Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
10296123
10296123
Aging Power Cable Maintenance
Guideline
Medium-Voltage Cables, 5 to 35 kV
1024044
10296123
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION NAMED BELOW AS AN ACCOUNT
OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH
INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE
ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:
(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)
WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR
SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR
INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL
PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER
(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR
SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
REFERENCE HEREIN TO ANY SPECIFIC COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY ITS
TRADE NAME, TRADEMARK, MANUFACTURER, OR OTHERWISE, DOES NOT NECESSARILY
CONSTITUTE OR IMPLY ITS ENDORSEMENT, RECOMMENDATION, OR FAVORING BY EPRI.
THE FOLLOWING ORGANIZATION PREPARED THIS REPORT:
Electric Power Research Institute (EPRI)
NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or
e-mail askepri@epri.com.
Electric Power Research Institute, EPRI, and TOGETHERSHAPING THE FUTURE OF ELECTRICITY
are registered service marks of the Electric Power Research Institute, Inc.
Copyright © 2012 Electric Power Research Institute, Inc. All rights reserved.
10296123
ACKNOWLEDGMENTS
This publication is a corporate document that should be cited in the literature in the following
manner:
Aging Power Cable Maintenance Guideline: Medium-Voltage Cables, 5 to 35 kV. EPRI, Palo
Alto, CA: 2012. 1024044.
iii
10296123
10296123
PRODUCT DESCRIPTION
Medium-voltage cables (5- to 35-kV rated cables) have provided reasonable service in power
plants. However, there is a concern that cables that have experienced long periods of wet
service might degrade and fail in service. Because most plants have had few problems with
medium-voltage cables, little on-staff experience with medium-voltage cables exists at most
sites. This report provides information that will be of practical use when questions concerning
medium-voltage cable longevity in adverse environments and service conditions arise at fossil-
fired plants.
Objectives
This report is meant for persons on plant staffs who are responsible for cable system
maintenance, operation, and design. This report describes the currently available information
concerning the types of cables in service in plants and the best way to address aging and
monitoring concerns.
Approach
This report is unique in that it focuses on the cable types used in the power industry and the
conditions that challenge them. The report supports the needs of plant staffs, allowing them to
understand the specifics of their cable systems without being confused by the much broader
range of cable information associated with power distribution systems that do not apply to
the plants.
The report was developed by consultants with extensive experience in cable design,
manufacturing, installation, troubleshooting, and replacement. They were guided by industry
task group meetings that helped maintain focus on the issues facing plant cable personnel. This
revision provides additional information and formalizes the report.
Results
The report provides information on cable system design, cable construction, insulation systems
and their aging characteristics, condition assessment, cable installation, and preparation for repair
and replacement. The report provides practical information to help choose the correct assessment
tests to apply and to help understand critical issues about the selection and installation of cables
and terminations that affect cable longevity.
Applications, Value, and Use
This report provides a wide range of information that can affect the aging of cables and the
management of that aging. Information is provided on grounding system designs that, if not
understood, could lead to the misapplication of cable types or increase the likelihood of a
potentially devastating phase-to-phase fault that could cause failures of transformers and buses
v
10296123
in addition to the cable. Guidance is provided on the selection of an appropriate test method,
depending on the cable design, and how the cable is expected to degrade with time. Effects of
aging of the insulation and metallic shield system must be considered to allow appropriate
selection of the test method.
Keywords
Aging
Cable insulation assessment
Cable systems
Cross-linked polyethylene (XLPE)
Ethylene-propylene rubber (EPR)
Medium-voltage cable
vi
10296123
CONTENTS
vii
10296123
3.2.4 Adverse Environments ........................................................................................ 3-5
3.2.4.1 Temperature-Related Aging .......................................................................... 3-5
3.2.4.2 Radiation-Related Aging ............................................................................... 3-5
3.2.4.3 High Conductor Temperature from Ohmic Heating ....................................... 3-6
3.2.4.4 High-Resistance Connections ....................................................................... 3-6
3.2.5 Surface Corona and Partial Discharge................................................................ 3-6
viii
10296123
5.6 Installation Considerations ......................................................................................... 5-18
5.6.1 Connecting the Conductors............................................................................... 5-18
5.6.2 Insulation for Splices ......................................................................................... 5-19
5.6.3 Semiconducting Insulation Shield Materials for Splices and Terminations ....... 5-19
5.6.4 Metallic Insulation Shield for Splices and Terminations .................................... 5-19
5.6.5 Jackets for Splices ............................................................................................ 5-20
5.7 Selection of Splices and Terminations ....................................................................... 5-20
ix
10296123
7 AGING AND DEGRADATION OF BUTYL, ETHYLENE-PROPYLENE RUBBER, AND
CROSS-LINKED POLYETHYLENE CABLES DUE TO ADVERSE ENVIRONMENTS ........... 7-1
7.1 Aging and Degradation of Butyl Rubber ...................................................................... 7-1
7.1.1 General ............................................................................................................... 7-1
7.1.2 Water-Related Degradation of Butyl Rubber....................................................... 7-1
7.1.3 Thermal Degradation of Butyl Rubber................................................................. 7-2
7.1.4 Radiation Degradation of Butyl Rubber............................................................... 7-2
7.1.5 Conclusions......................................................................................................... 7-3
7.2 Aging and Degradation of Ethylene-Propylene Rubber ............................................... 7-3
7.2.1 General ............................................................................................................... 7-3
7.2.2 Water-Related Degradation of Ethylene-Propylene Rubber ............................... 7-4
7.2.3 Thermal Degradation of Ethylene-Propylene Rubber ......................................... 7-6
7.2.4 Radiation Degradation of Ethylene-Propylene Rubber ....................................... 7-9
7.2.5 Conclusions......................................................................................................... 7-9
7.3 Aging and Degradation of Cross-Linked Polyethylene ................................................ 7-9
7.3.1 General ............................................................................................................... 7-9
7.3.2 Water-Related Degradation of Cross-Linked Polyethylene............................... 7-10
7.3.3 Thermal Degradation of Cross-Linked Polyethylene......................................... 7-11
7.3.4 Radiation Degradation of Cross-Linked Polyethylene....................................... 7-11
7.3.5 Conclusions....................................................................................................... 7-11
7.4 Other Degradation Causes ........................................................................................ 7-11
7.4.1 General ............................................................................................................. 7-11
7.4.2 Corona Discharge ............................................................................................. 7-11
7.4.3 Partial Discharge ............................................................................................... 7-12
x
10296123
8.4 Acceptance Tests ...................................................................................................... 8-12
8.5 Installation Tests ........................................................................................................ 8-12
8.6 Maintenance and In-Service Testing ......................................................................... 8-13
8.6.1 Introduction ....................................................................................................... 8-13
8.6.1.1 Historical Perspective on In-Service Testing ............................................... 8-14
8.6.1.2 Withstand Versus Diagnostic Testing.......................................................... 8-15
8.6.1.3 Global Versus Local Assessment................................................................ 8-15
8.6.1.4 60 Hz Versus Other Frequencies ................................................................ 8-15
8.6.2 Off-Line Diagnostic and Withstand Testing ....................................................... 8-16
8.6.2.1 Dissipation Factor (Tan δ) Testing .............................................................. 8-16
8.6.2.2 Dielectric Spectroscopy ............................................................................... 8-19
8.6.2.3 Off-Line Partial Discharge Measurement .................................................... 8-21
8.6.2.4 AC Withstand Testing.................................................................................. 8-25
8.6.2.5 Tests Under Development........................................................................... 8-27
8.6.3 On-Line Diagnostics Assessment ..................................................................... 8-27
8.7 Applicability of Tests .................................................................................................. 8-31
xi
10296123
10 RESPONSE TO CABLE FAILURES .................................................................................. 10-1
10.1 Electrical Tests of Circuit....................................................................................... 10-1
10.1.1 Logic for Testing to Be Performed ................................................................. 10-1
10.1.2 Logistical Issues for Electrical Testing ........................................................... 10-2
10.2 Fault Location........................................................................................................ 10-2
10.3 Forensics and Failure Assessments ..................................................................... 10-3
10.4 Repair Options ...................................................................................................... 10-6
10.4.1 Replacement of Failed Section ...................................................................... 10-6
10.4.2 Total Replacement ......................................................................................... 10-7
10.4.3 Acceptance Testing ....................................................................................... 10-7
xii
10296123
B RESULTS OF NUCLEAR ENERGY INSTITUTE SURVEY AND NUCLEAR
REGULATORY COMMISSION GENERIC LETTER 2007-01 .................................................. B-1
B.1 Nuclear Energy Institute Underground Medium-Voltage Cable Survey ...................... B-1
B.1.1 Survey Purpose.................................................................................................. B-1
B.1.2 Survey Scope ..................................................................................................... B-1
B.2 Survey Results Evaluation .......................................................................................... B-2
B.2.1 Contributors........................................................................................................ B-2
B.2.2 Underground Circuit Quantities .......................................................................... B-2
B.2.3 Installed Cable Types......................................................................................... B-2
B.2.4 Shielding ............................................................................................................ B-5
B.2.4.1 Underground Wet-Duty Failure Assessment ............................................... B-5
B.3 Results of Utility Responses to Generic Letter 2007-01 ........................................... B-13
B.3.1 Summary of Results ......................................................................................... B-14
B.3.2 Assessment Methods ....................................................................................... B-16
B.4 Comparison of the Nuclear Energy Institute Survey and Generic Letter
2007-01 Results ....................................................................................................... B-17
xiii
10296123
E OFF-LINE TESTS THAT ARE UNDER DEVELOPMENT .................................................... E-1
E.1 Introduction ................................................................................................................. E-1
E.2 Isothermal Return Current .......................................................................................... E-1
E.3 Return Voltage ............................................................................................................ E-2
E.4 Oscillating Wave ......................................................................................................... E-3
xiv
10296123
LIST OF FIGURES
xv
10296123
Figure 5-14 Another Layer of Semiconducting Tape Is Applied over the Insulating Tape ......... 5-9
Figure 5-15 A Metallic Braid Is Installed, a Ground Strap Is Attached, and the Jacket
Tape Is Installed over the Entire Splice Area ................................................................... 5-10
Figure 5-16 Cutaway of a Premolded Splice ........................................................................... 5-11
Figure 5-17 Cold-Shrink Splice, Showing the Direction of the Shrinking Process ................... 5-12
Figure 5-18 Carefully Position the Housing Before Removing the White Core Support .......... 5-12
Figure 5-19 Continue Removing the Core While Holding Its Position on the Cable ................ 5-13
Figure 5-20 High-Permittivity Mastic Material Is Placed over the Connector and
Conductor, Without Concern for Smoothness .................................................................. 5-14
Figure 5-21 The Semiconducting Tube Is Slid into Place and Heated Until Properly
Shrunk Down to the Cable ............................................................................................... 5-14
Figure 5-22 An Insulation Tube Is Slid into Place and Shrunk Down ...................................... 5-14
Figure 5-23 A Tube That Is Both Insulating on the Inside and Semiconducting on the
Outside Is Positioned and Shrunk into Place ................................................................... 5-14
Figure 5-24 Tinned Copper Braid is Wrapped Around the Splice to Replace the Metallic
Portion of the Insulation Shield ........................................................................................ 5-14
Figure 5-25 A Ground Strap Spring Is Placed Under One Side of the Cable’s Taped
Metallic Insulation Shields ................................................................................................ 5-15
Figure 5-26 The Ground Strap Is Placed Across the Splice and Connected to the Factory
Metallic Shield on the Opposite Side of the Splice ........................................................... 5-15
Figure 5-27 An Overall Rejacketing Tube Is Placed Around the Entire Area .......................... 5-15
Figure 5-28 Cold-Shrink Outdoor Termination with Rain Sheds or Skirts ................................ 5-16
Figure 5-29 Dead Break, Premolded Termination ................................................................... 5-17
Figure 5-30 Stub-Type Motor Connection ................................................................................ 5-17
Figure 5-31 In-Line Type Connection ...................................................................................... 5-17
Figure 5-32 Improperly Crimped Connector ............................................................................ 5-19
Figure 6-1 Butyl Rubber Molecule ............................................................................................. 6-4
Figure 6-2 Copolymer of Ethylene and Propylene ..................................................................... 6-6
Figure 6-3 Banbury Mixer Used for Preparing Ethylene-Propylene Rubber Compounds .......... 6-8
Figure 6-4 Various Representations of Crystallinity in Polyethylene ....................................... 6-14
Figure 7-1 Ethylene-Propylene Rubber Cable Failures as Function of Color of the
Ethylene-Propylene Rubber Insulations ............................................................................. 7-5
Figure 7-2 Generational Differences in Life Expectancy for Ethylene-Propylene Rubber
and Other Insulations ......................................................................................................... 7-6
Figure 7-3 AC Breakdown Strength of Medium-Voltage, Field-Aged Ethylene-Propylene
Rubber Insulated Cables at V0 ........................................................................................... 7-8
Figure 7-4 AC Voltage Breakdown Strength of Combined Ethylene-Propylene Rubber
Cables Aged in the Laboratory and in Service at Rated Voltage ....................................... 7-8
Figure 7-5 Unfilled and Mineral-Filled Cross-Linked Polyethylene Cable Failures
Compared to All Failures .................................................................................................. 7-10
Figure 8-1 Wafer Examination for Voids, Inclusions, and Conductor Shield Protrusions .......... 8-6
Figure 8-2 Hot Oil Test............................................................................................................... 8-6
xvi
10296123
Figure 8-3 Vertical Tray Flame Test .......................................................................................... 8-7
Figure 8-4 An Acceptable Partial Discharge Plot ....................................................................... 8-8
Figure 8-5 An Unacceptable Partial Discharge Plot ................................................................... 8-8
Figure 8-6 Partial Discharge Requirement History from AEIC 5 ................................................ 8-9
Figure 8-7 Dry Specimen Design Qualification Tests .............................................................. 8-10
Figure 8-8 Wet Specimen Design Qualification Tests ............................................................. 8-11
Figure 8-9 Derivation of Dissipation Factor (Tan δ) Measurement in Insulation ...................... 8-16
Figure 8-10 Typical Variable-Frequency, Very-Low-Frequency Portable Test Equipment
for Performing 0.1-Hz Dissipation Factor (Tan δ) Testing ................................................ 8-17
Figure 8-11 Voltage Dependence of Dissipation Factor for New and Aged Cross-Linked
Polyethylene Cable .......................................................................................................... 8-18
Figure 8-12 0.1-Hz Dissipation Factor of Cross-Linked Polyethylene-Insulated Cables.......... 8-18
Figure 8-13 Dielectric Spectroscopy Measurements for Shelf-Aged 15-kV Cross-Linked
Polyethylene Cable .......................................................................................................... 8-20
Figure 8-14 Calibration Equipment for Off-Line Partial Discharge Testing .............................. 8-21
Figure 8-15 Test Setup for Performing 60-Hz Partial Discharge Measurements ..................... 8-22
Figure 8-16 Test Profile for Short Duration Off-Line 60-Hz Partial Discharge Testing............. 8-23
Figure 8-17 60-Hz Partial Discharge One-Step Diagnostic Data Capture Approach ............... 8-23
Figure 8-18 Nominal Very-Low-Frequency (0.1-Hz) Sinusoidal Waveform ............................. 8-25
Figure 8-19 Trapezoidal (Bipolar Rectangular) Very-Low-Frequency (0.1-Hz) Waveform ...... 8-26
Figure 8-20 Sensors for On-Line Signal Detection and Data Acquisition System ................... 8-29
Figure 8-21 On-Line Condition Assessment Data Acquisition System .................................... 8-29
Figure 8-22 Influence of Defect Type on Signal Patterns Detected During On-Line
Evaluation ........................................................................................................................ 8-30
Figure A-1 External Cable Condition at the Location of the Fault ............................................. A-2
Figure A-2 Condition of the Shield After the Jacket and Burn Hole Were Removed from
the Fault ............................................................................................................................ A-2
Figure A-3 Damage to the Conductor from the Fault ................................................................ A-3
Figure A-4 Inner Surface of the Insulation with Severe Indentation from the
Semiconducting Tape ....................................................................................................... A-4
Figure A-5 Zinc Shield .............................................................................................................. A-6
Figure A-6 Tracking Path and Failure Location on Interior of Jacket ........................................ A-6
Figure A-7 Tracking Path and Failure Site on Insulation .......................................................... A-7
Figure A-8 Damage to Conductor from Fault ............................................................................ A-7
Figure A-9 Transformer Connection of One of the Cables ....................................................... A-8
Figure A-10 Circuit Breaker Connection of One of the Cables ................................................. A-9
Figure A-11 Cable Configuration in the Manhole Adjacent to the Transformer ........................ A-9
Figure A-12 Location of Failed, Faulted, and At-Risk Cables Within Duct Leading
to Building ....................................................................................................................... A-10
Figure A-13 Duct Sealing and Water Leaking from Failed Conductor B-5 ............................. A-11
xvii
10296123
Figure A-14 Cables A-5, B-5, and C-5 .................................................................................... A-12
Figure A-15 Close-Up of B-5 Cable Failure ............................................................................ A-12
Figure A-16 Comparison of Conductor Corrosion ................................................................... A-13
Figure A-17 Marked Low-Resistance Channel in the Insulation with Breakdown Hole
in Center .......................................................................................................................... A-13
Figure A-18 Micrograph of Insulation Wall at Low-Resistance Channel, Showing
Swelling and Fissures ..................................................................................................... A-14
Figure A-19 Temporary, Aboveground Cable System ............................................................ A-15
Figure A-20 Permanent Aerial Structure Routing on Second Unit .......................................... A-15
Figure A-21 Damage to the Copper Shield at the Damage Site ............................................. A-18
Figure A-22 Cross Section of Insulation at the Fault Tube, Showing Carbonization of the
Wall of the Fault Tube ..................................................................................................... A-19
Figure A-23 Large Water Tree Viewed Through Hot Oil Bath ................................................ A-20
Figure A-24 Cross Section of the Water Tree Shown in Figure A-23 ..................................... A-20
Figure A-25 Embedded Particle at Base of Water Tree Shown in Figures A-23 and A-24 ..... A-21
Figure A-26 As-Found Condition of Terminations ................................................................... A-22
Figure A-27 Burnthrough of Stress Relief Adaptor ................................................................. A-23
Figure A-28 Burned-Through Insulation Found After Removal of the Stress Relief
Adaptor ............................................................................................................................ A-23
Figure B-1 Manufacturers of 5-kV Cable .................................................................................. B-3
Figure B-2 Age Distribution of Units with No Failures ............................................................... B-6
Figure B-3 Age Distribution of Wet Cable Failures for All Insulation Types .............................. B-7
Figure B-4 Age Distribution of Wet Cross-Linked Polyethylene Cable Failures ....................... B-8
Figure B-5 Age Distribution of Wet Ethylene-Propylene Rubber Cable Failures ...................... B-9
Figure B-6 Age Distribution of Butyl Rubber Cable Failures ................................................... B-10
Figure B-7 Age of Wet Cable at Failure Versus Year of Failure ............................................. B-11
Figure B-8 Age at Time of Failure for Wet Ethylene-Propylene Rubber Cables ..................... B-12
Figure B-9 Age at Time of Failure for Wet Cross-Linked Polyethylene Cables ...................... B-13
Figure B-10 Inaccessible Cable Failures by Plant, from Responses to Generic Letter
2007-01 ........................................................................................................................... B-14
Figure B-11 Number of Failures of Wet Cable by Year .......................................................... B-15
Figure B-12 Wet Medium-Voltage Cable Failures Versus Age at Time of Failure .................. B-15
Figure B-13 Age of Cable at Time of Failure By Insulation and Design .................................. B-16
Figure C-1 AC Breakdown Strength Versus 60-Hz Tan δ Results............................................ C-2
Figure C-2 60-Hz Tan δ Results by Test Voltage ..................................................................... C-4
Figure C-3 Very-Low-Frequency Tan δ Result for a 5-kV Okonite Black Ethylene-
Propylene Rubber Cable ................................................................................................... C-5
Figure C-4 Very-Low-Frequency Tan δ Results for 15-kV Okonite Black Ethylene-
Propylene Rubber Cable ................................................................................................... C-6
Figure C-5 60-Hz Tan δ for Anaconda UniShield Pink Ethylene-Propylene Rubber ................ C-7
xviii
10296123
Figure C-6 Laboratory-Measured Very-Low-Frequency Tan δ for Anaconda Pink
Ethylene-Propylene Rubber UniShield .............................................................................. C-7
Figure C-7 Tan δ Results for Anaconda Pink UniShield Cables (Ordered by 8-kV result) ....... C-8
Figure C-8 Tan δ Results for Anaconda Pink UniShield Cables (Ordered by Difference in
16 kV and 8kV Results) ..................................................................................................... C-9
Figure C-9 Tan δ Results for Anaconda Pink UniShield Cables (Showing Moderate to
High Results) ................................................................................................................... C-10
Figure C-10 Tan δ Measurements for Anaconda UniShield Cables Taken One Year
Apart ................................................................................................................................ C-11
Figure C-11 Tan δ Measurements for Anaconda UniShield with Black Insulation .................. C-12
Figure C-12 Tan δ Versus Breakdown Strength for Okonex Butyl Rubber ............................. C-13
Figure C-13 Comparison of 60-Hz Tan δ Versus Breakdown Strength of Okonite Butyl
Rubber to Black Ethylene-Propylene Rubber ................................................................. C-14
Figure C-14 AC Breakdown Voltage (in V/mil) for Okonite Butyl and Black Ethylene-
Propylene Rubber Insulations ......................................................................................... C-15
Figure C-15 Tan δ Results for Mixed Shielded and Nonshielded Black Okonite Ethylene-
Propylene Rubber Circuits .............................................................................................. C-17
Figure C-16 Tan δ Results for Mixed Shielded and Nonshielded Black Okonite Ethylene-
Propylene Rubber Circuits .............................................................................................. C-18
Figure C-17 Average Tan δ Value Versus Percent of Nonshielded Cable Section ................ C-19
Figure D-1 A Cross-Linked Elastomer ...................................................................................... D-2
Figure D-2 Ethylene-Propylene-Diene Monomer Terpolymer ................................................... D-4
Figure E-1 Isothermal Return Current Plots for Cross-Linked Polyethylene Cable Having
Undergone Different Degrees of Aging ............................................................................. E-1
xix
10296123
10296123
LIST OF TABLES
xxi
10296123
10296123
1
INTRODUCTION
This report is based on the research performed for the Nuclear Sector and its membership.
