The Research Of Internal Model Control Method For Complex System

The non-minimum phase systems are widespread in the field of industrialprocess. The non-minimum phase factors cause serious influence for thestability of control system. Moreover, most of process objects also have thecharacteristics of time delay and strong coupling of multivariate. However, thetraditional control method will not be able to achieve the desired control result.Therefore, the research of such complex system has import theoreticalsignificance and application value.In this paper, based on multivariable frequency domain control theory,the internal model control is taken as the main means of control. Aiming at themultivariable system containing the time delay and right half plane zeros, thesimple and practical decoupling control method is proposed. The structureequivalent of the internal model control and Smith predictive control isdiscussed in detail. In addition, this univariate type of Smith internal modelstructure is extended to the multivariable system. Smith predictor is designedby internal model method, which not only solves the time delay problem, butalso effectively improves the weakness of Smith predictive control in the caseof model mismatch. For square system, static decoupling and dynamic decoupling are respectively designed. The structure of multivariabledecoupling control is given. The designed procedure of decoupling controlleris deduced in this paper. By using the Moore-Penrose inverse, the decouplingproblem of non-square system is better solved. The model order reduction forthe generalized controlled object simplified the design process of thecontroller. The simulation results show that this method effectively improvedthe dynamic performance and robust stability of control system. Meanwhile,this paper quantitatively evaluates the design method using four commonperformance indicators, which verifies the superiority of this method.Aiming at the high-order system in the industry, the internal modelcontrol method based on the Butterworth filter is proposed. In the paper, usingthe ideal low-pass characteristic of the Butterworth filter, as well as its moststable characteristics of ups and downs of amplitude-frequency characteristicswithin the whole frequency pass band, the optimized and improvedButterworth filter replaces the low-pass filter designed for conventionalinternal model controller, which makes the better control quality for controlsystem. The simulation studies show that the proposed method can effectivelyimprove dynamic quality and is robust against model mismatching. It issuitable for the control of high-order system.