基于RBF神经网络的PID控制在锅炉温度系统中的应用研究

Study on boiler temperature system PID control based on RBF neural network

目的提高电厂锅炉温度系统控制的可靠性和安全性,达到精确控制。方法提出一种基于RBF(Radial Basis Function)神经网络的PID控制器,建立3层神经网络模型。结果在RBF-PID控制过程中,由神经网络RBF在线辨识得到梯度信息,然后根据梯度信息对PID的3个参数进行在线调整,从而改善系统的控制品质。结论仿真结果表明,基于RBF神经网络的PID控制较传统PID控制有较强的鲁棒性,提高了实时性能,获得了更好的控制效果。

Aim To improve the reliability and security of the temperature control system of boiler in power plant and achieve precise control. Methods PID controller based on RBF（Radial Basis Func- tion）neural network is proposed. Three neural network layers are created. Results Gradient informa- tion can be obtained through identification of the RBF neural network on line in the RBF-PID control process and then three PID parameters are set according to the gradient information on line, thus im- proving the quality of the control system. Conclusion Simulation results show that the RBF-PID control can achieve a stronger robustness, a better adaptability and control effect than traditional PID control.