GenMAC acknowledges their work as comprising the content of this product and its
applicability to the Generation Sector. It is our intent to partner where ever possible on this
important research to avoid duplication of work or seek to overload limited resources providing
this critical information.
The medium-voltage cable system (cable rated from 5 kV to 35 kV, for the purpose of this
report) is composed of cable, terminations, splices, trays, conduits, ducts, and, in some cases,
trenches. Vertical support systems, manholes, fire stops, and water drainage systems also affect
the cable. All of these play a role in the longevity of the cable system.
This report provides a wide range of information in support of medium-voltage cable aging
management. Information is provided on topics such as electrical system grounding practices and
operational issues because they affect failure mechanisms, especially in the final stages of
failure. For example, some safety-related distribution systems are ungrounded or have high-
resistance grounds so that, during accident conditions, they could sustain a phase-to-ground fault
and continue to operate for some period of time. However, under normal operating conditions, a
cable with a suspected or identified ground fault should be de-energized as soon as possible to
prevent the fault from converting to a phase-to-phase fault, which is much more damaging. A
phase-to-ground fault in an ungrounded or high-resistance ground system is likely to generate
heat or increase voltage stress across the adjacent phase insulation so that deterioration of the
insulation of a second phase in a relatively short period of time is possible. After the fault
converts to a phase-to-phase fault, extremely high currents result, causing severe damage to the
cable and connected equipment. Understanding the need for de-energizing cables having a
phase-to-ground fault, even though the system is designed to withstand the condition for a short
period, is important to limiting damage and stress to the remainder of the system.
Topics included in this report related to cable aging management include the following:
The grounding system design, which determines the way in which ground currents and
transient voltages affect cables under fault conditions and affects operating practices at the
time of faults
Types of cables and cable designs, including nonshielded and shielded, and the effects of
different cable designs on aging and testability
The physical layout of the cable and the various service conditions and stresses imparted to
the cable beyond operating voltage and current
Cable design, aging mechanisms, and failure mechanisms
Splice and termination issues that can lead to early failure
1-1
10296123
Introduction
Cable insulation materials and changes and improvements that have been made to them since
the 1970s
Cable condition assessment tests
This report is an update of the Electric Power Research Institute (EPRI) report Medium Voltage
Cable Aging Management Guide (1016689), which was published in 2008 [1]. This update
reflects the consensus guidance on aging management of medium-voltage power cables
developed in 2009–2010, as well as other developments since the initial report was published.
Due to industry attention to submergence of cable, much of the report concerns aging of
medium-voltage cables under wet energized conditions. However, cables and terminations can
fail under dry conditions if defects exist in the insulation, if damage or installation errors occur,
or if the cable is affected by an adverse local environment. Appendix A provides insights into
some of the conditions that have resulted in failures under both wet and dry conditions.
1-2
10296123
Introduction
1-3
10296123
Introduction
1-4
10296123
Introduction
1.5 Terminology
Hydrophobic. A tendency of a polymer to reject and not to absorb or react with water.
Lossy. Referring to insulation having a somewhat higher leakage current. XLPE is a low-loss
material, whereas EPR is lossy.
1-5
10296123
10296123
2
UNDERSTANDING THE DESIGN OF POWER PLANT
CABLE SYSTEMS
Understanding the design of the cables and the cable system is important because it can influence
the rate of degradation and the ability to detect that degradation. This section describes cable
shielding, circuit grounding philosophy, cable ground insulation, and cable phase configuration
choices that can influence or affect cable aging and testability. Protective relays and annunciation
alarms, as well as actions that be taken in response to their actuation, are described.
2-1
10296123
Understanding the Design of Power Plant Cable Systems
With a shielded cable, a multi-grounded metallic shield eliminates any voltage or tracking on
the cable’s exterior surface. An additional safety advantage is developed because there is no
surface voltage present, and workers can come into close proximity to energized cables, such
as in manholes; however, contacting medium-voltage cables, whether shielded or
nonshielded, is not recommended while they are energized.
Shielded cables are more complex to splice and terminate.
Two shields exist in a majority of the medium-voltage cables used in power plants today (see
Figure 2-1). The shielding material next to the conductor is called the conductor shield. Medium-
voltage cables used in more recent plants have conductor shields. The conductor shield prevents
discharges between the conductor and the insulation during operation. The shield that is of
concern with regard to shielded and nonshielded cable is the insulation shield, which consists of
a semiconducting layer in contact with the outer portion of the insulation, as well as a metallic
portion on top of the semiconducting layer. The metallic portion must be grounded in at least one
point in every circuit for the shield to contain the voltage stress within the insulation system. If
the shield is not grounded, the cable will function just like a nonshielded cable and will have up
to 90% of conductor voltage on the cable’s outer surface.
Figure 2-1
Shielded Cable Components
2-2
10296123
Understanding the Design of Power Plant Cable Systems
2-3
10296123
Understanding the Design of Power Plant Cable Systems
Faults on power plant cables are expected to clear immediately by automatic means or within a
short, finite time. Care must be taken that operators understand that if a cable shows signs of a
phase-to-ground fault, such as a ground alarm or popping noises from a manhole or cubicle, the
cable is de-energized as rapidly as possible.
2-4
10296123
Understanding the Design of Power Plant Cable Systems
Figure 2-2
Single-Point Grounding of a Shield, Showing Voltage of Shield to Ground as a Function of
Distance
However, with single-point grounding, a voltage builds up on the shield as a function of the
current in the central conductor and the distance from the grounding point (see Figure 2-3).
Figure 2-3
Shield Losses and Voltages for Single-Conductor Cables [8]
2-5
10296123
Understanding the Design of Power Plant Cable Systems
This voltage can be as much as 100 V on a heavily loaded circuit and becomes a concern for the
integrity of the jacket as well as a concern for workers that are not aware of the voltages and
current that might be present on the shield. When current flows in the conductor of the cable, that
current produces electromagnetic flux in the metallic shield, if present, or in any parallel
conductor. This becomes a one-turn transformer when the shield is grounded two or more times,
because a circuit is formed and current flows when a person touches the cable surface at the
ungrounded end of the circuit. The cable circuits that should be considered for single-point
grounding are systems with conductors of 1000 kcmil (507 mm2) and larger and anticipated
loads of more than 500 A. The EPRI report Power Plant Electrical Reference Series, Volume 4:
Wire and Cable (EL-5036) should be reviewed before designing a single-point grounding
scheme [8].
In addition, if a discontinuity occurs in the metallic portion of the insulation shield that is
grounded at only one end, a discharge site is created that can lead to carbonization and tracking
at the point of the break leading to insulation failure. One such example is described in
Section A.2.3.
If the shield is grounded two or more times or otherwise completes a circuit, the magnetic flux
produces a current flow in the shield. The amount of current in the shield is inversely
proportional to the resistance of the shield. The voltage on the jacket surface stays at zero if two
or more grounds are connected but current is being carried in the shield. This is the reason that it
is not safe to cut a shield in half while the cable is energized, because the two open ends become
energized.
The distance between the grounds has no effect on the magnitude of the current. Whether the
grounds are 1 ft or 1000 ft (0.3 m or 305 m) apart, the current is the same—depending on the
current in the central conductor and the resistance and impedance of the shield. In the case of
multiple cables, the spatial relationship of the cables is also a factor. Figure 2-2 provides
equations for calculating shield currents and induced voltages.
2-6
10296123
Understanding the Design of Power Plant Cable Systems
2.3.1 Configuration
When a single-conductor cable carries current, it generates a magnetic field that radiates out and
alters the impedance of adjacent cables. If the configuration is not correct, a serious impedance
imbalance and, therefore, an imbalance in the current of parallel cables can result. If is the
configuration is done correctly, each cable making up one phase has an equal magnetic field
acting on it.
Parallel cables must have the same conductor size and metal, the same length, the same
insulation type, and the same connector type, and they should be larger than #1/0 (~0.08 in2
[~53.5 mm2]). Figures 2-4 through 2-6 show possible configurations for balanced magnetic
fields.
Figure 2-4
Acceptable Configurations: Two Cables per Phase
Figure 2-5
Acceptable Configurations: Three Cables per Phase
2-7
10296123
Understanding the Design of Power Plant Cable Systems
Figure 2-6
Acceptable Configurations: Four Cables per Phase
2-8
10296123
Understanding the Design of Power Plant Cable Systems
2-9
10296123
10296123
3
UNDERSTANDING THE PHYSICAL CONDITION OF
THE SYSTEM
The environmental and operating conditions of a cable circuit are a major factor in determining
their long-term reliability. Cables operated in benign environmental conditions (dry, low ambient
temperature, and low radiation dose levels) and benign operating conditions (lower voltage stress
and currents lower than rated) are highly likely to function reliably well beyond 40 years. Cables
in benign environments have failed in operation, but for the most part, those failures were
attributed to random defects or installation damage rather than an aging problem.
This section describes the adverse environments and operating conditions that must be
understood by personnel responsible for cable system aging management. Knowing the adverse
environments and conditions that affect cable longevity will allow personnel within a plant to
identify them and to select appropriate aging management methods for the cable exposed to
them.
3-1
10296123
Understanding the Physical Condition of the System
metal shields and conductors. These conditions can be controlled by design of the cable,
constructing the duct system to remain dry, or adding dewatering systems. The length of
exposure to the adverse condition will determine the ultimate life of the cable. Short or no
exposure to elevated temperature and wetting will have little or no effect on cable life. Extended
exposure to elevated temperature or wetting or both will shorten a cable’s life.
Natural rubber was the main source of insulation for medium- and low-voltage cables for power
plants from 1900 to 1950. Each cable manufacturer had its own formula for rubber
compounding. Testing was limited to taking a short sample of a manufacturer’s insulation,
weighing it, soaking it in hot water for about one week, and then weighing it again. If the weight
of the insulation had increased by less than 10% by the end of that week, the compound was
considered adequate for insulation used in wet environments.
World War II brought polyethylene (PE) to the cable industry—first for radar cables and, just
after the war, to low- and medium-voltage power cables. Using the same weighing method, they
found that they could not measure any weight increase. Electrical measurements were
incorporated to show any change in properties. These test results seemed to indicate that this new
material was unaffected by water. This was not the case, because even a small amount of
moisture could deteriorate PE by creating water trees when the cable was energized for a period
of time in a wet environment.
By the 1950s, researchers determined that moisture could penetrate the insulation wall of a cable
if a dc negative polarity was applied to the conductor. Test standards, therefore, specified that dc
test sets should be constructed to impose dc negative polarity on the conductor of the cable in a
test environment.
In the case of ac, the voltage is negative for only one-half cycle, and engineers originally
believed that during the positive half cycle, the water would be pushed away from the conductor.
However, degradation was still noted at the conductor-to-insulation interface. This condition led
to the theory that dielectrophoresis was applicable to water migration in ac cable circuits.
Electrophoresis is a term used to describe the movement of charged particles in an electric field.
Particles with a positive charge tend to move toward a negative electrode, and negative ions tend
to move toward the positive electrode.
Dielectrophoresis relates to the movement of an uncharged but polarized particle or molecule in
a divergent (ac) field. In the example of a single-conductor electrical cable, the field in the
insulation increases as a particle or molecule gets closer to the conductor. An uncharged particle
will be polarized at any given point in time, so that it will have a negatively charged dipole with
its negative side toward the conductor that is positive at that instant. Because the negative side of
this dipole exists in a stronger field than the positive side, the particle will be attracted toward the
field of greatest field intensity. In an ac system, as the conductor becomes negatively charged,
the polarization process is reversed. This means that the particle is still attracted to the conductor
with its higher electric field (see Figure 3-1).
3-2
10296123
Understanding the Physical Condition of the System
Dielectrophoresis
Figure 3-1
Electrostatic Flux Lines and Charged Particles
The practical effect of dielectrophoresis is that moisture is drawn to the higher dielectric field
regions, even in an alternating field. This high stress point is likely to be a small void that is a
portion of the initial tree formation. The void that was initially filled with gas now becomes
filled with water. Although this does not fully explain the formation of the water tree, it does
shed some light on the growth of such trees and the dispersion of moisture in an energized cable.
Dielectrophoresis provides a means for water to propagate through the insulation to feed water
tree formation, both in the insulation and at the interface between the conductor shield and the
insulation.
One way to solve the problem of moisture ingress is to use a jacket or impervious metal sheath
over the insulation shield system. When there is no voltage drop across the jacket, there is no
dielectrophoresis effect. All of this is possible only when the jacket is intact, of course. Putting a
jacket over nonshielded conductors does not accomplish this goal because the jacket material
does have a voltage drop across it. Even semiconducting layers of insulation do not stop the
process because there is some voltage drop across that portion of the cable as well. However,
some jacket materials are better than others at preventing water passage through to the insulation.
Dyed chlorosulfonated polyethylene (CSPE) jackets allow much more water transfer than do
black jackets in which carbon is the colorant. Accordingly, water transfer will occur through
some jacket materials without needing dielectrophoresis to cause it. This transfer occurs
relatively slowly, on the order of months or more.
3-3
10296123
Understanding the Physical Condition of the System
Figure 3-2
Example of Manhole Drainage System
A well-designed dewatering system does not require a pump in every manhole, but it can be
desirable to have a sump pump built into each manhole for occasional water intrusion. Automatic
sump pumps might be necessary. They use a float switch or, when the water rises, the entire
housing goes up to turn on the pump. Regular inspection and maintenance of sump pumping
systems are required; otherwise, they can fail relatively rapidly and allow the cables to be
submerged.
If pumping is to be performed, whether using automatic sump pumps or periodic inspections
with manual pumping, sampling of the water for environmental or radiological conditions might
be necessary to determine how to dispose of the water. Pumping oil-contaminated or tritium-
contaminated water into storm drains or into the ground could have serious consequences.
