Why Wait? Near Instant, AI-Powered Wind Estimates are Transforming Early Design

Session: Track B: Technology and the Next Generation Façades and Innovation Award

Jan BergstromProject image

Jan Bergstrom
General Manager, Orbital Stack

The design cycles for buildings are becoming shorter while performance requirements are getting more complex. City planners, developers and citizens are realizing the importance of exceptional urban spaces to achieve vibrant and livable cities. Rising construction costs have increased the focus on efficient and cost-effective design, and design complexity and higher urban density is undermining conventional intuition and rules of thumb.

These are all heavily influenced by wind and the climate. To optimize the multitude of potentially opposing objectives, designers need to integrate more quantitative information into the early design process. Historically this meant significant investments in both time and money, rendering them impractical in those dynamic, early stages. Recent improvements in computation methods have made CFD modeling, for example, quicker, but not quick enough for many.

Artificial intelligence (AI) tools and services are creeping into all aspects of our lives, from image recognition on cell phones and voice assistants in our homes, to automated traffic control systems in our cities. Innovations in 3D, and geometric focused machine learning have opened up the opportunity to apply these tools to wind analysis as well. Applying historical wind tunnel data to train these frameworks, it is now possible to predict wind speeds and pressures in a matter of seconds. This presentation will explore how these tools are developed, why they have the potential to be so transformational, and how they are already being applied to various aspects of microclimate and wind engineering analysis.