YAN Renwu, WU Huimin, LI Peiqiang. Ultra-short-term Wind Power Multi-step Forecasting Based on DT-KAN-iTransformer[J]. 2025, 53(10): 95-105. DOI: 10.20204/j.sp.2025.10013.
Ultra-short-term Wind Power Multi-step Forecasting Based on DT-KAN-iTransformer
To address the issues of data distribution shifts and modeling channel-wise dependencies among multiple variables in ultra-short-term wind power multi-step forecasting,a novel forecasting model named DT-KAN-iTransformer is proposed. First,a Dish-TS (DT) module is introduced to perform normalization and denormalization before and after model processing,respectively,to alleviate distribution shifts within and across spatial domains of wind power data and maintain data window distribution consistency. Second,an improved iTransformer network incorporating Kolmogorov-Arnold networks (KAN) is employed to establish channel-wise correlations among different variables through transposed embedding and nonlinear mapping,thereby capturing complex nonlinear relationships between multiple variables. Finally,comparative experiments with nine baseline models are conducted using actual data from a 200 MW wind farm. The results demonstrate that the proposed DT-KAN-iTransformer model achieves higher forecasting accuracy across three evaluation metrics.
LIU J,YANG F,YAN K,et al. Household energy consumption forecasting based on adaptive signal decomposition enhanced iTransformer network [J ] .Energy & Buildings,2024,324:114894.
YANG Z,YANG B,LIU W,et al. Short-term wind power forecast based on STL-IAOA-iTransformer algorithm:a case study in northwest China [J ] .Energy Engineering,2025,122(2):405-430.
ANBO M,ZIKANG X,HAO Y,et al. A novel network training approach for solving sample imbalance problem in wind power prediction [J ] .Energy Conversion and Management,2023,283:116935.
OGASAWARA E,MARTINEZ L C,DE OLIVEIRA D,et al. Adaptive normalization:a novel data normalization approach for non-stationary time series [C ] // The 2010 International Joint Conference on Neural Networks (IJCNN).Barcelona, Spain,2010:1-8.
PASSALIS N,TEFAS A,KANNIAINEN J,et al. Deep adaptive input normalization for time series forecasting .[J ] .IEEE Transactions on Neural Networks and Learning Systems,2020,31(9):3760-3765.
TAESUNG K,JINHEE K,YUNWON T,et al. Reversible instance normalization for accurate time-series forecasting against distribution shift .[C ] //International Conference on Learning Representations.Virtual Event.2022:1-2