Al-Cu-Mg-Ag合金热压缩变形行为的预测  

Behavior Prediction of Al-Cu-Mg-Ag Alloy during Hot Compression Deformation

作　　者：陆智伦[1] 潘清林[1,2] 曹素芳[1] 刘晓艳[1] 

机构地区：[1]中南大学材料科学与工程学院,湖南长沙410083 [2]有色金属材料科学与工程教育部重点实验室,湖南长沙410083

出　　处：《稀有金属》2011年第2期176-183,共8页Chinese Journal of Rare Metals

基　　金：国防科工委军品配套研制资助项目(JPPT-115-2-948)资助项目

摘　　要：采用了热模拟实验机研究了Al-Cu-Mg-Ag耐热铝合金的热压缩变形行为。实验的温度和应变速率分别为340～500℃,0.001～10 s-1。分别用了本构方程和人工神经网络来对Al-Cu-Mg-Ag合金的流变行为进行了分析和模拟。神经网络的结构是3-20-1;输入参数是温度,应变速率和应变;输出参数是流变应力。结果表明该合金的流变曲线出现加工硬化、过渡、软化和稳态流变这4个阶段,流变应力随着应变速率的增加而增大,随着变形温度的下降而减少。用所建立的神经网络模型预测了变形温度和应变速率对流变应力的影响,预测的结果与热压缩变形的基础理论吻合得很好,而且该模型可以很好地描述Al-Cu-Mg-Ag合金的流变应力,在应变速率为0.001～10 s-1的条件下,其平均相对误差分别为3.68%,3.98%,1.53%,3.53%和2.04%。这表明神经网络的预测性能优良,具有很强的推广能力。同时通过本构方程和神经网络的预测结果比较看出神经网络模型的相关系数比较高,而且神经网络比本构方程有更好的预测性能。神经网络可以预测不同应变下的相应的流变应力,但是本构方程只可以根据不同的应变速率和温度来预测峰值应力。The behavior of Al-Cu-Mg-Ag heat-resistant aluminum alloy during hot compression deformation was studied by thermal simulation test.The experimental temperatures and strain rates were 340~500 ℃ and 0.001~10 s-1.Constitutive equations and an artificial neural network（ANN） model were developed for the analysis and simulation of the flow behavior of the Al-Cu-Mg-Ag alloy respectively.The architecture of ANN became 3-20-1.The input parameters were temperature,strain rate and strain.The output parameter was the flow stress.The results showed that the flow stress curves of the alloy were composed of four stages,i.e.,work hardening stage,transition stage,softening stage and steady stage.The flow stress increased with increasing the strain rate,and decreased with increasing the deforming temperature.The proposed ANN model was used to predict effect of deformation temperature and strain rate on the flow behavior.And predicted results were consistent with the fundamental theory of hot compression deformation.The established model could also delineate the flow behavior of Al-Cu-Mg-Ag alloy precisely.When the strain rates were 0.001~10 s-1,average absolute relative errors were 3.68 %,3.98%,1.53%,3.53% and 2.04%,respectively,which showed that ANN had excellent prediction performance and strong generalization capability.Comparison between constitutive equation and ANN results showed that correlation coefficient of ANN model was higher.It had a better prediction power than constitutive equations.And ANN model could predict the corresponding flow stress at different strain,but constitutive equations could only predict the peak stress depending on different strain rates and temperatures.

关 键 词：AL-CU-MG-AG合金 本构方程 人工神经网络 热压缩变形 流变应力 

分 类 号：TG146.21[一般工业技术—材料科学与工程]