Data-driven Based Approach For Performance Assessment Of Control System Subject To Composite Disturbances

Concerning the performance assessment of SISO system,Harris’s approach of performance assessment takes MV as benchmark,and can estimate MV benchmark through FCOR algorithm.Only the priori knowledge of time delay and closed-loop routine operation data are needed to obtain the performance evaluation of the control system.The above advantages make approaches for performance assessment based on MV benchmark popular gradually,and they have been widely applied in practical industrial production.However,a lot of complex deterministic disturbances exist in the industrial scene,such as step signal,sinusoid,square wave and sawtooth wave,etc.,which makes conventional FCOR algorithm based on the assumption of stationary signals no longer suitable.This paper improves the original FCOR algorithm and restudies the performance assessment of SI SO system based on MV benchmark under composite disturbances,and proposes a method based on the improved FCOR algorithm to estimate the impulse response coefficient of closed-loop transfer function,so as to solve problems of performance assessment subject to complex disturbances.The effectiveness of this method is verified by a large number of examples.Compared with SISO system,it is more difficult to evaluate the performance of MIMO system,because it needs the priori knowledge of interactor matrix.In this paper,on the basis of the algorithm of performance assessment without interactor matrix,an algorithm based on the more easily obtained time delay matrix to calculate MV benchmark is proposed;At the same time,considering the complex disturbance circumstance of industrial scene,the conventional FCOR algorithm is improved,and the solution of MIMO system performance assessment is given.A large number of experiments demonstrate the effectiveness of this method.