An Approach To Performance Analysis Of Correlated Control Loops

It is practical that control performance assessment results are usuallyinfluenced by loop interactions. While the internal disturbance dynamicswithin correlated loops can be recognized as the assignable causes.This paper initially analyzes the disturbance transfer dynamicswithin correlated control loops, decomposing output variance of a controlloop into two parts which in turn stemmed from its own controller and theother loops.In order to eliminate the mutual influence upon evaluation results ofcorrelated control loops, an auxiliary evaluation benchmark which takesadvantage of the interactive disturbance information between controlloops is established. With help of this auxiliary evaluation benchmark, wecould possibly find out assignable causes of degrading controlperformance when some loops suffer performance problems in the sametime.After an in depth discussion on the impact of PID parameters oncontrol performance, a PID-parameter tuning method is suggested toreduce mutual interaction of control loops in terms of controlperformance. It reveals that both itself and other correlated loops should be considered when tuning PID-parameters of a controller. Simulationresults have supported this ideaAdditionally, the proposed approaches are applied to TE processwhich provides a rich environment to analyze performance of correlatedcontrol loops. The experimental results demonstrate that internaldisturbance analysis based control performance assessment method canhelp find out problematic loops so as to further improve overall controlperformance of industrial processes.