2IntroductionHigh Voltage Partial Discharge Ltd (HVPD) are experts in the growing technology field of On-line Partial Discharge (PD) test and monitoring technology for in-service high voltage cables and plant.We now have a significant body of experience in the On-line PD testing of offshore installations, primarily Oil and Gas, Renewables, and Sub-sea interconnector cables.The full scope of the medium voltage (MV) power network of offshore subsea and deep-water installations can be tested and monitored by on-line and off-line testing for PD related condition degradation.
3Land-sea export cables 33 kV – 132 kV HVAC and HVDC IntroductionHVPD have over 7 years of experience in the insulation condition testing and monitoring of subsea and land-sea HV cables:Land-sea export cables 33 kV – 132 kV HVAC and HVDCOffshore wind farm sub-sea MV cable arraysOil & gas platform MV interconnection cablesHVPD have carried out on-line condition testing of subsea cables using three main condition assessment techniques:On-line Partial Discharge (OLPD) testing, monitoring and mappingPower Quality/Harmonics/Earth Screen Current monitoringTime Domain Reflectometry (TDR) cable ‘fingerprinting’ tests
4HVPD Offshore Experience HVPD’s subsea global cable clients include:HVDC & HVAC Interconnector ownersOff-shore oil and gas extractionOff-shore wind farm operators
5Off-line and On-line PD Testing of Subsea Cables FactoryAll components tested separately along with factory-made cable jointsAt commissioningEnsure installation integrityEnsure no transport damageDuring the cable’s service lifeBaseline OLPD measurement and/or permanent monitoringRoutine maintenance/CBMPost Fault test and analysis (to ensure the repairs have worked)
6‘Drivers’ for Applying On-line Testing and Condition Monitoring Cables are essential for the effective operation of offshore installations.On-line condition testing and monitoring provide an early warning against HV insulation faults - to provide sufficient advance warning carry out preventative maintenance and avert unplanned outages.Most of these incipient insulation faults can be detected prior to potential catastrophic failure.State and condition monitoring technologies include using early-stage detection of PD activity, temperature hot-spots using DTS, power quality/harmonics, sheath currents and overvoltage/overcurrent events such as transient earth faults (TEF’s).
7Why Test Subsea HV Cables for PD? Long repair time and/or replacement time.Repairs are expensive as additional cable lifting vessels are required for subsea cable faults.Unplanned outages are much more expensive to repair than scheduled preventative maintenance during planned outages (outages can cost around times more).Operators need to ensure their HV cable networks have high reliability, good maintainability and maximum availability.Diagnostic On-line PD testing and monitoring can be done without outage, with the cable assessed under normal operating conditions
8Reliability Centred Maintenance ‘Bathtub Curve’ End of Life ‘Wear-out’Infant MortalitySteady State FailureEnd of Life ‘Wear-out’Failure RateSteady State Failure‘Infant Mortality’ Phase3 YearsTime20-50 Years
9Installation Lifetime Testing and Monitoring Philosophy DamageMistakeRepairAgingManufacturingTransportationInstallationOperationAcceptance TestingFactory TestingContinuous Monitoring
10Specific ‘Drivers’ for the Offshore Renewables Industry Offshore wind generation - the most expensive of the renewable energy options (at a total cost of £130/MWh over the 20-year asset life) c.f. Onshore wind generation - £75/MWh.The UK Government’s Department of Energy and Climate Change, DECC, are looking at technologies which can reduce this lifetime cost of offshore wind generation by 25% (to £100/MWh by 2020).More proactive, on-line condition monitoring and predictive asset management tools allow better scheduling of planned maintenance by detecting issues early and reducing expensive unplanned outages.These cost reductions will only be achievable through a operator company moving towards condition based management (CBM).
11MV and HV Cable Failure Rate Data for Offshore Wind Farm Networks Typical annual failure rate for land-based, UK distribution utility MV cable networks per 100 km.The annual failure rates of some of the recently installed UK offshore cable networks per 100 km.These failures typically occur within the first 3 years of service life, typically at their weak points at cable joints and cable terminations due mainly to incorrectly installed cable accessories.The failures during the first 3 years of service are referred to as the ‘Infant Mortality’ Phase, as they occur within the initial stages of service life.
