In a crude oil distillation unit, tests showed that the
flare system valve was not providing effective isolation and would require
eventual removal for overhaul at a scheduled shutdown of the flare. A ‘cold
work’ permit to work was issued two days prior to the incident. Alternate
flange bolts were removed and the other bolts lubricated as a standard practice
to save time. Sufficient bolts remained at all times to retain the flange
seals. There was at that time no need to verify line conditions. Two contractors
wearing breathing apparatus completed the work. When almost all the bolts were
undone, liquid leaked from the gap between the flanges and gas escaped from the
top of the joint.
The men stopped work, came down to ground level and sought advice. The
supervisor checked the
platform and saw gas issuing from the top and liquid leaking from the bottom of
the flange. He concluded that neither was under pressure and that the quantity
of liquid was small. Without any further tests assured the contractors that it
was safe for work to continue. The fitters remained concerned, thus asked and received
'spark proof' tools. Liquid continued to leak as more bolts were removed then,
as the last bolt was undone and the crane took the strain and started to lift
the valve, the spacer suddenly sprang upwards. A large quantity of liquid was
released, a flammable vapour cloud formed and ignited by the nearby compressor.
Two workers died in the incident.
Key learning points
A tower scaffold with a working platform and access ladder had been erected for
work on the valve but due to access restrictions, it was necessary to climb
over or under it. This seriously limited the route of escape. Work on the valve
should not started prior to verification of the isolation and should not have continued
after the first leak occurred until all doubt about the safety of the situation
had been resolved. The absence of the spark arrester on the compressor was not
known of until after the incident.
Source:IChemE
No comments:
Post a Comment