Short-term electric load forecasting using an artificial neural network: case of Northern Vietnam

Abstract

Short-term electric load forecasting is an important requirement for electric system operation. This paper employs a feed-forward neural network with a back-propagation algorithm for three types of short-term electric load forecasting: daily peak (valley) load, hourly load and the total load. The forecast has been made for the northern areas of Vietnam using a large set of data on peak load, valley load, hourly load and temperature. The data were used to train and calibrate the artificial neural network, and the calibrated network was used for load forecasting. The results obtained from the model show that the application of neural network to short-term electric load forecasting problem is very useful with quite accurate results. These results compare well with other similar studies. Copyright © 2004 John Wiley & Sons, Ltd.