Optimization of Rotary Friction Welding Parameters Through AI-Augmented Digital Twin Systems

Area of Science:

• Materials Science and Engineering
• Computational Engineering
• Artificial Intelligence in Manufacturing

Background:

• Rotary Friction Welding (RFW) is crucial for joining dissimilar materials, but precise temperature control is vital for weld integrity.
• Predicting and controlling peak temperatures during RFW of Ti Grade 2/AA 5005 joints is challenging.
• Existing methods lack real-time adaptive control for optimizing RFW parameters.

Purpose of the Study:

• To develop a Digital Twin (DT) of the Rotary Friction Welding (RFW) process using Artificial Neural Networks (ANN).
• To accurately predict peak temperatures during the welding of dissimilar Ti Grade 2/AA 5005 joints.
• To enable real-time optimization and adaptive control of RFW parameters for enhanced joint quality.

Main Methods:

• Development of ANN models to predict peak temperatures based on Finite Element Method (FEM) simulations in ADINA.
• Integration of a Genetic Algorithm (GA) with the DT for optimizing RFW process parameters.
• Real-time (less than 0.1 s) adaptive control of parameters like rotational speed, axial force, and friction time.

Main Results:

• Accurate prediction of peak temperatures within the 20-640 °C range for Ti Grade 2/AA 5005 joints.
• Successful dynamic adjustment of critical process parameters using AI-augmented DT systems.
• Demonstration of real-time optimization leading to improved joint quality and minimized defects.

Conclusions:

• The developed AI-augmented Digital Twin provides an effective tool for analyzing and optimizing RFW parameters.
• Real-time adaptive control significantly enhances weld quality, defect reduction, and process efficiency in RFW.
• This approach offers a pathway to advanced, intelligent manufacturing processes for dissimilar material joining.