Utilizing a combination of experimental and machine learning methods to predict and correlate between accelerated and natural aging of CFRP/AL adhesive joints under hygrothermal conditions

Abstract This study investigates how carbon fiber reinforced polymer (CFRP)‐to‐aluminum adhesive joints behave under accelerated aging conditions with hygrothermal exposure and compares these findings against naturally aged samples to evaluate material reliability in challenging environments. The CFRP‐to‐aluminum adhesive joints were manufactured and subjected to natural aging for durations ranging from 1 to 3 years with 6‐month intervals, as well as accelerated aging (hygrothermal) for periods ranging from 100 to 1200 h, with intervals of 50 h. Subsequently, the mechanical properties of the joints were evaluated using a three‐point bending test. To forecast natural aging times from accelerated aging data, five machine learning models were utilized: artificial neural network, support vector regression, linear regression, polynomial regression, and random forest regression. Hygrothermal aging significantly degraded the matrix, causing a shift in failure modes from cohesive to mixed types (cohesive, adhesive, and fiber tear failures), leading to a notable decline in bending strength. The study observed a 23.13% strength reduction in samples aged naturally for 3 years and a 24.33% decrease in those subjected to 1000 h of accelerated aging. The random forest regressor demonstrated superior accuracy in predicting natural aging times across different accelerated aging periods. Through the application of machine learning models, this study introduces a novel approach to forecast natural aging durations using data from accelerated aging experiments. This method shows potential for optimizing joints and composite structures, ultimately improving their durability and minimizing the likelihood of failures during operational use. Highlights Studied hygrothermal effects on accelerated aging of carbon fiber reinforced polymer/Aluminum (AL) adhesive joints. Noted strength reduction from hygrothermal aging. Used five machine learning models; random forest regression had the highest accuracy. Analyzed correlation between natural and accelerated aging of dissimilar adhesive joints.