vary at full load from 2 or 3 per cent., with alternating, to 8 or 9 per cent., with continuous current; but with lower loads the efficiency of the apparatus rapidly falls off, until with alternating current generation, unless the transformers can be switched out, the loss is most excessive. In addition to the loss in the transformers, there is, of course, that due to the resistance of the mains, but this is usually small. This last, of course, varies in the same manner as in the case of low pressure generation, but the loss in the transformers is the reverse, that is to say, the percentage loss is a minimum at full load and heaviest at light loads. Hence, with an ordinary lighting system, the waste is considerable. Although, through the efficiency being highest at full load, the capital cost of the generating plant is kept down, there has to be set against this the fact that transformers having an aggregate capacity equal to the maximum rate of output have to be provided. There is thus on the one side the cost of the low pressure mains, plus a certain increase of capacity of the generating plant, and on the other the cost of much smaller mains, but having more costly insulation, plus the cost of the transforming plaut, together with its switching gear, substation buildings, and attendants.

Up to a radius of a mile from the generating station, the loss of pressure in the low tension feeders can economically be kept down to 10 per cent. at full load, if generating at a pressure of from 400 to 500 volts, and experience shows that up to this distance low pressure is by far the most economical and convenient.

This is, however, by no means the limit of low pressure, a fall of potential double that named and even more being frequently allowed, and a radius of 1 miles and more being served.

If there be too much demand at a radius of 1 miles or upwards, it becomes advisable to adopt a high pressure system if it be required to supply from a single station. Whether a high pressure or an extra high pressure system be required will be determined by the extent of the area.

General considerations show that essentially a low pressure system is the best, if the district be suited for it. It is the simplest, safest, most reliable, and most economical. There are, however, many cases in which it is quite inadmissible, and there is no choice but to resort to high pressure, which is then in the nature of a necessary evil, but an evil it certainly is, as compared with low pressure.

The considerations which govern the choice of system will be dealt with in the next chapter.

Assuming that a high pressure system is unavoidable, it will be necessary to compare the advantages of the various systems so far as relates to economy of copper in the feeders; they have already been examined as to their general merits.

The choice lies between continuous current, and alternating single-phase, two-phase, or three-phase. In comparing these, it is important to define clearly the basis of comparison. For low pressures this may properly be the effective pressure, but for high pressures it should be the maximum pressure, since it is this which is usually considered to determine the breakdown point of the dielectric surrounding the conductors, though evidence on this point is not over plentiful.

Inasmuch as we are here concerned only with high pressures, continuous current having already been shown to possess overwhelming advantages over alternating for low pressure feeders, it is obvious that we have to make the comparison on the maximum pressure.

Here another consideration enters into the problem. The stress on the insulation of the conductors for a given difference of potential between them will depend upon whether any portion of the circuit, and, if any, which, is connected to earth. Thus, in a continuous current feeder operating at 2000 volts, the stress on the insulation to earth will only be 1000 volts, if both conductors are equally insulated, but if one be connected to earth, the stress on the other immediately rises to 2000 volts.

Various considerations point to the desirability of working at maximum stress at all times rather than running the risk of an excess pressure suddenly being put on the insulation, and hence it is usual to run with one pole earthed with high pressure continuous and single-phase alternating currents. With two-phase working, it is the practice to earth the middle conductor, while with three-phase the neutral point is usually connected to earth, thus reducing the stress on the insulation below what it would be if one of the conductors were at earth potential. This can, of course, only be done with star winding.

Inasmuch as practice varies with regard to earthing, and as the earthing cannot be applied in exactly the same manner to all systems, it will be most convenient to compare them on the assumption that all conductors are equally well insulated.

On this basis the proportionate amount of copper required by the various systems, taking continuous current as the basis of comparison, is as follows:

Clearly, then, on the basis of copper alone, continuous current is by far

the most economical, and next to that comes three-phase, while two-phase and single-phase are on the same footing as one another, but are both uneconomical.

The cost of insulation for continuous current will also be the least, since the copper is of the smallest section, and concentric mains can be employed; but the relative economy of three-phase will be somewhat diminished, as compared with two-phase or single-phase, since a three-core cable is more expensive to make than a concentric.

As has already been pointed out, however, there are important differences between these various kinds of currents, and it is impossible to ignore these considerations. For pressures up to which continuous current can be used, it should certainly be adopted, but this limit is reached, probably, at 1500 volts, almost certainly at 2000 volts, for ordinary working.

