First thing in the morning

The commissioning process began with the main night shift operator on March 23 carrying out the initial filling of the separation tower. As plant start-ups are particularly prone to unexpected situations, operational practice requires the application of a controlled and approved pre-start-up safety review (PSSR) procedure. BP had one, but it was not adopted in this case. The process control level transmitter was designed to indicate the raffine level at a range of 1.5 m (5 ft) from the bottom of the dividing tower to a level of 2.7 m (9 ft). A high level alarm dependent on this transmitter sounded as scheduled at 3:09 am, when a level of 7.6 feet (2.3 m) from the bottom was reached. However, during start-up it was common to ignore this alarm and fill to a 99% level (as indicated by the transmitter) to avoid damage to the furnace that heated the bottom of the divider. Unknown to the operators, the process control level transmitter, which was used to monitor the level in the divider throughout the start-up operation, was not calibrated and its readings were unreliable. A separate level alarm activated by a high level switch at 2.4 m (7.9 ft) should also have sounded, but did not. At 5:00 a.m., the main operator of the ISOM satellite control room reported to the central control room and went home early. At 6:00 am, the day shift operator (central control room operator) arrived and began his 30th consecutive day of work on a 12-hour shift. At the time, the level was still believed to be below the 9 ft mark, but it was already at 13 ft (4.0 m).

At 7:15 a.m., more than an hour late from the start of his shift (6:00 a.m.), one of the two day shift ("A") supervisors arrived at the central control room. Due to his delay, he was unable to make the necessary relief with the night shift. During the morning meeting on 23 March, it was discussed that the heavy raffine storage tanks were almost full and therefore a second day shift ("B") supervisor was told that the start-up procedure should not continue, but this information was not conveyed. Therefore, the start-up procedure was resumed just before 9:30 a.m. following the instructions of day shift Supervisor A. The separator level control valve on the heavy storage tank was placed in manual mode (rather than regulating its opening via the level transmitter) and set to 4,300 bpd. However, the flow transmitter on this line had an error and the valve had actually closed. This, together with the malfunction of all the instruments in the separation tower, allowed the liquid level in the tower to increase without anyone realizing it and caused the operators to lose all knowledge of the situation.[12] The circulation process was restarted just before 10 in the morning and raffin was reintroduced into the tower, although the level was already too high. As the level control valve was closed and therefore there was no circulation outside the tower (i.e. no heavy raffin was transferred to the storage tank), the separation tower inevitably began to fill. The faulty level transmitter continued to show a level below 100% and as the external sight glass was opaque, a visual check to verify the level in the dividing tower was not possible. Day shift supervisor A, the more experienced of the two, left at 10:50 a.m. due to a family emergency. Contrary to operating regulations, there was no supervisor left in the central control room. A single operator, unsupervised and very tired, had to supervise the ISOM and two other units during a critical start-up procedure.

Late morning

Two furnace burners had been turned on at 9:55 a.m. to preheat the raffinate entering the tower and to heat the raffinate at the bottom of the tower. Two others were lit at 11:16 a.m. The required tower reboiler return flow temperature was 135 °C (275 °F) with a rise of 10 °C (18 °F) per hour, but this procedure was not followed: during startup, the return flow temperature reached 153 °C (307 °F) at a rate of 23 °C (41 °F) per hour. The faulty level transmitter continued to erroneously indicate a safe level condition on the tower. However, there was still no flow of heavy raffin from the separation tower to the storage tank, as the level control valve remained closed; Instead of the hydrocarbon liquid level being at 8.65 feet (2.64 m), or 93% of the instrument's range, as indicated, it had actually reached 67 feet (20 m). Just before noon, with the heat in the tower increasing, the actual liquid level had risen to 98 feet (30 m). Pressure began to build up in the system as hydrocarbon vapors and nitrogen remaining in the tower and associated piping since it had been returned to service were compressed by the increasing volume of refining. Operations staff thought the pressure increase was due to overheating of the tower bottoms, as this was a known startup issue, so the pressure was released.

Burst

At 12:42 pm, the furnaces had been shut down and the level control valve finally opened, draining the heavy raffine from the separation tower. The gas feeding the furnace was turned off, but not the raffin feed to the separation tower. Operators believed the level transmitter reading, which was now 78% (7.9 ft [2.4 m]), but the fluid level in the 170 ft (52 m) high dividing tower had reached 158 ft (48 m).

Although the opening of the heavy raffinate flow should have lowered the level in the column, the fact that this hot flow was used to preheat the feed meant that the temperature inside the separator increased dramatically, causing a significant increase in vaporization and the rise of a slime of liquid above the top of the column to the collector line. By 1:13 p.m., the hydrostatic head of this liquid had increased to over 42 psi (290 kPa), which was enough to open the relief valves. With the relief valves fully open, more than 196 m3 (51,900 US gallons) of heated raffinate passed directly to the collection manifold over a 6-minute period before the valves closed, as the pressure within the splitter header pipe fell below its minimum shut-off set point (37.2 psi (256 kPa)). 256 kPa (37.2 psi) above atmospheric pressure) Hot raffinate flowed into the blowdown drum and stack, and as it filled, some of the fluid began to flow into the ISOM unit sewer system through a 6-inch (15 cm) pipe located at the base of the blowdown drum. As the blowdown drum and chimney filled, hot raffin was shot out the top of the chimney into the air, forming a 6 m "geyser". The equivalent of a nearly tanker truck full of raffin rained onto the ground, ran down the side of the blowdown drum and chimney, and accumulated at the base of the unit. A radio call was received in the control room reporting that the chimney was overflowing with hot hydrocarbons. The plant's evacuation alarm did not sound, which prevented people in the vicinity from being able to evacuate before the ignition occurred.

A diesel truck, with the engine idling, was parked about 8 m from the blowdown chimney. The vapor cloud reached the vehicle and hydrocarbon fumes entered the engine's air intake, causing it to accelerate. Nearby workers frantically tried to shut down the engine, to no avail. The expanding cloud of vapor forced workers who were trying to turn off the engine of the truck that was going too fast to retreat. The cloud continued to spread across the ISOM plant, across the pipe rack to the west and into the unobstructed trailer area. No emergency alarm sounded and, at approximately 1:20 p.m., the vapor cloud was ignited by a backfire that nearby witnesses observed came from the truck's engine overheating.

The congestion of equipment and pipes contributed to accelerating the flame front. This triggered a massive vapor cloud explosion that could be heard for miles. The pressure wave from the explosion hit the group of contractor trailers located just 37 m from the blowdown stack, completely destroying them. The explosion sent debris flying, instantly killing 15 people in the nearest trailers and injuring 180 others.[e] Workers in the trailers were injured up to 150 m (480 ft) from the purge drum, and some trailers were severely damaged up to 180 m (600 ft). More than 40 trailers were damaged.[25] All of the fatalities and many of the injured in the accident were contractors. Fifty storage tanks suffered structural damage, although most of the tank farm was more than 250 feet (76 m) from the explosion. More than 2,750 pounds (1,250 kg) of benzene leaked from one of the damaged tanks.

emergency response

The local emergency response team intervened immediately, mounting a search and rescue operation. A shelter-in-place order was issued for 43,000 people. Mutual aid provided by IMAS (Texas City Industrial Mutual Aid System) and Memorial Hermann Life Flight resources were mobilized at 1:45 p.m.[26] Power to the raffine splitter was not cut off, but stopped at 2:45 p.m. when power was lost. The fires were brought under control by 150-200 firefighters within two hours. The ambulances and lifesaving helicopters left at 4:44 p.m. The last body was found under a pile of rubble around 11:00 p.m.

Contenido

Los expertos internos de BP, así como diversas autoridades y comités, investigaron la explosión en relación con aspectos técnicos, organizativos y de cultura de seguridad. Las investigaciones internas de BP incluyeron un grupo (la investigación Mogford, por el nombre del investigador principal) encargado de reconstruir la cadena causal del accidente y realizar un análisis detallado de las causas subyacentes, y otros dos equipos (las investigaciones Bonse y Stanley) que se centraron en los factores subyacentes de procedimiento y culturales, así como en las responsabilidades de los directivos. BP encargó a un grupo independiente de alto nivel (el grupo Baker), por recomendación urgente de la Junta de Seguridad Química de EE. UU. (CSB), que examinara los problemas de gestión y cultura de la seguridad. La CSB llevó a cabo su propia investigación, detallada y exhaustiva, centrada tanto en aspectos técnicos como de procedimiento.

Los distintos paneles e investigaciones detectaron fallos organizativos, como la reducción de costes por parte de la empresa, la falta de inversiones en la infraestructura de la planta, la falta de supervisión por parte de la empresa tanto de la cultura de seguridad como de los programas de prevención de accidentes graves, la concentración en la seguridad laboral en lugar de en la seguridad del proceso, una gestión defectuosa del proceso de cambio (que permitió la ubicación de remolques de contratistas demasiado cerca de la unidad de proceso ISOM), la formación inadecuada de los operarios, la falta de supervisión competente de las operaciones de puesta en marcha, la mala comunicación entre personas y departamentos y el uso de procedimientos de trabajo obsoletos e ineficaces que a menudo no se seguían. Entre los fallos técnicos cabe citar el uso de un tambor de purga de tamaño insuficiente y anticuado, la falta de mantenimiento preventivo de los sistemas críticos para la seguridad y la inoperatividad de las alarmas y los sensores de nivel de la unidad de proceso ISOM.[27]​.

