection shews that the curved shape of type B gives rise to greater compressive stress at the centre on the combustion side than type A on account of the bending action which arises when a state of temperature stress comes into existence. Type D is very liable to failure unless a spreading flame plate is used. From this it would appear that of aU the types illustrated B and D are the most liable to failure. In practice types B and C are the most usual, and their inherent weakness in the larger sizes (small pistons rarely crack) is guarded against (1) Careful selection of material. (2) Providing a considerable thickness of metal. (3) Water cooling. (4) Providing loose crowns, or central cores. The provision of a great thickness of metal assists matters in two ways : — (a) By reducing the bending stresses due to both tem- perature (probably the most vital) and pressure. By giving additional area for heat flow to the walls of the liner. (b) Fig. 116, Fig. 116 is an attempt to give a diagrammatic representation of the heat flow by increasing the inten- sity of shading towards the parts having the higher temperatures. The usual form of water cooled Fig. 117. 164 DIESEL ENGINE DESIGN trunk pistons is shewn in Fig. 117, and the means adopted to convey the water to and from the water space wiU be discussed later. In the earlier types the space was intended to be full of water, but in more recent designs the tendency is to rely on a smaU flow of water, part of which is evaporated and absorbs a relatively large quantity of heat in latent form. Fig. 118 shews a loose piston top secured by four studs, the holes for which have sufficient clearance to allow for the expansion of the former. Fracture of such a loose top due to temperature efiects is improbable, and if it occurs the cost of renewal is trivial. Fig. 119 shews a similar design patented by Mr. P. H. Smith and applied by him to pistons in which cracks had Fia. 118. Fib. 119. already appeared. Proportions of Trunk Piston Bodies. — The usual proportions found in practice are discussed below with reference to Fig. 120, which is purely diagrammatic. The thickness (C) of the crown (with uncooled pistons) increases rapidly as the bore is increased, for reasons which have been noticed above, and the following figures are a guide to good practice : — Bore of cylinder Thickness C 10" . . li" 12" . . If" 14" . . 21" 16" . . 3" 18" . . 4" 20" . . 5" Fio. 120. The distance from the top of the piston to the first ring may con- veniently be made equal to C. With RUNNING GEAR 165 the above values of C this prevents the first ring being placed in too high a position where its proximity to the source of heat would cause it to become stuck with carbonised lubricating oil in a short time. The rings themselves may be of square section with R=0-025 to 0-033 B. The gap between the ends of the ring when free may be about 2-50 times R^. The number of rings fitted varies from about five in small to eight in large cylinders, and the space between consecutive rings is not as a rule less than R. The construc- tion of piston rings is rapidly be- coming a specialised branch of in- dustry and details will not be given here. A method of designing and manufacturing piston rings is de- ' j.^^ ^21 scribed by Guldner in considerable detail ("Design and Construction of Internal Combustion Engines "). It is important to prevent the joints of the rings working into line during service, and a method of fixing them is shewn in Fig. 121. Dimension " A " varies greatly in different cases. The maximum value found in practice, viz. A=2B gives a piston of ideal running properties, but is seldom fitted nowadays owing to considerations of first cost. A=1-4B to 1-6B repre- sents good average practice. Still smaller values of A are sometimes used, but are not to be recommended. The pro- vision of a long piston skirt is advantageous from the following points of view : — (1) Rendering possible a low and therefore comparatively cool location for the gudgeon pin bearing. (2) It minimises piston knocks. (3) It facilitates the flow of heat away from the crown by providing a large surface in contact with the cylinder walls. The upper part of the piston body is turned taper to allow for expansion, and the approximate allowances to be made on the diameter are given in Fig. 122, in which the taper is greatly exaggerated. The diameters of the piston ring grooves are figured by allowing all the grooves to have the same depth below the tapered surface. The clearance behind the grooves 1 R denotes the thickness of the ring measured radially. The bars K between the rings may be equal to or slightly greater than R. 166 DIESEL ENGINE DESIGN should be a practicable minimum. The gudgeon pin is usually located either at the centre or slightly above the centre of the parallel part of the body, and allowance is made for possible local distortion due to driving in the gudgeon pin or to ex- omtoo-ozo Fig. 123. Fig. 122. pansion of the latter by relieving the piston surface to the extent of about 20-thousandths of an inch, in the manner indicated in Fig. 123. Approximate figures for other main proportions are given below : — E=2G. G=0-4B. H=0-5B. C (see Table above). D=0-07B. E=0-033B. Gudgeon Pins. — Alternative forms of gudgeon pins are shewn in Fig. 124, type A being most generally used. The pin itself is of special steel, case-hardened and ground if working in a bronze bearing. If the bearing is white-metaUed the case- hardening is unnecessary. The pin is a driving fit in the body and is secured in the mann# indicated in the illustrations. Lubrication of the pin may be effected in various ways : — - (1) Forced lubrication by means of a pipe or drilled hole leading from the big end of the rod. (2) By means of a groove or pocket on the surface of the piston communicating with the surface of the gudgeon pin and fed by means of a fitting similar to that shewn in Fig. 105 (see Fig. 125). RUNNING GEAR 167 Assuming the proportions given in the previous article, the maximum bearing pressure works out to about : — 0-785x500 , xor^nmi, =about 2000 lb. per sq. m. 0-4X0-5 Fig. 124. Type A. and it is hardly surprising that good bronze appears to be preferable to white metal for this bearing. Tig. 124. Type B. 168 DIESEL ENGINE DESIGN Miscellaneous Points of Detail. — ^Where forced lubrication is employed it is important to prevent oil from splashing on to the hot piston crown and some arrangement of baffles is very desirable. This may take the form of a plate across the mouth of the piston body or a light guard over the cranks. A piston ring at the lower end of the piston is useful in removing superfluous oil from the liner and in effecting a good distribu- tion of the film. It is generally necessary to cast two shallow recesses in the /l i-X. H Fig. 126. Fig. 125. piston crown to clear the air and exhaust valve heads at the top dead centre, and the positions of these being remote from the point of greatest tempera||ire'are usually chosen for two tapped holes to receive lifting bolts. Holes should not be drilled in the centre of the crown, and if a turning centre is necessary a special boss should be cast for this purpose and turned off afterwards (see Fig. 126). So far nothing has been said about pressure stresses and the bearing pressure on the piston body considered as a crosshead, as these appear to be irrelevant. Temperature considerations determine the proportions of the piston body, and the gudgeon RUNNING GEAR 169 bearing is made as large as the limited space allows. The guide pressure between the piston body and the liner works out at a very moderate figure, and measurements of pistons after long periods of service fail to disclose any appreciable wear and justify the conclusion that under normal running conditions the piston body floats on a ,film of oil. In cases of seizure the cause of abrasion is usually traced to local distor- FiG. 127. Type A. Fig. 127. Type B. tion, sometimes assisted by the destruction of the oil film by the presence of viscous deposits. Water Cooling. — Two systems of conveying the water to and from the piston are shewn in Fig. 127. In both systems the aim is to render the success of the scheme independent of the water tightness of the various joints involved. In type B inaccuracies of alignment are allowed for by a ball joint at the foot of the stationary tube. Pistons for Four Stroke Crosshead Engines. — These are generally made of not much greater length than is necessary to accommodate the rings, eight to ten in number. The pro- 170 DIESEL ENGINE DESIGN vision of an extra number of rings above what is considered sufficient for a trunk piston may be attributed to : — (1) The throttling effect lost by discarding the piston skirt. (2) The lower speed of revolution usually associated with engines of the crosshead type. Cooling by means of a blast of air has been used (apparently successfully) in cylinders of medium size, but water cooling is now almost universal. Fig. 128 exhibits different forms of piston having one feature in common, viz., ribbed support of the crown. There is reason to believe that these ribs are con- ducive to cracking in uncooled pistons, and therefore the legitimacy of their use in cooled pistons would appear to depend on the reliability of the coohng system employed. Assuming that failure of the cooling system is a contingency to be reckoned with, there would appear to be some measure of prudence in proportioning the piston crown to be self- supporting and transmitting the pressure to the piston rod via the piston walls. Two systems of water cooling are shewn in Pig. 129, which requires no description. Pistons for Two Stroke Crosshead Engines. — ^The existence of ports in communication with the exhaust pipe at the lower end of a two stroke cyUnder necessitates the provision of a skirt or extension of the piston to prevent the uncovering of these ports when the piston is at the top dead centre. The skirt usually takes the form of a light drum secured to the piston by a number of weU-locked studs. It is common practice to arrange one or two inwardly expanding rings at the lower end of the cylinder to prevent leakage past the skirt (see Fig. 130). If the cyUnder liner is of sufficient length these exhaust rings may be located in the skirt itself, as in Fig. 131. Such an arrangement involves a higher engine than that of Fig. 130, but facilitates conduction oiheat from the piston and inci- dentally secures a lower mean temperature for the cylinder liner. The construction of the piston proper is generally similar to that of a four cycle engine, but it must be borne in mind that the conditions as to temperature are more severe than in a four stroke engine of the same size working at the same mean indicated pressure, so that the remarks of the preceding article in reference to ribs under the crown apply with still greater force to two stroke engines. It is significant RUNNING GEAR 171 FiQ. 128. Fig. 129. Outlet 172 DIESEL ENGINE DESIGN that pistons are now being fitted by one of the leading makers of large two stroke engines in which the crown is free from ribs and the skirt is extended to do duty for the piston rod, as shewn more or less diagrammatically in Fig. 132. Fig. 130. Piston Rods. are : — Fig. 131. Fig. 132. -The forces to which a piston rod is subject (1) The pressure load of the piston, which attains a maxi- mum of about 500 lb. per sq. in. of piston area, with an occasional explosive load of about double this intensity at the top dead centre. (2) A load due to the inmia of the piston (including the water therein) and the rod itself. This also attains its maximum at the top dead centre, but is opposite in direction to the pressure load. The maximum in- tensity of the inertia load seldom exceeds 70 lb. per sq. in. of piston area in commercial engines. (3) Friction of the piston rings on the liner. This effect has its maximum at the top firing centre and acts in the RUNNING GEAR 173 same direction as the inertia so far as the expansion stroke is concerned. It has been proved that piston ring friction absorbs about 5% of the indicated power, and assuming (as seems probable) that the friction is at every point proportional to the cylinder pressure, it appears that the maximum friction is equivalent to about 10 lb. per sq. in. of piston area. (4) Fluid friction due to shearing of the lubricating oil film. The joint effect of these forces amounts then to about 420 lb. compression at firing dead centre and about 70 lb. tensile at the end of the exhaust stroke per sq. in. of piston area. Unfortunately the fatigue stress of steel between limits of this sort appears not to have been determined yet, but the value is probably in the neighbourhood of 35,000 lb. per sq. in. for 30 ton steel. A table of buckling loads for circular mild steel rods with rounded ends is given below, from which it will be seen that the question of buckling appears to be irrelevant, in view of the fact that the ratio of length to diameter of rod does not usually exceed 15, and is usually less. In this case, and in cases of connecting rods also, the use of a strut formula, such as Euler's, in conjunction with a large factor of safety, would appear to be irrational (see page 184). L -^D =Length -H Diameter Buckling stress, Ib./sq. in. L-HD=Length 4- Diameter Buckling stress, Ib./sq. in. In current practice the diameter of the piston rod is made from 0-23 to 0-3 of the cylinder bore, corresponding to a maximum compressive stress under normal working conditions. 420 420 of about ^-^-^ = 8000 lb. /sq. in. to ^^:^, = '^^00 Ib./sq. in. The higher of these figures corresponds to a factor of safety under fatigue conditions of about 35,000-f8000 = 4-35, which appears ample. The stress under an exploaon of 1000 lb. per sq. in. would be =19,000 Ib./sq. in., and the factor of {O'Zoj safety against buckling (assuming ==15)=— ^——=2 -35, which would appear to be sufficient in view of the provision of a relief valve to prevent the pressure from exceeding greatly that for which the rod is designed. 5 10 15 20 30 63,000 56,000 45,000 33,000 21,000 40 50 80 100 13,000 9,000 4,000 2,500 174 DIESEL ENGINE DESIGN The upper end, of the rod usually ends in a circular flange for carrying the piston to which it is secured by a row of studs proportioned to the inertia load, of the piston with a very moderate stress allowance, in order to give a good margin for dealing with such emergencies as seized pistons. The lower end is sometimes secured to the crosshead by means of a flange, as in Fig. 133, or in the manner indicated in Fig. 134. In either case it is reasonable to make the connection of the same strength as that between the rod and the piston, assuming that ample provision has been made for the contingencies referred to. Grossheads and Guides. — For marine engines the slipper guide, shewn in Fig. 135, is the favourite. The bearing surface is usually made equal to the piston area,^ and the maximum bearing pressure with a con- necting rod 4-5 cranks long then has a value of about 55 lb. per sq. in. The slipper itself is of cast steel, white- metaUed on ahead and astern faces. The studs securing the slipper to the gudgeon block must be adequate to carry the maximum guide pressure when running astern. The area of the gudgeon bearing is based on a bearing pressure of about 1500 lb. per sq. in. The ahead guide face is of cast iron provided with water cooling. The aa|bern bars are frequently of forged steel, secured by fitting bolts. The stress Ta the latter is usually very moderate, as stiffness is the chief con- sideration. The proportions given in Fig. 135 are of course approximate only and subject to modification to suit different conditions. The type of guide block indicated on Fig. 132 is well known Fig. 133. Fig. 134. ^ i.e. the sectional area of the cylinder. RUNNING GEAR 175 in connection with paddle-steamers and locomotives, and needs no further description here. For land engines double semicircular guides are sometimes used, particularly when the cylinder and frame are cast in one piece. In general, the cross- Width of crosshead face, '75 B. Depth „ „ 110 B. Thickness „ plate, 010 B. Fig. 135. heads and guides used in Diesel Engine construction differ but little from those commonly fitted to steam engines and large gas engines, the most important point of difference being the gudgeon pin and its bearing, which require to be liberally dimensioned to withstand the high maximum pressures to which they are subject. It is also desirable to provide means to prevent carbonised oil from the cylinder from reaching the guide surface. Guide Pressure Diagrams, — A diagram shewing approximately the guide pressure at any crank angle is very simply obtained from the twisting moment curve in the manner described below, with reference to Fig. 136. P= Piston load in lb. R= Guide reaction in lb. T= Twisting moment in in. lb. S= Height of gudgeon pin centre above the centre of the crank-shaft. j'ie_ i3g_ 176 DIESEL ENGINE DESIGN Now R=-^ But T=P.k T Therefore R=-Fr b The rule is therefore : Divide the turning moment at any instant by the distance from the gudgeon pin centre to the crank-shaft centre, and the result is the guide reaction at the same instant. It would appear that the guide reaction and the twisting moment should change sign simultaneously. This is not quite the case, for the following reason : — The twisting moment curve contains an inertia element in which an approximation is obtained by dividing the mass of the connecting rod in a certain proportion between the revolv- ing and reciprocating parts. This approximation, though good so far as vertical forces are concerned, gives very inaccurate values for horizontal forces. Also the centrifugal effect of the revolving parts of the rod influences the guide reaction but not the twisting moment. These discrepancies are of very small importance with the piston speeds at present obtaining. For a full discussion of the influence of the connecting rod inertia forces on the guide reaction, the reader is referred to Dalby's " Balancing of Engines." A table of values of ^ where " 1 " = the length of the connecting rod, is given below for various crank angles, assuming a rod 4-5 cranks long. Crank angle 0° 20° 40° 60° 80° 100° 120° 140° 160° 180° Values of |- 1-22 1-21 M6 1-09 1-02 0-94 0-87 0-82 0-79 0-78 Connecting Rods. — The material for connecting rods is generally Siemens-Martin steel of the same quality as that' used for the crank-shaft, ^ampings are sometimes used for small engines, and if large quantities are made at a time this is an economical way of producing a rod of " H " section, if machining all over is not considered essential. Cast steel has been used on the Continent for gas engine connecting rods, but the author has not met with this practice in Diesel Engine construction, either British or continental. Connecting Rod Bodies. — The section of the body or shaft of the rod is generally circular, or part circular, with flattened RUNNING GEAR 177 sides. The latter section is slightly lighter for a given strength, but involves an extra machining operation. For extreme lightness an " H " section of rod milled from the solid (like a locomotive rod) would appear to be advantageous. The body usually tapers gently from the big to the small end, and it will be shewn later that this practice is a rational one from con- siderations of strength. It is interesting to note that the engineers of the North-East Coast, in their recent specification for standard reciprocating marine steam engines, recommend parallel rods, and it seems possible that the extra cost of Fig. 137. material involved is discounted by the reduced