结合电弧炉主电路系统,推导出电弧炉电气系统状态方程,在此基础上分析了电弧炉三相电流的耦合关系。使用PID神经元网络控制器对电弧炉电极系统进行解耦控制。给出了PID神经元网络控制器的权值初始化规则,其权值更新采用经典的BP算法。分别对于控制器的解耦性、抗干扰性和鲁棒性进行仿真,结果表明PID神经网络控制器具有很好的动态性能,能很好的控制电极调节系统。
According to the main circuit of the EAF, the electric state equation of electric arc furnace and the coupling characteristics among the three electrodes of electrode regulator system are analyzed.In this paper a PID neural network decoupling controller is designed for electrodes regulator system. How to initialize and update weights of neural network controller is also proposed. The simulation result shows that the decoupling performance, anti-disruptive ability and robust characteristics of controller are acceptable.