Performance Monitoring of MPC Based on Dynamic Principal Component Analysis

Abstract

Abstract:A unified framework based on the dynamic principal component analysis (PCA) is proposed for performance monitoring of constrained multi-variable model predictive control (MPC) systems. In the proposed performance monitoring framework, the dynamic PCA based performance benchmark is adopted for performance assessment, while performance diagnosis is carried out using a unified weighted dynamic PCA similarity measure. Simulation results obtained from the case study of the Shell process demonstrate that the use of the dynamic PCA performance benchmark can detect the performance deterioration more quickly compared with the traditional PCA method, and the proposed unified weighted dynamic PCA similarity measure can correctly locate the root cause for poor performance of MPC controller.