Data-Driven Fault Diagnosis Approach for Synchronous Generators

Abstract

This article presents a data-driven approach for diagnosing interturn short-circuit (ITSC) faults in the field winding of synchronous generators (SGs). A notable advantage of this method is its independence from the load’s linearity or nonlinearity. The method’s foundation is derived from analyzing the impact of ITSC faults on the state-space model of an SG, utilizing the SG equations in the dq rotor reference frame. Based on the state-space model, subspace identification and input–output data, including voltages and currents, are used to estimate the eigenvalues of the state matrix. The detection, isolation, and estimation of faults are achieved through the estimated eigenvalues, without relying on the model. Simulation and experimental results validate the effectiveness of this data-driven fault diagnosis methodology.