the line. They were very simple in construction and as they have no stuffing boxes they are supposed to require no attention. As a matter of fact, however, they did not prove to be durable for the particular service to which this line was subjected. The alternate heating and cooling seemed to render the diaphragms brittle and they gave out frequently. The insulation of this line was not very satisfactory; this, added to the fact that the water of condensation was largely lost, which involved a loss of heat estimated at eight to ten per cent in excess of that from radiation, led the board of trustees to consider the reconstruction of the line with a boiler house so located that the water of condensation would return by gravity to a tank or well from which it could be pumped while hot to the boilers. Such a system was constructed in 1894 by Mr. Andrew Harvey of Detroit. In the reconstructed system the method of insultation adopted was, first, a sewer pipe laid concentric with the steam pipe and inside a wooden tube similar to that first described. Heavy expansion joints placed at distances of about 200 feet were used to take up the Fig. 33.-WYKOFF BUILT-UP WOOD TUBING. expansion. This system has now been in use two winters and has proved to be of good mechanical construction. The heat insulating properties are not, however, superior to those of the old system which it replaced. The wooden pipes laid in 1886 and taken out in 1893, having been in use seven years, were in every case badly charred. Weather checks extended from the center to the outside and soil water could easily reach the main pipes. In both systems of pipe construction at Cornell University the question of soil drainage has been neglected with the result that a good deal of trouble and great loss of heat is frequently occasioned by water flowing onto the hot steam pipes. The writer measured at one time the loss of heat in a pipe line at South Plainfield, N. J., which I believe was constructed by one of the honorable members of our society for the purpose of conveying steam for power purposes from a boiler plant to several engines situated at considerable distance from the boiler. In this case the insulation was obtained by using the Wyckoff wood cov ering. This covering consisted of two thicknesses of inch board with tarred paper between, so laid as to break joints and form an octagonal pipe as shown in the sketch. This pipe, in the opinion of the writer is much superior to the solid wood pipe, since it is less likely to check and crack. In this case, however, the Heat-Loss between Boiler House and Engine 3000 lbs. 2000 lbs. 1000 lbs. 85 lbs. 80 lbs. 75 lbs. 100 lbs. 90 lbs. 80 lbs. 70 lbs. 60 lbs. 40° 20° 15° 4% 4 8 12 14 16 18 28 .-DIAGRAM SHOWING RESULTS OF TESTS TO DETERMINE HEAT LOSSES IN UNDERGROUND PIPE AT SOUTH PLAINFIELD, N. J. Fig. 34.-1 pipe was not buried under ground, but the whole was inclosed in a large plank box partly filled with saw-dust and placed on top of the ground. No expansion joints were used nor were collars employed to keep the pipe in the center of the wood casing. The ground in that locality was very swampy, and wet, and it would have been impossible to keep it dry, so the location of the pipe was the best possible under the circumstances. The results of the test showed that the loss of heat was comparatively small, and in this respect the covering was satisfactory. The writer has been informed that the line has since been reconstructed. The result of this test is shown clearly in Fig. 34. Various other systems of insulation have no doubt been used with greater or less success. The writer is of the opinion, however, that whatever system is adopted the important points to be considered are in every case as follows: first, protection of the pipe from surface water (it should be remembered that evaporation of water from the outside of the pipe is accompanied by the condensation of an equal amount of steam); second, provision for taking up the expansion and keeping the pipe line in proper alignment; and third, insulation of the pipe from heat losses. It is the writer's opinion that no matter what system of construction is adopted, means should always be provided for thoroughly draining the soil. The simple slip joint, when made sufficiently heavy and when thoroughly anchored, seems to provide for expansion and contraction very well. It needs, however, occasional inspection and repacking. A swinging joint has been used in some systems with success. On account of the bends in the pipe it produces some additional friction and slight loss of head. It also changes the lateral alignment to a greater or less extent and cannot be erected in as narrow a space as the straight line with expansion joints. The loss of heat depends entirely upon the insulation which is provided for the pipe. With the best quality of insulation the loss can be reduced to about 15 per cent of that from a bare pipe in still air. Tests of various insulating substances as taken from a table in a recent work by the author on "Heating and Ventilating Buildings" gives the results which may be expected in a very convenient form. It should be noted, however, in every case that the loss from a naked pipe is dependent principally upon the temperature of the steam and is independent of the velocity of the steam in the pipe. Under ordinary conditions the condensation loss. from a naked pipe in still air is about one-third of a pound of steam or about 320 heat units. This amount is increased by motion of the air, the results being substantially as shown in Fig. 35. It seems most convenient to refer all steam coverings to the bore of the pipe as a standard, and this method was adopted in the table quoted. It will be noticed that with some coverings having fairly good conducting power the loss of heat increases with increase of thickness. This is due to the fact that an increase of thickness is accompanied by an increase in surface; if the heat passes readily 259 Fig. 35.-LOSS OF HEAT FROM PIPES IN AIR UNDER VARIOUS CONDITIONS. through the covering the amount it gives off will depend largely on the nature of the surface and as this is increased the conducting power also increases. The following gives results of tests of various pipe-coverings made at Sibley College, Cornell University.* Kind of Covering. Naked pipe.. Relative amount of heat transmitted. 100. 15.2 Two layers asbestos paper, I in. hair felt, and canvas cover. Two layers asbestos paper, I in. hair felt. Hair felt sectional covering, asbestos lined. One thickness asbestos board.. Four thicknesses asbestos paper Two layers asbestos paper. Wool felt, asbestos lined. Wool felt with air spaces, asbestos lined. Wool felt, plaster paris lined. Asbestos molded with plaster paris. Asbestos felted, pure long fiber.. Asbestos and sponge. Asbestos and wool felt. Magnesia, molded.. Magnesia, sectional. Mineral wool sectional. Rock wool, fibrous.. Rock wool, felted. Fossil meal, molded, 34 inch thick. Pipe painted with black asphaltum.. Pipe painted with light drab lead paint. Pipe protected by wood tube, 3 feet earth. Pipe protected by Wyckoff tube, 3 feet earth. 15. 17. 18.6 59.4 50.3 77.7 23. I 19.7 25.9 31.8 20. I 18.8 20.8 22.4 18.8 19.3 20.3 20.9 29.7 103.5 106.7 32.0 18.0 The following table translated from Péclets Traite de la Chaleur gives in a general way the amount of heat transmitted through coverings of various kinds and of different thicknesses; the loss from a naked pipe is taken as 100. PERCENTAGE LOSS OF HEAT FOR VARIOUS PIPE COVERINGS. *From Heating and Ventilating Buildings, John Wiley & Sons, New York. |