not the only condition to be satisfied. With some kinds of storage the rate at which energy can be taken out of store is unlimited. In other cases it is limited, and then the storage must be so arranged that the rate at which energy can be taken out of store is equal to the difference between the maximum rate at which energy is required and the mean

rate.

Fly-wheel Storage.-All steam engines are provided with fly-wheels which store and re-store part of the energy generated in the cylinders. Let w be the weight of a fly-wheel in tons, its velocity of rim in feet per second; then the total kinetic energy is (approximately enough for the present purpose) 2,240 w (v2/2g). For a range of variation of velocity from v1 to v2 the amount of energy alternately stored and restored is

Suppose a fly-wheel weighs twenty tons, its rim velocity cannot generally exceed fifty feet per second, for reasons of strength. If a 5 per cent. variation of speed is permitted (which is more than is usually allowed), the amount of energy alternately stored and re-stored will be 169,800 foot lbs., or 0.086 h.p. hour, an insignificant quantity compared with the work which would be done by an engine with such a fly-wheel. If, however, the fluctuation of demand for energy occurred in half a minute, the fly-wheel would supply in that time 10-28 h.p., which might have a very useful effect in diminishing variation of speed. The function of a fly-wheel is therefore to meet the fluctuations of demand for energy in very short intervals of time, and it has no sensible effect in regulating the variation of demand and supply over longer intervals.

Gasholder Storage.-The distribution of gas is not strictly a distribution of energy, but only of the means of conveniently obtaining it. But a gas-lighting distribution is analogous to a distribution of energy, and the demand varies nearly as much as in an electrical distribution. The gas engineer is happy in having a convenient and cheap means of storage. Usually about twenty-four hours' supply of gas is stored in the gasholders at a gas-generating station.

Hence the gas-making plant can be worked at an almost uniform rate day and night. Taking 25 c. ft. of gas as capable of yielding one effective h.p. hour of energy, it appears that gasholders cost about 5s. 6d. per effective h.p. hour stored. Mr. Trewby puts the cost of gasholders at a London station at 10,000l. per million cubic feet of gas supplied per day. Taking these gasholders to contain twenty-four hours' supply, and reckoning thirty cubic feet of gas per h.p. hour, the gasholders cost only about 6s. per h.p. hour of storage capacity.

A station supplying one million cubic feet per day, considered as a power station, works virtually at 1,666 average effective h.p. during the whole twenty-four hours. The cost of the gasholders adequate to meet any fluctuation of demand comes to only 61. per average effective h.p. supplied. Allowing 10 per cent. for interest and depreciation, the storage adds about 12s. per effective h.p. to the annual cost of the power.

Hydraulic Storage.--Hydraulic storage will be discussed in another chapter. It is sufficient here to state that in hydraulic systems energy is stored in two ways, by hydraulic accumulators and by reservoirs. In the accumulators the total amount of energy stored is so small that it is only sufficient to meet momentary fluctuations of demand. The limitation is due to the great cost of accumulators, which amounts to something like 300l. per h.p. hour of storage capacity. The accumulator is like the fly-wheel of an engine. By pumping water to an elevated reservoir very large amounts of energy may be stored at not very great cost if local conditions are favourable. The pumped water descending again will re-store the energy by working hydraulic motors. A reservoir on an hydraulic system, like a gasholder on a gas supply, may be made large enough to completely meet all fluctuations of demand, so that the pumps can be worked at a uniform rate throughout the twenty-four hours.

Compressed Air Storage.—In systems for distributing energy by compressed air there is always more or less storage of energy, partly in special receivers, partly in the system of mains. The volume of compressed air, by expanding, gives up energy independently of a supply from the compressing plant. There is a fall of pressure, and there must be some limit, due to

this fall, at which further energy cannot usefully be drawn from store, but must be supplied by the compressors.

For the purpose of diminishing fluctuations of pressure, small reservoirs of a capacity about equal to the air supply in three to five minutes are sufficient. For storage of energy much larger reservoirs are required. A very large air reservoir (400,000 c. ft. capacity) was at one time projected for the Paris compressed air system, but it was never carried out. It was intended that this should fill with water under a head of 260 ft. as the air was drawn off, the water being again driven out when the air supply from the compressors exceeded the demand. In that case the pressure would remain constant. More commonly the reservoirs merely supply part of the air they contain by expansion with diminishing pressure.

