Dominant modes estimation using SCUSUM method along with SSI

Low-frequency oscillations (LFOs) in the power system have emerged as a significant possible cause of abrupt wide-area blackouts in several parts of the world. As consumer demand grows, so does the insecurity of the operational stability points of power systems as a result of unforeseen events such as connection and disconnection of loads, generators, or inter-area power networks. Although the stochastic subspace identification performed well in identifying certain LFO modes, it gives inaccurate results due to incorrect model order estimates. Incorrect estimation of model order results in the inclusion of trivial modes to essential subspace modes of a power system. The Sequential Cumulative Sum (SCUSUM) approach of order estimation detects changes in the mean with respect to the eigenvalues. This paper proposes the use of the SCUSUM method instead of estimating using the singular value decomposition of the weighted projection matrix in Stochastic Subspace Identification.