cost of forging and machining. Big Ends. — Fig. 137 shews two forms of big end, type A being the cheapest and that most commonly used. Type B is the strongest, but suffers from the disadvantage of providing no facility for adjusting the compression by means of liners. Returning to type A, the " brasses " are usually of cast steel lined with white metal. With a stronger section, as shewn in Fig. 138, cast iron may be used instead of cast steel, with satisfactory results, but the practice is uncommon. The bolts are frequently reduced to the core diameter between fitting lengths, as shewn in Fig. 137 ; but it appears that full diameter bolts are stronger under the conditions to which they are 178 DIESEL ENGINE DESIGN subject in trunk piston engines. In addition to tensile stresses the big end bolts have to resist shearing forces between the two brasses and also between the crown brass and the palm end of the rod. Partial relief of this duty is afforded by the following means, one or more of which are generally used in good designs : — (1) Spigoting the two brasses into each other, as in Fig. 139. This is very rarely done. (2) Spigoting the crown brass into the palm of the rod, as in Fig. 137, Type A. (3) Providing fitting rings, half in the palm and half in the crown brass at the bolt holes (Fig. 140). Fio. 138. Fig. 139. Fig. 140. Small Ends. — Various types of small end for trunk engine rods are shewn in Fig. 141. With type A the chief difficulty is to find room for bolts of adequate strength. If a big end of type B, Fig. 137, be fitted, it becomes necessary to maWe provision at the small end for adjusting the compression, as in Fig. 141, types B and D. Type C combines strength and adjustability of the bearing itself, but makes no provision for altering the compression. Type E contains a solid bush, which must be replaced when worn, and is therefore only suitable for small engines, and adequate section of metal round the bush must be provided to prevent the hole becoming enlarged (see approximate propor- RUNNING GEAR 179 tions on Fig. 141). The brasses are usually of phosphor bronze. The forked end of a marine connecting rod is shewn in Fig. 142, on which approximate proportions are noted in terms of the cylinder bore. It differs from the similar member of a marine steam engine chiefly in the follow- ing points : — • (1) The cap brass is not provided with a steel keep. (2) The fork gap is re- latively narrower, owing to the piston- rod nut having to deal with inertia only. (3) The brasses are of cast steel, instead of gun- metal. Points of Detail. — The lubrication of the big and smaU ends has been referred to under crank-shafts and pistons. Where forced lubri- cation is used an oil-hole is generally drilled in the rod to conduct the oil from the big to the small end in the case of high speed engines. External pipes are used for this purpose in large, slow- running engines, but are unsuitable for high speeds owing to the tendency of the joints to work loose. When air compressors, oil pumps, or other gear are worked by links from the connect- ing rod the connection to the latter should be made near the top end, as in Fig. 143, so that the strength of the rod to resist buckling is not impaired. 180 DIESEL ENGINE DESIGN Strength of Connecting Rods. — The forces acting on the connecting rod are : — (a) The joint effect of (1) The pressure load on the piston, (2) The inertia of the piston and crosshead, (3) The piston-ring friction, (4) The lubricated friction of piston and crosshead, all divided by the cosine of the angle of obliquity of the rod. (b) The longitudinal component of the inertia of the rod itself. (c) The transverse component of the inertia of the rod itself. (d) The friction of the top and bottom end bearings. Fig. 142. When considering the compressive stress of the rod on the expansion stroke one is on the safe side in neglecting item (4). The tensile forces attain their maximum at the top dead centre following the exhaust stroke, and the reciprocating parts being then at their position of minimum speed, item (4) may prob- ably be neglected with safety. Item (b) is estimated with sufficient accuracy by the usual procedure of dividing the mass of the rod between the recipro- RUNNING GEAR 181 eating and revolving parts in that ratio in which the centre of gravity of the rod divides its line of centres. Item (c) gives rise to a bending moment, the maximum value of which is given approximately by the formula : — F^ VlOO/ ,R.L2 -(1) 26d Where f =Bending stress. n= Revolutions per minute. R=Crank radius in inches. L=Length of connecting rod in inches. d=Diameter of rod in inches (mean). Fig. 143. Fig. 144. The sign of the bending moment is such that the latter always tends to bend the rod outwards to the pide on which the crank stands. The variation in the magnitude of the stress over the length of the rod for various positions of the crank relative to the top dead centre is shewn in Fig. 144 for a rod five cranks long. The stress varies as sin {d+(f>), where 0=the crank angle relative to top dead centre and =the angle of obliquity of the rod, and as Lx— y-, where L is the length of the rod and x is the distance from the small end to the section under consideration. The assumptions made use of in equation (1) are that the rod is of uniform section, and, as is usually assumed in books on appUed mechanics and machine design, that the influence of the rod ends is small. 182 DIESEL ENGINE DESIGN Item (d) may be estimated, on the assumption that the co- efficient of friction attains Morin's value of 0-15 for slightly greasy metal at the top or bottom end or at both ends simul- taneously. The effect of journal friction is to divert the line of thrust from the centre line of the rod, and the amount of this deviation is found by the well-known graphical construc- tion shewn in Pig. 145, in which the line of thrust is shewn to be tangential to two very small circles whose radii are Fig. U5. Fig. 146. equal respectively to the radii of the crank-pin and gudgeon- pin multiplied by the coefficient of friction. The deviation at any point of the rod of the liift of thrust from the line of centres will be denoted " d." The effect of this deviation is to bend the rod into an " S " shape, as shewn much exaggerated in Fig. 146. This form of failure is one consistent with the Eulerian theory of pure buckling, but usually regarded as an improbable solution. The author has actually seen one instance of a Diesel Engine rod failing in this way, and the effect was probably due to the RUNNING GEAR 183 causes indicated above, arising in acute form. To be on the safe side, it seems advisable to consider two cases : — (1) Coefficient of friction negligible at the gudgeon-pin and equal to 0-15 at the crank-pin. (2) Coefficient of friction negligible at the crank-pin and equal to 0-15 at the gudgeon-pin. The weak sections are clearly those (cf. Fig. 149, KK and SS) where the big and small ends merge into the shaft with a radius, and incidentally those selected by the draftsman when Fig. 147. FiQ. 148. giving dimensions for the diameters of the rod. The reason- ableness of making the rod tapered will now be apparent, as the bending effect is clearly greater at the large end of the rod on account of the crank-pin being of greater diameteir than the gudgeon-pin. In all probability these considerations had nothing to do with the practice, but illustrate a fact that has a very important bearing on the part of machine design, viz., that a construction which appears wrong to the eye of an individual gifted with a sense of form will usually, on investiga- tion be found unsound in principle, and vice versa. This theme is one that might profitably be made the subject-matter 184 DIESEL ENGINE DESIGN of a less specialised book tan this, but it may be worth mentioning here that the conscious or unconscious recognition of this principle appears to be responsible in part for the ascendancy of continental constructors in certain branches of mechanical design. Returning to Fig. 146, it is evident that the deflection at any point of the rod should, to be strictly accurate, be added to the deviation of the line of thrust at that point, in order to find the bending moment, and further, this new bending moment involves the construction of a revised deflection curve, and so on. This evidently calls for some form of mathematical treatment, which with certain approximations can readily be applied. It will be found, however, that the deflections involved are small compared with the deviation of the line of thrust, and whatever error may be incurred can be considered to be covered by the factor of safety. On these assumptions, " d " f or the weak sections KK and SS is given by the following : — Case (1), Fig. 147 . . d=0-15R<,. t^ (2) Case (2), Fig. 148 . . d=0-15Rg •?— ^ (3) And f=pQ+d)__(4) Where P=The thrust in the rod in lb. A=Sectional area of rod in sq. in. Z= Sectional modulus of rod in in. ^ f= Maximum compression stress in lb. per sq. in. Ro= Radius of crank-pin. ^g= >, „ gudgeon-pin. It will be noticed that the strut formulae of Euler, Gordon and Rankine and others have not been utilised above. It appears to the writer that these formfilse are irrelevant to the case of Diesel Engine connecting rods, for the following reasons : — (1) Euler 's formula is based on the calculation of the load required to produce elastic instability, and with short rods the stress commonly works out at a higher value than the ultimate strength. (2) The Gordon and Rankine formulse are based on experi- mental values of the buckling stress under static RUNNING GEAR 185 conditions, and give no indication of the strength under repetitions of stresses, which are generally only a fraction of the buckling load. It seems more rational, therefore, to calculate the maximum direct stresses as closely as possible and to apply to the approxi- mately known fatigue stress of steel a factor of safety of 2-5 to 3, which is known to be satisfactory in other cases. In view of occasional abnormal pressures^of about 1000 lb. per sq. in., it is interesting to see what factor of safety a given rod has for meeting such contingencies, and the table of buck- ling stresses given on page 173 may be used for this purpose. Example of Stress Calculation for Connecting Rod. — Four Stroke Cycle. Bore of cylinder . . 24 in. Stroke 30 in. Revolutions per minute 200 Weight of piston (trunk) 2200 lb. ,, connecting rod complete . . . 25001b. Main dimensions of rod, as in Fig. 149, under : — (1) Calculation of stress due to thrust 30° after firing centre. Piston pressure load=0-785 x 24^ x 500 = Inertia load 30° after dead centre Fig. 149. = 226,000 lb. K 277 X 200V 60 -; Xl5x 22004-0-35x2500 X(cos30°+icos60°) 386 = 51,000 lb. Resultant vertical force=226,000-51,000 = 175,000lb. At 30° after dead centre the obliquity of the rod is 6°. Connecting rod thrust = 175,000 -^cos 6° = 176,0001b.=P. At section KK : — Area = 33 -2 in. "= A Section modulus =27-0 in. '=Z 186 DIESEL ENGINE DESIGN Deviation of line of thrust = =0-52=d 75 and the stress f =^176,000 (5^+?^) = 8970 lb. /sq. in. At section SS : — Area = 38-5in.2=A Section modulus =33 -7 in.^=Z ■P, ... .,. ,,, , 0-15x6-25x55 -„_. Deviation 01 line of thrust = =0'69 in. 75 and the stress f = 1 76,000 (-^ +^^) = 176,000 x -0464 = 81601b./in.2 (2) Calculation of stress due to inertia bending at 30° after dead centre. Maximum inertia stress in rod, from equation (1), 2^x15x752 ,.--„ ,. 2 = 26x6-75 =1920 lb. /in.^ From Fig. 145 the fraction of this maximum applying to postion KK at 30° after dead centre is 0-37, .'. Inertia bending stress at section KK=0-37xl920 = 710lb./in.2 The fraction applying to section SS at 30°, after dead centre is 0-49. .". Inertia bending stress at section SS=0-49xl920 = 940 1b./in.2 (3) Resultant stress at KK=8970-710 = 8260 Ib./in.^ Since the bending actions due to inertia and eccentricity of thrust are of opposite sign. Resultant stress at SS = 8160+940=9100 Ib./in.^ since the two bending actions are of the same sign. (4) Tensile stress at SS at Jaeginning of suction stroke. Inertia force = (?^)\ 15 x?^5^±|f2i?^^ 1-2 = 63,200 lb. Stress at SS=63,20o(-i- +|^)=2930 Ib./in.^ The total range of stress is therefore 9100 + 2930 = 12,030 lb. /in.2 RUNNING GEAR 187 The range of stress required to produce fracture of mild steel by fatigue appears to be about 35,000 lb. /sq. in., so the factor of safety is about 3. Calculation on the above lines might with advantage be made for several different positions of the crank. It is evident that the results of the calculation depend very largely on the assumed conditions of journal friction, but it should be borne in mind that almost any possible combination of unfavourable conditions is a probable contingency in the combined lives of a number of similar engines. Proportions Found in Practice. — In the preceding example the mean diameter of the rod is approximately 0-28 of the diameter of the cylinder, a very favourite ratio in practice. In different designs this ratio varies from 0-26 to 0-30, and it is a rather curious fact that the two extreme figures are those which appear to be used by two of the leading makers in their respective practices. The maximum and minimum diameters are usually about 5% more and less than the mean. Connecting Rod Bolts. — In four stroke engines these are usually proportioned to the maximum inertia load with a nominal stress of 6000 to 8000 Ib./in.^ based on the inertia and centrifugal loads divided by the area of two bolts at the bottom of the threads. With trunk piston engines failure when it occurs is generally due to piston seizure, to which it would be difficult to apply definite rules of calculation. Danger of seizure is largely eliminated by the use of a cross- head. The strength of connecting rod bolts for four stroke Diesel Engines forms the subject-matter of a paper by Mr. P. H. Smith, read by him before the Diesel Users' Association and containing the results of several years' experience. For the big end it appears that the bolts seldom fail if made of a diameter 12 to 13% of the cylinder bore. For the small end, if bolts are used at all, the only safe rule is to make the bolts as large as the space available will allow. Mr. Smith also points out that the bolts, for both big and small ends, are not equally stressed, as may easily be seen by reference to Fig. 150. Owing to the deviation of the line of pull from the centre line of the rod, that bolt (No. 1 in the Fig.) which first passes the top dead centre at the beginning of the suction stroke is nearer the line of pull than the other bolt, and consequently more highly stressed. 188 DIESEL ENGINE DESIGN If Then P= Resultant pull in lb. S= Centres of bolts in in. d=0-15 radius of crank-pin. Pull in bolt No. 1 = Px - + ci Px Pull in bolt No. 2 = S With crosshead engines the small end bolts have to carry the inertia load due to piston, piston rod and crosshead, and also any frictional forces acting on the piston and crosshead. The Fig. 150. latter being more or less indeterminate, it is customary to allow a nominal stress on these bolts about 30% less than that allowed for the big end bolts. The bolts or studs connecting the crosshead to the piston rod and the latter to the piston are given a large margin of strength for the same reason. Connect- ing rods for two stroke engines are not as a rule distinguishable from those for four stroke engines, as the possibility of com- pression being lost has always to be kept in view. Indicating Gear. — ^The only satisfactory gear for obtaining accurate cards consists of a link motion directly connected to the piston. The usual arrangement is shewn diagrammatically in Fig. 151 for both trunk and crosshead engines. The condi- tions for giving an accurate r%roduction of the motion of the piston are :■ — (1) The line of the cord to be at right angles to the mean position of the short arm of the lever. The long arm of the lever to make equal angles of swing above and below the horizontal. The versine of the arc of swing of the drag link to be negligible in comparison with the stroke of the engine. (2) (3) RUNNING GEAR 189 The mechanical details of indicator gears are hardly of sufficient interest to require description here, but it may be well to mention that indicating is of far more importance in the successful running of a Diesel Engine than in that of a steam or even a gas engine, and consequently all the more care should be given to the design of the gear by which the indicat- ing is accomplished. Makeshift or temporary gear should not be tolerated, but the same attention paid to lubrication and bushing of joints, etc., as on other parts of the engine. Literature. — " Piston Cooling for Diesel Engines." Article in The Motor Ship and Motor Boat, July 18th, 1918, et seq. Trunl< Engine. Crosshelad Engine. Fig. 151. CHAPTER X FTJEL OIL SYSTEM For purposes of description the complete fuel oil system is conveniently divided, into two parts, the first consisting of those elements, such as tanks, etc., which are external to the engine, and the second of those organs of the engine itself which are directly concerned with the delivery of the fuel to the working cylinder. External Fuel Oil System. — Pig. 152 represents diagram- matically a fuel system for a small Diesel power station and Fig. 152. consists essentially of a main storage tank A, a ready-use tank B, a filter C, and a pump D, fm raising the oil from the storage tank. The storage tank is preferably arranged underground, as close as possible to the railway siding, so that oil can be run from the railway tank waggon to the storage tank by gravity, through a hose pipe. Some form of level indicator or a plugged hole for a sounding rod should be provided. The capacity of the tank will depend on the size of the station and the local conditions of supply. 