基于历史性能基准的模型预测控制性能监控  

Performance monitoring of model predictive controller based on historical benchmark

作　　者：张荣金[1] 张泉灵[1] 

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

出　　处：《计算机与应用化学》2011年第3期275-278,共4页Computers and Applied Chemistry

基　　金：国家高技术研究发展计划资助项目(863计划)(No.2009AA04Z162)

摘　　要：针对实际工业过程中希望只利用过程的日常运行数据对控制系统进行监控的情况,提出了1种用历史性能基准对预测控制系统进行性能评估、监控以及分析的方法。利用运行状态良好的1段历史数据计算出历史性能基准,并将历史性能基准与实际性能的比值用于控制器性能的实时监控。根据历史性能指标的残差监控检测出性能变化时,进一步通过历史数据协方差监控及广义特征值分析,区分出性能显著变差或变好的方向和子空间。为现场工程师提供性能变化的一些原因,用于系统维护。最后通过Shell重油分馏塔仿真证明该方法可以获得很好的评估与监控结果。To solve the problem that minimum variance control benchmark is not achievable by predictive control and the calculation of the interactor matrix is complicated, a novel method based on the historical benchmark for performance assessment and monitoring of model predictive control system is proposed ＂mthis paper. Firstly the benchmark is obtained by a period of＂golden＂ operation data. Then a performance measure based on theratio.of the historical benchmark and achievedperformance is used for the real-time monitoning. When performance changes are detected by residual momtoring of historical performance index, data-based covariance monitoring and generahzed eigenvalue diagnosis are used to determine the directions and subspaees with significantly worse or better performance versus tile benchmark. Scme reasons of the performance change are obtained tor control engineers. Finally, simulation was done aboute a heavy oil fractionator system and good results show goodperformances of this method.

关 键 词：模型预测控制 性能监控 历史性能基准 广义特征值分析 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]