MultivariateControl Performance Assessment and ControlSystem Monitoring via Hypothesis Test on Output Covariance Matrices.

Controlloops widely exist in industrial processes, whose performancedirectly influences the efficiency, safety, and product quality ofproduction plants. Therefore, control performance assessment (CPA)and control system monitoring (CSM) are critically important for industrialprocessing. In consideration of multivariate control systems, thecovariance matrix of closed-loop outputs plays an important role inboth CPA and CSM. Existing methods mainly focus on comparing tracesor determinants of the output covariance matrices, which only utilizepartial information contained in the matrices. As a result, the assessmentand monitoring results may be misleading. In this paper, a multiobjectivescheme is proposed for both CPA and CSM of multivariate control systems,which takes the entire covariance matrices into account by conductinga hypothesis test on the equality of the matrices. To fulfill thepresupposition of such test, autoregressive moving-average (ARMA)filters are established to remove the autocorrelation contained inthe closed-loop output data. The developed scheme can be divided intothree aspects: CPA using a minimum variance (MV) benchmark, CPA usinga user-specified benchmark, and CSM based on historical data. Casestudies show that, compared with the conventional approaches, theproposed method provides more abundant information and achieves betterresults. [ABSTRACT FROM AUTHOR]