BP internal investigations

A team of experts led by John Mogford, senior vice president of safety and operations at BP Group, examined the technical aspects of the explosion and proposed corrective measures. An interim report was published on May 12, 2005.[79] In response to the report, Scott Berger, director of the Center for Chemical Process Safety (CCPS) at the American Institute of Chemical Engineers (AIChE), expressed surprise at the report's emphasis on the individual responsibility of plant operators and supervisors.[28]​.

On December 9, 2005, BP published the final review of the Mogford report, which identified four critical factors without which the explosion would not have occurred or would have had a minor impact: “loss of containment; procedures for starting up the raffine separator and applying knowledge and techniques; control of work and location of the trailer; and blowdown stack design and engineering." Additionally, five critical underlying cultural issues were identified:

• - "Business context", which included a work environment that resisted change and was dominated by a lack of motivation, confidence and a sense of purpose. Added to this were unclear expectations regarding management and supervision.

• - "Safety as a priority", since management did not give process safety the necessary importance within the company.

• - "Complexity and organizational capacity", in the sense of a lack of clarity in responsibilities and poor communication.

• - "Inability to see risk", or a tendency to accept high levels of risk due to poor awareness of the dangers.

• - "Lack of early warning", or inability to recognize and act on signs that reveal deterioration of facilities and procedures.

However, Mogford's final report found no evidence that anyone intentionally made decisions or took actions that endangered others.

The company also assembled a team of BP and external experts to conduct an operational and process audit review of the refinery. James W. Stanley, former deputy director of the Occupational Safety and Health Administration (OSHA), was the team leader. The audit focused on processes and operations; incident management, work control, risk assessment and compliance assessment; employee and contractor management; and maintenance, reliability and integrity. The Stanley report was published internally on 15 June 2005. It concluded that a significant change in performance and behavior on the part of senior and extended management would be required, despite well-designed and documented procedures and processes and the workforce being generally capable of good performance. Among the issues that "impeded the successful execution of some key work processes," the team highlighted: leadership factors, such as lack of security accountability and silo mentality, among others; risk awareness, indicated by complacency and repeated failure to heed recommendations resulting from previous accidents; work control measures, which were insufficient and were not complied with; negative workplace conditions, as demonstrated by poor cleaning and insufficient maintenance of facilities; and a contractor management philosophy lacking diversity and inclusion values.

Another internal team was led by Wilhelm Bonse-Geuking [de], who was vice president of the BP Group in Europe. The Bonse team investigated plant managers' compliance with BP's management framework and corporate code of conduct. The final report - which was not made public until a court ordered it on May 3, 2007 - detected numerous management failures[74],[29][30]​ It further noted that management responsibilities within the BP Group were unclear and that the poor condition of the plant's equipment and insufficient maintenance spending were contributing factors to the accident.[30]​.

In summary, the Texas City refinery had a culture of risk-taking coupled with an inability to understand the process safety implications of past incidents [...], a long tradition of non-compliance with simple procedures, a desire to avoid conflict within its organization, and a penchant for placing people in key positions who lacked appropriate professional training.

The report singled out four executives for dismissal: Pat Gower, regional vice president of U.S. refining; Mike Hoffman, group vice president of refining and marketing in North America; Don Parus, Texas City complex director; and Willie Willis, Texas City West plant supervisor.[2] On the date of publication of the report, none had been fired.[6]​.

Baker Panel Report

Following the March explosion, two other serious process safety incidents occurred at the plant:

• - On July 28, 2005, a gaseous hydrogen heat exchanger pipe at the waste hydrotreatment unit ruptured, causing a hydrogen leak that erupted in a large jet of fire. The fire lasted about two hours. One person was slightly injured and property damage amounted to $30 million. The Chemical Safety Board discovered that a contractor had accidentally swapped a low-alloy steel elbow for a carbon steel pipe elbow during maintenance, causing a failure mode known as high-temperature hydrogen attack (HTHA). The CSB found that BP should have required positive verification of the materials using an

• - On August 10, 2005, a diesel hydrotreater suffered a leak caused by high-temperature corrosion and sulfidation that caused the release of toxic gases, such as carbon monoxide (CO), hydrogen sulfide (H2S), and sulfur dioxide (SO2), as well as flammable hydrocarbon vapors.[33] Material damage amounted to approximately $2 million.

Following these events, on August 17, 2005, the US Chemical Safety and Hazard Investigation Board (CSB) issued an urgent recommendation that BP commission an independent group to investigate the safety culture and management systems of BP North America.[34]​ A group of experts was convened led by former US Secretary of State James A. Baker III.[35][36]​ BP specifically singled out one figure. of this stature and resume to publicly demonstrate to US opinion makers that the company was willing to learn its lesson and take steps towards change.

The Baker Panel differed from the BP and CSB investigations in that it was not tasked with conducting a root cause investigation. Its scope was broader, focusing on BP's corporate safety oversight, corporate safety culture, and its process safety management systems, and included all five BP refineries in the United States.

The Baker Panel report was released on January 16, 2007[37]​[38]​ and cited a poor safety culture driven by poor process safety leadership and poor employee training as underlying causes of the accidents that plagued the refinery. It also highlighted that, despite BP's comprehensive safety management system, it was not sufficiently implemented at its refineries. It was also revealed that when measuring its safety performance, BP was focusing primarily on workplace safety indicators, to the detriment of monitoring process safety performance.[39] It was also highlighted that worker fatigue and a system that encouraged overtime had detrimental effects on the safe operation of the plant, and that the company had not addressed deficiencies arising from known incidents, risk assessments and audits. The lack of application of good engineering practices was also noted.

Additionally, the group interviewed several employees, managers and contractors at BP's five U.S. refineries. They also developed and administered a process safety culture survey to the refineries. They concluded that the Toledo and Texas City plants had the worst process safety cultures, while the Cherry Point refinery, located in Birch Bay, Washington, had the best process safety cultures. The survey results also showed that managers and white-collar workers generally had a more positive opinion of the process safety culture in their plants compared to the views of blue-collar operators and maintenance technicians.

CSB report

Given the magnitude of the disaster, the Chemical Safety Board examined both the safety management at the Texas City refinery and the role of the BP group and that of OSHA as a regulatory agency. The CSB investigation team arrived at the scene 48 hours after the accident. Some 13 CSB investigators remained on site for three months. For its investigation, the CSB used a budget of $2.5 million and reviewed more than 30,000 documents, interviewed 370 witnesses, and conducted computer modeling and testing. The results of the agency's investigation were published on March 20, 2007 in a lengthy 341-page report, the largest ever produced by the agency, which was then nine years old. The report's conclusions were presented the same day at a public meeting in Texas City.[41][42]​.

One of the main findings of the CSB was that the purge system used in the ISOM unit was outdated and totally inadequate, as it was located in the middle of the plant and could release heavy unignited vapors into normally occupied areas.[116] The CSB concluded that BP had failed to take into account and implement multiple warnings and safety recommendations issued before the explosion in relation to the ISOM purge system. Among them:

• - In 1991, Amoco's refining planning department proposed eliminating vented blowdown systems, but funding for this plan was not included in the budget.

• - In 1992, OSHA penalized Amoco for the unsafe design of a similar blowdown system elsewhere in the refinery. However, Amoco was able to convince OSHA to withdraw the citation based on the less stringent requirements of API Recommended Practice 521.[43]​.

• - In 1993, the "Amoco Regulatory Cluster Project" proposed eliminating atmospheric purge systems, but again its funding was not approved.

• -

A flare chimney, an inherently safer solution to remove excess flammable gas

Despite Amoco's "Process Safety Standard No. 6," which prohibited the installation of new atmospheric purge systems and required the phasing out of existing ones, in 1997 Amoco replaced the 1950s-era drum chimney/purge vent that served the raffin separation tower with an identical system, rather than upgrading it with recommended alternatives that were safer.

• - In 2002, the opportunity to connect the ISOM relief system to the naphtha desulfurization unit's new flare system was not taken due to an additional cost of $150,000.

• - Also in 2002, BP's "Clean Streams Project" proposed converting the blowdown drum into a flare discharge tank and directing discharges to a flare. When it was discovered that a necessary relief study of the ISOM system had not been completed due to budget constraints, Clean Streams decided not to pursue the option.

Between 1994 and 2004, there were at least eight similar cases in which flammable vapors were emitted from the ISOM drum/purge fan chimney, two of which resulted in fires. Furthermore, the purge drum was not designed to cope with overfilling of the container. This was not necessarily attributable to Amoco or BP, but rather to the lack of guidance in API Recommended Practice 521.[43]​.

Los daños materiales de la explosión ascendieron a 200 millones de dólares (322 millones en 2024)[46]​ La explosión provocó el cierre de varias unidades de la refinería. En previsión de la llegada del huracán Rita en 2005, se cerró toda la refinería. BP se centró entonces en reparar los daños causados tanto por la explosión como por el huracán. El reinicio de las unidades de proceso comenzó en marzo de 2006. Los costes de las reparaciones y la producción aplazada ascendieron a más de 1.000 millones de dólares.[47]​ BP se declaró culpable de delitos federales contra el medio ambiente, por los que pagó 50 millones de dólares.[48]​ La empresa también pagó al menos unos 2.100 millones de dólares en acuerdos civiles.[47]​Además, BP pagó 84,6 millones de dólares y 27 millones de dólares en multas al gobierno federal a petición de la OSHA y la EPA, respectivamente,[49]​[50]​[51]​ y una multa de 50 millones de dólares al gobierno de Texas por infracciones medioambientales.[52]​ El desastre es el accidente de refinería más costoso del mundo.

