摘要
针对现行挤压变形力计算的局限性 ,采用四层BP网络建立 15钢正挤压变形力预测的神经网络模型 ,其预测值与实验值吻合良好 ,为挤压力预测提供了新的途径。
Presents a new method for predicting the extrusion force,which is developed on the basis of the four layers of back propagation neural network.BP neural net is trained with gradient descent,so as to gain the non-linear relationship between forward extruded force of 15 steel and die angle,hollow diameter,the ratio of area reduction.The results show that the predicted values are well identical with the experimental ones.
出处
《现代制造工程》
CSCD
北大核心
2004年第2期56-58,共3页
Modern Manufacturing Engineering
关键词
神经网络
挤压
变形力预测
Artificial neural net Forward extrusion Deforming force prediction
