Research On Dynamic Decoupling Internal Model Control Method For Multivariable Time-Delay Systems

In the actual chemical industrial processes, multiple time delays, strong coupling and non-minimum phase are widely present in multivariable systems. Based on the number of inputs and outputs, multivariable systems can be divided into square system and non-square system in the analysis and design of control systems. The static decoupling method is used more widely among all the decoupling methods of multivariable system, since the method is simple to design, has fast response, but static decoupling method lacks consideration in the dynamic performance of the system, resulting in unsatisfactory decoupling effects and poor anti-disturbance recovery ability. Compared with static decoupling, the dynamic decoupling method can obtain completely dynamic decoupling effect in all frequency range and obtain better decoupling effect and dynamic response. However the research on the dynamic decoupling method of multivariable time delay systems with right phase zeros is very lack, especially non-square system. Therefore, the research on the dynamic decoupling method of multivariable time delay square and non-square system with non-minimum phase zeros, has very important significance and value.Based on the internal model control principle, traditional filter in internal model controller has been improved in this paper, a modified Butterworth filter is used, the process objects containing non-minimum phase zero is analyzed, and the detailed design procedure of the internal model controller based on a modified Butterworth filter is given. Simulation represent that this proposed method allows the system to obtain a better dynamic response, making the system improve the speed and robustness more effectively.Based on the multivariable control theory in frequency domain, for the multivariable time-delay square and non-square system containing non-minimum phase zeros, the corresponding dynamic decoupling control scheme is proposed. For the multivariable square system, this paper proposed a new dynamic decoupling internal model control method based on the improved Butterworth filter, which is based on the adjoint matrix of controlled object model. Based on the basic principle of inversion, decoupling compensator is proposed firstly, after that internal model controller is obtained for the decoupling generalized object. The filter of the internal model controller use Butterworth filter (maximum flat filter), then get the integrated decoupling internal model controller. The simulation represents that:the proposed method has fine dynamic decoupling effect and dynamic response characteristics. When the model mismatch or the system is disturbed, the proposed method still shows strong decoupling effect, robustness and disturbance rejection ability.For the multivariable non-square system. In this paper, the dynamic decoupling control method of square system is extended to the non-square system. Based on the related adjoint matrix of the object model, a new dynamic decoupling method is proposed to design the decoupling compen-sator internal model controller. Firstly, according to the principle of gene-ralized inverse matrix, the decoupling compensator is skillfully designed to eliminate the coupling between the input and output. Then, the integrated decoupling compensating internal model controller is designed to deal with the unstable poles, non-minimum phase of the generalized object and so on. Finally, the feedback filter is to improve the robustness of the system, and the analysis of robust stability and sensitivity are given. The simulation research show that this method has significant dynamic decoupling effect, which simplifies the controller design process of the dynamic decoupling, and increases the accuracy and robustness of the controlled system.