UNITED STATES DEPARTMENT OF LABOR
MINE SAFETY AND HEALTH ADMINISTRATION
METAL AND NONMETAL MINE SAFETY AND HEALTH
REPORT OF INVESTIGATION
Surface Metal Mine
Nonfatal Exploding Vessels Accident
July 5, 1999
Kaiser Aluminum and Chemical Corporation
Gramercy, St. James Parish, Louisiana
ID No. 16-00352
Lee D. Ratliff
Assistant District Manager
C. Okey Reitter
Assistant District Manager
John K. Radomsky
Mine Safety & Health Specialist
Clete R. Stephan
Principal Mining Engineer
Terence M. Taylor
Arlie B. Massey
Derrick M. Tjernlund
Fire Protection Engineer
Mine Safety and Health Administration
4015 Wilson Boulevard
Arlington, Va. 22203
Earnest C. Teaster, Jr.
Administrator for Metal and Nonmetal Mine Safety and Health
The Alumina Production Process
Pressure Relief Safety Systems
Plant Electrical System
INVESTIGATION OF THE ACCIDENT
Cause of the Power Failure
Impact of the Power Failure on Plant Operations
Impact of the Ineffective Pressure Relief Systems
Pressure Transmitter System
Pressure Relief Valve System
Blocked Pipe Between Blow-off Tank and Relief Tank
Blow-off Tank Temperatures on July 5, 1999
Impact of Inadequate Training of Plant Personnel
On July 5, 1999, at approximately 5:17 a.m.(1), an explosion occurred at the Gramercy Works Plant ("the Plant") operated by Kaiser Aluminum and Chemical Corporation ("Kaiser"; "the Company") in Gramercy, Louisiana, injuring 29 persons. The cause of the explosion was excessive pressure in several tanks in the digestion area of the Plant.
An unplanned electrical power failure occurred at the Plant at about 4:43 a.m., approximately 34 minutes before the explosion. This caused all the Plant's electrically-powered process machinery to stop, including the pumps that moved slurry and liquor through the digestion process, causing the flow and pressure to back up through the tanks. Further contributing to increasing pressure in the digestion area was the fact that the Plant's gas-fired boilers continued to deliver high-pressure steam to the pressure vessels in the digestion area. The Plant's system of relief piping and relief valves failed to control the increasing vessel pressures. Many of the relief valves failed to function because they had been impermissibly blocked out by Plant personnel. Further, some of the relief piping was clogged with scale, limiting the piping's ability to relieve pressure in the digestion process. Additionally, excessive accumulation of scale in the overflow pipe that connected the blow-off tank to the relief tank severely reduced the pipe's flow area, impeding its ability to relieve the increasing pressure in the vessels in the digestion area.
Management's failure to identify hazardous conditions and unsafe practices and to initiate actions to correct these conditions and practices contributed to the explosion. In addition, because of the inexperience of the employees on site at the time of the power failure and the employees' lack of training in procedures to be followed in the event of a power failure, they were unable to take timely action to prevent the explosion.
The Gramercy Works, a surface alumina operation owned and operated by Kaiser, is located on Airline Highway near Gramercy, St. James Parish, Louisiana. The Plant began production in 1958. The Plant normally operated two, 12-hour shifts per day, 7 days a week, and employed a total of 442 persons. On June 21, 1999, MSHA began a regular inspection of the Plant, which was still in progress at the time of the explosion.
Kaiser's principal operating officials at the Plant were William Kirsch, General Manager; Donald Philips, Assistant General Manager; Glen Lynagh, Engineering and Strategic Projects Manager; Seymour Brown, Operating Manager-Sections I and II; and George Guelfo, Operating Manager-Section III.
On September 30, 1998, the hourly workforce at the Plant, represented by the United Steelworkers of America ("the Steelworkers"; "the Union"; "the USWA"), went on strike after negotiations failed to produce a new contract with Kaiser. Before that date, Kaiser had contracted with Harmony Corporation ("Harmony") of Baton Rouge, Louisiana, to provide the Plant with operations personnel, maintenance personnel, and craftsmen in the event of a strike. As a result of this arrangement, Kaiser was able to continue to operate the Plant without interruption after the strike began, with Kaiser management providing on-the-job training for the replacement workers provided by Harmony.
On January 14, 1999, the USWA notified Kaiser that its members were offering to return to work while contract negotiations between the Company and the Union continued. That same day, Kaiser informed the Union that it was locking out the Steelworkers. Kaiser thereafter continued to operate the Plant with a workforce comprised of replacement workers and management personnel. Harmony Corporation employed 120 persons at the Plant at the time of the explosion.
The Alumina Production Process
The Plant processed bauxite ore into alumina using the "Bayer Process", which involves the caustic leaching of bauxite at elevated temperature and pressure. The bauxite ore processed at the Plant was extracted in Jamaica and transported by ship to Kaiser's Mississippi River dock.
