Process Fault Diagnosis Based on Bayesian Inference

Isolating faulty variables is a crucial step during the determination of the root causes of a process fault. Contribution plots, with their corresponding control limits, are the most popular tools used for isolating faulty variables. However, the isolation results may be misled by the smearing effect. In addition, the control limits of the contributions cannot be used to isolate faulty variables, since the control limits are obtained from the normal operating data, which lack any information about the faults. In chemical processes, process faults rarely show a random behavior; on the contrary, they will be propagated to varying variables due to the actions of the process controllers. During the evolution of a fault, the task of isolating faulty variables needs to be concerned with the faulty variables decided in the previous data; in addition, the current decisions should influence the isolation results for the next sample when the fault is constantly occurring. In the presented work, an unsupervised data-driven fault isolation method was developed based on Bayesian decision theory. The Tennessee Eastman (TE) process was used as a benchmark example to demonstrate how the different faulty variables were isolated when the fault was evolving.