Effect of Combined Damage (Crack and Delamination) and Thermoelastic Loading on Mechanical Deflections of Adhesively Single-Lap (Similar and Dissimilar) Composite Joints-Experimental and Artificial Neural Network Model

Abstract This research investigates the deflection response of multiple damaged (delamination and crack) adhesively bonded (similar and dissimilar material) single lap joints (SLJs). The glass fiber-reinforced polymer (GFRP) and AA2014-T6 materials are utilized for the joint components to prepare different configurations. Further, the intact, individual damage (delamination, longitudinal/transverse cracks) and multiple damages (delamination with cracks) are also prepared including variable overlapping lengths (25, 30, and 35 mm). Additionally, a DMA test is performed to verify the adhesive's glass transition (Tg) temperature to understand the strength under the influence of combined loadings. A few test specimens of bonded joints are fabricated to evaluate the properties and imported for the numerical analysis. Similarly, the inherent bonding strength (adherend and adhesive) behavior is understood through the field emission scanning electron microscope (FESEM) images. The responses of the damaged joint component are computed numerically via an abaqus model under the combined (thermal and mechanical) loadings. The computed deflections for similar and dissimilar joints are compared with the in-house experimental values for different temperatures (ΔT = 0, 10 °C, and 15 °C). The deviations between the similar and dissimilar joints are in the acceptable range (8.71% and 10.83%). Finally, an in-built artificial neural network (ANN) toolbox module from matlab is utilized to cross-verify the current predictions through the experimentally recorded 432 data samples and 36 bonded joint cases under unknown thermal conditions. The current ANN predictions of thermos-mechanical deflections do not deviate from the experimental data by a minimal amount, i.e., 3.43%.