基于RBF型人工神经网络的碳/陶瓷复合材料的化学成分对硬度的耦合影响分析  

Coupling Effects of Chemical Compositions on the Hardness of Carbon/Ceramic Composite Material by RBF Artificial Neural Network

作　　者：刘雅芳[1] 董万鹏[1] 由伟[2] 饶轮 

机构地区：[1]上海工程技术大学材料工程学院,上海201600 [2]华北科技学院机电工程系,廊坊065201

出　　处：《材料导报》2015年第12期153-157,共5页Materials Reports

基　　金：上海工程技术大学研究生创新项目(E1-0903-14-01131)

摘　　要：用RBF型人工神经网络研究了碳/陶瓷复合材料的化学成分对其硬度的影响。首先设计了RBF型神经网络模型,用"舍一法"进行了训练,使模型具有满意的预测性能。随后分析了化学组分对硬度的影响,包括单因素影响和双因素耦合影响。结果表明:材料的两种组分同时变化时,对硬度的影响更加复杂,呈现典型的非线性特征。RBF artificial neural network was developed to study the effects of chemical compositions on the hardness of carbon/ceramic composite material.The RBF neural network model was designed and trained by the“leave-one-out method”.After being trained,the model had satisfactory prediction performance.Then,the ANN model was used to analyze the effects of chemical compositions on the hardness of carbon/ceramic,including single ones of single factor and coupling ones of dual factors.The results showed that the coupling effects of dual factors on the hardness are more complicated,presenting typical non-linear characteristics.

关 键 词：碳/陶瓷复合材料 化学成分 硬度 RBF 型人工神经网络 耦合影响 

分 类 号：TQ174.75[化学工程—陶瓷工业] TP183[化学工程—硅酸盐工业]