PD Monitoring of MV/HV Power Cables December 2016

Transcription

1 PD Monitoring of MV/HV Power Cables December 216 Malcolm Seltzer-Grant

2 Introduction to HVPD Our global presence Introduction to HVPD Ltd HVPD are experts in the growing industry of on-line partial discharge (OLPD) condition monitoring and condition based maintenance (CBM) of high voltage networks. We supply portable and permanent OLPD surveying, diagnostic test and continuous monitoring solutions, and a complimentary range of on-site services, monitoring services and training. Over 4 customers in 1 countries trust our technology.

3 Contents Introduction to partial discharge in power cables Measurement equipment Continuous monitoring Partial discharge location methods Case studies

4 Partial Discharge Detection Theory Available Waveform Display PDs are incepted by the high voltage applied to cable. Chan PD pulses are short duration impulses (ns µs) that propagate in both directions away from PD site between cable core and sheath Time (msec) Signals can be detected on both the core and earth screen at terminations. End A End B

5 MV/HV Cable Faults - Causes and Effects Common Causes Poor workmanship (at cable accessories) Mechanical damage caused by poor installation practices (including damage to cable sheath during cable pulling and minimum cable bend radius being exceeded) Poor quality or poorly manufactured cables and cable accessories Aging of insulation Effects Electrical trees and interfacial surface tracking Localised heating/moisture ingress into the cable (caused by damage to the armour/outer sheath) Infant Mortality and premature failure within the firs years of operation

6 PD Damage to HV Cable Transformer Terminations Tracking on 11 kv Termination (PD detected before failure) Failed 11 kv Termination (same type as opposite)

7 Insufficient Mastic Around Connector in 33 kv Joint

8 Trees on 66kV paper cable

9 Why and When to Perform PD Testing At Manufacture Quality Assurance Type/routine tests, e.g. IEEE/IEC standard At Commissioning To check for transport damage To ensure the installation has been made to a good standard, for example attachment of terminations and joints to power cables Service Life Detection of issues that emerge over time Condition Based Maintenance

10 Survey - Detection Identify equipment with PD in network Simple instrumentation can be susceptible to noise Low level of training Usually only on-line Diagnostic/Location Testing Detailed test result more advanced instruments Higher level of training to perform Off-line and On-line Spot Test: Surveying vs Diagnostic Testing Use survey to identify equipment with PD and diagnostic testing only where PD is identified

11 Spot Testing vs Monitoring Spot Test Snapshot of the condition Doesn t take into account variations with operating stress Labour resource to perform testing Monitoring Continuous evaluation of the condition Detect variations with operating stresses (e.g. temperature/ humidity) Less labour requirement after set-up Usually deployed on critical plant or plant with high PD in spot test

12 Test Equipment Detection Simple handheld detectors, give indication if PD detected and its level Diagnostic More advanced detector, often with PC software More information type of PD, locations Monitoring Portable or permanent logging of PD Interface to plant control system (SCADA)

13 Aspects of Testing MV and HV Cables MV HV Equipment at primary substation and ring main units Cross-bonding less common Some tolerance to PD activity Strong safety motivation for sealing ends Equipment and substations, cable termination, cable joints Very little tolerance to PD activity

14 PD Detection Energies for Different Points in Cable System Corona at metal contacts Electrical charge RF Electromagnetic radiation Acoustic Ultraviolet Ozone Discharges on insulator surface Electrical charge RF Electromagnetic radiation Acoustic Ultraviolet Ozone Partial discharge in termination insulation system Electrical charge RF Electromagnetic radiation (local) Acoustic (local) Partial discharge in cable insulation or joint Electrical charge RF Electromagnetic radiation (local) Acoustic (local)

15 High Frequency Current Transformer (HFCT) Sensors Detect PD in cables and connected plant Wide bandwidth (from 1 khz to 2 MHz) Attach to power cables at terminations and earthing links of HV equipment Installation inside or outside of cable box Temporary or permanent

16 HFCT Sensor Attachment to Power Cables The HFCT sensor should be attached to intercept either the conductor PD current (i+) or the earth PD current (i-) 1 HFCT on Earth (i-) 1 HFCT on cable with Earth brought back through (i+) 2 HFCT around cable (i- + i+ = ) 3 2

17 HFCT Attachment at Cross-bond Points on 132 kv Cables

18 PD Against Phase for Power Cables On-line Detection Single Core Cable Available Waveform Display Three Core Belted Cable with HFCT on Common Screen Available Waveform Display Chan Time (msec) Chan Time (msec)

19 Modern PD Detection Systems PD Sensors Hardware Filtering/ Amplification Digitiser/DSO PD Data Analysis Software Trigger signal

