DACNN based predicting tensile strength of friction stir welded aluminium alloy joints

Aluminum alloys are difficult to weld using traditional fusion welding methods. Friction stir welding (FSW) offers a potential alternative to typical fusion welding methods for generating high-quality aluminum connections. Welding process factors significantly impact joint quality. This manuscript proposes a Double Attention Convolutional Neural Network (DACNN) for Predicting the Tensile Strength of Friction Stir Welded AA6061 Aluminium Alloy Joints. The primary objective is to predict the ultimate tensile strength (UTS) of AA6061-T6 aluminum alloy joints based on their mechanical properties and process parameters. The proposed DACNN method is used to predict the aluminum alloy's tensile strength. By then, the proposed technique is implemented in the MATLAB platform and the execution is computed with the existing method. The proposed technique displays superior outcomes in all existing systems as, the Marine Predators Algorithm with Random Vector Functional Link (MPA-RVFL), Artificial Neural Network Henry Gas Solubility Optimization (ANN-HGSO) and Random Vector Functional Link Based Hunger Games Search (RVFL-HGS). The existing methods demonstrate efficiencies of 85%, 75%, and 65%, respectively, while the proposed method achieves an efficiency of 95%. Regarding error rates, the existing techniques exhibit 2%, 3%, and 4%, whereas the proposed method shows a reduced error of 1%. These outcomes indicate that the proposed method outperforms the existing techniques with higher efficiency and lower error.