Sabry, I., salaheldeen, T., elzathry, N., hewidy, A. (2025). ARTIFICIAL NEURAL NETWORK PREDICTION OF CORROSION RATE IN FRICTION STIR-WELDED ALUMINUM FLANGE. International Journal of Materials Technology and Innovation, 5(1), 22-30. doi: 10.21608/ijmti.2025.337737.1116
Ibrahim Sabry; tarek salaheldeen; noah elzathry; ahmed hewidy. "ARTIFICIAL NEURAL NETWORK PREDICTION OF CORROSION RATE IN FRICTION STIR-WELDED ALUMINUM FLANGE". International Journal of Materials Technology and Innovation, 5, 1, 2025, 22-30. doi: 10.21608/ijmti.2025.337737.1116
Sabry, I., salaheldeen, T., elzathry, N., hewidy, A. (2025). 'ARTIFICIAL NEURAL NETWORK PREDICTION OF CORROSION RATE IN FRICTION STIR-WELDED ALUMINUM FLANGE', International Journal of Materials Technology and Innovation, 5(1), pp. 22-30. doi: 10.21608/ijmti.2025.337737.1116
Sabry, I., salaheldeen, T., elzathry, N., hewidy, A. ARTIFICIAL NEURAL NETWORK PREDICTION OF CORROSION RATE IN FRICTION STIR-WELDED ALUMINUM FLANGE. International Journal of Materials Technology and Innovation, 2025; 5(1): 22-30. doi: 10.21608/ijmti.2025.337737.1116
ARTIFICIAL NEURAL NETWORK PREDICTION OF CORROSION RATE IN FRICTION STIR-WELDED ALUMINUM FLANGE
2Department of Mechanical Engineering, Benha Faculty of Engineering, Benha University, Benha, Egypt
Abstract
This article outlines a methodical procedure for enhancing the aluminum alloy's friction stir welding process variables. Friction stir welding (FSW) is frequently used for challenging welding connections for aluminum alloys. Welding input parameters primarily determine weld quality. The rate of joint corrosion is greatly influenced by welding factors such as tool shoulder diameter, tool rotating speed, and welding speed. The aluminum alloy 6082 by FSW has been attempted to be joined in the current work utilizing a standard milling machine. FSW has been done on 6082 aluminum alloy pipes and plates that are 10 mm thick, 53 mm outer diameter, and have a 4 mm wall thickness. ANN has been created to predict the corrosion rate in FSW based on backpropagation (BP) of error. The tool diameter of the shoulder, speed of tool rotational, and speed of welding are the model's input parameters; the model's output is the joint corrosion rate. Following that, the ANN was trained in utilizing experimental data. Utilized data of experimental not utilized through training, the ANN is tested. The findings indicated that the constructed neural network could be used as a different method for determining corrosion rate for specified process parameters because the ANN's results perfectly harmonize with the experimental.
View on SCiNiTO
Top