Process monitoring and fault diagnosis based on a hybrid modeling technique

A hybrid model based process monitoring and fault diagnosis approach is developed, in order to make full use of the advantages of various modeling methods. An ANN-based hybrid modeling technique with serial configuration is adopted to develop an accurate process model. Based on the hybrid model, principal component analysis (PCA) is used to deal with the correlations among high-dimensional process variables for process monitoring; and a model-based fault estimation and diagnosis method is used to determine the cause to fault. The main advantage of the proposed approach over existing works lies in that, it has good performance both in process monitoring and fault diagnosis. The application results on a three-tank system show the effectiveness of the proposed approach.