BP神经网络PID控制对液压控制系统的改进

Improvement of hydraulic control system based on BP neural network PID control

闭式压力机的滑块在与工件接触过程中需要精确的位移控制,基于闭式压力机液压伺服系统的大惯性、时变性、高度非线性以及无法获得相对精确的数学模型等特点,单纯的PID控制器很难达到理想的控制效果。将BP网络与传统PID控制器结合,通过实时检测滑块在不同时刻的实际值、设定值和误差值,利用神经网络的在线自学习能力,在线调整PID参数KP、KI、KD,实现PID参数的最优组合,而且能够实现比较精确的滑块位移控制,控制性能明显优于传统PID控制器。通过仿真曲线对比分析可知,BP神经网络PID控制器对液压伺服控制系统的改进是非常有效的。

The slider of closed press needs precise control of displacement when contacted with the workpiece,based on the characteristics of the large inertia,time-varying,high nonlinera and differenty in obtaining relative precise mathematical model of hydraulic servo system for closed press,the ideal effect can not be achieved easily by traditional PID controller. Combined BP network with traditional PID controller,the optimal combination of parameters and more precise slider displacement control were realized by the real-time detection of slider＇s actual value,set value and error value at different time with the help of the on-line self-learning ability of neural network to adjust the PID parameters KP、KI、KD,online. The control performance of BP network is superior to traditional PID controller. The curve obtained by simulation comparison analysis shows that PID controller of BP neural network can effectively improve the hydraulic servo control system.