论文部分内容阅读
提出一种完全数据驱动的闭环子空间辨识及预测控制器设计方法.该方法完全由闭环系统的输入输出数据辨识子空间矩阵,通过子空间矩阵的拆分,排除了与扰动相关的模型输入,进而获取子空间矩阵参数的无偏估计;将辨识得到的闭环系统子空间矩阵描述直接作为预测模型,设计预测控制器;将其应用于某钢铁集团焦炉炭化室压力控制系统,取得了良好的控制效果.
A fully data-driven closed-loop subspace identification and predictive controller design method is proposed.The subspace matrix is identified completely from the input and output data of the closed-loop system, and the perturbation-related model inputs are excluded by the subspace matrix splitting, And obtain the unbiased estimation of the subspace matrix parameter. The closed-loop subspace matrix description of the closed-loop system is directly used as the predictive model to design the predictive controller. The predictive controller is applied to the pressure control system of a coke oven coking chamber in a steel group, and a good Control effect.