Model-free Adaptive Predictive Algorithm And Its Application In Nonlinear Process

Because most of the traditional predictive control requires the accurate model of the controlled plant, and the majority of control theory is for linear system, so when facing the complex industrial control process, which is difficult to establish its accurate mathematical model,the control performances is not ideal. In order to solve this problem, according to the model-free adaptive predictive control algorithm, this paper adopts the data driven control method to establish the dynamic process of the whole control system, which only using input and output data of the closed-loop system and has nothing to do with the middle of the state variables. This paper also has made improvement to the algorithm, realizing the model-free adaptive PI predictive control by combining the excellent characters of the generalized predictive control and model-free adaptive predictive control. The results of the theoretical simulation illustrate that the control effect is very ideal.In order to further verify the effectiveness and practicality of the theory, this paper also does the physical experiments which aim at the predictive control of the fluid temperature with the distribution parameters. Building the whole experimental platform by combining the PLC, OPC and MATLAB, and realizing the data exchange between PLC and MATLAB through the Simulink toolbox and OPC toolbox from MATLAB.From the results of sufficient experiments, comparing the performances of the model-free adaptive control and model-free adaptive PI predictive algorithm, it is observed that the control performances of the latter are much more ideal. It can be used in the industrial processes which are difficult to establish the mathematical models and have delayed systems.