Multichannel continuous wavelet transform approach to estimate electromechanical oscillation modes, mode shapes and coherent groups from synchrophasors in bulk power grids

Continuous Wavelet Transform (CWT) is a traditional single-channel method to estimate the dominant mode from measurements, but it is rarely applied to estimate mode shapes and coherent generators. On the other hand, estimation accuracy of traditional CWT is significantly affected by the observability of oscillations. This paper develops a multichannel CWT which is based on multichannel measurements and is less observability constrained so as to estimate not only dominant modes, but also mode shapes and coherent generators. First, wavelet power spectrum (WPS) is applied to wavelet coefficient matrices (WCMs) of multichannel measurements obtained by CWT to detect the critical scale ranges associated with the dominate modes. Then, the WCMs with the same scales in the detected ranges extracted from the raw WCMs are used to estimate the dominant modes and mode shapes. Meanwhile, the measurements that only contains the information of dominant modes are reformed by inverse CWT to detect the coherent groups of generators using direction cosines. The proposed approach is applied and evaluated with the simulation data from the 16-generator 68-bus test system and field measurements from Phasor Measurement Units (PMUs) in China Southern Power Grid (CSG). Results show that the proposed approach is accurate and efficient in estimating dominant modes, mode shapes and coherent groups of generators from synchrophasor measurements.