10/8/2004 1
MEDIUM VOLTAGE CABLE TESTING
New Technologies - New Methods
Michael T. PeschelHigh Voltage, Inc.Copake, NY. USA
www.hvinc.com
10/8/2004 2
This Can Be Prevented
In-service failures cause great damage to faulted cables and adjacent cables. Not so if failed under a VLF test.
10/8/2004 3
Cable Test Methods AC Power Frequency. AC Series & Parallel Resonant. DC Hipot. Very Low Frequency (VLF) AC Hipot. Tan Delta, Partial Discharge, And Other Diagnostic
Methods.
10/8/2004 4
Why Is DC Used DC hipots are small, portable, and economical. DC originally used with PILC, oil insulated cable,
worked well. Same techniques used when HMW, XLPE, EPR, and
other solid dielectric cables were installed. Years later DC was found to be damaging to
insulation and leakage current measurements often ineffective for exposing defects. VLF work started.
Until recently, AC field testing of cable was not practical.
10/8/2004 5
Worldwide consensus exists among engineering organizations, utilities,
and cable manufacturers that typical DC test voltages of 4 5 Vo damage
insulation and lead to failures.
DC Use Discouraged
10/8/2004 6
Avoid DC Voltage On Service Aged, Solid Dielectric Insulation
z DC Voltage Polarizes Cable. Water trees trap space charges lead to
future failures.z Leakage current measurements mean
little - not predictive of cable quality.
10/8/2004 7
Water Treesz Tree shaped channels found within the
insulation of operating cables resulting from the presence of moisture and electrical fields.
z Prevalent in solid dielectric cables.z Eventually leads to the inception of PD.z Leads to insulation failure.
Why is DC Harmful
10/8/2004 8
Water Trees
10/8/2004 9
Why DC Is Damaging
DC hipot input
Negative space charges
AC input
DC hipot output negatively charges up water tree areas.
These trapped space charges remain after test.
When AC is reapplied, theres a high difference of potential across very little of the insulation. Leads to electrical trees cable fails.
10/8/2004 10
Utilities Response To DC Concerns
z Typical DC test voltages were/are 4 - 5 times V0.z At these high voltages, damage is done.z Some reduce voltage to 15 kVDC on 15 kV cable.z 15 kV is only 30% above the peak AC stress in service.z At this voltage, leakage currents are meaningless.z Some do nothing, letting cables fail and then repair.z Many have adopted VLF AC hipoting to expose bad
insulation and accessories. VLF is the best splice checker.
10/8/2004 11
VLF Very Low Frequency AC Advantages:
Stress Similar to Service Conditions. Light Weight. Low Cost. Simple to Use.
Disadvantages: Voltage Waveform in Some Designs. Destructive Cable May Fail Under Test.
What To Do? Use AC Voltage
10/8/2004 12
WHAT IS VLF?Very Low Frequency: 0.1 Hz and lower.
By decreasing the frequency, it is possible to test miles of cable with a small and affordable unit.
Models range from 0.1 0.01 Hz.
A VLF AC hipot is just an AC hipot withan output frequency lower than 60 Hz.
10/8/2004 13
VLF ExplainedXc = 1
2 x pi x f x C
The lower the frequency, the higher Xc (capacitive reactance).
The higher Xc (or resistance across the power supply output),
the lower the current/power needed to apply a desired voltage.
At 0.1 Hz, it takes 600 times less power to test a cable, or any other high capacitance load, than at 60 Hz. At 0.01 Hz, 6000 times
higher capacitive loads can be tested than at 60 Hz.
10/8/2004 14
60 Hz vs. 0.1 Hz
At 60 Hz. a 1 F cable has an Xc of 2.65 kOhms. At 22 kV, it requires 8.3 amps of current to test.
Total power supply rating must be 183 kVA.
At 0.1 Hz, the Xc is 1.59 megohms. At 22 kV, the current needed is 14 mA. Total supply power needed is .304 kVA.
