改进多变量广义最小方差解耦控制方法及应用

Improved Multivariable Generalized Minimum Variance Decoupling Control Method and Its Application

针对现有多变量广义最小方差解耦控制方法不能直接抑制有界可测干扰的问题,以及另加前馈控制器带来的结构复杂且成本增加的难题,提出一种改进的多变量广义最小方差解耦控制方法。所提方法通过在加权多项式矩阵和丢番图方程中的多项式矩阵同时引入积分项,用于直接抑制工业过程中有界可测干扰对系统的影响,从而消除稳态误差,提高控制精度。数值仿真以及在双容水箱液位控制半实物仿真平台的实验验证表明了所提方法的有效性和实用性。

In view of the problem that the existing multivariate generalized minimum variance decoupling control method can not directly restrain the bounded measurable disturbance,and the complex structure and increased cost brought by feedforward controller,an improved multivariable generalized minimum variance decoupling control method is proposed.By introducing the integral terms in the weighted polynomial matrix and the polynomial matrix of the Diophantine equation,the proposed method is used to directly restrain the influence of bounded measurable disturbance on the system in the industrial process,thus eliminating the steady-state error and improving the control accuracy.The numerical simulation and the experimental verification on the semi-physical simulation platform for the control of the liquid level in the double capacity tank demonstrate the effectiveness and practicability of the proposed method.