Case Studies on Power Cables

1Prof. Charles Q. SuThe Petroleum Institute

Case Studies on Power Cables

Case study - 1

Quality Management of Distribution Cables

Professor Charles Q. Su( PhD, Fellow IET, SM IEEE, CIGRE A2 )

Prof. Charles Q. SuThe Petroleum Institute

About the workshop instructor- Prof. Charles Q. Su

Industrial experience

1970-1973 Operations engineer

1974-1978 HV testing engineer

2002-2006 Chief Technologist, Singapore Power (SPPG) Ltd

Research & teaching experiences

1985 Research Associate, University of Western Australia

1990-1991 Lecturer at University of NSW, Australia

1992-2001 Senior Lecturer & Associate Professor, Monash University

2007-now Professor, Chair of Research Committee (EE)

Petroleum Institute, UAE

Membership of professional organisations

Fellow of IET, Senior Member IEEE (91), member of CIGRE SC A2 (Transformer)


Ron E James & Q. Su Condition Assessment of HV Insulation in Power System Equipment - IET Power and Energy Series No.53, April 2008


Some important issues in distribution cable

management

Causes of distribution cable failure:

1. Damages (road digging, land movement etc)

2. Manufacture defects (material of quality control problems)

3. Poor workmanship (cable joints and terminals)

4. Insulation ageing (water seepage, water treeing etc)


Condition assessment of distribution cables

Criterion of condition assessment:

1. The total failure rate.

2. Frequency of a type of failure warrantees root cause analysis.

3. The consequences of failures.

4. Costs of repair or replacement.


Cable insulation ageing in the life span

Bathtub curve - Determined from the failure rate change (a number of the same insulation samples)

TwTs

0

Random or slowly increa sing fa ilure ra te

Bur n-in

period

Use ful life period We arout

period

Ope rating

life

Failure

rate


Background

There are over 3,700 km of 6.6kV cables in a utility. The average failure rate from 2000 - 2004 was 30 cables per year.

Serious consequences:

1. In-service failures of 6.6kV cables cause local blackout.

2. Due to the time of failure (e.g. at mid-night) and the possible bad weather conditions at failure (e.g. thunder storm), restoration of power supply is difficult and could take many hours.

There was an urgent need to reduce the in-service failures.

AGE PROFILE OF DISTRIBUTION CABLES (2004)

0.35% (13.109 km)

9% (323.093 km)

14% (541.152 km)

25% (954.976 km)

36% (1,349.877 km)

15% (583.758 km)

0

6.6kV

Cable Length

: 3,768.647 km

22kV

Cable Length

: 4,948.379 km

21 25 yrs

16 20 yrs

11 15 yrs

6 10 yrs

< = 5 yrs

6% (288.551 km)

13% (668.678 km)

25% (1,218.271 km)

34% (1,694.035 km)

22% (1,078.844 km)

> 30 yrs

26 30 yrs

21 25 yrs

16 20 yrs

11 15 yrs

6 10 yrs

< = 5 yrs

20 40 60 80 100 (%)

0.07% (2.682 km)

19% beyond the

age of 15

27% beyond the

age of 15


How to reduce the in-service failures of 6.6kV cables?

A. Cable replacements according to their designed life time?

B. HV tests on all cables to flash out incipient faults? (using DC, AC or VLF)

C. Replace the type of cable joints of high failure rate?

Which one would you prefer if you are the asset manager?

12 Prof. Charles Q. Su

0

10

20

30

40

50

60

70

80

Cable (XLPE) Cable (PILC) Joint

Year

Categories of 6.6kV Cable Failures( 2000-2005 )

39

25%

45

28%

76

47%

* Total 160 failures


0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Year

Age Profile of Failed 6.6kV XLPE Cables( 2000-2005 )

Average age is around 20

Implication: XLPE cable insulation is generally reliable within 15 years.


