摘要
在实验数据的基础上,利用人工神经网络建立了非调质钢的抗拉强度、屈服强度、断面收缩率和断后伸长率等力学性能与合金成分对应关系的模型。将合金成分作为网络的输入,非调质钢的力学性能作为网络的输出,来训练网络预测非调质钢的力学性能,与实测值比较获得了满意的结果,为高性能材料设计提供了一个辅助手段。
Based on experiment data, artificial neural network was used to build a model for the relation between the mechanical properties (tensile strength, yield strength, percentage reduction in area and rate of elongation after break) and alloy elements in non-quenched and tempered steel. Taking the partial alloy elements as the input of network and the mechanical properties as output of network, according to the data of the manual data, the network to predict the mechanical properties of non-quenched and tempered steel was trained. The satisfactory results are obtained compared with the measured date of the mechanical properties, which provides a theoretical aided tool for the design of high performance material.
出处
《热加工工艺》
CSCD
北大核心
2009年第4期108-110,115,共4页
Hot Working Technology
关键词
材料设计
人工神经网络
非调质钢
力学性能
material design
artificial neural network
non-quenched and tempered steel
mechanical property