3-4
10296123
Understanding the Physical Condition of the System
is warranted. Discussions with the manufacturer of the cable might provide insights regarding
the importance of the damage and any necessary corrective actions. Compression and damage to
cables can occur at the dropouts from trays to local conduits. The cable should be protected
from sharp edges of the conduit by a bell or other appropriate fitting, and padding might be
necessary on the rungs of the tray where the cable drops out to preclude excessive load on the
side of the cable.
3-5
10296123
Understanding the Physical Condition of the System
3-6
10296123
Understanding the Physical Condition of the System
Figure 3-3
White Powder Indicates a Location of Corona Discharge Between a Cable and a Ground
Cable in Close Proximity
3-7
10296123
10296123
4
CABLE DESIGNS
This section describes cable constructions that have been installed in typical power plants.
4-1
10296123
Cable Designs
Figure 4-1
Shielded, Single-Conductor, Medium-Voltage Cable Design
The conductor (A) is typically stranded copper or aluminum, with copper being more common in
power plants. The insulation (C) used in conventional and plant medium-voltage cables is EPR
or, to a lesser extent, XLPE. Medium-voltage cables in a few early plants had butyl rubber
insulation. Shields (B and D), composed of semiconducting polymer in modern designs, help
maintain a uniform voltage stress in the insulation. The metallic tape shield (E) provides a
continuous drain for the insulation shield and a return path for fault currents. The jacket (F) adds
mechanical protection as well as an additional barrier to moisture and external contaminants.
In early black EPR and butyl rubber cables, the insulation and conductor shields were formed
from helically wrapped carbon black–loaded cotton tape. By the early 1970s, the insulation
shield was made of a helically wrapped semiconducting polymer tape. The insulation shields
were in tape form to enable installers to differentiate between the black shield and the black
insulation. When strippable, extruded black semiconducting polymer shields became common,
the industry converted to gray, pink, or brown EPR by eliminating some or all of the carbon
black from the insulation. Extruded conductor shields became available for power plant use in
the mid 1970s.
A modern construction, similar to that shown in Figure 4-1, is the water-impervious design
shown in Figure 4-2. The sub-components are virtually the same, but the water-impervious cable
is constructed to limit absorption of moisture and external contaminants through additional radial
and axial design barriers. One design barrier is that the metallic tape shield (E) has been replaced
with a continuous, corrugated tape shield that is copolymer coated and sealed at the overlap. A
second design barrier is the use of water-swellable tapes or powders placed in the conductor
strands to prevent water migration through the stranded core. These two design changes provide
barriers to prevent moisture and external contaminants from penetrating into the layers below.
The corrugated metallic shield also should provide lower, more stable shield resistance compared
to the helically wrapped copper tape design, which should aid in long-term testability. Other
moisture-impervious designs exist, including those that use a fully sealed lead or aluminum
sheath instead of a corrugated copper sheath with a glued overlap. For more information on cable
design and selection, refer to the EPRI report Plant Support Engineering: Common Medium
Voltage Cable Specification for Power Plants (1019159) [12].
4-2
10296123
Cable Designs
Figure 4-2
Medium-Voltage Shielded Cable, Water-Impervious Design
Two disadvantages of the water-impervious design are the higher cable cost and the increased
difficulty in making field terminations. The cost differential might not be significant in
comparison to the overall cost of installation.
Using a lead sheath in place of the corrugated copper tape is also possible, but it has limited
availability and requires additional care in pulling to preclude damage to the lead sheath.
Installation of lead-sheathed cable might not be practical when replacing cables that had tape
shields or were nonshielded. The lead-sheathed cable will likely require larger ducts and larger-
radius bends.
Figure 4-3 shows a nonshielded cable design. In this design, there is no insulation shield. This
design has been applied in many instances for medium-voltage circuits of 4160 V and lower. The
reason is that non-grounded systems, when designed properly, can tolerate a single phase-to-
ground fault while remaining in service for a short period. However, the absence of a shield that
confines the voltage stress to the insulation makes testing of the insulation system in a
meaningful way quite difficult. Testing is possible in a laboratory by submerging the cable in
water and using the water for the ground plane. Submergence of an entire circuit is not practical
in a plant and can provide confusing results because the jacket of the cable is in series with the
insulation and can affect the test results.
Figure 4-3
Medium-Voltage Nonshielded Cable Design
4-3
10296123
Cable Designs
Figure 4-4 shows an example of the UniShield cable design. In this design, the outer polymer
layer serves as both jacket and semiconducting insulation shield for the cable. The shield
conductor system is composed of six longitudinal, corrugated, neutral wires embedded in the
jacket. The design includes a compacted conductor. The result of this design is a smaller-
diameter cable. The small-diameter design might be needed when existing ducts are relatively
small in diameter because smaller-diameter cable had been previously used. Other designs exist
that have small diameters, but they have not been commonly used in power plants to date.
Figure 4-4
UniShield Construction, Shield Wires Embedded in the Semiconducting Outer Jacket
4-4
10296123
Cable Designs
When a ground fault occurs on a power system, voltages higher than the cable’s voltage rating
can occur. To address this, ICEA and AEIC specify thicker insulation if the overvoltage can last
longer than 1 minute. The duration of the overvoltage depends on whether the neutral is solidly
grounded or ungrounded. The insulation and resultant thicknesses are classified as 100%, 133%,
or 173% insulation levels [8], as follows:
100% Level. Relay protection normally clears ground faults within 1 minute.
133% Level. The faulted cable de-energizes within 1 hour.
173% Level. The time needed to de-energize the fault is indefinite. This level is
recommended for resonant grounded systems.
Even if the 173% level of insulation is used, it does not mean that cables with a ground fault can
be left in service. If a ground is suspected (for example, an alarm is received), action should be
taken to remove the cable from service as soon as possible. The faulted phase can adversely
affect the adjacent phases and lead to a phase-to-phase fault that will have extremely large
currents, which could cause damage to the connected equipment. For plant use, some designers
opted to use the 133% level as conservatism.
4.2.2 Conductors
The conductor (see A in Figure 4-1) is typically stranded copper or aluminum, the former being
more common in power plants. The cross-sectional area size, the associated diameter, and
number of strands in the conductor are standardized and unaffected by material choice. The
conductor size is given in terms of American Wire Gauge (AWG) or kcmil in the United States
and in terms of mm2 elsewhere. The outside surface of the conductor can be smoothed to a near-
perfect circle by compacting the strands to a compact-round standard, which can result in up to a
3% diameter reduction. The compression also aids in attaining a smoother interface between the
strands and the conductor shield. Figure 4-5 shows a comparison of the conductor stranding
compression options. Although it reduces flexibility somewhat, the compact-rounding helps to
avoid potentially ionizing and discharging air gaps at the inside surface of the insulation, where
the voltage gradient is greatest. The smallest diameter stranded conductor configuration results
from compacting in which each strand is shaped to allow approximately 9% reduction in
diameter. This degree of diameter reduction is used in cables designed for smaller diameter ducts
and conduits.
Figure 4-5
Conductor Stranding Configuration, Showing Compressed and Compacted Conductor
Configurations
4-5
10296123
Cable Designs
Distribution and plant conductor design philosophies differ substantially. Distribution designers
frequently use aluminum conductors and typically use solid conductors for aluminum wires up
through 2/0 AWG (67.5 mm2) and copper up through 6 AWG (13.28 mm2). Stranded conductors
are used for all larger sizes. Although the solid conductor is somewhat stiffer, it does effectively
block the flow of water through the interior of the cable. In contrast, plants use stranded copper
conductors to allow greater flexibility for installation. Water blocking of stranded conductors can
be accomplished through the use of polymer fills, water-swellable powders, or a thixotropic gel
filling for the strand interstices. However, water-blocking technologies were not available at the
time of construction of most plants, but they are available for replacement cables and new
plant use.
Historically, most cables were not compacted, with the exception of the UniShield construction,
which is designed to be a small diameter for a given voltage rating.
4.2.4 Insulation
The cable’s primary insulation (see C in Figure 4-1) is manufactured of materials that are
designed with sufficient dielectric strength to withstand the voltage stress experienced during
normal operation, as well as unusual voltage spikes and surges. The insulating material for most
medium-voltage plant cables is either XLPE or EPR; however, some early plants have cables
with butyl rubber insulation. Although XLPE has additives for fire retardance and processing, it
4-6
10296123
Cable Designs
is mostly XLPE polymer, but a copolymer (additional plastic type) is often used. Recent
experience within the transmission and distribution utility industry has led to the development of
a tree-retardant enhancement of this insulating material. The additives to tree-retardant XLPE
(TR-XLPE) do not totally eliminate water trees; rather, they greatly reduce their rate of
generation and growth.
In contrast to the relatively limited number of variations in XLPE formulations, EPR insulations
are comparatively complex compounds that vary substantially among polymer suppliers and
cable manufacturers. Different improvements of the EPR materials have evolved and have been
distinguished from one another by color, such as black, gray, brown, or pink. The colors alone
are not directly related to the way these materials age, but they are indicative of the changes
made to the formulations that improved the longevity of cables. Black EPR cables are early-
generation EPR cables manufactured through the mid to late 1970s. Brown EPR appears to have
been resistant to water-enhanced aging effects throughout the period of its use. Pink (or red) EPR
cables are the more modern generation, available from the late 1970s through today. A few
manufactures changed from black EPR to gray EPR in the late 1970s. The major shift in EPR
formulation during that period was the transition from untreated clay to silane-treated clay. The
silane treatment caused the EPR to bind more tightly to the clay, and it sealed the clay so that
water uptake in the EPR was greatly reduced.
A typical dielectric constant for XLPE is ~2.3, whereas that for EPR is distinctly higher, at ~3.2.
Thus, EPR has higher dielectric losses per unit length of cable than XLPE, but it is generally
more resistant to voltage stress and discharges. The dielectric loss through the EPR insulation
drains any charge that could build in the insulation at imperfections and eliminates high localized
stresses that result in water-enhanced aging. Because medium-voltage cables in power plants are
usually much shorter than those in transmission and distribution applications, electrical system
losses are not paramount in the design of station cable systems. Thus, a majority of underground
medium-voltage cables at plants have EPR insulation, which has always had an expectation of
greater operating life. Rubber insulation systems also have been chosen for in-plant applications
due to their flexibility, which is important during installation in the tighter confines of power
plant applications.
When PE insulation first became available for medium-voltage applications, it was hailed as the
cure-all for many of the issues then facing the distribution industry. These PE materials had
extremely low losses in a high-stress electrical field, were easier to compound than rubber
systems, were lower in cost than either rubber or paper-insulated lead-covered cables, and were
quite hydrophobic (tended not to absorb water or moisture), whereas conventional rubber
systems were not. Thus, distribution utilities made widespread of various types of PE.
generating station designers gave relatively little weight to PE’s low loss characteristics because
of the insignificant circuit lengths involved. Those few plant designers who did not choose
rubber-insulated systems chose XLPE for its superior mechanical strength and thermal
endurance.
4-7
10296123
Cable Designs
4-8
10296123
Cable Designs
Care must be taken when splicing new cable to old black rubber cables with tape shield designs.
In modern cables, the insulation shield layer is easy to identify because it is black as opposed to
the gray, brown, or pink insulation. Current splicing crews are familiar with the modern cable
designs. However, they are unlikely to have seen woven fabric or polymer tape shields. They
must be trained in how to make splices and terminations to these old style cables. If the
semiconducting layer is not removed properly, failure is possible immediately on energization or
shortly thereafter because the separation between the conductor and the shield will barely
withstand operating voltage.
4-9
10296123
Cable Designs
4.2.7 Jacket
Jackets protect the power cable’s underlying insulation from mechanical and environmental
damage. Jackets are also used to maintain the configuration of multi-conductor cable. Although
metallic armor might be used on some medium-voltage constructions, most underground power
cables at plants have a nonmetallic jacket (see F in Figure 4-1) as their outer protective sheath
(with or without underlying armor). The jacket materials include polyvinyl chloride (PVC),
thermoplastic chlorinated polyethylene (CPE), neoprene, and CSPE. Replacement cables might
have thermoset CPE or thermoset low-smoke zero-halogen jackets. The transmission and
distribution industry also favors a linear low-density polyethylene (LDPE) jacket material for
moisture protection, but PE jacket material is not used in plants because it is not flame retardant.
Jackets slow moisture intrusion into the underlying cable core. In the case of the UniShield
cable, the jacket serves a dual role as the insulation shield and the overall protective sheath.
generating stations use flame-retardant designs that could be obtained only through the use of
specially modified jacketing systems. When non-flame-retardant cables have been used, flame-
retardant coverings have been applied to prevention the propagation of fire.
Although they are not impervious to water migration, the existence of the polymer jackets on
plant cables has impeded the ingress of moisture into the insulation. The presence of this
additional diffusion barrier (along with some of the installation attributes described in Section 6,
Fundamentals of Cable Insulation Systems) helps explain why moisture-related degradation of
medium-voltage cable is just now becoming a concern for plants, although the cables are 25–35
years old.
4-10
10296123
5
SPLICING AND TERMINATING
This section describes splicing and termination of medium-voltage cables. Splicing of medium-
voltage cable is not a common practice in U.S. power plants, but it must be used in long runs and
might be necessary when repairing a failed section of cable.
5.2 Gradients
5-1
10296123
Splicing and Terminating
Figure 5-1
Equipotential and Flux Lines in a Cable [13]
When the cable is cut so that the shield ends abruptly, the electrical stresses change from being in
the semiconducting material to being in the air, as shown in Figure 5-2.
Figure 5-2
Electrical Stress Fields, Shield Removed
To reduce the electrical stress at the end of the cable, the insulation shield is removed for a
sufficient distance to provide adequate leakage or creepage distance between the conductor and
the shield. The distance depends on the voltage involved as well as the anticipated environmental
conditions. The removal of the shield disrupts the coaxial electrode structure of the cable. In
most cases, the resulting stresses are high enough that they cause dielectric degradation of the
materials at the edge of the shield, unless steps are taken to reduce that stress. The concentration
of electric stress is now located at the conductor and end of the insulation shield. The stress lines
are the horizontal lines that curve upward at the end of the shield (identified by the arrow in
Figure 5-2), and the flux lines are the curved lines that are at right angles to the stress lines. The
stress lines are more closely spaced near the conductor, and the flux lines more closely spaced at
the end of the shield. These forces are strong enough to actually decompose the factory
insulation at that interface and ultimately cause the cable insulation to fail.
The stress at the insulation shield remains great because the electrical stress lines converge at the
end of the shield. The equipotential lines are closely spaced at the shield edge. If those stresses
are not reduced, PD can occur. Electrical stress relief is required in most medium-voltage
applications.
5-2
10296123
Splicing and Terminating
Figure 5-3
Termination of an Insulation Shield with a Stress Cone [14]
A stress cone increases the spacing from the conductor to the end of the shield, as shown in
Figure 5-3. This spreads out the electrical lines of stress as well as providing additional
insulation at this high stress area. The ground plane gradually moves away from the conductor
and spreads the dielectric field, thus reducing the voltage stress per unit length. The stress relief
cone is an extension of the cable insulation. Another way of saying this is that the electrostatic
flux lines are not concentrated at the shield edge as they are in Figure 5-2; it follows that the
equipotential lines are also spaced farther apart. Stress cones can be taped by hand or premolded.
Terminations that are taped achieve this increase in spacing by creating a lapped conical
configuration of tape followed by a semiconducting layer that is connected electrically to the
insulation shield, as shown in Figure 5-3. Premolded stress cones use the same concepts in the
construction.
The classic approach to the design of a stress relief cone is to have the initial angle of the cone be
a few degrees and take a logarithmic curve throughout its length. This provides the ideal
solution, but it was not usually needed for the generous dimensions used in medium-voltage
cables. There is such a little difference between a straight slope and a logarithmic curve for
medium-voltage cables that, for hand build-ups, a straight slope is acceptable. Premolded designs
usually maintain that logarithmic shape.
In actual design, the departure angle is in the range of 3° to 7°. The diameter of the cone at its
greatest dimension has generally been calculated by adding another insulation thickness to the
diameter of the insulated cable at the edge of the shield; therefore, at the maximum diameter of
the stress cone, the insulation thickness is twice that of the cable’s insulation. A major
disadvantage of such stress cones is that they require much more space between cables than the
voltage gradient types that are described in Section 5.2.3, Voltage Gradient Design.
5-3
10296123
Splicing and Terminating
Figure 5-4
Stress Relief with High Dielectric Constant or High Resistivity Materials
5-4
10296123
Splicing and Terminating
Figure 5-5
Projection from Semiconducting Layer and Cut into Insulation
5-5
10296123
Splicing and Terminating
Medium-voltage cables with extruded insulation are prepared for splicing and termination in a
similar manner, regardless of the type of insulation. The entire cable must be cut to the proper
length with a straight cut. The jacket, metallic shield, and semiconducting layer must then be
removed to the proper dimensions to fit the termination being used. Tools are available to score
the semiconducting layer for removal (see Figure 5-6). Spiral or longitudinal strips are scored
partway through the semiconducting layer but not into the insulation. These narrow strips are
then removed one at a time.
Figure 5-6
Two Types of Semiconducting Layer Scoring or Stripping Tools
Cables manufactured in the 1970s occasionally had small bits of semiconducting layer remaining
on the surface of the insulation when the strips of the semiconducting layer were removed. These
had to be removed by abrading the surface. For these cables, some light sanding might be
required to remove the small chunks of the semiconducting material that adhered to the
insulation surface. A fine, nonconducting, nonmetallic abrasive paper, such as 400–600 grit,
should be used to polish the insulation surface and remove these particles.
With present manufacturing technology, there should be no reason to sand or abrade the exposed
insulation. Care must be taken to avoid cuts or scrapes, which are places for moisture to be
trapped or for elevated electrical stress to occur that can lead to failure. One critical location for
cuts is where the semiconducting material is removed. A cut into the insulation at that point is
not easily seen. If left there, a cut creates a void that is likely to cause an early failure, especially
in XLPE insulated cables.
