丁凤娟,贾向东,洪腾蛟,徐幼林.基于GA-BP和PSO-BP神经网络的6061铝合金板材流变应力预测模型[J].稀有金属材料与工程,2020,49(6):1840~1853.[Feng-juan Ding,Xiang-dong Jia,Teng-jiao Hong,You-lin Xu.Prediction model on flow stress of 6061 aluminum alloy sheet based on GA-BP and PSO-BP neural networks[J].Rare Metal Materials and Engineering,2020,49(6):1840~1853.]
基于GA-BP和PSO-BP神经网络的6061铝合金板材流变应力预测模型
投稿时间:2019-10-26  修订日期:2020-05-01
中文关键词:  6061铝合金  流变应力  人工神经网络  遗传算法  粒子群优化  热处理工艺
基金项目:江苏省高等学校自然科学基金(18KJB460020),南京林业大学高水平(高等教育)科学基金(GXL2018020)和南京林业大学青年科技创新基金(CX2018027)
中文摘要:
      6061铝合金作为一种热可强化铝合金,具有良好的成形性能,但是其塑性流变应力受最终热处理工艺的加热温度、保温时间和冷却方式等参数的影响很大。因此,为了获得最终热处理工艺参数对6061铝合金板材的塑性性能及流变行为的影响,试验中以6061-T6铝合金板材为研究对象,通过单向拉伸试验、金相实验和硬度测试等方法研究不同热处理工艺参数(加热温度为500、530、560和590℃、保温时间2小时、冷却方式为空冷)对6061铝合金塑性性能和硬度的影响。通过单向拉伸试验获取不同热处理工艺参数条件下6061铝合金的真实应力应变曲线;借助BP、GA-BP和PSO-BP神经网络构建不同热处理温度条件下6061铝合金的本构关系模型。研究结果表明BP、GA-BP和PSO-BP神经网络模型均能较好的拟合不同热处理温度条件下6061铝合金的流变行为,但是PSO-BP神经网络模型对6061铝合金流变应力的预测精度更高,网络预测性能更优越,其平均绝对误差(MAE),平均相对误差(AARE)和相关系数(R2)分别为1.89,1.56%和0.9965。
Prediction model on flow stress of 6061 aluminum alloy sheet based on GA-BP and PSO-BP neural networks
英文关键词:6061Aluminum Alloy  Flow stress  Artificial neural network  Genetic algorithm  Particle swarm optimization  Heat treatment Process
英文摘要:
      6061 aluminum alloy, as a kind of heat strengthened aluminum alloy, has good formability, but its plastic flow stress is greatly affected by the final heat treatment parameters, such as heating temperature,holding time and cooling method. Therefore, taking 6061-T6 aluminum alloy cold-rolled sheet as the research object, the plastic deformation behavior of 6061 aluminum alloy under different heat treatment temperatures (500 °C, 530 °C, 560 and 590 °C) were analyzed through uniaxial tensile test, metallographic test and microhardness test. Combined with experimental data and BP, GA-BP and PSO-BP neural networks, the constitutive models of this material under different heat treatment temperature conditions were constructed. The results show that BP, GA-BP and PSO-BP neural network models can better fit the flow behavior of 6061 aluminum alloy under different heat treatment temperature conditions, but PSO-BP neural network model has higher prediction accuracy and performs well in predicting the flow stress of 6061 aluminum alloy , its average absolute error (MAE), average relative error (AARE) and the correlation coefficient (R2) are 1.89, 1.56% and 0.9965, respectively.
作者单位E-mail
丁凤娟 机械电子工程学院 南京林业大学 dingfengjuan2018@163.com 
贾向东 机械电子工程学院 南京林业大学  
洪腾蛟 机械电子工程学院 南京林业大学  
徐幼林 机械电子工程学院 南京林业大学  
摘要点击次数: 143
全文下载次数: 88
查看全文  查看/发表评论  
关闭