基于神经网络的激光加工PMMA微通道工艺研究
Research on Laser Processing of PMMA Microchannel Based on Neural Network
张瑶 1张秀丽 1郑宏宇 1王铭洋 1魏娟1
作者信息
- 1. 山东理工大学机械工程学院,山东淄博 255000;山东省精密制造与特种加工重点实验室,山东淄博 255000
- 折叠
摘要
为实现聚甲基丙烯酸甲酯(PMMA)微通道智能化激光加工,采用CO2激光器研究激光功率密度、扫描速度和扫描次数对PMMA微通道深度和宽度的影响.基于BP(back propagation)神经网络建立激光加工工艺参数的预测模型,采用实验数据进行网络训练,并通过粒子群优化算法进行寻优.结果表明,所建立的神经网络优化模型可以使PMMA微通道宽度加工误差控制在5％以内,深度加工误差控制在12％以内,具有良好的预测精度,将为智能化PMMA微通道激光加工参数选取提供依据.
Abstract
This study investigates the intelligent laser processing of polymethylmethacrylate(PMMA)microchannels,exami-ning the influence of laser power density,scanning speed,and scanning passes on channel dimensions utilizing a CO2 laser.A back propagation(BP)neural network model was developed to predict the laser processing parameters.The model was trained with empirical data and further optimized using a particle swarm optimization algorithm.The results showed that the neural network optimization model can control the PMMA microchannel width machining error within 5％and the depth machining er-ror within 12％.The model has good prediction accuracy and will provide a basis for the intelligent selection of laser processing parameters of PMMA microchannel.
关键词
激光加工/微通道/BP神经网络/粒子群算法Key words
laser processing/microchannel/back propagation neural network/particle swarm optimization引用本文复制引用
基金项目
国家重点研发计划政府间重点专项(2022YFE0199100)
国家自然科学基金(51905317)
山东省自然科学基金(ZR2020ME047)
出版年
2024
NETL
NSTL
万方数据