The first step in the alumina production process was the sizing of the bauxite ore by a "Bradford Breaker". The ore was then fed into "pug mills", which mixed the ore with a sodium hydroxide solution, referred to as "caustic liquor", along with slaked lime to form a slurry.
A majority of the slurry was then pumped into an enclosed pressure vessel, referred to as the "desilicator", which chemically removed unwanted silica from the slurry. In the desilicator, 650-degree Fahrenheit (F) steam was added and the slurry was agitated at high pressure to remove the silica. The slurry then moved from the desilicator into a series of four large pressure vessels, called "digesters". High-temperature steam and caustic liquor were added to the slurry in the digesters, and the pressure was increased. The slurry then passed from the last of the four digesters through a series of nine pressure vessels called "flash tanks", where the temperature and pressure were gradually reduced. Finally, the slurry moved to a "blow-off tank", where the pressure of the slurry was reduced to atmospheric levels. The blow-off tank was connected by an overflow pipe to an open-topped relief tank, which relieved excess vapor pressure in the system and received excess slurry from the blow-off tank.
Vapor pipes connected to the top of each flash tank carried steam from the tanks to heat exchangers, where the steam was used to reheat caustic liquor that was recirculated from the leaching process. The heat exchangers also collected the water that condensed from the steam and recycled it to the power house boilers.
A diagram of the main slurry flow in the digestion area is included as Appendix A.
The remaining slurry was pumped from the pug mills to an unpressurized atmospheric "pretreater," where caustic liquor was added and the slurry agitated. This slurry, referred to as "Sweeta," was pumped from the pretreater directly into Flash Tank No. 5 or 6, bypassing the high-temperature, high-pressure processing in the desilicator and digesters. This short-circuited process increased the Plant's daily alumina production, although it recovered a smaller percentage of the available alumina from the bauxite ore than did the high-temperature, high-pressure processing in the digesters.
The slurry exited the blow-off tank through a blow-off tank pump and moved into the "clarification" area, where the waste from the digestion process suspended in the slurry (called "red mud") was separated out. The resulting solution, which contained the dissolved alumina component (called "green liquor"), was pumped to the "precipitation" area of the Plant. Here, the green liquor was cooled and treated with aluminum trihydrate particles. This treatment caused formation of a crystalline structure that assisted in the precipitation of alumina. The larger aluminum trihydrate particles were then washed and filtered before being fed into three kilns, where the particles were heated to produce aluminum oxide. The remaining caustic liquor was recirculated in a closed loop, fortified with sodium hydroxide solution, and reheated by the heat exchangers before it reentered the digestion process.
The Plant processed about 9,000 tons of bauxite ore each day, producing 3,200 tons of alumina.
Pressure Relief Safety Systems
Two pressure relief safety systems were in place in the digestion area to protect the desilicator, the digesters, and the flash tanks from excessive pressure levels. These systems included a pressure transmitter system for the desilicator and digesters, and a pressure relief valve system for the flash tanks.
The pressure transmitter system was designed to automatically stop the flow of steam and slurry into the desilicator and digesters when the pressure level in any of the vessels reached the vessel's trip settings. Kaiser did not furnish MSHA investigators with the specific vessel trip settings. Key-operated selector switches, located in the digestion area control room, were designed to permit the control room operator to bypass the operation of the pressure transmitters on any vessel that was down for maintenance. However, Kaiser sometimes also used these switches to defeat the operation of the system and to allow steam and slurry to continue to flow into the desilicator and digesters, even though one of more of the vessels was experiencing overpressure. This allowed the digestion process and alumina production to continue even though pressure levels in a vessel or vessels were at or above the vessel trip settings.
The pressure relief safety system for the flash tanks consisted of spring-loaded pressure relief valves connected through manifolds. Each flash tank was protected by multiple relief valves that were designed to discharge steam into a network of relief piping leading to the relief tank.
The relief valve manifolds were connected directly to Flash Tank Nos. 1, 4, 5, and 7. These same manifolds were connected indirectly through vapor lines to Tank Nos. 2, 3, 6, 8, and 9.
The relief valve manifolds protecting Flash Tank No. 1 also protected Flash Tank Nos. 2 and 3. Eight relief valves were arranged in pairs on four separate manifolds, and a ninth relief valve was arranged on a separate manifold. According to Kaiser documents, eight relief valves were required to be in service at all times to adequately protect the three tanks from overpressure.
The relief valves protecting Flash Tank No. 4 did not protect any other flash tanks. There were a total of four valves on two manifolds, with all four valves required to be in service for adequate protection of the tank, according to Kaiser documents.
The relief valve manifolds that protected Flash Tank No. 5 also protected Tank No. 6. This included a total of six valves (two each on three separate manifolds), although, according to Kaiser documents, seven valves were required to be in service for adequate protection of the two tanks.
The relief valve manifolds protecting Flash Tank No. 7 also protected Tank Nos. 8 and 9. A total of seven relief valves were provided on three manifolds. Four relief valves were required to be in service for adequate protection of the three tanks, according to Kaiser documents.
