There are approximately 124 million residential and commercial buildings in the U.S. consuming $395 billion/year in energy bills. DOE’s Building Technologies Office (BTO) has the overarching goal to reduce energy use intensity (EUI) 30% by 2030 compared to a 2010 baseline. Despite building energy modeling’s ability to identify Energy Conservation Measures (ECMs) with less than 1-year simple payback and Scout-based savings of 2.4 Quads, modeling suffers from the high transaction costs necessary to create, refine, understand, and effectively utilize a building energy model.

DOE’s Office of Electricity (OE) has established the Grid Modernization Laboratory Consortium (GMLC) with the aim of making the nation’s electricity grid more resilient as a foundational infrastructure supporting the nation’s security, economy, and modern way of life. This resilience is facilitated by achievements in energy storage to handle the challenges of decentralized generation and the mismatch with time of use, and capabilities toward dispatching load to match generation via intelligent buildings and devices.

Buildings consume 73% of the nation’s electricity, approximately 80% during peak generation. BTO has established the Grid-interactive Efficient Buildings (GEB) initiative to help make buildings smarter regarding the amount and timing of energy use as an additional value proposition to energy efficiency. BTO and OE/GMLC have co-funded the “Virtual EPB” project which has created a digital twin of 178,368 buildings, empirically validated the virtual utility’s energy use with 15-minute from every building [in partnership with the Electric Power Board of Chattanooga, TN (EPB)], simulated nine monetization scenarios to quantify energy efficiency impacts on utilities and their ratepayers, and integrated results into EPB’s operational business systems to quantify the sub-hourly, building-specific energy, demand, emissions, and cost impacts.

The Virtual EPB project continues to extend several software capabilities detailing the data sources and algorithms referred to as Automatic Building detection and Energy Model creation (AutoBEM). This suite of software tools mine imagery (satellite, aerial, and street-level), LiDAR, cartographic data, tax assessor’s data, and building code assumptions with algorithms and engineering heuristics for assigning building-specific footprint, height, age, type, occupancy, window-to-wall ratio, and other simulation input parameters. The team’s AutoSIM software was developed on world-class high-performance computing machines and demonstrated as the world’s fastest building energy simulator on OLCF’s Jaguar and Titan systems. Once building energy models are created for an area of interest, AutoSIM uses DOE’s flagship whole-building simulation engine EnergyPlus to simulate energy and demand as a function of building use and weather, along with impact of energy efficient technologies.

With this ALCC project, the researchers will use DOE supercomputers to leverage existing organizational relationships, scalable data sources, and unique algorithms to attempt, for the first time, nation-scale building energy use models. The team is currently working with companies to make the resulting building energy models and analysis free and publicly available to stimulate private sector activity towards more grid-aware energy efficiency alternatives for the built environment.