摘要
以医用介入导管接头为研究对象,基于塑料成型理论在Moldflow软件中进行导管接头模流分析,通过正交实验极差分析,确定了注塑工艺参数对导管接头缩痕指数的影响趋势,得到最佳工艺参数组合。针对实际生产中出现的缩痕缺陷,建立导管接头缩痕指数的BP神经网络参数模型,并用遗传算法进行优化,同时对结果进行仿真模拟,得到缩痕指数为0.0752%,此时的最佳注塑工艺参数为熔体温度238℃、模具温度71℃、注塑压力68 MPa、注塑时间0.61 s、保压压力27 MPa、保压时间24 s,其结果比极差分析法的优化结果(0.088%)减少了14.5%。将遗传算法优化BP神经网络后的注塑工艺参数组合应用于导管接头加工试生产,得到产品外观无明显熔接痕,表面质量良好,满足企业设计要求。
Taking medical interventional catheter joint as the research object,the Moldflow analysis of the catheter joint was carried out in Moldflow software based on plastic molding theory.Through range analysis of orthogonal experiment,the influence trend of injection molding process parameters on the sink mark index of the catheter joint was determined,and the best process parameters combination were obtained.Aiming at the shrinkage mark defects in actual production,the BP neural network parameter model for the shrinkage mark index of the catheter joint was established,and the genetic algorithm was used to optimize and the result was simulated,it is found the sink mark index is 0.0752%,the corresponding best injection molding process parameters are as follows:The melt temperature is 238℃,the mould temperature is 71℃,the injection pressure is 68 MPa,the injection time is 0.61 s,the holding pressure is 27 MPa,and the holding time is 24 s.The result is 14.5%less than that(0.088%)optimized by the range analysis method.The combination of injection molding process parameters based on BP neural network optimized by genetic algorithm was applied to the trial production of the catheter joint processing,the results show that the appearance of the product has no obvious weld marks,and the surface quality is good,which meets the design requirements of the enterprise.
作者
廖生温
王玉勤
王可胜
刘婧
Liao Shengwen;Wang Yuqin;Wang Kesheng;Liu Jing(School of Mechanical Engineering,Chaohu University,Chaohu 238000,China)
出处
《工程塑料应用》
CAS
CSCD
北大核心
2021年第3期65-70,共6页
Engineering Plastics Application
基金
安徽省高校自然科学研究重大项目(KJ2019ZD47)
巢湖学院自然科学研究项目(XLZ-202003)。
关键词
导管接头
注塑
BP神经网络
缩痕指数
catheter joint
injection molding
BP neural network
sink mark index