BP response and fate of the refinery

BP launched a crisis management plan just six hours after the explosion. The next day, a website[53]​ was created to post updated information about the accident. The chief executive, Lord John Browne, visited the plant the day after the explosion.[54] In the months after the accident, BP tended to blame its workers and supervisors, which victims and union leaders saw as simple scapegoating. At the time, the company consistently chose not to publicly apologize for the accident. This changed on May 17, 2005, when Ross Pillari, president of BP Products North America, publicly apologized, saying: "We regret that our mistakes have caused so much suffering. We apologize to those injured and to the community of Texas City" and promised “financial assistance and compensation” to the injured and the families of the deceased.

On December 9, BP announced that it would implement a $1 billion budget over five years to improve the safety of its Texas City refinery.[55] BP also announced that it would remove all bleed drum/vent stack systems in flammable service, of which there were 11 at Texas City, and install new flares in accordance with its new policy prohibiting atmospheric venting of heavier light hydrocarbons. than air.[6]​[56]​They also relocated trailers away from areas where explosions could occur[56]​and initiated a trial of an integrated electronic system of safe work (ISSOW).[56]​ Group-wide initiatives were also adopted.[56]​.

However, shortly after the Texas City explosion and other accidents in 2005, BP's image in the United States was further tarnished by the near-sinking of the Thunder Horse PDQ semi-submersible oil rig in July of that same year[56] and, more seriously, in March 2006, when a pipeline spill was discovered in Prudhoe Bay, Alaska, while multiple investigations into the Texas explosion were still ongoing. City.[57]CSB Chairwoman Carolyn Merritt said there were striking similarities between the Texas City and Prudhoe Bay accidents, including "long delays in implementation, administrative documentation of closure even though no corrective action had actually been taken, or simple noncompliance," as well as "poor communication of lessons learned, excessive decentralization of safety functions, and high management turnover." [58] "Another serious accident occurred in 2007 at BP Texas City, when 143 refinery workers reported that they had been injured by inhaling toxic fumes released at the plant.[59] CEO John Browne resigned in 2007 for issues unrelated to the case,[9] although he was not spared criticism for the lax safety culture and budget cuts at US refineries in BP.[16]​[60]​Tony Hayward took over as head of the company. Hayward changed Lord Browne's emphasis on alternative energy and announced that safety would be the company's "number one priority."[61]​ By then, BP had already adopted a markedly apologetic stance for the recent accidents, especially the one in Texas City, and its executives and technical experts were making presentations about what had gone wrong and how they were working to prevent it from happening again. Then came the Deepwater Horizon explosion and oil spill in 2010, which had a very serious impact on the company globally, again stemming from BP's operations in the US. As a result, Hayward resigned and his position was taken over by US-born Bob Dudley.[62] Under Dudley, BP announced in 2011 that it was selling its Texas City refinery as part of its plan to divestment to pay for ongoing compensation claims and remediation activities following the Deepwater Horizon disaster. The sale of the refinery was successfully completed in early 2013 to Marathon Petroleum Corporation for $2.5 billion.[3]​[63]​Marathon already owned the adjacent Galveston Bay refinery and in 2018 merged both sites into a single refining complex.[64]​.

Two congressional hearings were held specifically on the Texas City disaster. Among those heard were representatives of API, Baker Panel, CCPS, CSB, EPA and USW, as well as relatives of the victims. In other congressional hearings devoted to subsequent BP accidents in the US, the story of the Texas City case was consistently presented within the pattern of degraded safety culture at BP.[65]​.

Agreements with victims

The case of Eva Rowe, a young woman who lost her parents in the explosion, attracted national attention. Rowe said he would not accept a settlement from BP and would take the group to justice. Ed Bradley, a well-known American journalist, broke his story on the television program 60 Minutes.[66] On November 9, 2006, BP settled with Rowe as the final plaintiff, after his lawyers attempted to invite John Browne, BP's CEO at the time of the accident, as a witness. The amount of compensation to Eva Rowe remained unknown. BP also paid $32 million to hospitals and educational and research institutions designated by Rowe, including the Mary Kay O'Connor Process Safety Center at Texas A&M University ($12.5 million), the University of Texas Medical Branch at Galveston and its Truman G. Blocker Adult Burn Unit ($12.5 million), the College of the Mainland in Texas City ($5 million), St. Jude Children's Research Hospital of Memphis, Tennessee ($1 million), and the Hornbeck, Louisiana school system ($1 million).[67]​ In addition, BP was forced to release some seven million pages of internal documents, including the Telos and Bonse reports.[68][69]Rowe would later participate in one of the congressional hearings on the accident.

As of September 2007, BP had settled at least 1,350 of the approximately 3,000 claims related to the accident.[70] By February 2008, some 4,000 claims had been filed, half of which had been settled, totaling more than $1.6 billion.[71] BP also stated that it had set aside another $525 million for other claims.[47] August 2008, only one of the approximately 4,000 lawsuits remained open.[72]​.

Criminal actions

On February 4, 2008, U.S. District Judge Lee Rosenthal heard arguments regarding BP's bid to plead guilty to a federal environmental crime for two violations of the Clean Air Act (CAA)[73] with a fine of $50 million. At the hearing, victims of the explosions and their families objected to the plea, calling the proposed fine "trivial".[74]​ However, the plea was eventually agreed to, along with a three-year probationary period for BP.[48]​ This was the first and, for several years, remained the only federal conviction for an accidental chemical release under the CAA.[75]​.

Fines

In September 2005, the Occupational Safety and Health Administration, which the CSB report described as lacking oversight and jurisdiction, fined BP a record $21 million for committing 301 violations of the Process Safety Management standard.[76][77]In October 2009, OSHA fined $87 million, shattering its record of 2005, after stating that BP had not implemented safety improvements after the disaster and noting that there had been four other fatal accidents at the refinery since the previous fine. In its new report, OSHA cited 709 safety violations.[78][79]​ BP announced that it would challenge the fine.[80]​ On August 12, 2010, BP announced that it had agreed to pay $50.6 million of the 2009 fine, while continuing to challenge the remaining $30.7 million (the fine had been reduced by $6.1 million between the time it was imposed and the time it was imposed). that BP paid the first portion).[81]​ In July 2012, OSHA and BP agreed that the $30.7 million outstanding would be reduced to $13 million, which BP paid.

Following the explosion, the US Environmental Protection Agency inspected the refinery for compliance with the Clean Air Act (CAA) and determined that BP had violated the CAA as well as its Risk Management Program (RMP) standard.[82] Most of the violations were not directly related to the explosion, but to other events, such as the two subsequent accidents in 2005. In February In 2009, the EPA imposed a first fine of $12 million on BP. [51]​In September 2010, EPA and BP reached a $15 million settlement on other violation charges,[33]​ which was the largest civil fine imposed for CAA violations at a single facility and the largest imposed for RMP civil violations.[83]​.

At the request of the Texas Commission on Environmental Quality (TCEQ), the Texas Attorney General opened proceedings against BP for violations of the Texas Health and Safety Code[84]​ and the Texas Water Code,[85]​including the release of hydrocarbons through the blowdown stack on March 23, 2005, the prolonged release of benzene from a tank damaged in the explosion, which lasted more than 25 days, and dozens of others. events,[23]​ In a 2011 settlement, BP agreed to pay a $50 million fine covering 72 other emissions that exceeded BP's operating permits. This included legal costs of $500,000. In exchange, the Department of Justice agreed not to admit additional criminal charges against BP in connection with the refinery explosion.[86]​.

La catástrofe tuvo una notable repercusión en el ámbito de la seguridad de los procesos. Texas City se ha convertido en un caso clásico utilizado para explicar los fallos tanto de gestión como de barreras técnicas en las plantas de procesos.[36]​[87]​[88]​.

El informe del Panel Baker se hizo muy conocido entre los ingenieros de seguridad de procesos, que consideraron que sus conclusiones eran relevantes para otras plantas y que era importante para reforzar la concienciación sobre la seguridad de los procesos en la industria de procesos químicos,[89]​ cumpliendo así el deseo del Panel recogido en el informe:.

• - Aunque necesariamente dirigimos nuestro informe a BP, pretendemos que llegue a un público más amplio. No nos hacemos ilusiones de que las deficiencias en la cultura de seguridad de los procesos, la gestión o la supervisión corporativa se limiten a BP. [90]​.

• - Otras empresas y sus partes interesadas pueden beneficiarse de nuestro trabajo. Instamos a estas empresas a que evalúen periódica y exhaustivamente su cultura de la seguridad, el rendimiento de sus sistemas de gestión de la seguridad de los procesos y su supervisión corporativa de la seguridad para detectar posibles mejoras. También instamos a las mismas empresas a que revisen detenidamente nuestras conclusiones y recomendaciones para aplicarlas a sus situaciones.[91]​[90]​.

• - Análisis de la deficiente aplicación de la gestión de la seguridad de los procesos.

• - El accidente ha sido ampliamente analizado en la literatura especializada, que ha puesto de relieve cómo varios elementos de la gestión de la seguridad de los procesos (PSM) se implantaron de forma deficiente y se gestionaron mal. Entre las observaciones realizadas sobre los elementos de la PSM definidos en la norma de la OSHA[92]​se incluyen:.

• - Participación de los trabajadores: La refinería adolecía del miedo de los trabajadores a ser castigados por notificar incidentes o peligros.[28]​.

• - Análisis de peligros del proceso: el análisis de peligros y operabilidad (HAZOP) de la planta no identificó el escenario de sobrellenado de la columna[90]​ y la evaluación de riesgos para la colocación de edificios temporales no fue completa y, en su mayor parte, errónea.[91]​[90]​.

