Abstract
Macrosegregation is of great importance to study due to its negative impact on the quality of casting. Although numerical models can predict macrosegregation during alloy solidification, solving the partial differential equations is rather time-consuming. Thus, numerical simulations are almost inoperable for the real-time online monitor-adjustment in industrial production, where the prediction is expected to be completed in an extremely short time. To overcome this challenge, a data-driven approach based on deep learning is proposed to predict the macrosegregation pattern under specific input parameter(s). Based on limited simulation results, this approach focuses on mining certain patterns within massive data, and thus enables fast predictions of macrosegregation, by incorporating a convolutional neural network autoencoder with a fully connected neural network. The best prediction accuracy is achieved after clarifying the effects of the error metric and the convolutional filter size. This method can predict the macrosegregation distribution in less than 0.1 second, and the accuracy is comparable to the conventional numerical simulations. The data-driven approach developed in this work shows instantaneity and adequate accuracy in the prediction of macrosegregation and could be a promising method for the application in the direct visualization and quality control of casting.
Graphical Abstract
Similar content being viewed by others
Abbreviations
- \({C}_{ls}^{p}\) :
-
Species exchange due to phase change (\(\mathrm{kg }{{\text{m}}}^{-3} \, {\text{s}}^{-1}\))
- \({c}_{0}\) :
-
Initial concentration of the alloy (-)
- \({c}_{l},{c}_{s}\) :
-
Species concentration (-)
- \({c}_{{\text{mix}}}\) :
-
Mixture concentration (-)
- \({c}_{p}^{l},{c}_{p}^{s}\) :
-
Specific heat (\(\mathrm{J }{{\text{kg}}}^{-1} {\text{K}}^{-1}\))
- \({D}_{l}\) :
-
Diffusion coefficient (\({{\text{m}}}^{2} {{\text{s}}}^{-1}\))
- \({\overrightarrow{F}}_{l}\) :
-
Buoyancy force (\({\text{N}}{{\text{m}}}^{-3}\))
- \({f}_{l},{f}_{s}\) :
-
Volume fraction (-)
- \(\overrightarrow{g}\) :
-
Gravity (\(\mathrm{m }{{\text{s}}}^{-2}\))
- \(h\) :
-
Heat transfer coefficient (\(\mathrm{W }{{\text{m}}}^{-2} {\text{K}}^{-1}\))
- \({h}_{l},{h}_{s}\) :
-
Enthalpy (\(\mathrm{J }{{\text{kg}}}^{-1}\))
- \(\Delta {h}_{f}\) :
-
Latent heat (\(\mathrm{J }{{\text{kg}}}^{-1}\))
- \(K\) :
-
Permeability (\({{\text{m}}}^{2}\))
- \(k\) :
-
Solute partition coefficient (-)
- \({k}_{l},{k}_{s}\) :
-
Thermal conductivity (\(\mathrm{W }{{\text{m}}}^{-1} {\text{K}}^{-1}\))
- \({M}_{ls}\) :
-
Solidification mass transfer rate (\(\mathrm{kg }{{\text{s}}}^{-1} {\text{m}}^{-3}\))
- \(m\) :
-
Liquidus slope of the binary phase diagram (\(\mathrm{K }{\left({\text{wt}}\,{\text{pct}}\right)}^{-1}\))
- \(p\) :
-
Pressure (\({\text{Pa}}\))
- \({Q}_{ls}^{d}\) :
-
Energy exchange by heat transfer (\(\mathrm{J }{{\text{m}}}^{-3} {\text{s}}^{-1}\))
- \({Q}_{l}^{p},{Q}_{s}^{p}\) :
-
Energy source term due to phase change (\(\mathrm{J }{{\text{m}}}^{-3} {\text{s}}^{-1}\))
- \(q,\widehat{q}\) :
-
True value and predict value
- \(T,{T}_{l},{T}_{s}\) :
-
Temperature (\({\text{K}}\))
- \({T}_{{\text{EXT}}}\) :
-
External temperature (\({\text{K}}\))
- \(t\) :
-
Time (\({\text{s}}\))
- \({\overrightarrow{U}}_{ls}^{d}\) :
-
Momentum change due to drag force (\(\mathrm{kg }{{\text{m}}}^{-2} {\text{s}}^{-2}\))
- \({\overrightarrow{U}}_{ls}^{p}\) :
-
Momentum exchange due to phase change (\(\mathrm{kg }{{\text{m}}}^{-2} {\text{s}}^{-2}\))
- \(\overrightarrow{u}\) :
-
Velocity vector of the melt (\(\mathrm{m }{{\text{s}}}^{-1}\))
- \({\beta }_{c}\) :
-
Solutal expansion coefficient (\({\text{wt}}\,{\text{pct}}^{-1}\))
- \({\beta }_{T}\) :
-
Thermal expansion coefficient (\({{\text{K}}}^{-1}\))
- \({\lambda }_{1},{\lambda }_{2}\) :
-
Primary and second dendrite arm spacing of columnar (\({\text{m}}\))
- \({\mu }_{l}\) :
-
Viscosity (\(\mathrm{kg }{{\text{m}}}^{-1} {\text{s}}^{-1}\))
- \({\rho }_{l},{\rho }_{s}\) :
-
Density (\(\mathrm{kg }{{\text{m}}}^{-3}\))
- \({\overline{\overline{\tau }}}_{l}\) :
-
Stress-strain tensors (\(\mathrm{kg }{{\text{m}}}^{-1} {\text{s}}^{-2}\))
- \(\varepsilon \) :
-
Error metric (-)
- \(l,s\) :
-
Indicate liquid and solid phases
- \(mae\) :
-
Indicate absolute error
- \(mse\) :
-
Indicate mean squared error
- \(gdl\) :
-
Indicate a gradient difference loss
- \(mse-gdl\) :
-
Indicate the combination of mean squared error and the gradient difference loss
References
Iron Steel Institute: J. Iron Steel Inst., 1926, vol. 103, pp. 39–151.
