↓ Expand ↓

Category → Chemical Safety & Hazard Investigation Board

Tesoro refinery fire caused by weakened steel

Last week, the U.S. Chemical Safety & Hazard Investigation Board released its draft report about a 2010 fire at a Tesoro refinery in Anacortes, Wash., that killed seven workers. The fire occurred when a naphtha heat exchanger ruptured, the report says. The cause of the rupture was high temperature hydrogen attack, which occurs when hydrogen atoms diffuse into carbon steel and react with the carbon to form methane. The methane accumulates in the steel and causes stress and fissures. CSB found that curves established by the American Petroleum Institute to predict high temperature hydrogen attack are inaccurate. “CSB has learned of at least eight recent refinery incidents where HTHA reportedly occurred below the carbon steel Nelson curve,” the report says.

Here’s CSB’s video about the incident:


The CSB found several indications of process safety culture deficiencies at the Tesoro Anacortes Refinery. Refinery management had normalized the occurrences of hazardous conditions, including frequent leaks from the [naphtha hydrotreater unit] heat exchangers, by using steam to mitigate leaks, ineffectively correcting heat exchanger design issues, commonly requiring additional operators during [naphtha hydrotreater unit] heat exchanger startups, and exceeding the staffing levels that procedures specified.


The refinery process safety culture required proof of danger rather than proof of effective safety implementation. For years, technical experts used design data to evaluate the B and E heat exchangers for HTHA susceptibility. Data for actual operating conditions were not readily available, and these technical experts were not required to prove safety effectiveness in reaching their conclusion that the B and E heat exchangers were not susceptible to HTHA damage.

CSB noted several similarities between the Tesoro fire and a Chevron refinery fire in Richmond, Calif., in 2012:

  • The Chevron “incident was also the result of a metallurgical failure caused by a well-known damage mechanism called sulfidation corrosion, and Chevron process safety programs failed to identify the hazard before the major incident that endangered the lives of 19 Chevron employees.”
  • “Mechanical integrity programs at both Tesoro and Chevron emphasized inspection strategies rather than the use of inherently safer design to control the damage mechanisms that ultimately caused the major process safety incidents.”
  • “Rather than performing rigorous analyses of damage mechanisms during the PHA process, both companies simply cited non-specific, judgment-based qualitative safeguards to reduce the risk of damage mechanisms.”

One of the recommendations CSB makes in its Tesoro report is that Washington state implement a “safety case” approach to regulation, in which companies develop their own process safety requirements that are closely overseen by state regulators. The agency made the same recommendation to California in its Chevron report, but that report fell to a divided vote by CSB board members, with board members Beth Rosenberg and Mark Griffon wanting CSB to study regulatory options further. An expected vote on the Tesoro report last week was delayed. What happens from here is an open question. Given criticism of CSB that the agency is taking too long already to complete its investigations, holding reports for further study seems untenable. The federal budget passed last month allotted CSB “$11 million for fiscal 2014, slightly less than previous years and below the Administration’s request of $11.5 million.”

Feb. 5, 2014: Title changed to reflect the fact that although I saw the heat exchangers as a bundle of tubes inside a larger tube, CSB tells me that engineers view the outer structure as a shell, not a tube. In any case, weakened steel from high temperature hydrogen attack was still the problem.

CSB report on Chevron refinery fire urges new regulatory approach

On Monday, the U.S. Chemical Safety & Hazard Investigation Board released its draft report on a 2012 Chevron refinery fire in Richmond, Calif. CSB recommended that the state switch from a “current patchwork of largely reactive and activity-based regulations” to a performance-based system, the agency’s press release says.

CSB released an interim analysis and video of the incident earlier this year. The fire was caused by rupture of a pipe in a crude oil processing unit; the pipe was first identified as corroded in 2002 but was never replaced.

The regulatory approach CSB now recommends is called the “safety case” system and is already used in the United Kingdom, Norway, and Australia. From CSB’s press release:

…the safety case regime requires companies to demonstrate to refinery industry regulators – through a written “safety case report” – how major hazards are to be controlled and risks reduced to “as low as reasonably practicable,” or ALARP. The CSB report notes that the safety case is more than a written document; rather, it represents a fundamental change by shifting the responsibility for continuous reductions in major accident risks from regulators to the company.

To ensure that a facility’s safety goals and programs are accomplished, a safety case report generated by the company is rigorously reviewed, audited, and enforced by highly trained regulatory inspectors, whose technical training and experience are on par with the personnel employed by the companies they oversee, the draft report says.

That will mean that the regulatory agencies involved will also likely have to pay their employees more. A table in the CSB report notes that refinery personnel have an average annual salary of $187,630, while inspectors for county, state, and federal agencies make $96,875-$125,000 (pdf page 81).

