基于BP神经网络的Mg-Al-Si合金力学性能研究  

Research on Mechanical Properties of Mg-Al-Si Magnesium Alloy Based on BP Neural Network

作　　者：吴新华[1] 伊厚会[1] 

机构地区：[1]滨州学院理论物理研究所,山东滨州256603

出　　处：《热加工工艺》2013年第20期51-53,共3页Hot Working Technology

基　　金：山东省自然科学基金资助项目(ZR2011AM001);山东省科技发展计划资助项目(2011YD18027);滨州学院科研基金资助项目(BZXYL1305)

摘　　要：研究了Sr含量对Mg-Al-Si系镁合金中Mg-6Al-0.7Si-1Zn合金力学性能的影响,采用BP神经网络法建立了Mg-6Al-0.7Si-1Zn-xSr合金组织与力学性能的关系模型。采用BP神经网络预测的该合金力学性能与实验值接近,相对误差较小,最大误差为4.896%,最小误差仅为0.271%。结果表明,该模型具有很好的预测精度和较快收的敛速度,此模型的建立为研究Mg-Al-Si系镁合金提供了参考。The effect of Sr content on the mechanical properties of Mg-A1-Si alloy was studied and BP neural network was used to establish a relationship model between microstructure and mechanical properties of Mg-6A1-0.7Si-1Zn-xSr alloy. The input of the model is Sr content and the output is yield strength, tensile strength and elongation. The results show that the model has good prediction accuracy and fast convergence speed. The model can be used to further research the properties of Mg-A1-Si alloy.

关 键 词：Mg-6A1-0 7Si-1Zn镁合金 力学性能 BP神经网络 Sr含量 

分 类 号：TG113.25[金属学及工艺—物理冶金]