(22 kV is the typical test voltage for 15 kV cable)
10/8/2004 15
50 kVAC @ 3 kVA
Can test ~ 50 of cable
60 Hertz
40 kVAC @ 1.2 kVA
Can test ~ 5 miles of cable
0.1 0.02 Hertz
60 Hz. vs. 0.1 Hz.
10/8/2004 16
Can Now AC Stress Test Cable
z With VLF units, utilities, testing services, industrials, and others can now AC stress test cables in the field.
z Just like with vacuum bottle or rubber glove testing, now a go/no-go test can be performed on power cable.
z If a cable cant hold 2 3 times normal voltage, its not healthy. Find out now, make the repair, and move on.
z At the very least, every newly installed and/or repaired cable should be VLF tested before energizing, since many failures are due to installation damage or faulty workmanship.
10/8/2004 17
The First VLF Models Original marketed designs were European. Very large, heavy, expensive. Thus, limited
practical application. 1997 new VLF technology was introduced.
Smaller - lighter - less expensive. Now used worldwide for cable testing.
10/8/2004 18
VLF with Sinusoidal Waveform Advantages
z Stress Similar to Service Conditions.z Light Weight.z Low Cost.z Simple to Use.z Can Be Used For Diagnostic Testing.
(tan-delta and partial discharge)
10/8/2004 19
VLF Applications Power Cable. Large Rotary Machines. Diagnostic Testing:
z Tan z Partial discharge
10/8/2004 20
North American Standards For VLF Testing Of Cable
IEEE 433-1974 covers VLF testing for rotating machinery.now being updated
IEEE 400-2001 overall cable testing standard sanctions VLF testing of cables.
IEEE 400.2 sub-standard specific to VLF cable testing
10/8/2004 21
Recommends test voltage of 3V0 peak(V0 equals line-to-ground voltage).Test duration is 15 60 minutes.
Recommendation is for 30+ minutes*
* Based on >15,000 tests by TNB Malaysia
IEEE VLF Std 400.2/D3 (11/03)
10/8/2004 22
System Voltage
phase to phase
kVrms
5
15
25
35
Installation
phase to ground
kVrms/kVpeak
9/12
18/25
27/38
39/55
Acceptance
phase to ground
kVrms/kVpeak
10/14
20/28
31/44
44/62
Maintenance
phase to ground
kVrms/kVpeak
7/10
16/22
23/33
33/47
---------------------- 0.1 Hz Test Voltage --------------------
Suggested IEEE Field Test Voltages
For Shielded Power Cable Systems
Using Sine Wave Output VLF
Test voltages are generally 2.5 3 time the line-to-ground system voltage.
The above per IEEE400.2/D3 dated 11/03.
10/8/2004 23
The World View Of VLF
IEEE 2 3Vo for 15-60 minutes30+ minutes @ 3Vo recommended.
Germany 3Vo for 60 minutes
Japan 3Vo for 15 minutes
Malaysia 3Vo for 30 minutes Over 40 countries have purchased the HVI VLF
10/8/2004 24
Cable German Test Voltage
Insulation DIN-VDE 0.1 Hz DC AC: 46-62 Hz
Type Spec. Test Level Duration Test Level Duration Test Level Duration
PVC0276-620
0276-10013 x U0 30 Min. 6/10kV:34-48kV
12/20kV:67-96kV18/30kV:76-108kV
15-30 Min.15-30 Min.15-30 Min.
2 x U0 30 Min.
XLPE0276-620
0276-10013 x U0 60 Min. DC Test was Withdrawn N/A 2 x U0 60 Min.
Paper/Oil0276-621
0276-10013 x U0 30 Min. 6/10kV:34-48kV
12/20kV:67-96kV18/30kV:76-108kV
15-30 Min.15-30 Min.15-30 Min.
2 x U0 30 Min.
VLF Test Duration for mixed paper/PE-XLPE insulated cables should be 60 minutes.Criterion for a successful test is no flashover.
AC VLF test level is 3 x U0 and test duration depends on type of cable insulation.
German Standard
10/8/2004 25
IEC 60060-3
Standard for field testing power cable.
Updated to add VLF Hipot testing.
2 3 times V0 for 15 minutes.