CSB is on the report until Jan. 3. The agency will formally adopt or modify the report at a public hearing in Richmond on Jan. 15.

The Associated Press also reported this week that the Environmental Protection Agency “filed a formal notice against Chevron finding 62 violations of federal environmental laws.” The story goes on to say that EPA may “pursue criminal charges or fines if the company fails to address the violations.” So far there seems to be no mention of the notice on EPA’s main or regional websites.

Lab safety is critical in high school, too

The U.S. Chemical Safety & Hazard Investigation Board yesterday released a new video, “After the Rainbow,” that features a student injured in a high school laboratory fire. Calais Weber was 15 years old when her teacher poured methanol on an open flame in front of students at her desk. An explosion and fire ensued, and Weber was burned over 40% of her body.

I haven’t tracked numbers, but in the few years I’ve been following chemical safety news, I feel like I’ve heard of a lot of alcohol fires in schools. I don’t know why people seem to discount the fire hazard of alcohols, but they do.

Weber has this message for students:

While it can seem daunting, it’s perfectly okay to speak up if you’re not feeling safe, to at least question. And if you’re given a piece of information on safety, read it.

Crystal plant accident caused by stress corrosion cracking

The U.S. Chemical Safety & Hazard Investigation Board last week released its report on a 2009 explosion at a quartz crystal manufacturing plant in Illinois operated by NDK America. CSB determined that a corrosive environment led to cracks in a pressure vessel’s steel walls, resulting in its failure. The blast blew a piece of steel 650 feet to a nearby gas station, where it fatally injured one truck driver.

When they were in operation, the six 50-foot tall crystallization vessels at the plant were loaded with raw mined quartz, 800 gallons of 4% sodium hydroxide in water, a “small amount” of lithium nitrate, and seed crystals of pure quartz, the CSB report says. Once sealed, they were heated to 370 °C and pressurized to 29,000 psig for 100 to 150 days.

The vessels’ steel walls were eight inches thick. Sodium hydroxide and silica will react with iron in steel to produce a layer of sodium iron silicate, or acmite. ADK believed that the acmite layer would prevent corrosion of the steel, but neither the company nor the state ever inspected the vessels’ interiors. In 2007, one of the vessels leaked through a connection in its lid. A consultant hired by NDK determined that the leak was caused by stress corrosion cracking and found cracks in three other lids. The company continued to operate the remaining vessels without inspections, the CSB report says.

For more detail and for CSB’s recommendations to the American Society of Mechanical Engineers, the National Board of Boiler & Pressure Vessel Inspectors, the Office of the Illinois State Fire Marshal, and NDK, see CSB’s report. The agency also produced its usual excellent summary video:

When is an explosion really an explosion, take two

A few weeks back, we had a letter to the editor in C&EN that took us to task for using “blast” and “explosion” to describe two industrial incidents. We have more in this week’s issue (which, I might add, is a particularly awesome one in celebration of C&EN’s 90th anniversary). The consensus? The rupture of a nitrogen line is a mechanical explosion, and C&EN used the words appropriately. Here’s what our readers said:

Regarding Richard Rosera’s letter “Choosing the Right Words,” explosion is the correct term (C&EN, Aug. 5, page 4). The definition of the word explosion is the rapid expansion of a gas.

To quote Rosera, the case at hand was “caused by the rupture of a nitrogen line or vessel by overpressure.”

Rob Kovacich
Hood River, Ore.

Recalling my years dealing with hazard evaluation led me to question Rosera’s letter. An explosion is defined as the rapid increase in volume and release of energy in an extreme manner. Or, as Frank T. Bodurtha explains in his book “Industrial Explosion Prevention and Protection,” “an explosion is the result, not the cause, of a rapid expansion of gases. It may occur from physical or mechanical change.”

Thus, the rupture of a nitrogen line or vessel does indeed result in an explosion, as does the rupture of an overfilled tire.

Robert G. Robinson
Lawrenceville, N.J.

As a chemistry educator and professional pyrotechnician, I answer myriad questions regarding explosions. If the term explosion is used to refer to “the rupture of a nitrogen line or vessel by overpressure,” it is more specifically a mechanical explosion, but it’s an explosion nonetheless.

The criticism leveled by Rosera is unwarranted. Both the mainstream press and C&EN are correct in addressing the CF Industries accident as an explosion. Although physical failure of materials containment may be due to either chemical or mechanical reasons, the result is still an explosion.

Kathleen Holley
Arlington, Texas

At the time of the original news story, the cause of the explosion at a Williams C. ethylene plant in Geismar, La., was unknown. A July 31 story in the New Orleans Times-Picayune indicates that it was still unknown by that time. I can’t find anything more recent, but the U.S. Chemical Safety & Hazard Investigation Board is investigating it, so we’ll have an answer eventually. The incident killed two workers.

Friday chemical safety round up

Chemical health and safety news from the past couple of weeks.

