基于神经网络的水轮机调节系统的自适应 PID 控制

The Adaptive PID Control of the Hydraulic Turbine Regulation System on the Neural Networks

基于有功功率反馈参与控制的水轮机调节系统，以几个典型工况下的最优ＰＩＤ系数作为训练样本，训练了一个三层ＢＰ神经网络，设计了一个用ＢＰ神经网络实现变参数的ＰＩＤ控制器；并构造了一个目标函数，设计了一个自适应神经元，利用神经元的自学习功能，在线优化控制器的输出，以期达到最优控制的目的。对简单电力系统的仿真结果表明，这种控制器可以达到较常规的变参数ＰＩＤ较好的控制效果，是实现水轮机调节系统自适应控制的一种可行的方法。

For the hydraulic trubine regulation system with real power feedback,by using the optimal PID parameters on the typical operating mode as the training samples,the paper has trained a three layers BP neural network,designed a variable parameters controller on the BP neural network, and constructed a object funciton,designed an adaptive neuron.For realizing the optimal control by using its self-learning function,the neuron has optimized the controller's output online.The computer simulation results for the simple electric power system have shown that the control effect of the controller is better than the traditional variable parameters PID controller.It is feasible to realige the adaptive control of the hydraulic turbine regulation system.