| Summary: | In the context of the clean energy revolution and the high penetration of renewables and power electronics, data-driven Transient Security Assessment (TSA) models can significantly reduce the computational burden of power system TSA and adapt to the quickly changing operating states of modern power systems. In this paper, a multi-channel time series data mining framework is proposed to enhance the performance of data-driven TSA models. During the training procedures, a Lagrangian dual framework is adopted to enhance the feature extraction ability of different types of disturbed system trajectories. The proposed method is adopted to an ordinary Long Short-Term Memory (LSTM) model and numerical tests are carried out in the IEEE-39 system. The test results show that the proposed method can effectively improve the performance and generalization ability of the data-driven TSA model.
|
|---|