CHALLENGE - MUSICAS, a multistakeholder project, aims to improve the capacity to predict the performance and the fatigue of welded structures. Also, to simulate their non-destructive testing from a fine multiphysics simulation of the welding process of a car. It also aims to demonstrate that the rational integration of existing software in business infrastructure can consider automating processes to correspont to effective methodologies.
SOLUTION - In this work, industrial test cases are used in order to improve the modeling of welding inside the available simulation tools. Input variables are torch angle, power, and mesh, while output parameters are stress and fillet radius. Simulation tools for the welding process include WPROCESS and SYSWELD. The optimization process is performed with modeFRONTIER. Testing of different algorithms is done to find the optimal values of the tuning parameters. CAD/Mesh analysis is done for different kinds of welding. Deployment on HPC Drive.
BENEFITS - The methodology presented shows the successful integration of CAE tools using modeFRONTIER. In this study, optimal values of the tuning parameters were found, different algorithms were tested (pilOPT, MOGA II), and tuning parameters are added and removed to improve the fitting.