276 



NA TURE 



{Jan. 20, 1887 



London and North- Western and the Midland Railways. 

 The Birmingham and Warwick Canal forms one boundary 

 of this plot. The fuel to be used under the boilers is gas 

 obtained from Wilson's eight-hundredweight producers. 

 Eighteen of Lane's water-tube boilers will supply si.\ 

 engines to produce the 6000 horse-power aimed at at first 

 Air is admitted to the furnaces through gridiron sliding 

 shutters, by means of which the supply is hand-regulated. 

 It mixes with the gas in a mi.xing-chamber immediately 

 below the front end of the furnace. The roof of this 

 mixing-chamber is an arch of perforated bricks, and these 

 bricks becoming highly heated the mixed air and gas is 

 raised to a high temperature before being ignited. No 

 special means have so far been considered necessary to 

 prevent risk of lighting back into the mixing-chamber. 

 The production of gas in the producers is controlled by 

 the steam jet blown in at the foot of each. The steam for 

 these jets is supplied from a special donkey boiler. The 

 whole of the steam jets are throttled down by the action 

 of a governor that runs, so to speak, in equilibrium with 

 the air-pressure in the mains. The engine drives a small 

 air-pump, which forces air into one end of a small cylinder, 

 to the other end of which the air from the mains is ad- 

 mitted. If the pressure rises in the mains above standard, 

 the piston of this cylinder is moved, and this movement is 

 communicated by suitable gearing to the throttle-valve 

 regulating the steam jet to the producers. The production 

 of gaS; and therefore the production of heat by its com- 

 bustion under the boilers, are thus automatically regulated 

 in accordance with the requirements, so that the air- 

 pressure in the mains is prevented from varying outside 

 certain narrow limits. In connection with this part of the 

 scheme we may point out that it seems to be a mistake 

 not to throttle the entrance-areas for the air to the fur- 

 naces automatically and simultaneously with the regulation 

 of the gas supply. The chief advantage in using gas 

 instead of solid fuel lies certainly in the power of obtaining 

 perfect combustion by thorough admixture and careful 

 proportioning of air to fuel. This advantage is sacrificed 

 if the air supply is not diminished and increased simul- 

 taneously with, and in the same proportion as, that of gas. 

 We suspect also that it will be found desirable not to rely 

 solely on the throttling of the steam blast as at present 

 intended ; the more direct and rapid action of a throttle- 

 valve between the producer and the boiler-furnace will be 

 highly advantageous, if not necessary. By means of 

 simple mechanical relays, actuated either by the steam or 

 by the compressed air, there can be no difficulty in con- 

 trolling these three sets of throttle-valves by the action of 

 a single governor. 



The steam-pressure is to be 160 pounds per square 

 inch. Each set of three boilers supplies an engine of 

 1000 horse-power. The engine is of the triple-expansion 

 type ; the high-, intermediate-, and low-pressure cylinders 

 having the diameters 20, 30, and 49 inches, and a common 

 stroke of 48 inches. The areas of the three pistons are 

 thus in the ratios i, i\, and 6. The cranks are at 120° to 

 each other. The high-pressure and intermediate cylinders 

 are steam-jacketed at the sides. The low-pressure cylinder 

 is not jacketed, but a novel arrangement of steam-jacket- 

 ing its piston is adopted. The piston is hollow, and 

 steam is led into its interior by a tube which is parallel to 

 the piston-rod, and moves to and fro through a stuffing 

 gland in the cylinder cover, projecting into a larger tube 

 screwed on the gland and supplied with steam direct from 

 the boiler. The argument in favour of this arrangement 

 is that side-jacketing of the low-pressure cylinder involves 

 a large absolute waste of heat that goes towards heating 

 the exhaust-steam as it leaves the cylinder on its way to 

 the condenser ; this loss of heat by the jacket steam being 

 noxious, not only because it is pure waste, but also because 

 it raises the back pressure against the piston. The fresh 

 steam in the hollow piston sweeps over the inside surface 

 of the cylinder just in front of the incoming working 



steam, and thus heats the metal and prevents undue con- 

 densation of the working steam, while it is comparatively 

 inactive in heating the back-pressure steam. In criticism 

 of this argument, it may be remarked that towards the end 

 of the stroke (during the last quarter of the stroke) this 

 piston-jacket surface giving heat to the exhaust-steam is 

 greater than a side-jacket would offer. For three-quarters 

 of the stroke, however, it is less. 



