-
合金凝固过程中的收缩行为是决定铸锭质量的关键因素之一，利用数值模拟方法可以预测铸锭缩孔。本文建立了一种基于机器学习的动网格模型，能够模拟铸件凝固过程的动态收缩行为。采用元胞自动机进行铸件凝固模拟，采用RBF算法(Radial Basis Function, RBF)和SVM算法(Support Vector Machines，SVM)计算凝固收缩过程的网格运动位移，从而对凝固过程收缩的动态模拟。采用该模型计算了Al-4.7%Cu合金铸锭的缩孔形貌，并进行了对应的浇铸实验验证，模拟结果与试验结果误差不超过2%，符合较好。说明该模型能够有效捕捉凝固收缩引起的铸件变形的动态过程，且能够捕捉固液界面复杂形貌的演变，为凝固过程数值模拟提供了一种新思路。Shrinkage cavities and porosity are the main defects generated during the solidification process of castings. The cause of these defects is the contraction of the alloy during solidification, with the last regions to solidify not receiving effective compensation from liquid metal, resulting in cavitation defects. Shrinkage cavities and porosity significantly reduce the mechanical properties of castings and shorten their service life, thus necessitating appropriate process elimination measures. Utilizing numerical simulation technology can effectively predict the shrinkage of castings during solidification and optimize the process based on simulation results, thereby reducing the occurrence of shrinkage defects, which is a low-cost and high-efficiency method. This paper presents a machine learning-driven dynamic mesh model to simulate the dynamic shrinkage behavior of castings during solidification. Cellular automata are used to simulate the solidification process of castings, dynamically marking the displacement of boundary points and calculating the displacement of other grids using RBF neural network algorithms and support vector machine algorithms, achieving dynamic simulation of the solidification process. The model was used to simulate the shrinkage cavity morphology of the Al-4.7%Cu alloy solidification process and corresponding casting experiments were designed for verification. Comparisons between simulation and experimental results indicate that this coupled method can effectively capture the deformation of castings caused by solidification shrinkage, the evolution of complex solid-liquid interface morphologies, and the deformation of internal grids within the castings. The simulation results have an error of no more than 2% compared to experimental results, providing a new approach for numerical simulation of the solidification process.
-
Keywords:
- machine learning /
- dynamic mesh /
- cellular automata /
- lattice Boltzmann
-
[1] Zhao J, Zhang Y 1985 Foundry 5 1(in Chinese)[赵健, 张毅 1985 铸造 5 1]
[2] Yu Z Y, Zhang H, Wang M L, Wang X B, Zhang K F, Wang F 2023 Special Casting & Nonferrous Alloys 41 1073(in Chinese)[俞占扬, 张慧, 王明林, 王学兵, 张开发, 王飞 2023 特种铸造及有色合金 41 1073]
[3] Niven R K 2002 Ground water 40 670
[4] Bhoraniya D, Dharaiya V, Sata A 2022 International Journal of Process Management and Benchmarking 12 395
[5] Jia B Q, Liu B C 1996 Hot Working Technology 2 34(in Chinese)[贾宝仟, 柳百成 1996 热加工工艺 2 34]
[6] Kang C, Son Y, Youn S 2001 Journal of Materials Processing Technology 113 251
[7] Carlson K D, Beckermann C 2009 Metallurgical and Materials Transactions A 40 163
[8] He D 2007 M.S. Thesis (Harbin: Harbin Institute Of Technology)(in Chinese)[何东 2007 硕士学位论文 （哈尔滨: 哈尔滨工业大学）]
[9] Khalajzadeh V, Carlson K D, Backman D G, Beckermann C 2017 Metallurgical and Materials Transactions A 48 1797
[10] Li K, Yin J, Lu Z, Kong X, Zhang R, Liu W Proceedings of the 21st International Conference on Pattern Recognition (ICPR2012) p170
[11] Li X, Wang H, Gu B, Ling C X Twenty-Fourth International Joint Conference on Artificial Intelligence
[12] Rendall T C, Allen C B 2009 Journal of Computational Physics 228 6231
[13] Beckert A, Wendland H 2001 Aerospace Science and Technology 5 125
[14] Dong S H, Zhang H W, Lv W M, Lei H, Wang Q 2024 Acta Metallurgica Sinica 60 388(in Chinese)[董士虎, 张红伟, 吕文朋, 雷洪, 王强 2024 金属学报 60 388]
[15] Yang C R, Sun D K, Pan S Y, Dai T, Zhu M F 2009 Acta Metallurgica Sinica 45 43(in Chinese)[杨朝蓉, 孙东科, 潘诗琰, 戴挺, 朱鸣芳 2009 金属学报 45 43]
[16] Yang Y Y, Li R, Zhou J C, Zhao C Y 2016 J. Eng. Thermophys 37 2613(in Chinese)[杨莹莹, 李日, 周靖超, 赵朝阳 2016 工程热物理学报 37 2613]
[17] JC Z 2017 M.S. Thesis (Tianjin: Hebei University of Technology)(in Chinese)[周靖超 2017 硕士学位论文 （天津: 河北工业大学）]
[18] Lian Q Q 2017 M.S. Thesis (Tianjin: Hebei University of Technology)(in Chinese)[连庆庆 2017 硕士学位论文 （天津: 河北工业大学）]
[19] Liu L 2018 M.S. Thesis (Tianjin: Hebei University of Technology)(in Chinese)[刘林 2018 硕士学位论文 （天津: 河北工业大学）]
[20] Ma W 2020 M.S. Thesis (Tianjin: Hebei University of Technology)(in Chinese)[马旺 2020 硕士学位论文 （天津: 河北工业大学）]
[21] Bai Y 2020 M.S. Thesis (Tianjin: Hebei University of Technology)(in Chinese)[白羽 2020 硕士学位论文 （天津: 河北工业大学）]
[22] Estruch O, Lehmkuhl O, Borrell R, Segarra C P, Oliva A 2013 Computers & Fluids 80 44
[23] Buhmann M D 2000 Acta numerica 9 1
[24] Vapnik V, Golowich S, Smola A 1996 Advances in neural information processing systems 9
[25] 郑楚光 2009 格子 Boltzmann 模型及数值仿真应用 (武汉: 华中科技大学)
[26] Wendland H 1998 Journal of approximation theory 93 258
计量
- 文章访问数: 173
- PDF下载量: 4
- 被引次数: 0
下载:
