Title :
Experience with artificial neural network models for short-term load forecasting in electrical power systems: a proposed application of expert networks
Author :
Asar, A.-u. ; McDonald, J.R. ; Rattray, William
Author_Institution :
Strathclyde Univ., Glasgow, UK
Abstract :
This paper investigates the feasibility of applying artificial neural networks (ANN) to short term load forecasting in electrical power systems. It describes ANN behaviour for various short term load forecast types and lead-times. These include peak load prediction, half hour ahead forecasts, prediction of load over a flexible time window ranging from a half hour to 24 hours ahead, and load profile forecasting a day in advance. The networks were trained and tested on actual power utility load data and weather data. The absolute average error in forecasting ranges from 0.5% to 2.5% for different cases and confirms the potential of the methodology for economic applications
Keywords :
expert systems; load forecasting; neural nets; power engineering computing; electrical power systems; expert networks; half hour ahead forecasts; lead-times; load profile forecasting; neural network models; peak load prediction; short-term load forecasting;
Conference_Titel :
Artificial Neural Networks, 1993., Third International Conference on
Conference_Location :
Brighton
Print_ISBN :
0-85296-573-7