Research On Linear Control Method Of Nonlinear Systems Using Disturbance Observer

Because of the complexity of nonlinear system and the universality of nonlinearphenomena, the traditional approximate processing method for nonlinear systems can notmeet the requirements of modern industrial control. Therefore, many control methods fornonlinear system have been proposed in recent years. The most commonly used method is toapproximate a nonlinear system by a linear system or convert a nonlinear system to a linearsystem, then the linear system theory is applied to finish the analysis and synthesis, and thiskind of method is called linearization method. While most of the linearization methods havesuch characteristics as algorithm complexity, inconvenience for engineering implementation,dependence on accurate mathematical model, etc, in varying degree. This thesis study alinearization method for nonlinear system, called linear time-invariant disturbance observerfeed forward control method. This method has many advantages, such as no dependence onaccurate mathematical model, no involvement of advanced mathematical tools, lineartime-invariant controller. Thus, it is easy-mastering for engineering technicians. This studyprovides a simple and effective design approach for the linearization of nonlinear system. Themain contents of the thesis are as follows:(1) Based on previous research work, the controller design issue of this method is studiedunder discrete-time system. And under the discrete condition, observability theorem ofsubstituting model is proved. The simulation results indicate that the design of discrete-timecontroller is correct and effective, making the method available in discrete-time system.(2) Based on the research of the discrete controller, the influence on dynamicperformance of composite system is researched with varying controller parameters. Thequalitative relationship between the controller parameters and the dynamic performance ofcomposite system is obtained, providing the basis for parameter selection of the controller,and the shortage of the previous controller parameter selection can be make up.(3) An important theoretical problem of this method, namely the existence of thesubstituting model, is studied. Through the study of reversible nonlinear systems, theexistence theorem of the substituting model is put forward and proved. Comparing thesimulation results, it shows that the reversible nonlinear system can be controlled by usingthis method, and takes good control effect.(4) Through the deformation of dynamic composite system model, the model issimplified as a feedback closed loop system. On this basis, there comes up with the stabilityconditions of the composite system when the controlled object is of a kind of non-memorynonlinear system. The simulation shows the validity of the stability conditions. (5) Through analysis of the vector control system, using the decoupled linear system byvector control (the equivalent DC motor model) as the target model, the AC speed regulationsystem is decoupled by the linear time-invariant disturbance observer feed forward controlmethod. The simulation results show the performances of this method and vector control arealmost the same, while the structure and realization of the controller of this method aresimpler.