Explosion at a bio-fuels research laboratory

The incident at a bio-fuels research laboratory occurred when a visiting research fellow was transferring hydrogen, oxygen and carbon dioxide from compressed gas cylinders into a low-pressure mixing tank. The mixture, consisting of 70% H2, 25% O2 and 5% CO2, was used as a bacterial growth medium.
The research fellow had just finished to fill the tank when an explosion occurred.
The explosion was the result of leaks which allowed the formation of a flammble and explosive mixture, sparked probably by the use of an inproper pressure gauge or by static electricity.
The researcher was severly injured by the explosion.

The INITATING CAUSE was the creation of a flammable mixture in the surrounding of the tank due to leak, ignited by a spark of the pressure gauge, possibly assisted by static electricity shock.

All filling operations, including their controls, were performed manually. Except for a measure of the internal tank total pressure by means of a portable pressure gauge, no other monitoring system was used, despite the specific hazards related to handling hydrogen and oxygen. In particularly, no flammable gas detector is mentioned having been used. Moreover, leak tests after assembly had not been performed at the nominal (rated) tank pressure, but at less than 2/3 of it.

The ROOT CAUSE was related to lack of proper risk assessment, lack of preventing and mitigating safety systems, and lack of duly supervision by university management of the activities if a visiting fellow.

Portable gas cylinder

The tank had been bought few months before, together with its accessories: a digital pressure gauge, a pressure relief valve and fittings. The tank was rated at 150 psi (10 bar). When the tank was assembled with its parts, a pressure test was done using the buildings compressed air, which could only reach 91.2 psi (6.2 bar). Several leaks were detected, so the tank assembly was taken to the university maintenance workshop for assistance in stopping the leaks.

For the purpose of her research, the composition of the gasses had to be changed from [70% H2, 20% O2 and 10% CO2] to [70% H2, 25% O2 and 5% CO2]. The premixed gas would then be connected to a reactor.
According to the visiting research fellow, this was the first time to use of a portable tank to premix the gases required for the experiments. Before, each individual gas would have a separate hose line and be premixed at the reactor eliminating the use of a portable tank.
According to the Laboratory Professor the tank that failed was a design of the visiting research fellow. According to the visiting research fellow, the design of the tank was that of the Laboratory Professor.

The research fellow also added that earlier in the week of the accident, she was conducting another experiment using a smaller one gallon size air tank assembly nearly as identical to the one that failed using similar components to include a digital pressure gauge and premixed gasses. The tank pressure was set to 27 psi (1.9 bar). After reading the gauge, she pressed the OFF button and a small internal explosion occurred, with evidence of soot and smoke stains. Static shock also appeared to have been a problem as the researcher would get shocked on occasion when touching the tank.

DESCRIPTION OF THE PTOCESS
The tank that exploded was a portable 13-gallon steel cylinder (50 litres). The tank was used to contain a mixture of gases to include: 70% Hydrogen (H2), 10% Carbon dioxide (CO2) and 20% Oxygen (O2) in that order and to a normal operating pressure of 50 psi (3.5 bar). The tank was not grounded and was normally moved from stored location to areas where it could be filled. The tank would be moved approximately 3 feet to fill it with H2 and CO2 and then moved approximately 13' so it can be filled with O2.

One person lost an arm.
The lab was destroyed.
Due to the explosion, the air tank was separated at two points. One end cap was completely severed from the tank and the weld that ran horizontally across the tank was split open; the opposite end cap remain intact but there was visual evidence of over-pressure as the end cap bulged outward.
There were four threaded holes in the tank: one was for the digital pressure gauge, one for the shut-off valve, one for the drain and the last for the pressure relief valve. Only the pressure relief valve and drain valve remained intact and affixed to the tank. The pressure gauge and shut off valves were separated from the tank when the explosion occurred.

Based on the fire brigades investigation which included but was not limited to the physical examination of the incident scene, observed fire patterns, photographs, witness statement and experience, it is with reasonable degree of fire investigative certainty that this incident was caused by an accidental explosive event.
The fire/explosion has been classified as accidental, pending any further investigation.

Update September 2016:
State occupational safety and health agency found 15 workplace safety violations associated with blast that severely injured a researcher and the University of Hawaii was fined $115,500
The researcher got a 6,700,000 US$ settlement
(https://www.khon2.com/local-news/former-uh-researcher-gets-6-7-million-settlement-for-lab-explosion-injuries/#:~:text=Former%20UH%20researcher%20gets%20$6.7%20million%20settlement%20for%20lab%20explosion%20injuries)

These are the finding of the legal investigation, issued by the [Hawaii Occupational Safety & Health Division Citation and Notification of Penalty to the University of Hawaii, issued Sept. 16, 2016].
The university workplace safety violations identified by HIOSH
1. The employer failed to provide a safe workplace by reducing employee exposure to potential explosion and fire hazards.
2. The employer did not ensure that its safety practices were followed by employees and underscored through training, positive reinforcement, and a clearly defined and communicated disciplinary system.
3. The employer did not ensure periodic in-house inspections were being performed in Hawaii Natural Energy Institute laboratories to determine new or previously missed hazards.
4. Laboratory personnel working under the principal investigator did not use the required personal protective equipment at all times.
5. Two exit routes were not available in the laboratory to permit prompt evacuation of employees and building occupants.
6. The exit door did not swing out in the direction of exit travel.
7. The employer’s emergency action plan(s) did not list the evacuation meeting point nor a way to account for the evacuees.
8. The employer did not review the emergency action plan when employees were initially assigned.
9. A fire prevention plan did not include specific provisions to address potential ignition sources in the presence of hydrogen and other flammable gases.
10. Activities performed in the laboratory by researchers with the potential exposure to explosion and fire hazards were not assessed for appropriate personal protective equipment.
11. Activities performed in the laboratory by researchers with the potential exposure to explosion and fire hazards were not assessed for appropriate glove protection to guard against static discharge and flame-retardant laboratory coats to guard against fire.
12. Where hazardous chemicals were used in the workplace, the employer did not carry out the provisions of a written Chemical Hygiene Plan, which were capable of protecting employees from health hazards associated with hazardous chemicals in that laboratory.
13. The employer’s Chemical Hygiene Plan did not include the standard operating procedures relevant to safety and health considerations to be followed when laboratory work involved the use of hazardous chemicals.
14. The employer’s Chemical Hygiene Plan did not include criteria to determine and implement controls relevant to the gas mixing operation (engineering controls, personal protective equipment, administrative).
15. The employer failed to review and evaluate the effectiveness of the Chemical Hygiene Plan at least annually and update it as necessary.