The Research Of The Model-free Adaptive Control Method Of Nonlinear Systems

Since 40 years from the 20th century, classical control theory developed fast, and formed a mature theory system. Classical PID controller has been widely used in the petrochemical, electric power, metallurgy, textile and other industries. However, with the development of the petrochemical, electric power, metallurgy, textile and other industries, equipment becomes more complex and system is bigger, so the traditional PID controller can not fit the requirements any more. With the development of the modern control theory, scholars from different countries proposed many advanced control algorithms, such as predictive control, adaptive control, neural networks and robust control. With the development and improvement of the advanced control algorithm, they have been applied in some areas.However, these advanced control methods are all based on the model of the controlled system. If we can not get the model of the controlled system, we can not design the advanced controller. For avoiding the problem, scholars proposed model-free control method. We named the controller which are designed only by input and output date without any information of the model of the controlled system as model-free controller. Until now, the main model-free control methods are model-free adaptive control, active disturbance rejection control, self-learning control, iterative feedback tuning, and so on.This paper first introduces the research significance and present situation of model-free control method and basic theory of model-free adaptive control algorithm. Then the model-free adaptive control algorithm is extended to the MIMO nonlinear systems. Simulation result shows that the proposed algorithm is an effective strategy with excellent tracking ability and strong robustness. Then a model-free predictive control algorithm based on linearization of partial format is proposed for a class of nonlinear systems which are described by NARMAX model. Simulation result shows that the proposed algorithm is an effective strategy with excellent tracking ability and strong anti-jamming capability. At last, model-free predictive control algorithm and model-free adaptive control algorithm are used to control CSTR. Simulation result shows that model-free predictive control algorithm is an effective strategy with better tracking ability and stronger anti-jamming capability than model-free adaptive control algorithm.