When a relief manifold was undergoing maintenance, the blocking valves leading into the manifold were closed to isolate the manifold. The blocking valves also allowed the relief manifolds to protect the other tanks in a bank when one of the tanks was out of service for maintenance.
Blocking valves upstream of each manifold (between the flash tank and the manifold) were manually operated. For Flash Tank Nos. 1 through 6, each relief valve also had a downstream in-line blocking valve for total isolation of each relief valve. Downstream blocking valves (from the manifold to the pipe leading to the relief tank) were not provided on the relief valves protecting Flash Tank Nos. 7, 8, and 9.
There was no instrumentation in the digestion area control room that indicated whether blocking valves were open or closed or which flash tanks were connected directly to the relief system at any given moment. Blocking valve closures were supposed to be tracked through daily check sheets completed by the digestion supervisor and the outside digestion operators, and through the hanging of unlabeled white tags on closed valves.
Plant Electrical System
The Plant produced its own electricity. The operating voltage was distributed to Plant motors at either 2,400 or 480 volts. A detailed description of the Plant's electrical system is included as Appendix B.
DESCRIPTION OF THE ACCIDENT
On July 4, 1999, the second shift began at its regular starting time of 6:00 p.m. Fifty-three Kaiser employees and 20 Harmony employees worked on the second shift. Walter Hansley supervised the crew working in the digestion area of the Plant on the second shift. David Steele was the second-shift digestion control room operator, and Terrence Hayes, Albert Jones, and Wayne Robins were the outside digestion operators.
Because the crew on the preceding shift had experienced problems with excess pressures in the digestion process, the slurry flow had been reduced from a normal flow of about 8,000 gallons per minute (GPM) to 5,600 GPM. The Plant continued to operate at this flow rate throughout the second shift.
At about 4:30 a.m. on July 5, Hayes, Jones, and Robins entered the change room to prepare to take showers, their normal practice at the end of the shift. Hansley had left the digestion area to make the shift relief report and to exchange information with Reuben Cole, the digestion supervisor who would be relieving him. Steele remained in the digestion control room, monitoring the instrumentation. At about 4:43 a.m., all areas of the Plant except the power house experienced a complete power failure, caused by an electrical fault in Switch House No. 2.
When the power failed, Hayes, Jones, and Robins returned to the digestion control room, where Steele was attempting to page Hansley. The instruments in the control room, which were battery-powered, continued to operate during the power failure, and several alarms on the control board had activated.
The power failure caused all of the electrically-powered pumps in the digestion area to stop, consequently stopping the movement of slurry through the digestion process. However, the gas-fired boilers in the power house continued to produce steam, which began to build up in the desilicator and digesters. To address this situation, Steele reduced the flow of steam to the digestion area and also closed the back pressure control valve, which prevented additional slurry from moving from the digesters to the flash tanks. Steele stated he did this because slurry levels and pressure levels had exceeded their operating limits in many of the vessels.
As a result of the power failure, the electrically-powered blow-off tank pump was not operating, and slurry was therefore not moving from the blow-off tank to the clarification area. The power failure also stopped the circulation of liquor through the heat exchangers, causing excessive heat to remain in the flash tanks. Finally, the blow-off tank overflow pipe was severely clogged with scale, which greatly restricted the relief of pressure from the blow-off tank into the relief tank. As a result of these developments, pressure built up in the desilicator, the four digesters, the nine flash tanks, and the blow-off tank. A chart of the pressure levels measured in the flash tanks is included as Appendix C.
As Hansley was leaving the Plant he observed that the Plant's lights had gone out. To determine what the problem was, Hansley returned to the digestion control room, arriving there about 10 minutes after the power outage. Upon arriving at the control room, Hansley took over operation of the digestion control board from Steele. Hansley immediately turned the switches to bypass the overpressure protection system for the desilicator and digesters, and he opened the back pressure control valve. This was consistent with Kaiser's policy that in the event of a power failure, the digestion control operator should slowly introduce steam to keep the slurry moving through the system. Otherwise, stationary slurry would solidify in the pressure vessels in the digestion area in as little as 30 minutes. The vessels would then need to be cleaned out, which would require the Plant to be shut down and production halted until the hardened slurry was removed.
Hansley also instructed Hayes, Jones, and Robins to go to the heater deck to close the vent valves for the heat exchangers, in order to prevent caustic slurry from being sprayed over the area. Kaiser did not provide these employees with special protective equipment or clothing for performing this task. The three men retrieved two flashlights from their lockers and traveled to the heater deck to manually close the valves. As they were closing the valves, they were sprayed with caustic slurry. They rinsed the material from their faces and necks and returned to the control room, arriving there about 15 minutes later. Cole directed the three men to return to the heater deck because they had failed to close one of the vent valves, which was spewing slurry. The three men returned to the heater deck, located the open valve, and were debating whether it was possible to close the valve when the explosion occurred.
