摘要
基于深孔加工中的BTA钻削方式,通过实验获取了不同转速、进给速度、切削深度下的钻孔粗糙度数据真实值。应用天牛须算法与BP神经网络相互结合,以转速、进给速度、切削深度为输入数据,以孔的表面粗糙度为输出建立了3-6-1的BAS-BP神经网络模型,并绘制折线图将优化前后的预测值进行对比,结果表明BAS-BP神经网络克服了训练时间长、收敛速度慢的缺点,预测精度明显提高。达到了较为理想的效果。也为深孔加工粗糙度研究提供了较好的思路。
Based on the BTA drilling method in deep hole machining, the actual values of drilling roughness data under different rotation speeds, feed rates and cutting depths are obtained through experiments. The application of Tianniu algorithm and BP neural network is combined with the rotational speed, feed rate and depth of cut as input data, and the surface roughness of the hole is used as the output to establish a 3-6-1 BAS-BP neural network model. The line graph compares the predicted values before and after optimization. The results show that the BAS-BP neural network overcomes the shortcomings of long training time and slow convergence speed, and the prediction accuracy is significantly improved. Achieved a better result. It also provides a good idea for the study of deep hole processing roughness.
作者
温静媛
苗鸿宾
刘晓峰
WEN Jingyuan;MIAO Hongbin;LIU Xiaofeng(North University of China,School of Mechanical Engineering,Taiyuan 030051,China;Shanxi Province Deep Hole Machining Center,Taiyuan 030051,China)
出处
《机械设计与研究》
CSCD
北大核心
2019年第5期80-83,共4页
Machine Design And Research
