Design of the Neural Network Angular Velocity Compensator based on Modeling of a Hydraulic Servo Motor System

Abstract

This paper presents the angular velocity control of a hydraulic motor by adaptive learning using the Neural Network, which is located in the feedback loop. This system features a precise steady performance, and has robustness. In the report, at first, the mathematical model is introduced from the basic equations on a servovalve and hydraulic motor. Based on this model, various measurement results of the static characteristic determine the design parameters that will be needed for the hydraulic servo control design. The open loop dynamic response experimentally reveals that the mathematical model is extremely appropriate. Furthermore, as it is necessary that the better performance will be accomplished, a Neural Network compensator is employed instead of the state feedback one. The experimental results demonstrate the effectiveness on robustness by means of the adaptive learning of Neural Network, even when the disturbances such as the leakage, additional torque, and supply pressure are intensively provided.