Abstract:

 In this paper, a mid-term load forecasting is performed using fractal theory as a new tool that can assist management in the distribution power network. Based on rescaled range analysis combined with fractal interpolation functions, an iterated function system is built to predict the power demand over a period of a week. The rescaled range analysis is performed to detect the degree of persistence of power demand for each week. The calculation of the Hurst exponent H is used to establish the iterated function system and built the IFS code. Illustration of the method is given for the power demand of a region of Tetouan City (North of Morocco). The results show the consistence and the accuracy of this prediction method.