Age Profile of Failed 6.6kV Cable Joints( 2000-2005 )

Average 28 years

Implication: Cable joint can fail at any time due to mainly poor workmanship, as

well as bad quality of materials and insulation ageing


EXISTING MAINTENANCE TESTS

for 6.6kV cables

Megger measurement - Resistance - Comparison between phases

Polarization index (R10/R1)


Effectiveness of Megger Test

Detect the leakage caused by terminal contamination (surface crapping resistance)

Water seepage to the joint

Insulation deterioration (ageing), especially paper/oil cables


Megger Test Alone Is Not Conclusive

For example:

Water tree contamination (before electric tree is established)

Bubbles and unbridged cracks in XLPE or epoxy insulation


Failures of cables with high megger readings

Case (1)

For example, cable A:

Megger readings 1G/1G/1G on 16 Nov 2005

The circuit failed on 17 Nov 2005 at 6:14am

Also, in this utility a number of 6.6kV cables of high megger reading failed in the past.


Case (2)

A 66kV XLPE cable under a bus stop failed;

It was found that the failure was due to an

early damage caused by sinking an earthing

rod;

Lost about 1/3 of the XLPE insulation


A close look

10


A surprise

It was found that the bus stop was built five years

before the failure.

So, after the bad damage, the poor cable survived five more years before its insulation broke down;

More surprisingly .

Its insulation resistance was measured three times

during the five years, always giving high megger

readings!


Why megger tests could not detect the

incipient fault (damage)?

XLPE insulation has a very large volume

resistivity of 1016 .cm.

The damage did not bridge the insulation.

Water trees do not affect insulation

resistance before electric treeing is

established across the electrodes.

11


How to use megger tests effectively?

Comparison of megger readings between phases;

Trend analysis;

Stability of insulation resistance reading under dc high voltage.

Add polarisation index measurement in the analysis (PI = R10min / R1min)

Prof. Charles Q. Su

ACTION PLAN VLF tests on selected cables

Selection criteria

Cables with seram joints (more frequent failures)

First leg feeders (important)

Megger readings VLF test is carried out if :

1. M < 50 M or

2. 50 M < M < 200 M and K > 1.5 or

3. 200 M < M < 1000 M and K > 5

Where M is the minimum megger reading for the three phases and K the ration between the maximum and minimum phases.

12

Prof. Charles Q. Su

VLF tests voltage and time duration

For cables less than 10 years old, 2Uo for 15 minutes

For cables older than 10 years, 1.7 Uo for 20 minutes

These test voltages and time are in line with the new IEEE Standard on VLF tests IEEE Std 400.2TM 2004. The IEEE/EPRI/CEA and other world engineering bodies recommended test level for MV extruded cables is two to three times line to ground voltage for 15-60 minutes.

Initial VLF Test

Flowchart

Note: M is the minimum megger

reading of the three phases

13


Incipient Faults Averted by VLF Tests (May 2003 Dec 2005)

0

5

10

15

20

25

30

35

40

45

50

2003 2004 2005

Num

ber

of fa

ilure

s


VLF Tests on 6.6kV Cables(May 2003 to Dec 2005)

Total Circuits Tested Failed During VLF Tests

(incipient faults averted)

540 97

100% 18%

14

Statistics of 6.6kV Cable Failures between 2000-2005( Total 160 cable and joint failures )

0

5

10

15

20

25

30

35

40

1999 2000 2001 2002 2003 2004 2005 2006

Year

Nu

mb

er

of

failu

res

Before 2004, the cable failure rate was around 30 per year.

In 2005, it dropped to 12, about 1/3 yoy


Why Cable Still Fails after Passing VLF Test?

Total Circuits

Tested

Failed During

VLF Tests

Failed in Service

after VLF Tests

540 97 20

100% 18% 3.7%

15


Profile of the 20 Cables Failed After VLF Tests

0

2

4

6

8

10

12

14

16

18

XLPE PILC Joint

17

2 1


Age Profile of XLPE Cables Failed After VLF Test(Average 19.5 Years)

0

5

10

15

20

25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Year

16


Why Cable Fails After VLF Test? ...

The possible causes:

More than one defects.

Bad water tree contamination.

VLF test time is too short.

