论文部分内容阅读
In this paper,a historical objective function benchmark is proposed to monitor the performance of data-driven subspace predictive control systems.A new criterion for selection of the historical data set can be used to monitor the controller s performance,instead of using traditional methods based on prior knowledge.Under this monitoring framework,users can define their own index based on different demands and can also obtain the historical benchmark with a better sensitivity.Finally,a distillation column simulation example is used to illustrate the validity of the proposed algorithms.
In this paper, a historical objective function benchmark is proposed to monitor the performance of data-driven subspace predictive control systems. A new criterion for selection of the historical data set can be used to monitor the controller performance, instead of using traditional methods based on prior knowledge. Undefender this monitoring framework, users can define their own index based on different demands and can also obtain the historical benchmark with a better sensitivity. Finally, a distillation column simulation example is used to illustrate the validity of the proposed algorithms.