1. SWEEP FREQUENCY RESPONSE ANALYSIS AS A DIANOSTIC TOOL TO DETECT TRANSFORMER MECHANICAL INTEGRITY by:Luwendran Moodley Brian de Klerk ETHEKWINI ELECTRICITY

2. INTRODUCTION Why assess the mechanical condition of a transformer? To detect core displacement and winding deformation due to: –Large electromagnetic forces from fault current –Winding shrinkage –Transformer transportation and relocation If these faults are not detected they will manifest into a dielectric or thermal faults Which normally results in the loss of the transformer Hence periodic testing is essential ETHEKWINI ELECTRICITY

3. DETECTION OF WINDING MOVEMENT Prior to SFRA the mechanical integrity of the transformer was assessed with the following methods: winding capacitance; excitation current; and leakage reactance measurements Each of these methods have drawbacks Winding Capacitance: successfully only if reference data is available Ideally measurements must be made on each phase ETHEKWINI ELECTRICITY

4. DETECTION OF WINDING MOVEMENT Excitation Current: Excitation current is an excellent means of detecting turn-to-turn failure as a result of winding movement if a turn-to-turn failure is absent, winding movements can remain undetected. Leakage Reactance: Per phase leakage reactance measurements generally shows no or little correlation between the phases and nameplate discrepancies from nameplate value of 0,5 % to 3 % can be a reason for concern The range of defect detection is to large for an accurate assessments Hence the move to SFRA ETHEKWINI ELECTRICITY

5. FUNDAMENTALS OF SFRA Transformers are made up of complex RLC networks Any physical damage results in changes to these RLC networks These changes in the RLC networks is what we are looking for. ETHEKWINI ELECTRICITY

6. FUNDAMENTALS OF SFRA How does it work? A low voltage signal with varying frequencies are applied to the transformer. The input and output signal is measured. The ratio of the two signal gives the frequency response of the transformer. This ratio is called the transfer function from which the magnitude and phase can be obtained. ETHEKWINI ELECTRICITY

7. FUNDAMENTALS OF SFRA The RLC network offers different impedance path at different frequencies. The transfer function of each frequency is the measure of the effective impedance of the RLC network. Any geometrical deformation, changes the RLC network, which in turn changes the transfer function at different frequencies. These changes gives an indication of damage within a transformer. ETHEKWINI ELECTRICITY

8. ANALYSIS OF SWEEP FREQUENCY RESPONSES Measured responses are analyzed for any one of the following key indicators : Starting dB values The expected shape of a star and delta configuration Comparison of responses to fingerprint Comparison of responses to the different phase of the same transformer Comparison of responses to sister transformer Creation of new resonant frequencies and the elimination of existing resonant frequencies ETHEKWINI ELECTRICITY

9. ANALYSIS OF SWEEP FREQUENCY RESPONSES A guideline for the analysis of SRFA has been in existence for many years 2 kHz scan is sensitive to core deformation, open circuit, shorted turns and residual magnetism. 20 kHz scan is sensitive to mainly bulk winding movement and clamping structure 200 kHz scan is sensitive to deformation within the main and tap winding 2 MHz scan is sensitive to movement of main and tap winding leads Different frequency bands have different sensitivities to different mechanical failure modes ETHEKWINI ELECTRICITY

10. ANALYSIS OF SWEEP FREQUENCY RESPONSES At low frequencies – inductive – magnetic circuit Significant difference in the responses between the outer two phases and the center phase ETHEKWINI ELECTRICITY

11. ANALYSIS OF SWEEP FREQUENCY RESPONSES At high frequencies – complex with a large number of resonant points - dominated by the winding inductance At highest frequencies - capacitive ETHEKWINI ELECTRICITY

12. TYPICAL RESPONSES Typical response of a healthy transformer –Starting dB value is what we would expect –The shape is consistent with a star configuration –All phase lineup with each other ETHEKWINI ELECTRICITY

13. TYPICAL RESPONSES Typical response of a transformer with huge problems –Starting dB value of the white phase is unexpected –The shape is not consistent with a star configuration –Phase do not lineup with each other –Clear frequency shift with the formation of new resonant point ETHEKWINI ELECTRICITY

14. ETHEKWINI’S CONDITION ASSESSMENT Our approach to condition assessment on transformers: –Monthly visual inspection –Oil analysis Oil screen DGA Furan analysis Tan δ of oil at 90ºC –Electrical tests Tan δ and capacitance measurements on windings and bushings 10 kV ratio measurements Excitation current measurements Leakage reactance measurements Sweep Frequency Response Analysis (SFRA) Insulation resistance measurements –Infrared scanning ETHEKWINI ELECTRICITY

15. SWEEP FREQUENCY RESPONSE ANALYSIS (SFRA) SFRA was introduced three years ago in eThekwini Since its introduction SFRA has proven to be a reliable and sensitive means of detecting winding movement SFRA measurements are performed under the following conditions: On all new transformers for fingerprinting purposes After relocation After long duration short circuits After any type of maintenance ETHEKWINI ELECTRICITY

16. SWEEP FREQUENCY RESPONSE ANALYSIS (SFRA) SFRA measurements are performed on the highest, lowest and nominal tap position An additional measurement is made on the faulted tap position Short circuit tests on HV and LV windings These short circuit tests acts as an indication and does not identify the area of concern ETHEKWINI ELECTRICITY

17. CASE STUDIES Four case studies will be presented on real defects Case studies will focus on problems in the low frequency range. ETHEKWINI ELECTRICITY

18. CASE STUDY 1 : Shorted Turn Failure These are the responses of the HV and LV winding of the same transformer Significant difference in the White phase Imbalance in the reluctance on one of the core limbs (white phase) as a result of shorted turns will produces this characteristic response in the low frequencies ETHEKWINI ELECTRICITY

19. CASE STUDY 2 : Multiple grounded neutral Figure A shows the LV winding response of a multiple grounded neutral and Figure B shows the response after the transformer was repaired. Significant difference in the starting dB values This difference in the dB or vertical shift in response A is attributed to the change in the resistance of the transformer ETHEKWINI ELECTRICITY

20. CASE STUDY 3 : Core Magnetization Figure A shows the fingerprint response of the transformer and Figure B shows the response as a result of a magnetized core This transformer was diagnosed with a tapchanger problem – DC test performed on contacts - inadvertently magnetized the core Figure B shows a shift and the elimination of existing resonance frequencies ETHEKWINI ELECTRICITY

21. CASE STUDY 4 : Extra Core Earth These are responses of the middle phase with (White trace) and without (Red trace) an extra core earth. Differences were noted on all three phases but the greatest difference was on the middle phase There is a difference in the capacitance response from about 1,5 kHz to 14 kHz The White trace has a resonant frequency at 1,6 kHz which is absent in the Red trace ETHEKWINI ELECTRICITY

22. CONCLUSION Sweep Frequency Response Analysis has proven itself within eThekwini Electricity to be a valuable diagnostic tool for the detection of winding movement and other faults that affect the transformers impedance An advantage is that reference responses are not required The results obtained are reliable, repeatable and unaffected by test lead position, weather and electromagnetic interference. The test is easy to perform, however measurements must be made confidently and conscientiously to ensure reliable and meaningful diagnosis SFRA when used in conjunction with other diagnostic tools can provide a complete condition assessment of the transformer and in so doing ensure that informed decisions are made by asset managers. ETHEKWINI ELECTRICITY