betaFeedback

WorksComputer Science

How ready is the United States commercial building stock for artificial intelligence? A national estimate from energy codes and survey data

Abstract

Buildings serve as platforms for artificial intelligence (AI) through the data they generate, but the building stock's readiness as an AI platform has not been assessed. A recent framework measures a building's AI readiness as the share of 76 answer levels, across 31 types of operators' questions, that its recorded data could reach, and has so far been applied only to individual buildings. This study applies it to the 2018 U.S. commercial building stock. It combines 6,436 records of the Commercial Buildings Energy Consumption Survey with ASHRAE 90.1's data requirements, state code adoption, advanced-meter statistics and 45 buildings' knowledge graphs, imputing each building's code edition and data recording. Weighted by floor area, the stock could answer on average 0.145 (95% interval 0.127-0.163) of these answer levels. For 92.0% of floor area its data answer less than half of the questions about current status, and most floor area stays below that mark under every tested assumption about compliance, description and the reference buildings. More than half of what installed equipment could answer is lost in recording, mostly because conventional thermostats store nothing and meters lack interval readings. Readiness rises with size, from 0.073 in buildings of at most 5,000 ft2^2 to 0.173 at 25,000 ft2^2 or more. Only 1.6% of floor area was built under a code edition requiring recording, introduced in 90.1-2013 for large buildings. The study gives the first national baseline of commercial-building AI readiness and a public-data method that can be reapplied to each new survey to track the stock.

Paper details

vlnc:W_zkm3ztgx
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC-BY-NC-ND-4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

Not recorded for this paper

View statement
Looking for related papers…