摘要
基于MATLAB采用BP神经网络 ,对铝合金腐蚀试验数据进行学习训练 ,建立了腐蚀时间、温度与最大腐蚀深度和疲劳性能的非线性映射关系 .结果表明 ,人工神经网络用于铝合金的腐蚀预测是可行的 .分别采用三层和四层神经网络进行预测 ,讨论了网络结构模型对预测精度的影响 .
A prediction model for corrosion damage of aluminum alloys was developed and the nonlinear relationship between maximum corrosion depth,fatigue performance and corrosion temperature,corrosion time was established based on MATLAB,using BP (Back Propagation) algorithm of neural network.The corrosion trend of aluminum alloys can be predicted by this means.Four-layer network and three-layer network were used,and the effects of structure model of neural network on the precision were discussed.The results show that the former is more precise than the latter.
出处
《中国腐蚀与防护学报》
CAS
CSCD
2004年第4期218-221,共4页
Journal of Chinese Society For Corrosion and Protection
基金
军队"十.五"预研项目资助
关键词
神经网络
疲劳
腐蚀预测
铝合金
neural network,fatigue,corrosion prediction,aluminum alloys
