Analytical Approach for Forecasting the Load Capacity of the EN AW-7075-T6 Aluminum Alloy Joints Created Using RFSSW...

Area of Science:

• Materials Science
• Mechanical Engineering
• Manufacturing Processes

Background:

• High reliability of aircraft structures necessitates robust welding processes with predictable performance.
• Refill Friction Stir Spot Welding (RFSSW) is critical for joining aluminum alloys in aerospace applications.
• Optimizing RFSSW parameters is essential for achieving high load capacity and minimal variability in weld strength.

Purpose of the Study:

• To identify optimal technological parameters for the RFSSW of EN AW-7075-T6 Alclad aluminum alloy sheets.
• To enhance the accuracy of predicting weld load capacity and process variability.
• To ensure uniform functional properties and high quality in aircraft structural connections.

Main Methods:

• Utilized Statistica 13.3 software for neural network calculations to model RFSSW parameters.
• Compared neural network model predictions against classical models and previous experimental data.
• Evaluated prediction accuracy using maximum error and standard deviation metrics for load capacity.

Main Results:

• Neural networks achieved significantly higher accuracy in predicting load capacity (max error 1.55%, std dev 3.74%) compared to classical models (max error 6.13%, std dev 14.51%).
• The neural model enables determination of parameters ensuring a high probability (P = 0.999935) of achieving required load capacity.
• Classical models are insufficient due to high error and sensitivity to input parameter variations.

Conclusions:

• Neural network modeling provides a superior method for optimizing RFSSW parameters in aerospace applications.
• This approach ensures enhanced reliability and quality control for aircraft structures manufactured using EN AW-7075-T6.
• The study demonstrates the capability of neural models to achieve near-perfect prediction accuracy for critical weld properties.