Mobile QR Code QR CODE : The Transactions of the Korean Institute of Electrical Engineers
Title Prediction of Special Day’s Hourly Load : Relative Coefficient Method vs Adjusted Coefficient Method
Authors 남영진(Young-Jin Nam) ; 조하현(Ha-Hyun Jo)
Page pp.1523-1533
ISSN 1975-8359
Keywords Load Pattern Analysis; Load Forecasting; Relative Coefficient; Adjusted Coefficient; Two-Step Procedure
Abstract In order to predict special day‘s hourly loads, it is important to identify the characteristics of the special days in advance. In other words, a basic analysis should be carried out, taking into account that special days generally show a lower or similar level of hourly load compared to normal days and include a small number of samples. In addition, it is important to look at patterns of each special day and use the same special day to reflect the calendar effects rather than simply using the latest identical special day information in analyzing the load pattern. This is defined in this study as a calendar effect or day effect adjustment. Deriving special day’s hourly load in this way is called the relative coefficient method. On the other hand, these relative coefficient method do not reflect other information such as weather or economic factors. Therefore there is a prediction error in the relative coefficient method, which is adjusted through regression to derive a new load forecast through the adjusted coefficient method. In other words, the two predictive methodologies of these relative coefficient and adjusted coefficient are defined in this study as a Two-Step Procedure and are intended to contribute to forecasting special day’s hourly load by comparing the predictability of each model.