In the Portsmouth Dockyard compressed air system, for instance, there are eight air receivers of a total capacity of 18,000 c. ft. The normal (gauge) pressure is 60 lbs. per sq. in. Let p1 = 75 lbs. (absolute); P1 = 15 lbs. ; v1 = 18,000. Then Ρι the work to fill the reservoirs, assuming isothermal compression and neglecting friction, would be—

Pa

144 p, V, loge (P1/P1) =311,100,000 foot lbs.,

or 157 h.p. hours. This agrees sufficiently with a statement by Mr. Corner that the receivers can be filled by the 200 i.h.p. compressing plant in one hour. Not all this work can be recovered by calling on the store in the receivers. Suppose the pressure can be reduced to p2 = 40 lbs. per sq. in. by gauge, or 55 lbs. absolute, before it is too low to work the motors. Then the work recovered would not exceed

144 v1 {p, loge (P1/Pa)—P2 log. (P2/Pa)}

or 63 h.p. hours.

Part of this would of course be lost in inefficient action of motors. Mr. Corner states that about half the machines on the air mains can be worked for about two hours with air drawn from the receivers, the compressors being stopped. This means probably that they are driven at their ordinary intermittent rate of working for two hours. It is obvious that such an amount of storage as this, though it does

not equalise supply and demand over the twenty-four hours, may have an important effect in regularising the working of the compressors, engines and boilers, and may not only be a convenience in permitting stoppage for slight repairs and in other ways, but may greatly reduce the waste of fuel and steam due to variation of demand for power.

Accumulator or Battery Storage. The electrical engineer would be glad to have a means of storage equivalent to a gasholder. For a time it was thought that such an equivalent had been found in the storage battery. The use of such batteries is limited to continuous current systems, and they have besides the practical defects (1) that the maximum rate of discharge is limited, and (2) that about one-fifth of the energy stored is wasted. . Nevertheless they would have been an extremely important factor in electric central station working but for their excessive cost. With a twenty-four hour load-line, such as that of most electric lighting stations, the amount of storage required to enable the generators to work at a uniform rate may be defined thus. The battery must be capable of supplying energy at a rate equal to three times the mean rate of supply for the twenty-four hours. Also it must be capable of storing during one part of the twentyfour hours, and re-storing in the other, about half the whole supply for the twenty-four hours. The cost of storage batteries prohibits their employment on this scale in large stations. Employed in a limited way, they serve some useful ends. In some stations they supply the energy required for ten to thirteen hours out of the twenty-four, during which time the engines are stopped. They diminish the fluctuation of load of the engines during the time in which they are running, storing energy not required in the external circuit. But they do not obviate the necessity for having a varying number of engines at work. Professor Kennedy puts the case well when he says that they enable the station to be shut down for some hours and act as fly-wheels, smoothing the irregularities of supply.' The accumulator battery, however, is inferior to the fly-wheel in the rate at which it will absorb and give out energy to meet momentary fluctuations of demand.

Cost of Accumulator Batteries.—From data given me by Professor Ayrton it appears that eight Epstein cells tested in the laboratory would work at one h.p. and store a charge for two and

a half h.p. hours. The cells cost, without allowance for buildings, insulation or switching arrangements, or for waste of energy, 201. That is, the bare cost of the cells amounts to 201. per h.p. reckoned on their maximum rate of working, or to 81. per h.p. hour stored. Suppose a station working at an average of 500 h.p. The maximum demand in the twenty-four hours would be 2,000 h.p., of which 1,500 would have to be supplied from the battery. The cost of the battery to supply energy at the necessary rate would be 30,000l. During twenty-four hours the quantity of energy supplied would be 12,000 h.p. hours, half of which must be stored. Batteries of sufficient capacity would cost 48,000l. Here the latter condition determines the cost. Taking interest at 5 per cent. and maintenance and depreciation at 12 per cent., the annual cost of the battery would be 8,4007., or nearly 177. per h.p. of average rate of working of the station. This is the bare cost of the cells, without buildings, adjuncts or

reserve.

In a project for lighting Frankfort-on-Main, Mr. Oskar von Miller and Mr. Lindley provided large secondary battery stations. The batteries had a capacity of 11,700 ampère hours, and were capable of supplying a current of 3,500 ampères at 100 volts. The batteries with wood platforms, insulation, &c., were taken to cost 25,1007., and the buildings for them 11,6007. This is equivalent to a capital cost of 231. per h.p. hour of storage capacity, or 781. per h.p. reckoned on the assumed maximum rate of working. The difference between this and the previous calculations is that it includes necessary adjuncts, buildings, and reserve of storage to meet contingencies.

Thermal Storage.-Secondary batteries being too costly as a means of storage, except on a very limited scale, the question arises, Is any other means of storage available in conjunction with steam engines? Some means of hydraulic storage will be considered later: such means are rarely applicable for the storage of steam power. Lately Mr. Druitt Halpin has proposed a system of thermal storage which appears in many respects to meet the conditions required.

Energy is first obtained in steam power stations in the form of heat. Can the heat be directly stored? Heat is a very unprisonable form of energy, escaping through all bodies