Residual Learning-Based Synthetic Data for Hybrid Metamodeling in the Stamping of an Automotive Door Panel

Article Preview

Abstract:

Hybrid twins and residual-learning strategies are increasingly used to reconcile the broad coverage of physics-based simulations with the fidelity of production measurements. In industrial stamping, however, truly matched simulation-experiment cases are scarce, while large-scale monitoring data are clustered around a narrow operating window. This work proposes an industrially practical hybrid metamodel in which the discrepancy (ignorance) model is trained exclusively from approximate residuals computed at in-domain production inputs, considering a surrogate of a thermal enabled AutoForm Sigma Design of Experiments (DoE). To prevent uncontrolled extrapolation, learning and evaluation are restricted to an explicitly validated domain defined through multivariate kNN support in standardized space derived from the DoE cloud. The residual model is selected through five-fold cross-validation on 10,446 in-domain production samples and then retrained on the full approximate-residual set. A small set of five matched cases is kept as an external check based on true residuals. The resulting hybrid predictor enables the generation of synthetic, experiment-informed corrected data while retaining the DoE coverage required for downstream modelling tasks.

You have full access to the following eBook

Info:

Periodical:

Pages:

103-111

Citation:

Online since:

April 2026

Funder:

The publication of this article was funded by the Mondragon Goi Eskola Politeknikoa, J.M.A. S.Coop

Export:

Share:

Citation:

* - Corresponding Author

[1] P. de Souza and B. Rolfe, "Characterising material and process variation effects on springback in sheet metal forming," Int. J. Mech. Sci., 2010.

DOI: 10.1016/j.ijmecsci.2010.09.009

Google Scholar

[2] M. Golmohammadi et al., "A machine learning-based model to predict residual stress distribution considering FE modelling and experimental uncertainties," Int. J. Mech. Sci., 2025.

Google Scholar

[3] L. Muñiz, L. Galdos, and J. Trinidad, "Metamodel-based control algorithms for the correction of bending angle after springback in an industrial U-Bending process," Int. J. Mater. Form., vol. 18, art. no. 44, p.1–20, May 2025.

DOI: 10.1007/s12289-025-01906-7

Google Scholar

[4] S. B. Choi and co-authors, "Replacing FEA for sheet metal forming by surrogate modeling," Cogent Eng., 2014.

Google Scholar

[5] F. Chinesta, A. Huerta, G. Rozza, and K. Willcox, "Virtual, Digital and Hybrid Twins: A New Paradigm in Data-Based Engineering and Engineered Data," Arch. Comput. Methods Eng., 2019 (online 2018).

DOI: 10.1007/s11831-018-9301-4

Google Scholar

[6] S. Torregrosa, F. Chinesta, and co-authors, "Hybrid twins based on optimal transport," Appl. Math. Comput., 2022.

Google Scholar

[7] A. Forrester, A. Sóbester, and A. Keane, "Engineering Design via Surrogate Modelling," Wiley, 2008.

Google Scholar

[8] AutoForm Engineering GmbH, "AutoForm-Sigma: Robust Processes," product brochure, 2024.

Google Scholar

[9] L. Muñiz, J. Trinidad, E. Garcia, I. Peinado, N. Montes, and L. Galdos, "On the Use of Advanced Friction Models for the Simulation of an Industrial Stamping Process including the Analysis of Material and Lubricant Fluctuations," Lubricants, vol. 11, art. no. 193, Apr. 2023.

DOI: 10.3390/lubricants11050193

Google Scholar

[10] C. Rasmussen and C. Williams, "Gaussian Processes for Machine Learning," MIT Press, 2006.

Google Scholar

[11] T. Chen and C. Guestrin, "XGBoost: A Scalable Tree Boosting System," in Proceedings of the 22nd ACM SIGKDD Int. Conf. Knowl. Discov. Data Min. (KDD '16), 2016, p.785–794.

DOI: 10.1145/2939672.2939785

Google Scholar