Cuts, scrapes, or gouges cannot be corrected by sanding off the defect or by reducing the
insulating wall thickness. Serious errors might require that the cable end be re-cut and
preparation restarted.
The acceptance criteria for cable end preparation depend on the methods and materials used to
construct the splice or termination. Common requirements include a cable insulation surface that
is free of contamination, imperfections, and damage. A smooth surface for extruded dielectric
insulations minimizes contamination and moisture adhering to the surface. If a rough surface
remains, it must be made smooth.
5-6
10296123
Splicing and Terminating
Connections to the metallic shield of the cable must not damage the underlying cable
components. Cutting into the strands of wire—or, of even greater importance, into a solid
conductor—cannot be tolerated. Figures 5-7 through 5-11 illustrate some of the steps involved in
cable preparation.
Figure 5-7
Measure Carefully to Achieve the Proper Length of Cable
Figure 5-8
Slip Applicable Sleeves over One or Both Cable Ends Before Installing Connector
Figure 5-9
Connectors of Several Lengths and Diameters for the Same Conductor Size
Figure 5-10
Shear-Bolt Connectors for Copper Conductor
5-7
10296123
Splicing and Terminating
Dampen a clean, lint-free cloth with the solvent, start the wipe
on the exposed insulation,and wipe toward the connector.
Figure 5-11
Clean the Insulation, Using an Approved Solvent
Figure 5-12
Semiconducting Tape Is Applied over the Connector to Form a Smooth Interface
5-8
10296123
Splicing and Terminating
Figure 5-13
Insulating Tape Is Applied Until the Proper Thickness Is Achieved over the Connector
Figure 5-14
Another Layer of Semiconducting Tape Is Applied over the Insulating Tape
5-9
10296123
Splicing and Terminating
Figure 5-15
A Metallic Braid Is Installed, a Ground Strap Is Attached, and the Jacket Tape Is Installed
over the Entire Splice Area
5-10
10296123
Splicing and Terminating
Figure 5-16
Cutaway of a Premolded Splice
Research sponsored by EPRI [17] showed that premolded splices should be selected so that they
fit tightly over the cables. Manufacturers show a range of sizes on each splice that the particular
molded device will accommodate. When cables are spliced in confined spaces, bending can
cause gaps between the premolded splice and looser-fitting cables; these gaps can allow moisture
to enter and cause a failure. The solution is to select premolded devices that are on the tighter
side of the manufacturer’s size range, even though splicers might not appreciate the extra effort
required to assemble the splice.
Advantages of premolded splices are the following:
They require the least time to complete.
They require the least skill to assemble.
They have the greatest margin for error, yet still work.
They are the least complex to make.
Disadvantages of premolded splices are the following:
Cable sizes must be close to the same on both sides.
The premolded device must be a tight fit, causing it to be hard to push on.
Premolded devices might require stocking several sizes because they are not capable of
fitting over a wide range of cable sizes.
Proper connectors and dies must be used. If the crimping process increases the diameter of
the flashing of the connector and it becomes too large, the inside of the molded rubber can be
damaged when slid into place.
5-11
10296123
Splicing and Terminating
Figure 5-17
Cold-Shrink Splice, Showing the Direction of the Shrinking Process
Figure 5-18
Carefully Position the Housing Before Removing the White Core Support
5-12
10296123
Splicing and Terminating
Figure 5-19
Continue Removing the Core While Holding Its Position on the Cable
5-13
10296123
Splicing and Terminating
Figure 5-20
High-Permittivity Mastic Material Is Placed over the Connector and Conductor, Without
Concern for Smoothness
Figure 5-21
The Semiconducting Tube Is Slid into Place and Heated Until Properly Shrunk Down to the
Cable
Figure 5-22
An Insulation Tube Is Slid into Place and Shrunk Down
Figure 5-23
A Tube That Is Both Insulating on the Inside and Semiconducting on the Outside Is
Positioned and Shrunk into Place
Figure 5-24
Tinned Copper Braid is Wrapped Around the Splice to Replace the Metallic Portion of the
Insulation Shield
5-14
10296123
Splicing and Terminating
Figure 5-25
A Ground Strap Spring Is Placed Under One Side of the Cable’s Taped Metallic Insulation
Shields
Figure 5-26
The Ground Strap Is Placed Across the Splice and Connected to the Factory Metallic
Shield on the Opposite Side of the Splice
Figure 5-27
An Overall Rejacketing Tube Is Placed Around the Entire Area
5-15
10296123
Splicing and Terminating
5.4 Terminations
Terminations are available in the four types: taped, premolded, cold-shrink, and heat-shrink.
They are similar to the splices with respect to installation practices (see Section 5.3, Splices for
Shielded Cables).
Cold-shrink and heat-shrink type terminations with skirts (see Figure 5-28) can be used indoors,
as long as there is sufficient space for the skirts. Having the additional creepage length might be
desirable because rain is not available to clean dust and contaminants from terminations in
cubicles.
Figure 5-28
Cold-Shrink Outdoor Termination with Rain Sheds or Skirts
The premolded types of terminations (see Figure 5-29) are known as separable connectors that
are useful in power plants. They are not exactly the same as those used in the distribution
systems because distribution generally uses the load break type. They both consist of a
semiconducting shield over the metallic connector system and insulating material, and they have
an overall grounded surface. They are used to provide the connection between the cable and the
electrical compartment of a transformer, switch, or other device.
5-16
10296123
Splicing and Terminating
An advantage of these devices is that, while they are de-energized, they can be taken apart at the
motor for testing purposes, if there is sufficient space in their connection box. Taped
terminations are destroyed during the disconnection process and require significant time to
rebuild.
Figure 5-29
Dead Break, Premolded Termination
Courtesy of Elastimold
The insulating portion of the elbow is made of EPDM rubber with an outer covering of similar
materials that contain carbon black to make them semiconductive.
Other termination types that are used to connect medium-voltage feeder cables to motors,
switchgear, or transformers include stub-type (or V-type) and in-line type kits.
For stub- or V-type kits, stress-control tubing is installed on shielded cable; compression lugs are
installed on each cable and bolted together as shown in Figure 5-30. A variety of shielded cable
designs can be accommodated. Sealant and insulating material complete the installation. Because
of the many cable sizes that can be involved, these kits do not contain the required hardware.
Figure 5-30
Stub-Type Motor Connection
Courtesy of Tyco
For in-line type kits, compression lugs are bolted together in a similar manner as for stub-type
kits, but they are connected end-to-end like a normal splice (see Figure 5-31).
Figure 5-31
In-Line Type Connection
Courtesy of Tyco
5-17
10296123
Splicing and Terminating
5-18
10296123
Splicing and Terminating
always specified over these connectors. When crimping a connector, the crimping tool must be
rotated around the connector about 90° after each crimp to maintain a straight configuration.
Parallel crimps result in a banana-shaped connection (see Figure 5-32). The flashing that results
on the connector can score the interior of a premolded connector.
Figure 5-32
Improperly Crimped Connector
5-19
10296123
Splicing and Terminating
5-20
10296123
6
FUNDAMENTALS OF CABLE INSULATION SYSTEMS
Cable insulation systems include the primary insulation system (EPR, XLPE, and butyl rubber
for nuclear plants) and the conductor and insulation shields. The shields equally distribute
voltage stresses within the insulation and prevent PD between the conductor and insulation and
between the insulation and the metal insulation shield. This section describes the properties of
the insulation and shields and the issues that could affect aging.
This section also contains a chronology of improvements made to the insulation and shielding
system since the late 1960s. This chronology does not imply that cables manufactured for
nuclear plants between 1968 and 1990 will have short lives; rather, it indicates that newer cables
should be expected to have lives much longer than those of older-generation cables.
More information on insulation materials, formulations, details of chemical processes in
manufacture, and aging of insulation polymers is provided in the appendices.
6-1
10296123
Fundamentals of Cable Insulation Systems
6-2
10296123
Fundamentals of Cable Insulation Systems
elastomer improve; the tensile strength, modulus, volume resistivity, and breakdown strength all
increase. The silane also serves to remove remaining water and provides the capability to link
between the polymer and the clay particles. Clay minerals have the ability to adsorb cations on
their surfaces [22]; therefore, silane treatments prevent the ions from participating in water
treeing. Several different types of silanes can be used. The modification of the clay, a major
component of EPR formulation, is a complex process.
The term compounding is used to describe the mixing of all the materials in an EPR formulation,
and the EPR that is blended in the soon-to-be cable insulation is called a compound.
6.2.3 Crystallinity
Crystallinity refers to chain alignment, a tendency of unbranched polyolefins to form structures
that are impermeable and exclude impurities [23]. In general, elastomers, including butyl rubber,
are not crystalline, but some types of EPR used for wire and cable can possess a low level of
crystalline regions as a result of high ethylene-to-propylene comonomer ratios. Branches inhibit
crystallinity, as they hinder the tendency of chains to align. To place the subject of crystallinity
in perspective, the highest level of crystallinity in an EPR wire or cable insulation material is
perhaps 6% to 8%, whereas the level of crystallinity for low-density PE is approximately 50%.
In general, the level of crystallinity is low in the EPRs used in medium-voltage cable.
Crystallinity plays a bigger role in XLPE insulation (see Section 6.6, Cross-Linked
Polyethylene).
6-3
10296123
Fundamentals of Cable Insulation Systems
Elastomeric cables with cable shields are intended to be discharge free, in the sense that they do
not exhibit discharges that exceed a specific allowable level on testing—the exact maximum is
defined in industry specifications. The requirements have forced lower and lower maximum
allowable discharge levels as acceptance criteria became more stringent over the years.
Figure 6-1
Butyl Rubber Molecule
The isobutylene portion of butyl rubber provides the desired insulation characteristics, and the
isoprene functionality is present solely to provide unsaturated sites (double bonds) along the
chain for facilitating cross linking. Vulcanization produces chemical links between the loosely
coiled polymeric chains; elasticity occurs because the chains can be stretched and the cross links
cause them to retract when the stress is released. The major portion of the butyl polymer consists
of saturated hydrocarbon chain lengths that impart oxidation, ozone, chemical, and moisture
resistance, as well as low gas permeability. A variety of commercial grades of butyl rubber exist.
In addition, several curing techniques were used in the past to cure butyl for wire and cable
application.
6-4
10296123
Fundamentals of Cable Insulation Systems
6-5
10296123
Fundamentals of Cable Insulation Systems
Propylene monomer differs from ethylene in that it possesses a -CH3 linkage in place of one -H
linkage. When propylene is copolymerized with ethylene, the resulting EPR copolymer is
rubbery and soft, can be deformed, and has relatively high elongation. EPRs possess many short
chain branches; this is what imparts the rubbery or elastomeric quality. In the absence of the
bulkiness provided by the repeating -CH3 units, the PE homopolymer has a tendency to
crystallize. That is why the 75:25 ratio for ethylene to propylene is slightly crystalline and the
50:50 ratio is not.
Figure 6-2
Copolymer of Ethylene and Propylene
6-6
10296123
Fundamentals of Cable Insulation Systems
Cross linking through the use of peroxides imparts improved properties; the resulting elastomer
insulation becomes tougher, resists softening at elevated temperatures, and facilitates
maintenance of form stability at elevated temperatures. This is particularly important for
elastomers (such as ethylene copolymers) as compared to homopolymers such as PE. Cross
linking the polymer chains improves strength and structure, but it does not improve the electrical
properties such as dielectric constant, dissipation factor, and dielectric strength.
6-7
10296123
Fundamentals of Cable Insulation Systems
Figure 6-3
Banbury Mixer Used for Preparing Ethylene-Propylene Rubber Compounds
In contrast to batch processing, continuous mixing has been applied using a Buss kneader. This
continuous process claims the following advantages: lower temperature levels and precise
temperature control during mixing, uniform shearing effect without temperature peaks, shorter
residence times of the components in the unit, and short product changeover times.
Newer, state-of-the-art elastomer technology uses either mixing method, depending on the cable
manufacturer’s mixing capability (or their materials suppliers’ techniques). However, older
cables (more than 30 years old) would have compounds blended with a Banbury mixer.
6-8
10296123
Fundamentals of Cable Insulation Systems
There are two basic types of conductor shields: conductive and emissive shields. An emissive
shield is one that uses a material with a high dielectric constant. However, the most popular type
of shield is the extruded, semiconducting conductor shield containing carbon black.
Cable types used for nuclear power plants have included 5-kV nonshielded and 5-, 8-, and 15-kV
shielded constructions [28]. The tape or extruded shield 5-, 8-, and 15-kV constructions used the
following:
Stranded copper conductors
Extruded strand shields or tape strand shields
EPR insulations of various colors
Extruded insulation shield or tape insulation shields (fiber or polymer)
Helical copper tape shields
Jackets composed of CPE, neoprene, or CSPE
The 5- and 8-kV wire-shielded construction (UniShield) used the following:
Stranded copper conductor
Extruded strand shield
Pink EPR
Semiconducting CPE jacket with an embedded wire shield
The 5-kV nonshielded constructions used the following:
Stranded copper conductor
Extruded strand shield
EPR insulation
CPE, neoprene, or CSPE jackets
6.5.1 General
Butyl rubbers were first commercialized in 1942 for tire inner tubes. Its use for wire and cable
followed that, and it was used for nuclear plant (and distribution) cables into the late 1960s and
early 1970s. PE and then XLPE replaced butyl for the distribution arena. When EPR was
introduced in 1962 and commercial production began in 1963 [30], it quickly replaced butyl
rubber for modern day cables for applications in which elastomers were preferred. Extruded
thermoplastic shields were introduced after 1965 to avoid the use of tapes. From the 1970s
6-9
10296123
Fundamentals of Cable Insulation Systems
onward, a series of improvements were imparted to semi crystalline PE and XLPE systems,
whereas EPR changes were less dramatic, but ongoing. Some of the changes to PE and XLPE,
although focused primarily in the distribution area, are relevant for nuclear cables. These overall
changes included the following:
Movement from thermoplastic PE to XLPE to improve high-temperature properties
Movement from thermoplastic shields (used for PE) to shields with better high-temperature
properties (to match the XLPE); first, deformation-resistant thermoplastic PE shields, and
later, extruded cross-linked shields
Improved extrusion processes in which the shields were extruded concurrently with the
insulation (to avoid contamination)
Development of TR-XLPE
Use of improved jackets
These changes were mostly directed toward distribution cables, but some affected nuclear cables,
as well. One major industry change that took place in the 1980s for XLPE that had little impact
on EPR was the shift in cable cross-linking technology from steam curing to dry curing. A major
change in the late 1980s–1990s that impacted all shielded cable systems (nuclear and
distribution) was the introduction of shield materials that possessed significantly reduced ion
components. By that time, ions had been identified as contributing to water treeing and it was
known that many harmful ions had been introduced into the cable system from the old carbon
blacks used to manufacture the shields.
While all these changes were taking place, EPR usage continued to grow steadily, and changes
were also occurring for EPR. A major advancement for EPR technology occurred in the mid to
late 1970s when a superior grade of EPR that could be extruded on the same type of equipment
used for XLPE was provided; others followed shortly. This was one factor that facilitated
movement from butyl to EPR. Other EPR changes were more subtle, described in terms of color
(change in components of the formulations) rather than content, as the components of EPR
formulations were considered proprietary, for the most part.
6-10
10296123
Fundamentals of Cable Insulation Systems
insulation materials in their constructions, and the difference would be solely in the extrusion
technology. However, some manufacturers prefer to mix their own EPR compounds; therefore,
the specific components would differ from those of their competitors and be proprietary. In these
cases, manufacturers provide their own unique cable systems and apply their own philosophy on
how to ensure longevity and reliability.
In the past and present, for applications in which elastomers were preferred, EPR cables have
been informally classified by the color of the insulation layer. In most cases, these color
classifications provide general indications of changes in formulation and capability. However,
manufacturers have continued to make changes in formulations and improvements to extrusion
practices that occurred well beyond the point at which the color of the polymer was changed.
Although it is convenient to classify EPRs by color, the color classification does not provide
complete information on the total compound package and the ultimate aging and life
characteristics of the material.
The following section focuses on the different EPR cables as provided at different times and by
different suppliers based on color. Some information on aging is necessarily included; more
information on aging of the different cables (butyl, EPR, and XLPE) is presented in Section 7,
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene
Cables Due to Adverse Environments.
6-11
10296123
Fundamentals of Cable Insulation Systems
Black EPRs were provided by Anaconda, Okonite, and General Cable. Some failures have
occurred before the end of the 40-year design life, especially under wet conditions. NEI 06-05
[11] describes the experience with wet medium-voltage cables at nuclear plants. For example,
in a plant that experienced 10 failures of early Okonite black EPR, it was concluded that most of
the problem was associated with manufacturing defects (inclusions in the insulation) combined
with severe operating conditions. Black EPRs were replaced by pink EPRs for new cables in the
mid-1970s, so most of these failures occurred long after the black cable insulation and
constructions were no longer being provided. Most of the failures appear to have been related to
multiple issues, such as voids and inclusions, or installation damage coupled with wetting and
somewhat higher susceptibility to water, rather than just the use of black EPR. Many of the black
EPR cables that were installed in the early 1970s remain in service and are in satisfactory
condition.
6-12
10296123
Fundamentals of Cable Insulation Systems
Table 6-1
Typical Components in Medium-Voltage Ethylene-Propylene Rubber Compounds
(Amounts Are Approximate)
Table 6-1 shows the approximate levels of the various ingredients in typical EPR formulations
using different EPR insulation polymers. These compounds were typically used from the 1970s
to 1990s. Each component has a specific purpose. Note also the presence of a small amount of
LDPE in the semi crystalline EPR compound. The functions of the fillers and additives are
the following:
Zinc oxide improves heat aging.
Lead oxide maximizes wet electric stability.
Silane-treated clay optimizes physical properties. The silane coating on the clay improves
interfacial contact with the polymer. This is necessary to prevent the formation of gaps or
weak boundary layers that could lead to the formation of voids.
Vinyl silane contributes to improved interfacial contact between the clay and the ethylene-
propylene.
Oil is a processing aid.
Wax is a release agent (keeps material from sticking in the extruder).
