It could one day be possible to create an energy model which includes every building in the country. All you need is a supercomputer, some time, and the Department of Energy's Oak Ridge National Laboratory's (ORNL) recipe for extracting high-level building data from public sources. To test the concept, ORNL has created an energy model for the buildings in Chattanooga, all 178,368 of them.
The Electric Power Board (EPB) of Chattanooga operates one of the nation's most advanced electric grids, according to ORNL, and gathers utility data every 15 minutes for all the facilities in its territory. EPB was interested to find out ways to reduce peak demand and ways to better balance the power grid.
Using the recently decommissioned Cray XK7 Titan supercomputer, ORNL crunched data such as floor area and building orientation and ran multiple simulations to find the building models which most closely matched actual energy use data. They were then able to evaluate the city's building stock through more than 2 million simulations, varying parameters such as infiltration, lighting, or insulation levels.
From the city-wide energy model, researchers could see that strategies such as switching to natural gas, sealing buildings properly, bringing insulation levels up to code, and using smart thermostats to preheat or precool homes could help manage peak demand, potentially saving the utility $11 to $35 million per year, according to ORNL, and potentially avoiding the need for new centralized power generation facilities.
ORNL intends to make the models openly available.
Naomi Millán is senior editor of Building Operating Management.
