An isolated departure from routine procedures caused the Oct. 23, 1989, explosions and fire that leveled Phillips 66 Co.'s Pasadena, Tex., polyethylene plant and killed 23 people, a Phillips study has concluded.
Other possible causes-such as flaws in design, construction, or materials quality or inadequate safety procedures-were not implicated, concluded a study of the accident by a special Phillips Petroleum Co. committee.
Based on its findings, Phillips is contesting all citations issued by the Occupational Safety & Health Administration against the company for alleged violations of federal safety law and regulations.
In April, OSHA charged Phillips with 556 willful violations-one for each plant employee-for fines totaling $5,666,200 (OGJ, Apr. 30, 1990, p. 46). Phillips also disputes legality of OSHA's method for determining the number of violations.
THE CAUSE
Phillips determined the explosions originated with a major release of hydrocarbons through an open 8 in. Demco ball valve in the No. 4 settling leg of reactor 6 in polyethylene Plant V.
The Demco valve was designed to isolate the leg and other equipment downstream from the reactor during maintenance on the leg. Procedure calls for the Demco to be closed and locked out and connecting air hoses removed.
"It has been established by statement of employees who worked in the polyethylene Plant V on shifts preceding the Oct. 23 day shift that, according to procedures, the Demco valve on the No. 4 leg of Plant V, reactor 6, was closed, the air lines to the cylinders that operated the Demco were disconnected, and the settling leg and transfer lines to the flash chamber were ready for maintenance," Phillips said.
"The lockout and air line disconnection had been performed on the preceding Saturday, but because of work priorities, maintenance did not begin on this leg until Monday, Oct. 23."
Phillips said, however, that examination of the evidence indicated the lockout device had been removed and the air hoses reconnected to the valve operator on the No. 4 leg Demco valve. Further, the valve and settling leg were exposed at the bottom of the leg, where a swedge spool leading to the product takeoff valve should have been connected. Because block valves for air lines to the Demco and connecting piping were damaged in the explosion and moved, their relative positions at the time of the accident could not be determined, Phillips said.
The only survivors in the immediate area of the accident were Fish Engineering & Construction Inc. employees performing maintenance work at the time. Fish has long been a service contractor for Phillips 66.
QUESTIONS REMAIN
Phillips was unable to interview Fish employees regarding the exact sequence of events leading to the hydrocarbon release.
But, the company said, available evidence indicated either of these scenarios:
- The lockout device was removed from the Demco valve, the air lines were connected, and the air line block valve was opened with the settling leg open to the atmosphere.
- The lockout device was removed from the Demco, air lines were reconnected, the air line block valve was opened while the leg was closed-and the leg was then opened without first relocking the Demco, closing the air line block valves, and removing the air lines.
While not speculating about who violated the procedure or why, Phillips said, "Either of these actions would have been a serious violation of well established and well understood procedures and would have created the conditions that permitted the release and subsequent explosion.
"There are several explanations for the sequence of events that actually opened the Demco valve, but the inability to interview those believed to have been in the area, plus the damage caused by the release and the explosion have prevented a determination of which sequence actually occurred.
"it is the understanding of the committee that the OSHA representatives took samples from the remnants of air lines that had been reconnected to the Demco and that a report from an FBI laboratory indicated that the lines were connected in the reverse order."
If that is the case, said Phillips, it would reduce the number of possible explanations for the valve being opened. Regardless of the sequence of events or whether the lines were connected incorrectly, activating the valve would not have resulted in a hydrocarbon release had proper procedures been followed, the company said.
OSHA DISPUTED
Phillips disputed what it said is an implication in a OSHA report that the Demco valve could have been opened inadvertently, constituting an extremely hazardous situation.
Before the Demco valve can be opened, Phillips said, the following must occur:
- The two air hoses must be reconnected to the pneumatic device that operates the Demco valve.
- The air line block valves must be opened to permit air to flow through the hoses.
- The mechanical stop, or lockout device, on the valve must be removed.
Phillips contends there is no plausible reason why a worker deliberately would circumvent the three tiers of protection provided by the established maintenance procedure for working on the reactor leg.
"In interviews with (Phillips) employees familiar with working on the reactor leg, it was evident that these individuals were familiar with the proper procedure and were aware of the serious consequences that could result if this procedure was not followed," the company said.
Phillips also took issue with allegations before the U.S. House government operations subcommittee on employment and housing.
"At least part of the allegations were misleading if not false. For example, the statement was made that 'the Phillips Petroleum lockout policy is inconsistent with OSHA standard 1910.147 and good industry practices.' This standard was not in effect at the time. It had been published in the Federal Register on Sept. 1, 1989, to become effective on Oct. 31, 1989, eight days after the accident. It actually became effective on Jan. 2, 1990."
RECORD DEFENDED
Phillips defended its record for safety procedures and process designs, noting that 74 reactors worldwide employing the same design as the Pasadena reactor have operated for more than 1,300 reactor years without an incident like the Oct. 23 blast.
In addition, a design hazards review and risk assessment by a major engineering company of a Phillips polyethylene plant of the same basic design showed the plant to be at least equal to or superior to accepted industry safety practices, Phillips said.
"Of greatest concern," Phillips said, "is OSHA's allegation that the company willfully endangered the safety and well-being of the people working at the Houston Chemical Complex. This simply was not and never will be the case."
The company said that since 1980, when it strengthened its company wide safety program, only five employee fatalities have been reported in Texas, Louisiana, Oklahoma, and Arkansas-four as a result of industrial accidents and one in a vehicle accident off company premises. Further, HCC workers reported no injuries resulting in lost work days in 1987-88.
"One death is too many," Phillips said. "Nevertheless, the fact remains that no matter how redundant the safety devices or how elaborate the safety procedures, they all can be defeated if they are not employed properly."
Copyright 1990 Oil & Gas Journal. All Rights Reserved.
