Skip to content
Open access

Physics-informed neural network framework for predicting drilling-induced delamination in GFRP composites

Sep 2026 · PLoS ONE · Vol 21 · 0 citations · 35 references
Medicine

Abstract

Drilling-induced delamination is one of the most critical defects that affect the structural integrity, dimensional accuracy, and assembly performance of Glass Fiber Reinforced Polymer (GFRP) composites. Accurately predicting delamination is challenging due to the complex nonlinear interactions between machining parameters and interlaminar damage mechanisms. In this study, we developed a Physics-Informed Neural Network (PINN)-based predictive framework to estimate drilling-induced delamination in GFRP laminates by integrating experimental data with fundamental principles of fracture mechanics. A full-factorial experimental design with 243 drilling trials have been conducted by varying drill point angle, drill diameter, laminate thickness, spindle speed, and feed rate. The delamination factor was evaluated experimentally using image-based measurement techniques. The physics governing thrust-force-induced delamination has been incorporated into the neural network’s loss function to ensure consistent and accurate learning. The proposed PINN architecture effectively captured the nonlinear relationship between drilling parameters and delamination behavior, achieving a coefficient of determination (R2) of 0.9769 and a low root mean square error (RMSE) of 0.0048. Sensitivity analysis revealed that the feed rate was the most influential parameter affecting delamination, followed by laminate thickness and drill diameter. Notably, higher spindle speed reduced damage formation due to improved cutting efficiency and decreased thrust force. Further interaction analysis demonstrated significant coupling effects between feed rate and spindle speed, as well as between drill diameter and laminate thickness. The integration of physics-based constraints significantly enhanced the robustness of predictions and the generalization capability compared to conventional data-driven approaches. Our findings confirm that the proposed PINN framework serves as a reliable and physically interpretable tool for predictive modeling, intelligent process optimization, and damage minimization in composite drilling applications.

Read PDF

Similar papers

Open access Sep 2026

Predictive modeling of hybrid-layer friction drill joining of cast aluminum via integrated finite element and deep learning techniques

Numerical results indicate that the presence of an upper sheet modifies heat generation and stress distribution, promoting more uniform material flow and improved bushing geometry under the investigated condition.

S. El-Bahloul · 0 citations
Open access Oct 2026

Universal physics-informed machine learning framework for the prediction of porosity defects in high-power laser beam welding with different metallic materials

High-power laser beam welding (LBW) is a widely used joining technique for metallic materials. However, porosity defects frequently arise during the process, leading to significant degradation of weld properties. Accurate prediction of porosity remains challenging due to the highly nonlinear and material-dependent unde...

Xiang-Meng Meng, M. Bachmann, Fan Yang et al. · 0 citations
Open access Oct 2026

Formation Mechanisms and Suppression Strategies of Delamination Defects in Drilling of Carbon Fiber Composite Materials

The delamination defects in the drilling of carbon fiber composite materials are the key problem restricting the hole-making quality of these materials. Systematically studied on the geometric characteristics, formation mechanism and suppression strategy for delamination defects. In terms of geometric characteristics,...

Qing-Hui Wang, Xin Mi, Bing Xiao et al. · 0 citations
Oct 2026

A Physics‐Informed Machine Learning Model for Predicting Fatigue Delamination Growth With Fiber Bridging‐Defined Confidence Intervals in Composite Laminates

In fatigue delamination growth (FDG), for different fiber bridging intensity and load ratios ( R ‐ratios), the Paris‐type law on FDG rate ( da/dN ) is formulated using strain energy release rate ( SERR ) on the crack tip Δ√ G tip and effective SERR Δ√ G eff as similitude parameters. The resulting da/...

Jie-Xiong Wang, Liao-Jun Yao, Yang Liu et al. · 0 citations
Open access Sep 2026

A Structure-Informed DNN Model for Welding Flux Viscosity Prediction

Viscosity plays a pivotal role in the performance of welding fluxes by directly affecting the weld pool and the resultant weld metal, both of which are largely determined by heat and mass transfer during the welding process. Accurate viscosity prediction is essential for developing high-quality welding fluxes. However,...

Hongyu Liu, Yan-Yun Zhang, Xiaoxi Song et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.