Modelling of Shear Strength of Single Lap Adhesive Joints using Neural Networks

This paper presents a study on the application of artificial neural networks to predict the shear strength of single lap adhesive joints. The study uses a Matlab environment to model the strength based on two main input parameters: the type of adhesive composition (including different fillers) and temperature. The neural network models trained with Levenberg-Marquardt, Scaled Conjugate Gradient, and Bayesian Regularization algorithms were compared. The study found that the Levenberg-Marquardt algorithm achieved the best prediction accuracy with the least computational complexity. The results indicate that it's possible to accurately predict shear strength in adhesive bonds, a finding that could significantly benefit the design of adhesive bonds.