• - Procedimientos operativos: Las desviaciones de los procedimientos clave, como la puesta en marcha de la planta ISOM, se habían convertido en rutina. Además, el mismo procedimiento de puesta en marcha carecía de instrucciones suficientes[90]​[29]​.

• - Formación: La formación de los operadores clave de la planta era muy insuficiente, especialmente en lo relativo a la gestión de situaciones anómalas y la verificación de los conocimientos y cualificaciones de los operadores.[30]​.

• - Contratistas: La dirección de las empresas contratistas hizo caso omiso de aspectos de seguridad importantes. Por ejemplo, no se informó de la puesta en marcha a quienes ocupaban remolques próximos a la central ISOM. Todas las víctimas mortales y muchos de los heridos del accidente eran contratistas.

• - Revisión de seguridad previa a la puesta en marcha: Esta revisión clave en la preparación de una operación especialmente peligrosa no se llevó a cabo.[93]​.

• - Integridad mecánica: Múltiples instrumentos y válvulas no funcionaban porque no se habían probado o mantenido.[90]​.

• - Gestión del cambio: Esta gestión falló en múltiples aspectos, a saber, los cambios en el procedimiento de puesta en marcha de la unidad, que obligaban a los operarios a hacer funcionar el divisor por encima de un nivel seguro,[94]​la falta de MOC o la existencia de MOC incompletos para la colocación de remolques en las proximidades de una unidad altamente peligrosa,[29]​ y la falta de un proceso de cambio formal que evaluara la eliminación de puestos de operarios críticos para gestionar la planta ISOM.[94]​.

• - Investigación de incidentes: El aprendizaje de incidentes y cuasiaccidentes pasados se vio perjudicado por la práctica ausencia de investigaciones internas y la consiguiente difusión de lecciones aprendidas útiles.[90]​.

• - Planificación y respuesta ante emergencias: No se hizo sonar la alarma de evacuación, lo que puede haber contribuido al número de víctimas mortales, ya que los contratistas que se encontraban en los remolques no tuvieron oportunidad de abandonar la zona.

Auditorías de cumplimiento: Se llevaron a cabo auditorías, pero, según admitió la propia BP, "no parece que se hiciera un seguimiento de los puntos de acción [de la auditoría] ni que se cerraran de forma efectiva".

Analysis of the poor application of process safety management

The accident has been widely analyzed in the specialized literature, which has highlighted how various elements of process safety management (PSM) were poorly implemented and poorly managed. Observations made on the elements of PSM defined in the OSHA standard[1]​ include:

• - Worker participation: The refinery suffered from workers' fear of being punished for reporting incidents or dangers.[90]​.

• - Process hazard analysis: the plant's hazard and operability analysis (HAZOP) did not identify the column overfill scenario[91]​ and the risk assessment for the placement of temporary buildings was not complete and, for the most part, erroneous[43][90]​.

• - Operating procedures: Deviations from key procedures, such as the start-up of the ISOM plant, had become routine. Furthermore, the start-up procedure itself lacked sufficient instructions.[91][90]​.

• - Training: The training of the main plant operators was very insufficient, especially in relation to the management of anomalous situations and the verification of the knowledge and qualifications of the operators.[90]​[91]​.

• - Contractors: The management of the contracting companies ignored important safety aspects. For example, those occupying trailers near the ISOM headquarters were not informed of the start-up. All of the fatalities and many of the injured in the accident were contractors.[65]​.

• - Pre-startup safety review: This key review was not carried out in preparation for a particularly dangerous operation.[95]​.

• - Mechanical integrity: Multiple instruments and valves were not working because they had not been tested or maintained.[4]​[3]​[91]​.

• - Change management: This management failed in multiple aspects, namely, the changes in the unit start-up procedure, which forced operators to operate the separator above a safe level,[94]​the lack of MOC or the existence of incomplete MOC for the placement of trailers in the vicinity of a highly dangerous unit,[3]​ and the lack of a formal change process that evaluated the elimination of operator positions critical to managing the plant ISOM.[5]​.

• - Incident investigation: Learning from past incidents and near misses was harmed by the virtual absence of internal investigations and the consequent dissemination of useful lessons learned.[3]​.

• - Emergency planning and response: The evacuation alarm was not sounded, which may have contributed to the number of fatalities, since the contractors who were in the trailers did not have the opportunity to leave the area.

• - Compliance audits: Audits were carried out but, by BP's own admission, "it does not appear that the [audit] action items were tracked or closed effectively."

New industry guidelines and initiatives

API implemented the CSB report's recommendations, creating new standards and guidelines for the industry:

• - Standard 521: API published a new revision of its document on Pressure Relief and Depressurization Systems and elevated its category from Recommended Practice to Standard. The new standard contained stricter requirements on atmospheric venting of process gas and an instruction to include overfilling of upstream equipment among the design cases of purge systems.[99]​[100]​.

• - Recommended Practice 753: API issued new guidance on locating trailers and portable buildings in hazardous process locations. The scope of Recommended Practice 752 was reduced to permanent buildings only beginning with the third edition.[101][102][103]​.

• - Recommended Practice 754: API addressed the need for process safety performance indicators in this new Recommended Practice on Process Safety Performance Indicators for the Refining and Petrochemical Industries. investigations.[106]​.

• - Recommended Practice 755: This new guideline was aimed at refineries and petrochemical plants and detailed how to implement a fatigue risk management system (FRMS). This document includes recommendations for rotating shift work, for the maximum acceptable number of overtime hours and the number of days that should be worked without interruption.[107]​[108]​[109]​.

The CCPS published extensive guidance on organizational change management to address another related recommendation from the CSB report.[110]​[111]​.

After being singled out for its lack of initiative and competence in the evaluation and inspection of large hazardous process plants, and refineries in particular, OSHA took action, initiating a National Process Safety Management Emphasis Program (NEP) for refineries and implementing a targeted inspection audit program between 2007 and 2011. This was the largest PSM enforcement action since the regulation[1]​ was issued in 1992.[112]​.

OSHA also issued an internal memorandum[113]​ to address the CSB's recommendation to update the PSM standard to include requirements for hazardous process facilities to expand their change management procedures to include organizational changes. However, the CBS did not consider this initiative sufficient to close the recommendation, which, as of January 2024, remained open.[111]​[114]​.

The accident has appeared in several documentaries:

• - National Geographic series Seconds from Disaster, season 3, episode 10 "Texas Oil Explosion", first aired on November 6, 2006.[27]​.

• - The History Channel series Modern Marvels, season 12, episode 56 "Engineering Disasters 20", first aired on December 6, 2006.[28][29]​.

• - The Science Channel series Engineering Catastrophes, season 4, episode 4 "Texas Terror", first aired on July 7, 2021.[30]​.

• - Seconds of the Disaster episodes from National Geographic.

• - Abílio Ramos, Marilia; López Droguett, Enrique; Mosleh, Ali; das Chagas Moura, Márcio; Ramos Martins, Marcelo (August 18, 2017). "Revisiting Past Refinery Accidents from a Human Reliability Analysis Perspective: The BP Texas City and the Chevron Richmond Accidents". The Canadian Journal of Chemical Engineering. 95 (12): 2293–2305. doi:10.1002/cjce.22996. eISSN 1939-019X. ISSN 0008-4034.

• - Barozzi, Marco; Melchiore, Alessandra Derudi, Marco; Copelli, Sabrina (2023). "Integrating Recursive Operability Analysis with Different Risk Assessment Methods: Analysis of the Historical BP American Refinery Explosion" (PDF). Chemical Engineering Transactions. 104: 163–168. doi:10.3303/CET23104028. ISBN 979-12-81206-06-9. ISSN 2283-9216. Archived (PDF) from the original on January 26, 2024. Retrieved on January 26, 2024.

• - ESReDA Project Group Foresight in Safety (2020). Enhancing Safety: The Challenge of Foresight. EUR 30441 EN. Luxembourg, Luxembourg: Publications Office of the European Union. doi:10.2760/814452. ISBN 978-92-76-25189-7. Archived from the original on January 26, 2024. Retrieved on January 26, 2024.

• - Hopkins, Andrew (July 1, 2008). Failure to Learn: The BP Texas City Refinery Disaster. Macquarie Park, NSW: CCH Australia. ISBN 978-1921322440.

• - Manca, Davide; Brambilla, Sara (2012). "Dynamic Simulation of the BP Texas City Refinery Accident". Journal of Loss Prevention in the Process Industries. 25 (6): 950–957. doi:10.1016/j.jlp.2012.05.008. eISSN 1873-3352. ISSN 0950-4230.

• - Saleh, Joseph H.; Haga, Rachel A.; Favarò, Francesca M.; Bakolas, Efstathios (October 2, 2013). "Texas City Refinery Accident: Case Study in Breakdown of Defense-in-depth and Violation of the Safety–Diagnosability Principle in Design". Engineering Failure Analysis. 36: 121–133. doi:10.1016/j.engfailanal.2013.09.014. eISSN 1873-1961. ISSN 1350-6307.

• - 2005 Houston Chronicle Special Report (archived March 12, 2008).

• - Baker Panel home page (archived July 26, 2007).

• - Explosion of the BP America refinery (Texas City) at the United States Chemical Safety Commission.

• - Monitoring the status of CSB investigation recommendations.

• - BP response corporate webpage (archived April 8, 2005).

• - Remember the 15 (archived March 12, 2008) - Site in memory of those who died in the explosion.

• - Texas City Explosion - Website managed by Brent Coon & Associates, lead attorney in the litigation that followed the explosion. It contains numerous pieces of evidence used in the case.