After the explosion, Hansley, Cole and Steele left the control room and eventually began searching for other employees.
The force of the explosion blew Hayes' goggles off and sprayed caustic slurry into his eyes. Because he was unable to open his eyes, he decided to remain in one place. Hansley eventually found Hayes and took him to a safety shower in one of the Plant's office buildings. Hayes received serious chemical and thermal burns to his eyes, and is now legally blind.
Jones was knocked down by the force of the explosion, and sprayed with caustic slurry. Unable to determine his location, Jones stayed in place and yelled for help until Steele found him and took him to the same shower where Hansley had taken Hayes. Jones received third degree burns to his head, neck, left forearm, and left calf.
Robins' safety goggles were blown off during the explosion. He crawled along the floor of the heater deck until he found a stairway and descended to ground level. Cole eventually found Robins and took him to the same shower where Hayes and Jones were located. Because the shower's water pressure was low, only one person could rinse off at a time. Robins received third degree burns to his face, legs, and groin area.
Hayes, Jones and Robins initially were transported by ambulance to a nearby hospital. All three were subsequently moved to a hospital that specialized in burn treatment.
At approximately 5:15 a.m., immediately before the explosion, Gary Guy, digestion area supervisor for the second shift, and Todd Landry, leadman, were in Switch House No. 1, attempting to locate the electrical fault that had caused the power failure. They were joined by Ken Hymel, Jr., who was scheduled to relieve Guy as digestion area supervisor on the next shift. Hymel was discussing the electrical fault by telephone with Matt Matthews, senior power house process engineer, when the explosion occurred. Hymel was blown out of Switch House No. 1 and sprayed with caustic slurry. Chris Dufour, power house boiler operator, Ellen Gauthier and Jerome Bougere, outside power house operators, and Earl Veal, generator operator, found Hymel wandering behind the power house. They took Hymel to a safety shower, where they helped him rinse off. Hymel suffered no burn injuries, but was treated for abrasions.
The force of the explosion also blew Landry out of Switch House No. 1. He was knocked to the ground and was covered with caustic slurry. He crawled away from the switch house as he searched for help. Dufour found Landry about 50 yards west of the power house, and took him to the same shower Hymel used. Dufour helped Landry shower and took him to the front office to wait for an ambulance. Landry received third degree burns to his arms, legs and trunk, affecting more than 80 percent of his body.
Guy was also blown out of Switch House No. 1 and covered with caustic slurry. He crawled about 30 yards, eventually saw a light, and called for help. Two employees helped Guy into a pick-up truck and drove him to the front office area. Guy was taken to a safety shower and rinsed before being transported to the hospital. Guy received third degree burns to his arms, legs, and trunk, affecting almost 50 percent of his body, and lost functional use of his right hand.
Kelly Duffy, maintenance supervisor, was completing paperwork in the maintenance trailer when the explosion occurred. The explosion threw Duffy against the wall and propelled a length of angle iron through the trailer, striking him. Duffy was eventually found by Marcus Wilson, Harmony welder, and Travis Revere, Harmony laborer, who helped Duffy into a pick-up truck and brought him to the front office area, where he was taken by ambulance to a nearby hospital. Duffy suffered fractures to his ankle, lower leg and ribs.
Fourteen additional employees received injuries that required medical treatment but not hospitalization. Appendix D contains a list of the persons whose injuries were required to be reported to MSHA under the reporting requirements found at 30 Code of Federal Regulations (CFR) Part 50. Nine other employees received minor injuries requiring first aid that did not meet the reporting criteria of 30 CFR Part 50.
The July 5th explosion caused catastrophic destruction to the digestion area. The forces of the explosion released more than 400,000 pounds of sodium hydroxide into the atmosphere. Residue settled on homes, buildings, and vehicles in the towns of Gramercy and Lutcher, located approximately 3 miles away. The concussion from the explosion also caused glass breakage and minor damage to structures in these towns. Several residents of the surrounding community complained of respiratory problems in the aftermath of the explosion.
The explosion blew portions of the blow-off tank and Flash Tank Nos. 6, 7, 8 and 9 hundreds of feet from the digestion area. The dome of Flash Tank No. 6, which weighed approximately 7,600 pounds, was propelled about 3,000 feet from its original location. The explosion also propelled sections of steel pipe, structural supports, and valves hundreds of feet in several different directions. The concussion from the blast wave and flying debris damaged equipment and structures throughout the Plant, including Switch Houses No. 1, 3, and 4; the maintenance trailer; Sections I and II office buildings; the power house; several nearby substations; the maintenance warehouse; and the digestion control room. An aerial photograph of the digestion area after the explosion is included as Appendix E. Photographs of digestion vessels before and after the explosion are included as Appendix F.
Initiation of MSHA's Investigation
MSHA was first informed of the explosion at 6:15 a.m. on July 5, 1999, by a telephone call from a locked-out employee of the Plant.