10/8/2004 26
If cable passes the above test, there is a 97% probability that cable will last 3 or 5 years.
2.8/3.0 Uo2.2/2.4 Uo1.8/1.9 Uo
10/8/2004 27
ANSI IEEE Std. 433-1974 (1974)z Recommended Practice For Insulation Testing of Large
AC Rotating Machinery with High Voltage VLF.z Refers to Large AC Machines 10 MVA/6 kV & above.z Test Waveform: Sinusoidal.z Test Frequency: 0.1 Hz.
North American Standards
10/8/2004 28
IEEE/EPRI RecommendationsVLF TEST PROCEDURE
1. Apply 2 3Vo voltage for 30+ minutes.2. If cable holds voltage, cable is good, test is over.3. If cable fails, repair and start test over.4. If cable then holds, cable is good, test is over.5. If cable fails again, repair and resume test or stop
test and replace cable soon, depending on nature of failures. Or, perhaps conduct a tan delta test.
10/8/2004 29
How to VLF Testz Isolate cables ends like with DC testing.z Connect HV lead to conductor ground to shield.z Apply HV to cable in several steps. Not to measure leakage
currents but helpful in determining where cables fail.z There are no leakage currents read. Test is go/no-go z Wait the desired test time.z If cable holds, test is over. De-energize.z If cable fails, make repairs and repeat test.
10/8/2004 30
Growth rate at 0.1-Hzsinusoidal test voltage
(mm/h)2.3
10.9-12.658.3-64.2
336
Test voltage factor(V/Vo )
2345
XLPE Tree Growth RatePer IEEE 400-2001
A 15kV 133% cable has an insulation thickness of 5.9 mm.In a 30 minute test, nearly all defects will grow to failure.
10/8/2004 31
The Move To VLF
Why Utilities Worldwide Have Switched to VLF
To Not Perform Any Cable Testing Is Unacceptable DC Testing Is Out AC Testing Is The Most Desirable
Most Diagnostic Methods Are Inconclusive & Expensive
A VLF Test Is Easily Performed With Results Certain
Worldwide Standards Exist
10/8/2004 32
VLF Vendors
High Voltage, Inc. NY, USA
Baur Austria
Seba/KMT (HDW) Germany
10/8/2004 33
Waveform Outputs
High Voltage, Inc. Sine Wave
Baur Sine Wave
Seba/KMT (HDW) Cosine Rectangular(Trapezoidal)
* Sine wave needed for pd and td diagnostics
10/8/2004 34
Model Sizes Available(all vendors included)
Voltage ratings from 20 kV 200 kV
Load ratings from 0.4 uF 50 uF
For a 15 kV cable, 0.4 uF can test ~4000, 50 uF can test >50 miles.
10/8/2004 35
Users Of VLF
VLF is now embraced worldwide as the most effective method of cable testing.
Over 350 VLF units have been shipped from High Voltage, Inc. over the last four years, delivered to 30+ countries and numerous US locations. Other VLF products have shipped from other vendors
for even more years.
10/8/2004 36
Worldwide Users of VLF
HVI has shipped over 350 VLF products to the following:USACanadaTaiwanIndonesiaSo. KoreaAustraliaSaudi ArabiaEnglandCosta Rica
So. AfricaSingaporeHong KongBelgiumPuerto RicoHollandNew ZealandJapanMalaysia
Slovak RepublicCzech RepublicChinaDubai, UAEVietnamPanamaJordanGermany
10/8/2004 37
In SummaryVLF testing:z Has the virtues of DC test equipment (low cost, small size,
light weight, easy to use) but is AC.z Does not have the negative consequences of DC.z Requires 1/600 1/6000 of the kVA of power frequency.z Can be used for breakdown tests and predictive tests such as
Tan Delta and Partial Discharge. (sine wave VLF units)z Internationally accepted Standards exist and more are in the
process of issuance.
10/8/2004 38
Conclusionz The surest way to verify voltage withstand capability of
any load is perform an AC stress test.z Its easy and certain. The load holds voltage or fails.z VLF testing is easily performed with minimal training.z Worldwide standards and years of history exist.z Some VLF models are very portable and affordable.z VLF makes Tan Delta diagnostic testing possible.