First thing in the morning

The commissioning process began with the main night shift operator on March 23 carrying out the initial filling of the separation tower. As plant start-ups are particularly prone to unexpected situations, operational practice requires the application of a controlled and approved pre-start-up safety review (PSSR) procedure. BP had one, but it was not adopted in this case. The process control level transmitter was designed to indicate the raffine level at a range of 1.5 m (5 ft) from the bottom of the dividing tower to a level of 2.7 m (9 ft). A high level alarm dependent on this transmitter sounded as scheduled at 3:09 am, when a level of 7.6 feet (2.3 m) from the bottom was reached. However, during start-up it was common to ignore this alarm and fill to a 99% level (as indicated by the transmitter) to avoid damage to the furnace that heated the bottom of the divider. Unknown to the operators, the process control level transmitter, which was used to monitor the level in the divider throughout the start-up operation, was not calibrated and its readings were unreliable. A separate level alarm activated by a high level switch at 2.4 m (7.9 ft) should also have sounded, but did not. At 5:00 a.m., the main operator of the ISOM satellite control room reported to the central control room and went home early. At 6:00 am, the day shift operator (central control room operator) arrived and began his 30th consecutive day of work on a 12-hour shift. At the time, the level was still believed to be below the 9 ft mark, but it was already at 13 ft (4.0 m).

At 7:15 a.m., more than an hour late from the start of his shift (6:00 a.m.), one of the two day shift ("A") supervisors arrived at the central control room. Due to his delay, he was unable to make the necessary relief with the night shift. During the morning meeting on 23 March, it was discussed that the heavy raffine storage tanks were almost full and therefore a second day shift ("B") supervisor was told that the start-up procedure should not continue, but this information was not conveyed. Therefore, the start-up procedure was resumed just before 9:30 a.m. following the instructions of day shift Supervisor A. The separator level control valve on the heavy storage tank was placed in manual mode (rather than regulating its opening via the level transmitter) and set to 4,300 bpd. However, the flow transmitter on this line had an error and the valve had actually closed. This, together with the malfunction of all the instruments in the separation tower, allowed the liquid level in the tower to increase without anyone realizing it and caused the operators to lose all knowledge of the situation.[12] The circulation process was restarted just before 10 in the morning and raffin was reintroduced into the tower, although the level was already too high. As the level control valve was closed and therefore there was no circulation outside the tower (i.e. no heavy raffin was transferred to the storage tank), the separation tower inevitably began to fill. The faulty level transmitter continued to show a level below 100% and as the external sight glass was opaque, a visual check to verify the level in the dividing tower was not possible. Day shift supervisor A, the more experienced of the two, left at 10:50 a.m. due to a family emergency. Contrary to operating regulations, there was no supervisor left in the central control room. A single operator, unsupervised and very tired, had to supervise the ISOM and two other units during a critical start-up procedure.

Late morning

Two furnace burners had been turned on at 9:55 a.m. to preheat the raffinate entering the tower and to heat the raffinate at the bottom of the tower. Two others were lit at 11:16 a.m. The required tower reboiler return flow temperature was 135 °C (275 °F) with a rise of 10 °C (18 °F) per hour, but this procedure was not followed: during startup, the return flow temperature reached 153 °C (307 °F) at a rate of 23 °C (41 °F) per hour. The faulty level transmitter continued to erroneously indicate a safe level condition on the tower. However, there was still no flow of heavy raffin from the separation tower to the storage tank, as the level control valve remained closed; Instead of the hydrocarbon liquid level being at 8.65 feet (2.64 m), or 93% of the instrument's range, as indicated, it had actually reached 67 feet (20 m). Just before noon, with the heat in the tower increasing, the actual liquid level had risen to 98 feet (30 m). Pressure began to build up in the system as hydrocarbon vapors and nitrogen remaining in the tower and associated piping since it had been returned to service were compressed by the increasing volume of refining. Operations staff thought the pressure increase was due to overheating of the tower bottoms, as this was a known startup issue, so the pressure was released.

Burst

At 12:42 pm, the furnaces had been shut down and the level control valve finally opened, draining the heavy raffine from the separation tower. The gas feeding the furnace was turned off, but not the raffin feed to the separation tower. Operators believed the level transmitter reading, which was now 78% (7.9 ft [2.4 m]), but the fluid level in the 170 ft (52 m) high dividing tower had reached 158 ft (48 m).

Although the opening of the heavy raffinate flow should have lowered the level in the column, the fact that this hot flow was used to preheat the feed meant that the temperature inside the separator increased dramatically, causing a significant increase in vaporization and the rise of a slime of liquid above the top of the column to the collector line. By 1:13 p.m., the hydrostatic head of this liquid had increased to over 42 psi (290 kPa), which was enough to open the relief valves. With the relief valves fully open, more than 196 m3 (51,900 US gallons) of heated raffinate passed directly to the collection manifold over a 6-minute period before the valves closed, as the pressure within the splitter header pipe fell below its minimum shut-off set point (37.2 psi (256 kPa)). 256 kPa (37.2 psi) above atmospheric pressure) Hot raffinate flowed into the blowdown drum and stack, and as it filled, some of the fluid began to flow into the ISOM unit sewer system through a 6-inch (15 cm) pipe located at the base of the blowdown drum. As the blowdown drum and chimney filled, hot raffin was shot out the top of the chimney into the air, forming a 6 m "geyser". The equivalent of a nearly tanker truck full of raffin rained onto the ground, ran down the side of the blowdown drum and chimney, and accumulated at the base of the unit. A radio call was received in the control room reporting that the chimney was overflowing with hot hydrocarbons. The plant's evacuation alarm did not sound, which prevented people in the vicinity from being able to evacuate before the ignition occurred.

A diesel truck, with the engine idling, was parked about 8 m from the blowdown chimney. The vapor cloud reached the vehicle and hydrocarbon fumes entered the engine's air intake, causing it to accelerate. Nearby workers frantically tried to shut down the engine, to no avail. The expanding cloud of vapor forced workers who were trying to turn off the engine of the truck that was going too fast to retreat. The cloud continued to spread across the ISOM plant, across the pipe rack to the west and into the unobstructed trailer area. No emergency alarm sounded and, at approximately 1:20 p.m., the vapor cloud was ignited by a backfire that nearby witnesses observed came from the truck's engine overheating.

The congestion of equipment and pipes contributed to accelerating the flame front. This triggered a massive vapor cloud explosion that could be heard for miles. The pressure wave from the explosion hit the group of contractor trailers located just 37 m from the blowdown stack, completely destroying them. The explosion sent debris flying, instantly killing 15 people in the nearest trailers and injuring 180 others.[e] Workers in the trailers were injured up to 150 m (480 ft) from the purge drum, and some trailers were severely damaged up to 180 m (600 ft). More than 40 trailers were damaged.[25] All of the fatalities and many of the injured in the accident were contractors. Fifty storage tanks suffered structural damage, although most of the tank farm was more than 250 feet (76 m) from the explosion. More than 2,750 pounds (1,250 kg) of benzene leaked from one of the damaged tanks.

emergency response

The local emergency response team intervened immediately, mounting a search and rescue operation. A shelter-in-place order was issued for 43,000 people. Mutual aid provided by IMAS (Texas City Industrial Mutual Aid System) and Memorial Hermann Life Flight resources were mobilized at 1:45 p.m.[26] Power to the raffine splitter was not cut off, but stopped at 2:45 p.m. when power was lost. The fires were brought under control by 150-200 firefighters within two hours. The ambulances and lifesaving helicopters left at 4:44 p.m. The last body was found under a pile of rubble around 11:00 p.m.

Contenido

Los expertos internos de BP, así como diversas autoridades y comités, investigaron la explosión en relación con aspectos técnicos, organizativos y de cultura de seguridad. Las investigaciones internas de BP incluyeron un grupo (la investigación Mogford, por el nombre del investigador principal) encargado de reconstruir la cadena causal del accidente y realizar un análisis detallado de las causas subyacentes, y otros dos equipos (las investigaciones Bonse y Stanley) que se centraron en los factores subyacentes de procedimiento y culturales, así como en las responsabilidades de los directivos. BP encargó a un grupo independiente de alto nivel (el grupo Baker), por recomendación urgente de la Junta de Seguridad Química de EE. UU. (CSB), que examinara los problemas de gestión y cultura de la seguridad. La CSB llevó a cabo su propia investigación, detallada y exhaustiva, centrada tanto en aspectos técnicos como de procedimiento.

Los distintos paneles e investigaciones detectaron fallos organizativos, como la reducción de costes por parte de la empresa, la falta de inversiones en la infraestructura de la planta, la falta de supervisión por parte de la empresa tanto de la cultura de seguridad como de los programas de prevención de accidentes graves, la concentración en la seguridad laboral en lugar de en la seguridad del proceso, una gestión defectuosa del proceso de cambio (que permitió la ubicación de remolques de contratistas demasiado cerca de la unidad de proceso ISOM), la formación inadecuada de los operarios, la falta de supervisión competente de las operaciones de puesta en marcha, la mala comunicación entre personas y departamentos y el uso de procedimientos de trabajo obsoletos e ineficaces que a menudo no se seguían. Entre los fallos técnicos cabe citar el uso de un tambor de purga de tamaño insuficiente y anticuado, la falta de mantenimiento preventivo de los sistemas críticos para la seguridad y la inoperatividad de las alarmas y los sensores de nivel de la unidad de proceso ISOM.[27]​.