20 Noise Reduction Performed with the Kronos Software Raw Data PD Event Recognition Apply Expert De-noising Rules De-noised PD Data Indicated Condition The Kronos automatically 1,121,68 21,183,18 (before more noise HVPD training) pulses (1%) implements recognises are rejected. data the 1,182,992 rules on the software 61,924 point (6%) PD acquired data points pulses Indicated (.2%) as possible correctly over 33,941 PD pulses identified and rejects Condition power the cycles remainder as noise

21 Data Analysis Data before analysis/noise rejection Data after analysis/noise rejection

22 Data Analysis

23 Continuous Monitoring

24 Continuous PD Monitoring Aspects Detect cyclic changes in activity Load varying activity on PILC cables Humidity related activity from surface discharges Detect changes that relate to incipient faults Gradual rise Sudden rise Sudden drop Carried out on: key circuits, circuits with suspected cyclic PD changes, circuits with high spot-test results

25 PD and Load Relations Although PD incepted by voltage, load can have effect Mostly on PILC cables Load variations Movement of oil/impregnant Expansion of conductors S S M T W T F S S M T W T F

26 1kV PILC Cables Prozess Datum : : : : : : : : : : : : : : : : : : : : : : : : :3 PD burns PD in burns a hot in cable: a cold electrodes cable (9% expand of cases): - possible fluid shrinks, movements voids inside appear, accessories local PD in lead voids to increased field strengths in dielectrics PD in accessories

27 Examples of PD rises to Failure 45, 4, 35, 3, 25, 2, 15, 1, 5, Number above threshold 1 (start date=12/7/99) Time (Days) Signal levels (mv) Peak, ave, and No above thresholds (start date=29/11/ Chan=18) 36, 34, 32, 3, 28, 26, 24, 22, 2, 18, 16, 14, 12, 1, 8, 6, 4, 2, Time (Days) 5 6 Thresh 1

28 On-line Cable Mapping (PD Site Location)

29 PD Detection Theory PD Pulse Propagation and Attenuation Cables act as waveguides for PD pulses and as low-pass filters. PD pulses attenuate and disperse as they travel down the cable PD sensor must have a good low frequency response to detect long distance PD. Increasing the number of test points gives more conclusive results. A study was carried out using ~5 pc calibration pulses injected into 2 km, 4 kv cable to determine attenuation and measurement range the pulse was successfully detected 2 km from the source. Reference: S.Sutton, R. Plath and G. Shröder, The St.Johns Wood Elstree Experience Testing a 2km Long 4kV XLPE- Insulated Cable System After Installation, Jicable 27-7 th International Conference on Insulated Power Cables, Paris Versailles, June 27.

30 PD Location in Power Cables Direct Pulse Reflected Pulse ΔT = Time difference between direct and reflected pulses. L = Cable Return Time for cable Measurement End PD event Remote End Direct pulse ΔT Reflected pulse L PD Site Location ΔT PD% = 1 1 L

31 Single-ended Cable Mapping Only possible for long distances when the tested cable s far end impedance change is HIGH (e.g. if the end of the cable goes into a transformer and/or the circuit breaker at the far end of the cable is OPEN).

32 OLPD Location on Power Cables In many on-line cases reflected PD pulses are often not visible: Attenuation is too large to measure reflected pulses from the far end (long cables) Waveforms too difficult to interpret (noisy signals) Teed or jointed cables Cables with many ring main units or switches Cables with no change in impedance at the far end Cross-bonded cable circuits multiple reflections

33 Double-ended Cable Mapping (Range up to 5 km or 3 RMUs) Necessary when the tested cable s far end impedance change is LOW (e.g. if the far end circuit breaker is CLOSED). The HVPD Portable Transponder system amplifies PD signals, allowing the HVPD Longshot at the other end of the cable to receive and interpret the relative arrival times of pulses at each end of the cable to give an accurate location. The cable earth strap must be accessible at both ends of the cable in order to perform double-ended mapping.

34 On-line PD Location on Power Cables Example of Usage Without Transponder ΔT Reflection may not be clearly visible (e.g. due to noise) With Transponder The large transponder pulse removes any confusion ΔT ΔT tr ΔT = Time between direct and reflected PD pulses ΔTtr = Transponder time delay

35 On-line PD Location on Power Cables Example Results 15 Direct PD Pulse Voltage (mv) Reflected PD Pulse Transponder Time Delay Reflectogram showing PD and transponder pulses Transponder Pulse Time (usec) All Phases PD (pc) PD location map for all PD pulses in cable section under test Location (% along cable)

36 CASE STUDY: OLPD TESTING AND CABLE MAPPING OF 33 KV XLPE CABLES IN METRO NETWORK

37 Case Study: Introduction OLPD testing was carried out in response to a number recent faults* of 33 kv cable joints within the customer s network. The faults led to disruption of the power supply to the rail system. The purpose of the testing was to measure and locate any PD activity within the cables with particular focus on the cable joints. * It should be noted that this was a newly installed cable system that had been inservice for just over 12 months before the faults started to occur.

