Bearing Response Prediction in Hydrothermal Aged Carbon Fiber Reinforced Epoxy Composite Joints using Machine Learning Techniques

Articles in Press

Document Type : Special Issue: Mechanics of Advanced Fiber Reinforced Composite Structures

Authors

1 Village Raogarh, P.O. Jyotisar,

2 Faculty of Geoengineering, Mining and Geology, Wroclaw University of Science and Technology, Na Grobli 15, 50-421 Wroclaw, Poland

3 Department of Electronics and Communication Engineering, Chandigarh University, Mohali, Punjab, India, 140301

Abstract

The work focuses on predicting the bearing response in hydrothermal-aged carbon fiber reinforced epoxy composite (CFREC) joints through the utilization of machine learning technique. CFREC are extensively employed in aerospace and other high-performance applications, and their long-term structural integrity is of paramount importance. The hydrothermal aging process can significantly affect the mechanical behavior of such composites, particularly in joint configurations. In this research, an innovative support vector regression approach is present that leverages machine learning algorithms to forecast the bearing response of CFREC joints after undergoing hydrothermal aging. The study encompasses the development of predictive models using a comprehensive dataset of experimental observations. The machine learning technique, support vector regression is trained and evaluated to assess their accuracy and reliability in predicting bearing response. The results shows that the overall percent reduction in bearing response, after 30 days pristine composite bolted joints at 0 Nm bolt torque show reductions of 23.22 % at 65°C, respectively. Conversely, under same conditons, MWCNTs added composite bolted joints exhibit only 9.2% reduction. The predictive models finds the value of 0.0081 RSME and 0.8 R2 respectively through support vector regression confirming that the predicted values lie in between the upper and lower bond.

Keywords

Main Subjects

Mechanics of Advanced Composite Structures

Articles in Press, Accepted Manuscript
Available Online from 15 August 2024

Files

Share

How to cite

Statistics