Explosion and fire of coke gas in a steel production plant

An explosion followed by a fire occurred at a tar precipitator of gas treatment of a coke oven. The extractors were manually activated. Ten minutes later, the operator activates the internal emergency plan. The fuel supply to the installation was interrupted. Approximately two hours later the fire was extinguished by nitrogen purging.
The operator estimated the maximum amount of material that reacted in the explosion of the tar precipitator at 170 kg in TNT equivalent. In total, the operator estimated that 36,790 t of coke oven gas were released and burned at the candles (torches above the oven batteries) and flares. The coke oven gas consisted mainly of hydrogen and methane.
A crack had been identified in the steam circuit eight months before the accident, which could not be repaired without a precipitator shutdown. As the monitoring of the oxygen measurement at the precipitator outlet did not show any deviation, the repair was scheduled for the normal shutdown 3 days after the day of the accident.
Several hypotheses as to the origin of the ignition energy were put forward: electrostatic due to the friction of the air/gas mixture (also composed of dust) on the metal plates, chemical due to the presence of pyrophoric materials or mechanical due to the fall of a metal part inside the precipitator.

The INITIATING CAUSE was a crack in the steam circuit, which probably allowed air access to the precipitator.
The third-party expert report concluded that the most probable hypothesis was that the mixture of coke oven gases came into contact with the oxygen in the air coming from the leakage of the steam network and ignited by the electrostatic energy generated by the friction of the gases on the metal walls.

The decision to continue operation after the detection of a crack in the steam circuit can be considered the ROOT CAUSE of this incident. This decision was based on the measure of oxygen content, which did not seem to increase despite the crack.

steel plant, tar precipitator, coke gases

At the moment of the incident, the precipitator was shutdown, being drained and washed.

DESCRIPTION OF THE PROCESS
a tar precipitator is an electrostatic filter which collects dust and tar contained in the gas during the transformation of the coal in coke. The coke oven gas can then be used as energy source in other site processes.

The tar precipitation unit was destroyed.
The estimated volume of coke oven gas burned at the candles and the flare was 3 776 600 Nm³.
Flaring of the coke oven gas followed for 90 days until the tar precipitator was replaced.
In total, the operator estimates that 36,790 t of coke oven gas were released and burned at the candles and flares. The operator carried out a health assessment and modelled the emission of SO2 which, was considered the only pollutant resulting from combustion which could have an impact on the population. They concluded that the daily average limit value for the protection of human health, the information and recommendation threshold and the alert threshold were not exceeded.

16h15, Explosion followed by fire
16:25, the operator activated the internal emergency plan. The personnel is evacuated. A security perimeter is set up. The fuel supply to the installation is interrupted.
20:00, the fire was extinguished by nitrogen purging. The weakened structure was in danger of collapsing.
At 8:30 pm, the upstream and downstream valves of the tar precipitator and scrubber were closed, coke oven gas was released, nitrogen was injected on the gas manifold and the structure was assessed. The operator installed solid gaskets on the flanges to isolate the tar precipitator units and the H2S scrubber.
Flaring of the coke oven gas then followed for 90 days until the tar precipitator was replaced.

Explosions in an electrostatic tar precipitator had already occurred on this site in July 2005 and on other sites of the group abroad in November 2017 and in 2019. However, these explosions always occurred during the restarting of the precipitator by powering the electric fields, never during the washing and draining phases when the electric fields are not powered (no source of ignition). There was therefore no feedback and this risk had not been identified by the operator.

Following this accident, the operator reviewed the risk analyses and took additional measures, including a review of the washing procedure for the tar precipitator by carrying out the washing and draining operations under pressure, to avoid air ingress.