7075/2A12异种铝合金搅拌摩擦焊搭接接头疲劳寿命预测研究

Fatigue life prediction of 7075/2A12 dissimilar aluminum alloys friction stir welding lap joints

对7075/2A12异种铝合金搅拌摩擦焊搭接接头进行了疲劳加载测试,并观察了疲劳断口特征。根据实际试件结构建立局部应力应变法和缺口应力法有限元模型,对焊缝区域进行应力应变分析;结合Soderberg修正方程对Manson-Coffin(MC)模型中应力部分进行修正,再根据有限元分析结果,使用局部应力应变法和缺口应力法对异种铝合金搅拌摩擦焊搭接接头进行疲劳寿命预测。试验结果表明,有效搭接厚度严重影响试件的疲劳寿命;位于受载7075前进侧的钩状缺陷比7075后退侧钩状缺陷对疲劳寿命的影响大。与试验结果相比,局部应力应变法中,2组试件最小板厚处疲劳寿命的预测结果都要比钩状缺陷根部处疲劳寿命的预测精度高。部分试件结合Soderberg方程修正的MC模型预测误差大部分都在2个因子之内,而Smith-Watson-Topper(SWT)模型和其他MC修正模型预测结果均在3个因子之内,Sachs修正的MC模型和Morrow修正的MC模型的预测结果相近;缺口应力法中,部分试件预测结果较好,误差在2个因子之内。

Fatigue loading tests were carried out on 7075/2A12 dissimilar aluminum alloy friction stir welding lap joints,and the fatigue fracture characteristics were observed.The finite element models for both local stress-strain method and notched stress method were established according to the actual specimen structure,and the stress-strain at the weld zone was analyzed.The stress component of Manson-Coffin(MC)model was modified by Soderberg correction equation.Combined with the results of finite element analysis,the fatigue life of dissimilar aluminum alloy friction stir welding lap joint was predicted by local stress-strain method and notched stress method.The experimental results showed that the effective lap thickness has a great effect on the fatigue life of the specimen,and the hook-like defects on the loaded forward side of 7075 have a greater effect on the fatigue life than that on the backward side of 7075.Compared with the experimental results,with the local stress-strain method,the predicted fatigue life at the minimum thickness of the two groups of specimens is more accurate than that at the root of the hook-like defect.The predicted error of MC model modified by Soderberg equation is mostly within two factors,while that of Smith-Watson-Topper(SWT)model and other MC models are all within three factors,the predicted results of MC model modified by Sachs and MC model modified by Morrow are similar;while with the notched stress method,some specimens predicted better results,and the error is within two factors.