基于LQG基准的预测控制器经济敏感度分析及调节准则

LQG-based Sensitivity Analysis and Tuning Guidelines in Economic Performance Assessment of Predictive Controller

为考虑控制变量方差变化对控制器经济性能的影响,提出一种基于线性二次高斯(Linear Quadratic Gaussian,LQG)基准的预测控制器经济敏感度分析方法及调节准则.首先由子空间辨识算法推导出带输入输出变量加权的LQG基准一般描述形式,在此基础上,构造了基于方差调节和基于约束松弛的两个优化问题进行敏感度分析,最终求解得到敏感变量的方差调节量和约束松弛量以提高控制器的经济效益,Shell塔仿真实验结果表明本文方法的有效性.

In order to take the inflnence of input wlriance on the economic performance of predictive controller into account, an LQG-based sensitivity analysis method and tuning guidelines have been proposed. First, a general formulation of LQG beneh- marking with weighting matrix for input and output variables is derived through the subspace identification algorithm. Fur- thermore, based on variability tuning and constraint loosening, two optimization formulas have been built, respectively. The variance overshoot and constraint relaxation resulting from the above optimization problems can improve the economic benefit. Simulation study of the Shell tower is carried out to demonstrate the efficiency of the proposed approach.