190 FUEL OIL SYSTEM 191 The pump D, by means of which the oil is pumped, to the ready-use tank, may be of the semi-rotary type, capable of being worked by one man in the case of small stations ; but where the daily demand is greater, a motor -driven rotary or reciprocating pump is generally fitted. The ready-use tank may have a capacity of say half a day's run, so that the routine of replenishing it will occur twice daily. Some form of float indicator should be fitted, so that the level of oil may be conveniently ascertained from the engine room floor. Other necessary fittings are an overflow pipe leading to the storage, or a special drain tank, and a drain valve communi- cating with the overflow pipe. The tank must be closed at the top to exclude dirt. The valves in connection with the fuel system are preferably of the gate or sluice type, as cocks are liable to leakage and globe valves tend to choke by accumula- tion of sediment. The tank only requires to be located a few feet above the level of the filter as the discharge is very small. The filter usually consists of a cylindrical tank of about 40 gallons capacity, located about two feet above the level of the cylinder cover and provided with a filter diaphragm at about a third of the height of the tank from the bottom. The diaphragm consists of a sheet of felt sandwiched between two sheets of wire gauze and reinforced by an angle iron ring. The fuel enters the filter at the bottom, passes through the diaphragm by virtue of its static head, and is drawn off by the engine fuel pump at a point a few inches above the diaphragm. The filter vessel is prevented from being overflooded by a ball float mechanism which closes the inlet cock when the oil reaches a predetermined level. The plug of this cock is kept fairly tight by means of a spring acting on the plug, but slight leakage is almost inevitable, so it is desirable to mount the filter on an oil-tight tray provided with a drain. It is very usual to provide a small reservoir of the same capacity as the filter arranged alongside the latter for the reception of paraffin, by means of which the piping leading to the engine, and also the fuel pumps and fuel valves, etc., may be cleansed from time to time by running the engine for a few minutes on this fuel before stopping the engine. Marine installations follow on similar lines with a few com- plications. The double bottom is used as a storage tank, and the fuel is raised to the ready -use tank by motor-driven pumps, when electric power from auxiliary engines is continuously 192 DIESEL ENGINE DESIGN available or by means of pumps driven ofi the main engine in cases where the main engine drives its own auxiliaries. In either case it is usual to install duplicate pumps to guard against breakdown. The motion of ships being unfavourable to the successful operation of float devices, the level of the oil in the ready-use tank has to be inferred from gauge glasses, test cocks and the like. For similar reasons, the filters must be totally enclosed and provided in duplicate with change-over cocks, so that they may be overhauled at any time. In addition, special requirements of the Board' of Trade and Lloyd's have usually to be complied with. Fuel System on the Engine itself. — The commoner arrange- ments fall into one of two broad classes : — (1) Those in which each cylinder has a separate fuel pump or separate plunger and set of valves to itself. In this case the oil is delivered direct from the pump to the injection valve by the most direct route possible. (2) Those in which one fuel pump plunger supplies the oil for a pluraUty of cylinders, usually a maximum of four. In this case the pump dehvers to a fuel main provided with a branch and distributing valve separate to each cylinder, whereby the amount of oil delivered to each cylinder may be equalised while the engine is running. Engines of six or eight cylinders are divided into two groups of three or four respectively, so far as the oil system is concerned. For marine engines the first system is at present the favourite, and has the advantage that the failure of one pump does not affect the working of the remaining cylinders. With the second system a stand-by pump is provided, ready to take over duty at a moment's notice. For land engines the two systems appear to be on an equality as nearly as can be ascertained by the reputations of repre- sentatives of both classes, but if anything system (1) appears to be slightly the more popular of the two. Figs. 153 and 154 illustrate the two systems diagram- matically. It will be noticed that in Fig. 153 the governor operates on all the pumps by means of a shaft extending nearly the whole length of the engine, and as the quality of the governing is dependent on the freedom from friction of the governing mechanism it is desirable to mount this shaft on ball bearings. The expense of providing separate pump bodies FUEL OIL SYSTEM 193 and, drives is sometimes reduced by grouping the pumps in the neighbourhood of the governor, even to the extent of driving all the plungers by a common eccentric. With the arrangement of Fig. 154 there is only one pump to regulate, and this renders possible the use of a type of governor Yue/ Inlet Pipe Fig. 153. which is probably unrivalled for sensibility and which will be described later. The distributors indicated in Fig. 154 are a special feature of this system and are illustrated to a larger scale in Fig. 155. A particularly neat arrangement of piping is obtained by combining the fuel distributor and blast air T- piece in one fitting. GovernorW Fig. 154. The inclusion of a non-return valve prevents in a great measure the oil being forced back through the pump by the blast air pressure in the interval elapsing between the turning on of the blast air and the attainment by the engine of full working speed. A non-return valve is sometimes fitted to the 194 DIESEL ENGINE DESIGN fuel valve itself for the same reason. Vent cocks are provided on the distributors, and sometimes on the fuel valves, to en- able the pipes to be primed before starting the engine. The priming may be effected in various ways : — (1) By gravity, means being provided for holding the fuel pump valves off their seats during the process. (2) By means of an auxiliary hand-operated and spring- returned plunger on the fuel pump. (3) By means of the fuel pump plunger itself, where provi- sion has been made for disconnecting the latter from its operating eccentric in order to enable it to be oper- ated by a hand lever provided for the purpose. To,Fuel Vshe To Fuel I/aha FueJ Brass Copoer Via. 156. The piping in connection with the high pressure fuel system deserves special attention, on account of the high pressures used, and the type of uniorrshewn in Fig. 156 is probably the most satisfactory that has yet been devised both for oil and high pressure air. Fuel Pumps. — ^A simple fuel pump for a large marine engine is shewn in Fig. 157, and is representative of a large class of pumps for both marine and stationary purposes. The operation of the pump is almost obvious from the figure, but may be described briefly as follows : — FUEL OIL SYSTEM 19S The eccentric A works the plunger B, which is guided, at C. E and E are the delivery and suction valves respectively, and the latter communicates with the suction chamber D, to which the fuel is led by means of a pipe not shewn in the figure. M is an auxiliary plunger operated from the crosshead C by links I, H, K, etc., and whose function is to keep the suction valve off its seat for a frac- tion of the delivery stroke, depending upon how much oil is required per stroke. The duration of this inopera- tive portion of the stroke is altered as required by raising or depressing the point K by means of an eccentric keyed to the shaft L, according as less or more oil is required. In the case of a governed engine the shaft L is con- trolled by the governor. On marine engines the shaft L is operated by hand gear, consisting of levers, rods, etc. Neglecting the obliquity of the eccentric rod the' main plunger describes simple har- monic motion of amplitude equal to half the stroke of the eccentric, and it will be clear from the drawing that the auxiliary plunger M will describe a similar motion exactly in phase with the first but of amplitude equal to stroke of main plunger x KH KI Fio. 157. When the main plunger is at the bottom of its stroke the auxiliary plunger is also at the lowest point of its travel, and the clearance between the top of the auxiliary plunger and the TCT suction valve multiplied by the ratio =5™- is equal to the effec- 196 DIESEL ENGINE DESIGN tive stroke of the pump, that is that portion of the strok< during which the suction valve is on its seat, as of course r must be (apart from viscosity effects) for delivery to take place The quantity of oil delivered per stroke therefore depends oi a certain clearance between the auxiliary plunger and th( suction valve, which clearanc( is readily adjusted by shorten ing or lengthening the rod LE when assembling or adjusting the engine and in the ordinarj course of running by the eccen trie at L. Constructional Details .— Th« pump body, plunger sleeve anc guide are of cast iron. Th« main and auxiliary plungers the crosshead pin and jointf in the linkwork are of case hardened steel. The valves may be either of steel or cas1 iron. If the latter, then the suction valve should be fittec with a hardened steel thimbh where it makes contact wit! the auxiliary plunger. Th« main eccentric and strap ma^ be of cast iron, the lower hal of strap being white-metaUec in some cases. It will b( noticed that no packing is pro vided for the main plunger but reliance has been placec on the fit of the plunger. Wit! good workmanship the leakage should not be excessive.* A Fig. 158. cast tray is provided to catch drips during working and the overflow at priming. A light sleeve encircling the auxUiar5 plunger is arranged for operation by external gear so that the suction valve may be lifted by an emergency governor in cases of excessive speed, and also by hand in case it is desired to cu1 any individual cylinder out of operation. Variations of this system, embodying the same principle * Except with tar oil, for which this arrangement is unsuitable. FUEL OIL SYSTEM 197 are shewn in Figs. 158 and, 159. The front view of the latter shews three pumps grouped together, but each worked by its own eccentric. Fig. 160 shews four plungers being operated by eccentrics in common. It is evident that with this arrange- ment the oil delivered to the pulverisers of the various cylinders wUl have different allowances of time in which to settle before injection into the cylinder. This appears to have no effect on the efficiency, but it is usual to space the eccentrics so that oil is not in process of delivery whilst a fuel valve is open. The pumps so far illustrated have been driven off the cam-shaft. That shewn in Fig. 161 is arranged with horizontal plungers CamshaFt Hand Pump Fig. 159. 198 DIESEL ENGINE DESIGN for driving off a vertical shaft. The auxiliary plunger is driven by a separate eccentric which on account of the intermediate lever L requires to be at 180 degrees or thereabouts to the main eccentric. The suction valve control may in this case be effected, in one of two ways. (1) By an eccentric movement of the lever L. (2) By advancing or retarding the auxiliary eccentric. " III ForGoyernor Control Fig. 160. The latter leads to a very neat and efficient arrangement of governor and fuel pump, to which reference has already been made. It wiU be immediately obvious that with a given maximum clearance betw^fci the suction valve and the auxiliary plunger, an angular movement of the auxiliary eccentric will have the effect of advancing or retarding the instant at which the suction valve comes on its seat, and con- sequently increasing or decreasing the amount of oil delivered per stroke. This angular movement is effected very simply by a type of governor which has been well known for a long time, in steam practice, and which is illustrated in Fig. 170. FUEL OIL SYSTEM 199 200 DIESEL ENGINE DESIGN Returning to Fig. 161, the use of this type of fuel pump is almost entirely confined to land, engines. The provision of an eccentric mounting for lever L enables the pump to be set in three different positions, apart from its normal running position, viz. : — (1) "Starting." In this position the lever is moved so that the maximum clearance under the suction valve is increased about 50%, so that the delivery of oil per stroke is increased correspondingly. (2) " Stop." In this position the suction valve is held continuously off its seat and no oil is delivered. (3) " Priming." In this position both suction and delivery valves are held off their seats and the oil has a clear passage through the pump. Figs. 163 and 164 wiU make this matter clear without further explanation. Details of Fuel Pumps. — ^The bodies are usually of cast iron, but solid slabs of steel are sometimes used. In designing the body three considerations should be kept in view : — (1) The shape to be favourable to sound casting. (2) As few machining operations as possible to be necessary apart from those which can be done on a drilling machine. (3) The pump chamber and passages to be free from air- locks. Owing to the costly precautions necessary to ensure the plunger and guide being concentric and in line it is convenient to allow some side play at the point where they join, as in Fig. 165. Some different forms of plunger packing are shewn in Fig. 166 and a selection of suction and delivery valves in Fig. 167. Fig. 168 shiews a hand-operated plunger for priming purposes. Calculations for Fuel Pumps. — The process of computing the capacity of a fuel pump for a proposed engine is most easily illustrated by an example, as follows : — B.H.P. of one cylinder (four stroke), 260. One plunger to each cylinder. Estimated fuel consumption, 0-4 lb. per B.H.P. hour. Revolutions per minute, 120. FUEL OIL SYSTEM 201 202 DIESEL ENGINE DESIGN Fig. 165. I i m ■h Fig. 166. 4 4 tIT mTm r Fie. 167. FUEL OIL SYSTEM 203 Estimated, quantity of fuel per cycle =-^7r — —-=0-0278 lb. Volume occupied, by 1 lb. of fuel=about 31 cub. in. Therefore volume of fuel per cycle=0-0278 x 31 =0-86 in. » Adding 50% to allow for overload, possible increase of fuel consumption, leakage of plunger, etc. : — Stroke volume of plunger = l-5x0-86 = l-29 in.^ Which is satisfied by a plunger diameter of | in. x 3 in. stroke. Fig. 168. This size of plunger would only be permissible on a marine engine. If the cylinder belonged to a governed engine the stroke volume of the fuel pump plunger would need to be about four times the above figure, as it is found that good governing at aU loads is only to be obtained by using about the last quarter of the stroke. This is probably due to the fact that the quantity of fuel consumed by the engine in a given time is not proportional to the load but more nearly propor- tional to the load plus a constant representing the engine friction. The actual position taken up by the governor and the effective stroke of the pump plunger at any specified pro- portion of full load are not easy to determine experimentally with great accuracy, but the angular positions indicated in Fig. 169, with reference to the fuel pump eccentric circle, are those generally used as the basis of calculation. When one plunger is used to supply several cylinders the length of effective discharge period is limited by the condition that the latter should not overlap the injection periods. In estimating the capacity of a fuel pump driven off a vertical shaft the speed of the latter must be kept in mind, being usually the same speed as the engine, and in some cases 50% more. The valves, hand plungers, etc., are suitable subjects for distributive standardisation. For example, a suction valve f in. in diameter would be quite suitable for all sizes of cylinder (assuming one plunger per cylinder) up to about 20 in. bore provided that the use of fuels of exceptional viscosity were not 204 DIESEL ENGINE DESIGN contemplated. For oils like crude Mexican, of the consistency (when cold) of tar, larger valves are probably advisable. With the valve arrangements in common use the diameter of the delivery valve is determined by that of the head of the suction valve plus adequate clearance for the flow of the oil round the latter. The general thickness of metal of the pump body is usually kept as uniform as possible to facilitate casting, and the nominal stress in the neighbourhood of the pump chamber based on a blast pres- sure of 1000 lb. per sq. in. is about 2500 lb. per sq. in. The design of a fuel pump affords ample scope for a draftsman's skill in many directions, in which numerical calculation plays a very small part, and the following suggestions are offered : — (1) The arrangement generally to be neat and substantial and presenting an external appearance in keeping with its surroundings. (2) The flanges and brackets by which it is secured to the framework of the engine to be unobtrusive and to have the appearance of growing as naturally as possible out of the main body of the casing, so as to convey an impression of rigidity and equilibrium. (3) Every detail to be carefully studied, both with regard to its special function and, also to economy in manu- facture, efficiency always having precedence over economy. In particular, case-hardoning and bushing of parts subject to wear must not be stinted, and provision should be made for lubrication of aU moving parts. (4) Valves and other internal mechanism to be easily accessible for inspection and overhaul. (5) Arrangements to be made to catch all drips, both of fuel and lubricating oil, avoiding the use of trumpery sheet-iron guards and the like. FUEL OIL SYSTEM 205 Fig. 170. 206 DIESEL ENGINE DESIGN Many of the above principles apply of course to the design of any part of any high-grade machine, and they are mentioned here because the matter on hand provides an excellent oppor- tunity of emphasizing them in a particular case, in which the subject is singularly free from the complications arising from calculations. When the discussion is transferred to some large part of a machine, in which the stresses are approximately determinate and the scope of the design appears to be limited by adjacent parts, it becomes increasingly difficult to reconcile the ideals of high-class design with the requirements of effi- ciency and economy and the skill of a designer may be gauged by the extent to which this difficulty is overcome. From this point of view no part of a design can be said to be finally deter- mined until the whole design is complete, as there is always the possibility that a design for a certain part, perfect in itself, may require to be modified subsequently on account of its relationship, perhaps remote, to some other part as yet un- determined. Governors. — It is not proposed to deal here with governor design generally, as that is a subject for a specialist in this particular department of mechanical design, but only to illustrate the application of governors to stationary Diesel Engines by means of a few examples, and to give the main lines of calculation for the type of governor shewn in Fig. 170, which is a type not usually standardised by governor specialists. The action of the weights in causing rotation of the central sleeve will be immediately obvious from the figure. The amount of this rotation between the limits of no load and full load should correspond with the angle ~ 60° of Fig. 169, but as a safety pre- caution it is advisable to give the governor sufficient range to give a complete cut-out, and the sleeve should therefore be free to describe an angle of about 70°. The first stage in the design of the governor is to rough out a drawing similar to Fig. 170, fulfilling all the reauirements as to space, accessi- bility, etc., and in which the aloove angular rotation is secured. As regards the size of the governor, it is generally wise to avail oneself of all the space obtainable. The next step is to find the mass and centre of gravity of the weights and the positions of the latter in the extreme in-and-out positions. A diagram similar to Fig. 171 should now be constructed, in which the abscissae are distances of the weight from the centre of the shaft in inches and the ordinates are the centrifugal forces at FUEL OIL SYSTEM 207 these distances at no load speed and full load speed respectively. Point " p " corresponds to "no load " speed and " no load " distance from centre, and point " Q " the same quantities for full load. The line PQ then determines the properties which the controlling spring would have to possess if it were con- nected to the weight at its centre of gravity. These properties are as follows : — (1) The initial tension, when the weights are full in, is equal to the centrifugal force corresponding to the point "Q." (2) The weight of the spring per inch extension is equal to the slope of the line PQ, that is the amount in lb. by which the ordinate increases as the abscissa increases by one inch. Actually the spring is attached to the weight at a point nearer to the fulcrum than the centre of gravity, and both the initial tension and the rate as found thus require to be increased in the ratio t-, where : — k=Distance of the weight fulcrum from the line joiniag the C.G. of the weight to the centre of the governor. 