Kaiser still had not notified MSHA of the accident at 7:30 a.m., when MSHA Inspector James Bussell arrived at the accident scene. MSHA began its investigation that same day. MSHA issued an order under Section 103(k) of the Federal Mine Safety and Health Act of 1977 ("Mine Act"), restricting access to hazardous areas to ensure the safety of persons in the Plant. A list of the persons who participated in the investigation is included as Appendix G.
Members of the Hazardous Materials Team ("the Haz Mat Team") of the Louisiana State Police arrived at the Plant on July 5th and secured the accident site. The Haz Mat Team was responsible for directing the clean up of the caustic slurry that was released in the explosion, to minimize the impact of the slurry on the environment. The Haz Mat Team completed its work and left the Plant on July 9, 1999.
MSHA's accident investigation team conducted a physical inspection of the accident site, interviewed people with knowledge of the circumstances surrounding the accident, and collected documents relevant to the operation and maintenance of the Plant and the training of personnel. Because of the impact of the explosion on the neighboring community, two community representatives, as well as two Kaiser representatives, two statutorily-designated "representatives of miners", and two representatives of the locked-out Steelworkers accompanied MSHA's investigation team during some phases of its physical investigation of the Plant. Each of these groups was also permitted to take photographs of the digestion area of the Plant.
MSHA also collected physical evidence for subsequent testing, including several relief valves, an electric bus (a rigid connection point for the distribution of electrical power) that was involved in the electrical fault, and various sections of failed tanks. Executive abstracts of these tests are provided in Appendices H, I, and J, respectively.
Section 103(b) of the Mine Act authorizes MSHA to hold public hearings and subpoena witnesses and documents as part of its investigation of mine accidents. On August 6, 1999, MSHA published a notice in the Federal Register that it would hold a public hearing into the July 5th accident. In the notice, MSHA announced that the purposes of the public hearing were to: (1) determine the cause(s), including possible contributory causes, of the explosion; (2) identify and develop corrective actions, procedures and strategies to prevent the occurrence of similar accidents; and (3) determine whether federal safety standards had been violated in relation to the explosion. MSHA invited representatives of Kaiser, miners' representatives at the Plant, and any other interested persons to identify potential witnesses and to suggest questions that the MSHA panel might ask witnesses.
A prehearing conference was held on September 3, 1999, in Convent, Louisiana. The public hearing was held at the St. James Parish Courthouse in Convent, during the period of September 8-17, 1999, at which time the hearing was temporarily recessed so that legal issues that had arisen could be resolved by the United States District Court for the Eastern District of Louisiana. After the issuance of a decision by the U.S. District Court on October 1, 1999, the public hearing reconvened on October 12, and continued through October 19. The hearing is currently in recess until further notice.
All questioning at the public hearing was conducted by a panel of MSHA investigators. The public hearing was chaired by Tony Oppegard, Advisor to the Assistant Secretary of Labor for Mine Safety and Health. A list of the public hearing panel is included as Appendix K. MSHA subpoenaed employees, including Harmony employees, who worked at the Plant at the time of the explosion, members of Kaiser's mine management, and employees who had worked in the Plant's digestion area prior to the commencement of the strike in September, 1998. MSHA also subpoenaed one of Kaiser's contract consultants. MSHA invited Kaiser to provide an explanation at the hearing of the causes of the accident, based on the company's investigation, but Kaiser declined to do so.
During the course of the public hearing, MSHA heard testimony from 25 witnesses, including 2 members of management. MSHA subpoenaed 30 Kaiser officials to testify at the public hearing, but 28 of these company officials refused to testify, invoking their privilege against self incrimination, under the 5th Amendment to the United States Constitution. Pursuant to an order of the United States District Court for the Eastern District of Louisiana, Kaiser's management personnel did not have to physically appear at the public hearing in order to invoke their 5th Amendment rights. Rather, witnesses were permitted to submit a letter to MSHA, indicating that they were invoking their 5th Amendment privilege and refusing to testify.
A list of the persons who testified at the public hearing or provided statements to MSHA is included as Appendix L.
In addition to testimonial evidence, the MSHA panel subpoenaed documentary evidence from Kaiser relevant to the cause of the explosion. In response to this subpoena, Kaiser indicated that it was unable to locate a number of documents that were otherwise routinely kept for June and July, 1999, which MSHA believes would have helped determine the cause of the accident. The missing documents include valve check sheets, which should identify the valves that were blocked out and for how long; supervisors' daily written reports, which should indicate any unusual conditions; and the reports of examinations to identify hazardous conditions.
Kaiser also withheld certain of the documents subpoenaed by MSHA, asserting that they were covered by evidentiary privileges. The United States District Court for the Eastern District of Louisiana rejected Kaiser's assertions of privilege, and ordered the company to provide the documents to MSHA. Kaiser has appealed this ruling of the U.S. District Court to the United States Court of Appeals for the 5th Circuit.
