ZA22/Al_2O_(3f) 复合材料力学性能的神经网络模拟研究  

Neural Network Simulation on Mechanical Properties of ZA22/Al 2O 3f Composites

作　　者：陈凯[1] 俞蒙槐[1] 胡上序[1] 于思荣[2] 何镇明[2] 

机构地区：[1]浙江大学 [2]吉林工业大学

出　　处：《材料工程》1997年第10期14-16,共3页Journal of Materials Engineering

基　　金：国家自然科学基金

摘　　要：采用人工神经网络的典型模型即Ｂ－Ｐ算法对ＺＡ２２／Ａｌ２Ｏ３ｆ复合材料的力学性能进行分析和预测，结果表明：Ｂ－Ｐ模型不仅适用于单变量输出非线性系统的拟合与预测，同时还适用于多变量输出的非线性系统；Ｂ－Ｐ模型对研究复合材料的力学性能具有良好的容错性。The typical model of neural network e.g. B P algorithm was adopted to analyze as well as predict the mechanical properties of ZA22/Al 2O 3f composites . The results show that B P model is not only suitable for the regression and prediction for the single variable output non liner system, but also suitable for the multi variables output non liner system. Perfect fault tolerance ability was also found in the B P model while it was used to analyze the mechanical properties of ZA22/Al 2O 3f composites.

关 键 词：锌铝合金 复合材料 力学性能 神经网络 

分 类 号：TB331[一般工业技术—材料科学与工程]