1 =Distance of the weight fulcrum from the line of action of the spring. Radius of C.G. oF Goifernor Weight Fig. 171. This very simple construction, repeated as often as may be necessary in the process of trial and error, contains all the dynamical calculation required to ensure sensibility and stabihty, but it is advisable to provide adjustments for spring tension, in the manner shewn in the figure, to allow for un- 208 DIESEL ENGINE DESIGN avoidable errors and routine adjustment on the test-bed. Strictly speaking, the diagram shewn in Fig. 171 should be corrected to allow for the versed sine of the arcs described by the points of suspension, and so on ; but these are practically negligible. Other types of governor are designed on similar lines, but are usually complicated by Unk mechanism, of which the variations of configuration are not negligible. Fig. 172. Fig. 173. Variation of speed during the running of the engine is readily seciu-ed by transferring a ^rt of the controlling force to an auxiliary spring, the tension of which can be varied by mechanism provided for the purpose, as shewn in Fig. 172, or by varying the tension of the main spring itself, as in Fig. 173. Some points to be observed in governor design are : — (1) Weights as heavy as possible, to give power and con- sequently render the effects of friction negligible. FUEL OIL SYSTEM 209 (2) Springs to be readily adjustable. (3) Small pin-links, etc., to be as substantial as conveniently possible. (4) Friction to be reduced to a minimum by ball-bearings. (5) Joints other than ball-bearings to be bushed and pro- vided with weU-hardened pins. (6) Lubrication, both as regards supply of lubricant to the working parts and systematic disposal of the surplus, to be considered carefully. The general disposition of the governor and fuel pump with respect to the framework of the engine is shewn in Figs. 174, Fio. 174. Fig. 175. 175 and 176, in three cases. In Fig. 174 the governor is of the angular movement type described above and illustrated in Fig. 170, driven off the vertical shaft from which the cam-shaft receives its motion. The fuel pump is of the horizontal plunger type, receiving its motion from eccentrics mounted on the same vertical shaft. The fuel pump body is supported by a facing on the lower side of the case containing the upper spiral gears. In the arrangement shewn in Fig. 175 the fuel pump is attached to the cylinder jacket, but in other respects the details are similar to those of Fig. 174. The governor shewn in Fig. 176 is of the more usual type characterised by a sleeve which is mounted on a feather and which rises as the engine's speed increases. The pump is of the vertical multi-plunger type, and regulation is effected by 210 DIESEL ENGINE DESIGN rotation of an eccentric shaft, on which are hinged the levers which operate the auxiliary plungers. The arrangements described, briefly cover the bulk of the Fig. 176. fuel pump and governor mechanisms found in practice, but mention must be made of some modern refinements which are coming increasingly to the fore. (1) Control of fuel valve opening. At light loads the dura- tion of opening of the fuel valve is greater than necessary if uncontrolled and the instant of opening which is most favourable for full load running is inclined to be late for light load running. At least one firm has attacked this problem of governor control of the fuel valve operating mechanism. (2) Blast pressure control. This question is closely allied with (1), as a shortened opening period would lead to increase of the blast pressure if the latter were not corrected. Apart from this the blast pressure has in any case to be altered in accordance with the load (unless cylinders are cut out of operation) if good combustion is to be secured at all loads, including no load. The blast pressure is placed under the control of the governor by means of a throttle slide on the compressor suction. (3) Pilot ignition. This refers to cases where exceptionally refractory oils are being used which require for their FUEL OIL SYSTEM 211 combustion a preliminary charge of a lighter oil, such as Texas oil, which is deposited in the pulveriser in advance of the main charge by a small auxiliary pump provided for the purpose. The necessity for this device appears likely to be obviated by improvements in fuel valves and the flame plates in particular. Fuel Injection Valves. — It now remains to deal with the valve by means of which the fuel is injected into the combus- tion space and to which oil is delivered by the fuel pump for this purpose. These may be broadly classified as the open and closed types respectively, and as the former form a relatively small class at present it is convenient to dispose of them first. Open Type Fuel Valves. — Fig. 177 is a diagrammatic view of such a valve, omitting all detail not required to illustrate Fig. 177. the bare principle. Oil is delivered to the space A past the non- return valve B by means of the fuel pump, and this type of fuel valve derives its name from the fact that the space A is in constant communication with the interior of the cylinder. It is to be noticed that the fuel pump is not required to deliver against the pressure of the blast air as the latter is restrained by valve C. The latter is opened by appropriate gear at the predetermined instant for injection and carries with it the fuel oil contained in the space A. The action appears to be highly eflftcient in pulverising effect and excellent fuel consumptions have been reported for engines in which these valves have been fitted. This type of fuel valve appears to have been devised in the first instance for use in horizontal engines in which it was anticipated that the more usual type of fuel valve would be at a disadvantage. A valve working on a somewhat similar principle has been tried, from all accounts successfully, on 212 DIESEL ENGINE DESIGN vertical engines, but has not yet, to the author's knowledge, become a standardised fitting. Closed Type of Fuel Valves .— In this type com- munication between the combustion space and the interior of the fuel valve only exists during the in- jection period, when the flow is always in the same direction, apart from such derangements as stuck valves or failure of the blast pressure. Fig. 178 shews what may not improperly be called the Augsburg type of fuel valve. Apart from the cast-iron body, the construction of which is sufficiently illustrated by the drawing, the principal parts are : — (1) Needle valve A. (2) Spring B. (3) Stuffing box C. (4) Pulveriser D. (5) Elame plate E. The needle valve is usually made of special steel, case - hardened in way of the stuffing box to prevent cutting by the packing. Accurate align- ment of aU parts of the needle is essential and readily secured by grind- ing between fixed centres. The lower part of the needle is preferably re- duced in diameter by a FUEL OIL SYSTEM 213 few thousandths, as a certain temperature gradient exists between the needle and the pulveriser tube which may lead to seizure if sufficient clearance is not allowed. The tip generally has an angle of about 40 degrees. The needle spring, in addi- tion to returning the needle to its seat against the pressure of the blast air, has to deal with the friction of the stuffing box, and may be figured out on the basis of a pressure of 1500 lb. per sq. in. over the sectional area of the needle at the stuffing box. The latter is usually provided with a screwed gland. Extra hole For Pi I Ignition charge .',:_ 00 1 to Q-0I3B Fig. 179. The pulveriser tube is held on its seat by a stiff spring, and serves the double purpose of affording some support to the needle and retaining in their relative positions the rings and the cone which play an important part in pulverising the fuel. It will be clear from the figure that the pulveriser is sur- rounded by blast air, which enters at F. The fuel is introduced by means of a narrow hole G, at a point H immediately above the top ring. If the point H is located too high the oil fails to distribute itself evenly round the pulveriser rings and in- efficient combustion results. Fig. 179 shews the injection end of the pulveriser, together 214 DIESEL ENGINE DESIGN with the flame-plate and nut, to a larger scale. The details shewn are those in most common use, but are subject to variation in the practices of different manufacturers. The proportions shewn are roughly indicative of good practice, but it must be admitted that the rule of linear proportionality does not appear to be rational in this case. Experience in this matter discloses two facts : — (1) That for a given engine there is a certain minimum diameter of pulveriser ring, below which results are not satisfactory (about 9% of the cyhnder bore). (2) That as cylinders are increased in size it becomes in- creasingly difficult to obtain a high M.I.P. These suggest the following hypothesis : — That the best results are to be obtained when the depth of oil in the pulveriser before injection is a certain amount, and the same for cylinders of all sizes. If this is true, then the ared, of the pulveriser ring should be in proportion to the cylinder volume. This would lead to the diameter of pulveriser rings being made proportional to the cylinder bore raised to the power of 1'5. Such a rule has not been adopted, and would probably lead to inconveniently large valves in the larger sizes of engines, but the question would appear to offer some induce- ment to research. A very large number of different types of pulveriser are in use, and have been described in the technical press ; but it still remains to be proved that they are more efficient than the common variety shewn in Pig. 179. A neat form of pulveriser tube, which dis- penses with the long narrow hole drilled in the fuel valve casting, is shewn in Eig. 180, from which it will be seen that the oil is led to an annular space A at the top of the tube, whence it flows down- wards to vke pulveriser rings via a number of grooves in the surface of the pulveriser tube. Holes B are provided to give passage for the blast air. Swedish Type of Fuel Valve.— Eig. 181 shews the construction of this type of valve, which has also been widely adopted and which is characterised by the fact that Fio. 180. "^ FUEL OIL SYSTEM 215 216 DIESEL ENGINE DESIGN M~ the needle is completely enclosed, within the casing and is subject on all sides except the extreme tip to the pressure of the blast air. On this account the spring A does not require to be as strong as that of an Augsburg type of valve of the same size. The needle is lifted in working by the lever B attached to a cross-shaft C, the end of which penetrates the casing through a stuffing box D. The mechanical means by which end thrust on the cross-shaft and bend- ing actions on the overhung end, due to the pressure on the external lever, are dealt with, will be clear from the figure without further explanation. The use of this type of valve appears to be limited at present to those designs in which the requirements of other parts of the valve gear necessitate the fuel valve operating lever being arranged off the centre line of the cylinder cover. Burmeister Fuel Valve. — ^The construction of this valve is shewn diagrammatically in Fig. 182, and its outstanding features are the use of a mushroom valve, the extreme sim- plicity of the whole arrangement, and the fact that the valve is opened by a downward movement. The latter is a particularly valu- able feature as it secures uniformity of valve gear and ease of withdrawal. The four classes of fuel valve described above include as members practically all the fuel valves in use on Diesel Engines at present. Each type has its advantages, but no one of them can be said to hold the field. Something similar might be said for the enormous variety of pulverisers patented and in actual use. It seems doubtful if any of these can claim out- standing efficiency. When *lot injection of a less refractory oil is used to facilitate the use of tar oil as fuel an additional hole has to be drilled in the fuel valve, as shewn dotted in Fig. 179. The question of burning tar oil is still in the experi- mental stage in this country, but the results so far obtained hold out hopes that it will be possible to dispense with pilot ignition in favour of special arrangements of a simpler char- FlG. 182. aoter in connection with the fuel valve details. FUEL OIL SYSTEM 217 Some of the arrangements by means of which fuel valves are operated are shewn in Figs. 183, 184, 185 and, 186. The long lever, which is a feature of all these schemes, is usually of cast steel, and should be of stiff construction. The fulcrum on which the lever hinges is common to the levers which operate Fig. 183. Fig. 184. the other valves, viz. air and exhaust valves in the case of four stroke engines and scavenge valves in the case of two stroke engines, and starting valves in both cases. With land engines, and many marine engines, it is usual to mount the fuel and starting valve levers on eccentric bushes mounted on the Fig. 185. Fig. 186. fulcrum shaft at such angles that the operation of putting the starting valve into gear automatically puts the fuel valve out of gear and vice versa. This is considered in detail in Chapter XI. The use of the needle type of fuel valve in conjunction with a single lever necessitates the latter being so disposed that its 218 DIESEL ENGINE DESIGN roller is rendered, more or less inaccessible by the cam-shaft, particularly if the latter runs in a trough (see Fig. 184). The difficulty may be got over by providing a small intermediate lever, as shewn in Fig. 185, to reverse the direction of motion. In spite of the objections which have been raised against this arrangement it appears to be satisfactory in practice. Design of Fuel Valves. — An approximate rule for the internal diameter of the body has already been given, being the same as the diameter of the pulveriser rings. The thick- ness of the walls (cast iron) may be from a third in large valves to a half in the case of small valves of the internal diameter. If the valve is of the Swedish type this thickness will be approxi- mately constant throughout the body of the valve, except in the neighbourhood of flanges, etc. If of the Augsburg type, those parts of the body not subject to pressure may be a little thinner. In all cases a good rigid job should be aimed at, as lack of alignment leads to sticking of the valve. The pulveriser tube is made of steel and the details, such as rings and cones, of steel or cast iron. The Swedish type of valve requires special care to be devoted to the design of the cross-shaft and its fit- tings, in order to obtain freedom under load, adequate bearing surface and accessibility of the stuffing box. As regards the valve as a whole, the designer should aim at shapely solidity and avoid flimsiness of detail. With the Augsburg type of valve (Fig. 178) the load necessary to lift the needle is the spring load less the product of the blast pressure and the area of the needle at the stuffing box (approx.) plus the gland friction. With the Swedish type (Fig. 181) the load may be taken as approximately equal to the spring pressure plus the product of the blast pressure and the area of the needle at its seat. This load evidently induces bending and twisting actions, which the cross-shaft should be pro- portioned to carry with a low stress. The weakest section is generally at the reduced diameter to which the external lever is keyed. The key itself shwild be amply proportioned, and is preferably made of tool steel. The ball thrust must be pro- portioned to the load obtained by the product of the maximum blast pressure into the sectional area of the cross-shaft at the stuffing box. The flame plate is of nickel steel and the dia- meter of the hole is usually about 1% of the cylinder bore, but the best size for any particular case must be found by experi- ment. The flame plate nut may be of steel or bronze secured FUEL OIL SYSTEM 219 to the fuel valve body by a fine thread and provided with flats to accommodate a spanner. The main points in the design of fuel valves may be sum- marised as follows : — (1) Rigidity and alignment of casing. (2) Alignment of the needle and its guide. (3) Freedom of all working parts. (4) Sturdy proportions for all small details. Literature. — Renolach, N. 0., " Tar Oils as Fuel for Diesel Engines." — Internal Gombusiion Engineering, July 15th, 1914, et seq. Porter, G., "Tar Oil Fuel and Diesel Engines." — Diesel Engine Users' Assoc, May 24th, 1917. Smith, P. H., "Two Essential Conditions for Burning Tar Oil in Diesel Engines." — ^Diesel Engine Users' Assoc, May 16th, 1918. CHAPTER XI AIR AND EXHAUST SYSTEM Four Stroke Engines. — So far as four stroke engines are concerned the parts included in this system are : — The air suction pipe. Air suction valve. Exhaust valve. Exhaust piping. Silencer. In the neighbourhood of cement works, or other sources of grit, a suction filter is sometimes added, with a view to preventing foreign matter from reaching the interior of the cylinder along with the indrawn charges of air. Air Suction Pipes. — The usual form of suction pipe for both land and marine engines " is shewn in Fig. 187, and is intended to serve the double purpose of muffling the sound caused by the rush of air past the inlet valve on the suction stroke and also to prevent in some measure the ingress of dust. To be of real service in either capacity the slots should not exceed about 40 /lOOO of an inch in width, and in use must be kept clear of dirt, other- wise the volumetric efficiency, and consequently the power of the engine, will be seriously impaired. To be reasonably effective as a silencer th^ slots should cease within a foot or so of the point where the pipe joins the cylinder cover. To prevent throtthng the aggregate area through the slots is usually made about 1-5 to twice the clear area of the pipe. Assuming the slots are spaced f " apart, the slotted length of the pipe works out to about 3i to 4 J times the bore. In very 220 Fio. 187. AIR AND EXHAUST SYSTEM 221 expert foundries these slots are formed in the casting (cast iron). Where the slots have to be milled aluminium is a convenient but expensive material for this purpose. Welded tubes of sheet iron are sometimes used, but are liable to become dented and unsightly. An efficient and durable arrangement, shewn in Fig. 188, consists of a common collecting pipe in communication with all the cylinders and ending in a trumpet-shaped piece which is very effective in muffling the sounds of suction. The trumpet fitting is similar to a cornet mute, and consists of internal and external members so arranged that the space between them has a sectional area increasing outwards whilst the distance between the two members diminishes. The noise question is most effectively disposed of by carry- -f 111 Fig. 188. ing the suction pipe outside the building in the case of a land engine or on deck in the case of a marine engine. If an air filter be added the arrangement is ideal but expensive. Where a separate suction pipe, as in Fig. 187, is fitted to each cylinder the bore is usually made equal to that of the suction valve or slightly larger. Where a common suction pipe is provided the bore may be anjrthing up to about double this size, according to the number of cylinders to be supplied and the length of the pipe. Suction Valves. — Atypical suction valve is shewn in Fig. 189. The valve proper may be of carbon or nickel steel, in the form of a drop forging. The upper end is guided by a little piston, which also takes the thrust of the spring. In the example shewn the guide piston is of chilled cast iron extended in the form of a nut and provided with a cup-shaped cavity to accom- modate the operating tappet. The casing being in two parts, it is necessary to finish each part on a mandril to ensure align- ment in position. It is not necessary at this point to describe 222 DIESEL ENGINE DESIGN variations in the details of suction valves, as the latter are usually made similar to the exhaust valves (which wiU be described in detail later) except for such special features as are necessary to deal with the heat effects to which the latter are subject. FiQ. 189. Dimensions of Air Suction Valves. — ^The requisite diameter for an air suction valve may be regarded as determined by the mean vacuum allowable on the suction stroke. Taking this to AIR AND EXHAUST SYSTEM 223 be 0-6 lb. below atmospheric pressure, the theoretical mean velocity is found from Fig. 13 to be about 280 feet per second, and taking the mean coefficient of discharge to be 0-70 this gives a mean apparent velocity of 195 feet per second. Now as regards the mean opening area of the valve, if we assume a harmonic cam opening and closing exactly at the upper and lower dead centres respectively, then the mean opening would be just half the maximum opening if the maximum Hft is made = J of the value diameter. Actually the valve is always arranged to open before top centre and close after top dead centre, so that the mean area is usually more like 0-6 of the maximum area. Adopting this figure, we may write : — V^ ^P^0-65x0-785d2 Where Vj= Apparent mean velocity of air in feet per second. Vp=Mean piston speed in feet /seconds. B =Bore of cylinder. d= Diameter of suction valve. From which, g-^I-54-^-^ p_ 127 Values of = calculated from this formula for various piston X) speeds are given below and agree well with average practice. Piston speed in ft. permin. . 