基于加权偏离度统计方法的预测控制性能评估算法

Control performance assessment of model predictive control based on statistics analysis of weighted points

针对带区域约束条件的预测控制系统性能评估问题,在考虑过程输出变量约束类型的基础上,提出了基于加权偏离度统计方法的控制性能评估算法。该方法依据控制要求的不同,将输出变量分为质量变量和约束变量,并结合工程经验合理选择变量的权重。基于系统闭环运行数据和约束设置,通过计算变量的加权偏离度得到控制系统的性能评估指标,从而为预测控制器的参数调整和性能提升提供了决策依据。系统仿真实例和工程应用证明了该评估算法对区域预测控制系统性能评估的有效性。

With the goals of optimal performance,energy conservation and cost effectiveness of process operations in industry,controller performance assessment of process control have received great attention in both academia and industry.Controller performance monitoring and assessment are necessary to assure effectiveness of model predictive control systems and consequently safe and profitable plant operation.Taking region constraint properties of model predictive control（MPC）into consideration,a novel approach based on the weighted point statistics is developed for the performance assessment of region control（RC）problems in this study.All controlled variables are grouped into two categories：constrained variables（CVs）and quality variables（QVs）,according to the importance of the controlled variables in MPC systems.By introducing the weighted points for all controlled variables,the performance index is calculated based on the statistics analysis of closed-loop data sets,which can be used to assess the performance of an MPC control system.The important advantage of the proposed approach is that just the routine closed-loop operation data of the system and constrained region of each CV are required,which is convenient for the industrial applications.Simulation example and industrial case study illustrate the applicability of the proposed approach.