skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Quantifying and Reducing Uncertainty in Correlated Multi-Area Short-Term Load Forecasting

Abstract

In this study, we represent and reduce the uncertainties in short-term electric load forecasting by integrating time series analysis tools including ARIMA modeling, sequential Gaussian simulation, and principal component analysis. The approaches are mainly focusing on maintaining the inter-dependency between multiple geographically related areas. These approaches are applied onto cross-correlated load time series as well as their forecast errors. Multiple short-term prediction realizations are then generated from the reduced uncertainty ranges, which are useful for power system risk analyses.

Authors:
; ; ; ; ;
Publication Date:
Research Org.:
Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
1339048
Report Number(s):
PNNL-SA-114275
TE1103000
DOE Contract Number:
AC05-76RL01830
Resource Type:
Conference
Resource Relation:
Conference: IEEE Power and Energy Society General Meeting (PESGM 2016), July 17-21, 2016, Boston, MA
Country of Publication:
United States
Language:
English
Subject:
load forecast; principal component analysis; sequential Gaussian simulation

Citation Formats

Sun, Yannan, Hou, Zhangshuan, Meng, Da, Samaan, Nader A., Makarov, Yuri V., and Huang, Zhenyu. Quantifying and Reducing Uncertainty in Correlated Multi-Area Short-Term Load Forecasting. United States: N. p., 2016. Web. doi:10.1109/PESGM.2016.7741272.
Sun, Yannan, Hou, Zhangshuan, Meng, Da, Samaan, Nader A., Makarov, Yuri V., & Huang, Zhenyu. Quantifying and Reducing Uncertainty in Correlated Multi-Area Short-Term Load Forecasting. United States. doi:10.1109/PESGM.2016.7741272.
Sun, Yannan, Hou, Zhangshuan, Meng, Da, Samaan, Nader A., Makarov, Yuri V., and Huang, Zhenyu. 2016. "Quantifying and Reducing Uncertainty in Correlated Multi-Area Short-Term Load Forecasting". United States. doi:10.1109/PESGM.2016.7741272.
@article{osti_1339048,
title = {Quantifying and Reducing Uncertainty in Correlated Multi-Area Short-Term Load Forecasting},
author = {Sun, Yannan and Hou, Zhangshuan and Meng, Da and Samaan, Nader A. and Makarov, Yuri V. and Huang, Zhenyu},
abstractNote = {In this study, we represent and reduce the uncertainties in short-term electric load forecasting by integrating time series analysis tools including ARIMA modeling, sequential Gaussian simulation, and principal component analysis. The approaches are mainly focusing on maintaining the inter-dependency between multiple geographically related areas. These approaches are applied onto cross-correlated load time series as well as their forecast errors. Multiple short-term prediction realizations are then generated from the reduced uncertainty ranges, which are useful for power system risk analyses.},
doi = {10.1109/PESGM.2016.7741272},
journal = {},
number = ,
volume = ,
place = {United States},
year = 2016,
month = 7
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:
  • A model for hourly load (hourly energy consumption) forecasting is developed. The method is based on using the linear estimation theory to obtain four different prediction models for the load at a particular hour corresponding to the four correlation periods, namely an hour, a day, a week, and a year. The four predicted values with their respective variances are combined optimally to obtain a composite forecast for the hour. The hourly forecasts are improved by selectively rejecting one or more of these four predicted values. Actual data from the Hydro-Quebec system load file for the period of September 28, 1969more » to September 26, 1971 are used as data for identification of the model. The load forecasts for the period of September 27, 1971 to September 26, 1972 are compared with the loads actually realized during the same period. (auth)« less
  • This paper reports on the developments and findings of the Distribution Short-Term Load Forecaster (DSTLF) research activity. The objective of this research is to develop a distribution short-term load forecasting technology consisting of a forecasting method, development methodology, theories necessary to support required technical components, and the hardware and software tools required to perform the forecast The DSTLF consists of four major components: monitored endpoint load forecaster (MELF), nonmonitored endpoint load forecaster (NELF), topological integration forecaster (TIF), and a dynamic tuner. These components interact to provide short-term forecasts at various points in the, distribution system, eg., feeder, line section, andmore » endpoint. This paper discusses the DSTLF methodology and MELF component MELF, based on artificial neural network technology, predicts distribution endpoint loads for an hour, a day, and a week in advance. Predictions are developed using time, calendar, historical load, and weather data. The overall DSTLF architecture and a prototype MELF module for retail endpoints have been developed. Future work will be focused on refining and extending MELF and developing NELF and TIF capabilities.« less
  • The development of feed-forward artificial neural network based models to predict gas consumption on a daily basis is the subject of this paper. The discussion concerns an iterative process based on network sensitivities and intuition regarding selection of proper input factors. The method is applied in forecasting gas consumption for two regions in the State of Wisconsin, namely, a portion of metropolitan Milwaukee and a region near Fond du Lac. The investigation includes a study of the effects of various network sizes and training algorithms given a limited availability of relevant historical data. The effects of using multiple sources ofmore » weather data are also investigated for regions without a centrally located weather recording station. The training results indicate that feed-forward artificial neural network based models reduce the residual predicted consumption root mean squared errors by more than half when compared to models based on linear regression using identical input factors. The inclusion of weather data from multiple sources leads to further error reduction. The models were then used for load forecasting starting in Summer 1994. The static models performed well from the Summer 1993 until the severe cold weather of mid-January 1994. This poor performance was scrutinized; investigations point to probable causes that include severe weather patterns non-existent in the training data and the excitation of consumer behavioral modes either not evident in the training data or not well represented by time-of-the-year input factor proxies. The search for a solution lead to dynamic models that were implemented by updating the neural networks with new data as it became available. These models performed significantly better but still poorly during and after the severe weather of January 1994.« less
  • The accurate prediction of hourly electrical demand one or more days ahead is of great economic importance to electric utilities for generation unit dispatch and unit commitment. Artificial neural networks for pattern recognition are developed to identify days in the historical record that are most similar to the days being forecasted, to use for load prediction. Artificial neural networks are also used to generate linear and nonlinear multivariate time series models, to project demands forward in time. The genetic algorithm is used to select the optimal set of independent variables for forecasting. Techniques are developed to combine forecasts derived frommore » independent methods, to achieve better accuracy than any single forecast. In this way, artificial neural networks can be used to generate practical, accurate short-term electrical load forecasts.« less
  • Short-term electric load forecasting (STELF) plays an important role in electric utilities, and several techniques are used to perform these predictions and system modelings. Recently, artificial neural networks (ANN`s) have been implemented for STELF with some success. This paper will examine improved STELF by optimization of ANN techniques. The strategy for the research involves careful selection of input variables and utilization of effective generalization. Some results have been obtained which show that, with the selection of another input variable, the ANN`s use for STELF can be improved.