Lead Performer: Lawrence Berkeley National Laboratory (LBNL)—Berkeley, CA
-- FirstFuel—Lexington, MA
-- NYSERDA—Albany, NY
-- Signetron—Berkeley, CA
-- TRC Energy Services—Oakland, CA
Project website: https://buildings.lbl.gov/emis/machine-learning
DOE Funding: $500,000 in FY19
Project Term: October 1, 2018—September 30, 2021
Funding Type: Direct Funded
State-of-the-art analytics software and modeling tools can provide valuable insights into efficiency opportunities. However, prior research has shown that key barriers include relatively limited data sources (smart meters and weather being most common in commercial tools), or reliance on user-provided inputs for which default values may be the fallback. There is great opportunity to apply techniques based on multi-stream data fusion and machine learning to overcome these challenges, particularly for those involved in the development and use of efficiency analysis tools, as well as broader research and development in this field.
This project will develop automated approaches to determine building characteristics and efficiency opportunities using unstructured data from a subset of three categories:
- Public data – disclosure and permit records.
- Imagery – RGB (red, green, blue), thermal and LIDAR (light detection and ranging), acquired via satellite or aerial methods (airplanes or drones).
- Meter and weather data.
The first two categories of unstructured data can be used to determine building characteristics: type, geometry, tightness, material, and orientation. Combined with the third category, meter and weather data, these can also define operational measures. Prior research on permit data and analysis of images using machine learning will be expanded, and new methods will be created and validated to compare their efficacy with the current state of the art and industry practice.
This project will leverage advances in machine learning with unstructured data sources to improve the inputs to and performance of state-of-the-art building efficiency tools. This new approach has the potential to improve identification capabilities for efficiency measures and building characteristics in comparison to remote meter analytics tools and in-person audits. Overall, this project will provide more accurate building energy data analytics for the building industry.
DOE project manager: Harry Bergmann
Principal Investigator: Jessica Granderson, LBNL