Research Andapplication Of Model-free Dynamic Matrix Control Method

The establishment of controlled system mathematical model is thebasis of modern control theory applications. Mathematical model of thecontrol system establishment, especially for nonlinear system modelingproblem, is still a thorny issue nowadays. Even if a mathematical modelof the control system is managed to be established with many simplifies,it is difficult to achieve the desired control effect. Model-free adaptivecontrol method is an effective control method which has overcome thisessential problem in modern control theory. In this paper, more in-depthresearch has done based on model-free adaptive control algorithm andseveral achievements has been conquered:1. In discrete-time nonlinear SISO system, to face the defect ofMFAC algorithm only using the zero-order information of thecontrolled system I/O data fully but without it system information, amodel-free dynamic matrix control method based on linearization oftight format is proposed. This method uses the tight format linearizationmethod to turn the nonlinear system into a linear system described bythe pseudo-gradient vector. With the help of the rolling optimizationand feedback correction strategy, the analytic form of the model freedynamic matrix control rate is derived on this basis. Because of acontrolled system I/O data within the system with information, themethod has improved the accuracy of the prediction model andenhanced the robustness of the system greatly in order to improve theperformance of the controller effectively.2. On the basis of MFAC algorithm research for nonlineardiscrete-time SISO system, the proposed model-free dynamic matrix control method is extended to nonlinear MISO systems, anddemonstrates its validity by MATLAB simulation.3. To face the problem of delay and uncertainty which is exist inthe process of pH neutralization with obvious nonlinear characteristics,model-free dynamic matrix control method in pH neutralization processis proposed and Simulation result verifies its effectiveness andsignificantly better than traditional PID control method.