In recent years,with the development of information technology,the requirements of industrial control system are becoming higher and higher.In practice,the characteristics of strong coupling and nonlinearity in each process of the control system make it difficult to establish an accurate model of the controlled object.It becomes more difficult to design modelbased control methods.Model-free adaptive iterative learning control as a data-driven approach has attracted more and more attention.Combining the advantages of model-free adaptive control and iterative learning control,the design task of the control system can be realized by using only input and output data.The error convergence rate of model-free adaptive iterative learning control has always been the focus of research.The main factors affecting the error convergence rate are initial parameters,control signal design,disturbance sum,data loss,etc.This paper studies the parameter estimation,data loss,and interference that affect the rate of error convergence.The main contents are as follows:(1)Aiming at the problem that different initial values of pseudo-partial derivatives have a great influence on the convergence rate,a high-order model-free adaptive iterative learning control method is proposed.The proposed method uses multiple historical data to estimate the value of the pseudo-partial derivative under the current batch,and then updates the controller algorithm with the estimated pseudo-partial derivative value.This method improves the estimation accuracy of the pseudo-partial derivative and weakens the influence of the initial value of the pseudo-partial derivative on the error convergence rate,so as to obtain a better control effect.Numerical simulation and permanent magnet linear motor simulation verify the effectiveness of the method.(2)Aiming at the influence of the initial value of partial derivative on the convergence rate of the modeler adaptive iterative learning control algorithm,a modeled adaptive iterative learning control method based on high-order partial derivative estimation is proposed.The proposed method uses multiple historical data to estimate the pseudo-partial derivative values under the current batch,and then updates the controller algorithm with the estimated pseudopartial derivative values.This method improves the estimation accuracy of pseudo-partial derivatives,and can converge quickly even if different initial values of pseudo-partial derivatives are selected,so as to obtain better control effects.Numerical simulation and displacement control simulation of permanent magnet linear motors verify the effectiveness of the method.(3)A model-free adaptive iterative learning control method based on iterative extended observer was proposed to reduce the control performance of unsaturated polyester resin due to interference.Firstly,a nonlinear term containing interference is added to the dynamic linearization,and then an iterative extended observer is introduced to estimate the nonlinear term as disturbance compensation and eliminate its influence in the control.The proposed algorithm can guarantee a good control effect under strong interference.Finally,the effectiveness of the method is verified by simulation experiments. |