Publication

Abstract : The modern power grid protection system should have considerable operational flexibility and resiliency to hedge the variability and uncertainty of high dimensional dependencies. The use of wide-area monitoring systems (WAMS) in the smart grid enables the real-time supervision of power system oscillations. With the help of advanced signal processing methods and big data analytics, time-synchronized phasor measurements can be used to extract valuable information concerning the electromechanical modal properties of power system oscillations. This chapter introduces a novel method for identifying electromechanical inter-area oscillation modes with the help of wide-area measurement data. Variational mode decomposition (VMD) can be considered as a flexible signal processing technique on the wide-area phasor measurements in power oscillation analysis. For the real-time operation, it is challenging to preset the value of the mode number in the VMD process. This issue has been addressed by improving the strategy for VMD, which is presented in this chapter. The first stage involves the use of Complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) technique to generate intrinsic mode functions and gives indexing based on the correlation factor. Depending on the indexing, the mode number is selected for the second stage VMD process. Techniques such as spectral analysis and Hilbert transform are quite suitable for the estimation of modal parameters. The study is based on significant features of power oscillations, such as determination of damping ratio, amplitude, and frequency. The identification and estimation of low-frequency modes have been performed using this improvised VMD technique, and the results have been compared with those obtained using empirical mode decomposition approaches. The proposed approach is also validated using real-time data obtained from load dispatch centers. The results indicate the effectiveness of non-linear, nonstationary analysis methods for analysing the low-frequency modes and provide reliable validation of these algorithms in analysing real-time data patterns.