摘要
应用球头铣刀对压铸模具型腔曲面铣削精加工前,通过预测其表面粗糙度,可以辅助工艺人员合理配置铣削工艺参数组合。依托GA算法强大的全局寻优能力,优化ELM固有缺陷,建立遗传算法-极限学习机(GA-ELM)模型,并应用该模型预测压铸模型腔曲面铣削精加工表面粗糙度。训练本模型需要选定输入参数及输出参数,选定输出参数为表面粗糙度,输入参数选定为与之相关的铣削工艺参数,通过可变轴精加工铣削实验获得训练集和测试集数据样本,并将本模型预测结果与其他模型对比,对比模型选择未优化的ELM模型及GA-BP模型,对比结果验证了本模型的优越性。最后,再次通过可变轴精加工铣削实验验证本模型预测结果的可靠性,并基于MatLab环境开发可视化的表面粗糙度预测系统。结果表明:本模型相比ELM模型及GA-BP模型,在铣削精加工表面粗糙度预测方面,预测精度及效率优势明显。
Before milling the die casting mould cavity surface through ball milling cutter,surface roughness prediction can help technical personnel arrange the milling process parameter combination reasonably. In this work,the genetic algorithm—extreme learning machine(GA-ELM)model is used to predict the surface roughness of die casting die cavity surface milling.The input layer weight matrix and hidden layer threshold matrix of ELM are random. GA-ELM model is established by optimizing the input layer weight matrix and hidden layer threshold matrix of ELM through GA algorithm. With the surface roughness as the output parameter and the relevant milling parameter as the input parameter,the corresponding data are obtained through the variable axis milling experiment,and the GA-ELM model is conducted with case analysis. Similarly,the prediction results of GA-BP neural network model optimized by genetic algorithm and original ELM model are compared.Finally,reliability of the prediction result of GA-ELM model is verified through surface roughness measurement experiment,and a visual surface roughness prediction system is developed in MatLab environment. The results demonstrate that GA-ELM model has high accuracy and efficiency in predicting surface roughness of the die casting model cavity. Compared with other algorithms,GA-ELM model has certain advantages.
作者
孙全龙
梅益
杨幸雨
SUN Quan-long;MEI Yi;YANG Xing-yu(College of Mechanical Engineering,Guizhou University,Guizhou Guiyang550025,China)
出处
《机械设计与制造》
北大核心
2020年第8期188-191,196,共5页
Machinery Design & Manufacture
基金
贵州省科技计划项目(黔科合支撑[2017]2010)
贵州大学研究生创新基地(研理工2017056)。
