| Title |
Monthly Electricity Load Forecasting Using a Multiple Fuzzy Forecasting System |
| Authors |
방영근(Young-Keun Bang) ; 이철희(Chul-Heui Lee) ; 박하용(Ha-Yong Park) |
| DOI |
http://doi.org/10.5370/KIEE.2019.68.4.558 |
| Keywords |
TSK fuzzy logic model ; Multiple fuzzy predictors ; interpolation and trend analysis ; Segmented data ; K-means clustering |
| Abstract |
This paper presents design methods of a fuzzy forecasting system to forecast the monthly peak electricity load data in south korea. In the proposed forecasting system, multiple fuzzy predictors are combined by parallel so that each predictor can perform suitable forecasting for corresponding to segmented data set. To segment the original peak electricity load data, data interpolation and trend analysis methods are used. The segmented data set is used as input data, to design each predictor, the TSK fuzzy logic model and the least square method are used for linguistic rule base and parameter identification. Also the K-means clustering algorithm is used to generate suitable fuzzy sets and tune their membership function. Using monthly peak electricity load data from Feb. 2009 to Feb. 2018 in south korea, in simulation section, the forecasting performance and advantage of the proposed system are verified and explained. |