基于RBF神经网络预测2219铝合金多层TIG焊接头拉伸性能研究
|
摘要
针对焊接过程的严重非线性和焊接工艺参数的复杂交互作用,使得接头抗拉强度的准确估算十分困难的问题,利用K-均值聚类算法建立焊接接头抗拉强度RBF (Radial basic function)神经网络预测模型。以2219铝合金多层TIG焊的焊缝形状参数(正面余高、背面余高、背面宽度、正面宽度、首层盖面熔深)作为预测模型的输入参数,以焊后接头的抗拉强度作为预测模型的输出参数。利用22组焊接试验数据对所建的预测模型进行网络训练,用另外6组焊接试验数据进行验证。结果表明,所建的RBF神经网络预测模型具有适应性强、精度高,预测的平均绝对误差仅为1.94%。通过预测模型,得到焊缝几何形状参数与接头抗拉强度之间的影响规律以及最佳的接头焊缝形状,从而对调控焊接接头的强度提供了依据。
引文
[1] Dengkui Zhang, Quan Li, Yue Zhao, et al. Microstructure and mechanical properties of three-layer TIG-welded 2219 aluminum alloys with dissimilar heat treatments[J]. Journal of Materials Engineering and Performance, 2018, 27(6):2 938-2 948.
[2] LI Quan, WU Ai-ping, LI Yan-jun, et al. Influence of temperature cycles on the microstructures and mechanical properties of the par-tially melted zone in the fusion welded joints of 2219 aluminum alloy[J]. Materials Science&Engineering A, 2015, 623:38-48.
[3]王国庆,熊林玉,田志杰,等.不同热处理状态2219铝合金TIG焊接头组织性能分析[J].焊接学报, 2017, 38(1):121-125.
[4] Chan B, Pacey J and Bipbi M. Modeling gas metal arc weld geometry using artificial neural network technology[J]. Canadian Metallurgical Quaterly, 1999, 38(1):43-51.
[5] Casalino G, Facchini F, Mortello M, et al. ANN modeling to optimize manufacturing processes:the case of laser welding[J]. IAFC,2016:378-383.
[6]董志泼,魏艳红,占小红,等.遗传算法与神经网络结合优化焊接接头力学性能预测模型[J].焊接学报, 2007, 28(12):69-74.
[7] Anand K, Birendra Kumar Barik and Tamilanman K. Artificial neural network modeling studies to predict the friction welding process parameter of Incoloy 800H joint[J]. Engineering Science Technology, 2015(18):394-407.
[8] Pal S, Pal S K and Samantalay A K. Radian basic function neural network model based prediction of weld plate distortion due to pulsed metal inert gas welding[J]. Science and Technology of Weld-ing and Joining, 2007, 12(8):725-731.
[9]唐正魁,董俊慧,张永志.混合聚类RBF神经网络焊接接头力学性能预测[J].焊接学报, 2014, 35(12):105-108.
[10]张永志,董俊慧.基于模糊C均值的模糊RBF神经网络焊接接头力学性能建模[J].机械工程学报, 2014, 50(12):58-64.
[11] Dengkui Zhang, Guoqing Wang, Aiping Wu, et al. Study on the inconsistency in mechanical properties of 2219 aluminium alloy TIG-welded joints[J]. Journal of Alloys and Compounds, 2019, 777:1 044-1 053.
[12] Sagai Francis Britto A, Edvin Raj R and Carolin Mabel M. Prediction of shear and tensile strength of the diffusion bonded AA5083 and AA7075 aluminium alloy using ANN[J]. Materials Sicence and Engineering A, 2017:1-8.
[13]张彬.基于聚类算法的RBF神经网络设计综述[J].微型机及应用, 2012, 31(12):1-3, 7.
[2] LI Quan, WU Ai-ping, LI Yan-jun, et al. Influence of temperature cycles on the microstructures and mechanical properties of the par-tially melted zone in the fusion welded joints of 2219 aluminum alloy[J]. Materials Science&Engineering A, 2015, 623:38-48.
[3]王国庆,熊林玉,田志杰,等.不同热处理状态2219铝合金TIG焊接头组织性能分析[J].焊接学报, 2017, 38(1):121-125.
[4] Chan B, Pacey J and Bipbi M. Modeling gas metal arc weld geometry using artificial neural network technology[J]. Canadian Metallurgical Quaterly, 1999, 38(1):43-51.
[5] Casalino G, Facchini F, Mortello M, et al. ANN modeling to optimize manufacturing processes:the case of laser welding[J]. IAFC,2016:378-383.
[6]董志泼,魏艳红,占小红,等.遗传算法与神经网络结合优化焊接接头力学性能预测模型[J].焊接学报, 2007, 28(12):69-74.
[7] Anand K, Birendra Kumar Barik and Tamilanman K. Artificial neural network modeling studies to predict the friction welding process parameter of Incoloy 800H joint[J]. Engineering Science Technology, 2015(18):394-407.
[8] Pal S, Pal S K and Samantalay A K. Radian basic function neural network model based prediction of weld plate distortion due to pulsed metal inert gas welding[J]. Science and Technology of Weld-ing and Joining, 2007, 12(8):725-731.
[9]唐正魁,董俊慧,张永志.混合聚类RBF神经网络焊接接头力学性能预测[J].焊接学报, 2014, 35(12):105-108.
[10]张永志,董俊慧.基于模糊C均值的模糊RBF神经网络焊接接头力学性能建模[J].机械工程学报, 2014, 50(12):58-64.
[11] Dengkui Zhang, Guoqing Wang, Aiping Wu, et al. Study on the inconsistency in mechanical properties of 2219 aluminium alloy TIG-welded joints[J]. Journal of Alloys and Compounds, 2019, 777:1 044-1 053.
[12] Sagai Francis Britto A, Edvin Raj R and Carolin Mabel M. Prediction of shear and tensile strength of the diffusion bonded AA5083 and AA7075 aluminium alloy using ANN[J]. Materials Sicence and Engineering A, 2017:1-8.
[13]张彬.基于聚类算法的RBF神经网络设计综述[J].微型机及应用, 2012, 31(12):1-3, 7.