Metric Based Mesh Adaptation Using Ensemble Learning
Please login to view abstract download link
Mesh adaptation is vital in CFD for refining grids to capture flow features accurately. Metric-field-based mesh adaptation offers a mathematically robust method for mesh refinement to capture characteristics of fluid flows. However, the algorithm relies on computationally expensive adjoint solutions for error estimation. This study aims to build upon the research by Ghosh for the prediction of adjoint weighted error estimators using sequential ensemble models. The results highlight the potential of machine learning to bypass adjoint computations, enabling efficient simulations for complex flows.