基于神经网络的带钢跑偏电液伺服系统研究

Study on Electrical-Hydraulic Servo System for Steel Strip Deviation Based on Neural Networks Control

带钢跑偏电液伺服控制系统的非线性和时变性使得传统的PID控制很难达到理想的控制效果,将神经网络与普通PID控制相结合形成神经网络自适应PID控制策略,应用于该系统实现其良好控制。为提高系统的动态响应速度及性能,采用RBF神经网络对系统进行辨识预测。首先建立带钢跑偏电液伺服系统数学模型,然后利用AMESim和Simulink软件对传统PID控制和神经网络自适应PID控制进行联合仿真。结果表明,神经网络自适应PID控制系统响应速度快、超调量小、鲁棒性强,并具有良好的稳定性和控制精度。

The nonlinearity and time-variance of electrical-hydraulic servo system for steel strip deviation make traditional PID control hard to achieve the ideal control effect it. The combination of neural networks control and general PID can form neural network self-adjusting PID control strategy to realize good control of electrical-hydraulic servo system for steel strip deviation. To improve the dynamic response speed and performance of the system, RBF neural network is adopted to indentify and predict the system. The mathematical model of the electrical-hydraulic servo system for steel strip deviation is established, then AMESirn/Simulink software is used to do the united simulation of the traditional PID and neural network adaptive control. The result shows that neural network adaptive PID control system has fast response, small overshoot, strong robustness and good stability and control precision.