• US Department of Energy unlocks innovation in smart building products

    Releasing the world’s largest validated dataset on building HVAC operations

A first of its kind resource for building technology development

The DOE and Berkeley lab have partnered across the national laboratory complex and with the research community to curate, validate, and publish the world’s largest set of labeled time-series data representing commercial HVAC systems operating in faulted and fault-free states. 

The data spans:

7
common HVAC systems and configurations
257
faulted or fault-free condition states, over a full year of operation
8B
time series points

Get access to the data today!

Access now

To learn more about fault detection, view our publications.

Publications >

Open data drives value, but is hard to come by in buildings

In today’s world, data is a commodity – the “open data” movement has gained traction as a powerful lever to drive innovation democratizing access to information enhances value across industries.

Buildings are becoming increasingly data rich environments but, in contrast to other domains such as image processing or government operations, open datasets representing commercial building operations are nearly non-existent. Even more rare is ground-truth verified data for building system and equipment operations.

Why HVAC faults are important

Software-based analytics represent one of the fastest growing markets in commercial building technologies, and as data science comes to buildings, the industry has seen an explosion of interest in the development of advanced analytics and controls.

In particular, FDD (fault detection and diagnostics) technology enables average savings of 9% with 2-year paybacks, by using building operational data to identify system or equipment level faults, and isolate their causes.

Putting the HVAC fault data to use

This resource is being used by industry, research, and academia to:

  • Create new analytics and fault diagnostic algorithms
  • Benchmark performance
  • Improve product efficacy and reliability
  • Train tomorrow’s workforce, bring data science to building science