Abstract
Wide-area monitoring systems (WAMS) based on phasor measurement units (PMUs) provide the essential information of an electric power system (EPS) in order to face the challenges of generation, operation, and planning. The PMUs are used to obtain the power system ringdown responses. These signals can be analyzed by advanced signal-processing techniques for the estimation of electromechanical modes. This paper proposes a new spectral fitting (SF) approach to estimate these electromechanical oscillation modes and mode shapes by using vector fitting (VF) under noisy conditions. This methodology combines the numerical Laplace transform (NLT) with the VF method. A participation factor is also proposed to identify and detect the system's dominant oscillatory modes. The proposed methodology is demonstrated for a synthetic multi-component signal, a set of noisy synthetic multi-component signals, and numerical simulations of a 16-machine and 68-bus test power system. Results indicate that the proposed technique has the ability and accuracy to carry out a complete identification of electromechanical modes and mode shapes from several ringdown responses, under noisy conditions. © 2019 Elsevier B.V.