摘要
以钢材中C和Cr的成分、变形程度和单位挤压力的关系为研究对象,在Matlab语言环境下,以正挤压试验数据作为训练和预测样本,用5节点的单隐含层BP型神经网络对钢材单位挤压力进行了预测。结果表明,采用正交试验和人工神经网络相结合的方法来预测单位挤压力是有效和可行的。
The paper researches on the relationship between C and Cr in steel product,deformation extent and unit extrusion pressure,taking experimental data of forward extrusion as the samples for training and predicting,predicts unit extrusion pressure using single hidden layer BP model with five nodes under Matlab language environment.The results indicate that combination method of orthogonal test and artificial neutral network is a valid and feasible way for predicting the unit extrusion pressure.
出处
《汽车工艺与材料》
2004年第7期13-15,共3页
Automobile Technology & Material
关键词
神经网络
预测
单位挤压力
artificial neutral network
predict
unit extrusion pressure
