Model-based fault detection and isolation for chemical processes: Application to the prevention of thermal runaway

Over the past decades, serious industrial events affecting lives, facilities and environment have occurred in chemical industry. Among the most critical events are the phenomena of thermal runaway that often lead to important damages. The rapid detection of thermal runaway events and the isolation of the most probable cause are major challenges to ensure safe operations. The objective of this work is to propose a method for early detection of faults in a chemical reactor based on a reference model and for isolating the fault cause. The reaction of perhydrolysis of formic acid to peroxyformic acid by hydrogen peroxide is used as a test case. The kinetic model of the reaction was computed in order to simulate the reaction in normal and abnormal modes. In a first stage, the method was validated by simulation data and an experimental validation is now in progress. As a consequence, the proposed method can contribute to the safety of chemical reactors in the chemical industry.