500 600 700 800 900 1000 1100 1200 Ratio d-f-B . -257 -281 -303 -324 -343 -362 -380 -397 In the best practice the maximum lifts of the air and exhaust valves are frequently made as much as 0-35 of the valve dia- meter, as although the extra lift does not increase the maximum available area the mean area is increased and the opening at the dead centre is augmented without adopting an unduly long opening period or an awkward shape of cam. The air and exhaust valves are usually operated by the lever arrangement shewn in Fig. 186 with reference to fuel valves. Exhaust Valves. — In some small engines the exhaust valves are similar to and interchangeable with the air suction valves, but with medium and large size engines, owing to the larger dimensions of the exhaust valves, and in consequence their slower rate of cooling by conduction to the surrounding media, 224 DIESEL ENGINE DESIGN special arrangements have to be made to conduct heat away from the valve seat which would otherwise become pitted and grooved in a short time. With cyhnders up to about 16 in. bore the trouble can be reduced to reasonable proportions by providing the valves with cast iron heads, which are less liable to become pitted than those of nickel steel. Ventilation of the casing by means of perforations is also found useful. Some types of cast-iron exhaust valve heads which have proved successful in practice are shewn in Fig. 190. With cylinders much in excess of 16 in. bore some form of water-cooled casing is desirable, and the simple arrangement shewn in Fig. 191 has proved most Cast Iron Cast Iron f Iron Fig. 190. Fio. 191. AIR AND EXHAUST SYSTEM 225 effective with the very largest cylinders without resorting to the expedient of coohng the valve itself by direct means. It appears that the flow of heat from the seating to the water- jacket is suflBicient to keep the temperature of the former at a suitable low value so long as the mean indicated pressure does not exceed a reasonable figure dictated by other considerations. The valve spindle is liable to become stuck by carbonaceous deposit, and should therefore be about 20/1000 slack in the casing. Furthermore, arrangements should be made to feed a little paraffin to this point from time to time. The upper or guide end requires some care in detailing, Fig. 192. particularly in high speed engines, to prevent the nut from slacking back or the threads becoming stripped. The arrange- ment shown in Fig. 189 is simple and effective ; other arrange- ments are shown in Fig. 192 for slow speed engines. Valve Casings. — Two types of valve casings have been illustrated in Figs. 189 and 191, and further modifications are shown in Fig. 193. The valve seating is made in a separate piece, which is readily replaceable when worn by an inter- changeable spare. After repeated regrindings a ridge is formed which has to be turned off, a process more conveniently carried out on a light ring than on a whole casing. If no loose seat is provided it becomes necessary in time to turn down the under surface of the casing flange, in order to bring the valve seat 226 DIESEL ENGINE DESIGN flush with the under side of the cover. The port in the easing provided for the flow of exhaust gas should accurately coincide with the corresponding port in the cover, and it is usual to provide an equal and opposite dummy port to preserve the symmetry of the casting and diminish chances of distortion due to expansion or growth. The depth of the metal under the ports should not be reduced to too fine a limit, otherwise there is danger of leakage on the compression and ex- pansion strokes, due to deflection or even cracking of the metal at this point. The number of studs (two to four) used to secure the casing to the cylinder cover is purely a matter of convenience in arrangmg the gear on the cover or cylinder head. Two studs properly proportioned are sufficient for the largest valves, given an adequate depth of flange and good connection of the latter to the body. Exhaust Lifting Devices. — Some form of exhaust lift for breaking com- pression is usually fitted to both land and marine engines. With the former the use of such a device on shutting down the engine obviates the ten- dency of the engine to swing in the reverse direction to that for which it was designed, just before stopping, and also facilitates turning the engine by hand. With marine engines some such device should come into opera- tion automatically on reversing to present the possibility of compressing an unexhausted charge when motion is begun in the reverse direction. An expansion stroke executed in the ahead direction becomes a compression stroke in the astern direction and vice versa. Two hand devices are shown in Fig. 194. Tjrpe A consists of a substantial steel lever pro- vided with a slotted end which normally keeps it in a vertical position clear of the gear. To bring the lever into operation it Fig. 193. AIR AKD EXHAUST SYSTEM 227 is lifted to the extent of the slot and allowed to fall forwards. A projection on the lever then slips under an extension of the roller-pin on the first occasion of the valve being lifted and pre- vents its return. Type B consists of a link and screw, by means of which the valve may be depressed during the running of the engine and which normally lies alongside the casing. Fig. 195 shows three arrangements suitable for marine engines. In type A the valve is depressed by a pneumatic cylinder arranged above Fig. 194. and in line with the valve. Type B consists of a series of cams (one cam over each exhaust valve lever) mounted on a shaft running from end to end of the engine. A turning movement of this shaft simultaneously depresses all the exhaust valves. Type C is appropriate to those engines in which the valve levers are operated by push rods. The first movement in reversing consists of swinging the lower end of the push rods out of the range of operation of the cams. By a scheme of linkwork, which is obvious from the illustration, the same movement introduces a lever with an incUned face under a roller provided for this purpose attached to the valve lever, with the result that the valve is held off its seat until the push-rods regain their normal working positions. Proportions of Exhaust Valves and Casings. — The diameter of the exhaust valve is almost invariably made the same as that of the air valve, in order to simplify manufacture and give 228 [DIESEL ENGINE DESIGN a symmetrical arrangement of valve gear. On the exhaust stroke the area through the valve is more than sufficient, but a fairly early opening (40 to 50 degrees before bottom dead centre) is necessary to effect a rapid fall of pressure. The dimensions of the various parts of the valve and casing are with a few exceptions matters of experience only, and the approximate proportions given below with reference to Fig. 196 are representative of average practice. The figures are ex- pressed in terms of the valve diameter as unit. ■ B" Fig. 195. The larger figures refer generally to the smaller sizes of valve. The spring, which wiU be considered later, has usually between •twelve and twenty-two turns. A large number of turns reduces the range of variation of stress, and consequently increases resistance to fatigue. On the other hand, a spring with a small number of turns has less tendency to buckle^ The seating of the casii^ in the cylinder cover is almost always made square now instead of conical as formerly. The width of seating need not exceed a quarter of an inch in the largest sizes. The seating of the valve in the casing is usually made at an angle of 45 or 30 degrees, and the width of the seating varies greatly in different designs. Narrow seatings about an eighth of an inch in width appear to be least subject to pitting, and seatings as narrow as l/25th of an inch have AIR AND EXHAUST SYSTEM 229 been used successfully. For large marine engines seats about f inch wide seem to be preferred. The spindle clearance may be about 20/1000 and the guide piston clearance about 10/1000 in all sizes. The size of the holding-down studs may be found from the standard table (page 130), the total load being based on a pressure of 500 lb. per sq. in. over the least area of the casing where it makes joint with the cylinder cover. The casing lugs shoiold be amply proportioned, particularly in cases where the castings are not above average quality, and they should have a good hold on the cylindrical part of the casing, reinforced if necessary by internal ribbing. It is not ■h 12 to?2 turns a 6 c e / h 015 0-20 046 007 046 006 to to to to to to 018 0-24 056 OIT 0-64 008 Fig. 19 unknown for these lugs to break off in tightening up the studs, so it is as well to err on the strong side. Exhaust Valve Springs. — Apart from the weight of the valve in itself there are causes tending to open the exhaust valve when it should be shut, viz. : — (1) The vacuum on the suction stroke. (2) With four cylinder engines especially, the first rush of exhaust from the neighbouring cylinder if the latter discharges into the same collecting pipe. In addition, the inertia of the valve and any levers, rods, etc., in connection therewith, tend to make the latter lose contact with the operating cam in the neighbourhood of maximum lift. These various influences are overcome by fitting a spring, the normal load of which is equivalent to a 230 DIESEL ENGINE DESIGN pressure of about 10 lb. per sq. in. of valve area in the case of slow speed engines, and anything up to about 20 lb. per sq. in. or more in the case of high speed engines. A method of comput- ing the inertia effect will be dealt with in some detail as there is a tendency for higher speeds to be used in practice, and the principles involved have a wide application in the design of high speed machinery generally. Inertia Effect of Valves. — Consider the simple arrangement shown in Fig. 197, consisting of a valve directly operated by a cam without intermediate members. Let the weight of the valve guide and roller, etc., be " W" lb. The effect of the inertia of the spring may be allowed for by adding one-third of its weight to that of the other parts. Let " X " be the distance in inches of the valve from its seat at any instant. If the shape of the cam and the speed of the cam-shaft be known it is possible to ex- press " X " in terms of the time " t " in seconds counted from the instant at which the valve begins to lift, by means either of an equation or a graph exhibiting the lift on a time base. If this equation (equation of motion) is available, then one differentiation with respect to "f' gives an expression for the velocity denoted by " x," and a second differentia- tion gives the acceleration denoted by " X." If the relation between x and t is given by means of a graph, then the differentiation may be done by one or other of the graphical methods explained in books on practical mathematics. In either case, x being measured from the valve seat outwards, positive values of x denote inertia effects tending to press the roller against the cam, and negative values of x denote inertia effects tending to cause the roller to lose contact with the cam. Here we are only concerned with the negative values of x. If X denotes the resultant force on the valve, neglecting all Fio. 197. effects except those due to the inertia, W Then, X = — .XntrA-g. (1) g=386in./sec.« AIR AND EXHAUST SYSTEM 231 Equation (1) gives the minimum value of the spring pressure to prevent the roller jumping due to inertia. The value of x (max.) is very easily calculated in one case, viz., when the cam is so designed that the valve describes simple harmonic motion ; that is, when the graph of x and t is a sine curve. In this case it is convenient to measure x from the position of mid-Uft positive outwards and negative inwards. The equation of inotion is then : — x=A.sin pt — — (2) where A = Half the maximum lift in inches, and p is a constant such that pT=27r, where T is the whole period of opening in seconds. If the exhaust valve is open for 240 crank-shaft degrees, then : — _ 60^240 , 27r.nx360 „ _„ ^=n ^360' ^""^ P=^60^^240-='-^''^— ('^ " n " being the number of revolutions of the engine per minute. From (2) x= — Ap^.sin pt. And X(jj^x)= - Ap"./ W Substituting in ( 1 ) X = — ^ . A.p ^ (4) Example: W = 10 lb. A=Half lift=0-375". n=400 R.P.M. From (3) p=0-157 x 400 = 62-9. From (4) X=;^ X 0-375 x62-92 = 38-5 lb. The valve would therefore require to be provided with a spring capable of exerting a force of 48-5 lb. to deal with inertia and dead weight only, apart altogether from gas pressure and friction. The use of the harmonic cam, to which the above figm-es apply, has not become very general in Diesel Engine practice, the tangent cam shown in Fig. 228, Chapter XIII, being more commonly employed. Typical velocity and acceleration curves for a tangent cam are shown in Fig. 198. Effect of Levers and Push Rods, etc. — The simple case considered above of a cam operating directly on the end of the valve is seldom realised in practice, and a somewhat more general case, illustrated diagrammatically in Fig. 199, will be considered. We have now several different members, all participating in the motion and acquiring momentum which 232 DIESEL ENGINE DESIGN AIR AND EXHAUST SYSTEM 233 must be overcome by the spring. The problem is a simple case of the general theory of a system of one degree of freedom simplified by treating the " coefficient of inertia " as a constant instead of a function of x. Consider any particle of the lever AB situate at a distance " r " from the fulcrum E and having" k E ^-jtAwtts^ ^b\ PK^-J_ J _C_ ^ h "^ \j ^■ i Fig. 199. a weight " w " lb. If s denote the speed of this particle during any small displacement of the system, then : — s=x.- and s=x.- a a The effective force acting on the particle is therefore equal to •X— and the reaction Xj at A due to all such particles of the g a lever is given by ^ 2w.r2=?.Wi^i' -(5) ga^ g a^ Where Wi is the weight of the lever and k^ is its radius of gyration about E. W k 2 The expression — i-|- is the inertia co- g.a efficient of the lever with respect to the co-ordinate x, and may be denoted by A^. The total reaction X at A, due to the inertia effects of the valve itself, aU the levers, push-rods, etc., is the sum of all the reactions due to the individual members, and therefore, -j^^^ (Ao+A^+A.+A,) (6) Wn Aq being= — where Wo=weight of valve. 234 DIESEL ENGINE DESIGN and Aj is the inertia coefficient of the push-rod, and A» is that of the link. By similar reasoning to that given above for A i it is found W,/b\2 that A2= — -I - ) where W2=weight of push-rod. and A,=^ — -I -^ ) , where W3=weight of link, and k3=its g \ ac / radius of gyration about its axis. It will be seen at once that equation (6) is similar to (1), with inertia coefficient substituted for mass. The assumption made is that the angles a ^ydo not deviate far from 90 degrees. For ordinary practical purposes a deviation of 10 or 15 degrees on either side involves a negligible error. Example: Wo= 6 lb. a=10". kx=7" Wi=151b. b = 12". ka=5" W2= 81b. c= 7". W3= 51b. x = 1600in./sec.2 X=x(Ao+Ai+A2+A3) 1600r„ , 1 ./ 7 Ws/12\' , /12X5Y- = 118 lb. For slow running engines the inertia does not usually amount to more than two or three lb. per sq. in. of valve area. In most cases the spring may be based on the inertia load plus about 6 lb. per sq. in. of valve area, to deal with the other effects which enter into the question. Strength and Deflection of Springs. — ^The usual formulae for the safe load and the deflection of springs made of steel wire of circular section are given below for handy reference. 'P=Safe load (maximum) in lb. Tg f= Safe stress, usually about 60,000 lb. P=0-2f — Where i gper sq. in. ^ I d=Diameter of wire in inches. r =Mean radius of coils in inches. S =Deflection in inches. n= Number of turns (free). G=Modulus of rigidity, usually taken to be about 12,000,000 lb. per sq. in. And 64.n.r^ P "^ d* G Where AIR AND EXHAUST SYSTEM 235 Exhaust Piping. — A common arrangement of unjacketed cast-iron exhaust piping is shown in Fig. 200. The flexibility of the system renders any special provision for expansion un- necessary. The piping itself being out of reach need not be lagged, and may be as light as casting considerations will allow. Fig. 200. The connecting pieces between the various covers and the common discharge pipe may be made equal in bore to the diameter of the exhaust valve. The bore of the collector pipe joining the silencer may be proportioned with reference to the nominal velocity of the exhaust gases as follows : — Let Vp =Piston speed in feet per second. Vj=Nominal speed of exhaust in feet per second. B =Bore of cylinder in inches. d= Bore of exhaust pipe in inches. n=Number of cylinders. T"™ v.=y.(2)-x5 In using this formula " n " should be put equal to 4 in all cases where the number of cylinders is equal to or less than 4. The reason for this is that the gases discharge intermittently, and an engine of one, two or three cylinders requires approxi- mately the same size of pipe as a four-cylinder engine of the same size and speed. The value of V^ varies from about 70 in 236 DIESEL ENGINE DESIGN small engines to 110 in large. The pipe leading from the silencer to the atmosphere may be made about 25% larger in the bore. A somewhat neater arrangement, involving an inter- mediate collector under the floor, is shown in Fig. 201. The individual exhaust pipes must now be water-cooled, but there is no objection to short, uncooled sections in way of the flanges of sufficient length to accommodate the bolts. If the pipes are cast, the thickness of the outer walls need not exceed about J" to f " with good foundry work. The jackets may also be of welded steel tubes or sheets. Fig. 201. In marine installations the exhaust is sometimes used to furnish a supply of hot water for heating and other purposes, and this may be achieved by providing the exhaust collector with nests or coils of tubes through which water is circulated. As a rule the exhaust from a four stroke marine engine is not sufficiently hot to necessitate the provision of a water-cooled silencer apart from the arrangements which have just been mentioned. Silencers. — As a rule four stroke Land Diesel Engines are supplied with a cast-iron silencer, having a capacity of about six times the volume of one of the working cylinders, and a common type is shown in F^. 