E. Marburg: JOM, 1953, vol. 5, pp. 157–72.
A. Olsson, R. West, and H. Fredriksson: Scand. J. Metall., 1986, vol. 15, pp. 104–12.
R.J. Mcdonald and J.D. Hunt: Trans. Metall. Soc. Aime, 1969, vol. 245, pp. 1993–97.
R.J. Mcdonald and J.D. Hunt: Metall. Trans., 1970, vol. 1, pp. 1787–88.
T.F. Bower, H.D. Brody, and M.C. Flemings: Trans. Metall. Soc. Aime, 1966, vol. 236, pp. 624–34.
H.D. Brody and M.C. Flemings: Trans. Metall. Soc. Aime, 1966, vol. 236, pp. 615–24.
M.C. Flemings and G.E. Nereo: Trans. Metall. Soc. Aime, 1967, vol. 239, pp. 1449–61.
M.C. Flemings, R. Mehrabian, and G.E. Nereo: Trans. Metall. Soc. Aime, 1968, vol. 242, pp. 41–49.
A. Ludwig and M. Wu: Mater. Sci. Eng. A, 2005, vol. 413, pp. 109–14.
M.H. Wu and A. Ludwig: Metall. Mater. Trans. A, 2006, vol. 37A, pp. 1613–31.
H.H. Ge, F.L. Ren, J. Li, X.J. Han, M.X. Xia, and J.G. Li: Metall. Mater. Trans. A, 2017, vol. 48A, pp. 1139–50.
M. Wu, A. Ludwig, and A. Kharicha: Appl. Math. Model., 2017, vol. 41, pp. 102–20.
H.J. Yang, S.P. Vanka, and B.G. Thomas: Isij Int., 2019, vol. 59, pp. 956–72.
X.W. Xu, J. Li, Z.H. Dong, L. Zeng, Z.M. Xu, and J.G. Li: J. Mater. Res. Technol., 2022, vol. 18, pp. 830–40.
J. Li, M.H. Wu, A. Ludwig, and A. Kharicha: Int. J. Heat Mass. Trans., 2014, vol. 72, pp. 668–79.
M. Xie and H. Shen: Front. Mater., 2020, vol. 7, p. 577290.
H.H. Ge, J. Li, Q.T. Guo, F.L. Ren, M.X. Xia, J.H. Yao, and J.G. Li: Metall. Mater. Trans. B, 2021, vol. 52, pp. 2992–3003.
R. Ramprasad, R. Batra, and G. Pilania: NPJ Comput. Mater., 2017, vol. 3, p. 54.
K.T. Butler, D.W. Davies, H. Cartwright, O. Isayev, and A. Walsh: Nature, 2018, vol. 559, pp. 547–55.
A. Agrawal and A. Choudhary: MRS Commun., 2019, vol. 9, pp. 779–92.
J. Schmidt, M.R.G. Marques, S. Botti, and M.A.L. Marques: NPJ Comput. Mater., 2019, vol. 5, p. 83.
C. Yang, Y. Kim, S. Ryu, and G.X. Gu: Mater. Des., 2020, vol. 189, p. 108509.
Z.B. Wei, J.X. Yu, Y. Lu, J.J. Han, C.P. Wang, and X.J. Liu: Mater. Des., 2021, vol. 198, p. 109287.
Q. Xie, M. Suvarna, J.L. Li, X.Z. Zhu, J.J. Cai, and X.N. Wang: Mater. Des., 2021, vol. 197, p. 109201.
Z.R. Pei and J.Q. Yin: Mater. Des., 2019, vol. 172, p. 107759.
B. Meredig, A. Agrawal, S. Kirklin, J.E. Saal, J.W. Doak, A. Thompson, K. Zhang, A. Choudhary, and C. Wolverton: Phys. Rev. B, 2014, vol. 89, p. 094104.
Y. Liu, T. Zhao, Ju. Wangwei, and S. Shi: J. Materiom., 2017, vol. 3, pp. 159–77.
J.R. Li, R. Jin, and H.Z. Yu: Mater. Des., 2018, vol. 139, pp. 473–85.