Each cylinder is connected with the fly-wheel shaft by 

 a cross-beam. Over each end of each beam stands a 

 single-acting, air-compressing cylinder of 26 inches dia- 

 meter and 4S-inch stroke. Each engine thus drives six 

 of these air-pumps ; and, since the speed is ninety double 

 strokes per minute, the volumetric capacity of the com- 

 pressors of each engine is close on 8000 cubic feet per 

 minute. 



The pressure in the mains is to be 45 pounds per square 

 inch above the atmosphere, and the delivery-valves are ex- 

 pected to lift a little before three-quarters of the compressor 

 piston-stroke is finished. Thus the volume of air com- 

 pressed to the above pressure delivered per minute 

 by each engine is taken as about 2000 cubic feet. 



The ratio of pressures is iiZ = 4-06. Thus, if the com- 

 147 

 pression curve were isothermal, the valves would lift, as 

 above assumed, at 75 of the stroke. If it were adiabatic, 

 this pressure ratio would correspond to a ratio of final to 

 initial volume of '367, and the valves would lift at '63 of 

 the stroke. If the curve lay exactly mid-way between 

 these two, or were according to the law / oc t' " '=, the 

 ratio of final to initial volume would be ^i, and the valves 

 would lift at '69 of the stroke. In the latter case the 

 volume delivered would be '31 :■■ 8000 = 2480 cubic feet 

 per minute. Calculating simply from the product of this 

 volume by 45 pounds per square inch pressure {i.e. from 

 the work the air could do in an air-engine without clear- 

 ance, without expansion, and without more than atmo- 

 spheric back pressure), this would give about 487 horse- 

 power delivered in the consumer's engines for each engine 

 developing 1000 horsepower at the central station. Two 

 indicator-cards taken from two air-compressing cylinders 

 at Frood Colliery, near Wrexham, give very different 

 results, possibly because one compressor was near the 

 steam-engine cylinder, and was heated by it, while the 

 other was not. The compression-curve from the one 

 cylinder corresponds with the relation pa-. ?' " ' '3', while 

 that from the other corresponds to p x 1' - "'''. The latter 

 curve is thus much steeper even than the adiabatic, and 

 would indicate that the air was actually heated by conduc- 

 tion or radiation during its compression. Such heating 

 could hardly have taken place to such an extent as to 

 account for the above very high index, and the more pro- 

 bable explanation is that the air was steam-laden as it was 

 taken in, and that the extra rise of pressure is really due 

 to that of the steam in the mixture consequent on the rise 

 of the temperature. 



It is desirable to keep down the compression-curve as 

 nearly as possible to the isothermal line, because by doing 

 so the area of the compressor indicator-card, and there- 

 fore the work to be done by the engines, is kept down to 

 its minimum ; whereas no advantage can be derived from 

 the increase of temperature obtained by adiabatic com- 

 pression, because this is rapidly lost by cooling in the 

 pipes long before the air is utilised in the air-engine it 

 drives. It is worth noticing that, because of the air being 

 discharged from the compressors through valves which 

 niitoiiiatically lift when a certain designed pressure is 

 reached, this loss of power due to cooling in the pipes is 

 effected rather by a contraction of vohtme than by a 

 diminution o{ pressure. The decreasing-pressure gradient 

 along the pipes is very small, and is due solely to fric- 

 tional and viscous resistance to the flow, and to variation 

 of velocity consequent on variation of section. 



In order to approximate to isothermal compression, 



