Condensate-induced water hammer is the most frequently reported type of water hammer at DOE facilities and was the cause of all of the fatalities and injuries and most of the monetary damage attributed to water hammer. The distribution of water hammer occurrences reported in ORPS is shown in Figure 7.
Figure 7. Types of Water Hammer Occurrences Reported in ORPS
1986 Event at Brookhaven National Laboratory
On October 10, 1986, a condensate-induced water hammer at Brookhaven National Laboratory injured four steamfitters, two of them fatally.4 One of the steamfitters attempted to activate an 8-inch steam line located in a manhole. He noticed that there was no steam in either the steam line or the steam trap assembly and concluded that the steam trap had failed. Steam traps are devices designed to automatically remove condensate (liquid) from steam piping while the steam system is operating in a steady state. Without shutting off the steam supply, he and another steamfitter replaced the trap and left.
Later the first steamfitter, his supervisor, and two other steamfitters returned and found the line held a large amount of condensate. They cracked open a gate valve to drain the condensate into an 8-inch main. They cracked the valve open enough to allow water to pass, but this was too far open to control the sudden movement of steam into the main after all the condensate had been removed. A series of powerful water hammer surges caused the gaskets on two blind flanges in the manhole to fail, releasing hot condensate and steam into the manhole. A photograph of one failed gasket is shown in Figure 8. All four steamfitters suffered external burns and steam inhalation. Two of them died as a result.
(CLICK BELOW FOR LARGE IMAGE)
Figure 8. Failed Gasket on Blind Flange
A Type A Accident Investigation Board determined that the probable cause of the event was a lack of procedures and training, resulting in operational error. Operators had used an in-line gate valve to remove condensate from a steam line under pressure instead of drains installed for that purpose.
The board also cited several management problems. There had been no Operational Readiness Review prior to system activation. Laboratory personnel had not witnessed all the hydrostatic and pressure testing, nor had all test results been submitted, as required by the contract. Documentation for design changes was inadequate. The board also determined that Brookhaven management had not been significantly involved in the activities of the steam shop.
1993 Event at the Hanford Site
On June 7, 1993, a condensate-induced water hammer fatally injured a journeyman power operator at Hanford.5 The operator was in the process of opening a steam system valve that had been closed for 8 months. The line was 840 feet long and the valve was inside a steam pit at the low end of the line, 11 feet below the high end.
The operator cracked open the valve but did not hear any flow in the pipe. He then opened the valve further, so that it was about 50 percent open. Hearing water hammer sounds, he attempted to close the valve but was unable to. A blanked-off 6-inch valve ruptured from the water hammer, releasing steam and water. The operator climbed out of the pit but later died from the effects of inhaling steam.
A Type A Accident Investigation Board determined that the pipe was full of water when the operator opened the valve and allowed steam into the line. They also determined that the resulting water hammer caused a pressure transient, which caused the blanked-off 6-inch valve to rupture.
The board determined that the direct cause of the accident was the operator opening the steam valve too quickly. He apparently found the valve easy to manipulate, and probably opened the valve over 50 percent in 1 to 2 minutes. They also found deficiencies in operational procedures and training, system design, equipment, and management systems. These deficiencies included inadequate training on condensate-induced water hammer, failure to follow procedures, inability to manipulate valves remotely, inadequate configuration management, inadequate attention to and resources for the safety of non-nuclear facilities, and failure to comply with Occupational Safety and Health Administration regulations.
1991 Event at a Georgia Hospital
In June of 1991, a valve gasket blew out in a steam system at a Georgia hospital.6 Operators isolated that section of the line and replaced the gasket. The section was closed for two weeks, allowing condensate to accumulate in the line. After the repair was completed, an operator opened the steam valve at the upstream end of the section. He drove to the other end and started to open the downstream steam valve. He did not open the blow-off valve to remove condensate before he opened the steam valve. Water hammer ruptured the valve before it was 20 percent open, releasing steam and condensate and killing the operator.
Investigators determined that about 1,900 pounds of water had accumulated at the low point in the line adjacent to the repaired valve, where a steam trap had been disconnected. Because the line was cold, the incoming steam condensed quickly, lowering the system pressure and accelerating the steam flow into the section. This swept the accumulated water toward the downstream valve and may have produced a relatively small steam-propelled water slug impact before the operator arrived. About 600 pounds of steam condensed in the cold section of the pipe before equilibrium was reached.
When the downstream valve was opened, the steam on the downstream side rapidly condensed into water on the upstream side. This flow picked up a 75 cubic foot slug of water about 400 feet downstream of the valve. The slug sealed off a steam pocket and accelerated until it hit the valve, causing it to rupture.
Investigators concluded that the accident could have been prevented if the operator had allowed the pipe to warm up first and if he had used the blow-off valve to remove condensate before opening the downstream valve.
1994 Event at the Savannah River Site
On June 16, 1994, a building supervisor opened a manual steam supply valve from the header to heating coils for a tank at Savannah River H-Canyon.7 A control room operator then opened a remotely operated valve downstream of the manual valve, allowing steam to enter the heating coils. About 10 minutes later, the building supervisor heard a hammering sound and saw steam coming from the area. He investigated and found steam leaking from the steam header isolation valve. He notified the operator, who closed the remotely operated valve.
The building supervisor then instructed the crane supervisor to close the building main isolation valve. Operators opened blowdown valves to drain condensate and reduce steam pressure. About 950 gallons of condensate were drained, which stopped the water hammer. The event lasted between 10 and 15 minutes and caused approximately $250,000 in damage to the steam header.
During the subsequent system walk-down, operators discovered that sections of the 10-inch steam header were displaced about 29 inches at a bellows-type expansion joint. A photograph of the displaced expansion joint is shown in Figure 9. Angle-iron pipe supports had failed during the water hammer, allowing the movement. The steam header and expansion joints were permanently deformed. Additionally, the steam header isolation valve gaskets failed and two cooling-water lines were damaged, as were instrument cables and cable trays.
(CLICK BELOW FOR LARGE IMAGE)
Figure 9. Displaced Steam Header at Bellows Expansion Joint
A Type B Accident Investigation Board determined that the direct cause of the accident was the opening of steam supply valves, allowing steam to flow through a water-filled steam header, resulting in condensate-induced water hammer. The board also determined that the header had been full of water because personnel had rendered inoperable the only steam trap for the involved section of the header when they closed a steam trap valve without authorization sometime between May 28 and June 13, 1994. The board determined that the root cause was management failure that allowed the steam system to be operated without adequately trained operators and without technical, administrative, and procedural controls in place that would have assured safe operation. Contributing causes included a lack of understanding of the principles of conduct of operations, failure to recognize unacceptable operating and maintenance procedures, lack of understanding of the fundamentals of steam theory and steam system operations, ineffective training in steam system operations, failure to provide some required procedures and to enforce procedure use, failure to upgrade procedures following the water hammer fatality at Hanford, and lack of an effective lessons-learned program.
Water hammer events continue to occur. The most recently reported DOE water hammer event occurred on October 25, 1998, at the Hanford Site while restoring steam following a three-day steam outage. On July 20, 1998, at the Hanford Site, a water hammer event occurred because an operator failed to open steam line drains before introducing steam into the line, as required by operating procedures.