Condition monitoring artefacts for detecting winding faults in wind turbine DFIGs

1
Condition monitoring artefacts for detecting
winding faults in wind turbine DFIGs
By
(1) School of Electrical and Electronic
Engineering, The University of Manchester
(2) School of Engineering, Durham University
2
Introduction

• Doubly fed induction generator (DFIG) is the most
  commonly used type of generator in contemporary
  large variable speed wind turbines
• One of the most common induction machine faults
  are winding faults due to short-circuit or
  open-circuit or abnormal connection of phase
  windings
• A common fault diagnostic method is the
  investigation of frequency spectrum content of
  machine steady state electrical quantities such
  as current and instantaneous power - non invasive
  as stator windings are used as search coils.

3
DFIG modelling for condition monitoring purposes

• Based on coupled-circuit approach
• A circuit is defined as any series connection of
  coils
• Coupling inductances are calculated between
  circuits
• This approach makes it possible to analyze an
  arbitrarily connected n-phase machine while
  taking into account higher order field space
  harmonics

4
DFIG test rig diagram

5
Test Rig Description
Laboratory test bed (viewed from above)
6
Test Rig Description
DFIG Terminal box
Rotor back-to-back converter
7
Healthy and faulty DFIG winding configurations
used in this presentation

a) Healthy
b) Open-circuit DFIG stator
winding configuration

8
Frequency content of predicted DFIG stator line
current for healthy and faulty DFIG operation
a) Predicted current spectrum for DFIG operating
with healthy windings
b) Predicted current spectrum for DFIG operating
with faulty stator windings
9
Frequency content of predicted DFIG total stator
instantaneous power for healthy and faulty DFIG
operation
a) Predicted power spectrum for DFIG operating
with healthy windings
b) Predicted power spectrum for DFIG operating
with faulty stator windings
10
Predicted and measured DFIG stator line current
spectra for DFIG operation with stator
open-circuit fault
a) Experimental current spectrum
b) Predicted current spectrum
11
Predicted and measured DFIG stator instantaneous
power spectra for DFIG operation with stator
open-circuit fault
a) Experimental power spectrum
b) Predicted power spectrum
12
Conclusions

• A DFIG analytical model was developed and a DFIG
  test rig built for the purpose of this study
• Research demonstrates that there are harmonic
  components in DFIG steady state stator current
  and total power spectra that are directly related
  to existence of winding fault. These frequencies
  are slip dependant.
• Simulation data indicate that the power signal
  spectrum carries more fault specific information
  when compared to the current signal. Measurement
  of the power spectrum is however shown to be
  influenced by high noise levels.

13
Thank You