基于人工神经网络的FGM数据库设计

Design and Study on FGM Database Based on Artificial Neural Network

采用数据库管理技术和人工神经网络技术建立了FGM数据库,开发了管理和操纵该数据库的应用程序.利用该数据库可以查询编辑已知FGM材料组分、分布和性能之间的关系,预测未知FGM在给定材料组分和分布条件下的材料性能.作为应用实例,预测了SiC/C FGM的弯曲强度,并计算了预测误差,分析结果表明预测值与实验值较为吻合.

A FGM database and its management and controlling application programs are developed using the technologies of database management system and artificial neural network.With the help of this database,the material relationship among constituents,distributions and properties for known FGMs can be inquired and edited.Material properties of unknown FGMs with the given material constituents and distributions can be predicted.As an application example,the bending strengths of SiC/C FGMs are predicted,and the prediction errors are also calculated.Analysis result shows that the prediction values are consistent with the experiment values.