Machine Learning-Based Prediction and Classification of Hygrothermal Degradation in Fibre-Reinforced Composites across Multiple Architectures and Ageing Protocols

Aug 4, 2026

This study investigates prediction of water uptake and mechanical performance of fibre-reinforced composites under hygrothermal aging using machine learning. Eight models are used for water absorption, while six datasets support mechanical property prediction across varied aging and manufacturing conditions. Fiber orientation, pattern number, curing, and aging temperature are identified as key drivers of model accuracy. Gaussian process regression, SVR, XGBoost, and BRT reach R2 above 0.95 with relative errors under 5%. Random forest and ANN also capture tensile, compressive, hoop, deflection and load–displacement behavior, enabling durability assessment, though generalization is limited by dataset size and diversity.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages77

Reading Time3 h 14 min

AccessFree to download

Abstract

Acknowledgements

Symbols and abbreviations

Declarations

1 Introduction

2 Objectives of the research

2.1 Objectives

2.2 Research problem and research question

2.3 Research methods

2.4 Scope

2.5 Research gap

3 Literature review

3.1 Fibre-Reinforced Composites and Hygrothermal Ageing

Machine Learning-Based Prediction and Classification of Hygrothermal Degradation in Fibre-Reinforced Composites across Multiple Architectures and Ageing Protocols Lappeenranta–Lahti University of Technology LUT Master’s Programme in Mechanical Engineering, Master’s thesis 2025 Zhicheng Huang Examiner(s): Assistant Professor Humberto Almeida Jr, D.Sc. (Tech.) Assistant Professor Guilherme Gomes, D.Sc. (Tech.) ABSTRACT Lappeenranta–Lahti University of Technology LUTLUT School of Energy Systems Mechanical Engineering Zhicheng Huang Machine Learning-Based Prediction and Classification of Hygrothermal Degradation in Fibre-Reinforced Composites across Multiple Architectures and Ageing Protocols Master’s thesis 2025 77 pages, 41 figures, 6 tables Examiner(s): Assistant Professor Humberto Almeida Jr, D.Sc. (Tech.) and Assistant Professor Guilherme Gomes, D.Sc. (Tech.) Keywords: Machine learning, Composite material, Hydrothermal aging, Filament winding pattern This study investigated the prediction of water uptake and mechanical performance of fibrereinforced composites under hygrothermal aging using machine learning. Eight models (ANN, BPNN, SVR, RF, RT, BRT, XGBoost, and GPR) were applied for water absorption prediction, while six datasets were used for mechanical property prediction under varying aging and manufacturing conditions. Key factors such as fiber orientation, pattern number, curing, and aging temperature strongly influenced the results. Models like GPR, SVR, XGBoost, and BRT achieved R2 > 0.95 and relative errors below 5%, outperforming classical Fick diffusion models. RF and ANN accurately predicted tensile, compressive, and hoop strength, deflection, and load–displacement behavior. These findings demonstrate that machine learning effectively captures complex relationships in composites under hygrothermal aging, providing a reliable tool for material design and durability assessment, though generalization is limited by the size and diversity of the datasets. ACKNOWLEDGEMENTS At first, I would like to express my sincere gratitude to my supervisor, Assistant Professor Humberto Almeida Jr., for his invaluable guidance and insightful advice throughout my thesis work. His profound knowledge, patient guidance, and thoughtful advice not only greatly enhanced my understanding of the discipline but also helped me cultivate critical thinking and research skills. Throughout the research process, he consistently provided constructive feedback, helping me stay focused and motivated. His passion for research and dedication to academic excellence deeply inspired me and set an example for my future academic and professional development. I am profoundly grateful for the opportunity to work under his guidance and for the trust he placed in me throughout this journey. At the same time, I would like to express my deepest gratitude to my parents, Huang Feng and Chen Chunhong, for their selfless and unconditional love. Although we are separated by thousands of miles, their unwavering support and encouragement during times of disappointment, uncertainty, and confusion have been the driving force that keeps me moving forward on this journey. Finally, I would like to thank my friends—Yuxuan Guo, Zihan Ren, Jinwei Lu, Minoo Yadi, and Md Kabir—for their support and companionship. Their kindness, encouragement, and help in daily life have made this journey more manageable and truly enjoyable. I am lucky to have shared this experience with them. Zhicheng Huang 07.10.2025 Lappeenranta Finland SYMBOLS AND ABBREVIATIONS Abbreviations ANN BPNN SVR RF RT BRT XGBoost GPR CFRP GFRP MAE RMSE R2 PC FC FRCs FW ML CNN LSTM Artificial Neural Network Backpropagation Neural Network Support Vector Regression Random Forest Regression Tree Boosted Regression Tree Extreme Gradient Boosting Gaussian Process Regression Carbon Fiber Reinforced Polymer Glass Fiber Reinforc

FAQs

What does the thesis predict for fibre-reinforced composites under hygrothermal aging?

It predicts water uptake (water absorption) and mechanical performance, including tensile, compressive, hoop strength, deflection, and load–displacement behavior.

Which machine learning models are used for water absorption prediction and mechanical property prediction?

Water absorption uses eight models: ANN, BPNN, SVR, RF, RT, BRT, XGBoost, and GPR. Mechanical properties are predicted using six datasets under varying aging and manufacturing conditions.

What factors most strongly influence the prediction results?

Fiber orientation, pattern number, curing conditions, and aging temperature are highlighted as key factors affecting the results.

Aug 4, 2026

Upload by Theodora