Neural Network-Based Machine Learning Model for Spatiotemporal Prediction of Temperature and Fraction Solid in Low-Pressure Sand Casting

Abstract:

This paper presents a development methodology for a metamodel-based machine learning approach for spatiotemporal prediction of temperature and solid fraction evolution of aluminum castings during cooling under low pressure sand casting (LPSC) conditions, for various pouring temperatures (T_p) ranging from 614°C to 720°C. High-fidelity finite element (FE) simulations were performed, based on a representative case study produced by the LPSC process to generate a comprehensive database, recording nodal temperatures across the casting symmetry plane at different cooling times, and for several T_p values within the studied domain. Three different machine learning (ML) algorithms were evaluated using comparative metrics (R², MAE and MSE). Among the evaluated algorithms, the artificial neural network (ANN)-based ML model was selected for its superior predictive accuracy and robustness. The accuracy of the selected ML-model was assessed by comparing predicted and FE-simulated temperature fields. The results indicate that the predicted temperature error within the cast symmetry plane remains below 1%. Furthermore, a graphical user interface (GUI) was developed to visualize the predicted casting temperature field for different T_p values not used during the learning stage, as well as the corresponding solid fraction, which is computed based on the solidification curve of the aluminum alloy AlSi₇Mg_0.3 given by the ProCAST^® database. This methodology could provide a fast, robust, and scalable framework for extending predictive models to higher-dimensional cases and diverse LPSC casting process configurations.

• Periodical: Materials Science Forum (Volume 1188)
• Pages: 1-10
• Online since: April 2026
• Authors: Ahmed Ktari*, Souhail Housni, Jérémie Bourgeois, Mohamed El Mansori
• Keywords: Cast Solidification, Fe Simulation, Low-Pressure Casting, ML Prediction

References

[1] I. Jalilvand, J. Jang, B. Gopaluni, A.S. Milani, VR/MR systems integrated with heat transfer simulation for training of thermoforming: A multicriteria decision-making user study, Journal of Manufacturing Systems 72 (2024) 338–359.

DOI: 10.1016/j.jmsy.2023.11.007

[2] A. Ktari, M. El Mansori, Digital twin of functional gating system in 3D printed molds for sand casting using a neural network, J Intell Manuf 33 (2022) 897–909.

DOI: 10.1007/s10845-020-01699-3

[3] F. Deng, X. Rui, S. Lu, Z. Liu, H. Sun, W. Volk, Deep learning–based inline monitoring approach of mold coating thickness for Al-Si alloy permanent mold casting, Int J Adv Manuf Technol 130 (2024) 565–573.

DOI: 10.1007/s00170-023-12709-y

[4] A. Ktari, M. El Mansori, Intelligent approach based on FEM simulations and soft computing techniques for filling system design optimisation in sand casting processes, Int J Adv Manuf Technol 114 (2021) 981–995.

DOI: 10.1007/s00170-021-06876-z

[5] A. Ktari, M. Elmansori, Bridging FEM and Artificial Neural Network in gating system design for smart 3D sand casting, Procedia Manufacturing 51 (2020) 795–800.

DOI: 10.1016/j.promfg.2020.10.111

[6] I. Pastor-López, B. Sanz, A. Tellaeche, G. Psaila, J.G. De La Puerta, P.G. Bringas, Quality assessment methodology based on machine learning with small datasets: Industrial castings defects, Neurocomputing 456 (2021) 622–628.

DOI: 10.1016/j.neucom.2020.08.094

[7] Z. Duan, W. Chen, X. Pei, H. Hou, Y. Zhao, A multimodal data-driven design of low pressure die casting gating system for aluminum alloy cabin, Journal of Materials Research and Technology 27 (2023) 2723–2736.

DOI: 10.1016/j.jmrt.2023.10.076

[8] S. Shahane, N. Aluru, P. Ferreira, S.G. Kapoor, S.P. Vanka, Optimization of solidification in die casting using numerical simulations and machine learning, Journal of Manufacturing Processes 51 (2020) 130–141.

DOI: 10.1016/j.jmapro.2020.01.016

[9] G.-C. Vosniakos, V. Galiotou, D. Pantelis, P. Benardos, P. Pavlou, The scope of artificial neural network metamodels for precision casting process planning, Robotics and Computer-Integrated Manufacturing 25 (2009) 909–916.

DOI: 10.1016/j.rcim.2009.04.018

[10] A.P.M. Diniz, P.M. Ciarelli, E.O.T. Salles, K.F. Coco, Use of deep neural networks for clogging detection in the Submerged Entry Nozzle of the continuous casting, Expert Systems with Applications 238 (2024) 121963.

DOI: 10.1016/j.eswa.2023.121963

[11] G.T. Kridli, P.A. Friedman, J.M. Boileau, Manufacturing processes for light alloys, in: Materials, Design and Manufacturing for Lightweight Vehicles, Elsevier, 2021: p.267–320.

DOI: 10.1016/B978-0-12-818712-8.00007-0

[12] A. Ktari, M.E. Mansori, Towards remanufacturing of failed parts through rapid low-pressure sand-casting (LPSC) process, Int J Adv Manuf Technol 139 (2025) 2569–2588.

DOI: 10.1007/s00170-025-16049-x

[13] D. Rajput, W.-J. Wang, C.-C. Chen, Evaluation of a decided sample size in machine learning applications, BMC Bioinformatics 24 (2023) 48.

DOI: 10.1186/s12859-023-05156-9

[14] O.A. Montesinos López, A. Montesinos López, J. Crossa, Overfitting, Model Tuning, and Evaluation of Prediction Performance, in: Multivariate Statistical Machine Learning Methods for Genomic Prediction, Springer International Publishing, Cham, 2022: p.109–139.

DOI: 10.1007/978-3-030-89010-0_4