CUSUM based Fault Detection of Stator Winding Short Circuits in Doubly-Fed Induction Generator based Wind Energy Conversion Systems

Abstract

This paper investigates the detection of short-circuit faults in the stator winding of a Wind Energy Conversion System (WECS) based on a Doubly-Fed Induction Generator (DFIG). The Cumulative Sum (CUSUM) algorithm is used to detect the change in the stator negative-sequence current magnitude which is used as the fault residual. The grid voltage unbalance is considered along with the fault position and its severity to evaluate the effect on the residual. The generator is modeled using the winding-function approach. Simulation results are presented for a 15 kW system operated, under a power control strategy, at rated conditions. It is shown that the fault-detection method is able to detect the fault rapidly and is not influenced by the grid voltage unbalance within normal limits.

Wavelet‐based index to discriminate between minor inter‐turn short‐circuit and resistive asymmetrical faults in stator windings of doubly fed induction generators: a simulation study

“…In the proposed approach, the frequencies of interest under both stator winding fault conditions are twice and four times the fundamental frequency in the spectrum of rotor-reference-voltage space-vector magnitude signal calculated according to (13)- (15). DWT analysis is a signal analysis based on a combination of approximation (A j ) and detail (D j ) coefficients, via a mother wavelet and scaling functions for J decomposition levels.…”

“…The frequencies of interest are around twice the fundamental frequency. The other diagnostic techniques reported for this type fault of WRIMs or DFIGs in the literature are based on phase angle difference of three-phase stator currents [11], rotor current spectrum [12][13][14], stator negative sequence current magnitude [15,16], stator and rotor current spectrum [17][18][19], and artificial neural network [20]. The multiple-coupled circuit model based on winding function approach [11-13, 15, 21], finite-element method [14,16], magnetic equivalent circuit model (MEC) [22,23], abc time-domain model [17][18][19], hybrid qd model [24], and qd model [25] are the models that have been used to simulate ITSCF of stator windings.…”

Multi-Physical Signature Analysis of Induction Machines under Unbalanced Supply Voltage

“…Besides the classical fast Fourier transform (FFT) more advanced spectral-based techniques, proposing a finer resolution or applicable to noisier signals are reviewed in [2], [9]. Methods tracking the Park's vector module or pattern [10], [11], or the negative-sequence current [12], as well as using Hilbert transform are also proposed [2], [9].…”