基于BP神经网络的注塑变频控制系统设计

Control Method of Injection Molding Machine Based on Neural Network

针对注塑机存在严重能耗,提出了基于BP神经网络的注塑变频控制系统设计方案。该方案采用ARM Cortex-M3内核微控制器采集压力、流量信号,利用BP算法计算出对应的控制量,改变变频器的输出频率,进而控制油泵电机,实现注塑节能。实验结果表明,该系统节约注塑机能耗可达30%—60%。

In view of the serious energy consumption of injection molding machine, variable frequency control method based on the back propagation （BP） of neural network is provided. ARM Cortex-M3 Kernel is adopted, which can exactly calculate frequency of inverter on the basis of pressure and flow from injection molding machine controller. Test results show that the system can effective reduce energy consumption of injection molding machine to 30% --60%.