{"title":"Short-term Wind Power Local Forecast based on Morphological High-frequency Filter and Local-sensitive Hashing","authors":"L. Liu, T. Ji, M. S. Li, Q. Wu","doi":"10.1109/GTDASIA.2019.8715901","DOIUrl":null,"url":null,"abstract":"As wind power becomes an important part of power systems, more accurate forecast is critical to high-efficiency application of wind power. The paper proposes a novel forecast model of MHF-LSH-SVR based on morphological high-frequency filter (MHF) and local-sensitive hashing algorithms (LSH), focusing on short-term wind power forecast. To mitigate the effect of high volatility of the raw wind power time series on forecast, the MHF filter is used to extract the mean trend time series, which indicates the mean tendency. In order to perform local forecast, the mean trend time series is reconstructed into mean trend segments in the higher-dimension phase space. Afterwards, to obtain more accurate forecast result, the LSH algorithm is applied to pick out the similar segments of the forecast mean trend segment, which are then employed to perform local forecast by the algorithm of support vector regression (SVR). Simulation studies are carried out on the raw wind power time series collected from a real wind farm, and the final results demonstrate that the MHF-LSH-SVR model outperforms the models of persistence (Per.), SVR and MHF-SVR in terms of both accuracy and stability.","PeriodicalId":435708,"journal":{"name":"2019 IEEE PES GTD Grand International Conference and Exposition Asia (GTD Asia)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 IEEE PES GTD Grand International Conference and Exposition Asia (GTD Asia)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/GTDASIA.2019.8715901","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
As wind power becomes an important part of power systems, more accurate forecast is critical to high-efficiency application of wind power. The paper proposes a novel forecast model of MHF-LSH-SVR based on morphological high-frequency filter (MHF) and local-sensitive hashing algorithms (LSH), focusing on short-term wind power forecast. To mitigate the effect of high volatility of the raw wind power time series on forecast, the MHF filter is used to extract the mean trend time series, which indicates the mean tendency. In order to perform local forecast, the mean trend time series is reconstructed into mean trend segments in the higher-dimension phase space. Afterwards, to obtain more accurate forecast result, the LSH algorithm is applied to pick out the similar segments of the forecast mean trend segment, which are then employed to perform local forecast by the algorithm of support vector regression (SVR). Simulation studies are carried out on the raw wind power time series collected from a real wind farm, and the final results demonstrate that the MHF-LSH-SVR model outperforms the models of persistence (Per.), SVR and MHF-SVR in terms of both accuracy and stability.