基于遗传算法的改进BP神经网络在中空成型机型坯壁厚控制系统中的应用  

Application of Improved BP Neural Network Based on Genetic Algorithm in Blank Wall-Thickness Control System of Blow Molding Machine

作　　者：程凤芹 王正通 曲娜 CHENG Feng-qin;WANG Zheng-tong;QU Na(School of Electrical and Information Engineering,Jilin University of Architecture and Technology,Changchun 130114,China;School of Electrical and Electronic Engineering,Changchun University of Technology,Changchun 130012,China)

机构地区：[1]吉林建筑科技学院电气信息工程学院,吉林长春130114 [2]长春工业大学电气与电子工程学院,吉林长春130012

出　　处：《塑料科技》2020年第7期114-117,共4页Plastics Science and Technology

摘　　要：中空吹塑成型技术受到型坯自重的影响造成容器内壁厚度不均匀,影响了塑料产品的质量和生产成本。针对此问题,将遗传算法与BP算法相结合,设计了能够收敛到全局最优解的改进BP神经网络。分别对壁厚控制系统的各元件建立传递函数模型,进行Simulink仿真实验,结果表明,该方法不仅能够提高系统的控制精度,并且提高了系统收敛速度。Hollow blow molding technology is affected by the weight of the blank,resulting in uneven thickness of the inner wall of the container,which affects the quality and production cost of plastic products.In order to solve this problem,the paper combined the genetic algorithm and BP algorithm to design an improved BP neural network which can converge to the global optimal solution.The paper established transfer function model for each component of the wall thickness control system separately,and carried out Simulink simulation experiment.The results showed that this method could not only improve the control accuracy of the system,but also improved the convergence speed of the system.

关 键 词：吹塑成型 壁厚控制 神经网络 遗传算法 

分 类 号：TQ320.52[化学工程—合成树脂塑料工业]