The Performance Diagnosis Of Multivariable Control System Based On Covariance Analysis

First of all, the paper makes analysis of the performance assessment benchmark and method on the multivariable control system. And then use the covariance matrix of closed-loop control system output data for multivariable control system performance assessment based on the historical data benchmark. The performance assessment index I_v, used to detect and quantify the degree of change of the control system performance. Make a matlab/simulink experiment on a general four-input and four-output multivariable control system to verify the validity and feasibility of this assessment method. Make analysis of two sets of operating data from a unit of a combustion control system before and after the thermal power plant maintenance to verify the calculation results using the proposed method is the same with the actual situation. And then use the cosine of the angle between the control loop variables and the performance subspace to diagnose the performance of each control loop changes.We use the historical data benchmark for performance assessment, the benchmark data could be a period of"golden"operation data from the process with satisfactory control performance. For example, the benchmark data could be a period of operation data after a satisfactory controller tuning or updating, which may make the benchmark data does not representative the best control performance and then may lead to wrong conclusion in control system performance assessment. Through the benchmark optimization method, replace the benchmark data with better data, so as to find a more ideal new benchmark data, which makes the control system performance assessment more credible.The algorithm,which takes Visual C++ as the development platform and performance evaluation software as the base, has been programmed to realize the function and embedded the nonlinear diagnosis sub-module into the performance evaluation software.After that,test the overall performance of the software to reach the goals to be achieved and complete the functional requirements.