Activity Number:
|
519
- SPEED: Methodological Advances in Time Series: BandE Speed Session, Part 2
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, July 31, 2019 : 10:30 AM to 11:15 AM
|
Sponsor:
|
Business and Economic Statistics Section
|
Abstract #307896
|
|
Title:
|
Forecasting Daily Electricity Load Profile Using Functional Principal Components and Transfer Function Models
|
Author(s):
|
Abdelmonaem Jornaz* and V A Samaranayake
|
Companies:
|
Northwest Missouri State University and Missouri University of Science and Technology
|
Keywords:
|
Functional Data Analysis;
Time series;
Electricity Load
|
Abstract:
|
Accurate modeling and forecasting of the 24-hour daily electricity load profile has remained an important issue for stakeholders in the business of buying and selling electricity. Practitioners have employed many approaches, including regression analysis with a myriad of input variables to the use of B-splines to model the daily electricity use profile. We introduce an approach based on functional principal components, where the daily use profile is treated as a realization of a stochastic process observed over the 24 hour period. In the proposed method, the smoothed 24-hour electricity load data for each day are subjected to functional principal component analysis and the Eigen scores associated with the Eigen functions (harmonics) are modeled using transfer function models to obtain forecasts of use profiles for future days. The proposed methodology is applied to 20 years of data from the Pennsylvania-New Jersey-Maryland (PJM); the data cover Atlantic Electric zone (AE) in southern New Jersey.
|
Authors who are presenting talks have a * after their name.