A Multi-Fidelity Gaussian Process Framework for Predicting Process-Induced Deformations in Composites

A Multi-Fidelity Gaussian Process Framework for Predicting Process-Induced Deformations in Composites

K. JOHNSON, M. SANTORI, E. ZAPPINO, N. ZOBEIRY, M. PETROLO

Abstract. This work presents a multi-fidelity machine learning strategy for efficiently modeling the mechanical response of composite laminates during curing. A Gaussian Process Regression (GPR) framework is adopted to link the laminate stiffness parameters (B_11, D_22) to the stress component σ_xx, combining data from numerical models of different accuracy. High- and low-fidelity datasets are generated using Layer-Wise (LW) and Equivalent Single Layer (ESL) theories derived from the Carrera Unified Formulation (CUF) and finite-element analysis. The proposed approach employs Bayesian Optimization to adaptively select the most informative high-fidelity samples, thereby achieving accurate predictions at reduced computational cost. The model’s predictive validity for stress distributions is demonstrated for L-shaped cross-ply composite laminates.

Keywords
Multi-Fidelity Surrogate Models, Gaussian Process Regression, Bayesian Optimization, Carrera Unified Formulation, Process-Induced Deformations, Composite Design

Published online 7/20/2026, 5 pages
Copyright © 2026 by the author(s)
Published under license by Materials Research Forum LLC., Millersville PA, USA

Citation: K. JOHNSON, M. SANTORI, E. ZAPPINO, N. ZOBEIRY, M. PETROLO, A Multi-Fidelity Gaussian Process Framework for Predicting Process-Induced Deformations in Composites, Materials Research Proceedings, Vol. 69, pp 1220-1224, 2026

DOI: https://doi.org/10.21741/9781644904251-210

The article was published as article 210 of the book CEAS – AIDAA Conference 2025

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 license. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

References
[1] E. Zappino, M. Santori, R. Masia, N. Zobeiry, M. Petrolo. Numerical Analysis of the Impact of Process Parameters on the Residual Stress of a Flat Composite Part. Aerotecnica Missili & Spazio (2025). https://doi.org/10.1007/s42496-024-00231-7
[2] C. Schoenholz, E. Zappino, M. Petrolo, N. Zobeiry. Efficient analysis of composites manufacturing using multi-fidelity simulation and probabilistic machine learning. Composites Part B: Engineering, Vol. 280, 111499 (2024). https://doi.org/10.1016/j.compositesb.2024.111499
[3] E. Zappino, N. Zobeiry, M. Petrolo, R. Vaziri, E. Carrera, A. Poursartip. Analysis of process-induced deformations and residual stresses in curved composite parts considering transverse shear stress and thickness stretching. Composite Structures, Vol. 241 (2020). https://doi.org/10.1016/j.compstruct.2020.112057
[4] E. Carrera, M. Cinefra, M. Petrolo, E. Zappino: Finite Element Analysis of Structures through Unified Formulation. John Wiley & Sons (2014). https://doi.org/10.1002/9781118536643