A Particle Filter Based Dynamic Gaussian Mixture Model for Process Fault Detection and Diagnosis

Complex multimode processes may have dynamic operation scenario shifts and strong transient behaviors so that the conventional monitoring methods become ill-suited. In this article, a new particle filter based dynamic Gaussian mixture model (DGMM) is developed by adopting particle filter resampling method to update the mixture model parameters in a dynamic fashion. Then the particle filtered Bayesian inference probability index is established for process fault detection. Furthermore, the particle filtered Bayesian inference contributions are decomposed among different process variables for fault diagnosis. The proposed DGMM monitoring approach is applied to the Tennessee Eastman chemical process with dynamic mode shifting and the results show its superiority to the regular Gaussian mixture model in terms of fault detection and diagnosis accuracy.