12Subsea MV/HV Cable Export Cable Networks Site Overview Offshore Platform or MV/HV SubstationMain Components (6):Land terminations, Land XLPE cables, Land joints, Land-subsea joints, Subsea XLPE cables (with factory joints), Offshore terminations
13Subsea MV and HV Cable Networks – Key Components Site Overview Offshore Platform or MV/HV SubstationDeep Water Wind TurbineSubsea XLPE CableUp to 10 kmor 3x Turbine RMUSubsea XLPE CableUp to 10 kmSubsea export cable to land(typically 132 kV)Main Components (6):Land terminations, Land XLPE cables, Land joints, Land-subsea joints, Subsea XLPE cables (with factory joints), Offshore terminations
14Subsea MV and HV Cable Networks Key Components Tower Base J-Tubes – Protect and secure cables up height of pile / tower.Subsea Cable – XLPE, armoured, single cable for three phases with fibre-optic communication. Jointed frequently.Offshore Platform or MV/HV SubstationDeep Water Wind TurbineTower(deep water only)J-TubesSubsea XLPE cableUp to 10 kmUp to 10 kmor 3x Turbine RMUTowards substationTowards end of stringSubsea Cable JointFoundation Pile
15Subsea MV and HV Cable Networks Key Components Nacelle and Transformer Generator / Power Converter (690 V)Generator – Power is generated and converted to low voltage AC.Transformer – Fed directly from the generator and mounted either at the top or bottom of the tower.33 kV typical voltage.690 V Cables690 V – 33 kV Transformer33 kV Cables to RMU
16Subsea MV and HV Cable Networks Key Components Tower Base 33 kV Cables from TurbineRing Main Unit – Connection point for adjacent turbine subsea cables and feed from transformer aboveAccess Lift – For moving equipment or personnel up tower.Access LiftRing Main UnitSubsea CableJ-TubeSubsea Cables run down J-Tubes
17The Weak Points within the Subsea MV and HV Cable Networks PD Hazards The most likely points of failure of the subsea cables are at the cable joints and terminations as shown below.Offshore Platform or MV/HV SubstationPrior to failure, many ‘incipient’ cable insulation faults produce PD activity which can be detected and located on-line, with the cable remaining in service, to enable preventative maintenance to be made.
18The Weak Points within the Subsea MV and HV Cable Networks PD Hazards Key points of potential failure of the wind farm network are as follows:Turbine Generatorand TransformerTest locationOffshore Substation switchgear / Export TransformerTest locationRing Main UnitTest locationSubsea Cable JointsJ-tube Cable Entry PointsMain Components (6):Land terminations, Land XLPE cables, Land joints, Land-subsea joints, Subsea XLPE cables (with factory joints), Offshore terminations
19Most Likely Causes of Subsea Cable Faults Poor installation causing partial discharge (PD) and tracking in cable joints and terminations.Inadequate cable/cable accessories design, inadequate mechanical protection for duty in the sea producing sheath faults from external abrasion.Mechanical wear and bending caused by movement of subsea cables, this can cause mechanical movement of the cable joints.Due to the nature of offshore cable networks, the cost of an unplanned outage can range from anything from 10x to 100x of the failure of a similar onshore cable, depending on where the faulted cable is!
20Causes of Subsea MV and HV Cable Faults Thermal – incorrect design or installation of cross-bonding or earthing bonds (at cable joints) can cause ‘thermal runaway’, tracking and failure.Electrical – incorrect installation and/or poor workmanship of the cable accessories are the No.1 cause of faults within the ‘infant mortality phase’.Ambient – the effects of mechanical wear and tear from bending caused by movement of the subsea cables with tidal and current changes in the sea.Mechanical – inadequate mechanical protection of the cables for duty in the sea, particularly at the land-sea cable intersection.
