Model-free Predictive Controller Design For Several Classes Of Nonlinear Systems

In modern control theory,the controller design is usually founded on the precise mathematical model of the controlled process.However,the model-based control algorithms often cannot achieve the expected effect in the actual production application due to the unmodeled dynamics and uncertain various factors in the modeling process.In the wake of the progress of science and technology,it is difficult to get an accurate system model because the scale of industrial production is becoming larger and larger,the degree of non-linearity in the actual production process is becoming more and more complex.Model-free predictive control uses the current and historical data of the controlled system to design and analyze the controlled system.Therefore,this article consider designing control methods which are not based on precise mathematical models.In this article,suitable model-free predictive controllers are designed for several special classes of discrete high-order nonlinear systems to make the system stable.The work content of this paper can be summarized into the following three parts:First,a model-free predictive control method is presented towards a type of high-order discrete-time nonlinear systems whose models are undefined.At first,an estimation model is given by using the improved projection algorithm to approach the controlled nonlinear system.Then,on the basis of the estimation model,a predictive controller is designed by solving the finite-time domain rolling optimization quadratic function,and the explicit analytic solution of controller is also obtained,thus ensuring the stability of the system.Finally,the results of simulation reveal that the presented control strategy has excellent stability and robustness,and exists a faster convergence speed as well as more stable dynamic property compared with the model-free sliding mode control.Second,a model-free predictive control method is proposed for a kind of high order discrete nonlinear switched systems whose models are undefined.Firstly,the compact dynamic linearization technique is adopted in allusion to the subsystems,and the estimation models are designed by using the improved projection algorithm to approach the controlled nonlinear subsystems.Then,on the basis of the estimation models,the controller's explicit analytic solutions are also obtained by solving the finite time domain rolling optimization quadratic functions.Finally,an appropriate switching law is designed to make the subsystems switch reasonably in different subsystems,thus ensuring the stability of the whole system.Simulation results show the effectiveness of the control strategy.Third,a model-free predictive control method is proposed for a kind of high order discrete nonlinear asymmetric systems whose models are undefined.According to the dynamic characteristics of the asymmetric system,the compact scheme dynamic linearization technology is adopted for the positive and negative models respectively,and the improved projection algorithm is used to approximate the controlled nonlinear system and the estimation models are obtained.Then,based on the estimation models,the explicit analytical solution of the controller is obtained by solving the quadratic function of finite time domain rolling optimization.Finally,a switching law is designed to make the system switch reasonably between the positive and negative models,thus ensuring the stability of the whole system.The effectiveness of the proposed method is verified by a simulation example.