10/8/2004 39
PAUSE
10/8/2004 40
Dissipation Factor or Loss Angle Measurement For Power Cables
Non Destructive Testing To Determine Insulation Quality
(Similar to Power Factor Testing)
Tangent Delta or Tan
41
Tan Delta MeasurementUsing VLF @ 0.1 Hz
z Excellent predictive tool for determining the integrity of insulation z Absolute values, variations, and trending of values are of interest for
predicting insulation integrityz Evaluates over all condition of cable (rather than localized problems
as with PD measurement) z Tan Delta is more easily measured at VLF (magnitude increases as
frequency decreases)z Requires VLF sinusoidal applied test voltagez Excellent way to evaluate Water Trees
42
Simplified Cable Model and Phasor DrawingTan Delta = IR/ IC - measured in radians
With perfect insulation, a cable is a near perfect capacitor, with a 90 phase shift between voltage and current. Less than 90 indicates insulation degradation. Cables can be rated good, marginal, or bad.
The tangent of this angle is calculated
I
V
IC
IR
IR IC
= tangent of C R
Cable insulation Cable Cross Section
43
Characteristics of Water Trees
z Addition of a new parallel R- C component
z Water trees increase resistive current through insulation
z The R component is voltage dependent (nonlinear I = V/R)
z Tan Delta becomes a function of voltage
10/8/2004 44
Are There Established TD Numbers?(Table from IEEE400-2001)
Above numbers are at 20C. Numbers vary with temperature. Consult your cable vendor.
10/8/2004 45
What do the numbers mean?
Absolute and differential values are meaningful.
10/8/2004 46
Tan Delta Test ProcedureMeasurements are recorded at
numerous voltage levels. Loss angle versus voltage important.
Recommended voltage test points are:
0.5 Uo, .75 Uo, 1.0 Uo, 1.25 Uo,
1.5 Uo, 1.75 Uo & 2 Uo.
47
Tan Delta vs.Voltage for New and Aged XLPE Cables
New and Aged 15 kV XLPE Cable (Nov 2000)
0
0.01
0.02
0.03
0.04
0.05
0.06
0 2.5 5 7.5 10VLF Voltage (kV rms)
L
o
s
s
A
n
g
l
e
(
T
a
n
D
e
l
t
a
)
Aged
New Cable
New cablelinear tan delta #s
versus voltage
Aged cablenon-linear tan
delta #s versus voltage
10/8/2004 48
0.1 Hz phase 1 phase 2 phase 3 phase 1 phase 2 phase 3x Vo tgdelta L1 tgdelta L2 tgdelta L30.5 0.0018 0.0008 0.0013 (2Uo-Uo) (2Uo-Uo) (2Uo-Uo)1.0 0.0019 0.0009 0.0014 0.0035 0.0018 0.00251.5 0.0026 0.0011 0.0019 0.0019 0.0009 0.00142.0 0.0035 0.0018 0.00252.5 0.0044 0.0026 0.0032 0.0016 0.0009 0.0011
0.0000.0020.0040.006
0.0 1.0 2.0 3.0Voltage x Vo
t
a
n
d
e
l
t
a
L2: tg delta 0.1Hz L3: tg delta 0,1HzL1: tg delta 0.1Hz
10/8/2004 49
TD at 0.1 Hz vs. 60 Hz.
Why Do We Use VLF?TD numbers higher at 0.1 Hz versus 60 Hz
Equipment far smaller, lighter, and less expensive, VLF can test longer cable runs than 60 Hz systems.
VLF may be more effective for TD measurements of unheated cable (most test conditions).
