Real-time and high-fidelity urban wind simulation

We are developing methodologies to obtain a high-fidelity/detailed wind field over a region without the need of expensive CFD simulations. We combine emerging AI/ML techniques jointly with more traditional CFD and Numerical Weather Prediction techniques (mesoscale simulations) to do this. These methodologies have numerous applications, including obtaining wind statistics in areas with complex orography or urban areas or developing digital twins of wind farms, where these models can provide high-resolution wind predictions in real-time.