📢 Excited to share that I will be presenting a poster on May 15 at the Data-Driven Urban Tech Workshop at Cornell Tech in NYC! 📝

I’ll be presenting our work on real-time environmental performance prediction of outdoor thermal comfort in cities. We’ve developed a surrogate model for urban CFD simulations that has been trained on thousands of building massing models. This approach significantly reduces computational costs and increases accessibility in the early design process.

💡 Highlights:

1️⃣ The model is trained using an end-to-end pipeline based on #Eddy3D and implemented within the Rhino and Grasshopper environment.

2️⃣ The model provides real-time simulation feedback within Rhino/Grasshopper software, reduces the risk of user error, and allows for appropriate spatial resolution in early urban design.

3️⃣ We show very promising accuracy on 1300+ geometries and 130k+ training pairs.

🚀 Although our model has limitations in predicting some very complex urban wake regions, it significantly facilitates the incorporation of urban airflow analysis in early design with minimal effort and real-time feedback.

You can find the initial preprint here: A Gan-Based Surrogate Model for Instantaneous Urban Wind Flow Prediction

If this piqued your interest, please come talk to me next week 5/15 01:15-02:45pm! 🏙️

More info, agenda, and lineup

Timur Dogan Environmental Systems Lab, Cornell AAP Cornell University College of Architecture, Art, and Planning

#UrbanTech #UrbanDesign #CFD #GANs #MachineLearning #UrbanMicroclimate #PedestrianComfort #CornellTech