针对工业过程中常见的多变量时滞输入输出系统，基于内模控制结构提出了一种解析解耦内模控制器矩阵的设计方法，该方法设计的内模控制器同时具有解耦器和控制器的作用。其优点是能够实现标称系统输出响应之间的近似或完全解耦,且用遗传算法对于解析模型进行降阶或近似处理，从而简化了推导解耦控制器矩阵的运算量；并可以通过调整内模控制器的滤波器时间常数来补偿被控过程的模型不确定性所导致的控制品质下降。用该方法对国产200MW火电机组带汽—汽换换热器的多变量强耦合再热汽温控制系统进行了控制系统的设计和仿真研究，仿真结果证明了该方法的有效性。并且在过程模型和过程失配时显示了较强的鲁棒性和抗干扰能力。

A new method of controller decoupling based on the internal model structure has been proposed for applications in multi-input multi-output (MIMO) time delay systems in industrial processes. The internal model controller (IMC) has been designed with both decoupler and controller. The advantages of the new method are that it can realize the almost or complete decoupling of system outputs, and the process for calculating the load on the decoupling controller matrix is simplified by using model reduction based on a genetic algorithm. As result, the poor control performance caused by the mismatch between plant and the process model can be improved by adjusting the time constant of the filter. The method was used in a multivariable strong coupling reheating temperature control system with a steam-steam heat exchanger in a 200 MW unit. Computer simulation results show that even when the model is mismatched, the system can still be highly robust and possess anti-interference properties and demonstrate the efficacy of the new method.