摘要
针对滚挤压工艺参数难于选取、已加工表面质量难以控制的问题,采用增加动量项和学习率自适应调整的BP 神经网络建立了滚挤压加工表面质量的预测模型,并以表面粗糙度的预测作为实例进行研究,试验结果表明该模型可用于 滚挤压加工表面质量的预测。
An artificial neural network (ANN) model for predicting the surface integrity of burnishing process was established,which was trained by the momentum method and the self-adaptive adjusting strategy of learning rate.As an example,the prediction of surface roughness was presented.The result shows that the artificial neural network method and the established model are efficient for the surface integrity prediction of burnishing process.
出处
《机床与液压》
北大核心
2005年第3期17-19,共3页
Machine Tool & Hydraulics
基金
国家自然科学基金重点项目(50135020)
关键词
神经网络
预测
滚挤压加工
表面粗糙度
Artificial neural network
Prediction
Burnishing process
Surface roughness