BP internal investigations

A team of experts led by John Mogford, senior vice president of safety and operations at BP Group, examined the technical aspects of the explosion and proposed corrective measures. An interim report was published on May 12, 2005.[79] In response to the report, Scott Berger, director of the Center for Chemical Process Safety (CCPS) at the American Institute of Chemical Engineers (AIChE), expressed surprise at the report's emphasis on the individual responsibility of plant operators and supervisors.[28]​.

On December 9, 2005, BP published the final review of the Mogford report, which identified four critical factors without which the explosion would not have occurred or would have had a minor impact: “loss of containment; procedures for starting up the raffine separator and applying knowledge and techniques; control of work and location of the trailer; and blowdown stack design and engineering." Additionally, five critical underlying cultural issues were identified:

• - "Business context", which included a work environment that resisted change and was dominated by a lack of motivation, confidence and a sense of purpose. Added to this were unclear expectations regarding management and supervision.

• - "Safety as a priority", since management did not give process safety the necessary importance within the company.

• - "Complexity and organizational capacity", in the sense of a lack of clarity in responsibilities and poor communication.

• - "Inability to see risk", or a tendency to accept high levels of risk due to poor awareness of the dangers.

• - "Lack of early warning", or inability to recognize and act on signs that reveal deterioration of facilities and procedures.

However, Mogford's final report found no evidence that anyone intentionally made decisions or took actions that endangered others.

The company also assembled a team of BP and external experts to conduct an operational and process audit review of the refinery. James W. Stanley, former deputy director of the Occupational Safety and Health Administration (OSHA), was the team leader. The audit focused on processes and operations; incident management, work control, risk assessment and compliance assessment; employee and contractor management; and maintenance, reliability and integrity. The Stanley report was published internally on 15 June 2005. It concluded that a significant change in performance and behavior on the part of senior and extended management would be required, despite well-designed and documented procedures and processes and the workforce being generally capable of good performance. Among the issues that "impeded the successful execution of some key work processes," the team highlighted: leadership factors, such as lack of security accountability and silo mentality, among others; risk awareness, indicated by complacency and repeated failure to heed recommendations resulting from previous accidents; work control measures, which were insufficient and were not complied with; negative workplace conditions, as demonstrated by poor cleaning and insufficient maintenance of facilities; and a contractor management philosophy lacking diversity and inclusion values.

Another internal team was led by Wilhelm Bonse-Geuking [de], who was vice president of the BP Group in Europe. The Bonse team investigated plant managers' compliance with BP's management framework and corporate code of conduct. The final report - which was not made public until a court ordered it on May 3, 2007 - detected numerous management failures[74],[29][30]​ It further noted that management responsibilities within the BP Group were unclear and that the poor condition of the plant's equipment and insufficient maintenance spending were contributing factors to the accident.[30]​.

In summary, the Texas City refinery had a culture of risk-taking coupled with an inability to understand the process safety implications of past incidents [...], a long tradition of non-compliance with simple procedures, a desire to avoid conflict within its organization, and a penchant for placing people in key positions who lacked appropriate professional training.

The report singled out four executives for dismissal: Pat Gower, regional vice president of U.S. refining; Mike Hoffman, group vice president of refining and marketing in North America; Don Parus, Texas City complex director; and Willie Willis, Texas City West plant supervisor.[2] On the date of publication of the report, none had been fired.[6]​.

Baker Panel Report

Following the March explosion, two other serious process safety incidents occurred at the plant:

• - On July 28, 2005, a gaseous hydrogen heat exchanger pipe at the waste hydrotreatment unit ruptured, causing a hydrogen leak that erupted in a large jet of fire. The fire lasted about two hours. One person was slightly injured and property damage amounted to $30 million. The Chemical Safety Board discovered that a contractor had accidentally swapped a low-alloy steel elbow for a carbon steel pipe elbow during maintenance, causing a failure mode known as high-temperature hydrogen attack (HTHA). The CSB found that BP should have required positive verification of the materials using an

• - On August 10, 2005, a diesel hydrotreater suffered a leak caused by high-temperature corrosion and sulfidation that caused the release of toxic gases, such as carbon monoxide (CO), hydrogen sulfide (H2S), and sulfur dioxide (SO2), as well as flammable hydrocarbon vapors.[33] Material damage amounted to approximately $2 million.

Following these events, on August 17, 2005, the US Chemical Safety and Hazard Investigation Board (CSB) issued an urgent recommendation that BP commission an independent group to investigate the safety culture and management systems of BP North America.[34]​ A group of experts was convened led by former US Secretary of State James A. Baker III.[35][36]​ BP specifically singled out one figure. of this stature and resume to publicly demonstrate to US opinion makers that the company was willing to learn its lesson and take steps towards change.

The Baker Panel differed from the BP and CSB investigations in that it was not tasked with conducting a root cause investigation. Its scope was broader, focusing on BP's corporate safety oversight, corporate safety culture, and its process safety management systems, and included all five BP refineries in the United States.

The Baker Panel report was released on January 16, 2007[37]​[38]​ and cited a poor safety culture driven by poor process safety leadership and poor employee training as underlying causes of the accidents that plagued the refinery. It also highlighted that, despite BP's comprehensive safety management system, it was not sufficiently implemented at its refineries. It was also revealed that when measuring its safety performance, BP was focusing primarily on workplace safety indicators, to the detriment of monitoring process safety performance.[39] It was also highlighted that worker fatigue and a system that encouraged overtime had detrimental effects on the safe operation of the plant, and that the company had not addressed deficiencies arising from known incidents, risk assessments and audits. The lack of application of good engineering practices was also noted.

Additionally, the group interviewed several employees, managers and contractors at BP's five U.S. refineries. They also developed and administered a process safety culture survey to the refineries. They concluded that the Toledo and Texas City plants had the worst process safety cultures, while the Cherry Point refinery, located in Birch Bay, Washington, had the best process safety cultures. The survey results also showed that managers and white-collar workers generally had a more positive opinion of the process safety culture in their plants compared to the views of blue-collar operators and maintenance technicians.

CSB report

Given the magnitude of the disaster, the Chemical Safety Board examined both the safety management at the Texas City refinery and the role of the BP group and that of OSHA as a regulatory agency. The CSB investigation team arrived at the scene 48 hours after the accident. Some 13 CSB investigators remained on site for three months. For its investigation, the CSB used a budget of $2.5 million and reviewed more than 30,000 documents, interviewed 370 witnesses, and conducted computer modeling and testing. The results of the agency's investigation were published on March 20, 2007 in a lengthy 341-page report, the largest ever produced by the agency, which was then nine years old. The report's conclusions were presented the same day at a public meeting in Texas City.[41][42]​.

One of the main findings of the CSB was that the purge system used in the ISOM unit was outdated and totally inadequate, as it was located in the middle of the plant and could release heavy unignited vapors into normally occupied areas.[116] The CSB concluded that BP had failed to take into account and implement multiple warnings and safety recommendations issued before the explosion in relation to the ISOM purge system. Among them:

• - In 1991, Amoco's refining planning department proposed eliminating vented blowdown systems, but funding for this plan was not included in the budget.

• - In 1992, OSHA penalized Amoco for the unsafe design of a similar blowdown system elsewhere in the refinery. However, Amoco was able to convince OSHA to withdraw the citation based on the less stringent requirements of API Recommended Practice 521.[43]​.

• - In 1993, the "Amoco Regulatory Cluster Project" proposed eliminating atmospheric purge systems, but again its funding was not approved.

• -

A flare chimney, an inherently safer solution to remove excess flammable gas

Despite Amoco's "Process Safety Standard No. 6," which prohibited the installation of new atmospheric purge systems and required the phasing out of existing ones, in 1997 Amoco replaced the 1950s-era drum chimney/purge vent that served the raffin separation tower with an identical system, rather than upgrading it with recommended alternatives that were safer.

• - In 2002, the opportunity to connect the ISOM relief system to the naphtha desulfurization unit's new flare system was not taken due to an additional cost of $150,000.

• - Also in 2002, BP's "Clean Streams Project" proposed converting the blowdown drum into a flare discharge tank and directing discharges to a flare. When it was discovered that a necessary relief study of the ISOM system had not been completed due to budget constraints, Clean Streams decided not to pursue the option.

Between 1994 and 2004, there were at least eight similar cases in which flammable vapors were emitted from the ISOM drum/purge fan chimney, two of which resulted in fires. Furthermore, the purge drum was not designed to cope with overfilling of the container. This was not necessarily attributable to Amoco or BP, but rather to the lack of guidance in API Recommended Practice 521.[43]​.

Los daños materiales de la explosión ascendieron a 200 millones de dólares (322 millones en 2024)[46]​ La explosión provocó el cierre de varias unidades de la refinería. En previsión de la llegada del huracán Rita en 2005, se cerró toda la refinería. BP se centró entonces en reparar los daños causados tanto por la explosión como por el huracán. El reinicio de las unidades de proceso comenzó en marzo de 2006. Los costes de las reparaciones y la producción aplazada ascendieron a más de 1.000 millones de dólares.[47]​ BP se declaró culpable de delitos federales contra el medio ambiente, por los que pagó 50 millones de dólares.[48]​ La empresa también pagó al menos unos 2.100 millones de dólares en acuerdos civiles.[47]​Además, BP pagó 84,6 millones de dólares y 27 millones de dólares en multas al gobierno federal a petición de la OSHA y la EPA, respectivamente,[49]​[50]​[51]​ y una multa de 50 millones de dólares al gobierno de Texas por infracciones medioambientales.[52]​ El desastre es el accidente de refinería más costoso del mundo.

