基于遗传算法优化BP神经网络的控制面板注塑成型工艺优化
Optimization of Control Panel Injection Molding Process Based on Genetic Algorithm Optimized BP Neural Network
以某公司生产的家电控制面板为例,对控制面板注塑成型后的翘曲变形量进行研究。设计正交试验,运用Moldflow进行模流分析,构建一个3层网络结构4-5-1的反向传播(BP)神经网络模型,对模型进行训练与预测,得到工艺参数和翘曲变形量之间的非线性关系。通过遗传算法优化BP神经网络模型及全局寻优,得到最优翘曲变形量预测值(0.476 3 mm)及对应的工艺参数组合(模具温度58 ℃、保压时间13 s、料筒温度230 ℃、保压压力60 MPa)。对遗传(GA)算法优化工艺参数组合进行Moldflow模流分析,翘曲变形量为0.480 0 mm,与神经网络预测值0.476 3 mm十分接近,误差在合理范围内。GA算法优化工艺参数得到的翘曲变形量与初始工艺参数得到的翘曲变形量相比,降低30%。结果表明,神经网络GA算法用于注塑工艺参数优化设计能够有效提升制品质量,减少试模成本。
A household appliance control panel produced by a certain company was taken as an example to investigate the warpage deformation of the control panel after injection molding. Orthogonal test were designed, and Moldflow was employed for mold flow analysis. A back-propagation (BP) neural network model with a three-layer network structure of 4-5-1 was constructed, and the model was trained and predicted to establish the nonlinear relationship between process parameters and warpage deformation. The BP neural network model was optimized through a genetic algorithm with global optimization, yielding an optimal warpage prediction value of 0.476 3 mm and the corresponding process parameter combination (mold temperature of 58 ℃, holding time of 13 s, barrel temperature of 230 ℃, and holding pressure of 60 MPa). Moldflow mold flow analysis was performed on the process parameter combination optimized by genetic algorithm (GA) algorithm, resulting in a warpage deformation of 0.480 0 mm, which was very close to the neural network prediction value of 0.476 3 mm, with the error within a reasonable range. Compared with the warpage deformation obtained from the initial process parameters, the warpage deformation achieved by the GA-optimized process parameters was reduced by 30%. The results indicated that the neural network GA algorithm applied to injection molding process parameter optimization could effectively improve product quality and reduce trial molding costs.
| [1] |
张苏新. ABS电子继电器注塑成型优化分析[J]. 塑料科技, 2024, 52(8): 109-112. |
| [2] |
潘莎, 李紫蔓. 基于CAE技术的铜排壳体注塑工艺优化设计[J]. 塑料科技, 2025, 53(9): 174-178. |
| [3] |
刘春晓. ABS/Al免喷涂材料注塑成型黑线流痕的研究[D]. 郑州: 郑州大学, 2023. |
| [4] |
邱琳芝. 模内电子注塑成型缺陷分析及工艺参数优化研究[D]. 东莞: 东莞理工学院, 2025. |
| [5] |
曾文浩, 魏刚, 黄亚, |
| [6] |
陈妍. 注塑成型工艺参数的优化研究进展[J]. 合成树脂及塑料, 2020, 37(3): 87-90. |
| [7] |
郑方莉. 注塑成型工艺参数智能设置及优化[D]. 北京: 北京化工大学, 2022. |
| [8] |
方明月, 张宇, 王彻, |
| [9] |
黄海跃, 范希营, 李赛, |
| [10] |
郝彦琴, 龙春光. 基于BP神经网络和遗传算法的机箱壳注塑工艺参数多目标优化[J]. 工程塑料应用, 2018, 46(3): 69-75. |
| [11] |
郭旭. 注塑模具随形冷却水路优化及增材制造工艺研究[D]. 沈阳: 沈阳工业大学, 2024. |
| [12] |
吴俊超,陈爱霞,夏志平.光伏移动电源电池仓上盖热流道倒装注塑模设计[J/OL]. 塑料科技,1-11[2025-11-12]. |
| [13] |
张博, 黄志高, 金瑞民, |
| [14] |
吴其焦, 任思猛, 吴朋余. 基于MoldFlow分析的断路器外壳注射模冷却系统研究与优化[J]. 模具制造, 2026, 26(1): 27-30. |
| [15] |
李志恒. 汽车散热器水箱上盖注塑成型及翘曲优化[D]. 青岛: 山东科技大学, 2018. |
| [16] |
童希. TWIP钢热处理工艺参数优化及微观组织研究[D]. 长沙: 中南林业科技大学, 2019. |
| [17] |
何广川. 汽车薄壁塑件快速热循环注塑成型模拟及翘曲与缩痕控制工艺优化[D]. 重庆: 重庆交通大学, 2021. |
| [18] |
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. |
| [19] |
程效锋. 汽车雾灯盖板饰条注塑成型分析及工艺优化[D]. 马鞍山: 安徽工业大学, 2022. |
| [20] |
张永超. 基于遗传算法的车窗玻璃导槽的注塑成型设计及参数优化[D]. 天津: 天津职业技术师范大学, 2021. |
| [21] |
张海鹏. 轿车头枕罩盖注塑研究与仿真分析[D]. 大连: 大连交通大学, 2022. |
| [22] |
王晓衬. 基于多目标遗传算法电子适配器成型工艺参数优化[J]. 塑料科技, 2023, 51(5): 90-94. |
| [23] |
王晓东. 注塑成型过程能耗分析及多目标优化研究[D]. 天津: 天津职业技术师范大学, 2024. |
湖南省自然科学基金项目(2021JJ50079)
/
| 〈 |
|
〉 |