202. The result is not always satisfactory, and better iresults are obtained by using a large underground brick or concrete chamber having twenty or thirty times the volume of one cylinder. Wrought iron silencers, unless water-jacketed or buried underground, usually give out a ringing noise, unless the gases on entry are made to difEuse through a trumpet arrangement similar to that described above under suction pipes. AIR AND EXHAUST SYSTEM 237 Two Stroke Engines. Scavengers.— If the efficiency of the scavenging process could, be definitely ascertained in every case, it would be a simple matter to calculate a suitable capacity for the scavenge pump. In those engines of which the results on trial suggest that this process is nearly perfect the scavenge pump appears to have a stroke volume capacity of about 1 -4 times the aggregate stroke volume capacities of the cylinders which it feeds. It hardly appears safe, however, to assume that this allowance will necessarily be suffi- cient in any proposed design, and some constructors have adopted a higher figure, 1-8 to 2, in their first attempts at two stroke design. The scavenge air pressure is a factor which decides itself when once the details of port and valve openings have been fixed, and check calculations must be made on some such lines as those indicated in Chapter III, to make sure that reasonable limits of pressure will not be exceeded. The pressures obtained in practice vary from about 3 to 7 lb. per sq. in. The work done by the scavenger is all lost work and must therefore be kept at a low figure. On the other hand, the super-pressure above that of the atmosphere with which a controlled scavenge engine starts the compression stroke, is a valuable factor in increas- ing the power of a given-sized cylinder. If simplicity is the main consideration and uncontrolled port scavenge is adopted, it becomes necessary to be content with a very moderate M.I.P., as the super-pressure obtainable at the beginning of compression appears to be very limited with the arrangements hitherto adopted on such engines. Construction of Scavengers. — It is not necessary to deal exhaustively with the details of scavenge pumps as these differ little from the corresponding parts of L.P. steam engines. The scavenger, or scavengers, are preferably driven off cranks provided for the purpose on the main shaft, as in Fig. 203. In Fig. 202. 238 DIESEL ENGINE DESIGN marine designs the link-drive shown in Fig. 204 has also been used, but as usually carried out is open to the following objections : — (1) The arrangement does not lend itself readily to the closed engine type of framework which appears to have every advantage for Diesel Engine work. (2) The side levers involve cantilever connections, which lack rigidity. (3) The heavy reversals of thrust are liable to cause knock- ing and vibration, owing to small bearing surfaces and poor lubrication. Fig. 203. Fig. 204. Fig. 205. A scheme suitable for land work but a trifle inaccessible for marine purposes consists of a tandem arrangement of scavenger and low pressure stage of blast air compressor, as in Fig. 205. Valve Gear. — ^The earlier Diesel scavengers were fitted with piston valves single or double ported, and this type of gear is still retained in some des^s. For reversing such valves the Stevenson link motion appears to be the best solution on hand at present. In other designs automatic disc or plate valves are being increasingly used, with a gain in simplicity and the advantage in marine work that no reversing gear is necessary. The liability of the valves to failure by fatigue seems to be their chief drawback. Scavenge Air Receivers. — From the scavenge pump the air AIR AND EXHAUST SYSTEM 239 passes to a receiver in. communication with the cylinder covers or scavenge air belts of the working cylinders. With the usual arrangement, where one, or at most two, double-acting scavenge pumps are used to supply a number of working cylinders, say three to eight, it is advisable to make the capacity of the receiver large compared with the stroke volume of one scavenger, in order that the pressure in the receiver may remain sensibly constant, and a suitable capacity may usually be secured by making the diameter of the receiver about 1 to 1-3 times the cylinder bore. In any proposed case it is a simple matter to construct a diagram showing on a time base the rate at which air is being passed to the receiver by the scavenger and carried away from it by the working cylinders. The resultant effect in creating fluctuations in a receiver of any proposed capacity is then easily calculated. The effect of too small a receiver capacity is to give those cylinders which begin compression at the instant of maximum scavenge pressure an advantage over those less favourably timed. The effect is readily studied in practice by means of light spring diagrams taken from the receiver and the working cylinders respectively. By way of example, if two cylinders start compression with absolute pressures of 21 and 20 lb. per sq. in. respectively, then the first wiU (other things beiag equal) have a maximum power capacity 5% greater than the second, or if they are worked at the same power the second cylinder will work with a mean absolute charge temperature 5% greater than the first, which is no small evil. In some two stroke designs one double-acting scavenger is provided for each pair of working cylinders, and if the deliveries are correctly timed with respect to the scavenge periods the receiver capacity does not require to be very large. Scavenge receivers are usually made of riveted or welded sheet steel, having a thickness of about one per cent of the diameter. A disastrous explosion at Nuremberg in 1912, traceable to the ignition of lubricating oil vapour in the scavenge receiver of a large two stroke engine, points to the desirability of providing drain cocks and a relief valve or safety diaphragm of large area. The area of the trunk communicating with each cylinder should be well in excess (usually about double) of the maximum aggregate area of the valves or ports which it supplies. Scavenge Valves. — In different designs embodying the principle of scavenging through the cover one, two, three and four valves have been employed. In particular, if two valves 240 DIESEL ENGINE DESIGN are used they may be identical with the air and exhaust valves used in four stroke engines of the same size, with the piston speeds at present customary. With any other number of valves the area may be made equivalent. In view of the Fig. 206. relatively light duty which devolves upon them, scavenge valves are usually made of somewhat simpler construction than exhaust valves, as shown in Fig. 206. To prevent loss of scavenge air past the spindle the latter is usually made a good fit. With controlled port scavenge, double-beat valves, piston AIR AND EXHAUST SYSTEM 241 valves and Corliss type valves have all been used in different designs, as shown diagrammaticaUy in Fig. 207. The use of the valve is exclusively to regulate the instant at which air is admitted to the cylinder, the point of cut-off being determined by the piston covering the ports on the up stroke. It therefore foUows that so long as the valve is full open at the instant when the ports are covered, the point at which the valve seats again, may be determined arbitrarily by other considerations affecting the valve gear. Exhaust System. — The chief evil to be guarded against in the design of the exhaust system of a two stroke engine is the interference of the exhaust rush of one cylinder with the scavenge process of another. An obvious but inconvenient J way of avoiding this trouble is to provide separate exhaust Fig. 207. pipes and silencers for each cylinder. Practically the same effect can be achieved by providing common exhaust systems for pairs of cylinders whose cranks are at 180°. This, however, does not appear to be necessary, and satisfactory results with an exhaust system common to all cylinders are obtained by the use of arrangement similar to that shown in Fig. 201 for four stroke engines. The essential point is that the pipes which connect each cylinder to the common collector should be of large diameter (about three-quarters of the cylinder bore), and that the collector itself should have a volume large in comparison with the stroke volume of one cylinder. It does not follow, however, that a free passage for the exhaust is a necessary condition for efficiency, as the latter has sometimes been improved by the insertion of a throttling diaphragm in the exhaust passage at the point where the pipe joins the cylinder. 242 DIESEL ENGINE DESIGN Silencers. — The sudden release by the uncovering of ports of a pressure of 40 lb. per sq. in. and upwards produces a noise which in the absence of a silencer can be heard some miles off. A type of sUencer which renders the exhaust inaudible a few yards away without imposing any back pressure, is shown in Fig. 208. This type of silencer is most effective when water- Diameter of Silencer . 3-5 B Length ,, ,, . 5-6 B Bore of Inlet Pipe . . OSS B EEt Fig. 208. jacketed. Approximate main dimensions are given in terms of the bore of the engine cylinder on the assumption of a piston speed of about 800 feet per minute. For land purposes a large pit without special baffles would probably serve equally well. Literature. — For information on the mechanics of cam- operated mechanism, see : — Goodman, J., Mechanics Applied to Engineering (Longmans). CHAPTER XII COMPRESSED AIR SYSTEM As mentioned in Chapter I, the injection of fuel by means of an air blast is one of the outstanding characteristics of the Diesel Engine, and it seems probable that its use in the early experimental engines was suggested by the compressed air apparatus used to start the engine, and which still appears to be the most practicable method of doing this. The utihty of the air blast is by no means confined to its function of injecting the fuel ; in fact the widespread use of mechanical means of injection in other types of oil engine clearly indicates that effective atomisation can be obtained otherwise. As Guldner has pointed out, the use of an air-blast probably secures a more efficient mixing of the cyhnder contents than could be obtained in any other practicable manner. The advantages in efficiency which such mixing secures are easily appreciated on examination of the thermodynamic principles involved. With good mixing the combustion proceeds rapidly, and reduces after-burning to a minimum ; and, further, the whole charge tends to remain homogeneous as to temperature, a necessary condition for maximum efficiency. In practice there are two aspects from which the efficient utilisation of heat should be viewed, viz. : — (1) Economy in fuel consumption. (2) The effect of efficient combustion in keeping the mean cycle temperature to a minimum. The last consideration is a vital one from the point of view of rehabUity and durability, and experience abundantly proves that a relatively small increase of the cycle temperature, due to overloading, leakage past valves, loss of volumetric effi- ciency or other causes is sufficient to convert an otherwise reliable machine into a source of continual trouble. 243 244 DIESEL ENGINE DESIGN The blast air has a pressure which varies from about 900 to 1000 lb. /sq. in. at full load, to about 600 lb. /sq. in. at no load, and in land engines is usually supplied by a comprescor form- ing an integral part of the engine. In marine installations the compressors are sometimes driven by separate auxiliary engines. The arrangement adopted by one maker is to drive the lower stages of the compressors by auxiliary engines, the last stage being performed by a high pressure plunger driven by the main engine. From the compressor the air passes, via coolers, to a blast reservoir or bottle of sufficient capacity to absorb fluctuations of pressure, and fitted with suitable distributing valves, one of which communicates with the fuel injection valves and another enables surplus air to be passed to the storage reservoirs provided for starting purposes. Cam- operated valves in the covers of one or more cylinders enable the stored air to be used, to give the engine the initial impetus which is necessary before firing can begin. With land engines the starting bottles are generally charged to a pressure of about 900 lb. per sq. in., and with the fly-wheels commonly used it is not necessary to provide starting valves for more than one cylinder out of three. With marine engines, storage pressures of about 300 lb. per sq. in are more common, on account of the difficulty of making high pressure reservoirs of large size, and in order to secure prompt starting from any position starting valves are fitted to every cylinder. The air system also includes certain servo-motors or air engines,, frequently used to perform operations of reversing the valve geiir. The various organs will now be considered in more detail. Air Compressors. — Four stroke land engines of the slow speed type as at present constructed require compressors having a capacity of about 15 cubic feet per B.H.P. per hour, which assuming a volumetric efficiency of 80% corresponds to a stroke volume capacity|rf)f about 19 cubic feet per hour. High speed engines appear as a rule to require about 25% more than this allowance. The above method of basing the com- pressor capacity on the B.H.P. is not a very satisfactory one, as different makers have different views as to power rating. A better plan is to express the L.P. stroke volume as a per- centaige of the aggregate cylinder volume, and the following figures are representative of average practice for land engines. In view of the demands made on the system when manoeuvring, COMPRESSED AIR SYSTEM 245 marine engines ai;e usually provided more liberally, to the extent of 50 to 100%. Bore of Ratio L.P. stroke vol. -H Stroke vol. of working cylinders. working cylinders. Four Stroke Engines. Two Stroke Engines. 10 15 20 0-08 0-07 0-05 0-16 0-14 0-09 Number of Stages. — ^For small slow-running compressors two stages are sufficient, but a 9-inch diameter of low pressure ■cylinder appears to be about the safe limit ; and even with this restriction it appears wise to abandon the principle of equal distribution of work between the stages. The small diameter of the H.P. cylinder affords little cooling surface for the dissipation of heat, and this consideration points to the advisability of arranging for the greater part of the work to be done in the L.P. stage, a conclusion which has been anticipated by experience. Three-stage compressors are being increasingly used, even the smaller sizes, four stages being required in the very largest installations only. With three or four stages the principle of equal division of work is open to less objection, owing to the smaller ratio of compression in each stage. Compressor Drives. — ^Almost every conceivable type of drive has been adopted at one time or another, and only the commonest are mentioned below : — (1) Tandem two or three stage compressor driven off the crank-shaft. This arrangement appears to have the balance of advantages for most purposes. (2) Tandem two-stage compressor driven by links and levers from each connecting rod or crosshead. This arrange- ment is expensive, but has the advantage of distribut- ing the work amongst a number of small compressors, which are subject to less heat trouble than one com- pressor of the same capacity. The suction pressure of the H.P. stage is usually sufficient to prevent reversal of thrust due to inertia, and consequently sweet run- ning is secured. 246 DIESEL ENGINE DESIGN (3) Similar to (2), but stages separate. This arrangement is bad, as tlie cooling siirface is less than in case (2), and the L.P. gear is subject to reversal of thrust due to inertia. (4) Twin tandem cylinders driven by links off the crank- shaft. This arrangement gives good results, as the load is divided between two units and the pressure on the crank-pin is reduced by the leverage of the link- work. Constructive Details. — ^The construction of air compressors being a specialised branch of mechanical engineering, it is not proposed to give here more than a very brief reference to the subject . The cylinders of tandem two and three stage machines are frequently cast in one piece, including the water-jacket. The relatively low temperatures obtaining justify this pro- cedure, provided sound castings can be obtained with reason- able regularity. The foundry work may be simplified in the case of two-stage compressors by the following division of material :— L.P. Cylinder and Jacket — one casting. L.P. Cover and H.P. Jacket — one casting. H.P. Liner and H.P. Cover — separate castings. One advantage of this scheme is the possibility of renewing the H.P. Liner when worn. The latter is peculiarly subject to rapid wear on account of the high pressure behind the rings. Similar arrangements are of course possible with three-stage machines. The possibility of increasing the cooling surface of the H.P. liner by means of ribs does not appear to have received very much attention. The trunk pistons serve as admirable crossheads, being almost entirely free from the heat troubles to which the pistons of internal combustion engines are subject. For the L.P. and inter- mediate stages Ramsbottom rings are usually fitted. In j^ery large machines the latter can also be fitted to the H.P. plunger. Small H.P. plungers are usually fitted with some arrange- ment similar to that shown in Fig. 209. The only thing which need be said about the connecting rods is that on account of the thrust being always in one direction, special care is required in the details of lubrication. The design of valves has an important bearing on the success COMPRESSED AIR SYSTEM 247 or failure of a compressor. The chief evils to be avoided are : — (1) Sticking of the valves off their seats, due to deposits of carbonised oil. (2) Damage to valves or valve seats, due to hammering. The first is influenced more by the efficiency of the cooling arrangements and the compression ratio than with the design of the valves themselves. For obvious reasons the H.P. valves are most subject to this trouble. The second trouble is usually due to the valves being too heavy, having too much lift, or the failure to provide adequate cushioning, and in successful designs is avoided by one or more of the following means : — (1) Making the valves in the form of very light plates or discs - with a very small lift. (2) Providing a large number of very small valves in place of one or two large ones. (3) Where large valves of considerable weight are used, arranging for some sort of dash-pot action. All the valves should be easy of access and removal. Ex- periments with existing types of compressor seem to indicate that makers are inclined to base their valve dimensions on an air speed very much lower than is necessary. One or two per cent loss of efficiency is of small importance, if such a sacrifice enables the size of the valves to be reduced. In some designs the intercoolers are separate from the compressor cylinder, and in others take the form of pipe-coil^ arranged round the compressor cylinders inside a removable water-jacket. Vibration of the coUs should be prevented by adequate clamps and stays, and no sharp bends are allowable, on account of a scouring action (presumably due to turbulent flow) which in acute cases may cause fracture of the pipe in a short time. L.P. intercoolers are sometimes made similar to tubular condensers, and in other designs take the form of a cast-iron vessel provided with internal helical baffles which give rise to turbulent flow and increase greatly the efficiency of the cooling surface. It is desirable in all cases to fit a final cooler, to reduce the temperature of the fully compressed air before entering the blast receiver. Each receiver should be fitted with safety valve and drain. One or two isolated cases of explosion, traceable to accumulation of lubricating oil in the intercooler system, emphasise the necessity for these fittings. 