摘要
分析了注塑机的机械结构以及工作原理,利用基于神经网络的自学习、自适应能力,将传统比例积分微分(PID)控制与神经网络结合,采用梯度下降法实时调整PID的三个参数,从而实现了注塑机温度的自适应控制,提高了注塑机料筒多段温度控制精度以及稳定性,进而提升塑料制品质量。通过仿真可知,与传统PID控制相比,该控制策略的超调量更小,抗干扰能力更强,温度控制精度更高,系统能以较快速度实现对各料筒温度的控制。
This paper describes the mechanical structure and working principle of injection molding machine.The self-learning and self-adaptive ability based on neural network were combined with the traditional proportional integral differential(PID)to adjust three parameters of PID in real time by gradient descent method,which realizes the self-adaptive control of the temperature in injection molding machine and improves the precision and stability of multistage temperature control in the barrel,thus the quality of the products are improved.The simulation results show that this novel control strategy has less overshoot,stronger antiinterference ability and higher precision of temperature control than traditional PID control.The system can realize the temperature control for each barrel at a faster speed.
作者
党霞
Dang Xia(Henan Industry and Trade Vocational College,Zhengzhou 450000,China)
出处
《合成树脂及塑料》
CAS
北大核心
2018年第6期83-86,共4页
China Synthetic Resin and Plastics
关键词
注塑机
温度控制
神经网络
比例积分微分
仿真
injection molding machine
temperature control
neural network
proportional integral differential
simulation
作者简介
党霞,女,1980年生,硕士,讲师,主要研究方向为机电一体化。E-mail:19794518@qq.com。