BP response and fate of the refinery

BP launched a crisis management plan just six hours after the explosion. The next day, a website[53]​ was created to post updated information about the accident. The chief executive, Lord John Browne, visited the plant the day after the explosion.[54] In the months after the accident, BP tended to blame its workers and supervisors, which victims and union leaders saw as simple scapegoating. At the time, the company consistently chose not to publicly apologize for the accident. This changed on May 17, 2005, when Ross Pillari, president of BP Products North America, publicly apologized, saying: "We regret that our mistakes have caused so much suffering. We apologize to those injured and to the community of Texas City" and promised “financial assistance and compensation” to the injured and the families of the deceased.

On December 9, BP announced that it would implement a $1 billion budget over five years to improve the safety of its Texas City refinery.[55] BP also announced that it would remove all bleed drum/vent stack systems in flammable service, of which there were 11 at Texas City, and install new flares in accordance with its new policy prohibiting atmospheric venting of heavier light hydrocarbons. than air.[6]​[56]​They also relocated trailers away from areas where explosions could occur[56]​and initiated a trial of an integrated electronic system of safe work (ISSOW).[56]​ Group-wide initiatives were also adopted.[56]​.

However, shortly after the Texas City explosion and other accidents in 2005, BP's image in the United States was further tarnished by the near-sinking of the Thunder Horse PDQ semi-submersible oil rig in July of that same year[56] and, more seriously, in March 2006, when a pipeline spill was discovered in Prudhoe Bay, Alaska, while multiple investigations into the Texas explosion were still ongoing. City.[57]CSB Chairwoman Carolyn Merritt said there were striking similarities between the Texas City and Prudhoe Bay accidents, including "long delays in implementation, administrative documentation of closure even though no corrective action had actually been taken, or simple noncompliance," as well as "poor communication of lessons learned, excessive decentralization of safety functions, and high management turnover." [58] "Another serious accident occurred in 2007 at BP Texas City, when 143 refinery workers reported that they had been injured by inhaling toxic fumes released at the plant.[59] CEO John Browne resigned in 2007 for issues unrelated to the case,[9] although he was not spared criticism for the lax safety culture and budget cuts at US refineries in BP.[16]​[60]​Tony Hayward took over as head of the company. Hayward changed Lord Browne's emphasis on alternative energy and announced that safety would be the company's "number one priority."[61]​ By then, BP had already adopted a markedly apologetic stance for the recent accidents, especially the one in Texas City, and its executives and technical experts were making presentations about what had gone wrong and how they were working to prevent it from happening again. Then came the Deepwater Horizon explosion and oil spill in 2010, which had a very serious impact on the company globally, again stemming from BP's operations in the US. As a result, Hayward resigned and his position was taken over by US-born Bob Dudley.[62] Under Dudley, BP announced in 2011 that it was selling its Texas City refinery as part of its plan to divestment to pay for ongoing compensation claims and remediation activities following the Deepwater Horizon disaster. The sale of the refinery was successfully completed in early 2013 to Marathon Petroleum Corporation for $2.5 billion.[3]​[63]​Marathon already owned the adjacent Galveston Bay refinery and in 2018 merged both sites into a single refining complex.[64]​.

Two congressional hearings were held specifically on the Texas City disaster. Among those heard were representatives of API, Baker Panel, CCPS, CSB, EPA and USW, as well as relatives of the victims. In other congressional hearings devoted to subsequent BP accidents in the US, the story of the Texas City case was consistently presented within the pattern of degraded safety culture at BP.[65]​.

Agreements with victims

The case of Eva Rowe, a young woman who lost her parents in the explosion, attracted national attention. Rowe said he would not accept a settlement from BP and would take the group to justice. Ed Bradley, a well-known American journalist, broke his story on the television program 60 Minutes.[66] On November 9, 2006, BP settled with Rowe as the final plaintiff, after his lawyers attempted to invite John Browne, BP's CEO at the time of the accident, as a witness. The amount of compensation to Eva Rowe remained unknown. BP also paid $32 million to hospitals and educational and research institutions designated by Rowe, including the Mary Kay O'Connor Process Safety Center at Texas A&M University ($12.5 million), the University of Texas Medical Branch at Galveston and its Truman G. Blocker Adult Burn Unit ($12.5 million), the College of the Mainland in Texas City ($5 million), St. Jude Children's Research Hospital of Memphis, Tennessee ($1 million), and the Hornbeck, Louisiana school system ($1 million).[67]​ In addition, BP was forced to release some seven million pages of internal documents, including the Telos and Bonse reports.[68][69]Rowe would later participate in one of the congressional hearings on the accident.

As of September 2007, BP had settled at least 1,350 of the approximately 3,000 claims related to the accident.[70] By February 2008, some 4,000 claims had been filed, half of which had been settled, totaling more than $1.6 billion.[71] BP also stated that it had set aside another $525 million for other claims.[47] August 2008, only one of the approximately 4,000 lawsuits remained open.[72]​.

Criminal actions

On February 4, 2008, U.S. District Judge Lee Rosenthal heard arguments regarding BP's bid to plead guilty to a federal environmental crime for two violations of the Clean Air Act (CAA)[73] with a fine of $50 million. At the hearing, victims of the explosions and their families objected to the plea, calling the proposed fine "trivial".[74]​ However, the plea was eventually agreed to, along with a three-year probationary period for BP.[48]​ This was the first and, for several years, remained the only federal conviction for an accidental chemical release under the CAA.[75]​.

Fines

In September 2005, the Occupational Safety and Health Administration, which the CSB report described as lacking oversight and jurisdiction, fined BP a record $21 million for committing 301 violations of the Process Safety Management standard.[76][77]In October 2009, OSHA fined $87 million, shattering its record of 2005, after stating that BP had not implemented safety improvements after the disaster and noting that there had been four other fatal accidents at the refinery since the previous fine. In its new report, OSHA cited 709 safety violations.[78][79]​ BP announced that it would challenge the fine.[80]​ On August 12, 2010, BP announced that it had agreed to pay $50.6 million of the 2009 fine, while continuing to challenge the remaining $30.7 million (the fine had been reduced by $6.1 million between the time it was imposed and the time it was imposed). that BP paid the first portion).[81]​ In July 2012, OSHA and BP agreed that the $30.7 million outstanding would be reduced to $13 million, which BP paid.

Following the explosion, the US Environmental Protection Agency inspected the refinery for compliance with the Clean Air Act (CAA) and determined that BP had violated the CAA as well as its Risk Management Program (RMP) standard.[82] Most of the violations were not directly related to the explosion, but to other events, such as the two subsequent accidents in 2005. In February In 2009, the EPA imposed a first fine of $12 million on BP. [51]​In September 2010, EPA and BP reached a $15 million settlement on other violation charges,[33]​ which was the largest civil fine imposed for CAA violations at a single facility and the largest imposed for RMP civil violations.[83]​.

At the request of the Texas Commission on Environmental Quality (TCEQ), the Texas Attorney General opened proceedings against BP for violations of the Texas Health and Safety Code[84]​ and the Texas Water Code,[85]​including the release of hydrocarbons through the blowdown stack on March 23, 2005, the prolonged release of benzene from a tank damaged in the explosion, which lasted more than 25 days, and dozens of others. events,[23]​ In a 2011 settlement, BP agreed to pay a $50 million fine covering 72 other emissions that exceeded BP's operating permits. This included legal costs of $500,000. In exchange, the Department of Justice agreed not to admit additional criminal charges against BP in connection with the refinery explosion.[86]​.

La catástrofe tuvo una notable repercusión en el ámbito de la seguridad de los procesos. Texas City se ha convertido en un caso clásico utilizado para explicar los fallos tanto de gestión como de barreras técnicas en las plantas de procesos.[36]​[87]​[88]​.

El informe del Panel Baker se hizo muy conocido entre los ingenieros de seguridad de procesos, que consideraron que sus conclusiones eran relevantes para otras plantas y que era importante para reforzar la concienciación sobre la seguridad de los procesos en la industria de procesos químicos,[89]​ cumpliendo así el deseo del Panel recogido en el informe:.

• - Aunque necesariamente dirigimos nuestro informe a BP, pretendemos que llegue a un público más amplio. No nos hacemos ilusiones de que las deficiencias en la cultura de seguridad de los procesos, la gestión o la supervisión corporativa se limiten a BP. [90]​.

• - Otras empresas y sus partes interesadas pueden beneficiarse de nuestro trabajo. Instamos a estas empresas a que evalúen periódica y exhaustivamente su cultura de la seguridad, el rendimiento de sus sistemas de gestión de la seguridad de los procesos y su supervisión corporativa de la seguridad para detectar posibles mejoras. También instamos a las mismas empresas a que revisen detenidamente nuestras conclusiones y recomendaciones para aplicarlas a sus situaciones.[91]​[90]​.

• - Análisis de la deficiente aplicación de la gestión de la seguridad de los procesos.

• - El accidente ha sido ampliamente analizado en la literatura especializada, que ha puesto de relieve cómo varios elementos de la gestión de la seguridad de los procesos (PSM) se implantaron de forma deficiente y se gestionaron mal. Entre las observaciones realizadas sobre los elementos de la PSM definidos en la norma de la OSHA[92]​se incluyen:.

• - Participación de los trabajadores: La refinería adolecía del miedo de los trabajadores a ser castigados por notificar incidentes o peligros.[28]​.

• - Análisis de peligros del proceso: el análisis de peligros y operabilidad (HAZOP) de la planta no identificó el escenario de sobrellenado de la columna[90]​ y la evaluación de riesgos para la colocación de edificios temporales no fue completa y, en su mayor parte, errónea.[91]​[90]​.