248 DIESEL ENGINE DESIGN Some makers fit special " purge-pots " in communication with each receiver for the collection and discharge of condensed water and oil. Calculations for Compressors. — In calculating the L.P. stroke volume required to furnish a given free air capacity, allowance must be made for the volumetric efficiency, which depends mainly on the clearance space and the delivery pressure. For example, suppose the clearance to be 3% of the stroke volume and the receiver pressure to be 150 lb. per sq. in. On the suction stroke, the suction valve will not begin to lift until the air left in the clearance space has expanded down to , atmospheric pressure. If this clearance air expands according to the law : — P.V.i-2=constant, then its expanded volume expressed as a percentage of the stroke volume wiU be : — /I HA .n\ 1 = 22-5% Subtracting its original volume, viz., 3%, the amount by which the effective stroke is shortened is 19-5%, and the volumetric efficiency is 80-5%. A further deduction should strictly be made for the fact that at the beginning of the compression stroke the cylinder contents are in general at a pressure slightly less than atmospheric. One or two per cent will usually cover this contingency. This example is sufficient to show the importance of reducing the clearance volume of the L.P. cylinder to a minimum. The efficiencies of the L.P. or H.P. cylinders may be found similarly, but only influence the volumetric efficiency of the compressor as a whole indirectly by raising the receiver pressure above the value it would have if there were no clearance. It will be evident on reflection that leakage past the H.P. delivery valves will also raise the receiver pressures, and for this reason it is desirable to fit pressure gauges to all receivers so thsft the condition of the valves may be inferred from the gauge readings. Assuming perfect intercooling and equal volumetric effi- ciency in all the stages, the pressure of the atmosphere and the receiver pressures (absolute) will be in inverse ratio to the stroke volumes of the cylinders, and equal division of work between the stages will be secured by the proportions given below ;■ — COMPRESSED AIR SYSTEM 249 Two f Atmospheric Pressure. Intermediate Pressure. High Pressure. .\ 1 at. -14-7 lb./in.2 8 at. =118 Ib./in.^ 64 at. -940 Ih./in.' T L.P. volume. -8 H.P. volume. Atmospheric 1st Intermediate 2nd Intermediate High Pressure. Pressure. Pressure. Pressure, lat. =14-71b./ 4at. -58-81b./ 16 at. =235 lb./ 64at. =940 lb./ in.* in.= in.^ in.^ Three Stages.' L.P. volume : I.P. volume : H.P. volume =16 : 4 : 1. In practice better results are obtained with two-stage machines by the following proportions : — Atmospheric Pressure. Intermediate Pressure. High Pressure. 1 at. =14-7 lb./in.» 12 at. =177 Ib./in.^ 64 at. -940 Ib./in.^ L.P. volume -12 H.P. volume. Assuming that the L.P. stroke volume has been determined by some such considerations as the above, the actual cylinder dimensions are found by selecting suitable values for the piston speed and the L.P. stroke to bore ratio. The piston speeds commonly used lie between about 300 to 600 feet per minute, the lower speeds being usually associated with small machines. With a little care it is possible, by making small variations in the strokes, to design a series of four or five compressors suitable for engines covering a wide range of powers. Such a scheme involves sacrifices in some cases which, however, would appear to be quite outweighed by the advan- tages of standardisation. The ratio of stroke to bore of the L.P. cylinder wiU generally lie between about 0-7 and 1-7. The valve areas for each stage are based on some figure for the mean velocity obtained, in the case of suction valve, by multi- plying the mean piston speed by the ratio of piston to valve area, and in the case of delivery valves the mean piston speed during the delivery period by the same ratio. Certain con- tinental authorities recommend speeds not exceeding 80 and 115 feet per second for the suction and delivery respectively, but it appears that these figures may be doubled or even trebled with impunity, and sometimes to advantage. The calculations of the strength of the various parts are straightforward, involving no special principles, and are there- fore passed over. The cooling surface to be provided for inter- cooling is a very important matter, and the following figures from a eucoessf ul design may be useful as a basis of comparison in the absence of first-hand experimental data. 250 DIESEL ENGINE DESIGN Three-stage Compressor, Cooling surface, copper-pipe coils. Free air capacity, 130 ft. ^ /min. L.P. 8-7 ft. 2 L.P. 3-6ft.2 H.P. 3-6ft.2 Fig. 210. The subject of heat transmission being a very important one ■^ ^ in connection with internal combustion engines, a brief reference to the usual theory is inserted below. Transmission of Heat through ■ Plates. — Referring to Fig. 210, the r^ direction of heat flow is indicated by ;. ■ the arrow and the symbols t,^, t^, tg, t^ denote the temperatures of the hot fluid, the hot side of the plate, the cold side of the plate, and the cold fluid respectively. The total heat drop con- sists of three stages : — (1) An apparently sudden drop from the hot fluid to the plate. (2) A steady gradient across the thickness of the plate. (3) An apparently sudden drop from the plate to the cold fluid. The assumption is that the rate of heat flow is dependent only on the temperature drop, the thickness of the plate, and the particular fluids employed. On this assumption, if Q is the amount of heat transmitted per hour per unit of area, then : — Q=ai(ti— t2)=^(t2- t3)=a2(t3-t4) (1) from which Q= (ti-t*) -(2) 1 +i-F- Where d=thickness of plate and oi, 02, and X are constants. According to Hiitte, for air at atmospheric pressure and small velocities ai=about 0-4:-{-1-1-\/y. Where v= velocity in ft. /sec, and for water (not boiling and without turbulence) a2=about 100 B.T.U. /ft.^ deg. F. COMPRESSED AIR SYSTEM 251 Values of X are given below for various metals : — Iron . . . 460] Mild steel . . 320 1 B.T.U. /ft. ^ deg. F. per inch of C!opper . . . 2100 j thickness. Brass . . .740/ The values of X are well determined, but unfortunately the term involving this constant is the least important of the three as the bulk of the heat drop occurs at the surfaces of the plates. The values of oi and a 2 must be used with caution as the small amount of published data indicates that these constants are subject to enormous variation under different circumstances. In particular, the value of a 2 is greatly increased by the eddying motion produced by the introduction of spiral baffles in condenser tubes and the like, and is increased about tenfold if the water boils. The value of a^ for air increases greatly with rise of temperature and pressure. Apart from these uncertainties, formula (2) suggests the foUoTving corollaries : — (1) The rate of heat transmission is but little influenced by the thickness of the plate in most practical cases, such as cylinder covers, etc. (2) The temperature drop across the plate, and consequently the temperature stress, is proportional to the rate of heat transmission and to the thickness of the plate ; this suggests one reason why it is advisable to work large engines at a smaller mean cycle temperature than small ones. For further information on the subject of heat transmission, the reader is referred to the sources of information mentioned in the footnote.^ Air Reservoirs. — The usual arrangement of air reservoirs for land engines is shown in Fig. 211. This scheme was devised in the very early days of the development of the Diesel Engine and no substantial improvement has been made on it in recent years. Two starting and one blast-air bottles are provided, all designed for a working pressure of about lOOO'lb. per sq. in. One of the starting bottles serves as a reserve, in case of a 1 " High-speed Internal Combustion Engines," Judge. " Heat Trans- mission," Report by Prof. Dalby to the Inst. Meoh. Engs., 1909. "Notes on Recent Researches," paper by Prof. Petavel : Manchester Assoc, of Engineers, Oct., 1915. "The Laws of Heat Transmission," Lecture by Prof. Nicholson : Junior Inst., Jan., 1909. 252 DIESEL ENGINE DESIGN failure to start the engine, due to any derangement. In the event of such failure, every care is taken to make certain that the engine is in perfect order before using the reserve bottle, and it seldom happens in practice that the bottles require to be replenished from outside sources of supply. The connections between the bottles, the air compressor and the engine should Starting Pipe jmpressor- To Fuel Injection Valves Fig. 211. be quite clear from the diagram. Only one or two points will be mentioned. Before starting up, it i^ossible to ascertain the pressure in each of the three bottles by opening up the appro- priate valve on each bottle-head in rotation In each case the pressure is recorded on the left-hand gauge.. The pressure in any pair or all three bottles may be equalised by opening up a pair or all three such valves. The right-hand gauge registers the blast pressure on the engine side. By throttling the blast control valve on (1) (2) (3) COMPRESSED AIR SYSTEM 253 the blast bottle-head the injection pressure may be regulated below that of the bottle. This is done when replenishing the starting bottle on light load. It is thus possible to pump up the starting vessel to 1000 lb. /in. 2 whilst the blast pressure is only 600 lb. /in. 2, as required for light running. FlQ. 212. The bottle-heads containing the various valves are usually machined from a solid block of steel A detail of one of the valves is shown in Fig. 212. The bottles themselves are of weldless steel, and a neck is frequently screwed on, as in Fig. 213. Some idea of the capacities of the bottles commonly provided may be gathered from the following table : — 254 DIESEL ENGINE DESIGN Total Capacity of H P Starting Air Bottles (four stroke engines) Engines of about 9" bore, having one to six cylinders — about fourteen times the stroke volume of one cylinder. Engines of about 24" bore, having one to six cylinders — about seven times the stroke volume of one cylinder. Owing to the expensive machinery required to manufacture weldless reservoirs, only a certain limited number of standard sizes are available at reasonable prices, and in very large instal- lations it is sometimes necessary to provide groups of four or more starting vessels. The usual working stress is about Fig. 213. 8000 lb. /in.^, and it is customary to specify a water-test pressure of double the working pressure. Riveted Air Reservoirs. — For marine installations of high power it is usual to use a lower air pressure of about 300 lb. /in.^ for starting purposes. The air reservoirs now require to have a very much larger cubic capacity, but the reduced pressure permits of the employment of riveted reservoirs. The con- struction of the latter need not be dealt with here, being com- parable with that of the steam drums of modern water-tube boilers. Adequate drainaps for condensed water and oil vapour, and also a manhole for inspection and cleaning, should be provided. These matters, as well as others dealing with the strength of the riveted joints, the quality of material to be used and the tests to be carried out on completion, form the subject-matter of regulations by the various insurance societies and the Board of Trade. Blast Piping System. — From the blast bottle the injection COMPRESSED AIR SYSTEM 255 air passes to a main running along the back of the engine, where it is distributed by short lengths of pipe to the several fuel valves, as in Fig. 214. Where one fuel pump is provided for a number of cylinders the fuel distributors may be made to serve as distributing tee-pieces for the blast air. In marine ^ Fig. 214. engines it is usual to provide a shut-down valve, as in Fig. 215, to each tee-piece, so that the supply to any individual cyhnder may be cut off, to enable the fuel valves to be changed without stopping the engine. In addition, it is sometimes necessary (see Chapter XIII) to provide a valve whereby the whole supply of blast air is automatically cut off when the manoeuvring gear is put into the stop position. The bore of the blast air supply pipe' to each cylinder need not exceed about 2% of the cylinder bore, but is usually greater than this in small engines, to avoid the multiplication of standard sizes of unions. The blast air main may be about 4% of the cylinder bore for any number of cylinders up to about six. The same type of union may be used as has already been - illustrated in Fig. 156 in connection with the fuel system. Other types of union are in use, notably the Admiralty Cone Union, which is also very serviceable*. Fiq. 215. Blast Pipe 256 DIESEL ENGINE DESIGN The Starting Air Pipe System. — With land engines it is quite common to provide one cylinder only with a starting valve when the number of working cylinders does not exceed miiiuuk , from ^/r Bottles Fig. 216. \l.Vl.\^VW Fig. 217. four. With six cylinders and upwards two and sometimes three units are provided with air-starting arrangements. A neat arrangement of the starting pipe is shown in Fig. 216 for a four-cylinder engine. In this case the design of the starting valve is such that air is admitted through a port cast in the side of the cylinder cover. Any arrangement of piping is to be avoided which renders difficult the removal of a cylinder cover, hence the provision of an elbow on the latter. In other designs this elbow is cast integrally with the cover itself. COMPRESSED AIR SYSTEM 257 With marine engines all cylinders are provided with starting valves, to which the air is led through a steel main pipe line running the whole length of the engine. Fig. 217 shows the type of pipe flange most commonly used, the material being steel. The tee-pieces for distribution to the several cylinders may be of cast iron or cast steel. If the former material is used, the design should be very substantial, as in Fig. 218. The pipe lines should be securely clipped to the framework of the engine, otherwise there is liable to be severe vibration, due to the surging of pressure within the pipe. In large slow speed engines the diameter of the starting pipe may be about 0-07 to 0-1 of the cylinder diameter. In small high speed engines it is advisable to give the main distributing pipe a diameter of about 0-15 to 0-17 of the bore, in order to secure rapid acceleration. Starting Valves. — These are usually located in the cylinder cover and operated by cams and levers, in the same way as the other valves. An arrangement less frequently used con- sists of a centralised air distributing box of rotary or other type remote from the cyUnder covers but connected to them by distributing pipes. Loss of compression is obviated by the pro- vision of non-return valves in the ■ cylinder cover. The centralised dis- tributing box may consist of a sleeve rotating in a casing in such a manner that a slot in the sleeve admits air successively to a number of ports com- municating with the several cylinders. In other arrangements a set of cam- operated mushroom valves is used. These schemes have not become com- mon practice and will not be discussed here in further detail.. ^'''- ^l^- A common type of starting valve is shown in Fig. 219. No provision has been made here to prevent leakage past the spindle, and if the latter is a good ground fit in the casing the leakage should not be serious in amount. In large sizes of valve additional tightness may be secured to advantage by 258 DIESEL ENGINE DESIGN the provision of a number of small Ramsbottom rings. It is usual to make the diameter of the piston part of the spindle the same as the smaller diameter of the valve head. The minimum spring compression should be equivalent to the maximum starting air pressure acting on an even area equal to that of the valve seat. The valve casing should be sub- stantially proportioned, to prevent distortion and consequent leakage at the seat or binding of the spindle. It will be noticed that with this design of valve, joints have to be made at A and B simultaneously. There is no practical difficulty about this. Joint A is usually made with a copper or white-metal ring. A slight modification is sometimes made by the introduction of two cone joints, as in Fig. 220. This also works well. Fig. 221 shows a type of starting valve in which the air is led to the top of the valve casing, instead of being introduced through a port in the cylinder cover. A useful type of starting valve, devised by the Burmeister & Wain Company for Diesel Marine Engines, is illustrated diagrammatically in Fig. 222. With this design the valve becomes inoperative when the air pressure is removed and resumes working as soon as the pressure is restored. This results in a great simplification of the manoeuvring gear (see Chapter XIII) by the elimination of mechanism which in some other designs is provided for the purpose of throwing the starting valves out of gear when the fuel is turned on. The desired result is achieved by attaching to the upper end of the valve spindle an air cylinder and piston, kept in constant communication with the air supply by means of holes through the spindle. In the absence of air pressmre, the spring A is sufficiently strong to keep the piston B at the bottom of the cylinder C, thus removing the roUer from the range of opera- tion of the cam D. When pressure air is turned on the piston is forced to the top of the cylinder, and the valve remaias operative so long as the force required to open the valve is less than the difference betweeft the pressure load and the spring load on the piston B. Diameter of Starting Valves. — On theoretical grounds, the necessary diameter of startiag valves would appear to depend on the pressure of the air supply, amongst other things. It so happens, however, that in those cases where a low pressure air system is the most convenient (viz. in large marine instal- lations) the multiplicity of cylinders to which starting air is COMPRESSED AIR SYSTEM 259 II w §• Fig. 220. Fig. 221. Fig. 222. 