Study On Neural Network Self-Tuning PID Control Method For Electro-Hydraulic Servo System

With the development of automation technology and continuous improvement of industrial automation, higher control quality requirements to the electro-hydraulic servo system are being put forward, such as stability, rapidity, accuracy and self-adaptive control. In order to meet these requirements, on the one hand, the technological level and the quality characteristics of the manufacturing process of hydraulic components should be improved; on the other hand, the designing of a high quality controller has become the key technology. PID control as a main control method is widely used in the field of the practical engineering systems. But when these systems have large changes in model parameters or has a strong influence of nonlinear or the system is time-delayed, the drawbacks of PID control method are becoming obvious, parameters tuning is difficult,control quality and system robustness can not attain a satisfactory result.Depending on characteristics of electro-hydraulic servo control system, a study on neural network intelligent PID control strategy is proposed, setting PID parameters online using neural network for the sake of improving the dynamic state quality.Firstly, a study on the relevant theoretical issues of the neural network self-turning PID controller is conducted. The controller is composed of the classical PID and neural network. The classic PID is charged with the direct object of the closed-loop control; while the neural network system is running according to the status, regulating weight coefficients through online self-study and parameters of K_p, K_i and K_d real-time, so that the neural network controller can output the optimal PID controller parameters.Secondly, it is set up of the mathematical model of position servo electro-hydraulic material testing system, relevant parameters of the system are selected and calculated, simulation models are set up using Simulink software package. Two different strategies of conventional PID and neural network self-tuning PID control are used separately to get theoretical analysis and experimental study of the performance of the system.Thirdly, the computer system is controlled by applying LabVIEW graphical programming software. Separately using two different strategies of conventional PID and neural network self-tuning PID control, to compare the performance and investigate the experiment.Theoretical analysis and experimental results show that, neural network self-turning PID control system can effectively improve the dynamic quality of the system, highlighting the advantages of improving dynamic quality of the control system under the control strategy.