Explosion in waste unit of a chemical plant

The event occurred in an atmospheric pressure tank storing effluents. The plant was in the production phase. The head of the overfill prevention valve located at the top of the tank was thrown a distance of 77 m. The tank’s shell was deformed (curved downward) but not leaking. Employees noted a noise similar to the opening of a valve, and a low-pressure alarm in the tank appeared in the control room, despite the fact that the level and temperature were normal.

The tank was filled with i) organic compounds normally producing hydrogen, and ii) formic acid containing a minor faction of hydrogen peroxide.
Because of the presence of hydrogen the tank was maintained under positive nitrogen pressure to make chemically inert its vapour space. Two days before the incident, the second effluent was slightly contaminated with suspended solid particles of a catalyst used in the production unit feeding the storage tank. The catalyst consist of palladium on carbon in very fine size se particles (< 10 ppm).
The palladium catalysed the decomposition of hydrogen peroxide into water and gaseous oxygen. The hydrogen and oxygen mixture in the vapour space then exploded in contact with an undefined heat source, causing an overpressure in the tank resulting in its deformation and the ejection of the liquid seal cap.

The INITIATING CAUSE was the catalysed splitting on hydrogen peroxide into hydrogen and oxygen, with consequent internal explosion in the effluent tanks.
The presence of palladium catalysts in the tank was abnormal and the result of a process malfunction.
Despite the presence of nitrogen in the vapour space, this additional supply of oxygen lowered the lower flammability limit of the hydrogen, which in air is 4.1% vol.

ARIA does not provide and explicit root cause analysis and nothing is known on the design and/or operative aspects which brought to the malfunctioning. Nevertheless, the ROOT CAUSE could be attributed to inadequate safety design, because a post-accident study showed that, even in the absence of catalysis, oxygen production in the tank through decomposition of hydrogen peroxide was higher than initially estimated, representing a higher hazard than previously assumed.

storage tank

The facility had been operating in the same configuration for roughly 30 years.
Two days before the incident, an effluent sent to the tank was slightly contaminated with suspended solid particles. These particles, in small quantities (< 10 ppm), consist of palladium on carbon and serve as a catalyst in the workshop that produces the effluent.

DESCRIPTION OF THE FACILITY
The tank was fed by two main flows of aqueous effluents:
1) The first effluent is a solution containing an organic component that generates hydrogen gas under normal storage conditions.
2) The second effluent consists of a solution of formic acid and a minor by-product of the process. This by-product is 2% hydrogen peroxide (a strong oxidant) which, under normal storage conditions, generates carbon dioxide as a result of decomposition, with traces of oxygen.

Only damage to he storage tank

The ARIA report (see references) mentions that the operator considered the installation of a permanent nitrogen purge in the tank, to guarantee a continuous removal of the hydrogen from the top.
This would help zeroing possible production malfunctioning. However, as additional preventive measure, a detection system for the formation of explosive atmosphere could be another suggestion.