Research On Control Method Of Five-cylinder Hydraulic Press Based On Model-free Adaptive Prediction Algorithm

In our country’s modern industrial production,hydraulic presses are used for processing and manufacturing in many fields of mechanical production due to their advantages such as sensitive response,fast speed,large output force,and good safety performance.It is one of the main tools for forging and making components.According to the production requirements of the process,it can adjust the pressure required for the sliding block to descend,and adjust the movement range of the sliding block during operation.Most of the existing research results are to first establish an accurate mathematical model for the hydraulic press system,and then design a corresponding control algorithm based on the mathematical model,and use this algorithm to control the slider motion of the hydraulic press.The multi-cylinder hydraulic press is a complex nonlinear system.Due to nonlinear factors such as the hydraulic system and hydraulic components,the model structure is complex,strong coupling,and there are problems such as unknown disturbances.Although the requirements for control accuracy and response speed are relatively high,it is difficult to establish an accurate mathematical model,and the parameters are not only easy to change but also difficult to determine.Therefore,it is usually difficult to control the multi-cylinder hydraulic press.The model-free adaptive control method is a data-driven control method that only uses the input and output data obtained from the controlled system to design and analyze the controller,instead of relying on the mathematical model of the controlled system for controller design.It is of great theoretical and practical significance to design a model-free adaptive predictive control method by synthesizing the respective advantages of the model-free adaptive control method and the predictive control method.This method can use dynamic linearization technology to create a virtual dynamic linearized data model at the operating point of the controlled system,and then on the basis of the obtained virtual data model,the idea of rolling time domain optimization is used to design model-free adaptive.Predictive controller which can effectively avoid the problem that the multi-cylinder hydraulic press is difficult to accurately model.Aiming at the above problems of the five-cylinder hydraulic press,this paper studies the application of the model-free adaptive predictive control in the five-cylinder hydraulic press.The main research contents are as follows:（1）According to the characteristics of the five-cylinder hydraulic press model,a model-free adaptive predictive controller based on full-format dynamic linearization is proposed,and the convergence proof of the proposed controller is given.（2）Considering that five-cylinder hydraulic presses are widely used in industry and are generally repeatable,the model-free adaptive predictive control and iterative learning control are combined to propose a full-format model-free adaptive control with an iterative learning outer loop.It also prove the convergence of the proposed control scheme.（3）For a model of a five-cylinder hydraulic press,in the case of unknown model information such as piston rod mass and viscous damping coefficient,the input and output data are used to conduct MATLAB simulation verification of the above control strategy.On this basis,the LABVIEW development environment is used to control the five-cylinder hydraulic press.By observing the obtained results,it can be confirmed that the proposed control strategy can achieve accurate displacement tracking and diagonal cylinder leveling,and obtain better control performance.