摘要
根据人工神经网络 (ANN)的BP(backpropagation)算法 ,建立了快速凝固Cu Cr Zr铜合金时效温度和时间与硬度和导电率的神经网络映射模型。预测值与实际情况吻合良好 ,硬度和导电率最大误差分别为 4 1%和 1 9%。通过对样本集的学习 ,建立了快速凝固时效工艺知识库 。
A model is established to predict the aging properties of rapdily solidified Cu-Cr-Zr alloy based on the back propagation(BP) algorithm of the artificial neural network(ANN). The non-linear relationship between parameters of aging processes and mechancal and electrical properties of rapidly solidified Cu-Cr-Zr alloy is available. The predicted values of the ANN are in accordance with the tested data. The maxmium errors of hardness and conductivity are 4.1% and 1.9%. A knowledge base on the aging processes of rapidly solidified Cu-Cr-Zr alloy is established via sufficient data mining by the network. With the help of the knowledge base, the properties of rapidly solidified Cu-Cr-Zr alloy and its aging technology can be effectively predicted and controlled.
出处
《材料热处理学报》
EI
CAS
CSCD
北大核心
2003年第3期88-90,共3页
Transactions of Materials and Heat Treatment
基金
国家高技术研究发展计划 (863计划 ) (2 0 0 2AA331 1 1 2 )
河南省重大科技攻关项目 (0 1 2 2 0 2 1 30 0 )
