A Frequency-based RF Partial Discharge Detector for Low-power Wireless SensingPartial dicharge

A Frequency-based RF Partial Discharge Detector for Low-power Wireless Sensing

1. A Frequency-based RF Partial Discharge Detector for Low-power Wireless Sensing

2. index INTRODUCTION Rf pd monitoring PD frequency characteristics Wireless sensor network Detector requirement DETECTOR OVERVIEW OUTPUT DATA Laboratory case study CONCLUSION Reference

3. INTRODUCTION PARTIAL DISCHARGE “Partial discharge is a localized dielectric break down of a solid or fluid electrical insulation system under high voltage stress” Degradation in dielectric insulation during operational life time or manufacturing time Leads to internal arcing NEED FOR DETECTION To avoid equipment failure Ensure equipment uptime M itigate business risk

4. Rf pd monitoring Analyzing unit Wide band UHF signal capture unit Conditioning unit Coupled PC Visualizing unit User interface

5. Disadvantages It requires a large processing capability to capture UHF signals Not viable to install in all high voltage plants Advantages Effective & accurate tool. GIS & oil filled transformer.

6. PD frequency characteristics Defect-specific information used for classification based on frequency spectrum. The relationship between PD source and sensor in terms of geometry and distance due to complex propagation effects

7. Wireless sensor network wide-range of monitoring applications. Ad-hoc network with redundant links. Sensory data is passed back through data aggregation nodes to a wired network Data is presented to monitoring engineers. Integrated computing platform; Encapsulating sensing, processing, data storage, communications & power components in a single compact Package.

8. Disadvantages Degrading effect of Impulsive noise emitted by PDs on wireless data channels. Advantages Low cost Absence of potentially hazardous cabling Reduced bandwidth requirements High capacity communication links b/w substations &corporate networks.

9. Detector requirement Functional requirements Sensing & recording intensity of PD signals & relative spectral magnitudes Differentiation of PD from RF noise Relatively small Interface with RF sensors ( plant enclosures) A Low-power 3channel detector-

10. DETECTOR OVERVIEW DETECTOR FREQUENCY BANDS Band Range Lower 0 – 450 MHz Middle 400 – 750 MHz Upper 700 – 3200 MHz

11. 3 identical channels Schottky diode detector 5MHz LPF Amplifier PHYSICAL DEVICE OVERVIEW Sampling by an ADC output WB RF signals Output is three pulse envelopes ( the relative energy within the three frequency bands)

12. OUTPUT DATA Peak value of the PD envelope for three discrete frequency bands. Values are then normalized into a proportional form, It represents the relative spectral energies within the PD signal. The total magnitude of the captured sample is also included as a feature, as the sum of the three channels.

13. Test tank with three monopole RF sensors. Addition panel with RF PD sensor types. Three test cells. A 50kV foster transformer, energized up to 15kV to generate PDs within each of the cells. Cells filled with SF6, pressurized at 2 bar . Laboratory case study

14. Two of the test cells. On the left is the floating electrode in SF6 test cell, and on the right is the rolling particle in SF6 test cell. . Internal view of RF test tank, Defect positions within the test tank

15. Two positions are available within the test tank-in front of the sensor hatch and in the center of the test tank. limiting factors -position of the transformer and the length of the high-voltage cables . Safer operation directly beneath the sensor hatch, . The floating electrode test cell was oriented in three planes to simulate the RF emission of an individual defect propagating in different directions .

16. EXPERIMENTAL RESULTS Protrusion in sf6: 1

17. The spectra varies in intensity between the lower and middle bands, but have less than a 10% proportion of high-frequency content. The results fall into two distinct clusters, corresponding to two positions. That multiple protrusion defects in SF6 & can be distinguished based upon frequency spectra.

18. 2

19. PDs generated by this test cell were found to form a Tight cluster. The defect orientation have little effect on the recorded RF spectrum. higher proportion of spectral energy in the >700 MHz band, differentiating them from the floating particle and protrusion.

20. 3 Rolling particle in sf6

21. The results from both positions fall within the same region of the chart, with half the spectral energy falling within the middle 400MHz-750MHz band. The remaining spectral energy varies between the upper and lower bands based upon position. Results from each position are uniform with minor variance, falling in tight clusters. defect test cell is uniform, and the measured spectrum varies only with location.

22. CONCLUSION: Novel approach to RF PD monitoring using a low-powered detector employing frequency-based technique. Capable of determining the presence of multiple defects, & rudimentary defect classification. Ternary plots have been used for the presentation and analysis of the PD data, allowing linear separation of defects

23. Reference P. C. Baker, S. D. J. McArthur, and M. D. Judd, “Data Management of On-Line Partial Discharge Monitoring Using Wireless Sensor Nodes Integrated with a Multi-Agent System”, Intern. Conf. Intelligent Systems Applications to Power Systems (ISAP), pp. 1–6, 2007. Z. Tang, C. Li, X. Cheng, W. Wang, J. Li, and J. Li, “Partial discharge location in power transformers using wideband RF detection”, IEEE Transactions on Dielectrics and Electrical Insulation Vol. 17, No. 1; February 2010 ” A Frequency-based RF Partial Discharge Detector for Low-power Wireless Sensing” P. C. Baker, M. D. Judd and S. D. J. McArthur

24. Thank you