urbanFlowGen - Generating urban flows fields

urbanFlowGen is a framework for the automated generation of diverse urban flow configurations. It enables parametrized creation of synthetic and real-city building layouts and automates mesh generation and simulation setup, supporting scalable high-resolution CFD studies and high-throughput dataset generation.

[GIT]

urbanPathGen - Drag-efficient UAV path planning 

urbanPathGen is a data-driven framework for energy-aware UAV path planning in previously unseen urban environments. A GNN, trained on flow fields from lattice-Boltzmann simulations with diverse building configurations and inflow conditions, enables rapid flow prediction and efficient route planning.

[GIT] [PAPER]

UrbanFlow-3K: A Dataset of

 3,000 2D LB Simulations

UrbanFlow-3K is an open dataset of 3,000 2D urban flow simulations generated with the lattice-Boltzmann method across three Reynolds numbers. It provides a diverse, publicly available benchmark for developing and evaluating ML models for urban flow prediction before scaling to computationally expensive three-dimensional simulations.

[DATA] [PREPRINT]

Planning of urban wind 

turbines and PV systems

With an increasing global population, urban planning becomes more and more important for a habitable world.

This research focuses on a key challenge in modern urban planning, i.e., net zero cities. The main objective is to develop a tool that assists city planners in finding the best combination and locations of urban wind turbines and PV systems.