21Examples of Insulation Faults in a Medium Voltage Land-Sea Export Cables
22HVPD’s On-line PD Measurement Technology for Subsea Cables HVPD Longshot™- Spot-testing and locationOptions for long-term condition monitoring:HVPD Multi™ Portable- Short-term monitoringHVPD Multi™ Permanent- Long-term/permanent monitoring
23Case Study 1:Typical On-line Partial Discharge (OLPD) Test of 132 kV and 33 kV Assetson 500 MW Wind Farm
24BackgroundThe proposal for expansion of Wind Farm submitted for completion in 2016 will increase the number of operating wind turbines from 140 to 280 generating up to 1,008 MW of power.The growing number of assets will require effective and non-intrusive methods of testing to ensure a reliable long-term operation, reducing costs of maintenance as well as avoiding downtimes.
2545 km 132 kV Export Cables are connected to Windfarm via Offshore Substation Platform
26Two main areas of test on main export offshore substation are:- 132 kV GIS Switchgear under test using HVPD Longshot and HFCT on available cable terminations
27High levels (at this voltage) of wide frequency spectrum background noise were detected with HFCT sensors using HVPD Longshot™ and PDS Air™.The noise exists in the frequency spectrum occupied by genuine Cable PD events. The results of FFT on Noise waveform are shown below:
28An attempt to reduce some of the noise was made using 100 kHz and 200 kHz High Pass Filters as shown in Table 6 below.Due to levels of noise testing with Longshot is recommended at this location going forwards.
29…and 33 kV GIS Switchboard using HVPD Longshot and TEV Sensors.
30Cables were not accessible for testing due to requirement for isolation. Noise was investigated at site. Two distinctive groups of intermittent high frequency content pulses with magnitude of up to 41dB were detected within the switchboard room using TEV sensors as shown below:
31Low Levels of typical Electronic Switching Noise ResultsEM noise was associated with audible electro-mechanical sounds in the room.Data collected when intermittent background noise pulses disappeared:The exact location of the noise should be pinpointed with HVPD Longshot™ unit and distributed TEV sensors. The noise source should be then isolated if possible for duration of PD testing.If any TEV readings with PDS Air™ are to be made on the switchboard, they should be made when intermittent noise is not present i.e., the audible electro-mechanical noise is not heard.Phase PatternLow Levels of typical Electronic Switching Noise
3233 kV Turbine Switchgear initially screened with HVPD PDSAir™
33PD Sensing Techniques and Measurement Systems HFCTs on 33 kV feeder cables to turbine switchgear Non-intrusive PD sensors are used for on-line PD detection in MV/HV cable networks:Inductive, wideband, high frequency current transformer (HFCT) sensors (1 per phase) are used to detect PD in the cables and remote plant.Transient earth voltage sensors (TEV) are used for detection of electromagnetic radiation from ‘local’ PD activity nearby the sensor from sources in the cable termination or switchgear.By combining these sensors, sensitivity to different types of PD can be obtained and the measurements can also be correlated to aid diagnosis.
34HFCT PD Sensors on 33 kV and 132 kV Cables Non-intrusive, inductively coupled Further examples
3533 kV Turbine Switchgear cable arrangement – install HFCTs around each cable. Cables to Transformer(unarmoured)Cables to Array String(armoured)
36HVPD Longshot with TEV sensor and HFCT. Test feeders to turbine, and both feeders along cable string
37Access to nacelle by service lift or ladder climb. If testing of plant at top required, equipment is loaded into lift and hoisted.
38ResultsModerate levels of wide frequency spectrum background noise were detected with HFCT sensors using HVPD Longshot™ and PDS Air™.Switchgear, Array Cables and Transformer incomer cables were tested for Local and Cable PD as well as to assess the condition of the 33 kV Transformer windings.Due to levels of noise testing with Longshot™ is recommended to assess the insulation condition of the cables and 33kV Transformer windings. Site is suitable for cursory screening of the Switchgear panels with PDSAir™ using the handheld’s inbuilt TEV sensor.
39Logistics for Personnel and Equipment Access is often restricted due to weather conditions and boat availability. Average daily boat costs are in the region of £5k with boats limited to 12 men including crew.Boat transfer or Helicopter transfer to assets.
40Ladder climb to platform and Entrance via Transition Piece. Equipment hoisted onto platform in 20 kg hoist bags.
