2024 DAFoam Workshop

This presentation focuses on two key areas of research: optimal sensor placement for turbulence model augmentation and aerodynamic shape optimisation. Using the eigenspace perturbation method, uncertainty maps are generated to strategically place sensors in regions of high uncertainty, enabling efficient data assimilation to improve Reynolds-averaged Navier-Stokes (RANS) predictions. Additionally, the study employs a machine-learning-augmented turbulence model for aerodynamic shape optimisation, achieving significant drag reduction (25–27%) compared to baseline models. The results demonstrate the effectiveness of data-driven techniques in enhancing turbulence models and improving predictive accuracy for complex aerodynamic flows.