Determination of household energy using ‘fingerprints’ from energy billing data

Electric and gas utilities (in the U.S.A.) bill their customers on a regular basis, usually monthly or bimonthly. These data provide a truly valuable information resource for energy conservation programme analysts and evaluators. This paper discusses ways to analyse such billing data. The starting point is the Princeton University score‐keeping model, which permits decomposition of total household energy use into its weather‐and non‐weather‐sensitive elements; the weather‐sensitive portion is assumed to be proportional to heating degree days. The score‐keeping model also allows one to compute weather‐adjusted energy consumption for each household based on its billing data and model parameters; this is the model's estimate of annual consumption under long‐run weather conditions. The methods discussed here extend the score‐keeping results to identify additional characteristics of household energy use. The methods classify households in terms of the intensity with which the particular fuel is used for space heating (primary heating fuel vs. supplemental heating fuel vs. no heating at all with the fuel). In addition, households that use the particular fuel for air conditioning are identified. In essence, the billing data and model results define household energy use ‘fingerprints’. The billing data and model results can also be used to identify and correct anomalous bills. Finally, the methods permit careful examination and analysis of changes in energy use from one year to another. They help explain why some households show anomalously large energy savings (e.g. they began using wood as a heating fuel during the second year) or negative energy savings (e.g. very high air conditioning energy use during the second year).