41Case Study 2:Experiences Testing and PD Location for a 33 kV Offshore Wind Farm Oil & Gas Supply Cable Network
42BackgroundTwo platforms, A and B, are separated by a subsea cable of approx. 5 km length.Two wind turbines supply power exclusively to satellite oil extraction platform A, 1.5 km away.Several tests were performed to assess condition of key circuits
431 1 2 2 Site Overview Tests were performed at: Platform A 33 kV SwitchgearTurbine A RMUPlatform A 33 kV switchgear120 km Incomer from beach1Approx. 1.5 km subsea cable to wind farm2Two 5 MW wind turbines Water depth – 50 metres2Approx. 500 metre cable linking the two turbines
44Test 1PD Test and Mapping performed with Longshot™ on Platform A’s 33kV 4x feeder switchboard.
45PD Detected in cable splice Test 1PD Test and Mapping performed with Longshot™ on Platform A.Significant PD detected within red phaseOn-line TDR and single-ended PD mapping located incipient fault to the cable splice on the platform, 15 m out from the switchgear.PD Detected in cable splice
46PD Test and Mapping performed with Longshot™ on Platform A. Significant PD detected within red phaseOn-line TDR and single-ended PD mapping located incipient fault to the cable splice on the platform, 15 m out from the switchgear.Results (shown right)
47PD Test and Mapping performed with Longshot™ on Platform A. Significant PD detected within red phaseOn-line TDR and single-ended PD mapping located incipient fault to the cable splice on the platform, 15 m out from the switchgear.Mapping data (shown right)
48Test repeated 2 years later PD still present, remedial work had not been performed.Activity had not increased, but risk of failure still present, compromising a key transformer’s reliability.
49PD Detected 140 m from Wind Farm Test 2PD Test and Mapping performed with Longshot™ on Platform A.Significant PD detected within yellow phase cable joint.Single-ended mapping located PD to the subsea joint, 140 m from wind farm.Platform APD Detected 140 m from Wind Farm
50PD Test and Mapping performed with Longshot™ on Platform A. Significant PD detected within yellow phase cable joint.Single-ended mapping located PD to the subsea joint, 140 m from wind farm.
51PD Detected 140 m from Wind Farm Test 2Test repeated 2 years later, from the wind turbine A end of the cable.PD still present in the cable joint, remedial work had not been performed.Activity had increased, but risk of failure still present, again threatening power supply.Platform APD Detected 140 m from Wind Farm
52PD Detected on 33 kV Transformer feeder cable Test 3PD measured on the Yellow phase of the Turbine A feeder.Peak magnitudes of PD have been measured in excess of 6000 pC, with average levels of 1871 pC.The overall activity is moderately high, activity levels typically 30 nC/cycle have been measured.This PD located on the Yellow phase (L2) Wind farm Turbine A 33 kV cable joint at the 33 kV transformer.Platform APD Detected on 33 kV Transformer feeder cable
53PD measured on the Yellow phase of the Turbine A feeder. Test 3PD measured on the Yellow phase of the Turbine A feeder.This PD located on the Yellow phase (L2) Wind farm Turbine A 33 kV cable joint at the 33 kV transformer.
54PD measured on the Yellow phase of the Turbine A feeder. Test 3PD measured on the Yellow phase of the Turbine A feeder.This PD located on the Yellow phase (L2) Wind farm Turbine A 33 kV cable joint at the 33 kV transformer.
55PD measured on the Yellow phase of the Turbine A feeder. Test 3PD measured on the Yellow phase of the Turbine A feeder.This PD located on the Yellow phase (L2) Wind farm Turbine A 33 kV cable joint at the 33 kV transformer.
56Joint location confirmed Test 3Measurement of PD pulses observed on Turbine A cable feeder, showing a clear direct & indirect reflected pulse.Joint location confirmedPD located to 52 m, agreeing with visual estimates.
