材料超塑变形后室温机械性能的变化和预测  

Predicting the Mechanical Behavior of Superplastically Deformed Materials with Artificial Neural Networks

作　　者：刘马宝[1] 高宏 李淼泉[1] 吴诗[1] 

机构地区：[1]西北工业大学材料科学与工程学院,西安710072

出　　处：《热加工工艺》1998年第3期3-5,共3页Hot Working Technology

基　　金：国家自然科学基金!59471071

摘　　要：超塑性材料在变形过程中往往空洞化。空洞的存在严重降低超塑成形零件的室温使用性能，因此必须建立超塑变形后材料室温机械性能变化的理论预测模型。本文以铝合金LY12CZ为例，以实验数据为基础，利用人工神经网络首次建立了预测经超塑变形后的材料室温机械性能变化的理论模型。所建模型不但可以预测铝合金LY12CZ超塑变形后的刚度．强度以及韧性等室温性能指标，而且亦能充分反映超塑变形工艺参数对其室温机械性能变化的影响规律。同时，由于本文建模方法具有通用性，因此，该模型的建立为超塑成形零件的使用性能提供了理论依据和一般方法。Superplastic materials generally cavitate during superplastic deformation.The cavitate durign superplastic deformation. The eavitation leads to the decreasing of froming limit and to the worsening of the superplastically deformed materials mechanical properties. We deem that description of superplastically deformed materials mechanical properties at room temperature requires a suitable model. Up to now,owing to the complexity of the problem,such a model has not been found in the public published literatures. we present a model which is established by training the artificial neural network(ANN)with the mechanical properties obtained on duralumin LY12CZ sheet under different processing parameters of superplastic deformation. It is found that the mechanical properties predicted by the trained artificial neural network agree closely with the actual experimental results. The trained arrificial neural network also reveals the influence of processing parameters on the material mechanical properries after superplastic deformation. The results from the network are very encouraging.

关 键 词：超塑变形 空洞 机械性能 人工神经网络 铝合金 

分 类 号：TG301[金属学及工艺—金属压力加工] TG146.21[一般工业技术—材料科学与工程]