Dicumyl peroxide is a cross-linking agent.
All elastomers, regardless of the level of crystallinity, require compounding to properly mix the
components. The major difference between the noncrystalline and semi crystalline elastomeric
compounds is in the level of clay (see Table 6-1). The degree of crystallinity of a semi crystalline
EPR is estimated to be in the vicinity of perhaps 5%. Calcined clays that worked so well for
6-13
10296123
Fundamentals of Cable Insulation Systems
butyl rubber were inferior for electrical stability in EPRs; this led to the application of silane-
coated clays for EPRs. The clay improves the modulus of elasticity and the physical properties,
such as tensile strength, tear resistance, or abrasion resistance, of the final product. The influence
of silane coating of the clay in improving properties [21] is reviewed in Appendix D.7.
Figure 6-4
Various Representations of Crystallinity in Polyethylene
6-14
10296123
Fundamentals of Cable Insulation Systems
The crystalline regions are the parallel lines in all depictions, and the noncrystalline or
amorphous regions are represented by the curved lines or folds. The presence of crystallinity
means that PE need not be cross-linked to be useful as insulation, but the presence of the cross
links imparts slightly improved high-temperature properties and renders the semi crystalline
polymer more useful as insulation. Also, the presence of crystalline regions means that it is not
necessary to use mineral fillers or many of the other additives required in an elastomer recipe,
but occasionally this has been done.
From this fundamental perspective, several key points relevant to XLPE are the following:
The degree of crystallinity can vary, depending on the degree of chain alignment.
Increased crystallinity means increased density for the PE and also a higher melting point.
Because high-density PE has greater crystallinity than LDPE, it is tougher, but it is also more
challenging to extrude.
Crystallinity is diminished as the temperature is increased and completely disappears at a
temperature related to the density; for LDPE, this is approximately 223°F (106°C).
Any inorganic mineral fillers present will reside only in the amorphous regions.
Cross linking of medium-voltage cables with XLPE insulation is induced by peroxides, the
same technology used for EPR elastomers.
Peroxide-induced cross linking of XLPE takes place only in the amorphous regions;
crystalline regions are melted during the cross-linking process and re-form after the cable is
cooled.
Any impurities—for example, foreign contaminants, ions, or water—are located in the
amorphous regions, and moisture transport takes place by movement through those regions.
For additional information, see Electrical Power Cable Engineering [23].
Some cable manufacturers used mineral fillers, such as clay, in their XLPE insulated cables.
When filler additives are used, all the issues described in Section 6.1, Primary Insulations, for
elastomers apply to this type of XLPE. The difference, as noted, is that the fillers are now
inherently located in the amorphous regions of the XLPE (rather than uniformly dispersed
throughout the matrix). In principle, mineral filler additives provide a back-up source for
imparting toughness, certainly at elevated temperatures at which the crystallinity is reduced or
absent. As with EPR, mineral fillers do not improve the electrical properties.
Conventional XLPE was provided by many suppliers. Mineral-filled XLPE called Vulkene was
provided by General Electric; later, the supplier provided a second grade of Vulkene, called
Vulkene II. This technology was later made available by Raychem. As with butyl rubber,
mineral-filled XLPE is no longer being provided commercially. Only one plant is known to have
Vulkene insulation. Several Vulkene failures occurred under wet conditions at that plant.
Table 6-2 lists the percentage of nuclear plants that use XLPE insulated cables, by the voltage
rating of the cables.
6-15
10296123
Fundamentals of Cable Insulation Systems
Table 6-2
Percentage of Nuclear Plants with Cross-Linked Polyethylene Insulated Cables by Cable
Voltage Rating
8 kV 9%
15 kV 12%
25–35 kV 3%
6-16
10296123
7
AGING AND DEGRADATION OF BUTYL, ETHYLENE-
PROPYLENE RUBBER, AND CROSS-LINKED
POLYETHYLENE CABLES DUE TO ADVERSE
ENVIRONMENTS
7.1.1 General
Butyl rubber will degrade due to the long-term effects of wetting, thermal stress, and radiation.
The loss of physical properties during aging, such as changes in tensile strength or elongation, is
normally caused by elastomeric chain scission or, to a lesser extent, by additional chain cross-
linking. Scission refers to elastomeric chain cleavage, and reversion generally refers to
destruction of the cross links rather than elastomer backbone cleavage. External stress-inducing
agents include oxygen, ozone, thermal energy (most heavily), ultraviolet light, environmental
chemicals, and high-energy radiation. Voltage stress is anticipated to become a factor only as the
voltage rating of the cable increases or the insulation system is wetted [34]. Additives to counter
the influence of some of these stresses were sometimes incorporated into the elastomer. These
additives, such as amines or phenolics, interact with free radicals (electrons) generated by these
outside sources, and are referred to as antioxidants or antiozonants. As long as a small
percentage of the original antioxidants and antiozonants are available in the polymer, polymer
degradation proceeds at a slow pace. When these materials are consumed, degradation can
proceed quite rapidly, especially at higher stress levels.
7-1
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
Voltage stress influence was described by Hiranandani [34], who reported on a 30-year-old,
nonshielded, 5-kV butyl rubber insulated cable with a neoprene jacket that possessed areas in
which the jacket was not bonded to the insulation. The cable had areas that showed discharge
activity, but the cable had not failed (before removal and analysis). Surface discharge–induced
degradation, which can exist for an extended period, would result from voltage-induced
phenomena in the absence of water. Although degradation of butyl rubber is expected to increase
the rate of electrical property deterioration, that rate remains low, with electrical degradation
taking as long as three or more decades to occur.
7-2
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
rather slight initial improvement in properties of the butyl compound would be expected during
radiation aging, ultimately followed by softening during continued aging. Sheirs [41] suggests
the opposite with respect to properties, as it states that, with radiation, “butyl rubber crosslinks to
become stiffer, lose elongation and turn friable and powdery.” Sheirs refers to gamma radiation,
and although some cross-linking is to be anticipated, the presentation makes no mention of dose
or dose rate. One would expect other changes to be noticeable before those described by Schiers.
It was shown in past years that polyisobutylene cross linking could be enhanced under gamma
radiation by use of certain additives [42], but it is not known whether this technology was
applied to butyl rubber for cable insulation.
In general, few if any butyl rubber insulated medium-voltage cables are expected to be in high
radiation zones. The other stresses appear to be more significant at low anticipated radiation
doses. Thermal stress is likely to be of the most importance in plant applications.
Overall, the direction of physical property changes reported is not unreasonable, because the
polyisobutylene portion of a butyl rubber copolymer would be expected to ultimately degrade;
however, the radiation doses required for such degradation should far exceed what is seen in
practical butyl rubber usage for commercial nuclear power station applications.
7.1.5 Conclusions
Butyl rubber was a reasonable choice for general power station applications based on the state of
knowledge in the 1960s. However, history has shown that harmful environmental issues, some
not foreseeable at the time, influence long-term reliability. These include allowable moisture
pickup in butyl rubber cable compounds after extrusion; thermal degradation due to the nature of
the cross links in butyl rubber and their influence on thermal stability (not an issue for EPR—see
Section 7.2, Aging and Degradation of Ethylene-Propylene Rubber); and possible overall
radiation-induced degradation after long aging times. Also, for shielded butyl rubber cables, the
issue of contaminants and impurities in the carbon black and shield compounds, discovered in
the 1980s (and corrected at that time), also will contribute adversely to the rate of aging and loss
of life of butyl cables
7.2.1 General
A number of types of EPR insulations and EPR cable designs exist that are partially identifiable
by the color of the EPR (black, gray, pink, and brown). Industry operating experience indicates
failures of the cable in certain of the black, gray, and pink EPR types under wetted conditions.
Some types of discharge-free pink insulation and all the discharge-resistant insulations have been
free of water-related insulation failures.
U.S. power industry operating experience indicates that some EPR insulation systems are
affected by wetting, but the extent to which an individual manufacturer’s compound is
susceptible to water-related degradation differs from one EPR to another. Regardless, failures of
EPR cables in wetted environments have occurred and must be addressed as an aging concern.
7-3
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
In some instances, adverse thermal conditions resulted in failures of EPR insulated cables.
Identifying those adverse thermal conditions (such as external radiant energy, ohmic heating, or
high-resistance connections), determining the effect on the cable and correcting the condition are
required to maximize cable reliability. Thermal runaway and embrittlement due to insulation
hardening are the effects of long thermal degradation. Thermal runaway occurs when the heat
generated in the cable and insulation system are not dissipated by the surrounding environment,
causing the insulation resistance of the insulation to decrease. This results in higher leakage
currents in the insulation, which, in turn, increases the temperature of the insulation. The effect
feeds on itself to the point at which breakdown of the insulation occurs.
In general, medium-voltage cables are not exposed to radiation levels that would affect the long-
term reliability of EPR cable compounds. However, if medium-voltage cables have been exposed
to both higher than normal temperature (122°F [>50°C]) and radiation doses of >5 Mrad
(>50 kGy), a loss of elongation and a hardening of the jacket and/or insulation should be
expected. Identifying cables in these adverse conditions and verifying the electrical and physical
condition is recommended. More information can be found in the EPRI report Aging
Management Program Guidance for Medium-Voltage Cable Systems for Nuclear Power Plants
(1020805) [5].
7-4
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
Figure 7-1
Ethylene-Propylene Rubber Cable Failures as Function of Color of the Ethylene-Propylene
Rubber Insulations [11]
Figure 7-2 provides an indication of onset and expected onset of initial failures associated with
water-induced degradation for the various insulations used in medium-voltage cables. Because
EPR is the dominant insulation type, Figure 7-2 applies mostly to EPRs. One manufacturer’s
insulations (both black and brown discharge-resistant Kerite insulation) has not experienced a
water-related failure to date. The black discharge-free insulation was installed in some plants in
the early 1970s, with the brown version replacing the black discharge resistant-insulation in the
mid 1970s. A significant change in the compounding of EPR insulations occurred in the mid
1970s. Before that time, the clays were heat treated to dry them. In the mid 1970s, the dried clays
were also treated with silane, which improved the bonding with the EPR and reduced the water
absorption. This change to treated clay greatly reduced the rate of water-related degradation. As
a result, the Okonite pink insulation produced since that time has not experienced water-related
failures. Accordingly, two of the most dominant cable types in the U.S. nuclear industry where
records are more formally maintained (Kerite black and brown EPRs and Okonite pink EPR)
have not experienced water-related failures. Failures have been experienced in compact design
EPR cables; however, it is not clear at this time whether the failures of these cables is from
insulation degradation, design and manufacturing problems, or installation problems.
7-5
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
Figure 7-2
Generational Differences in Life Expectancy for Ethylene-Propylene Rubber and Other
Insulations
7-6
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
(breakdown) strength is compared to aging conditions [46]. Aging at slightly accelerated voltage
stress at ambient temperature (~86°F [~30°C]) has also been performed. Numerous EPR
insulated cables of various compound formulations and constructions have been studied and
compared using these methods.
What is clear from these studies is that pink EPRs do not behave in a manner similar to XLPE or
TR-XLPE. For the latter two, the dielectric strength drops continuously on aging (at a slower rate
under reduced accelerated conditions, with TR-XLPE having a slower rate than XLPE) with
failure occurring only after long aging times. For EPRs, the dielectric strength drops initially,
stabilizes, and then remains essentially the same without the cable failing within the timeframe
of the test programs. For XLPE, the higher the voltage stress, the more rapid the loss of life; the
higher the load cycle temperature, the more rapid the loss of dielectric strength and the more
rapid the failure during aging in the water-filled tanks. The EPRs do not follow this trend, and
the actual aging conditions are not as significant as for XLPE. Differences in the shields of the
EPR cables can also play a significant role in the results obtained.
So, although one can obtain an apparent projected life for XLPE and TR-XLPE by using the
ACLT, the same cannot be achieved for the pink and brown EPRs. The procedure is not suitable
for brown EPRs (see Section 6.5.2.3, Brown Ethylene-Propylene Rubber) and there is little
interest in performing such tests on older, black EPRs. Years of study by General Cable
Corporation [47–49] have confirmed that ACLT test results for pink EPRs cannot provide
absolute predicted service life—there is no correlation between time to failure and predicted
service life. It has also been demonstrated that these EPR cables have a lower ac breakdown
strength than TR-XLPE cables after aging but that both have long service lives. The ACLT is
apparently of greater value for XLPE and TR-XLPE than for pink EPRs. However, ACLT
testing of pink EPRs might provide information on 1) whether tested cables comply with
industry standards, 2) breakdown strength retention, and 3) relative rankings of different EPRs
(or EPR insulation-shield compositions).
A comprehensive, comparative study of various EPRs was performed under EPRI sponsorship in
which five commercial EPR grade cables of the same configuration, from five different
manufacturers, were studied both in the laboratory and in service (underground). Both operating
and accelerated voltage stresses were used, with cables being aged at ambient. In service, normal
operating conditions were applied; in the laboratory, the cables were aged in tanks at ambient
temperature (~86°F [~30°C]) with water in the strands. Normal and elevated voltage stresses
were applied in both the laboratory and in service. Katz described details of the tests [50].
Figure 7-3 shows the results from aging the five different pink and brown EPRs (and TR-XLPE)
for up to 70 months at rated voltage while in service. Note the differences in initial ac breakdown
strengths, as well as the changes on aging; clearly, the EPRs do not respond alike.
7-7
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
Figure 7-3
AC Breakdown Strength of Medium-Voltage, Field-Aged Ethylene-Propylene Rubber
Insulated Cables at V0 [51]
Figure 7-4 shows the comparative information from aging the same EPRs [50] in service and in
the laboratory at rated voltage (the latter procedure used water in the strands and outside the
cable core). Results at elevated voltage stresses applied in both the laboratory and in the field
have also been generated [50]. All the EPRs are combined in Figure 7-4, solely for the purpose
of easily comparing the in-service and laboratory aging results. The figure shows that water
presence alone can accelerate the aging relative to service aging (although the temperature in
service was likely different from in the laboratory). Figure 7-4 also shows that, in both the field
and laboratory aging, the breakdown strengths remains six to seven times greater than operating
voltage, even though saturation with water has caused a significant drop in breakdown strength
from the dry condition.
Figure 7-4
AC Voltage Breakdown Strength of Combined Ethylene-Propylene Rubber Cables Aged in
the Laboratory and in Service at Rated Voltage
7-8
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
Finally, a legitimate question exists as to whether it is technically appropriate to test all EPR
insulated cable under the same accelerated temperature conditions. Kerite cable is much lossier,
and as a result, temperature control in tanks is difficult. Furthermore, even if adjustments could
be made, this set of accelerated test conditions remains of questionable validity for that
significantly different set of cable materials and construction technology.
Further information on EPR aging can be found in the literature [43, 50, 52–54]. The earlier
studies show that response to aging under water, thermal, and voltage stress yields different
results for different EPRs and for different shield types and that application of accelerated
thermal and voltage stress conditions must take into account the nature of the EPR compound
being evaluated. Valuable information has been obtained from past studies, but the application of
this information requires good judgment.
7.2.5 Conclusions
The reliability of EPR compounds in general, even the early black EPR compounds, have been
quite good. The major reliability challenge to date comes from slow, long-term degradation due
to installation-induced or manufacturing-induced defects, often combined with adverse
conditions (long-term wetting, thermal, and radiation). To date, research has not indicated when
these cables will reach an end of life condition due to normal thermal aging. The main challenge
at this time is to identify and correct or manage the adverse environments that have been
identified as contributors to the failures evaluated to date.
7.3.1 General
NEI 06-05 [11] noted the history of XLPE failures, both unfilled and mineral-filled; 12 XLPE
failures occurred at four plants, eight were at one plant that used a “filled XLPE that was
used only at that plant.” Figure 7-5 shows failure information for these XLPE cables relative
to all cables.
7-9
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
Filled XLPE failures are shown in blue, conventional XLPE failures (“Others”)
are shown in red, and all non-XLPE failures are shown in gray.
Figure 7-5
Unfilled and Mineral-Filled Cross-Linked Polyethylene Cable Failures Compared to All
Failures
7-10
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
degradation continues over time and, over decades, the water tree grows to the point at which the
insulation becomes sensitive to voltage surges that would either break down the polymer or
cause the water tree to convert to an electrical tree, which would grow more rapidly and result in
electrical failure of the insulation.
7.3.5 Conclusions
All XLPE insulated cables for power plant application require oversight and attention, with
particular emphasis on mineral-filled XLPEs with respect to long-term wetting. Dry XLPE
cables should have a long life, provided they are not exposed to high temperatures for extended
periods.
7.4.1 General
Additional degradation modes can affect both shielded and nonshielded cables and could result
in cable failures, regardless of insulation type. Corona discharge, surface PD, and internal PD, as
well as their effects on the cable, are described in the following subsections. In some cases,
visual inspections can identify the degradation (corona discharge damage), whereas in other
cases, maintenance or in-service testing can be used to detect PDs.
7-11
10296123
Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to
Adverse Environments
that forms adjacent to the discharge site or in a dead air space if an air flow is present. The
discharge site is often a grounded surface that is just touching a nonshielded cable or a shielded
cable with an ungrounded shield. Periodic visual inspections for powder residue from corona
discharges could be performed as a required preventive maintenance task for nonshielded cable
in circuit breaker housings and termination boxes and cabinets. These inspections could be done
in conjunction with inspections and maintenance of the equipment being fed by the cable.
Corona discharge is a subset of PD. Corona discharge occurs only at the surface of the insulation.
Section 3.2.5, Surface Corona and Partial Discharge, describes corona discharge and shows a
picture of the powder that indicates its presence.
7-12
10296123
8
TESTING: MANUFACTURING, INSTALLATION, AND
MAINTENANCE OR IN-SERVICE TESTS
This section describes the tests performed on cable during design and manufacture, after
installation, and during operation. It supports the development and revision of purchase
specifications and receipt inspection of the cable. Section 8.6, Maintenance and In-Service
Testing, provides a basis for application of the appropriate test protocol for a particular cable
type and primary failure mechanism for a particular adverse condition.
