Y Aslan - Electrical Engineering Department, University of Dumlupinar, Kutahya, Turkey
S Yavasca
C Yasar

In the system planning, control and management of power distribution companies the long term peak load demand projections has an important role. In general, the load forecasting is performed by studying the past events. Long term forecasting of future peak load demand is very important for the economic and secure operation of power systems. In this study the long term peak load forecasting is performed for the city of Kutahya with the least squares regression based methods and artificial neural networks (ANN) using the load, temperature and population growth data from 2000 to 2008. The results attained are validated with the real data obtained from the Turkish Electricity Distribution Corporation (TEDAS) which represents the monthly peak load electric consumption in Kütahya, Turkey. By comparing the forecasted results with the real data the most suitable method is proposed.

Load Forecasting, Least Squares Regression,Artificial Neural Networks