If the whole supply is to be converted into continuous current for distribution, as is in the Author's opinion the best course, then three-phase is certainly to be preferred to two-phase or single-phase. If, on the contrary, it be determined to distribute alternating current, then the difficulties of balancing the lighting circuits with three-phase current are probably so great as to more than counterbalance the advantage of saving in copper, and two-phase is then to be preferred. In such a case, the lighting supply would be single-phase, a portion being supplied from one of the two phases and a portion from the other, while motors would be supplied from both. Incidentally, the two-phase system admits of old single-phase mains being employed in two-phase work, and, in the case of change of system, this one consideration is usually conclusive in favour of two-phase. Single-phase transmission is undoubtedly the least desirable of any; its only merit being simplicity.

CHAPTER VI.

THE CHOICE OF A SITE FOR THE GENERATING STATION, AND OF THE SYSTEM OF SUPPLY.

THE position of the site and the system of supply to be adopted are so closely connected that it is most difficult to consider them independently. It is usually impossible to choose the site for a generating station with reference to engineering considerations alone; in practice there are most frequently several sites only available, and the balance has to be struck between the advantages and disadvantages of each. It will be well, however, to consider the features which an ideal site should possess, and it can then be seen how nearly an actual site approximates to the ideal.

Similarly, it will be convenient to study what is the best system for a given area on the assumption that the site can be obtained where wanted. The result in practice will be a compromise; but it may be remarked, that wherever the site is obtained it will be possible to select an efficient system, though it may not be ideally perfect.

It is necessary first to distinguish between the two cases of districts embracing large and small areas respectively to be supplied.

For a small area, i.e., one in which the supply can be given at low pressure from a single generating station (see above, p. 30), every effort should be made to secure a site as near as possible to what may be called the centre of gravity of the demand; that is to say, the point in which there is the greatest demand for energy within unit radius. If the area be one which cannot be served from a single station at low pressure, it must first be determined whether one or several generating stations are to be adopted; if several, they must be placed in the densest portions of the area; if a single station working at high pressure, the element of position of densest demand will be of less, but still not of negligible importance. Whatever the system of generation, the following points are material :— (1) Accessibility.

(a) For the delivery of coal and removal of cinders.

(b) For the delivery of heavy machinery and of stores of all

kinds.

с

() For connection to mains.

(d) For officials and workmen. (2) Proximity to water.

(a) For condensing.

(b) For boiler feeding.

(3) Stability of foundations.

(4) Isolation, to secure freedom from causing nuisance by noise, vibration, smoke, dirt, fuel, carting, etc.

(5) Facility for extension.

(1) Accessibility.—(a) Seeing that coal is the usual source of the energy delivered by central stations in this country, it is of primary importance that the price paid should be the lowest possible. For a given town, this will depend chiefly upon the cost of carriage to the works. A surprising difference is often caused by a distance of only a few miles. The most advantageous position is adjoining a railway from which a siding can be taken into the works, for by this means a much larger selection of collieries can be obtained than in any other way. It might be thought that a canal or other waterway would be preferable; but this is not so, for although water carriage is usually cheaper than rail, the number of collieries having rail, but not water, connections, enormously preponderates, and hence the lower rates are discounted by restricted competition. Again, it is usually easier, and therefore cheaper, to unload railway trucks than barges. A further very weighty argument in favour of railway carriage is that there is much less liability to interruption of service-a canal being subject to freezing in winter, and shortness of water in summer in times of drought; while rivers, though less liable to interruption from frost, may become unnavigable in times of flood. If possible, both rail and water carriage should be available.

The removal of cinders may be a most costly item, and it is important to see that the projected site affords conveniences for getting rid of them.

(b) The machinery likely to be used in most central stations in the near future is very bulky and massive, and its first cost will be considerably augmented if ready means of transport do not exist. The same considerations as to carriage of coal do not apply here with equal force. The size of an article that may come by rail is limited by the height of bridges and tunnels, and in some cases it may be much easier to bring the plant by water, or even by road; moreover, since the place where the machinery is made may be situated at a very long distance, the saving effected by the low rates for water carriage may be great. For the delivery of stores it is important that there should be good facilities; the same conditions here obtain as in the case of coal.

(c) Care should be taken to see that there is a clear run, with ample space for the laying of mains. The number of conductors leaving the