First up, on the West Fertilizer explosion in Texas:

  • The Chemical Safety Board launched a Facebook page for its investigation into the West Fertilizer explosion
  • Sustained Outrage posted about various familiar issues surrounding the disaster
  • At a Texas House committee hearing, many agencies many agencies said “not my job” regarding lack of oversight and allowing large quantities of ammonium nitrate to be stored near a residential area
  • And the Center for Public Integrity reported on concerns about the pace of CSB investigations


  • In honor of Workers’ Memorial Day, the National Council for Occupational Safety & Health released “Preventable Deaths: The Tragedy of Workplace Fatalities” and the Centers for Disease Control & Prevention devoted its Morbidity and Mortality Weekly Report to worker concerns.
  • The Bureau of Labor Statistics released its final count of fatal work injuries in 2011: 4,693, “the third lowest annual total since the fatal injury census was first conducted in 1992.” That’s 3.5 fatal injuries per 100,000 full-time equivalent workers.
  • The Berkeley Science Review published a long article on the lab safety changes in the University of California system in the wake of Sheri Sangji’s death
  • The May issue of the Process Safety Beacon looks at “Pressure relief valve bonnets–to plug or not to plug?“
  • A Florida high school student experimented with combinging aluminum foil and toilet bowl cleaner in a water bottle on campus before school. She subsequently was expelled from school and charged with possessing and discharging weapons and a destructive device on school grounds. Yes, gas pressure built up in the bottle so it exploded, but really, it seems ridiculous to expel a student for this. From all reports, she was just curious and didn’t intend to harm anyone.
  • Janssen chemist Ramineh Behbehanian, on the other hand, perhaps did want to harm people by putting rubbing alcohol-contaminated orange juice onto the shelves of a San Jose, Calif., Starbucks. An alert customer saw her do it.
  • Norway orders BP safety review after leak
  • The Las Vegas Sun looked back at a 1988 explosion at ammonium perchlorate manufacturer Pacific Engineering Production Co. of Nevada that killed two people and injured more than 300 (C&EN archive story here, paywall-free link! coming), and explored what hazardous materials plants are in the area today
  • And WSYR in New York looked back at a fire from a flame test demonstration that left a teacher and three students badly burned
  • U.K. authorities fined SAFC Hitech $190,000 for a 2012 incident in which trimethylindium caught fire and badly burned one worker

Fires and explosions:

  • Three workers were killed in an explosion in a fireworks factory in India
  • Also in India, and explosion and fire from some sort of chemical transfer at Ganesh Plasto injured one
  • A fire at a Formosa Plastics plant in Texas involved ethylene and injured at least nine people (another story says a dozen)

Leaks, spills, and other exposures:

  • One worker died and six others were treated for exposure after breathing hydrogen sulfide fumes while cleaning pipes at a wastewater treatment plant at the Port of Tampa, in Florida
  • Something “in the ‘cyanide’ family” spilled at metal finisher Kocour in Illinois, sending one person for medical treatment
  • Phenol spilled at a medical clinic in Iowa, sending 13 people to two local hospitals, and also at a U.K. high school
  • Hydrogen peroxide leaked from equipment at the College of Nanoscale Science & Engineering in New York
  • Chemicals stored by a deceased fireworks enthusiast in a residential shed led to the evacuation of 49 neighboring houses while the bomb squad investigated

Not covered (usually): meth labs; ammonia leaks; incidents involving floor sealants, cleaning solutions, or pool chemicals; transportation spills; things that happen at recycling centers (dispose of your waste properly, people!); and fires from oil, natural gas, or other fuels.

CSB releases interim report on Chevron refinery fire

The Chemical Safety & Hazard Investigation Board last week released an interim analysis of the Aug. 6, 2012 refinery fire at a Chevron refinery in Richmond, Calif. The fire started when a corroded pipe ruptured in a crude oil processing unit. CSB identified several failures that led to the fire: “Missed opportunities to apply inherently safer technologies, a failure to recognize and correct workplace hazards, and the lack of industrial safeguards,” reported C&EN’s Jeff Johnson on Friday. CSB’s report is available here, and a Chevron internal report is available here.

One notable set of images in the CSB report is on pdf page 11. The set includes four photographs taken from across the Bay that show the initial hydrocarbon vapor cloud that formed when the pipe burst, followed by the black smoke from the fire. Amazingly, no one was killed when the leaking material ignited. From CSB’s video about the incident and its findings:

A decision was finally made to begin an emergency shut down of the crude unit. But it was too late. Suddenly, the pipe ripped open. A vapor cloud formed and rapidly expanded, as the large inventory of hydrocarbons in the distillation tower started to vent through the ruptured pipe. The vapor cloud immediately spread over hundreds of feet, engulfing all 19 people who had gathered nearby. The firefighters and operators struggled to escape through the dense hydrocarbon cloud, unable to see. They had to feel their way out, some on their hands and knees. At approximately 6:30 p.m., two minutes after the huge vapor cloud formed, the hydrocarbons ignited. One firefighter was trapped inside a fire engine when it was suddenly engulfed in flames. He radioed for help. … But when he received no response, he assumed everyone else was dead. To escape the inferno, he fled through what witnesses described as a wall of fire.