Electrical Tree Grow During VLF Tests(In case of two large water trees)

Cable sheath

Conductor

XLPE

insulation

1. Star VLF test at 1.7Uo which may initiate electrical treeing on some large water trees.

No electrical treeing is triggered on small water trees.

2. The electrical trees start to grow until the largest one bridges across the insulation and

causes flashover.

No electrical tree is initiated on small water trees and

defects

Electrical trees are initiated at large water trees

17

Initial and Modified VLF Test Criteria

Note: M is the minimum megger reading of the three phases


Electrical Tree Growth under Different Voltages(IEEE Standard)

Voltage Tree Growing Speed( mm/hour )

50Hz 1.7

0.1Hz

Cos-rectangular

7.8

0.1Hz Sine 12.3

Implication: The VLF test time should be sufficiently long.

18

6.6kV Cable Failures in 2005

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

Year

5 PILC

Ave Age = 30

4 XLPE

Ave Age = 22

3 Joints

Ave Age = 6


HV Oscillating Wave Tests

Energise cable by DC voltage source.

After the voltage reaches a certain level, discharge through an inductor to ground.

A damped oscillating voltage is established which may last for a few mini-seconds.

Detect partial discharges and dielectric dissipation factor during OW tests.

Locate PDs using PD mapping techniques.

19


PD mapping under OW tests.

Some defects, especially those in cable joints, could be detected by PD mapping

Detector sensitivity is not better than 100pC at site in noisy environment

Not suitable for the detection of defects in XLPE insulation (very low PD level, normally

20


PD MAPPING TEST RESULTS


PD MAPPING TEST RESULTS

21

Prof. Charles Q. Su

Selection consideration:

Recommended by standards

From utilities experiences

TEST VOLTAGES

FOR 6.6KV AND 22KV XLPE CABLES

Prof. Charles Q. Su

Test Standards for 6.6kV and 22kV XLPE Cables

IEC Standard 60502-2: 2005Power cables with extruded insulation and their accessories for rated voltages from 6 kV up to 30 kV

European Standard CELENEC HD 620 S1 and HD 621 S1

IEEE Standard 400.2-2004IEEE guide for field testing and evaluation of the insulation of shielded power cable systems using VLF

EPRI report RP 3392-01/CEA 200-D-780A (1996)Trial guide for high voltage 0.1Hz tests on power cable systems in the field

22

Prof. Charles Q. Su

IEC RECOMMENDED ELECTRICAL TESTS( For new 22kV cables after installation )

AC 50/60Hz 1.7 Uo for 5 minutes or 24 hours under system voltage *

AC 50/60Hz test voltage and time are determined by agreement between the purchaser and the contractor

Other test methods (VLF, OW etc) are under consideration

* IEC Standard 60502-2: 2005Power cables with extruded insulation and their accessories for rated voltages from 6 kV up to 30 kV

EUROPEAN STANDARD(for PE and XLPE cables from 6kV to 36kV)

Frequency Test voltage (rms)

Test time

0.1 Hz 3 x Uo 60 minutes

50 Hz 2 x Uo 60 minutes

European Standard for cable after laying test CENELEC HD 620 S1 AND 621 S1

15 European countries signed the harmonization document 620 S1 and 621 S1 in 1996

23

VLF test voltage and duration

adopted by some utilities in North America

Age of cable Test voltage (RMS)

6.6kV 22kV

Newly installed 12kV (3.1Uo) 38kV (3.0Uo)

1~10 years old 9.5kV(2.5Uo) 32kV (2.5Uo)

10~30 years old 6.5kV(1.7Uo) 22kV (1.7Uo)

Note: 1. Test duration is always 15 minutes. 2. The data is from HV Inc, America.