Since the 1970's, MSHA and its predecessor agency, the Mining Enforcement and Safety Administration (MESA), have enforced the provisions of the federal mine safety and health law at all alumina processing facilities, such as the Gramercy Works Plant. In 1975, a federal district court upheld MESA's assertion of jurisdiction over an alumina processing operation. Aluminum Company of America v. Rogers C. B. Morton, (D.D.C. No. 74-1290, November 4, 1975) (unpublished).
In 1977, Congress specifically included operations engaged in "mineral milling" within the definition of "mine" found in Section 3(h) of the Federal Mine Safety and Health Act. A 1979 Interagency Agreement between MSHA and the Occupational Safety and Health Administration ("OSHA") clarified that the alumina refining process constitutes "mineral milling" and that alumina facilities are subject to MSHA jurisdiction.
In addition, an Administrative Law Judge of the Federal Mine Safety and Health Review Commission concluded in a 1981 decision that MSHA had jurisdiction over the entire Gramercy Plant. Kaiser Aluminum & Chemical Corp. v. Secretary of Labor, 3 FMSHRC 2296 (ALJ, 1981). Kaiser did not appeal this decision, and since that time MSHA has regularly inspected the Plant and Kaiser has paid civil penalties assessed by MSHA for violations issued during these inspections. Kaiser did not further question MSHA's exercise of jurisdiction at the Plant until some time after the July 5, 1999, explosion.
After MSHA convened the public hearing into the cause(s) of the Gramercy Plant explosion, Kaiser and some of its individual management officials challenged MSHA's jurisdiction over the Plant in a proceeding before the United States District Court for the Eastern District of Louisiana. The district court rejected Kaiser's claim that MSHA lacked jurisdiction, and that decision has been appealed and is currently pending before the United States Court of Appeals for the 5th Circuit.
The following discussion addresses the key findings of MSHA's accident investigation team and describes the pertinent factors that the team determined caused or contributed to the explosion.
Type of Explosion
MSHA investigators found no flame damage after a thorough examination of the digestion area of the Plant. The investigators therefore concluded that the explosion was not initiated or intensified by combustible materials. The investigators determined that the explosion occurred as a result of a build-up of excessive pressure within a vessel or vessels in the digestion area of the Plant, and the subsequent rupture of the vessel or vessels. Rupture of the vessels exposed the superheated liquid contents to atmospheric pressure resulting in a boiling liquid expanding vapor explosion. At this time, MSHA cannot conclusively determine which of the five failed vessels ruptured first.
Cause of the Power Failure
The electrical fault that occurred at the Plant at approximately 4:43 a.m. on July 5, 1999, was detected by the phase-to-ground relays located in the power house. The relays then caused circuit breakers to trip, de-energizing circuits and resulting in the loss of all power in the Plant except in the power house.
An examination of the auxiliary cabinet in Switch House No. 2 indicated damage to the current transformer leads and the cabinet wall. The examination also showed damage to the phase "A" bus, which is a rigid connection point for further distribution of electrical power. An oval-shaped portion of the bus on the right side, approximately 1-1/2 inches by 2 inches, had been vaporized by an arc. Approximately 2 to 3 inches of the phase "B" and phase "C" current transformer leads had also been vaporized. The cabinet had heat damage that burned a 3-inch by 5-inch area of paint off the wall inside and outside the cabinet. The examination further indicated that the current transformer leads had not been adequately secured to prevent movement. The leads had been attached to the cabinet wall by being tie-wrapped to stick pads. Three of the stick pads had come loose from the wall, permitting the leads to sag down very near the bus, possibly touching it. There was no evidence that any other factors, such as an animal or a misplaced tool, contributed to or caused the fault. For these reasons, the accident investigation team concluded that the leads contacting the bus was the most likely cause of the fault that resulted in the power failure on July 5, 1999.
Several eyewitnesses stated that they believed that the explosion was immediately preceded by lightning strikes near the Plant. To determine whether lightning may have played a role in the accident, MSHA obtained a report from Global Atmospherics, Inc. The report indicates that there were two lightning strikes on July 5, 1999, in the vicinity of the Plant. The first strike occurred at 5:02 a.m., 4.1 miles from the digestion area. The second strike occurred at 5:38 a.m., 4.8 miles from the digestion area. MSHA has concluded that lightning was not a factor in the electrical fault or the explosion, based on the facts that (1) the strikes were several miles from the Plant, and (2) the times of the strikes do not coincide with either the time of the electrical fault (approximately 4:43 a.m.) or the explosion (about 5:17 a.m.).
Impact of the Power Failure on Plant Operations
The power failure caused all of the electrically-powered pumps in the digestion area to stop, consequently stopping the movement of slurry through the digestion process. However, the gas-fired boilers in the power house continued to produce steam, which began to build up in the desilicator and digesters.
