Data-driven prediction and optimization of mechanical performance in multi-material extrusion additive manufacturing of laminated polymer composites: a hybrid response surface method and machine learning approach
e-Polymers · 23.01.2026 · 26 (1)
DOI: 10.1515/epoly-2026-0041 ↗Kurz gefasst
Fachartikel in der Zeitschrift e-Polymers, online erschienen am 11.11.2026. Verfasst von Imran Khan und 2 weiteren Autor:innen. Die Arbeit nennt 68 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Qatar National Library.
Abstract (Originalsprache)
Laminated polymer composites offer enhanced properties for various applications. Polylactic acid/acrylonitrile butadiene styrene (PLA/ABS) laminates were produced by fused filament fabrication (FFF). The effects of infill density (ID), layer thickness (LT), print speed (PS), and infill pattern (IP) were examined for compressive modulus (CM), compressive strength at 10 % strain (CS10), flexural modulus (FM), and flexural strength at 10 % strain (FS10). Response surface methodology (RSM) and a face-centered central composite design were used for statistical modeling and multi-response optimization. Analysis of variance identified ID as the dominant factor. Six machine learning models were compared. Polynomial ridge regression gave the best held-out results for CM ( R 2 = 0.858), CS10 ( R 2 = 0.978), and FS10 ( R 2 = 0.978), whereas Gaussian process regression gave the most consistent predictions across the additional experimental assessment. Optimization identified 90 % ID, 0.20 mm LT, 80 mm/s PS, and gyroid IP, predicting CM = 1,237.9 MPa, CS10 = 29.04 MPa, FM = 337.97 MPa, and FS10 = 64.77 MPa. The combined RSM-ML framework provides an interpretable and predictive basis for selecting PLA/ABS FFF conditions that balance compressive and flexural performance.
Förderung
- Qatar National Library
Zitieren
APA
Khan, I., Rashid, A. A., & Koç, M. (2026). Data-driven prediction and optimization of mechanical performance in multi-material extrusion additive manufacturing of laminated polymer composites: a hybrid response surface method and machine learning approach. e-Polymers, 26(1). https://doi.org/10.1515/epoly-2026-0041
Vancouver
Khan I, Rashid AA, Koç M. Data-driven prediction and optimization of mechanical performance in multi-material extrusion additive manufacturing of laminated polymer composites: a hybrid response surface method and machine learning approach. e-Polymers. 2026;26(1). doi:10.1515/epoly-2026-0041
BibTeX
@article{Khan2026,
author = {Khan, Imran and Rashid, Ans Al and Koç, Muammer},
title = {Data-driven prediction and optimization of mechanical performance in multi-material extrusion additive manufacturing of laminated polymer composites: a hybrid response surface method and machine learning approach},
journal = {e-Polymers},
year = {2026},
volume = {26},
number = {1},
doi = {10.1515/epoly-2026-0041}
}
EUVIMED