摘要
BP神经网络用于预测储氢合金的电化学性能。通过稀土族中金属Nd部分替代AB5型储氢合金中的La,研究了Nd的变化对合金的初始放电容量的影响。详细介绍了BP网络的结构确定、参数选择及训练方法,并利用该网络得出了较好的预测结果。
Back propagation network method was applied to property prediction of hydrogen storage alloy. The influence of rare earth Nd taking place of some La on the initial discharge capacity among AB_5-based hydrogen storage alloy was studies. The building of network structure, choose of network parameters and methods of training were introduced in details, and good predication results were obtained by use of the network.
出处
《功能材料》
EI
CAS
CSCD
北大核心
2004年第2期174-176,182,共4页
Journal of Functional Materials
基金
国家科技部基础研究重大项目前期研究专项资助项目(2001CCA05000)
国家重点基础研究发展规划资助项目(2002CB211800)
