Mobile QR Code QR CODE : Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
Title Analysis of Short-Term Load Forecasting Accuracy Based on Various Normalization Methods
Authors Bo-Sung Kwon ; Rae-Jun Park ; Kyung-Bin Song
Page pp.30-33
ISSN 1225-1135
Keywords Short-Term Load Forecasting ; Normalization ; Exponential Smoothing Method
Abstract The short-term load forecasting is necessary for stable and smooth power system operation. The accuracy of short-term load forecasting for weekdays according to various normalization methods is analyzed. The normalization methods to be analyzed are maximum and minimum normalization, maximum normalization, and Z-Score normalization. And the model used for 24-hours load pattern prediction is the exponential smoothing technique. In order to return the normalized 24-hour load value to the load value, the predicted maximum load, minimum load, average load and standard deviation of load are estimated using exponential smoothing. In the recent three-year case studies, the accuracy of the short-term load forecasting applying maximum normalization, maximum and minimum normalization and Z-Score to the exponential smoothing technique is analyzed based on the mean absolute percentage error(MAPE). The test results show that the maximum and minimum normalization method is better than the others.