The character of the water main, like that of many of the other auxiliaries, is determined primarily by the size of the plant. For small plants probably nothing but cast-iron pipe is ever considered, but for larger plants riveted steel mains have sometimes been used, and it is possible that in the larger sizes they are cheaper for a given strength. One point should, however, always be kept in mind in designing mains, and that is that the water which is pumped through them varies from just above the freezing point to eighty or ninety degrees in some districts. The expansion therefore is considerable and means must be provided to take care of it. At one large furnace plant a large riveted steel main was pulled in two by contraction after being several years in service, and the job of making even a temporary repair was a very difficult one, while it could not be permanently repaired except by having the whole plant shut down, a condition which should not exist for many years at a stretch.

This illustrates a point which it is well to bear in mind in connection with auxiliaries, that while each furnace and its immediate accessories are shut down completely at periods ranging from one to three or four years, the plant as a whole, even though it be not a very large one, is not likely to be shut down so completely as to permit stopping the auxiliaries for many years at a stretch. This fact should be borne in mind in laying out water mains, gas mains, electric lines, and sewer systems, particularly the first and last on account of their great size, their being buried underground, the great expense of doing anything with them, and the vast difficulty of rigging up temporaries to permit repairs to the regular ones.

Returning to the matter of the kind of pipe to be preferred, the cast iron has two very great advantages.

First

Cast iron is less subject to corrosive influences than any of the other ferrous metals, the extent to which it is eaten away being only a fraction of that at which steel is attacked under the same conditions.

Second:

A bell and socket cast-iron pipe is laid with a lead joint every twelve feet, this lead being cast and calked to place on a backing of oakum in a way that makes it virtually an expansion joint, so that the expansion of each section is taken up at its ends. No heavy contraction stresses can arise in this kind of pipe, nor can any distortion occur due to cumulative expansion from one end.

Steel pipe, on the other hand, is rigidly fastened together in one piece, and if special precautions be not taken to prevent, the expansion or contraction is liable to be concentrated at one point and cause the failure of the pipe at that point. For this reason if steel pipe be used as a main water-supply system, well-made expansion joints, preferably made on the principle of a huge stuffing box filled with oakum, hemp or other packing, and set up with a gland, should be inserted at such distances as to localize and limit the expansion and prevent its doing damage.

Where large plants are concerned, it is a matter for very serious consideration whether two mains should not be used instead of one, the two together being large enough to supply the whole plant, and one of them able to furnish an emergency or partial shut-down supply. By the introduction of a comparatively small number of duplicate valves permitting both mains to supply all parts of the plant, and at the same time enabling large sections of either to be cut out, provision can be made for repairs to these important accessories without much difficulty or expense, something virtually impossible when a single main is used.

Furnace Main Connections

The connections from the water main to the circulating system around the furnace are a matter of the greatest importance, and no amount of care, and no reasonable amount of money is too great to secure freedom from failure here.

It has been my misfortune to experience three partial failures of this kind, and while in every case we escaped without serious damage to the furnace, we were forced to desperate expedients, in one case being compelled to pull the tuyeres from the furnace as the emergency supply was insufficient to supply these as well as the coolers and cooling plates; the latter could not be removed without serious delay and so were left in and fed with the minimum quantity of water which would prevent boiling. In this way we managed to prevent the burning of any cooling member while we changed the fitting in the main supply line which had split but still remained in place, although its limit of life was evidently to be measured by minutes rather than hours.

It is my judgment that two independent supply lines should be used, equipped with the modern type of non-return valves close to the furnace main, these valves being bronze-fitted throughout so as to prevent sticking in the hour of need. These supply lines should be of moderate size so that normally some water would flow through each and keep them both active, while in case of trouble with one the other would deliver an emergency supply. The non-return valves in case of breakage would prevent the water delivered by the unbroken main from flowing back out through the one which failed, and are an indispensable feature of such a design.