Building energy contributes approximately 40% of U.S. greenhouse gas emissions and 75% of emissions in some urban areas. Evaluating modifications to existing building stocks is essential to a proper assessment of GHG reduction policy at various levels. With deeper penetration of intermittent renewable energy resources, supply and demand effects at a high resolution (e.g. hourly) will become more important as variations in grid emissions will become more significant. City-level hourly electricity load data is available; however, effects of building stock changes on usage profiles are not easily analyzed, and on-site fossil fuel usage — the dominant loads in many urban areas — are generally only available annually. Building energy models allow for detailed simulation of building systems, but existing building models must be calibrated to actual energy usage to predict the effects of energy conservation measures.
Reference building models developed by the U.S. Department of Energy for the EnergyPlus software tool were used as the basis for a set of calibrated building energy models to perform community-scale energy conservation measures on the dominant building classes in NYC (i.e. residential and office buildings). A statistical analysis of zip code-level annual electricity and fuel usage data was performed to determine electricity, space heating fuel and domestic hot water (DHW) fuel usage intensities (EUIs) for three broad building categories encompassing these building types in New York City. Several parameters were adjusted for each model until simulations produced the EUIs from the statistical analysis: Thermal envelope characteristics, peak electric equipment and lighting loads, DHW flow requirements, cooling equipment coefficient of performance and heating equipment efficiency. Cooling energy demands were adjusted based on the electricity demand vs. temperature behavior during the cooling season. The hourly daily usage schedules of internal electric and lighting loads were then adjusted for all models, targeting the actual hourly electricity demands for NYC. Because hourly changes affect annual EUIs, the calibrations were performed iteratively until the model outputs, weighted by each building type’s total NYC square footage, equaled the annual EUIs for each building type and the hourly electricity demand data.
This paper shows that this comprehensive calibration approach can achieve root-mean-square deviation (RMSD) of 7% from the average annual electricity demand for these building types, compared to a 31% RMSD for an approach using annual energy calibration only.