260 DIESEL ENGINE DESIGN supplied, affords adequate starting torque with a relatively low mean starting pressure in each cylinder. The result is that roughly the same diameters of starting valve are used in either case, i.e. whether a low or high pressure starting system be adopted, typical figures being from about 0-1 of the bore in the case of large engines to 0-13 in small engines. With slow speed land engines it is very desirable to obtain a " fat " starting card, to overcome the inertia of the heavy fly- wheels which are usually necessary. An engine in good work- ing order should start firing in the first or second revolution on starting up cold. It seems probable that in the event of low pressure air being used for such engines, it might be necessary to fit starting valves to aU the cylinders or to make the dia- meter of the latter larger than is customary with the high pressure starting air systems at present in use. M Oil Cyjinder Cocks for changing over From Pneumatic to Hand Operation Hand Pump Oil Tank Air Cylinder From Compressed Air Supply ^m. 223. Air Motors. — In large marine engines the work required to effect reversal of the valve mechanism when going from ahead to astern, or vice versa, is generally too great to be done with sufficient rapidity by a hand gear, except in case of a break- down of the air motor which is usually provided for the purpose . In different designs the air motor takes various forms, of which some are mentioned below : — COMPRESSED AIR SYSTEM 261 (1) (2) A small reciprocating engine (double acting), with two cylinders and cranks arranged at right angles. The arrangement is almost exactly similar to the small auxiliary steam engines used on steamships for the reversing gear or the steering gear. Low pressure air is used, and the air motor is geared down by worm and worm-wheel, so that it makes a considerable number of revolutions for one movement of the reversing gear. A single cylinder, with piston and rod, the reversiag motion being performed in one stroke. This arrange- ment is suitable for high or low pressure air, and in either case the piston-rod must be extended into an oil dashpot cylinder to reduce shock. If the two ends of Fig. 224. the oil cylinder be connected to a hand pump, the latter may be used for reversing, in the event of a failure of the air cylinder. This scheme is illustrated diagrammatically in Fig. 223. The reciprocating motion of the piston-rod may be converted into rotary motion (one complete revolution or more) by a rack and pinion. (3) A rotary engine of the type which is frequently used as a pump in connection with machine tools, motor-cars and other small machines, and which is shown dia- grammatically in Fig. 224. This type of motor is only suitable for low pressures and is arranged to do its work in a considerable number of revolutions by means of worm gearing. Types (1) and (3) would appear to have the advantage of greater adaptability to varying pressures. It is an easy matter 262 DIESEL ENGINE DESIGN to gear the motor down so that it will turn under the lowest air pressures anticipated. At higher pressures wire drawing at the ports prevents the attainment of an undesirably high speed. In type (2) hydraulic leather packings are used, and consider- able care should be taken in the design and the workmanship to eliminate aU unnecessary sources of friction. The strict alignment of the two cylinders and the guided end of the rod deserve special attention. Literature. — ^Ford, J. M., " High Pressure Air Compressors." — Paper read before the Greenock Assoc, of Shipbuilders and Engineers. See Engineering, October 20th, 1916, et seq. CHAPTER XIII VALVE GEAR Gams. — ^With few exceptions the valves are operated by external profile cams made of cast iron chilled and ground on the face. In order to facilitate the removal of valves and cylinder covers, it is usual to arrange the cam-shaft to one side of the cylinders and to transmit motion from the cams to the valves through levers or a combination of levers and push-rods or links. The arrangements in general use give an approximate one to one leverage between cam and valve, and the figures I Intone ■ OSS 0. Fis. 225. given above for the width of cam face are based on this pro- portion. Two forms of cam body for the air and exhaust valves of four stroke non-reversible engines are illustrated in Fig. 225, the dimensions being expressed in terms of the cylinder bore. Fig. 226 shows a combined ahead and astern cam for a large marine engine. The bosses should be bored a hard-driving fit on the cam-shaft, and their lengths should be machined accurately to dimensions, so that the complete group of cams required for one cylinder give correct spacing when driven hard up side by side. The fuel cam has to be of special construction, on account of 263 264 DIESEL ENGINE DESIGN the necessity for precise adjustment of the timing, and a typical form is shown in Fig. 227. The toe-piece is preferably made of hardened steel, but chilled cast iron is sometimes used. Profile of Cams. — ^The design of cam profiles for air, exhaust and scavenge values is a matter of reconciling the claims of the following desiderata : — (1) Rapid and sustained opening. (2) Absence of wear and noise. In slow speed engines the question of noise hardly arises, and wear is easily kept to a reasonable minimum by adequate width of face. For such engines the tangent cam shown in Fig. 228 is suitable. For high speeds a smoother shape, as shown in Fig. 229, is desirable, and the relatively slow opening may be compensated by earlier timing. Such profiles are easily drawn by deciding on some arbitrary smooth curve of roller hft, and J)lotting corresponding positions of the roller with respect to the cam, as in Fig. 230. Some designers are in favour of a sinusoidal form of roller-lift curve. With these smooth profiles peripheral cam speeds of five feet or more per second can be used with quite sweet running. On theoretical grounds the cam profile should be based on the roller clearance circle, as in Fig. 231, but it does not yet appear quite clear whether the procedure has the practical advantages claimed for it. The starting air cams are best given a sudden rise on the opening side to minimise wire-drawing (Fig. 232). Combustion valve cam profiles are a study in themselves, and the final decision rests with the test-bed engineers. The effective period is usually about 48 or 50 crank-shaft degrees, or 24 to 25 cam-shaft degrees. The cam-piece should, how- ever, give a range about 25% in excess of this, after allowing for the normal roller cleara^nce to allow for lost motion in the gear. The tangent prorae shown in Fig. 227 is usually found quite satisfactory, but other shapes are used. In order to avoid noise in two stroke engines, it appears necessary either : — (1) To make the cams of smaller diameter than those of a four stroke engine of the same size, or (2) To provide double-faced cams mounted on a half-speed cam-shaft. VALVE GEAR 265 Fig. 230. Fig. 231. 266 DIESEL ENGINE DESIGN The first procedure is the more usual, but the second would appear to have much in its favour. In one marine design advantage is taken of this arrangement to work the engine on the four stroke cycle at slow speeds. Cam Rollers. — Engines having longitudinally fixed cam- shafts are usually provided with cam rollers of steel case- FlG. 232. Fio. 233. hardened and ground inside and out (Fig. 233) and having a diameter of about one-third that of the corresponding cams. The grooves provided for hand lubrication of the pin should be noted. In marine engines in which reversal of rotation is effected by the provision of ahead and astern cams mounted on a longitudinally movable shaft, the roUers require to be of large diameter (about 60% of the cam diameter), in order that the idle cam may clear the lever, as in Fig. 234. RoUers of this size may be of cast iron bushed with phos- phor bronze. Valve Levers. — A common arrangement of valve levers and lever fulcrum shaft for four stroke land-engines is shown in Fig. 235. The fulcrum brackets are secured to the cylinder cover, and the latter may be lifted complete, with all valves and gear, and re- placed without disturbing the valve settings. With the arrajigement shown it is necessary to lift away the fulcrum shaft and levers before the various valves can be removed Fig. 234. for regrinding. This very slight inconvenience VALVE GEAR 267 is sometimes overcome by means of split levers or by provision of horse-shoe shaped distance collars on the fulcrum shaft, which when removed leave sufficient space to aUow the levers to be moved sideways clear of the valve casings. These devices are desirable in the largest engines only. Referring to Fig. 235 below, it will be noted that the fuel and starting levers are mounted on an eccentric bush A, connected to the handle B. The latter is provided with a spring catch engaging with notches in the fixed disc C, in accordance with the following scheme and the diagram shown in Fig. 236. Top notch. — Running. — Fuel lever in its normal running position. Starting valve roller out of range of cam. Middle notch. — Neutral. — Both fuel and starting valve rollers out of range of cams. Fulcrum Spindle For Valve Levers " Fig. 235. 268 DIESEL ENGINE DESIGN Fuel Needle Valve N> Starting 1 Starting Position Neutral Position (Running Position Fig. 236. VALVE GEAR 269 Bottom notch. — Starting.- -Starting air valve lever in its working position. Fuel valve roller out of range of cam. Sometimes this arrangement is modified, by kejdng the eccentric bush and handle to the fulcrum shaft and allowing the latter to turn in its supports. This scheme is useful when the disposition of the gear is such that an air suction or exhaust lever separates the fuel and start- ing levers. This eccentric mount- ing of levers may also be used ' for exhaust lifting or to remove all the levers out of range of the cams during the axial displace- ment of the cam - shaft of a reversing engine. Certain well-known marine makers of great repute do not consider it necessary to put the fuel valve out of action whilst the engine is running on compressed air, and content them- selves with suspending the supply of fuel to the valves during this period. The levers are generally of cast steel or malleable iron, but good cast iron may be used if the stress is confined to about Fio. Fig. 238. 2000 lb./in.2 Some alternative sections are shown in Fig. 237. The forked end of the lever calls for very httle comment. Type A (Pig. 238) is a good design, but expensive. Type B is very commonly fitted and is open to little objection. Type C is the cheapest existing construction, and if accurate castings (machine moulded) are obtainable the only machining opera- 270 DIESEL ENGINE DESIGN tion required is to driU and reamer the hole for the roller-pin. The two grooves for the taper -pin may be cast. In small engines the tappet-end may consist of a plain boss screwed to receive a hardened tappet-screw and lock-nut, as in Fig. 239. In larger engines the more elaborate arrange- ments shown in Fig. 240 are usually adopted. The bosses of the levers should be bushed with good phosphor bronze and provided with a dustproof oil cup of some description, in order to reduce wear to a minimum. With these precautions the bush should only require renewal at widely distant intervals Glass Hard Fig. 239, Fig. 240. and means of adjustment are unnecessary even in the largest sizes of engines. Strength of Valve Levers. — In four stroke engines the exhaust- valve lever is the most heavily loaded. Although the force required to operate the suction valve is relatively small, it is usual to make the inlet valve lever of the same section as the exhaust lever for the sake of uniformity of appearance, and the same pattern may frequently be used for both. The loads imposed on the tappet-ends of the various levers at the points of valve-opening are given bdlow : — FoTTE Stroke Engines Exhaust Valve. About 45 lb. per sq. in. of exhaust valve area+spring load+inertia of valve. Suction Valve. Spring load +inertia of valve -j-a maximum of about 5 lb. per sq. in. of valve area if the exhaust valve happens to be closing too early, due to excessive roller clearance. VALVE GEAR 271 Starting Valve. 500 lb. per sq. in. of valve area +spring load. Fuel Valve (a) Swedish type. 1000 lb. per sq. in. of needle area+spring load+inertia, all reduced by the leverage employed . (b) Augsburg type. Difference between the spring load and 1000 lb. per sq. in. of needle area at stuffing-box. Two Stroke Engines Scavenge Valves. Spring load less scavenge air pressure into area of valve +inertia of valve. Fuel and Starting Valves. As for four stroke engines. All the above are of course subject to sHght correction for friction. By way of example, the main dimensions of the exhaust valve lever for a 20" four stroke cylinder are calculated below : — Data : Diameter of exhaust valve, 6-5 in. Fulcrum spindle and ex- haust valve lever centres, as in Fig. 241. Pressure load on exhaust valve =0-785 X 6-52 X 45 = 1490 lb. Spring load at, say, 8 lb. per in.^ of valve area =0-785x6-52x8= 2651b. Inertia load, say . . 40 lb. Total load to open ex- haust valve 17951b. Fig. 241. Reaction at fulcrum spindle, about 3600 lb. Bending moment at fulcrum spindle = —. in. lb. Allowing a stress of 6000 Ib./in.^, d^ 3600x6x18 24 10 6000 X 24 d=3in. :2-7 272 DIESEL ENGINE DESIGN Allowing for a bush J" thick and about |" metal at the boss the external diameter of the latter will be 6". Sketching in the approx- imate outhne of the lever, as in Fig. 242, it is seen that at the weakest section AA, the bending moment is about 1800 X 18-5 in. lb., and taking a stress of modulus Z of the section AA Fig. 242. 5000 should be lb. /in. for cast steel, 1800x18-5 5000 the 6-66 in. s If the section AA is approximately T-shaped, as in Fig. then Z=b.t.h. nearly, "h" is 5-75, and therefore 243, 5-75 16 in.2 Fig. 243. which is satisfied by b=2-25 and t =0-515". The lever may be made of approximately uniform strength by tapering towards the -g ends, both in width and depth, as in Fig. 244, whilst the flange and web thicknesses are kept constant. If a double bulb or other section is required, for the sake of appearance, and on casting considerations, it is a simple matter to sketch in such a section approximately equivalent to the simple I-section to which the calculation apphes. Push-rods. — In some designs a push-rod is introduced between the lever and the cam-roUer, as shown diagram- matically in Fig. 199 ante, in order to enable the cam-shaft to be located at a low level. For this purpose bright hoUow shafting, or even black lap - welded steam tubes are suitable, if not too highly stressed. For handy reference in designing such pnsh-rods the follow- ing table, taken from Prof. Goodman's " Mechanics Applied to Engineering," is given for the buckling loads of tubular struts. In using these figures it is advisable to use a factor of safety of Fig. 244. VALVE GEAR 273 not less than about 3 or 4, and in no case to employ stresses exceeding 10,000 Ib./in.* Buckling Stress (Free Ends), Lb. per Sq. In. Ratio l^"g**^- Mild Steel. diameter. 10 59,000 20 42,000 30 29,000 40 20,000 50 14,000 60 10,500 70 8,200 80 6,500 90 5,500 100 4,500 The jointed ends of the rods may be of forged steel bar or malleable cast iron bushed with bronze, as in Fig. 245. Cam-shafts. — ^In modern shops the cam-shaft may be rapidly and cheaply ground to size from black bars. In order to facilitate the driving on of the cams for a multi-cylinder engine the enlarged diameters are usually made of increasing sizes, differing by successive thirty-seconds of an inch, or thereabouts, as shown exaggerated in Fig. 246. The same figure which represents the cam- shaft for a four stroke generating set of three cylinders also shows the method of supporting the shaft by means of a continuous trough with one bearing between each bank of cams. This arrangement has a very neat appearance and makes pro- vision for catching the oil which drips off the cams and rollers. In some designs the cams are allowed to dip into an oil-bath, the level of which is maintained constant by a small pump Fig. 245. 274 DIESEL ENGINE DESIGN provided for the purpose, or by a connection taken from the forced lubrication system. A copious supply of oil to the cams has the advantage of securing quiet running. In other designs the cam-shaft is supported by bearing brackets secured to the cyHnders as in Fig. 247. In this case it is very desirable to fit light cast or sheet iron guards round each bank of cams. The cam-shaft bearings are divided horizontally for adjustment, and the shells may be of cast iron lined with white-metal, solid gun-metal, or in small engines, where the cost of material does not outweigh the advantage of simplicity, of solid die-cast white-metal. Owing to the slow peripheral speed and the intermittent character of the loading. G. 247. grease lubrication by Stauffer boxes is quite adequate, although ring and syphon are more commonly used. A slightly different arrangement is shown in Fig. 248. Here the shaft is supported by a series of cam-troughs, one to each cylinder, each trough having two bearings. The extra rigidity of this arrangement allows of the cam-shaft diameter being reduced below the figure required with the other arrange- VALVE GEAR 275 ments described. This division of the trough into segments is advantageous from the manufacturing point of view as the smaller parts are easier to cast and handle in the shops ; also one pattern serves for engines of any desired number of cylinders. Strength of Cam-shafts. — The size of cam-shaft required for a given engine would appear to depend not so much on the stresses to which it will be subject, as on the rigidity necessary to secure sweet running of the gear. For a four stroke engine the opening of the exhaust valve against the terminal pressure in the cylinder is the severest duty which the cam-shaft is called Fie. 248 upon to perform. The load is applied and released fairly suddenly, and a cam-shaft lacking in torsional and transverse rigidity would undoubtedly be subject to oscillations, which in an acute case would give rise to the following evils :— (1) Noisy action of cams, due to torsional recoil of shaft after each exhaust lift. (2) Interference with the timing of valves (particularly the fuel valves) of cylinders remote from the gearing end of the cam-shaft. (3) Chattering of the gear-wheels by which the shaft is driven. In view of the fact that as shaft diameters are increased the stiffness increases at a greater rate than the strength, it seems just possible that strength considerations might outweigh those of stiffness in very large engines. On the other hand. 276 DIESEL ENGINE DESIGN if angular deflection of the shaft between contiguous cylinders be accepted as the criterion, then considerations of similitude give shafts of diameters bearing a constant ratio to the cylinder bores (or rather exhaust valve diameters) and constant stresses in all sizes if the terminal pressure is always the same. The fact that in practice relatively thinner cam-shafts are used in large engines may be due to the lower terminal pressures obtaining in the cylinders of the latter. The following table shows the approximate diameters of cam-shafts used in practice on four stroke engines of different sizes : — Bore of Cylinder in inches . Diameter of Cam-shaft in inches 6 10 15 20 25 30 U 2i, 2| 3f 3i 4i CamshaFt The above figures hold for any number of cylinders up to four with the cam-shaft drive at one end, or eight with the cam-shaft drive at the centre. For two stroke engines these diameters may be materially reduced on account of the absence of exhaust valves. Average figures for existing practice appear to be about 25% lower than those given above for four stroke engines. The fuel pumps and cylinder lubricating pumps are fre- quently driven off the cam-shaft, but any auxiliary gear, such as circulating pumps, etc., requiring appreciable power are precluded. Cam-shaft Drives. — For non-reversible engines, the spiral drive shown diagram- matically in Fig. 249 is the favourite. This drive comprises the following com- ponents : — (1) 4