| description abstract | This paper presents a study on the interlaminar shear strength (ILSS) retention of three types of glass fiber–reinforced polymer (GFRP) bars with different surface textures subjected to four types of conditioning environments (alkaline, alkaline, salt, acidic, and water) at two temperature levels (ambient laboratory and high temperature) for 3, 6, and 12 months. The conditioning temperature plays a critical role in reducing the strength of the bars. Scanning electron microscopy revealed the extent of damage to the fibers, resin, interface, and fracture morphologies in the cross sections. The causes of fiber cracking and lower strength upon exposure were validated by point energy-dispersive X-ray spectroscopy analyses, which detected the leaching of silicon from the fiber structure. Prediction models using multiple linear regression (MLR) and artificial neural networks (ANNs) were developed using Matrix Laboratory (MATLAB R2023b) software and compared. The coefficients of determination of the MLR and ANN prediction models were found to be 0.29 and 0.90, respectively, indicating the superiority of machine learning–based models in identifying and accounting for nonlinearities and highlighting their potential application in GFRP bars. Finally, the correlation between the transverse shear strength (TSS) and ILSS of the tested GFRP bars was identified. The ILSS of the bars was found to be approximately 0.26 times the TSS for any given conditioning scenario. | |