The power failure stopped operation of the electrically-powered blow-off tank pump and, as a result, the movement of slurry from the blow-off tank to the clarification area. The power failure also stopped the recirculation of liquor through the heat exchangers, which caused excessive heat to remain in the flash tanks, resulting in a pressure build-up.
Because of these various factors, pressure built up in the desilicator, the four digesters, the nine flash tanks, and the blow-off tank.
Impact of the Ineffective Pressure Relief Systems
The increased pressures in the tanks in the digestion area were not relieved quickly enough because the two pressure relief systems in the digestion area were not functioning properly. The pressure transmitter system that protected the digesters did not function properly because it had been turned off shortly before the explosion. The pressure relief valve system that protected the flash tanks did not function properly because some of the relief valves were blocked out, and also because some of the relief piping had excessive accumulations of scale. Finally, the overflow pipe that connected the blow-off tank to the relief tank was severely clogged with scale, greatly restricting the release of excess pressure from the tanks in the digestion area. The failure of the pressure relief systems is demonstrated by real-time data. These data show that pressure in each of the vessels in the digestion area increased rapidly after the power failure, and that at the time of the explosion, the pressure levels in all of the flash tanks except Tank No. 3 exceeded the pressure relief valve settings. Pressure level measurements for these vessels are included in Appendix M.
Pressure Transmitter System
Steele indicated in his testimony that the selector switches for the pressure transmitter system, regulating steam flow to the desilicator and digesters, were in the "on" position at the time the power failure occurred. Steele also indicated, however, that Hansley changed these switches to the "off" position when he took charge of the digestion control board shortly before the explosion. Switching the system to the "off" position permitted Hansley to introduce steam into the desilicator and the first digester without the possibility of the pressure transmitter system shutting off the steam flow.
Pressure Relief Valve System
Normally, only steam would discharge through the relief valves into the relief system piping. However, spikes in pressure could occur in the digestion process, resulting in the sudden filling of a flash tank with slurry. In such cases, called "upsets," slurry could be ejected through the pressure relief valves into the pressure relief system piping. The slurry would then build up and harden in the valves and piping. Kaiser failed to properly maintain the relief valves and piping, which MSHA's examination revealed had become clogged with hardened slurry. As a result, relief valves would not seat properly, reducing the pressures in the vessels below optimal levels and slowing the digestion process.
Upsets in the digestion process and slurry discharges were not uncommon. In fact, hearing testimony revealed that slurry discharge into the vapor lines was a constant concern of Kaiser's management, because the discharge contaminated the condensed steam that was collected by the heat exchangers.
To keep relief valves from leaking in such situations, Plant personnel routinely isolated the valves by closing the blocking valves around them. This also disabled the valves and had the effect of reducing the protection that could be provided by the pressure relief valve system.
Testimony indicated that the middle bank of relief valves for Flash Tank No. 7 were blocked out for months before the explosion because the valves were relieving at low pressures (in the range of 30 pounds per square inch gauge (PSIG). Documentary evidence also indicates that the blocking out of valves was not an uncommon practice at the Plant. As described earlier, Kaiser reportedly kept daily check sheets that recorded whether the pressure relief valves on the flash tanks were blocked or unblocked. Although the daily check sheets provided by Kaiser for the 12 months prior to the explosion were incomplete, the check sheets that were provided show a regular practice of blocking out relief valves. The most recent dated check sheet provided by Kaiser was of June 7, 1999, which indicated that five out of seven valves were blocked out on the manifold for Flash Tank No. 7, 8, and 9.
Other daily check sheets provided by Kaiser revealed that there were several occasions during the same 12-month period when Flash Tank No. 7 was being operated with all seven of its relief valves blocked. Finally, an August 5, 1999, letter from Kaiser to the United States Coast Guard and the United States Environmental Protection Agency states that the relief valves for Flash Tank
Nos. 6 through 9 were blocked out at the time of the accident. A copy of this letter is included as Appendix N.
Evidence showed that, prior to the explosion, Kaiser's management knew of and condoned the practice of allowing the digestion process to continue, even though pressure in one or more of the vessels exceeded the operating maximum allowable working pressure (MAWP). Real-time data from June 26, 1999, to the time of the explosion indicates that the operating pressure in Flash Tank No. 7 continuously exceeded the tank's MAWP. Further, an internal Kaiser memorandum indicates that Kaiser management had knowledge that Flash Tank No. 7 had been operating above its MAWP. This practice likely reduced the amount of time that Plant employees had to respond to excessive vessel pressures before the pressures reached critical levels.
> Normally open vent valves were provided on the steam side of the flash tank heat exchangers for relieving noncondensible gases, such as air, from the system. Just prior to the explosion, three outside digestion operators were sent to the heater deck to close these vent valves. Testimony indicated that, when these employees arrived, the vents were already discharging and spraying caustic slurry.
In addition to affecting the operation of the relief valves, hardened slurry adversely affected the relief piping itself. Management was aware that slurry within the relief piping had solidified on the piping's inner surfaces. Over time, flow restrictions, or even total plugging of the relief piping from scale, occurred. This further limited the ability of the pressure relief valve system to reduce vessel overpressure.