• - Procedimientos operativos: Las desviaciones de los procedimientos clave, como la puesta en marcha de la planta ISOM, se habían convertido en rutina. Además, el mismo procedimiento de puesta en marcha carecía de instrucciones suficientes[90]​[29]​.

• - Formación: La formación de los operadores clave de la planta era muy insuficiente, especialmente en lo relativo a la gestión de situaciones anómalas y la verificación de los conocimientos y cualificaciones de los operadores.[30]​.

• - Contratistas: La dirección de las empresas contratistas hizo caso omiso de aspectos de seguridad importantes. Por ejemplo, no se informó de la puesta en marcha a quienes ocupaban remolques próximos a la central ISOM. Todas las víctimas mortales y muchos de los heridos del accidente eran contratistas.

• - Revisión de seguridad previa a la puesta en marcha: Esta revisión clave en la preparación de una operación especialmente peligrosa no se llevó a cabo.[93]​.

• - Integridad mecánica: Múltiples instrumentos y válvulas no funcionaban porque no se habían probado o mantenido.[90]​.

• - Gestión del cambio: Esta gestión falló en múltiples aspectos, a saber, los cambios en el procedimiento de puesta en marcha de la unidad, que obligaban a los operarios a hacer funcionar el divisor por encima de un nivel seguro,[94]​la falta de MOC o la existencia de MOC incompletos para la colocación de remolques en las proximidades de una unidad altamente peligrosa,[29]​ y la falta de un proceso de cambio formal que evaluara la eliminación de puestos de operarios críticos para gestionar la planta ISOM.[94]​.

• - Investigación de incidentes: El aprendizaje de incidentes y cuasiaccidentes pasados se vio perjudicado por la práctica ausencia de investigaciones internas y la consiguiente difusión de lecciones aprendidas útiles.[90]​.

• - Planificación y respuesta ante emergencias: No se hizo sonar la alarma de evacuación, lo que puede haber contribuido al número de víctimas mortales, ya que los contratistas que se encontraban en los remolques no tuvieron oportunidad de abandonar la zona.

Auditorías de cumplimiento: Se llevaron a cabo auditorías, pero, según admitió la propia BP, "no parece que se hiciera un seguimiento de los puntos de acción [de la auditoría] ni que se cerraran de forma efectiva".

Analysis of the poor application of process safety management

The accident has been widely analyzed in the specialized literature, which has highlighted how various elements of process safety management (PSM) were poorly implemented and poorly managed. Observations made on the elements of PSM defined in the OSHA standard[1]​ include:

• - Worker participation: The refinery suffered from workers' fear of being punished for reporting incidents or dangers.[90]​.

• - Process hazard analysis: the plant's hazard and operability analysis (HAZOP) did not identify the column overfill scenario[91]​ and the risk assessment for the placement of temporary buildings was not complete and, for the most part, erroneous[43][90]​.

• - Operating procedures: Deviations from key procedures, such as the start-up of the ISOM plant, had become routine. Furthermore, the start-up procedure itself lacked sufficient instructions.[91][90]​.

• - Training: The training of the main plant operators was very insufficient, especially in relation to the management of anomalous situations and the verification of the knowledge and qualifications of the operators.[90]​[91]​.

• - Contractors: The management of the contracting companies ignored important safety aspects. For example, those occupying trailers near the ISOM headquarters were not informed of the start-up. All of the fatalities and many of the injured in the accident were contractors.[65]​.

• - Pre-startup safety review: This key review was not carried out in preparation for a particularly dangerous operation.[95]​.

• - Mechanical integrity: Multiple instruments and valves were not working because they had not been tested or maintained.[4]​[3]​[91]​.

• - Change management: This management failed in multiple aspects, namely, the changes in the unit start-up procedure, which forced operators to operate the separator above a safe level,[94]​the lack of MOC or the existence of incomplete MOC for the placement of trailers in the vicinity of a highly dangerous unit,[3]​ and the lack of a formal change process that evaluated the elimination of operator positions critical to managing the plant ISOM.[5]​.

• - Incident investigation: Learning from past incidents and near misses was harmed by the virtual absence of internal investigations and the consequent dissemination of useful lessons learned.[3]​.

• - Emergency planning and response: The evacuation alarm was not sounded, which may have contributed to the number of fatalities, since the contractors who were in the trailers did not have the opportunity to leave the area.

• - Compliance audits: Audits were carried out but, by BP's own admission, "it does not appear that the [audit] action items were tracked or closed effectively."

New industry guidelines and initiatives

API implemented the CSB report's recommendations, creating new standards and guidelines for the industry:

• - Standard 521: API published a new revision of its document on Pressure Relief and Depressurization Systems and elevated its category from Recommended Practice to Standard. The new standard contained stricter requirements on atmospheric venting of process gas and an instruction to include overfilling of upstream equipment among the design cases of purge systems.[99]​[100]​.

• - Recommended Practice 753: API issued new guidance on locating trailers and portable buildings in hazardous process locations. The scope of Recommended Practice 752 was reduced to permanent buildings only beginning with the third edition.[101][102][103]​.

• - Recommended Practice 754: API addressed the need for process safety performance indicators in this new Recommended Practice on Process Safety Performance Indicators for the Refining and Petrochemical Industries. investigations.[106]​.

• - Recommended Practice 755: This new guideline was aimed at refineries and petrochemical plants and detailed how to implement a fatigue risk management system (FRMS). This document includes recommendations for rotating shift work, for the maximum acceptable number of overtime hours and the number of days that should be worked without interruption.[107]​[108]​[109]​.

The CCPS published extensive guidance on organizational change management to address another related recommendation from the CSB report.[110]​[111]​.

After being singled out for its lack of initiative and competence in the evaluation and inspection of large hazardous process plants, and refineries in particular, OSHA took action, initiating a National Process Safety Management Emphasis Program (NEP) for refineries and implementing a targeted inspection audit program between 2007 and 2011. This was the largest PSM enforcement action since the regulation[1]​ was issued in 1992.[112]​.

OSHA also issued an internal memorandum[113]​ to address the CSB's recommendation to update the PSM standard to include requirements for hazardous process facilities to expand their change management procedures to include organizational changes. However, the CBS did not consider this initiative sufficient to close the recommendation, which, as of January 2024, remained open.[111]​[114]​.

The accident has appeared in several documentaries:

• - National Geographic series Seconds from Disaster, season 3, episode 10 "Texas Oil Explosion", first aired on November 6, 2006.[27]​.

• - The History Channel series Modern Marvels, season 12, episode 56 "Engineering Disasters 20", first aired on December 6, 2006.[28][29]​.

• - The Science Channel series Engineering Catastrophes, season 4, episode 4 "Texas Terror", first aired on July 7, 2021.[30]​.

• - Seconds of the Disaster episodes from National Geographic.

• - Abílio Ramos, Marilia; López Droguett, Enrique; Mosleh, Ali; das Chagas Moura, Márcio; Ramos Martins, Marcelo (August 18, 2017). "Revisiting Past Refinery Accidents from a Human Reliability Analysis Perspective: The BP Texas City and the Chevron Richmond Accidents". The Canadian Journal of Chemical Engineering. 95 (12): 2293–2305. doi:10.1002/cjce.22996. eISSN 1939-019X. ISSN 0008-4034.

• - Barozzi, Marco; Melchiore, Alessandra Derudi, Marco; Copelli, Sabrina (2023). "Integrating Recursive Operability Analysis with Different Risk Assessment Methods: Analysis of the Historical BP American Refinery Explosion" (PDF). Chemical Engineering Transactions. 104: 163–168. doi:10.3303/CET23104028. ISBN 979-12-81206-06-9. ISSN 2283-9216. Archived (PDF) from the original on January 26, 2024. Retrieved on January 26, 2024.

• - ESReDA Project Group Foresight in Safety (2020). Enhancing Safety: The Challenge of Foresight. EUR 30441 EN. Luxembourg, Luxembourg: Publications Office of the European Union. doi:10.2760/814452. ISBN 978-92-76-25189-7. Archived from the original on January 26, 2024. Retrieved on January 26, 2024.

• - Hopkins, Andrew (July 1, 2008). Failure to Learn: The BP Texas City Refinery Disaster. Macquarie Park, NSW: CCH Australia. ISBN 978-1921322440.

• - Manca, Davide; Brambilla, Sara (2012). "Dynamic Simulation of the BP Texas City Refinery Accident". Journal of Loss Prevention in the Process Industries. 25 (6): 950–957. doi:10.1016/j.jlp.2012.05.008. eISSN 1873-3352. ISSN 0950-4230.

• - Saleh, Joseph H.; Haga, Rachel A.; Favarò, Francesca M.; Bakolas, Efstathios (October 2, 2013). "Texas City Refinery Accident: Case Study in Breakdown of Defense-in-depth and Violation of the Safety–Diagnosability Principle in Design". Engineering Failure Analysis. 36: 121–133. doi:10.1016/j.engfailanal.2013.09.014. eISSN 1873-1961. ISSN 1350-6307.

• - 2005 Houston Chronicle Special Report (archived March 12, 2008).

• - Baker Panel home page (archived July 26, 2007).

• - Explosion of the BP America refinery (Texas City) at the United States Chemical Safety Commission.

• - Monitoring the status of CSB investigation recommendations.

• - BP response corporate webpage (archived April 8, 2005).

• - Remember the 15 (archived March 12, 2008) - Site in memory of those who died in the explosion.

• - Texas City Explosion - Website managed by Brent Coon & Associates, lead attorney in the litigation that followed the explosion. It contains numerous pieces of evidence used in the case.