246 – Contributed Oral Poster Presentations: Business and Economic Statistics Section
A Multi-step Approach to Modeling the 24-hour Daily Profiles of Electricity Load Using Seasonally-Varying Splines
Abdelmonaem Jornaz
Missouri University of Science & Technology
V. A. Samaranayake
Missouri University of Science & Technology
A multi-step approach to modeling hourly electricity load and estimating the 24-hour daily profile of electricity use is proposed. The methodology is motivated by the time-varying spline model introduced by Harvey and Koopman in 1993, but includes several important modifications. One is the multi-step approach to modeling daily load and then using the residuals to model the hourly data. The proposed method also use variables based on temperature to modify the model parameters, and in addition express these parameters as a time series, thus producing a model that is more dynamic and attuned to temperature related changes. Data from the Atlantic Electric Zone of the PJM market is used to illustrate the proposed method.