Thermal Simulation of TIG Welded Ti-6Al-4V Plates using Ansys and Modelling using ML Techniques

Titanium weldments are extensively used across industries due to their exceptional physical and mechanical properties, particularly in aerospace, medical, and industrial applications. However, challenges such as the Heat Affected Zone (HAZ), distortions, and alterations in mechanical properties and microstructures can result from improper welding processes or incorrect process parameters. These issues can lead to product failure, increased costs, longer timelines, and, in critical cases, safety risks.

To mitigate these challenges, it is essential to determine the optimal welding process and parameters to ensure high-quality titanium weldments. Understanding the impact of each process parameter on the final product quality is crucial for this optimization. By accurately simulating the temperature and heat flux distributions during the welding process, we can predict the formation of different zones within the weldment, such as the fusion zone, HAZ, and unaffected zone, using phase diagrams of Ti-6Al-4V alloy.

In this project, Thermal Simulation of TIG Welded Ti-6Al-4V Plates using Ansys and Modelling using ML Techniques, we use simulation tools to analyze the welded Ti-6Al-4V plates, focusing on temperature and heat flux distributions—parameters that are difficult to measure experimentally. The Goldak double ellipsoidal moving heat source model is applied to simulate heat distribution accurately, incorporating essential factors like thermo-mechanical properties, mesh quality, and heat source movement. The simulation results are validated against experimental data.

Additionally, the ability to generate large datasets through simulations offers a more efficient and cost-effective alternative to traditional experimental methods. These data are then utilized to train various machine learning (ML) models and algorithms to identify the most accurate ML technique for predicting temperature and heat flux distributions. The ML models are also used to determine the most significant process parameters influencing these distributions, providing valuable insights for process optimization.