8.1 Introduction
Tests are performed on cables in the design stage, during and after manufacture, after
installation, and after being in service for some time. The objective of the tests is to verify that
the cable meets the requirements of the applicable standards or specifications and to ensure
reliable operation. The tests can further be described as follows:
Manufacturing tests are conducted in connection with the design of a cable, at various stages
during the manufacture of the cable, and on the completed cable before shipment.
Acceptance tests are conducted when the material arrives at the users’ facilities.
Installation tests are conducted to verify that the cables, accessories, and associated
equipment have been properly installed and can be placed in service.
Maintenance tests are conducted to ensure that the installed cable continues to be suitable for
service and has not aged or been damaged to the point at which in-service near-term failure is
expected.
8-1
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Final tests verify that the product has been properly manufactured and exceeds the requirements
of the specifications to which it has been manufactured.
From the user’s standpoint, the purpose of manufacturing tests goes beyond verification that
specifications have been met. The test data generated form the base from which maintenance
tests can be selected and conducted to track aging and determine if the cable characteristics have
been deteriorated seriously. A lack of knowledge of the as-manufactured state can result in the
selection of maintenance tests that are meaningless, or possibly destructive. For instance,
discharge-resistant medium-voltage cables are not deteriorated by PD at operating stresses and
are not subjected to PD testing during manufacture. A user unaware of this fact could select PD
testing for maintenance purposes. The results would be meaningless and would cause confusion,
at best, as well as leading to possible replacement of a cable that was in good condition.
Acceptance tests are performed by the user before installing the cable and are performed to
partially verify that the relevant specifications have been met and that shipment damage has not
occurred. These tests are generally not as stringent as the post-manufacturing tests.
Installation tests verify that the installation is correct, damage free, and suitable for placing in
service. These tests verify that no gross installation errors that would lead to early failure have
occurred.
Maintenance testing determines whether the cable is satisfactory for continued service. Two
different approaches can be used. Withstand testing represents a go/no-go challenge to the
dielectric integrity of the system, whereas diagnostic testing is intended to characterize the
condition of the aged dielectric. Although these two approaches can be used independently, they
can also be applied in a complementary manner when a higher degree of confidence in the cable
circuit is required. These technologies are described in Section 8.6. More exhaustive descriptions
of their strengths and weaknesses can be found in IEEE Std. 400-2001 and related subdocuments
[51–55, 56].
Although the IEEE standards provide insights into the various test technologies, plant personnel
must be aware that most techniques (and the related literature) were developed for the power
distribution industry. Although much of the technology readily translates to the power generation
industry, the differences in cable construction, preferred dielectric, shield construction, modes of
degradation, and cable system architecture can impact or even negate practices that are
acceptable in the distribution industry.
8-2
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Table 8-1
History of Specifications and Testing of Shielded and Nonshielded Medium-Voltage Cables
Year Standards
Butyl insulated, IPCEA S 19-81
EPR insulated, IPCEA S 19-81
XLPE insulated, IPCEA interim versions of S 66-524 (issued May 71)
1969
XLPE insulated, AEIC 5-69 (shielded medium-voltage cables only)
It is improbable that AEIC 5-69 was called for by many (if any) plant specifications, as it was
generally viewed as an underground distribution specification.
Butyl insulated, IPCEA S 19-81
EPR insulated, IPCEA S 19-81 and interim versions of S 68-516
1972
XLPE insulated, IPCEA S 66-524
XLPE insulated, AEIC 5 (shielded medium-voltage cables)
EPR insulated, IPCEA interim versions of S 68-516
EPR insulated, AEIC 6 (shielded medium-voltage cables)
1975
XLPE insulated, IPCEA S 66-524
XLPE Insulated, AEIC 5 (shielded medium-voltage cables)
EPR insulated, IPCEA S 68-516
EPR insulated, AEIC 6 (shielded medium-voltage cables)
1979
XLPE insulated, IPCEA S 66-524
XLPE insulated, AEIC 5 (shielded medium-voltage cables)
EPR and XLPE (also TR-XLPE)
ANSI/ICEA S-94-649, Concentric Neutral Cables Rated 5-46 kV
ANSI/ICEA S-97-682, Utility Shielded Cables Rated 5-46 kV
2008 NEMA WC 74/ICEA S 93-639, 5–46 kV Shielded Power Cable for Use in the Transmission and
Distribution of Electric Energy
NEMA WC 71/ICEA S 96-659, Standard for Non-Shielded Cables Rated 2001–5000 Volts for Use
in the Distribution of Electric Energy
AEIC CS8, Specification for Extruded Dielectric Shielded Power Cables Rated 5 Through 46 kV
In addition to these standards, U.S. nuclear plants must meet the additional requirements of
IEEE Std. 383 for environmental qualification testing [58, 65] and of IEEE Stds. 308, 603, and
1185 [59–61] for installation requirements.
To list and describe all of the manufacturing tests for medium-voltage cables would result in a
volume beyond the user’s interest. Refer to the appropriate standard (and desired edition) for a
summary of all production and qualification tests. Table 8-2 lists the more common production
tests found in ICEA S-97-682-2000, “Standard for Utility Shielded Power Cables Rated 5,000–
46,000 Volts” [62].
8-3
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Table 8-2
Common Production Tests [62]
Thickness
Volume Resistivity
Thickness
Voids and Protrusions
Stripping Tension
Wafer Boil
Diameter
The test methods and the minimum frequency sampling plan are listed in the standard for each
test in the table. The word partial is used because additional tests unique to the manufacturer’s
process and quality assurance methods are not included. The tests are designed to ensure that the
cable received by the end user fully meets the specifications (including customer specifications)
involved. The following section concentrates on those tests of greatest interest to power plant
operators and those tests that might have an impact on maintenance and field testing.
8-4
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Insulation Type
Ethylene- Ethylene-
Properties Tree-Resistant
Cross-Linked Propylene Propylene
Cross-Linked
Polyethylene Rubber, Rubber,
Polyethylene
Class I, II, and III Class IV*
Dielectric Constant 3.5 3.5 4.0 4.0
2.0 (5–28 kV)
Dissipation Factor (%) 0.1 0.5 1.5
1.5 (>28 kV)
* Discharge resistant
Because these values cover a range of materials, they might not be representative of a specific
material in the category. For instance, the dielectric constant most common for TR-XLPE in use
is 2.4. Should a manufacturer actually obtain a value of 3.5 for a common TR-XLPE, it would be
cause for great concern, investigation, and likely scrapping the cable. Conversely, should a
special TR-XLPE consistently have a dielectric constant of 3.5, it would not be a negative
indicator of tree resistance and might well be superior. Clearly, plant personnel who wish to
know the actual values for the material purchased should require test reports or, at the very least,
consult the manufacturer’s literature for the specific product.
Discharge resistance tests, in accordance with ASTM D 2275-89 with a 60-Hz test voltage of
21 kV for 250 hours, are conducted on discharge-resistant materials [63]. No failure or surface
erosion of the insulation sample is permitted to occur. Although only cables classified as
discharge resistant must pass this test, it has been applied to other materials [33].
Void, contaminant, and conductor shield protrusion tests are conducted on a production basis.
These are destructive tests performed on samples of production cable that allow the insulation to
be physically examined under a microscope. These tests can also be used in forensic
examinations of the cable insulation or shielding when a failure occurs in service. For forensic
8-5
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
examinations of XLPE cables, the wafers of the cable are dyed to allow a water treeing
examination to be performed. For this type of test, a common method of examination involves
“slinkies” cut on a lathe, or simply cutting individual wafers after the conductor has been
removed. A typical examination is shown in Figure 8-1.
Figure 8-1
Wafer Examination for Voids, Inclusions, and Conductor Shield Protrusions
Unfortunately, the process is tedious and involves a small sample length relative to the total
length of the cable for each examination. For translucent insulations such as unfilled PE, XLPE,
and TR-XLPE, the examination includes the entire thickness of the wafer. For opaque materials
such as EPR, only the surface of the wafer can be examined.
Manufacturers take advantage of the fact that when unfilled XLPE and TR-XLPE are heated
sufficiently, they become transparent (virtually clear). With the insulation shield removed and a
sample heated (often in oil such as vegetable oil), a trained observer can see contaminants in the
insulation and the smoothness of the conductor shield–insulation interface. For forensic
assessments, this can also be done using a sample removed from a faulted cable to assess the
relative quality of the cable near the fault. This test has been named the hot oil test and is shown
in Figure 8-2.
Figure 8-2
Hot Oil Test (Insulation Is Clear in the Oil Bath)
8-6
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Filled-strand or blocked-strand conductors are becoming more popular since water penetration
tests on blocked conductors showed their effectiveness. Testing of blocking materials is
conducted in accordance with ICEA T-31-610 [64] and demonstrates the water pressure
capability of the blocked conductor strand in the completed cable.
Flame tests for nuclear plants are called for as required by IEEE 383-2003 [65], which in turn
requires the test method of IEEE 1202 [66]. IEEE 383-1974 defined two methods of flame
testing—a vertical flame test and a horizontal test using an oil-soaked burlap bag. The revision to
the standard requires that the flame test be conducted as defined in IEEE 1202. The test is a
vertical tray-type fire test with flames impinging on the cables near the bottom of the tray. Pass–
fail criteria involve the extent of burning upward in the tray. A typical tray-type fire test is shown
in Figure 8-3.
Figure 8-3
Vertical Tray Flame Test
8-7
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
PD testing is conducted in accordance with ICEA T-24-380 [67] on all cable types except
discharge-resistant cables. The voltage is raised to the maximum ac test voltage. During this
time, the voltage at which PD (if any) begins is called the partial discharge inception voltage
(PDIV). The voltage is then slowly lowered. The voltage at which the PD disappears is the
partial discharge extinction voltage (PDEV). Except for rare occasions when they are equal, the
PDEV is always lower than the PDIV. PD is measured in picocoulombs (pC). Accurate
measurement requires test room shielding and highly sensitive measuring equipment.
Improvements in measurement and cable manufacturing technology have allowed continuous
improvement in the PD requirements with time. A graphic plot of PD tests (called an x-y plot)
has been common since the 1970s. Figure 8-4 shows a typical passing result for a modern cable
in which no PD is measured (5 pC is the accepted lower limit of sensitivity).
Figure 8-4
An Acceptable Partial Discharge Plot
Figure 8-5 shows a failing result. The PDEV does not occur until the cable operating voltage is
reached.
Figure 8-5
An Unacceptable Partial Discharge Plot
8-8
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
In 1994, AEIC 5 [68] was revised for XLPE cables with extruded insulation shields to allow no
measurable PD in cables manufactured to the specification. Before that time, discharge was
allowed, as long as the level did not cross the pass–fail line versus test voltage. The history for
PD requirements in AEIC 5 is shown in Figure 8-6.
Figure 8-6
Partial Discharge Requirement History from AEIC 5
Cables not manufactured to AEIC specifications might have met less stringent PD requirements.
Plant personnel must consult the proper specification and edition to determine the requirements
for the cable of interest.
One special case involves cables that used semiconducting tape insulation shields. These were
subject to less stringent requirements, even in AEIC specifications.
Discharge-resistant cables (Kerite black and brown EPR) are not PD tested in accordance with
industry standards. The insulation is not deteriorated by PD at and above operating voltage and,
if tested, it would likely show some discharge.
Production PD testing techniques date back to the 1950s [33], and the acceptance criteria have
become tighter with years of progress. This has resulted in the emphasis for most EPR
insulations shifting from greater discharge resistance to constructions demonstrating less
measureable PD [33]. Thus, the PD history for older EPR cables is not as great a cause for
concern. This reported history seems to be supported by the fact that AEIC 6-73 [69] required
the U-bend discharge test as an EPR qualification test, but it was dropped from following
editions. Modern EPR insulations are, in general, more discharge resistant than XLPE. Plant
personnel considering the use of PD maintenance tests (in-plant field tests) for diagnostic
purposes might obtain meaningless results unless the test is appropriate for the cables involved.
Insulation resistance is now an optional test appearing as a qualification test in both ICEA S-93-
639 and ICEA S-97-682 [57, 62]. In the past, this completed cable test, conducted with a dc
voltage source of 100–500 volts applied for 60 seconds was used to calculate an insulation
resistance constant. The required specification constants were so low as compared to the actual
constants for medium voltage XLPE and EPR cables that they were meaningless. A difficulty
arose when some field personnel attempted to use these constants to determine a passing
insulation resistance for dc high-potential tests.
8-9
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
8.3.1.3 Final Electrical Tests for Nonshielded Cables, 2001–5000 Volts Without
Metallic Sheath or Armor
The ac voltage wet test is the most common test for nonshielded cables. It requires water
immersion for a sufficient time to ensure complete water penetration into the cable reel to
provide an electrode to test from the conductor to the made electrode. An ac voltage of 13 kV is
applied for five minutes. An alternate dc wet test may be substituted.
An equivalent high-potential maintenance or field test is not possible unless a made electrode,
such as flooding the ducts or conduits enclosing the cables, is practical. This is seldom the case.
Figure 8-7
Dry Specimen Design Qualification Tests
Several manufacturers regularly conducted wet tests on cables, but these were not standardized.
AEIC CS-5-79 for XLPE first required a wet qualification test, called the accelerated
water/electrochemical treeing test [74]. This was quickly followed in AEIC CS-6 for EPR [69].
The current flowchart found in ICEA S-97-682 is shown in Figure 8-8 [27].
8-10
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Figure 8-8
Wet Specimen Design Qualification Tests [27]
These wet (or treeing) qualification tests had a good history of identifying combinations of
insulation and shielding materials in completed, unjacketed, 1/0 AWG (53.5 mm2) aluminum or
copper, 100% insulation level, 15 kV cable that resists water treeing.
Other qualification tests in the standard include jacket material qualification, continuous
vulcanized extrusion qualification, and an “other” qualification test category.
ICEA S-93-639 [57] covers many of the cable designs found in ICEA S-97-682 [27]. However,
ICEA S-93-639 for medium-voltage cables does not have the same qualification test regimen as
ICEA S-97-682. This should not necessarily be a cause for concern, as it will have been qualified
by default when the same core is used to make cables to both specifications. The exception is for
core designs in ICEA S-93-639 that do not appear in S-97-682 (such as having a semiconducting
tape insulation shield). Such designs are not likely to be desired for future needs of plants. ICEA
S-93-639 does have several cable features that users may wish to use [57]. This makes it
necessary to reference both specifications to cover all of the desired requirements.
Low-smoke zero-halogen jackets are growing in popularity. ICEA T-33-655 provides specific
tests for acid gas, halogen content, and smoke generation [75].
The importance of requesting and retaining manufacturer’s qualification and test reports
concerning the basic cable design and the specific manufacturing run cannot be overemphasized.
Industry standards might contain pass–fail criteria that cover a range of materials and cable. In
that case, the pass–fail number might not be representative of the cable the user has purchased.
Tests reports can be requested, including qualification test data and showing actual test data
numbers. These may provide important baseline information for future on-site field testing.
During the construction, it was common practice for the consulting firms involved to require
technical data from the manufacturer, much of which is retrievable from plant records. That data
is often helpful in gaining an understanding of the characteristics of the installed cable.
Manufacturers’ literature is an excellent source of test results developed with different cable
materials and designs. The literature should be for the cable and time vintage of interest. Some
might object that it could have some commercial bias. Although that might be true, the data will
at least be representative of the cable and far superior to using generic information or data from
an entirely different era.
8-11
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
8-12
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
8.6.1 Introduction
Various in-service tests have been used in recent years to gain an understanding of the state of
aged, installed cable systems. However, only a limited set of tests is applicable to the cable types
used in plants. This tests include tan δ, dielectric spectroscopy, VLF and 60-Hz withstand, and
PD. The selection of the appropriate test depends on the insulation, the type and condition of the
insulation shield, and the rating of the cable. The selection also depends on the particular failure
mechanism that is of concern.
The tests described in this section apply to cables with insulation shields. An insulation shield
provides a consistent ground plane that enables electrical testing that indicates the condition of
the insulation. These tests do not apply to nonshielded cables. The lack of a ground plane for
nonshielded cables has led to a search for techniques that might provide an indication of the state
of the insulation. Plant cables that are nonshielded can be located in trays or in conduit. The
technique of filling the duct or conduit with water that serves as the ground plane has
been studied in the laboratory, but it has not been applied to field testing. One problem with
using water as a ground plane is that the jacket of the nonshielded cable is included in the
electrical test. The presence of a jacket can result in confusing test results.
On-line diagnostics testing (see Section 8.6.3, On-Line Diagnostics Assessment) purports to be
suitable as an electrical diagnostic test method for both shielded and nonshielded cable because
the diagnostic technology approach is different—it detects the magnetic field resulting from
current flow in either the shield or the conductor of a cable. For nonshielded cable testing,
special sensors are required to eliminate discharges between the sensor and the cable under test,
as well as for safety [76]. For 4-kV circuits, PD-related degradation might not be the dominant
failure mechanism due to the low electrical stresses in the insulation system and its interfaces.
Accordingly, there might be no signal to detect (that is, no PD or pre-PD) through on-line
assessment.
Montinari [77] warns that the applicability of either on-line or off-line PD measurement is of
questionable value for wetted medium-voltage cables. PD measurement would detect a failure of
a wetted cable only after a water tree converts to an electrical tree. In the case of on-line
measurements, the need for high signal-to-noise ratio detectors to enhance noise detection can
lead to identification of false negatives, whereas external noise detected by the detectors can
result in false positives. Montinari states that the value of on-line detection might be as a screen
to identify potential problems that should be followed by off-line validation combined with tan δ
testing for distributed effects, especially for cables that have been wetted.
Thermal and radiation aging (state of the cable) caused by the environment can be assessed by
indenter testing, which evaluates compressive modulus. As elastomers age, their physical and
mechanical properties change, and the materials generally harden. The indenter is
nondestructive. In the test, a small, instrumented probe is pressed against the side of the cable at
a constant velocity while force is measured. The modulus is the change in force divided by the
change in probe travel. As the material hardens, the modulus increases. Test data are available
8-13
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
for common rubber materials in the EPRI Cable Polymer Aging Database (1011874) [78]. The
EPRI report Initial Acceptance Criteria and Data for Assessing Longevity of Low-Voltage Power
Cable Insulations and Jackets (1008211) [79] provides a basis for establishing the degree of
aging of the polymer by using the indenter and other technologies.
