A Bottom-Up Model of Residential Electricity Demand in North and South Carolina

Abstract

Residential electricity is a significant component of total electricity use in the United States, and the residential market is also a key demographic for energy efficiency (EE) and distributed generation growth. Uncertainty in residential load growth is driven by the impact of changes in technology, policy, demographic and life-style changes. Using a bottom-up engineering model, we begin the construction of a tool to facilitate analyzing the effects of these factors. We use data from the EIA’s Residential Electricity Consumption Survey (RECS), in conjunction with EnergyPlus and BEopt, publicly available software from NREL, to construct 22 archetypical residential buildings characterizing North and South Carolina’s housing stock. We then model energy use for these buildings, and extrapolate these results to the larger housing stock. Projections are accurate for a benchmark year using actual weather data. We identify a number of potential improvements to the model and ways in which the uncertainty on future projections of energy use can be bound. Our conclusions follow:

The archetypical model is a reasonable solution for regional scale residential electricity modeling which minimizes computational needs. The model delivers disaggregated energy demand, and hourly demand, estimates which are useful for future analysis of energy policy cost effectiveness. Weather data is a driving source of uncertainty, and hence input weather data should be carefully considered. Projections should use varied weather data to bound uncertainty. Despite being less computationally demanding than other methods, this model would benefit from an automated method of archetype alteration to ease sensitivity analysis. BEopt supports this through python and XML input files.