Research On Internal Model Control And Disturbance Rejection For Multivariable System

With the speeding up of China's elimination of backward production capacity and technology innovation, industrial field is moving in a more efficient and secure direction. Then it puts a higher demand for industrial control technology. In the practical industrial field, multivariable systems exist widely in production process such as petrochemical, metallurgical and power production. Due to the existence of complex coupling within the system, a slight move in one part may affect the situation as a whole. So it gives a great deal of trouble to the control of multivariable system. With the industrial upgrading and the improvement of industrial product quality requirements, it also puts forward higher requirements to the accuracy of the control system.However, the existence of external disturbance will affect the control performance and accuracy of the system. The main content of this paper is to research the effective control method of multivariable systems in the case of external disturbances.This paper firstly introduces the basic principle of internal model control and disturbance observer, and then combines the internal model control structure and the disturbance observer structure to design the system which has two degrees of freedom so that the setpoint tracking performance and disturbance rejection performance could be adjusted separately. Through the compensation effect, the control system can overcome the influence of model uncertainty and have good robustness. For periodic disturbance, a modified disturbance observer design method is proposed, which effectively solves the problem of periodic disturbance rejection.This paper extends the disturbance observer structure and presents the multivariable internal model control system with the disturbance observer.The analytic method is used to design the decoupling internal model controller,and then a good decoupling control performance is achieved. Simultaneously,a multivariable fractional order disturbance observer is proposed to improve the disturbance rejection performance of the system. Through the robustness analysis of the system, the parameters of the fractional disturbance observer could be determined, which provides the basis for parameter tuning.To improve the decoupling control performance of multivariable system,a multivariable internal model controller design method based on inverted decoupling is presented in this paper. The paper promotes the inverse decoupling method to the non-square system, gets the general expression form of the decoupling controller element through deduction, and analyzes the realizability of the inverse decoupling. In addition, a modified filter is added into the feedback loop to improve the performance of disturbance rejection.The design process of this method is not complicated with the increase of multivariable system dimension, and the controller parameter is simple to solve, with the result that it gets good controls and disturbance rejection performance.