The Weekly Reflektion Week 32 / 2020

Trevor Kletz was an industrial safety advisor in Imperial Chemical Industries (ICI). He published a monthly newsletter describing incidents that had occurred in the company and actions needed to prevent them. Soon he was sending out 2000 copies a month to both internal and external recipients. After he retired be published many of these in a book called ‘What Went Wrong? Case Histories of Process Plant Disasters’, well worth a read. After a few years he realized that a further book was required since the first one obviously had not had the desired effect across the process industry. The title of the book was appropriately ‘Still Going Wrong!’. This week’s reflektion is from Trevor’s book and concerns one of the major hazards in the process industry, explosions, and one of the common causes for disasters, failure to manage changes.

How do you manage changes? Do you consider all the ways a change may affect your operations and systems? Are you systematic, methodical, and concerned with identifying potential hazards introduced by the change? We hope you are.

On November 2nd, 1997, an explosion occurred in a fixed roof gasoil tank located in the tank farm of the Ashdod Oil Refinery in Israel. The explosion killed one person and caused fires in two adjacent tanks located in the same dike that lasted more than 2 hours. The explosion occurred when a laboratory technician was sampling the contents of the tank. The explosion was unexpected as the gasoil should have had a flash point of 66 C and therefore should not have been able to form an ignitable mixture in air. 

Before storage, the gasoil was stripped to remove lighter hydrocarbons and ensure a flash point above 60 C. The process of stripping the gasoil was changed from steam to hydrogen twenty years before the incident. The refinery did not have a formal management of change process at the timeand no one had considered that hydrogen may have been carried with the gasoil into the storage tanks. Calculations following the incident showed that 90% of the hydrogen dissolved in the gasoil following the stripping would be released immediately when it was pumped into the storage tanks. The remaining 10% would slowly be released from the gasoil and diffuse through the atmospheric vents. This means that for some time following filling of the storage tank there was an explosive mixture of air and hydrogen in the space above the gasoil. The sample was taken 10 hours after filling and this was not enough time for the hydrogen to be released from the gasoil and be vented to atmosphere. The regular samples taken for testing of flash point did not register any low flashpoints since any hydrogen in the samples taken would have evaporated before they were analysed. No one had considered the possibility of hydrogen in the space above the gasoil. The explosion was just an accident waiting to happen.

The ignition source was most likely a spark from static electricity discharge from a nylon rope that was used to lower the sampling bottle into the storage tank. During the investigation, the standard for tank sampling was found. It stated, ‘In order to reduce the potential for static charge, nylon or polyester rope, or cords or clothing should be used’. A copy of the accident report was sent to the originators of the standard. They replied by apologizing for the omission of the word ‘not’ when the standard was last updated.

Further investigation into the explosion revealed that a similar explosion in a tank containing gasoil contaminated with hydrogen was reported 14 years earlier. Unfortunately, no-one at the refinery seemed to remember the incident and no precautions based on this explosion were included in relevant operating procedures and sampling standards.

A change in the stripping process, a change in a sampling standard and no changes implemented following a similar incident. One person dead, significant material damage and 15000 m3 of gasoil burned up. Still going wrong folks!!

Reflekt AS