连续时间域内具有积分作用的广义预测控制（CIGPC）

文献1]推导了连续时间域内的广义预测控制规律,但由该方法所获得的控制作用不具有积分作用。因此,对于大量具有自平衡能力的过程,当设计模型的比例增益同实际过程存在误差时,将导致系统输出的稳态偏差,本文在文1]的基础上提出了一种具有积分作用的广义预测控制方法,仿真结果表明,本文的方法不仅能够消除系统的稳态偏差,而且可使系统具有较好的控制品质。

Continuous-Time Generalized Predictive Control with Inherent Integral Action (CIGPC)

The reference 1] proposed a method of continuous-time generalized predictive control (CGPC). However, the controller of CGPC has no inherent integral action. Hence, for the majority of self-balanced processes , CGPC will cause the system offset closed-loop behaviour while the static gains of the designed model and the practical process model are different. Based on the reference 1], this paper presents a novel approach of continuous-time generalized predictive control with inherent integral action (CIGPC). The simulations show that the controller proposed in the paper not only can eliminate the static deviation of the output but also maintain the system to have good control performance.