基于模糊神经网络的注塑机料筒温度控制策略  

Temperature Control Strategy of Injection Molding Machine Barrel Based on Fuzzy Neural Network

作　　者：胡亚南[1] 张洪兴 丰会萍 HU Yanan;ZHANG Hongxing;FENG Huiping(Xijing University,Xi'an,Shaanxi 710021 ,China;Shaanxi Xiwei Aotumation Control Engineering Limited,Xianyang,Shaanxi 712000,China)

机构地区：[1]西京学院,陕西西安710021 [2]陕西西微测控工程有限公司,陕西咸阳712000

出　　处：《塑料》2019年第4期91-95,共5页Plastics

基　　金：西京学院科研基金(XJ170129)

摘　　要：针对注塑机料筒温度控制系统的耦合性和非线性,将模糊算法与神经网络算法的优点相结合,设计了模糊神经网络解耦控制器,提高了料筒的控制精确性和稳定性。将料筒其它区间温度的耦合影响作为系统干扰,然后利用模糊控制对温度误差进行调节和温度耦合干扰进行补偿,同时利用神经网络进行学习和优化模糊控制规则,提高了系统控制精度和补偿精度,成功地将料筒温度控制过程转变为无耦合系统。MATLAB仿真结果表明:模糊神经网络控制器能够补偿温度区间耦合和不确定因素所产生的干扰影响,具有很好的解耦能力、抗干扰能力和模型失配鲁棒性,提升了注塑机料筒温度控制的精确性和稳定性,验证了该策略的实用性和有效性。Aiming at the coupling and nonlinearity of the temperature control system of the injection molding machine barrel, combining the advantages of fuzzy algorithm and neural network algorithm, a fuzzy neural network decoupling compensator was designed to improve the accuracy and stability of the control of the temperature of the barrel. The coupling effect of other interval temperatures in the barrel was regarded as system interference. Then, the temperature error was adjusted and the interference of temperature coupling was compensated by the fuzzy control. At the same time, neural network was used to learn and optimize fuzzy control rules to improve system control accuracy and compensation accuracy. The fuzzy neural network could successfully transform the barrel temperature control process into a non-coupling system. MATLAB simulation results showed that fuzzy neural network controller could compensate the disturbance caused by temperature coupling and uncertain factors, and had good decoupling ability, anti-interference ability and model mismatch robustness. The accuracy and stability of barrel temperature control of injection molding machine were improved, and the practicability and effectiveness of the strategy were verified.

关 键 词：注塑机 料筒温度 模糊控制 神经网络控制 MATLAB 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]