Experimental Study of Single-Lap Adhesive Joints to Analyze and Predict the Tensile Strength Values of Aluminum Alloy 6061 Substrates using Artificial Neural Networks

Abstract

Adhesive bonding is one of the essential methods applied in wide fields, mainly automotive and aerospace, because the adhesive can be used with various materials, weighs less compared to other methods, is easy to work with, and does not require many tools. The present research focuses on determining and predicting the ultimate tensile values for single-lap adhesive joints. The mathematical models and artificial neural network (ANN) method predict the tensile strength values. Two variables were used: the surface roughness and the bonding area. To determine tensile test values, ten samples were used with different surface roughness and an overlap distance of 25 and 40 mm. The results showed that the bonding distance had more effect than the surface roughness on the ultimate tensile load. Also, the predicted error values through mathematical models did not exceed 3.209% for the samples, while the ANN samples' error values did not exceed 8.312.

Keywords

• Adhesive bonding
• Surface roughness
• Overlap distance
• Adhesive strength
• Artificial neural network
• Prediction values