Superheater “A” on vapour cracking unit no. 1 at a petrochemical platform exploded around 3 pm. Of the 8 personnel present at the time, 2 were killed and the other 6 injured. Projectiles, composed mainly of refractory material, were sent distances on the order of 100 metres and some pieces close to 50 cm in size fell in the immediate vicinity of the superheater; a dust cloud was visible directly above the site.
The Internal Emergency Plan was activated : the plant was evacuated and 70 fire-fighters arrived at the scene. Of the 6 injuries, all requiring hospitalisation (including 2 subcontracted employees working onsite), 5 were released the same evening. No damage or other impact was recorded offsite, outside of the sound of the blast. Cylindrical in shape with a 5-m diameter and some 20 meters high, this water vapour superheater, tied to a chimney of the same height via a connecting cone, did not contain any toxic product. The explosion was not followed by a fire outbreak. Subsequent to violent atmospheric storms during the night of July 13th to 14th and with water infiltration entering a utility room disturbing the digital control system, vapour cracking line 1 had been shut down and placed in safety mode. This line’s restart procedure was launched in the morning on the day prior to the accident. This procedure was long to carry out, since the start-up routine had to be performed section by section. On 15 July, superheater A was reset around 3 pm with the intention of a manual ignition. A technician had shown up with an adjustable pole to light the pilots when the superheater exploded. The bodies of this technician and a second employee were found underneath rubble from collapse of the superheater floor.
According to the site operator, this accident resulted from various causes, namely :
- an accumulation of flammable gas, still below the flammability limit : investigations conducted forwarded the hypothesis of a gas flow towards a burner during both the start-up phase and ignition step ;
- ignition of the dust cloud by the lighting pole or by a hotspot inside the superheater convection zone. Other ignition sources could be hypothesised as well (e.g. electric spark, static electricity), although the two identified above appear to be the most plausible. A number of circumstances facilitated the occurrence of this accident, whose severe consequences were due to the presence of personnel in the vicinity at the time of powering up the facility ;
- failure to proceed with a vapour cleaning of the superheater prior to restart, in violation of operating protocol ;
- gas intake through a burner in the absence of a flame on the corresponding pilot ;
- the technical safety barrier, according to which it is prohibited to supply burners without a visible flame on the pilot, was not operational. This barrier was composed of an automated mechanism that closes gas feed valves if the flame detector is not signalling the presence of a flame 10 seconds after valve opening. Following a number of erratic detection alerts shortly after installation, this automated detector mechanism was deactivated due to the limited number of shutdowns/restarts planned for the unit over its operating cycle.
Download the detailed report in .pdf format (720 Kb)
