EDK All work

Research · Princeton University

Design Wind CRIP

What good is a climate model if the people designing buildings cannot actually use it?

Read the IEEE publication

In brief

Building wind standards are still largely based on historical data and the assumption that extreme-wind risk stays constant over time. Researchers had already developed non-stationary models for a changing climate, but the methods lived in papers and code—not in a form most engineers, policymakers, or communities could readily use. I built Design Wind CRIP to close that gap: a county-level interface that turns climate-risk research into specific design wind speeds. The work was presented at an IEEE conference and published through IEEE Xplore.

Wind-test the page

Still · 0 mph. The page is operating under stationary assumptions.

Design Wind CRIP interface with building inputs and a county-level wind speed map
01The interface converts model inputs—risk category, build year, lifespan, and calculation method—into county-level design guidance.
Red county maps comparing ASCE 7-16 wind speeds with non-stationary calculations
02Difference from ASCE 7-16. The largest gaps concentrate along hurricane-exposed coasts and grow for longer-lived structures.
Blue county maps comparing non-stationary and stationary wind speed calculations
03Non-stationary versus stationary calculations, isolating the effect of a changing climate rather than differences between model families.

What I did

  1. 01Translated probabilistic research using synthetic tropical-cyclone data and non-stationary climate projections into a usable calculation workflow.
  2. 02Built the front end in HTML, CSS, and JavaScript and a Flask/Python back end that executed R models on demand.
  3. 03Let users vary building risk category, construction year, lifespan, and calculation method, then compare stationary and non-stationary results against ASCE guidance.

Key finding

For extended-lifespan structures, the non-stationary LEP-AEP estimates were 37.8% higher than ASCE guidance on average—and as much as 116% higher in some coastal counties.

Recognition

  • Presented at the IEEE MIT Undergraduate Research Technology Conference
  • Published through IEEE Xplore