Process Monitoring in Chemical Industries – A Hidden Markov Model Approach

The detection of abnormal events in advance is still a challenge in chemical industries. The earlier it is done, the greater the possibility of at least mitigating losses. This study investigates the performance of a signal processing tool so-called hidden Markov model (HMM) in accomplishing detection tasks. The case study is based on a chemical recovery boiler that belongs to a Kraft pulping mill in Brazil. The identified model, characteristic of normal operating conditions, was exposed to usual situations, namely multiple normal operating states, transient periods, and abnormal events. The detection system was able to reach both mandatory requirements, i.e. early identification of abnormal situations and minimization of false alarm rates.