On Performance Assessment Of Switched Control Systems Based On Tensor Space

Industrial control systems performance assessment is the key to guarantee process stationarity and safety running.Control systems performance assessment can detect and diagnose the abnormal states and faults of process and control equipments in advance,and implement effective maintenance of control systems and optimization of controller parameters,which the timely and appropriate measures can be taken to avoid the nonstationarity and unsafety of process.The research on control systems performance assessment have been obtained great development since 1989.However,the study on performance assessment of hybrid automatic control systems being consisted of control systems and protection systems still remains an open topic so far.In this dissertation,the performance assessment problem of switched control systems has been studied based on tensor space theory,which has the important theory significance for design and optimization of control systems and protection systems and facilitates to improve the safety and reliability of safety-important systems.The main contents are summarized as follows:(1)Based on tensor space,a performance benchmark estimation method is proposed for switched control systems.In the tensor space,the interacting relation of control systems and protection systems can be synthetically represented by adding the logical switching space,which can describe the high coupling interaction of control systems and protection systems.The data-driven tensor space algorithm based on higher-order singular value decomposition has been constructed to assess the performance of switched control systems.Simulation examples demonstrate that the tensor space approach is more rational and effective than the multiple model approach.(2)The switched systems tensor-product representation theorem has been proposed for multivariate switched linear systems.The stability property of switched systems tensorproduct representation is investigated.The minimum variance control benchmark is constructed in the tensor projection.The framework for describing the switched systems dynamic behavior and performance analysis of switched control systems has been presented.(3)The covariance based tensor method is proposed for switched systems controller performance assessment.The covariance tensor-based performance assessment method has been constructed and the calculation procedures have been derived from the sets of singular values of the sampling data tensor.Simulation examples and an industrial application have validated the effectiveness of the proposed method.(4)Based on tensor space approach,the effects of logical switching sequences between control systems and protection systems on the process safety performance have been studied.The corresponding computing procedures based on tensor space approach for estimating the performance indices are implemented.Simulation results demonstrate the feasibility and effectiveness of the tensor space approach for protection systems performance assessment.