10/8/2004 50
10/8/2004 51
Tan Delta Loss Analyzer
HV divider Signal analyzer/controller
60kV model
Used with VLF as voltage source
10/8/2004 52
Loss AngleAnalyzer
VLF Control
VLF
M
e
a
s
u
r
i
n
g
U
n
i
t
Cable Under Test
Fiber Optic Cables
Tan Delta
Test Set Up
10/8/2004 53
Tan Delta TestingAdvantagesz Less destructive than VLF or 60 Hz testingz Allows prioritization of cable replacementz Easier to use and interpret than other diagnostic methods
Disadvantagesz Can be destructive if cable very degraded.z Gives overall condition of cable, not singularitiesz Not best for mixed type cable runsz More useful with historical data
10/8/2004 54
Other Diagnostic Methods
z Partial Discharge On-Linez Partial Discharge Off-Linez Current Relaxationz Voltage Recoveryz Oscillating Wave Test (Partial Discharge)
10/8/2004 55
Explanation of Methods
z Except Pd testing, methods are based on predictable properties of capacitance
z Voltage/Current charging & discharging is predictable in a perfect capacitor
z Imperfections in a cable alter this normz Measurements can determine level of cable
insulation impuritiesz For partial discharge testing, location and
amplitude of pd can be measured
10/8/2004 56
Advantages/Disadvantages
Advantagesz Non destructive (although not in all cases)
Disadvantagesz Results highly interpretive and often unrepeatablez Most methods are very expensivez Some methods unproven for accuracy in the fieldz Require skilled operatorz Insufficient data of acceptable/unacceptable Pd levels
and other parameters to make test results conclusive
10/8/2004 57
Utilities and others need a method that is:
Affordable
Portable
Easy to use
Offers conclusive results
10/8/2004 58
Conclusion
VLF is the most effective method of exposing cable and accessory defects. With the addition of Tan Delta and Pd testing, hipoting and diagnostics are achieved.
VLF is suitable for use on cables and rotating machinery.
Worldwide standards exist for both. IEEE, VDE, CEA, EPRI, other countrys engineering organizations, cable manufactures, and major utilities all recommend VLF.
Hundreds of users worldwide have embraced VLF
10/8/2004 59
Thank You
Mike Peschel - High Voltage, Inc.
www.hvinc.com
10/8/2004 60
10/8/2004 61
10/8/2004 62
VLF Cable Check
Model VLF-25CM
0 - 25 kV Peak0.1 Hz @ 0.4 f Load
~ 4000 ft of 15 kV CableSingle Piece Design
Weight: 69 lbs/30 kgONLY
10/8/2004 63
VLF AC Hipot
Model VLF-4022CM
0 - 40 kV Peak
0.1 Hz @ 1.1 f Load0.05 Hz @ 2.2 f Load0.02 Hz @ 5.5 f Load
Controls: 50 lbs/22 kgHV Tank: 72 lbs/31 kg
10/8/2004 64
VLF AC Hipot
Model VLF-6022CM
0 - 60 kV Peak
0.1 Hz @ 1.1 f Load0.05 Hz @ 2.2 f Load0.02 Hz @ 5.5 f Load
Controls: 55 lbs/25 kgHV Tank: 100 lbs/45 kg
10/8/2004 65
VLF AC Hipot
Model VLF-12011CM
0 - 120 kV Peak
0.1 Hz @ .55 f Load0.05 Hz @ 1.1 f Load0.02 Hz @ 2.75 f Load0.01 Hz @ 5.5 f Load
Weight: 565 lbs/257 kg
10/8/2004 66
Other VLF ModelsVLF-50CM 0 50 kV @ 50 F @ 0.01Hz 0.1 0.01 Hz
Tests up to 50 kilometers of cable.
VLF-65CM 0 65 kV @ 2.2 F @ 0.1Hz 0.1 0.01 Hz0 65 kV @ 22 F @ 0.01Hz
VLF-200CM 0 200 kV @ 0.55 F @ 0.1Hz 0.1 0.01 Hz0 200 kV @ 5.5 F @ 0.1Hz
10/8/2004 67
VLF THUMPERURD Cable Care System
Only Combination VLF AC Hipot and Thumper
VLF Output: 30 kVAC peak
Load Rating: 1.0 uF @ 0.1 Hz ~ 1.5 miles of 15 kV cable
Discharge: 0 8 kV @ 500 J
VLF Cable Burning Mode
Radar Ready
10/8/2004 68
Top Related