Monitoring process transitions by Kalman filtering and time-series segmentation

Time-series segmentation algorithms, such as methods based on Principal Component Analysis (PCA) and fuzzy clustering, are based on input-output process data. However, historical process data alone may not be sufficient for the monitoring of process transitions. Hence, the key idea of this paper is to incorporate the first-principle model based state estimation into the segmentation algorithm to detect changes in the correlation among the state-variables. For this purpose, the homogeneity of the segments is measured using a PCA similarity factor calculated from the covariance matrices given by the state-estimation algorithm. The whole approach is applied to the monitoring of the industrial production of high-density polyethylene.