The American Society of Mechanical Engineers Code for Low Pressure Heating Boilers requires that welding "shall insure a joint of sound metal thoroughly fused and to a thickness in excess of the maximum thickness of the plate." The weld that failed showed practically no fusion of metal and it was not more than onequarter the thickness of the plate. Report of Mr. Nicholas P. Setchkin, National Bureau of Standards, on the faulty welding is attached. 2. Inadequate relief valve.—The American Society of Mechanical Engineers Code requires that each boiler of the size in this plant must have a 12-inch diameter relief valve set to open at pressure not above 30 pounds per square inch and without a shut-off between boiler and relief. The system which failed had one relief valve for both boilers set to open at 100 pounds per square inch pressure located in the connection to expansion tank and with boiler stop valves between the boilers and the relief. The original installation had a 4-inch relief valve which failed about 2 weeks before the explosion and was replaced by a 2-inch relief by one of our operating engineers. See sketch of Boiler Room Piping Detail. 3. Complicated type of automatic heating system control. The automatic control for each individual zone has advantages in fuel saving. However, it also has certain dangers from a safety standpoint. When the temperature in the boiler is low the three-way valves are wide open to flow from the boiler. As the temperature in the boiler rises and the zone flow-water controllers are satisfied, the three-way valves gradually change position until they are almost closed on the flow and almost open on the return. If the stoker control does not operate, the three-way valves will assume a full closed position in the flow main cutting off circulation through the boiler. If this occurred, and it is entirely possible, the boiler water would have become dangerously hot, contributing to the boiler failure. 4. Uninformed operating personnel. It appears from Mr. Huckaba's statements at the conference in Mr. Barber's office that he was not familiar with the details of the system and also was lax in checking the stoker control after Mr. Douglas reported the stoker would not come on. Mr. Huckaba said he "first tested the fuses; they were all right. Then I suspected the building was too warm for the stoker to come on. I turned the aquastat to 180°, about, and I turned it back down to original position. It went off. I decided it was the room temperature that had cut the stoker off." To check the aquastat merely by moving the adjustment both ways without knowing the temperature of the water in the boiler is an incomplete test. If the heat-sensitive element which actuates the aquastat had been out of commission this test would not necessarily have detected the failure. With this element out, there was nothing to stop the stoker from running away. In addition to the aquastat, the stoker was provided with a timer whose function was to start and run the stoker a short time at predetermined intervals regardless of room or boiler water temperature to prevent the fire from becoming too low or going out entirely. With the timer in proper working order it should have come on often enough during Fireman Douglas' shift to eliminate his handfiring of the boiler. When Mr. Douglas reported that the stoker would not come on it is believed that Mr. Huckaba, among other tests, should have determined whether or not the timer was functioning properly, which he did not do. Other possibilities which were considered.-(a) Gas explosion in firebox; (b) low water in boiler; (c) main stop valve on boiler shut. Each of these possible causes is commented on in order: (a) The condition of the tubes, tube sheets, crown sheet, and outer shell after the explosion was conclusive that the explosion was caused by internal pressure. (b) With low water, steam would have been generated for some time before the explosion. This would have caused considerable noise which undoubtedly would have been heard by the fireman in the boiler room and guards in the main building. Fireman Bates testified there was no unusual noise in the system previous to the explosion. (c) Fireman Bates left the boiler room about 5 minutes before the explosion. There was insufficient time for anyone to have closed these valves and escaped in that time. Operating conditions.