Q. Zhu, A. Samanta, B. Li, R.E. Rudd, and T. Frolov: Nat. Commun., 2018, vol. 9, p. 467.
W.S. Yan, J. Melville, V. Yadav, K. Everett, L. Yang, M.S. Kesler, A.R. Krause, M.R. Tonks, and J.B. Harley: Mater. Des., 2022, vol. 222, p. 111032.
K. Ryan, J. Lengyel, and M. Shatruk: J. Am. Chem. Soc., 2018, vol. 140, pp. 10158–68.
S. Feng, H. Zhou, and H. Dong: Mater. Des., 2019, vol. 162, pp. 300–10.
G.W. Song, B.A. Tama, J. Park, J.Y. Hwang, J. Bang, S.J. Park, and S. Lee: Steel Res. Int., 2019, vol. 90, p. 1900321.
S.Y. Lee, B.A. Tama, C. Choi, J.-Y. Hwang, J. Bang, and S. Lee: IEEE Access, 2020, vol. 8, pp. 21953–65.
J. Li, M. Wu, J. Hao, A. Kharicha, and A. Ludwig: Comput. Mater. Sci., 2012, vol. 55, pp. 419–29.
J. Li, M. Wu, J. Hao, and A. Ludwig: Comput. Mater. Sci., 2012, vol. 55, pp. 407–18.
M. Bellet, H. Combeau, Y. Fautrelle, D. Gobin, M. Rady, E. Arquis, O. Budenkova, B. Dussoubs, Y. Duterrail, A. Kumar, C.A. Gandin, B. Goyeau, S. Mosbah, and M. Zaloznik: Int. J. Therm. Sci., 2009, vol. 48, pp. 2013–16.
C. Beckermann, J.P. Gu, and W.J. Boettinger: Metall. Mater. Trans. A, 2000, vol. 31A, pp. 2545–57.
Y.F. Cao, Y. Chen, D.Z. Li, H.W. Liu, and P.X. Fu: Metall. Mater. Trans. A, 2016, vol. 47A, pp. 2927–39.
Q. Li, J. Shen, L. Qin, and Y. Xiong: Mater Charact, 2017, vol. 130, pp. 139–48.
Q. Li, J. Shen, Y. Xiong, L. Qin, and X. Yue: Mater. Lett., 2018, vol. 228, pp. 281–84.
C. Dong, C.C. Loy, K. He, and X. Tang: IEEE Trans. Pattern Anal. Mach. Intell., 2016, vol. 38, pp. 295–307.
H.C. Shin, H.R. Roth, M. Gao, L. Lu, Z. Xu, I. Nogues, J. Yao, D. Mollura, and R.M. Summers: IEEE Trans. Med. Imaging, 2016, vol. 35, pp. 1285–98.
V. Badrinarayanan, A. Kendall, and R. Cipolla: IEEE Trans. Pattern Anal. Mach. Intell., 2017, vol. 39, pp. 2481–95.
A. Esteva, B. Kuprel, R.A. Novoa, J. Ko, S.M. Swetter, H.M. Blau, and S. Thrun: Nature, 2017, vol. 542, pp. 115–18.
A. Krizhevsky, I. Sutskever, and G. Hinton: Commun. ACM, 2017, vol. 60, pp. 84–90.
C. Farabet, C. Couprie, L. Najman, and Y. Lecun: IEEE Trans. Pattern Anal. Mach. Intell., 2013, vol. 35, pp. 1915–29.
G. Hinton, Li. Deng, Yu. Dong, G. Dahl, A.-R. Mohamed, N. Jaitly, A. Senior, V. Vanhoucke, P. Nguyen, T. Sainath, and B. Kingsbury: IEEE Signal Process. Mag., 2012, vol. 29, pp. 82–97.
A.B. Nassif, I. Shahin, I. Attili, M. Azzeh, and K. Shaalan: IEEE Access, 2019, vol. 7, pp. 19143–65.
M.T. Anvarjon and S. Kwon: Sensors, 2020, vol. 20, p. 5212.
B.J. Abbaschian, D. Sierra-Sosa, and A. Elmaghraby: Sensors-Basel, 2021, vol. 21, p. 1249.
K. Hasegawa, K. Fukami, T. Murata, and K. Fukagata: Theoret. Comput. Fluid Dyn., 2020, vol. 34, pp. 367–83.
D. Chicco, M.J. Warrens, and G. Jurman: Peer J. Comput. Sci., 2021, vol. 7, p. 623.
S. Lee and D. You: J. Fluid Mech., 2019, vol. 879, pp. 217–54.
Acknowledgments
This work was sponsored by National Natural Science Foundation of China (No. 52074182 and No. 91860121), Natural Science Foundation of Shanghai (No. 22ZR1430700) and the Guided Local Science and Technology Development Projects from the Central Government (No. 2021FRD05006).
Competing Interest
The authors declare that they have no conflict of interest.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Xu, X., Ren, N., Lu, Z. et al. A Data-Driven Approach for the Fast Prediction of Macrosegregation. Metall Mater Trans A (2024). https://doi.org/10.1007/s11661-024-07381-0
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s11661-024-07381-0