Performance Assessment Of Control Systems With Varying Time-delay

There are hundreds of control loops in a typical industrial process,which are not fully adjusted,while market competitiveness strives the process of control systems to achieve optimal state.So the performance assessment of control systems is expected to a kind of effective tool to monitor performance for normal production.In this thesis,the performance assessment of control systems with varying time-delay is researched so that effective ideas and methods of assessment are provided.The main contents of this thesis include as follows:(1)Under the condition of unknown time delay information,the impulse response coefficients of process model,impulse response coefficients of noise model and the noise covariance can be identified by subspace identification algorithm with closed loop operating data.Then LQG benchmark is applied to assess performance.It is verified by simulation examples that the LQG benchmark with subspace identification can be used for performance assessment without the delay information.However,this method can not avoid the effect of varying delay on the system performance completely,it can only be used as an approximate estimation.(2)In order to accurately evaluate the variable time delay process,minimum variance performance benchmark is introduced.And considering the delay information is one of the key elements of the minimum variance benchmark,so strong tracking filtering algorithm is given to estimate variable delay.It can not only track the change of the varying delay,but also know the change time,which lays the foundation for the on-line evaluation of control systems with varying time-delay.Also,It is also extended to the method of combined estimation of time-varying delay and system parameters.It is verified the accuracy of the algorithm.(3)Based on the minimum variance benchmark of linear regression algorithm,the minimum variance benchmark of recursive linear regression algorithm with forgetting factor is improved.It is effectively decreases the computational complexity,realizes the online assessment,but also enhances the current data information in the algorithm,makes error variance estimation accurate.And combined improved algorithm and varying delay estimation based on strong tracking algorithm proposed in the third chapter,the variable delay process of control system online assessment algorithm is suggested.the simulation verifies the effectiveness and the use of the algorithm.(4)With the combination of performance assessment for single loop control system and performance assessment for control system with variable delay,a set of assessment performance software for control system is developed by MATLAB/GUI.