Identification method of temporal fluctuation characteristics of new energy generation under full time domain analysis

The power generation output of the new energy power grid has fluctuating characteristics under the introduction of background harmonics in the grid, making it difficult to accurately identify or match the installed capacity. To this end, a method for identifying the temporal fluctuation characteristics of new energy generation under full time domain analysis is proposed. By combining multi-scale time convolutional networks and multi head attention mechanisms under full time domain analysis, key features such as new energy generation power, capacity factor, and generation efficiency are extracted at multiple scales for temporal fluctuations; Using temporal feature decomposition techniques to decompose the temporal fluctuation characteristics of new energy power generation into IMF components at different time scales, and based on clustering algorithms, adaptively select a classifier for temporal fluctuation characteristics of new energy power generation to complete the classification of temporal fluctuation characteristics of power generation; Realizing the identification of temporal fluctuation characteristics in new energy generation based on the BiGRU network model. The test results show that this method can effectively extract the temporal fluctuation characteristics of new energy generation through full time domain analysis. Accurately decompose the temporal fluctuation characteristics of new energy generation into three time scale components. Under adverse weather conditions, this method can still accurately identify the power fluctuation status of the system.