基于优化BP神经网络的PID控制研究与仿真

The Study and Simulation of PID Control Based on Optimized BP Neural Network

PID控制要取得较好的控制效果,就必须通过调整好比例、积分和微分三种控制作用,形成控制量中既相互配合又相互制约的关系。优化BP神经网络是一种前向神经元网络,具有学习速率快、振荡小、精度高的优点,将其隐含层单元分别作为比例(P)、积分(I)、微分(D)单元,可以建立参数自学习的PID控制器。仿真结果表明基于优化BP神经网络的PID控制器具有较好的自学习和自适应性。

The best effect of PID control based on the best adjustment of the proportion, integral and differential parameter. Optimized BP neural network has the capability of expression nonlinearity and also has the self study and adaptive function,and thus, it can realize the best parameter combination of PID control. The computer simulation shows its advantage in self study and adaptability.