基于子空间方法的模型失配检测研究  

Research on model-plant mismatch detection based on subspace approach

作　　者：陈贵[1] 杨江[1] 谢磊[1] 张荣金[1] 

机构地区：[1]浙江大学智能系统与控制研究所,工业控制技术国家重点实验室,浙江杭州310027

出　　处：《化工学报》2011年第9期2575-2581,共7页CIESC Journal

摘　　要：引言 随着工业过程的不断复杂化和对控制系统高性能的要求,控制系统性能评估与监控以及故障诊断成为现代控制系统不可或缺的要素。过去的二十年中,性能监控与评估技术已经取得了很多成果,但是分析性能下降的原因及故障诊断仍然有很多问题亟待解决。性能下降的原因主要可分为以下几个方面：As a key factor in model-based control technique,model fidelity has significant influence on control performance.Model-plant mismatch（MPM）is an important step in the procedure of control performance monitoring and system maintenance.A subspace-based MPM approach for control systems was proposed.First,the residual was constructed through subspace transformation.Then the residual was analyzed by using statistical local approach,from which an indicator could be constructed to detect the mismatch.Using orthogonal subspace,it was not needed to estimate system states and residual could be obtained only by using input and output data.Furthermore,based on the state space model,the proposed approach was suitable for MIMO process.The effectiveness of the proposed approach was illustrated by both simulation and industrial applications.

关 键 词：模型失配检测 子空间 统计局部方法 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]