Actions for Weather normalization of electricity sales. Final report

Email RIS file
Bookmark
Report an Issue

Weather normalization of electricity sales. Final report

Author
Train, K.
Published
United States : [publisher not identified], 1983
Springfield, Va.: National Technical Information Service, [approximately 1983]
Physical Description
microfiche : negative ; 11 x 15 cm
Additional Creators
Engle, R., Granger, C., Ignelzi, P., Ramanathan, R., and Rice, J.

Availability

Browse Nearby on Shelf
I Want It
I Want It

Finding items...

Summary
Methods for quantifying the relation between electricity sales and weather are investigated, and techniques that utilities can use to adjust (or normalize) their sales for abnormalities in weather are developed. Based on utilities' comments in the early phase of the project, three major issues, and numerous minor issues, are addressed: (1) accurate matching of daily weather data to sales, (2) the nonlinear response of sales to weather, and (3) the lack of data on factors conditioning weather sensitivity. An appropriate method for matching weather data and sales is specified, and it is shown that models incorporating this matching are generally very accurate for weather normalization. A nonparametric method is developed for identifying the nonlinear response of sales to weather; this method is an advance over standard methods in that it does not require a priori information or trial-and-error experimentation. Finally, a test is developed to determine whether weather sensitivity has changed significantly over time due to changes in unobserved conditioning factors, and models with time-varying parameters are developed to account for any significant changes without requiring data on the conditioning factors.
Report Numbers
DE83902539; EPRI-EA-3143
Other Subject(s)
Collection
NTIS collection.
Note
OSTI Identifier 5928502
Research organization: Cambridge Systematics, Inc., Berkeley, CA (USA).
Research organization: Quantitative Economic Research, Inc., San Diego, CA (USA).
View MARC record | catkey: 47335822