The indenter cannot provide any indication of electrically induced degradation but rather would
indicate whether the cable had degraded from thermal or radiation exposure. The indenter might
be of limited use in evaluating ohmic heating, given that the jacket of the cable is in cooler
ambient air. If the ambient temperature is normally cool, hardening of the jacket is a strong
indication that ohmic heating is significant.
8-14
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
As a result of this experience, the industry has followed two paths: 1) seeking a substitute for dc
withstand testing that would perform the same role without causing a loss of service life, and
2) seeking alternative diagnostic tools that can estimate future performance again without
shortening the life of the tested system. Current industry standards discourage the use of dc high
voltage for all types of extruded cable as a means of testing aged insulation; however, it is still
recommended for use on paper-insulated, lead-covered cable [51, 55].
8-15
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
reliable a tool as the use of 50-Hz or 60-Hz test sets. Considering that there is comparable value
in methods, the portability issues for line frequency testing (which can make it impracticable for
testing in many locations) makes VLF and variable-frequency testing an attractive choice.
Figure 8-9
Derivation of Dissipation Factor (Tan δ) Measurement in Insulation
As with other testing methods, tan δ testing can be performed using line frequency, VLF, or
variable-frequency methods. EPRI report 1020805 recommends the VLF method and provides
assessment criteria for this method [5].
VLF tan δ measurements are desirable because small portable test sets can be used. VLF tan δ
measurements are made at several predetermined voltage level steps, starting at 0.5 V0 and
proceeding in 0.5 V0 increments up to a maximum of 2.0 V0, provided no unacceptably high
values are encountered. Preprogrammed equipment can be used or test operators can change the
voltages at specific times. In a 0.1-Hz test, one cycle takes 10 seconds. Accordingly, the test
duration must be long enough to take sufficient measurements to give valid results. Step
durations vary from 30 seconds to 5 minutes. The test equipment automatically calculates and
records the average and standard deviation of the readings. Although the absolute value of tan δ
8-16
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
provides some indication of aging, the difference in the tan δ readings at 2 V0 and 1 V0, the
degree of degradation, is considered a stronger indicator. Table 8-4 provides criteria for XLPE
insulation as being good, aged, or highly deteriorated. EPRI report 1020805 provides assessment
criteria for EPR and butyl insulations [5].
Table 8-4
IEEE Standard 400 Criteria for Assessment for Cross-Linked Polyethylene Insulated
Cables
Figure 8-10 shows a typical portable test device that is suitable for testing the relatively short
cables in power plants.
Figure 8-10
Typical Variable-Frequency, Very-Low-Frequency Portable Test Equipment for Performing
0.1-Hz Dissipation Factor (Tan δ) Testing
Courtesy of HV Diagnostics, Inc.
8-17
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Figure 8-11 shows the voltage dependence of the dissipation factor at 0.1 Hz for new and
service-aged XLPE-insulated cables.
Figure 8-11
Voltage Dependence of Dissipation Factor for New and Aged Cross-Linked Polyethylene
Cable
Another way of viewing the data is shown in Figure 8-12. Here the increase of dissipation factor
with voltage stress is shown as a bar (the color code is the same as that used in Figure 8-11).
Figure 8-12
0.1-Hz Dissipation Factor of Cross-Linked Polyethylene-Insulated Cables
8-18
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
This test method is global and, therefore, determines the general condition of the cable under test
rather than identifying and localizing weakened sites. One drawback is that numerous small
water-aging degradation sites can respond in the same manner as one or a few severe (long)
water-aging degradation sites, with the latter being the condition of most importance. It is
believed that the severe degradation sites will cause the differential measurement between V0
and 2 V0 to be larger, rather than just causing the overall measurement to be larger but stable
with increasing voltage. Advantages of 0.1-Hz dissipation factor testing include the following:
The test identifies the existence of water trees and water-related degradation before the point
of conversion to electrical trees.
Test equipment is portable and does not require the use of a van.
Disadvantages of 0.1 Hz dissipation factor testing include the following:
Because an overvoltage is applied, a failure of a severely aged insulation system is possible
during testing.
Because tan δ is an off-line test with elevated voltage, the cables must be disconnected from
their loads (such as motors and transformers).
Testing of dissimilar, interconnected cable types, such as XLPE and EPR, will require
separation to determine the state of both insulations correctly. In this case, normal EPR
results could mask problems in the XLPE segment.
8-19
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
The data developed have been characterized in several manners in terms of dielectric response
of cables without water trees, and several categories for XLPE with water trees XLPE [89],
as follows:
Low-loss linear permittivity. Characterized by an almost frequency-independent
capacitance. A linear response across a wide frequency band indicates that little or no aging
has occurred.
Voltage-dependent permittivity. Characterized by increases in both capacitance and
dissipation factor with increasing voltage, but essentially, the increase is independent of
frequency. This response is characteristic of cables in which water tree deterioration is
significant, but the trees have not penetrated the insulation wall.
Transition to leakage current. At low voltage levels, the response is similar to the voltage-
dependent permittivity response, but at higher voltage levels, the dissipation factor losses
increase with voltage and higher leakage current occurs at both high and low voltages.
Numerous dielectric spectroscopy curves are presented by Werelius et al. [89].
EPRI report 1001727 [88] refers to responses defined as different types:
Type 1. A voltage-dependent increase of dissipation factor and capacitance, almost
independent of frequency
Type 2. At low voltage levels, a type 1 response, but at higher voltage levels, a nonlinear loss
versus frequency characteristic
Type 3. At low voltage levels, a nonlinear type 2 loss versus frequency characteristic
A set of curves for a shelf-aged cable (never energized) is shown in Figure 8-13 [88, 89]. In this
example, there is a slight change in dissipation factor with frequency, but change with voltage is
essentially constant at any frequency.
Figure 8-13
Dielectric Spectroscopy Measurements for Shelf-Aged (Never Energized) 15-kV Cross-
Linked Polyethylene Cable
8-20
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
In essence, for unaged extruded cables, capacitance and losses are linear with voltage increases,
but higher losses and increasing nonlinear responses are associated with aged cables containing
water trees. Further details are provided in the EPRI report Guide for Non-Destructive Diagnosis
of Distribution Cable Systems (1001731) [90]. Dielectric spectroscopy is currently offered by at
least one testing company. Skilled practitioners are required to interpret the results.
Figure 8-14
Calibration Equipment for Off-Line Partial Discharge Testing
8-21
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Figure 8-15
Test Setup for Performing 60-Hz Partial Discharge Measurements
A typical test involves the exposure of the cable system to a brief overvoltage, up to 2V0. During
the period of voltage increase, the system looks for the onset of PD. The voltage at which this
occurs is known as the partial discharge inception voltage (PDIV). The elevated voltage should
be maintained no longer than is required to obtain the necessary partial discharge data. A second
and equally important point in the test is the partial discharge extinction voltage (PDEV). In a
good cable, the PDIV is greater than the peak test voltage, so that PD is not initiated. In an aged
cable, PD, if present, will initiate well above the operating voltage but at less than the peak test
voltage. The PDEV should also be well above operating voltage, but it will be significantly less
than the PDIV. If the PDIV is too close to the operating voltage and the cable is returned to
service, a small voltage surge during operation can start the PD, and the PDEV is likely to be
below the operating voltage, so that PD does not stop. Insulation failure could then occur in
weeks to months.
The level of overvoltage applied and the duration of its application is important. The overvoltage
and duration should be as low and as short as possible—but not so low or so short that the test
misses the defect. Figure 8-16 shows a voltage test profile for the short duration test, with an
overvoltage of 2.5 per unit and a test time of <12 seconds. This is as short as is reasonably
possible. Not all testing is performed in such a short period. Figure 8-17 shows the diagnostic
data capture approach using reduced dwell time at test voltage.
8-22
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Figure 8-16
Test Profile for Short Duration Off-Line 60-Hz Partial Discharge Testing
Figure 8-17
60-Hz Partial Discharge One-Step Diagnostic Data Capture Approach
A sinusoidal 0.1-Hz test set can also be used to evaluate PD at defect sites. As with 60 Hz, the
traveling wave is used to determine the location and magnitude of the discharge and, as with
60 Hz, defect locations in the cable, splice, or termination can be located. The PD acquisition
system is coupled to the cable to be tested through a coupling capacitor that detects the high-
frequency electromagnetic impulses created by any PD. The PD signature and PD site location
can be detected using this method.
The VLF sine wave generator is portable. The cable system must be isolated for testing. A
typical test will last for less than 10 minutes with peak applied voltage levels of 1V0 to 2V0.
Compared to 60-Hz test equipment, the required reactive power for testing a cable at 0.1 Hz
decreases by a factor of 600. A reduction in the kilovolt-ampere requirements for the test
transformer and a corresponding weight reduction occur; therefore, improved portability makes
the 0.1-Hz dissipation factor equipment of practical interest for in situ measurements. In
addition, very low frequency leads to more direct and quicker channel growth of electrical trees
[86].
8-23
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
If the failure mechanism for a cable issue results in PD, the PD test can determine whether the
cable is free from PD at the test voltage, and if not, it will indicate the PDIV, PDEV, and
location. In addition, some testers use phase-resolving methodology that provides an indication
of where the PD occurs on the ac wave. The phase angle location and density provide insights
into whether the discharge is occurring along a surface, at a shield–to-insulation interface, or in
the insulation itself. This information helps determine the importance of the PD and whether it
will cause failure in a short period.
Advantages of the VLF PD test are the following:
The location of PD activity can be detected and measured.
Cable system insulation condition can be graded as good, defective, or highly deteriorated
when the measurement data are compared against historically established cable system PD
data that have been developed by the same test method.
Compared to 60-Hz ac PD detection, the VLF test equipment is more compact and
transportable (no van is required).
Power requirements are comparable to standard cable fault-locating equipment.
If desired, a combined VLF withstand test and PD test system can be used to monitor the
dissipation factor or PD activity, or both, during a 15- to 60-minute withstand test procedure.
Phase-resolution technology allows the type of PD to be identified, which helps in
understanding the nature of the failure mechanism.
Potential limitations and disadvantages of this test method are the following:
There are length limitations (1–2 miles) for accurate measurements. These should not cause a
problem for most plant circuits, given their relatively short lengths.
Deteriorated helically wrapped metal tape shield conditions and cable insulation attenuation
characteristics can influence the propagation of PD pulses along the cable, thereby
influencing detection and locating of PD sites. This could significantly affect detection of PD
in wet segments of cable, where wetting would eventually cause corrosion of the tape shield,
a common plant shield configuration.
PD data collected with VLF might not be comparable to what is obtained with 60-Hz power
frequency data. Because of the slow repetition rate, fewer cycles occur during the test, and
fewer PDs are observed per unit period.
Testing of mixed systems (XLPE joined to EPR, for example) can complicate interpretation.
Detecting the existence of water trees during PD testing is unlikely. The string of micro-
voids connected by oxidized channels that make up a water tree will not discharge under
elevated voltage stress when ion-containing water resides in the tree channel. In addition,
discharges require an air gap that could not exist if water filled the void.
The cable must be taken out of service for testing to be performed.
8-24
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
The test method and the significance of proper grounding are described in IEEE Std. 400.2-2004
[56]. The test times and voltage stresses have been questioned and are the subject of ongoing
study; however, the voltage levels are generally between 2V0 and 3V0 for cables rated between 5
and 35 kV, and the test times range from 15 minutes to one hour.
The two main types of 0.1 Hz waves generated—sinusoidal and trapezoidal—are described in
the remainder of this section. For further details see IEEE Std. 400.2-2004 [56] and the EPRI
report High Voltage, Low Frequency (0.1 Hz) Testing of Power Cable (TR-110813) [92].
The sinusoidal voltage wave is generated by amplitude modulation of a higher-frequency signal
transforming up to the required high voltage and demodulation of the signal to produce the
0.1-Hz wave. A schematic of the sinusoidal wave is shown in Figure 8-18.
Figure 8-18
Nominal Very-Low-Frequency (0.1-Hz) Sinusoidal Waveform
8-25
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
The test set voltage is raised to the specified value [56, Table 5] for the specified period. If no
failure occurs, the voltage is returned to zero, the cable system and test set are discharged, and
the cable system is grounded. If a failure occurs during the test (that is, the test voltage drops) the
test set should be turned off to discharge the cable system and the test set.
The trapezoidal wave (cosine rectangular) is generated as follows: A dc test set forms the high
voltage source with a dc to ac converter (consisting of a high-voltage inductor and a rectifier),
changing the dc voltage to the VLF ac signal. The signal has polarity reversals at five-second
intervals, which generates a 0.1-Hz bipolar pulse waveform. When testing with a pulse bipolar
waveform, the test voltages should be peak voltages rather than rms. A schematic of the
trapezoidal wave is shown in Figure 8-19.
Figure 8-19
Trapezoidal (Bipolar Rectangular) Very-Low-Frequency (0.1-Hz) Waveform
8-26
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
8-27
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
runaway at the degradation site. Thermal runaway would occur if the leakage current in a low-
resistance zone in the insulation increased to the point at which the heat generated in the
insulation could no longer be dissipated by the surrounding medium and excessive insulation
temperature occurred.
Montanari [77] cautions that false positives could occur due to the dependence on high signal-to-
noise ratios required for the sensors used for on-line evaluations. In addition, errors can occur
due to the discontinuous nature of PD signals. Extending sampling time or continuous
monitoring can reduce these errors, but that is not necessarily a practical solution. The
occurrence of false positives can be caused by external noise detected by the sensors. The on-line
assessment physically places sensors at readily accessible locations along the cable circuit. For
example, for underground testing, access through manholes is used. Data from the signals are
transmitted to the home office of the vendor and, after analysis, the condition of the cable system
between sensors is determined. The sensors are moved to different locations as testing continues,
until the entire desired length of cable is assessed. This approach allows localized regions to be
evaluated (segments between sensor placements). Because cables age at different rates along
their lengths, independent of age, the technology is able to distinguish between the degrees of
degradation in segments of the cable. The technology estimates the condition of all equipment
between sensors, such as joints or terminations.
High-frequency signals during cable system operation result from current flow in the cable
shield and conductor. The magnetic field resulting from the current flow is measured. The
sensors are designed to pick up energy from the magnetic field and detect signals in the high-
frequency range.
The sensors use both inductive and capacitive coupling, which allows signal detection over a
wide frequency range. The sensitivity of the inductive coupler depends on the construction and
condition of the shield and the size of the cable. This is in contrast to capacitive coupling, in
which the signal sensitivity depends primarily on the cable capacitance. The sensors are
U-shaped, and various sizes exist to cover commonly used cable configurations. Slight corrosion
of helical tape shields will lead to significant attenuation of high-frequency signals along the
length of the cable.
Signals detected can be due to PD or pre-discharge events. The literature generally refers to all
signals detected in this manner as partial discharge detection. Significant signals can result from
mechanical damage, foreign impurities, poor splice or connection regions, water intrusion into
splices or terminations, or voids. This approach can, therefore, be described as seeking to
1) acquire signal information due to defects, 2) process the signal information to reject noise,
and 3) analyze the signals. The U.S. supplier of this technology provides results in terms of the
quality of the cable. Technical detail concerning the nature and amplitude of the signals is not
provided.
Figure 8-20 shows the sensors for the detection system and the data acquisition system.
Figure 8-21 shows the on-line condition assessment data acquisition system. Figure 8-22 shows
signal patterns and the influence of various defect types.
8-28
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Figure 8-20
Sensors for On-Line Signal Detection and Data Acquisition System
Figure 8-21
On-Line Condition Assessment Data Acquisition System
8-29
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
If pre-PD or PD is present, the patterns indicate the causes of the signals, allowing personnel to
determine their significance relative to future performance. The pulse pattern is more significant
in determining relevance to future performance than is the apparent discharge magnitude. The
latter, while of interest and value, does not allow distinguishing between defect types that cause
the specific signal.
After evaluation by the vendor’s technical personnel, the condition of the cable system is placed
into categories. The most significant level suggests that immediate action is required, whereas
intermediate levels suggest potential retesting after various timeframes. The lowest level
represents cables that show little or no aging.
In addition to the benefit of detecting defects at operating voltage while the cables are energized,
the on-line method purports to provide the following advantages compared to off-line testing
methods:
The type of defect can be identified.
Testing is performed at operating temperatures because cables undergo physiochemical
changes as the temperature is increased. Accordingly, on-line assessment evaluates the cables
under the conditions of operation, not under conditions following cooldown.
Testing does not require disconnecting and reconnecting of terminations.
Disadvantages of the on-line diagnostic testing are as follows:
Defects such as water treeing in wetted cables cannot be detected.
Analysis can be performed only by the vendor, so that the technology cannot be transferred
to the utility.
8-30
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
This methodology does not overstress the insulation, so that a gross defect close to
breakdown but not producing PD signals might not be detected. Such defects will not be
affected by operating voltage but can cause failure if a significant surge voltage occurs.
If the ultimate failure mechanism is thermal runaway and no PD is expected, this
methodology will not detect the aging of the cable.
This assessment method attempts to identify extremely small signals, in particular, pre-PD
signals. Accordingly, attenuation of the signal can impede the ability to identify adverse
conditions. When attenuation is an issue, the sensors must be applied at several spots along the
length of the cable to ensure that attenuation does not mask problems. Except for short cables,
making measurements at both ends of the cable is highly desirable. Even for cable lengths in the
vicinity of common large power plant cables (500 ft [164 m]), testing from one end of the circuit
might not fully assess the circuit.
