基于正交试验与BP神经网络的椰壳纤维复合材料注塑工艺优化
Optimization of Injection Molding Process of Coconut Shell Fiber Composites Based on Orthogonal Test and BP Neural Network
针对椰壳纤维复合材料在注塑成型中的翘曲变形问题,提出一种正交试验与反向传播(BP)神经网络相结合的注塑工艺优化方法。以副驾驶储物门为例,以5%椰壳纤维/10%氢氧化镁(MH)/聚丙烯(PP)复合材料为实验基材,采用Moldflow仿真初始参数翘曲变形量为1.609 0 mm。设计模具温度(A)、注射时间(B)、保压压力(C)和保压时间(D)四因素四水平正交试验,分析工艺参数对翘曲变形的影响规律,得到最佳工艺参数组合为A3B4C4D3,此时翘曲变形量为0.675 4 mm,较初始参数翘曲变形量降低58.02%。再基于正交数据建立BP神经网络模型,预测最优工艺参数组合,仿真验证翘曲变形量为0.479 0 mm,较初始翘曲变形量降低70.23%,满足工业生产基本要求,相对误差仅5.65%。纤维取向分析表明,椰壳纤维在制品边缘均匀分布,能够有效提升局部力学性能。
A process optimization method combining orthogonal test and back propagation (BP) neural network was proposed to address the warpage deformation of coir fiber-reinforced composites during injection molding. Taking the glove box door as an example, 5% coir fiber/10% magnesium hydroxide (MH)/polypropylene (PP) composite was employed as the experimental substrate. The initial warpage deformation obtained from Moldflow simulation was 1.609 0 mm. A four-factor four-level orthogonal experiment was designed with mold temperature (A), injection time (B), packing pressure (C), and packing time (D) as variables to analyze the influence of process parameters on warpage deformation. The optimal parameter combination was determined as A3B4C4D3, yielding a warpage deformation of 0.675 4 mm, which represented a 58.02% reduction compared to the initial value. A BP neural network model was then established based on the orthogonal data to predict the optimal process parameters. Simulation verification indicated a warpage deformation of 0.479 0 mm, corresponding to a 70.23% reduction from the initial warpage deformation, which met the basic requirements for industrial production with a relative error of only 5.65%. Fiber orientation analysis revealed that coir fibers were uniformly distributed at the edges of the molded part, effectively enhancing local mechanical properties.
| [1] |
李相旭. 纤维素纤维增强天然橡胶复合材料制备机理及实验研究[D]. 青岛: 青岛科技大学, 2025. |
| [2] |
王峰, 张家辉, 孙文革, |
| [3] |
张爱龙. 基于FDM成形连续碳纤维增强复合材料工艺与性能研究及卷尺结构制备[D]. 福州: 福建理工大学, 2024. |
| [4] |
高鹏堂. 椰壳纤维/聚丙烯复合材料应用于车载气囊盖板的可靠性研究[J]. 装备维修技术, 2022(2): 62-68. |
| [5] |
戴彬, 钱应平, 薛杭, |
| [6] |
武金炎. 仿生纤维增强复合材料梁板结构静动态力学性能研究[D]. 呼和浩特: 内蒙古师范大学, 2025. |
| [7] |
黄关山, 朱金婷. 基于响应面代理模型及遗传算法的汽车引擎盖注塑成型工艺优化[J]. 塑料科技, 2024, 52(12): 125-128. |
| [8] |
娄艳华, 周建华. 基于正交试验的高性能排水管注塑成型工艺优化[J]. 塑料科技, 2024, 52(11): 136-139. |
| [9] |
黄关山, 王新艳. 基于DOE设计及MOPSO算法的汽车滤清器外壳多目标优化分析[J]. 塑料科技, 2024, 52(8): 105-108. |
| [10] |
刘海波, 张睿. 基于计算机数值模拟技术的汽车内饰面板注塑成型工艺优化[J]. 塑料科技, 2023, 51(11): 89-93. |
| [11] |
杨明, 刘赛, 刘巨保, |
| [12] |
林煌旭, 孔选, 陆将男, |
| [13] |
杨明, 侯健超, 刘巨保, |
| [14] |
高洪晨, 薛松, 肖梦曦, |
| [15] |
孙铭泽. 椰壳纤维增强PP材料汽车安全气囊盖板工艺优化与研究[D]. 海口: 海南大学, 2023. |
| [16] |
胡家乐. 金属嵌件式汽车前端模块轻量化设计及注塑成型工艺优化[D]. 芜湖: 安徽工程大学, 2025. |
| [17] |
董坤. 基于CAD/CAE的手机外壳注塑模设计及成形工艺优化[D]. 芜湖: 安徽工程大学, 2016. |
| [18] |
费晓瑜, 刘新新, 闫长飞, |
| [19] |
|
| [20] |
孙文革, 程莉, 雷皓, |
| [21] |
朱梦萱. 基于多特征和BP神经网络的拷贝数变异检测方法研究[D]. 聊城: 聊城大学, 2025. |
| [22] |
皮卫, 熊建武. 基于RBF神经网络断路器注塑成型工艺优化[J]. 塑料科技, 2023, 51(6): 70-74. |
| [23] |
姜旭, 解锴, 王晓宏, |
| [24] |
管涛, 李元庆, 郭方亮, |
海南省自然科学基金项目(521RC497)
广东中贝能源科技有限公司项目(HD-KYH-2024215)
/
| 〈 |
|
〉 |