Fire at a fire heater of an hydrocracking unit

The incident occurred in the fired heater of a hydrocracking unit. Due to overheating, about 250,000 pounds of hydrogen and naphtha mixture were accidentally released into the firebox of the fired heater, where it ignited, creating a major fire.
The technical cause of the overheating was the reduction in flow of the process gas due to spurious materials left in one of the fire heater’s tube. Due to the reduced flow, the process gas had already reached three days before the temperature of 600 C, which was the operation limit. To reduce the gas temperature, the operators reduced the fuel supply to the burner. On the day of the incident, the fire heater reached again the temperature limit. Despite a second intervention of the operators, two of the tubes ruptured releasing hydrogen and naphtha into the firebox.
The post-accident metallographic analysis concluded that the tubes ruptured due to a combination of creep damage and short-term overheating.

The investigation found that fire blankets and insulation material were present inside the failed tubes, very likely left inside the tubes during the preventive maintenance work performed three weeks earlier.

The INITIATING CAUSE was the failure of the fire heater tubes due to overheating.

The ROOT CAUSE was an error occurred during the maintenance of the installation, which left inside the tubes spurious materials which caused reduction of gas flow and overheating. The diagnosis of the problem was hindered by the fact that the fired heater was not equipped with individual pass flow instrumentation. This made difficult to identify the local problem of only two tubes. This hint at shortcoming in the design.

heater, tubes

Maintenance had occurred 3 weeks before.
3 days before, the first high-temperature alarm had been reached.

No injury.

This incident, with only financial consequences, was caused by inaccurate maintenance works. It is unclear how the thermal isolation materials entered the tube, but it is probably something that the maintenance instruction do not consider. Anyhow, a final check before declaring the equipment ready for operation should have detect the spurious material.
The operative control system could also be improved, by allowing a more detailed data collection. The operators could not detect a reduction of flow in the tubes, because the unit was measuring only the total flow. Moreover, it seems also that the temperature measurement and diagnosis system was not allowing for an early detection of hot spot.