38 Case Study : OLPD Testing Equipment and Methodology On-line Cable PD Mapping using the HVPD Longshot test unit and Portable transponder. Tests started with calibration testing with pulse injection HFCTs.

39 Case Study: Test Results Cable PD signals have been detected on Blue Phase with cross-talk (lower magnitude) on Red and Yellow phases. The source of PD was located to Joint Number 2 (Jt2) using the on-line PD mapping technique. The faulty joint on this cable was replaced and re-tested using the HVPD Longshot test unit to verify the repair was good

40 Case Study: Top 2 Worst Performing Circuits Out of the 5+ circuits tested, Major PD was detected within cable accessories on the three of the circuits (6%) as shown in RED in the Table below. The levels of discharges detected put these 33 kv cables into RED category, Major concern, locate PD and then repair or replace. Criticality Number Circuit Comments Peak Cable PD Level (pc) Local PD Level (db) Cumulative Cable PD Level (nc/cycle) OLPD Criticality (%) Maintenance Action 1. DUB to MPS1 C2 B Phase < ABS to AH C2 B / Y Phase 9729 < Major concern, 3. BUR to HCC C2 B / Y Phase 3781 < locate PD and 4. BUR to HCC C1 B / Y Phase 3245 < then repair or 5. ABS to AH C1 B / Y Phase 292 < replace. 6. NHD to QYD C2 R Phase 2849 < ALQ to AHS C2 B Phase 1733 < Some concern, 8. MPS3 to BNS C2 R / B Phase 1337 < repeat test and 9. NHD to QYD C1 R Phase 887 < regular 1. HCC to CRK C1 Y / B Phase 759 < monitoring 11. AHS to SLD Y / R Phase 75 < recommended. 12. STD to ABH Y Phase 238 < ALR to BNS C1 B Phase 184 < ALR to BRJ No PD detected <1 15. ALG to PMD No PD detected <1 16. ALG to KBW No PD detected <1 17. AQD to AQ2 No PD detected <1 18. JDD to CRK No PD detected <1 19. ODM to JDF C1 No PD detected <1 2. ODM to JDF C2 No PD detected <1 Re-test in 12 months.

41 CASE STUDY: OLPD Testing, Location, Monitoring with Preventative Maintenance on a 33 kv Land- Sea Offshore Wind Farm Export Cable (UK)

42 Case Study: Circuit Details 33kV Grid Substation Offshore Wind farm 33kV Switching Substation Circuit 1 Circuit 2 Offshore 33kV GIS Switchgear Land Cables 3 x single core Land-Subsea Cable Joints 3 core Subsea Cables 1.7 km single core XLPE land cable 9.6/11.5 km 3 core XLPE subsea cable

43 Case Study: OLPD Test and Mapping Data L1 L2 L3 PD Magnitude (pc) Cable PD Phase of Pow er Cycle (deg) PD Magnitude (pc) Cable PD Phase of Pow er Cycle (deg) PD Magnitude (pc) Cable PD 1, 5, -5, -1, Phase of Pow er Cycle (de High levels of PD (of up to 1, pc / 1 nc) measured on Circuit B, Phase L3.

44 Case Study: PDMap Graph Showing PD Location Switching Substation Joint Pit 7 Land-sea Transition Joint , Location (meters) 1,2 1,4 1,6

45 Case Study: PD Signals Before and After Joint Replacement BEFORE High PD detected on L3 PD Located , 1,2 Location (meters) 1,4 1,6 Joint 7 with PD removed and replacement cable section installed Lower-level sporadic PD signals from different site after joint replacement AFTER

46 Case Study: Circuit B Evidence of Surface Degradation Due to Bad Fitting Heatshrink Stress Control

47 Case Study: OLPD testing of 11 kv XLPE cables and terminations for oil refinery client (Slovakia)

48 Condition Assessment Following the failure of two 11 kv transformer cable terminations, OLPD testing on the other cable terminations of same type was carried out.

49 Failed 11 kv Transformer Cable Termination

50 Test Set-up and Results PD Magnitude (pc) Cable PD Phase of Pow er Cycle (deg) 36 PD Magnitude (pc) Cable PD Phase of Pow er Cycle (deg) PD Magnitude (pc) Cable PD Phase of Pow er Cycle (deg) 36 HFCT L1 Earth Strap HFCT L2 Earth Strap HFCT L3 Earth Strap Cable PD Segment Waveform Significant levels of PD activity (of up to 4 pc) were detected on Phase L2. Volts (mv) Time us 2 3

51 Forensic Investigation L2 termination was eventually replaced 15 months after the initial tests were made. Investigation showed evidence of severe tracking. The cable had not yet failed i.e. the OLPD testing gave a very good early warning of 15 months against the incipient fault.

52 End of Presentation Thank you for your time Any Questions? Malcolm Seltzer-Grant