GMC2000A加工中心热误差建模方法研究  

Research on thermal error modeling method of GMC2000A machining center

作　　者：李有堂[1] 汤雷武 黄华[1] 吴荣荣 Li Youtang;Tang Leiwu;Huang Hua;Wu Rongrong(School of Mechanical and Electrical Engineering,Lanzhou University of Technology,Gansu Lanzhou,730050,China)

机构地区：[1]兰州理工大学机电工程学院,甘肃兰州730050

出　　处：《机械设计与制造工程》2023年第7期61-67,共7页Machine Design and Manufacturing Engineering

摘　　要：热误差是影响机床加工精度的主要因素之一,为减小热误差对机床精度的影响,提出萤火虫算法结合BP神经网络建立热误差模型。使用萤火虫算法对BP神经网络进行优化,对隐含层神经元个数进行优化取值,确定网络结构,并对网络初始权值和阈值进行了优化。以GMC2000A机床为试验对象,误差模型的输入为模糊C-均值聚类选取的机床关键位置的温度向量,输出为Y轴定位误差,通过均方根误差值RMSE、决定系数R~2和预测精度η三项指标对误差模型预测效果进行评估。结果表明,萤火虫算法优化BP神经网络误差模型取得了较好的预测结果,且在恶劣的工作环境中仍能保持一定的预测精度。Thermal error is one of the main factors affecting the machining accuracy of machine tools.In order to reduce the impact of thermal error on machine tool accuracy,the Firefly algorithm combined with BP neural network is proposed to establish a thermal error model.In this model,BP neural network is optimized firefly algorithm,the number of hidden layer neurons is optimized,the network structure is determined,and the initial weight and threshold of the network are optimized.Taking GMC2000A machine tool as the test object,the error model input is the temperature vector of the key position of the machine tool selected by fuzzy C-mean clustering,and the output is the Y-axis positioning error.Finally,the root mean square error value RMSE,the determination coefficient value R 2 and the prediction accuracy valueηevaluation predicts the effect of error model with three indicators.The results show that the BP neural network is optimized Firefly algorithm error model to achieve a significant prediction effect,and it can still maintain a certain prediction accuracy in a harsh working environment.

关 键 词：萤火虫算法 BP神经网络 热误差 误差建模 误差预测 

分 类 号：TG659[金属学及工艺—金属切削加工及机床]