(710d) A Novel Framework for Multi-Mode Process Modeling and Monitoring

With vast amounts of process data collected by the distributed control systems, multivariate statistical process control (MSPC) has been gaining significant momentum as a viable response to process monitoring and fault diagnosis issues in process industries. In most of these applications, the process is often assumed to operate at a single design point subject to disturbances and faults. In practice, however, chemical processes frequently operate at multiple regimes associated with product changeovers, capacity modifications and others. In the presence of such dynamic transitions that are part of the normal process behavior, detection and diagnosis of faults often become confounded. 

In this paper, a novel framework for process pattern construction and multi-mode monitoring is proposed. To identify process patterns, the framework utilizes a clustering method that consists of an ensemble moving window strategy along with an ensemble clustering solution strategy. Three graphical visualization tools are proposed to validate the ensemble clustering solution. Next, a new k-principal component analysis-independent component analysis (k-ICA-PCA) modeling is developed to capture the relevant process patterns in corresponding clusters and facilitates the validation of ensemble solutions. Following pattern construction, the proposed framework offers an adjoined multi-ICA-PCA model for detection of faults under multiple operating modes.

The Tennessee Eastman (TE) benchmark process is used to demonstrate the salient features of the method. Specifically, the proposed method will be shown to have superior performance compared to the previously reported k-PCA models clustering method.