That fire engine was destroyed by the fire.The Chevron report argues that the white vapor cloud itself was not flammable. Instead, material still leaking from the pipe either auto-ignited or dislodged a light fixture, exposing wiring that could have ignited the stream.  Chevron’s internal report also says, “The response and assessment after the discovery of the leak did not fully recognize the risk of piping rupture and the possibility of auto-ignition” (Causal Factor 1, pdf page 21).

From pdf page 36 of the CSB report, here’s a graphic that illustrates the number of times someone at Chevron said the pipe should be replaced, and the refinery didn’t do it:
CSB-Chevron-key events

Last but not least, here’s CSB’s video:

Letter on Donaldson Enterprises fatal fireworks incident

A letter to the editor in this week’s C&EN focuses on a fatal fireworks disposal incident in 2011, when five Donaldson Enterprises died in an explosion and fire in a storage magazine after disassembling contraband fireworks:

As a chemist with more than 50 years’ involvement with display fireworks, I find it appalling that in the Donaldson Enterprises Inc. incident the safest and most obvious means of disposal was apparently never considered (C&EN, Jan. 28, page 26). Simply firing the materials normally and allowing them to function as designed in a safe place would have been a far better course of action.

Display fireworks are fundamentally different from munitions and other classes of explosives in too many ways to list here. But following are a few of the more salient differences applicable to disposal: They are often complex in construction, not designed with disassembly in mind, and widely varied in the number of different pyrotechnic compositions that might be present in a single device. They are not reliably destroyed by water or other liquids, are perilous to cut into, and are dangerous to mass-incinerate whether wet or dry. Disposal involving such methods requires great caution and a full knowledge of the product and should be reserved only for situations where conventional firing is impossible.

It appears that the materials in this case were not damaged or defective, but were merely mislabeled. Had they been properly marked and classified for professional use, they would have been perfectly suitable for that purpose. Therefore, there was no practical necessity for disposal by unusual means.

This raises the question of whether the root cause of this tragedy was, in fact, bureaucratic: Might arbitrary yet rigid protocols have precluded a far safer and simpler disposal? It would not be the first time that safety has been sacrificed upon its own altar by misguided policy.

I’m not sure it’s safe to assume that the fireworks were neither damaged nor defective or that “arbitrary yet rigid protocols…precluded a far safer and simpler disposal.” I don’t think the Chemical Safety Board addressed whether the fireworks were in good condition–from what I understand, no one with fireworks expertise ever looked at them, which frankly seems to be the whole reason those five workers died.

As for whether bureaucracy was at fault, CSB actually pointed to a lack of regulations and protocols as contributing to the incident. From the CSB report (pdf):

Contractor Selection and Oversight Findings

  • The Federal Acquisition Regulation (FAR), which governs federal agencies’ acquisition of goods and services, does not specifically require a federal contracting officer to consider safety performance measures and qualifications when determining the “responsibility” of a potential government contractor or subcontractor to handle, store, and dispose of hazardous materials such as fireworks.
  • The Department of the Treasury Acquisition Regulation (DTAR), the Department of the Treasury’s supplement to the FAR, does not impose sufficient requirements for safe practices and subcontractor selection and oversight with respect to the unique hazards associated with handling, storing, and disposing of hazardous materials.

Regulatory and Industry Safety Standard Findings

  • The CSB found a lack of regulations or industry standards that adequately address safe fireworks disposal. Federal or local codes, regulations, or industry standards do not establish safety requirements, provide guidance on proper ways to dispose of fireworks, or address the hazards associated with the disassembly of fireworks and the accumulation of explosive fireworks components.
  • While OSHA’s [Process Safety Management (PSM)] standard applies to fireworks manufacturing, OSHA has determined that the regulation does not apply to work activities related to fireworks disposal. Therefore, DEI was not required to implement a more robust PSM system for its fireworks disposal process. For example, DEI’s change to its disposal process led to the accumulation of material that created a mass explosion hazard. PSM elements such as Management of Change (MOC) would have required a safety review of this change.
  • Emergency hazardous waste disposal permits are granted in Hawaii and throughout the country to entities seeking to dispose of seized contraband fireworks because they are considered an imminent threat to human health and the environment. However, the Resource Conservation and Recovery Act (RCRA) does not incorporate PSM-type elements in its hazardous waste permitting process, despite the extremely hazardous nature of the materials covered by these permits.