IEEE Standard 400.2-2004IEEE guide for field testing and evaluation of the insulation of

shielded power cable systems using VLF

System Voltage

rms in kV

Acceptance Test

rms or (peak)

Maintenance Test

rms or (peak)

5 10 (14) - 3.5Uo 7 (10) 2.4Uo

8 13 (18) 2.8Uo 10 (14) 2.2Uo

15 20 (28) 2.3Uo 16 (22) 1.85Uo

25 31 (44) 2.15Uo 23 (33) 1.6Uo

35 44 (62) 2.2Uo 33 (47) 1.6Uo

Test duration : 60 minutes

24

Prof. Charles Q. Su

RECOMMENDATIONS(for 6.6kV and 22kV cable after laying tests)

Insulation resistance test at 5kV

- Purpose: detect poor workmanship and joint/terminal insulation leakage

VLF tests at 2Uo RMS for 60 minutes

- Purpose: flush-out insulation defects. If failed during VLF test, after repair the cable should be VLF tested again regardless of the insulation resistance.

If necessary*, oscillating wave and PD mapping tests could be carried out at the following peak voltages: 1 Uo, 1.5 Uo and 2 Uo.

- Purpose: detect and locate defective joints and insulation weakness

* Criteria of PD level to be determined

Prof. Charles Q. Su

VLF and OW PD mapping tests flow chart

25

Prof. Charles Q. Su

NEW DIAGNOSTIC TESTS( maintenance tests )

DC component in AC leakage current

- water tree detection

Propagation characteristic spectroscopy

- LV pulse attenuation versus frequency

- For insulation ageing detection

- Can apply to in-service cables

AC superposition test (101Hz)

- Detect the 1 Hz component

- Detect water tree


SUGGESTIONS

Apply VLF tests to old PILC cables (age>20)

For XLPE cables if 200M

26


CONCLUSIONS

VLF Test has been successful in reducing 6.6kV cable failures and should be used according to the total insulation condition of the cable and joint assets.

Review the test procedure and failures every two years.

Some defects, especially those in cable joints, could be detected by PD mapping, during either VLF or OW tests.


Case study - 2

Failure Analysis of a 230kV/200MVA

Transformer-Cable Termination


Training Course for Continuous Education

27


Background

A new installation of transformer and cable termination

The failure of yellow phase terminal occurred only 10 days after commissioning

The failure caused an explosion and fire

The transformer/cable terminal box was destroyed

The transformer was significantly damaged

About a quarter of the city was blackout


Case study - 3

Three 230kV Cables Failed After Only 3 Years

Operation - Caused by a Design Problem



28


Background

230kV 2000 mm2 XLPE cable, circuit length 7.2 km

Installed in the middle of 2000 by a consortium of

three manufacturers

Loading was around 40% of rating

F1 failed on 12 September 2003

Only three years new

Serious impacts to customers due to voltage dips

Investigators of OEM insisted that the cable was damaged


Background cont

On 6 June 2004, another cable F3 failed

Again, serious impacts to customers

In June 2004, off-line PD measurement was carried

out on feed 2 (F2)

Large partial discharges (>100pC) were detected and located

A 10m long cable was cut and sectionised

Burnt damages to water swellable tape and semicon screen

were found.

29


Questions

Are the failures due to mechanical damage?

Are they isolated failures?

If not due to cable damages, what are the

possible root causes?

How to prevent the recurrence of the type of

failures


Case study - 4

230kV Cable Joint Failures Due to

Poor Workmanship



30


Background

Failures of two 230kV XLPE cable joints during HV accommissioning tests. The cable and joints were madeby different manufacturers. The cable joints wererubber pre-moulded joints.

Cable Joint A: circuit I, Red phase, joint bay 5/6: PDs were detected under 1.1Uo, PD inception voltage 120kV (0.9 Uo).

In Red phase, circuit II, joint bay 2/3: The joint failed at 27kV (0.2 Uo) during HV ac tests.


Wrong position the gripping shield is shifted out of the semi-conductive

electrode, as shown by the

corresponding mark left on the

internal wall of the rubber moulding

The mark on the internal wall of

EPR rubber moulding

Case Studies on Power Cables

Documents

suthe petroleum instituteabout

suthe petroleum institutehow

time of failure

kv cables

total failure rate

average failure rate

profile of distribution

cable replacements