57Case Study 2:PD Testing, Location, Monitoring and Preventative Maintenance of a 33 kV Land-Sea Wind Farm Export Cable
58Case Study 2: 33 kV Wind Farm Export Cables Background Off-shore wind farm with two 33 kV export supply cable circuitsProject carried out in follow up to a number of faults (at joints) on land cable sectionOn-line PD testing, PD Mapping and short-term (1 week) monitoring was carried out to assess insulation condition of circuits
59Case Study 2: 33 kV Wind Farm Export Cables 33 kV Export Cable Circuit Details1.7 km single core XLPE land cable9.6/11.5 km 3 core XLPE subsea cable
60Case Study 2: 33 kV Wind Farm Export Cables On-line PD test equipment installation at switching substationHFCT SensorsHVPD Multi™ PD Monitor
61Case Study 2: 33 kV Wind Farm Export Cables On-line PD test results - Phase Resolved PD PatternsL1L2L3Circuit AL1L2L3Circuit B
62Case Study 2: 33 kV Wind Farm Export Cables Circuit A: On-line PD monitoring resultsPD Activity Over One Week – Load related PD on Phase L3Circuit 1L1L2L3
63Case Study 2: 33kV Wind Farm Export Cables Circuit B: On-line PD monitoring resultsPD Activity Over One Week – high levels of hourly PD variationCircuit 2L1L2L3
64Case Study 2: 33 kV Wind Farm Export Cables High PD consistently detected on three cables.PD Mapping (PD site location) carried out.Initial focus on Circuit B, L3 phase (highest PD, up to 10,000 pC).PD MappingStep 1: On-line TDR for cable return time.Step 2: Location of measured PD pulses.
65Case Study 2: 33 kV Wind Farm Export Cables On-line TDR Set-UpNote: PD pulses take similar propagation path
66Case Study 2: 33 kV Wind Farm Export Cables On-line TDR WaveformInjected pulseMeasured Return Time = 20.9µsReflection from land-subsea joint.Reflection from grid substation
67Case Study 2: 33 kV Wind Farm Export Cables PD MappingDue to reflection from land-subsea joint pulse reflection was observed on larger PD pulses.Possible to perform single-ended PD location on Circuit B L3 phase, land section from on-shore substation.Direct PD pulsePD pulse reflection from land-subsea joint.Single PD pulse and reflection measured at switching substation
69Case Study 2: 33 kV Wind Farm Export Cables High PD Detected on L3PD Located2004006008001,0001,2001,4001,600Location (meters)Lower-level sporadic PD signals from different site after joint replacementJoint with PD removed and replacement cable section installed
70SummaryPulse reflections from land-subsea joint allowed locations to be made of large PDs on land cableRegular testing of other cables recommended with PDMapping possible if PD level increasesPD phase patterns indicate different types of PDDefective joint removed on circuits A and B, L3A- Defects due to insufficient mastic around connectorB- Defects along XLPE surface due to bad fitting stress control
71Case Study 3:Cable Failure on a 12.2 km, 6.6 kV Oil & Gas Platform Feeder Subsea Cable
72Main Gas Production Platform 56 metres from Satellite Platform Case Study 3: High Cost Cable Failure on a 12.2 km, 6.6 kV Subsea Feeder CableMain Gas Production PlatformSatellite ProductionPlatform6.6 kV Subsea Cable(12.2 km)Fault Location -56 metres from Satellite PlatformThe main gas production platform had a single 12.2 km, 6.6 kV interconnecting submarine cable to the satellite production platform, this being the sole supply to all of the MV motors and MV plant on the satellite.An earth fault occurred on the blue phase cable core which meant that the satellite gas platform lost power. Off-line, pinpointing TDR measurements were used to locate the fault and direct the cable repairs.
73Case Study 3: High Cost Cable Failure on a 12. 2 km, 6 Case Study 3: High Cost Cable Failure on a 12.2 km, 6.6 kV Subsea Feeder CableThe off-line TDR testing was conducted at both ends of the 12.2 km long cable. The ‘Return Speed’ of the cable was measured using calibration pulse injections and was found to be 92 m/microsecondTDR measurements from the main platform showed the location of the fault at 99.54% of the cable length out from platform (or % of the total cable length out from the satellite platform).Using the total cable length and the return speed of PD pulses along the cable provided a fault location of around 56 m out from the satellite platform.This position on the cable coincided with the entry point of the cable to the subsea cable pipe.
74Case Study 3: High Cost Cable Failure on a 12. 2 km, 6 Case Study 3: High Cost Cable Failure on a 12.2 km, 6.6 kV Subsea Feeder CableIt was found that the cause of the fault had been due to the (mechanical) stress-relieving ‘bung’ at the end of the pipe becoming dislodged.The cable had failed due to abrasion due to mechanical movement, with respect to the fixed pipe that led to the cable failure through a sheath fault to earth.The cable fault was repaired with new cable joints and OLPD tests were carried out at both ends of the cable circuit after the repair to ensure that the new cable joints had been installed correctly.These tests showed the re-instated cable was discharge-free and thus suitable for service.
