Explosion and fire at Buncefield Oil Storage Depot – Five companies to face prosecution

Hazardous Area Classification Industrial Accidents Safety Devices Safety Instrumented Systems
Spread the love

It is now almost three years since the Buncefield oil storage depot explosion took place and finally the authorities have declared, that they will be pressing criminal charges against five companies, ostensibly who have been found guilty of acts of omission.

For those of you who do not remember the case, here is a short overview. There were a number of loud explosions ( I mean really really loud-reportedly people in Netherlands and France heard it and it was recorded also a seismic event! ) and a massive fire at the Buncefield Oil Storage Depot in Hemel Hempstead, Hertfordshire, UK. Over 40 people were injured in the accident, fortunately there were no fatalities. Following the explosion, a Major Incident Investigation Board (MIIB) was established by the Health and Safety Commission, supported by the Board of the Environment Agency, UK.

This is a photo of the Bruncefield oil storage depot explosion

There were a series of investigations and reports being published from time to time by this board and some of the significant findings were as follows -my summary of a rather long series of detailed reports:

a) No consequence analysis was done by any of the design engineers or safety experts, as to what could be the severity of the possible explosions of the flammable vapors generated from the petroleum storage tanks.

b) The level control loop, (that is supposed to control the level in the tank and prevent overfilling) on one of the storage tanks failed. It consisted of a Servo tank gauge connected to a series of valves. This failure led to overfilling and spillage of massive amounts of petroleum into the dikes surrounding the storage tank. Petroleum was being pumped in at a rate of about 550 m3/hr for more than three hours, yet the servo level gauge indication, failed to record any change at all!

However the DCS trend records could be salvaged and the above information was gleaned from them. Apparently the CCTVs were working and the footage showed petroleum overfilling and flowing into the tank dikes, but nobody was watching it at the time.

c) The overfill protection was provided by a point level switch which was supposed to be independently connected to an alarm/annunciator panel (separate circuit from the DCS loop). The panel had an override switch and it may be that the interlock was bypassed  (no conclusive evidence since everything got burned in the subsequent fire, this may never be known). However it is warning to design engineers who think that by merely having a redundant level switch is good enough. Were there any common cause failures that both the continuous indication, as well as the interlock failed? Not known for sure.

d) The operators apparently did not notice anything amiss and neither was the control system very sophisticated, to tally the pumping rate into the tank to the rate of change of the level. Now here’s the cake. The pumping rate now increased to 890 m3/hr leading to the petroleum overflowing from the tanks, filling up the bunds and secondary containment areas and forming large vapor clouds. It seemed this occured because the inlet lines were common to all the tanks and the other tanks level indications were working, so the system diverted their inlets into this tank that appeared about half full (due to the faulty level indication). There must have been thousands of gallons of the stuff overflowing from all directions and nobody could notice anything! (Yes, it was about 3:00 am in the morning-but so what-¬† were there no operator rounds of the premises or anything like that?- or it doesn’t happen on the night shifts at all?!)

e) Apparently the hazardous area classification which may have been done during the initial stages, may not have considered wind directions. The entire vapor cloud was carried across the road from the tank farm to an emergency generator building,about 100 meters away, where it is thought to have been ignited. The building apparently was not a classified (hazardous) location.

No doubt this entire catastrophic incident and the consequent investigations will have a major impact on how instruments and controls are designed and maintained in petrochemical/hydrocarbon processing plants, how operator alertness and awareness is important and so on.

More details are available at the Buncefield investigation site.

Hertfordshire ExplosionNote: All images have been sourced from the Buncefield investigation site and all copyrights belong to that site.