Reducing Residential Plug-Load Energy Use Through Nonintrusive Submetering and Personalized Feedback

You are here

Lead Performer: Columbia University — New York, NY
Partners:
-- Lucid — Oakland, CA
-- Siemens — Washington, DC
DOE Total Funding: $1,534,397
Cost Share: $398,297
Project Term:  October 1, 2016 – September 30, 2020
Funding Type: Building Energy Efficiency Frontiers and Innovations Technologies (BENEFIT) – 2016 (DE-FOA-0001383)

Project Objective

Essential data from pervasive low-cost submetering with sufficient accuracy for equipment and plug loads is necessary to maximize and verify energy savings, as well as to provide critical information on the state and usage patterns of specific equipment to enable monitoring-based commissioning and facilitate the optimization of fault detection and diagnostics of operational faults along with control strategies and integration with the electric grid. This project will leverage existing nonintrusive submetering in developing a human-in-the-loop approach and investigating occupant feedback strategies to change electricity use by reducing load or shifting usage to non-peak hours.

Appliance-level consumption will be obtained via statistical disaggregation of apartment-level metering in the especially challenging and underserved multifamily residential building sector to eliminate the need for costly and intrusive plug load monitors and reduce electric bills. To target usage feedback to the desired consumption change, the project will utilize natural language processing (NLP) based approaches that automatically generate feedback messages that contain various types of illustrations and text. Statistical analysis will be used on disaggregated appliance-level consumption to determine how much reduction or load-shift in electricity use and in electric bills each type of feedback can achieve, and how this may vary by demographic.

Although the focus will be on multifamily dwellings, the technology can also be extended to single-family homes. A unique multiyear and anonymized dataset of residential electricity use, broken down by demographics and appliance end use, including which user types best respond to what type of feedback message (e.g., monetary saving, energy saving, environmental impacts), will be made publically available to inform the research community, as well as utilities on their investment strategies. 

Contacts

DOE Technology Manager: Erika Gupta
Lead Performer: Patricia Culligan, Columbia University

Publications