基于模糊神经网络的电阻焊机恒电流控制研究

由于建立实际电阻点焊过程精确的数学模型比较困难,使得常规的人工调节PID控制器参数较难实现良好的匹配,从而难以获得满意的控制效果。针对该问题,将智能调节与PID控制方法相结合,利用模糊神经网络设计了参数kp、ki、kd自适应调整的PID控制器,构建了逆变电阻点焊电源的系统模型,通过对系统运行状态的在线学习智能化地修正PID的三个参数。PID控制输出量通过PWM发生器产生四路独立的、占空比实时变化的PWM波形,进而控制逆变器的功率开关器件导通时间,最终实现对系统恒电流的输出控制。仿真结果表明,该方法能根据系统的运行状态自行匹配对应最优控制规律下的PID三个参数,能有效地控制焊接电流的恒定,达到满意的效果。

Research on constant current of resistance spot welding based on fuzzy neural network

In allusion to the problem that it is so difficult to establish the real-time mathematic model of resistance spot-welding process accurately that the conventional PID control can’t get better performance,the method of combining intelligent regulation and PID control was put forward.This article built a model of inverter resistance spot welding power.It intelligently corrected the three parameters of PID through the online learning of the running system.The PID controller could control the on-off time of the inverter’s switches by the four independent PWM waveforms.The duty ratio of the PWM waveforms were changed real time with feedback,and eventually realized the constant current output control of the system.The simulation results show that; the method could following the optimal control law and automatically match the three parameters of the PID controller according to the system’s operation state.Then it could effectively keep the current of the welding to be constant,to achieve satisfactory results.