CONDITION MONITORING OF MODEL PREDICTIVE CONTROL SYSTEMS USING MARKOV MODELS

Control monitoring has been an active research over the last decade. However, most work has focused on control performance monitoring with certain optimal control as the benchmark, and can not be applied directly to the condition monitoring of model predictive control systems. In this paper, a Markov modeling approach for control monitoring is presented. Not only does it address control performance monitoring problem but it also monitors stability and economic performance of model predictive control systems. The basic concepts of Markov chains are reviewed. The applications of Markov chains to control condition monitoring problem are addressed. An industrial case study is demonstrated and the value of the proposed approach is evaluated.