Hydrogen fire in a chemical plant

The incident occurred at a hydrogenation reactor producing alcohols. At the time of the fault, the plant was in normal operation. A sudden leak of hydrogen occurred on the flange of the upper cover of a high-pressure trickle bed reactor. The escaping hydrogen ignited and a fire broke.
The plant sensors for flammable gases went off in the control room a few seconds after the leak began and automatically alerted the plant fire brigade.
(1) The affected part of the plant and adjacent parts were shut down using the plant shutdown switch.
(2) The residual pressure in the affected hydrogenation reactor was released to the flare via the automatic rapid relief
(3) The reactor was purged with nitrogen, which extinguished the flame after less than 5 minutes.
(4) The BASF plant fire brigade was on site and cooled the affected parts of the plant.
There were no injuries.

The INITIATING CAUSE was a hydrogen leak on a flange with a deteriorated sealing ring and a low tightening torque.

The ROOT CAUSE was the failing of upgrading the correct tightening torque for this type of flanges, and the fling of inspect possible damage to the sealing ring. This can be attributed to shortcoming in the operative procedures.

flange

According to ZEMA, at the time of the fault, the plant was in normal operation at an operating pressure of approx. 250 bar and an operating temperature of approx. 135°C.

DESCRIPTION OF THE PROCESS
The plant produced a C13 alcohol by hydrogenation with hydrogen.

There were no injuries. The excess hydrogen was burned off using the flare. Defective cables, sensors and wet or damaged insulation had to be replaced.

The mitigating strategy worked as planned. The flammable gas detectors alarmed the operative room and the plant was shut down, the reactor purged with nitrogen, and the affected reactor components were cooled by the fire brigade. The fire was extinguished by 5 minutes.

There was no evidence of similar previous events. The post-incident investigation reported by ZEMA found that the following factors enabled the leak:
(1) the tightening torque of the screw connection of the affected flange prescribed by the original drawings from 1994 was at the lower limit according to the present regulations.
(2) The double-cone sealing ring was deformed in the tenth of a millimetre range, due to multiple opening and closing of the required by internal inspections and catalyst replacement.
(3) Use of graphite tape as an assembly aid during assembly of the reactor cover after the last catalyst replacement in 2020.

[Note of the event validator: It is unclear how the third factor (3) mentioned by ZEMA could have facilitated the leak, because it has the use of graphite tapes has been experienced as positive over the past 35 years. In fact, ZEMS reports that the influence of graphite tape as an assembly aid has still being investigated].

The department responsible for the design of the equipment checked the tightening torque of similar sealing systems of the high-pressure devices, according to the current state of knowledge.
The operative procedure for the opening and re-closing of similar high-pressure devices was modified so that the sealing surfaces are checked for deformations, every time, using suitable measuring technology. If necessary, the surface will have to be reworked.