Two-dimensional dynamic batch processes modelling and monitoring

Dynamics are inherent characteristics of batch processes, and the dynamic behavior may exist not only within a batch run, but also from batch to batch. Recently, a two-dimensional (2D) autoregressive model has been used to formulate the dynamic batch processes framework. For such two-dimensional (2D) dynamic batch monitoring, a statistical online process monitoring scheme is presented in this paper. The proposed method consists of two phase: on-line two-dimensional (2D) autoregressive model building and process monitoring via SPC. In the model building phase, an adaptive lasso method is used to identify the order and coefficients of this 2D autoregressive model. In the process monitoring phase, a fault can be detected by applying SPC to the model coefficients. The simulation results show that the coefficients of 2D autoregressive model are sensitive to the faults in batch processes, verifying the effectiveness of the statistical online process monitoring scheme.