基于Moldex 3D与GA-BP的MIM微小齿轮工艺参数优化
Optimization of MIM process for micro gear based on Moldex 3D and GA-BP
齐广浩 1王训伟 2雷保珍 2王青云 3罗子洪 4邱耀弘5
作者信息
- 1. 北京联合大学,北京市信息服务工程重点实验室,北京 100020
- 2. 北京联合大学,北京市信息服务工程重点实验室,北京 100020;北京联合大学,北京市智能机械创新设计服务工程技术研究中心,北京 100020
- 3. 北京京城增材科技有限公司,北京 101116
- 4. 型创科技顾问股份有限公司,广东东莞,523000
- 5. 昆山耀德企业咨询有限公司,江苏昆山,215300
- 折叠
摘要
为了降低MIM微小齿轮注射坯翘曲总位移,寻找较优的工艺参数组合,提出了基于Moldex 3D与GA-BP遗传神经网络的金属注射成形微小齿轮工艺参数优化方法.首先,以注射成形中喂料温度、模具温度和保压时间等因素建立全因子实验,对各实验点的翘曲总位移以Moldex 3D软件加以仿真,进而分析得出较优工艺参数组合.随后设计了 3-4-1型BP神经网络,以最小翘曲总位移作为适应度值,结合遗传算法,得到优化的工艺参数组合.相对于全因子实验的分析结果,GA-BP模型的寻优结果与其相近,准确度有98.09％.本文的方法为MIM工艺参数优化提供了新的思路,有助于提升微小齿轮制品的成形质量.
Abstract
In order to reduce the total warpage of MIM micro-gear injection parts and find a better combination of process parameters,a method for optimizing the process parameters of metal injection molding micro-gears based on Moldex 3D and GA-BP genetic neural network is proposed.Firstly,a full factorial experiment is estab-lished considering factors such as feed temperature,mold temperature and holding time in injection molding.The total warpage of each trial is simulated by Moldex 3D software,and then the better process parameter combina-tion is determined.Subsequently,a 3-4-1 BP neural network is designed,and the minimum total warpage is used as the fitness value.Combined with the genetic algorithm,the optimized process parameter combination is obtained.Compared with the analysis results of the full factorial experiment,the optimization results of the GA-BP model are similar to it,with an accuracy of 98.09％.This method provides a new approach for optimizing MIM process parameters,contributing to the improvement of the forming quality of micro gear products.
关键词
金属注射成形/Moldex/3D/BP神经网络/遗传算法/工艺参数优化Key words
metal injection molding/Moldex 3D/BP neural network/genetic algorithm/process parameter opti-mization引用本文复制引用
出版年
2024
NETL
NSTL
万方数据