A machine learning-driven approach to predict mechanical degradation associated with matrix cracks in fiber-reinforced composite laminates

Aug 5, 2026

Matrix cracking is an early, critical damage mechanism in fiber-reinforced plastic laminates, strongly impacting stiffness and residual strain. This study develops machine learning models to predict stiffness reduction and residual strain for carbon fiber- and glass fiber-reinforced composite laminates with off-axis plies. Models are trained under limited experimental data and many input features. Random Forest and LightGBM use Bayesian hyperparameter tuning and RFECV feature reduction, achieving high cross-validation accuracy (up to 0.83 for stiffness and 0.88 for residual strain) and robust generalization to unseen virgin and cracked datasets.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages15

Reading Time38 min

AccessFree to download

Introduction

Background on FRP applications and laminate complexity

Matrix cracking as an early damage mode

Indicators of degradation: stiffness reduction and residual strain

Next Materials 9 (2025) 101209 Contents lists available at ScienceDirect Next Materials journal [homepage:](homepage: www.sciencedirect.com/journal/next-materials)[ www.sciencedirect.com/journal/next-materials](homepage: www.sciencedirect.com/journal/next-materials) | Research article
A machine learning-driven approach to predict mechanical degradation associated with matrix cracks in fiber-reinforced composite laminates
M.J. Mohammad Fikrya,* , Jason P. Mackb, Faizan Mirza b , Niken Prasasti Martonoc , K.T. Tan b , Vladimir Vinogradov d, Shinji Ogihara a
a Department of Mechanical and Aerospace Engineering, Faculty of Science and Technology, Tokyo University of Science, Chiba, Japan b Department of Mechanical Engineering, College of Engineering and Polymer Science, The University of Akron, OH, United States c Department of Industrial and Systems Engineering, Faculty of Science and Technology, Tokyo University of Science, Japan
d School of Engineering, Newcastle University, Newcastle upon Tyne, United Kingdom | | | | | --- | --- | --- | --- | | A R T I C L E I N F O | | A B S T R A C T | | | Keywords:
Matrix cracking Stiffness reduction Residual strain
Data-driven approach Random Forest regression LightGBM regression Composite laminate | | Matrix cracking is an early, critical damage mechanism in fiber-reinforced plastic laminates, significantly affecting the mechanical properties such as stiffness and residual strain. This study aims to develop machine learning models capable of predicting stiffness reduction and residual strain in carbon fiber-reinforced plastic and glass fiber-reinforced plastic laminates containing off-axis plies. This work focuses on modeling under conditions with limited experimental data and a high number of input features. Two tree-based machine learning models, namely Random Forest and Light Gradient Boosting Machine (LightGBM), were trained using experimental datasets derived from laminates with various stacking configurations and material systems. Input features included material properties, laminate configurations, applied loading conditions, and crack density. Bayesian optimization was used for hyperparameter tuning, and Recursive Feature Elimination with Cross-Validation (RFECV) was applied to reduce the number of input features while preserving model performance. Both models achieved high predictive accuracy in cross-validation and performance evaluations, with mean crossvalidation R² values of up to 0.83 for stiffness reduction and 0.88 for residual strain. LightGBM performed well with the full feature set, whereas Random Forest benefited significantly from feature selection, leading to improved generalization. Validation using previously unseen experimental data confirmed that both models accurately predicted the mechanical behavior of virgin and cracked laminates. Overall, the results suggest that the combination of Random Forest and RFECV is especially effective when working with small datasets and highdimensional inputs. | | 1. Introduction Fiber-reinforced plastics (FRP) have gained significant attention in industries such as aerospace, automotive, wind energy, marine engineering, and construction owing to their high strength-to-weight ratio, durability, and corrosion resistance [1,2]. Particularly in aerospace applications, the relatively low weight of FRPs contributes to improved fuel efficiency and increased payload capacity [3]. Although quasi-isotropic laminates are widely used in practical applications because of their balanced mechanical properties, studies have often focused on simpler stacking configurations such as cross-ply and angle-ply laminates [4,5]. This is because the mechanical behavior and damage mechanisms in quasi-isotropic structures are highly complex, making it difficult to isolate and understand specific degradation processes [6]. Simpler configurations allow for clearer investigations of how fiber orientation and stacking seq

FAQs

What degradation effects does the study aim to predict for composite laminates?

The study predicts stiffness reduction and residual strain associated with matrix cracking in carbon and glass fiber-reinforced composite laminates with off-axis plies.

Which machine learning models are used in the work?

Two tree-based models are used: Random Forest regression and LightGBM (Light Gradient Boosting Machine).

How does the approach handle limited experimental data and many input features?

Bayesian optimization is used for hyperparameter tuning, and Recursive Feature Elimination with Cross-Validation (RFECV) reduces the number of input features while maintaining predictive performance.

Aug 5, 2026

Upload by Seraphina