A Modified Holt-Winter Forecasting Approach with consideration of Multi-Holiday Effects

Load forecasting is critical for power systems because it sheds insight on dispatching load, cost optimization, and grid efficiency, and while great progress has been made in reducing forecast errors by tackling numerous factors, the impact of holiday, has received less attention in research studies. To overcome these limitations, this study tackles the significant influence of the holiday on power load, which can lead to inaccurate forecasts. In order to utilize the holiday variable, the researchers opted to make a study of two important process, and these are the developing the forecasting method that can utilize the holiday variable and implementing the developed method by forecasting the Philippine electrical load data. Thus, this study focuses the first process, which by studying the Holt-Winters forecasting method, the seasonality correction rules were developed to allow correcting of not only one but multiple holiday effects in a year in the original forecasting method. As a result, the study successfully developed a model algorithm (MA-2) that can incorporate the correction equation in the Holt Winter in order for that original method to be modified. Thus, the results of this pre-liminary study on utilizing multiple holiday variable in Holt-winter forecasting method are relevant in order for the researchers to confidently continue and put up the second part of the research study, which is to perform the forecasting process and implementing it on the Philippine Electrical Load Data and evaluate and validate if the modified method improves the performance of original Holt-Winter forecasting method.