| Summary: | The most classical subsynchronous oscillation (SSO) mode extraction methods have some shortcomings, such as lower mode identification and poor anti-noise properties. Thus, this paper proposes a new time-frequency analysis method, namely, synchrosqueezed wavelet transforms (SWT). SWT combines the advantages of empirical mode decomposition (EMD) and wavelet, which has the adaptability of EMD, and improve the ability of anti-mode mixing on EMD and wavelet. Thus, better anti-noise property and higher mode identification can be achieved. Firstly, the SSO signal is transformed by SWT and its time-frequency spectrum is obtained. Secondly, the attenuation characteristic of each intrinsic mode type (IMT) component in its time-frequency spectrum is analyzed by an automatic identification algorithm, and determine which IMT component needs reconstruction. After that, the selected IMT components with divergent characteristic are reconstructed. Thirdly, high-accuracy detection for mode parameter identification is achieved by the Hilbert transform (HT). Simulation and application examples prove the effectiveness of the proposed method.
|
|---|
