An improved TLS-ESPRIT-based mode identification technique for low-frequency oscillations in power system using synchrophasor measurements

The accurate assessment of low frequency oscillations (LFOs) in power system is crucial in order to implement corrective measures to ensure its stable operation. The Total Least Square Estimation of Signal Parameters Via Rotational Invariance Techniques (TLS-ESPRIT) is one of the prominent mode estimation techniques that have been developed to characterize these modes based on their dominance in the signal, which requires prior ideas about the numbers of frequency components constituting the signal (model order). The literature proposes many methods for signal model order estimation, which use tolerance values to determine probable separation boundaries between two subspaces. As in real life, the tolerance value fluctuates with system nonlinearity and noise level, making online estimation procedures inefficient and difficult to automate. In order to deal with the aforementioned limitation, the prime contribution of the suggested study is to develop an efficacious model order estimation technique, which uses the convex combination of eigenvalues weightage (CCEW) to scale the dominancy of modes in the trace autocorrelation matrix (ACM). Thereafter, a smart clustering algorithm is used to segregate these modes based on their dominance on the trace ACM into two opponents: signal subspace and noise subspace. The suggested strategy utilizes a two layers of clustering approach to provide a precise estimate of model order by preventing the insignificant eigenvalues from getting clustered in the signal subspace. The performance validation of the proposed technique is done by conducting comparative study with some recently developed techniques for simulated signal, two-area system, real-time probing data of Western System Coordinating Council (WECC), oscillatory data obtained for Western System Coordinating Council (WSCC) 9 bus and IEEE39 bus system simulated on Real-time digital simulations (RTDS) and PMU data connected at Indian bus system.