Bessel and cosine filtering approaches for electromechanical modes identification

The real-time system monitoring of electromechanical oscillations in power grids is essential to maintain operational control in an optimal state and to make the right decisions in any adverse situation. In this context, this investigation proposes two filtering approaches to identify modal patterns (damping and frequency) from multivariate and multi-modal oscillation signals stemming from multiple PMUs and locations. The key idea behind this proposal lies in introducing the cosine- and Bessel-based filtering to conduct a time–frequency analysis for dealing with the modal identification of nonlinear signals that arise after disturbance in the grid. Thereby, this research develops the theoretical foundations of both filtering perspectives forming bi-orthogonal bases that support their analysis and synthesis equations. Then, mono-component signals are extracted by spectral analysis, and their modal patterns are extracted via Hilbert transform. Numerical experiments conducted on synthetical oscillating signals, simulated-based signals extracted from a well-known power grid, and synchrophasors captured by the Mexican WAMS, demonstrate the applicability of the proposal. • First-kind Bessel and scaled cosine functions are used as resonator filters. • Bessel and cosine filtering approaches decompose interarea oscillations. • The proposed filtering approaches no require frequency adjustments. • The optimization of the filters’ length achieves the separation of mono-components.