-The exact operating condition under which the boiler exploded could not be determined. Normally, the boiler was operated under a pressure of about 25 feet of water at gage, equal to a pressure of about 10.8 pounds per square inch. The boiling point of water under this pressure is approximately 241° F. It is obvious that the temperature of water in the boiler reached this point and steam was generated. The expansion of the steam created an increase of pressure in the boiler to a point beyond its holding capacity and caused it to rupture. As the boiler ruptured some of the water was released, and the pressure in the boiler was lowered. Due to this lowering of the pressure much of the entire 3,933 pounds of water in the boiler was immediately converted into steam with the resulting explosion, which loosened an enormous amount of energy that was sufficient to cause damage. This conversion of water in the boiler into steam would take place instantaneously and would cause no noise in the system previous to the explosion. Pertinent testimony: Mr. Dowdell, fireman on duty from 7 a. m. to 3 p. m., stated that about 11 a. m. the altitude gage on the boiler indicated 20 feet. He fed water to the boiler until gage indicated a little over 25 feet. Temperature of water was around 100°. Mr. Douglas relieved Mr. Dowdell and was on duty from 3 p. m. to 11 p. m. He stated the temperature of water was between 25 and 30 (he, no doubt, meant feet instead of temperature). Mr. Huckaba, engineer of Federal Office Building No. 2, stated the altitude gage at 8:30 p. m. indicated about 27 feet. Mr. Bates relieved Mr. Douglas at 10:40. He stated that altitude gage indicated 20 to 25 feet, nearer 25, which he said meant a temperature of water about 110° to 115°. Mr. Douglas stated that around 6:30 or 6:45 p. m. the motor on the stoker shut down and as it was so long coming back he called Federal Office Building No. 2 and asked that one of the operating engineers come over and look at it. Mr. Huckaba, operating engineer at Federal Office Building No. 2, responded to this call and arrived at the boiler house around 8:30 p. m. Mr. Huckaba stated he tested the fuses and found them all right. He turned the aquastat up to 180° and the stoker came on, then turned it back to its original position. This was about 9:15 p. m. Mr. Douglas stated that from the time Mr. Huckaba left until he was relieved by Mr. Bates, about 10:50 p. m., the stoker did not run and he fired by hand. Mr. Bates stated that when he came on duty at 10:50 p. m. the line switch of the stoker motor was open and the fire was low. He closed the switch and the motor started. He then threw the stoker motor on high speed and let it run that way for about 35 minutes, then threw it on low speed in which it was running for about 10 minutes when he left the boiler room with stoker running on low speed. The explosion occurred about 5 minutes after Mr. Bates left the boiler room. Design of heating system: The heating system was designed for a differential of 70° (zero outside and 70° inside). Heat loss was estimated as approximately- 225 gallons per minute=13,500 gallons per hour=112,500 pounds 13,500 square feet radiation, at 150 British thermal units per Stoker and furnace: (Assume over-all efficiency of 65 percent): B. t. u. per hour 2, 000, 000 2, 000, 000 1,995, 000 14,500 British thermal units per pound of coal×0.65=9,425 British thermal units per pound of coal fired. Low speed 132 pounds X9,425=1,244,000 British thermal units per hour÷ 60=20,730 British thermal units per minute. Intermediate 176 pounds 9,425=1,658,700 British thermal units per hour+ 60=27,640 British thermal units per minute. High speed 220 pounds 9,425=2,073,500 British thermal units per hour+ 60=34,570 British thermal units per minute. The full content of the boiler was 3,933 pounds of water, which raised 1° would require 3,933 British thermal units. Temperature rise in boiler with no circulation through the heating system is estimated as At low speed 20,730÷3933-5.3° per minute. At intermediate speed 27,640÷3,933-7° per minute. At high speed 34,570+3,933=8.9° per minute. While the heating system was designed for 0° outside to 70° inside, when the explosion occurred the outside temperature was 20°. Theoretically, with a differential of 50° (70-20) between the inside and the outside temperatures the heat transmitted from the boiler to the radiation to compensate for heat loss from building would be 71 percent of 2,000,000 British thermal units per hour, which is 1,420,000 British thermal units per hour, equal to 24,000 British thermal units per minute. Allowing 30 minutes of high speed of stoker, during which time the heat transmitted to water in the boiler was 34,570 British thermal units per minute, and assuming that 24,000 British thermal units per minute was dissipated by the water circulated through the radiation, there would remain in the boiler 34,570-24,000=10,570 British thermal units per minute minus 81° rise in 30 minutes. It is not possible without knowing the temperature of the water leaving the boiler to determine the quantity circulated through the boiler and the quantity circulated through the three-way valves. Roughly, as assumed above, there was transmitted to the radiators a total of 24,000 British thermal units per minute from the boiler, and there was a differential of 17.7° between incoming and outgoing water at the boiler, there would be 24,000 (8% 17.7)=162 gallons of water pumped through the boiler and 225-162-63 gallons pumped through the three-way valves per minute. These quantities would change with a higher or lower temperature of water leaving the boiler which would also change the assumed differential of 17.7° between incoming and outgoing water at the boiler. The amount of water pumped through the boiler and the differential between incoming and