75TDR ‘Fingerprinting’ on the Isle of Man Case Study 4:TDR ‘Fingerprinting’ on the Isle of Man90 kV, 108 km AC Interconnector Cable
76Case Study 4: TDR Fingerprinting on IOM 90kV Interconnector Cable Background108.7 km, 90 kV AC subsea cable1.94 km land cable (England), 105 km of subsea cable and 1.75 km land cable (Isle of Man).The TDR ‘fingerprinting’ tests were made during a routine maintenance outage to establish cable return time, pulse propagation speeds and joint positions
77Case Study 4: TDR Fingerprinting on 90 kV Cable Recorded TDR trace showing return timeMeasured Return Time = 1.28msInjected Pulse60 km jointSecond reflectionFar end
78Case Study 4: TDR Fingerprinting on 90 kV Cable Zoom to show land joints and the land-sea transition jointsLand-Sea Transition JointLand Joints
79Case Study 4: TDR Fingerprinting on 90 kV Cable Comparison of TDR Measurements to route records for jointsFeatureTime, msLength, mCalculatedFrom recordsDifference% errorStartN/ALand joint 17.19640623172.7%Land joint 214.7213101270403.1%Land/Subsea transition21.619231940-170.9%‘10 km' BICC/ Pirelli Joint13111661115081531.3%‘60 km' Pirelli/ BICC Joint7216417864300-1220.2%
80Case Study 4: TDR Fingerprinting on 90 kV Cable The TDR records showed a good match to the cable route records e.g. 122 metres in 60 km (0.2%).TDR proved achievable even on these very long circuits (this technique was also applied successfully on the 290 km long 400 kV ‘Basslink interconnector’ from Victoria, Australia across the Bass Strait to Tasmania).The cable owner will use the pulse propagation speed data obtained from this test to quickly and accurately locate any fault or ‘incipient fault’ using PD Mapping.
81The Future - HVPD’s Proposed Development of a ‘Holistic’ Subsea Cable Monitoring (HSCM) SystemIt is proposed that the HSCM system will provide complete ‘state and condition’ monitoring for subsea high voltage cable circuits and networks.This will be achieved by combining a number of existing technologies into one acquisition platform.The system would require low power, distributed monitoring node units transmitting data via fibre-optic cable.
82The Future - HVPD’s Proposed Development of a ‘Holistic’ Subsea Cable Monitoring (HSCM) SystemThe HSCM system will combine the following 5 technologies.Distributed PD Monitoring (DPDM) module for the detection and location of early-stage PD activity.Distributed Temperature Sensing (DTS) module.Power Quality Monitoring (PQM) module.Sheath Current Monitoring module.Overvoltage/overcurrent recorder – to record transient earth faults (TEF’s).
83HVPD Subsea Holistic HSCM Condition Monitoring System HVPD Proposed ‘Holistic’ Subsea Cable Monitoring (HSCM) System Layout(Subject to Cable Layout)Offshore HSCM Master UnitOnshore HSCM Master Unit20 km20 kmEvery 20 km20-40 km20-40 km20-40 km20-40 kmSubsea Node units
84ConclusionsThe Purpose of any condition monitoring to provide an advanced ‘early warning’ against ‘incipient’ HV insulation faults to enable planned, preventive maintenance to be carried out and to avert unplanned outages.With cross correlation of a number of cable condition and state parameters, it is then possible to provide the necessary level of detailed diagnostic data required for implementing Condition Based Management (CBM) schemes.The subsea HV cable owners need to consider radically better Power Quality Monitoring (PQM) and condition monitoring systems to help drive down the related high operational and maintenance costs.A move towards CBM is viewed as key for operating costs reduction.Effective CBM can only be implemented if there is detailed, real-time (continuous) diagnostic intelligence and data on both the state and condition of the HV subsea cables
85Do you have any questions? End of PresentationThank you for your timeDo you have any questions?