Examination of the relief piping after the explosion indicated varying amounts of scale build-up on the inside of the piping. The degree of scaling in the relief piping ranged from minimal to extreme, in some cases resulting in total blockage of the piping. Sections of 20-inch diameter relief piping were plugged solid. Design drawings for the Plant reflect that 20-inch diameter piping was used for the pressure relief system for the digesters and for Flash Tank Nos. 1, 2, and 3.
Blocked Pipe Between Blow-off Tank and Relief Tank
The blow-off tank was connected to the relief tank by a 36-inch diameter overflow pipe. Examination of this pipe during the accident investigation showed that scale build up had reduced the diameter of the pipe in at least one area by about 75 percent. This accumulation further impeded the release of slurry and excess pressure from the digestion pressure vessels.
Blow-off Tank Temperatures on July 5, 1999
From 4:00 a.m. until 4:53 a.m. on July 5, 1999, the temperature reading for the blow-off tank remained constant at 223oF. Beginning at 4:54 a.m., the tank temperature began to increase steadily, reaching 304oF, the instrumentation limit, at 5:10 a.m. The increasing temperature in the blow-off tank indicates that the pressure level within the tank also was increasing.
Impact of Inadequate Training of Plant Personnel
Although training records reflect that the employees at the Gramercy Works had nominally received the training required by 30 CFR Part 48, that training was inadequate. (The Accident Investigation Data-Victim Information sheets, which include training information, are included as Appendix O [file is pdf]).
The investigation revealed that because Plant employees had not been adequately trained, many of the employees did not even have a superficial understanding of the operation of the digestion system. These employees were unable to operate the digestion system without constant supervision. Hearing testimony established that both the permanent and replacement employees hired to operate and maintain the digestion system had not been trained to respond to unanticipated power failures. These employees also had not been trained to recognize hazards, such as non-functioning pressure relief valves. They had no knowledge of the manually operated valves that controlled the flow of steam to the pressure vessels in the digestion area, nor were they aware of the maximum operating pressures of each of these vessels.
Adequate training in procedures to be followed in the event of a power failure could have greatly reduced the likelihood of the occurrence of this accident.
Additionally, because Kaiser's supervisors on site at the time of the explosion were unfamiliar with the Plant's electric distribution system, they were unable to identify the location of the electrical fault and quickly restore power to the Plant.
The explosion occurred as a result of excessive pressure in several downstream pressure vessels in the digestion area.
Kaiser demonstrated disregard for the operating pressure limits established for the digestion area pressure vessels. Evidence showed that Kaiser management knew of and condoned the practice of allowing the digestion process to continue operating although pressure in one or more of the vessels exceeded the operating maximum allowable working pressure.
Kaiser management also ignored the accepted industry standard that requires functional pressure relief safety systems to be maintained for the digestion area pressure vessels. Additionally, Kaiser management knew of and condoned the practice of disabling these safety systems in order to operate the pressure vessels at excessive pressures, and thus maintain production.
Prudent engineering practices were not followed in that the integral piping for the pressure relief safety system and the discharge pipe connecting the blow-off tank to the relief tank were not maintained. Kaiser management knew upsets routinely occurred, causing debris to enter these pipes, yet they failed to ensure the pipes were inspected, maintained and free of obstructions.
Kaiser failed to conduct workplace examinations to identify conditions and practices that posed hazards to employees, and did not promptly correct the hazardous conditions and unsafe practices.
Kaiser also failed to provide adequate training for employees on the safety and health aspects and safe operating procedures of their assigned tasks. The failure to train employees in procedures to be followed in the event of a power failure contributed significantly to the accident.
Order No. 7869285 was issued on July 5, 1999, under the provisions of Section 103(k) of the Mine Act:
Digestion Area Main Slurry Flow Diagram
Description of the Plant's Electrical System
Chart of the Pressure in Flash Tanks
List of Persons With Reportable Injuries
Aerial Photograph of Site - Post Explosion
Pre-and Post-Explosion Photographs of Digestion Vessels
Personnel Who Participated in the Investigation
Executive Abstract Pertaining to Relief Valve Testing
Executive Abstract Pertaining to the Bus Testing
Executive Abstract Pertaining to Metallurgical Testing
Public Hearing Panel Members
Persons Who Provided Statements Or Testimony During the Investigation
Recorded Pressure Levels for Vessels in the Digestion Area
August 5, 1999, Letter from Kaiser to the U.S. Coast Guard
APPENDIX O - (version is PDF)
Accident Investigation Data-Victim Information Sheets
Single Line Drawing of Electrical Loop Distribution System
Thickness Measurements of Failed Vessels
Asbestos Fiber Sample Results
1. For purposes of this report, the time of the power failure and the time of the explosion were determined from documentation provided by Louisiana Power and Light.