Fire on a heat exchanger at a benzene manufacturing plant

The event occurred at a high-temperature heat of at a benzene manufacturing plant using a dealkylation reaction. High-pressure hydrogen leaked from a flange of the heat exchanger, during start-up The hydrogen ignited, and a fire followed.
The leakage occurred for two reasons:
(1) An insufficient tightening torque control performed during hot-bolting, which generated inhomogeneous force distribution among bolts.
(2) A out-of-design temperature rise of the heat exchanger caused by a revamping during a turnaround shutdown.

Event sequences:
A turnaround shutdown of the benzene manufacturing plant had been carried out 3 months before.
A gas-tightness test with real gas (hydrogen and another gas) was completed in September.
The plant was started by increasing temperature and pressure
During the operation, an operator at an adjoining plant found a fire at the upper part of the vertical reactor.

The INITIATING CAUSE was a leak caused by loosening of the tightness of a flange, due to inhomogeneous torque among bolts.

The IGNITION SOURCE was attributed to static electricity generated at leakage.

The shell-side inlet temperature had been increased from an initial value of 400°C to 420°C after a revamping aiming at energy-saving. This added an additional driver for the flange loosening, because a decrease in tightening torque was locally generated by a difference in thermal expansion between bolts made of stainless steel and the flange made of carbon stee (thermal expansion of stainless steel is greater than that of carbon steel).
This suggests ROOT CAUSE components related to inadequate management of changes and shortcoming in operative procedures unable to control the tightness of the joint.

reactor, heat exchanger

DESCRIPTION OF THE PROCESS
In this plant, benzene was produced by a dealkylation reaction using toluene and C9 aromatic hydrocarbon as raw material.
The overall reaction is: C6H5CH3 + H2 -> C6H6 + CH4.

HOT-BOLTING: it is the operation ensuring tightness of flanges operating at high temperatures. If bolted only at room temperature, these flanges will become loose at high temperature, because the tightening force decreases with increasing temperature. Re-tightening of bolts at high temperature is necessary. In this event, the design conditions of the flange were 2.4 MPa, 3.9 k and 500 °C at the shell side, and 2.8 MPa, 300 °C at the tube side, and it required hot-bolting.

About 4.9 square meters of the top of a feed evaporator of a reactor burned.

An emergency shutdown was executed and relevant equipment were purged with nitrogen.

The plant affected by this event did not have in place a mechanism able to guarantee the correct execution of installation works. The report does not specify who was responsible for the tightening works: in the majority of these plants, this type of work is performed by contractors. It is critical, for a plant operator, to be able to assess independently the correct execution of the maintenance, repair and installation works.

One element which contributed to exacerbate the potential leak from lack of proper tightness, was the decision to increase the gas inlet temperature to gain in energy efficiency. All modification should be ‘tested’ against the original design specification, to identify potential additional hazards introduced by the modifications.

The plant adopted the following measures:
(1) The quantitative control of initial axial force on tightening bolts was improved.
(2) A field patrol was established and the monitoring of the system evolution the during a temperature rise was intensified.