基于RBF神经网络的电弧炉电极调节系统PID参数整定

PID Parameter Setting on Electrode Adjustment System of Electric Arc Furnace Based on RBF Neural Network

由于常规PID控制难以满足电弧炉电极调节系统复杂的工况,本文将径向基函数(RBF)神经网络与PID控制相结合,提出RBF-PID参数整定方法。通过RBF神经网络对控制对象Jacobian信息的辨识,采用增量式PID梯度下降算法整定已有的PID参数,设计了RBF-PID电极调节系统控制器。仿真结果验证了RBF-PID控制器能够实时整定PID参数,实现电极调节系统快速、准确地控制。

Because conventional PID control is difficult to meet the complex conditions for electrode adjustment system of electric arc furnace, the radial basis function （RBF） neural network is introduced into the PID control, a setting method of RBF-PID is proposed in this paper. By RBF neural network identifies Jacobian information of control object, selecting incremental PID gradient descent algorithm sets existing PID parameters, a RBF-PID controller of electrode regulating system is designed. The simulation results verify that RBF-PID controller can real-time set the PID parameters, realize quickly and accurately control to electrode regulating system.