outgoing water is important only in estimating the time required for the temperature of water in the boiler to reach a dangerous point. However, it is obvious that the temperature of the water leaving the boiler reached a point so high that the automatic controls on the heating system practically closed circulation from the boiler altogether. Assuming this to be the case the 34,570 British thermal units per minute delivered to the boiler would be absorbed by the 3,933 pounds of water in the boiler, and would raise the temperature of this water approximately 9° per minute. Conclusion: In the absence of evidence to the contrary, the theory that seems most plausible is that the water circulation in the boiler was greatly reduced by the automatic controls on the three-way mixing valves, or was entirely stopped by the shut-down of the circulating pump. The failure of the stoker control to limit the water temperature resulted in the pressure increasing until the boiler burst. The bursting of the boiler lowered the pressure and the water immediately turned to steam. Following this line of reasoning, the relief valve was of inadequate discharing capacity and unable to relieve the pressure at the high temperature which, undoubtedly, had been reached when the boiler exploded. Boiler inspection: While the boiler was structurally weak and may have failed at any time, had it been provided with a relief valve in accordance with the American Society of Mechanical Engineers Code, the failure would not have been nearly so destructive. The inadequate relief valve would have been readily detected by a competent inspector. Recommendations: The Board concurs in the recommendations in the report, dated March 29, 1944, submitted by Mr. John A. Dickinson, Chief, Section of Safety Codes, Bureau of Standards, which follow: "1. That firemen in charge of heating boilers be given a simple training course covering the reading of gages and thermometers. It should stress the importance of watching water temperatures on hot-water heating boilers. "2. That operating engineers be given detailed instruction on the functioning, maintenance, and adjustment of all temperature-control equipment under their jurisdiction. "3. That each hot-water boiler be equipped with a large-size thermometer which can be read easily from the firing floor. (Thermometer on damaged boiler was not over 3 inches in length and was located 8 feet or more above floor.) "4. That the use of stoker equipment with automatic temperature-control systems be checked carefully to make sure that the automatic valves cannot shut off the delivery of all hot water from the boiler. With coal fires much heat can be delivered to the water in the boiler after the stoker has shut off. "5. That each boiler be provided with a relief valve of adequate size attached to the boiler without intervening valves. "6. That a complete piping diagram be made of all boiler layouts with all valves properly located and their functions described. This is particularly important where existing plants are taken over by the Branch and new personnel is placed in charge of equipment." The Board also recommends that all heating plants and all domestic hotwater-supply heaters be inspected periodically, at least once a year, by a competent mechanical engineer to see that all plants are in first-class operating condition. He should give special attention to automatic controls and the relief valves. He should consult the operating personnel to make certain they understand fully the operation of the plant and what to do in an emergency. Domestic water heaters are included in this recommendation because a boiler or heater containing steam and water, or water only, possesses considerable more potential energy and can cause much more damage in case of an explosion than a boiler containing steam only. Mr. Gordon Barber, manager of Virginia district, Public Buildings Administration, Mr. John A. Dickinson, and Mr. Nicholas F. Setchkin, of the National Bureau of Standards, were of great assistance to the Board in making this investigation, and their cooperation is appreciated. Enclosures: Reports of the following-Roy D. Hopgood, inspector of the guard, Virginia district; L. C. O'Neal, sergeant of the guard, Columbia Pike group; M. E. Graves, guard, 219 North Lee Street; C. D. Kline, guard, 219 North Lee Street; William Bates, fireman, 219 North Lee Street; George W. Giddens, assistant fire marshal, Protection Division; William P. McDonald, superintendent, Columbia Pike group; C. A. Kizer, chief engineer, Columbia Pike group; Thomas H. Douglas, fireman, 219 North Lee Street; J. W. Troy, operating engineer, Columbia Pike group; Roland Huckaba, operating engineer, Columbia Pike group; Mr. John A. Dickinson, National Bureau of Standards; and Mr. N. P. Setchkin, National Bureau of Standards. Also: Notes from meeting in Mr. Gordon Barber's office; nine photographs showing results of explosion; and print of boiler room piping detail. O PETER J. FURLONG, Office of Supervising Architect, Office of Supervising Engineer, Office of Buildings Manager, Public Buildings Administration. |