Fault Locating
Manufacturing Acceptance Installation Diagnostic
And
Test Test Test Test 1
Troubleshooting
Insulation Spark
X
Gap Test
Conductor DC
X
Resistance
Partial Discharge X2
AC Withstand Test X3
DC Insulation
X4 X4 X4 X4 X4
Resistance Test
DC Withstand Test X5
60-Hz or Very-Low-
Frequency X5, 6 X5, 6 X5. 6
Withstand Test
On-Line Partial
X
Discharge / Signal
8-31
10296123
Testing: Manufacturing, Installation, and Maintenance or In-Service Tests
Fault Locating
Manufacturing Acceptance Installation Diagnostic
And
Test Test Test Test 1
Troubleshooting
Off-Line Partial
X X7 X
Discharge
Resistance Bridge
X8 (Fault Location)
(Murray Loop Test)
Time Domain
X8 (Fault Location)
Reflectometry
Notes:
1. For initial diagnostic testing of aged cables, it is recommended that spare cable, accessories, and installation
procedures be available before testing should a replacement be necessary.
2. This test is generally performed between 500 and 5000 Vdc. Its value is in troubleshooting and verifying that the
cable under test is capable of withstanding higher test voltages used for PD, tan δ, and high-potential tests. It
should not be used as proof of suitability for service.
3. This is a 5-min, 200 V/mil test.
4. This test, described in Section 8.3.1.2, is not performed on Kerite brown EPR cables because the lossy nature of
the design makes the test meaningless. An acceptance criterion is <5 pC, but some exceptions are noted in the
text.
5. An ac withstand test would be the preferred method, except for paper-insulated, lead-covered cables. VLF
requires a smaller test set than 60 Hz testing. DC high-potential testing of new installations is acceptable from the
standpoint of not damaging the cable, but it is not recommended for service-aged cables. Test values should be in
accordance with IEEE Std. 400.
6. Test values after manufacture are generally 80% or factory test value for acceptance and installation tests and
60% for maintenance tests.
7. PD testing will identify degradation of cables. Test success depends on the quality of the metal shield, so that
attenuation due to the high impedance of the shield will not preclude measuring the PD signal.
8. If the cable fault is solidly grounded or can be carbonized sufficiently with a high-potential set so that it is solidly
grounded, this test is highly effective in determining the location of a fault.
9. The capacitance ratio test is an effective means of locating a fault for a cable with an open circuit.
8-32
10296123
9
CABLE AGING MANAGEMENT PROCESS
The cable management process involves obtaining information on the cables installed at the site,
determining the adverse environments and service conditions, knowing how those conditions
affect the specific cable designs, and applying the best test methods available to determine the
degree of degradation. This section describes the diagnostic approaches and the actions to be
taken to manage cable aging.
9-1
10296123
Cable Aging Management Process
9-2
10296123
Cable Aging Management Process
PD testing can also be useful for 13 kV UniShield cables that have distributed shield wires that
should not be subject to a high-impedance shield condition if the wires corrode slightly. In
addition, PD testing can be useful in identifying long-term aging (after 50–60 years) in dry cable
sections of all types of shielded cables. However, such testing is not needed until operating
experience indicates that the dry sections of cable are exhibiting signs of aging.
In the case of most EPR cable designs, PD might never occur or might occur only when the
cables are close to failure. In the latter case, no PD would be detectable for decades in the cable’s
life. Detection would be possible only during a short period near the end of the cable’s life.
Periodic testing for PD with frequencies on the order of once every six years would have a low
probability of usefully detecting PD, because the window between the start of partial discharging
in the insulation and the point of failure is a fraction of the period between tests. Detection of PD
would be happenstance rather than an expected outcome. If detected, the cable in question would
likely need immediate or near-term replacement because the presence of PD in the insulation
would indicate that the final stage in the degradation mechanism is in progress.
Brown EPR insulated cables are designed for discharge resistance, and some of these cables
exhibit levels of PD that would be unacceptable in other cable designs but cause no significant
degradation. Use of PD testing on these brown EPR insulated cables can be of little use and can
cause needless concern about the longevity of the cable.
9-3
10296123
Cable Aging Management Process
9-4
10296123
Cable Aging Management Process
An acceptable alternative might be to use industry operating experience of forensic analysis done
on similar cables under similar conditions. Forensic analysis performed at another plant on
shielded cables with the same insulation design and similar service conditions can also be a
suitable source of information for evaluating the condition of a site’s nonshielded cables.
Alternatively, the third option, time-based replacement, could be used. A time-based approach
for nonshielded cable could be broken down into different strategies as follows:
Replace all cables.
Prioritize replacement based on various risk factors (safety, business, regulatory).
Prioritize replacement based on environmental risk factors such as thermal, radiation,
chemical, or wetted environments (localized adverse environments).
Replace all cables at an arbitrary end-of-life. This strategy could be used in plants with a
small population of wetted medium-voltage cables. Prioritized replacements could be used
for plants with larger populations of wetted cables. It would be prudent to perform laboratory
evaluations of some of the removed cables to determine whether a delayed replacement
schedule were possible based on as-found condition of the removed cable.
When replacing a nonshielded system, replacement with a shielded cable design should be
considered. However, the design of the system must be considered to ensure that use of a
shielded cable is acceptable. In addition, water-impervious cables should be considered as
replacements. The use of shielded or water-impervious cables can be precluded by the design of
the duct system, pulling restrictions, and cable availability. Water-impervious designs are
available that use a corrugated tape shield that is folded over the insulation system and glued
shut. Additional information is provided in the EPRI report Plant Support Engineering: Common
Medium Voltage Cable Specification for Nuclear Power Plants (1019159) [12].
EPRI will begin submergence qualification for certain cable types based on operating experience
and on-going qualification as described in IEEE Std 323-1974 [98], Sections 5.2 and 5.6. A
cable that has been submerged for nearly its entire life will be removed from service and
subjected to electrical and physical characterization to provide the operating experience portion
of the qualification. Cable specimens will then be prepared from the remainder of the cable and
will be subjected to submerged, energized accelerated aging in a laboratory to provide the
ongoing qualification. The rate of acceleration is expected to reasonably low and require an
extended period to complete. The submergence qualification is expected to yield a basis for the
expected (qualified) life of the cables so that a basis for continued use or an appropriate
replacement interval will be determined.
None of these strategies preclude the possibility of a cable failure, so it is highly recommended
that plant personnel take action to be prepared should an in-service failure occur.
9-5
10296123
Cable Aging Management Process
discharge-resistant insulation (cables produced by Kerite) could have lower priority. Normally
energized cables having limiting conditions for operations, such as those connecting startup and
auxiliary transformers to busses, should have high priority, as should safety-related cables,
especially those that are continuously energized.
The prioritization method should include consideration for criticality of the connected
components, as defined by INPO AP-913, Equipment Reliability Process Description, [99] and
risk significance, as defined by 10CFR50.65 (Maintenance Rule) [97]. Cables that are not
Maintenance Rule risk-significant can be given low priority. However, care must be taken that
protective relaying and circuit breakers associated with non-risk-significant cables and loads are
included in plant maintenance and test programs to ensure that they function when called on
should a non-risk-significant cable fail. Prioritization could be further associated with
environmental stressors (wetted environment, elevated temperature) and their severity, the cable
and accessory design (water impervious, insulation type, and so on), site-specific and industry
operating experience, and duty cycle (normally energized and loaded, normally energized and
unloaded, periodically energized, or normally de-energized).
Weighting factors can be assigned for each of the factors that plant personnel consider important
to longevity. The summation of the weighting factors would indicate the relative priority of
cables with respect to aging. This prioritization could be used as a method to prioritize cable
testing, refurbishment, and replacements.
9-6
10296123
Cable Aging Management Process
9-7
10296123
Cable Aging Management Process
9-8
10296123
10
RESPONSE TO CABLE FAILURES
This section describes the actions and decisions that might be required if a medium-voltage
cable fails.
10-1
10296123
Response to Cable Failures
In some cases, acceptance criteria for a specific EPR formulation can be difficult to obtain.
When no data are available for a particular cable manufacturer’s insulation, testing of a cable in a
dry application is a possible way to establish a baseline for comparison. For example, a dry cable
that connects to buses can be relatively easy to test.
10-2
10296123
Response to Cable Failures
If PD testing is used and attenuation of the signal is not an issue, PD test set will provide the
location of the fault.
If a fault causes a solid grounding of a cable, a time domain reflectometer (TDR) can be used to
detect the location of the fault. However, the fault often burns out the insulation shield in its
vicinity, leaving a high resistance to ground and limiting the ability to be detected by the TDR.
High-voltage test sets are often used to burn the fault in so that it can be detected by a TDR or a
fault-locating bridge. A thumper can also be used. A thumper charges a high-voltage capacitor
and then discharges it into the cable. The thumper causes an audible sound (a thump) that can be
heard from above ground at the location of the fault. However, a thumper will do additional
damage at the failure site and sometimes destroy the insulation to the point at which
identification of the cause of failure is not possible. Other techniques, such as the use of a
Murray bridge or a capacitance bridge, can also be used [6].
10-3
10296123
Response to Cable Failures
The forensics laboratory will use the dry or unaged cable to determine the characteristics of a
good section of cable for comparison to the damaged section. The section adjacent to the fault
will allow the laboratory to determine whether the adjacent phases had similar deterioration or
whether the faulted phase had a unique problem. Removal and handling of the specimens should
be supervised to ensure that the specimens are properly handled and preserved. Electricians are
experienced in installing cables but are often not trained to preserve cables or terminations that
have failed.
In addition to the cable specimens, the laboratory should be provided with the physical and
operational conditions that the cable or cables experienced in service. The following information
should be provided to the failure analyst, if possible:
Cable manufacturer
Cable ratings
Circuit length
Cable routing details (penetrations, bends, cable tray, conduit, direct buried, ducted, and
so on)
Number of conductors per phase
Accessories and splices
Time in service
Operating voltage
Operating current
Operating duty cycle (continuous or standby)
Environmental conditions (wetted, flooded, near heat source, and so on)
Operating parameters at the time of failure (plant or system transient, equipment startup,
shutdown, change in demand, and so on)
Related operating experience
Forensic analysis of the cable should be performed by a laboratory that specializes in cable
analysis. Although many sites have access to corporate laboratories that provide failure analysis,
these labs might not have the knowledge or capability to evaluate the physical and electrical
properties of the cable specimens. For example, corporate laboratories might have experience
only with XLPE insulation used in distribution systems rather than EPR with helically wrapped
tape shields used in power plants.
10-4
10296123
Response to Cable Failures
The properties of the cable that should be evaluated by the laboratory will vary by the cable type,
the manufacturer’s design, and the nature of the failure. The following physical attributes should
be measured, if applicable, depending on the design of the failed cable:
Visual condition of the cable and the failure location
Jacket type, thickness, and condition
Metallic shield type, if applicable (copper tape, corrugated tape, drain wires, and so on) and
condition
Insulation shield type and condition
Conductor shield type and condition
Insulation type, thickness, and condition
Conductor type and condition
Tensile strength of jacket and insulation
Elongation of jacket and insulation
Moisture content of jacket, insulation, and conductors
Microscopic examination of insulation for imperfections
In addition to the physical properties, the following electrical properties of the cable should be
gathered, as applicable:
Jacket electrical resistance (ohms/cm)
Semiconducting shield resistance (if shield is different from jacket)
Dissipation factor at 60 Hz or VLF
PD
Insulation resistance (normalized to 1000 foot of cable) at 5000 Vdc (see Appendix F,
Insulation Resistance Test Measurements: Their Value and Limitations)
AC voltage breakdown test or dissection of failure areas
Evaluation of local insulation resistance of the insulation between the conductor and the surface
of the insulation can be useful in assessing water-related deterioration of rubber insulation. The
insulation resistance between the conductor and a probe run over the surface of the insulation is
used to detect areas of low resistance through the insulation. Such testing has identified areas of
lower insulation resistance that were several orders of magnitude lower than that of the
surrounding insulation. This activity showed that several pockets of degradation were distributed
along and around the insulation. The insulation consistently failed in these low-resistance
pockets when breakdown tests were subsequently performed. Even though the insulation was not
uniformly degraded, the identification of several low-resistance pockets indicated that distributed
general degradation existed, rather than a unique degradation site. See Section A.3.1, Failure
of a 38-Year-Old Butyl Rubber Cable Due to Water-Induced Degradation, for an example of
this technique.
10-5
10296123
Response to Cable Failures
1
Temporary is a relative term. Depending on the environmental stressor, it could take many years, even decades, for
the cable to degrade again to the point of failure. However, if the causes of the original failure cannot be eliminated
from consideration for the replacement cable, the repair should be considered temporary, and another option must be
implemented later to permanently resolve the problem.
10-6
10296123
Response to Cable Failures
Fire barriers and containment penetrations can also prevent using the original routing path of
replacement cables if disturbing the barriers would create issues for adjacent circuits or affect
their ability to continue to perform their design function.
Access to the cables can be another issue. Safety concerns for working in a confined location
such as a vault or manhole with other energized conductors can become a factor in the repair
decision. Corporate safety procedures for working in confined spaces and in areas with the
potential for arc flash must be followed.
10-7
10296123
Response to Cable Failures
If replacement of the cable is on the critical path in an outage, diagnostic testing could be
deferred to a more convenient time. Aging of a medium-voltage cable takes a long period.
Accordingly, baseline testing of the cable a few years into operation would be acceptable.
When old and new segments are to be combined, both segments should be tested separately
before splicing to verify the adequacy of the segments and then again after splicing to verify the
adequacy of the splices and terminations of the completed cable. At minimum, the completed
cable should withstand testing to verify that no serious damage has occurred during installation
and that no splicing or termination errors have occurred.
10-8
10296123
11
REFERENCES
1. Medium Voltage Cable Aging Management Guide. EPRI, Palo Alto, CA: 2008. 1016689.
This product has been superseded by Product 1021070, Medium Voltage Cable Aging
Management Guide, Revision 1.
2. R. B. Blodgett and R. G. Fisher, “A New Corona- and Heat-Resistant Cable Insulation
Based on Ethylene Propylene Rubber,” Paper No. 63-162, AIEE Transactions on Power
Systems and Apparatus, pp. 980–990 (December 1963).
3. J. C. Carroll, A. R. Lee and R. B. McKinney, “Design and Evaluation of Butyl-Rubber-
Insulated Power Cables,” AIEE Transactions on Power Systems and Apparatus, Vol. 74,
pp. 1204–1214 (December 1955).
4. Generic Letter 2007-01, “Inaccessible or Underground Power Cable Failures that Disable
Accident Mitigation Systems or Cause Plant Transients,” U.S. Nuclear Regulatory
Commission, Washington, DC (February 7, 2007).
5. Aging Management Program Guidance for Medium-Voltage Cable Systems for Nuclear
Power Plants. EPRI, Palo Alto, CA: 2010. 1020805.
6. W. A. Thue, editor. Electrical Power Cable Engineering: Second Edition. New York:
Marcel Dekker, Inc. (2003).
7. AEIC CS1-90, “Specification for Impregnated Paper-Insulated Metallic Sheathed Cable,
Solid Type.” 11th ed. Association of Edison Illuminating Companies (1990).
8. Power Plant Electrical Reference Series, Volume 4: Wire and Cable. EPRI, Palo Alto,
CA: 1987. EL-5036-4.
9. Nexans. “Configuration of Single Conductor Cable to Obtain Equal Load Sharing.”
www.nexans.ca/egy/tecdoc/8.htm.
10. Nexans. “Installation of Single Conductor Cable to Prevent Eddy and Circulating
Currents.” www.nexans.ca/egy/tecdoc/7.htm.
11. NEI 06-05, “Medium Voltage Underground Cable.” Nuclear Energy Institute, Washington,
DC (April 2006).
12. Plant Support Engineering: Common Medium Voltage Cable Specification for Nuclear
Power Plants. EPRI, Palo Alto, CA: 2009. 1019159.
13. T. P. Arnold and C. D. Mercier, editors. Power Cable Manual. Southwire Company (1991).
14. J. D. Medek, adapted from class notes of “Understanding Power Cable Characteristics and
Applications.” University of Wisconsin–Madison (2001).
11-1
10296123
References
15. G. Virsberg and P. H. Ware, “A New Termination for Underground Distribution,” IEEE
Transactions on Power Apparatus and Systems, Vol. PAS-86, pp. 1129–1135 (1967).
16. T. A. Balaska, “Jointing of High Voltage, Extruded Dielectric Cables, Basics of Electrical
Design and Installation.” IEEE UT&D Conference Record, pp. 318–326 (1974).
17. Evaluation of Pre-Molded and Field Assembled Joints for Extruded Dielectric Medium
Voltage Cables. EPRI, Palo Alto, CA: 2000. 1000045.
18. Electrical Connections Application Guide. EPRI, Palo Alto, CA: 2002. 1003471.
19. L. Mullins, “Effects of Filler in Rubber.” In The Chemistry and Physics of Rubber-like
Substances: Studies of the Natural Rubber Producers Research Association, edited by
L. Bateman, p. 311. Scientek Books (1963).
20. A. Schallamach, “Abrasion and Tire Wear.” In The Chemistry and Physics of Rubber-Like
Substances: Studies of the Natural Rubber Producers Research Association, edited by
L. Bateman, p. 374. Scientek Books (1963).
21. J. D. Hogan and L. A. Berry, “From Dust to Quality Dielectrics—The Formulation of
Dielectric Clays,” IEEE Electrical Insulation Magazine, Vol. 21, No. 5. pp. 25-28
(September/October 2005).
22. T. L. Brown, H. E. LeMay, and B. E. Burstein, Chemistry: The Central Science. 6th ed.,
p. 876. Prentice Hall Publishing Co. (1994).
23. B. S. Bernstein, “Electrical Insulating Materials.” In Electrical Power Cable Engineering,
edited by W. A. Thue, pp. 65–114. Marcel Dekker, Inc. (2003).
24. R. Webb, T. Shafer, and A. Tsou, “Butyl Rubber.” In Encyclopedia of Polymer Science and
Technology, edited by H. F. Mark, pp. 356–381.Wiley Interscience (2004).
25. Exxon. “Butyl Rubber.” In Materials for the Electrical Industry, publication SYN-73-1730,
pp. 5–16 (1973).
26. Exxon, “EPR.” In Materials for the Electrical Industry, publication SYN-73-1730,
pp. 19–51 (1973).
27. ICEA S-97-682-2007, “Standard For Utility Shielded Power Cables Rated 5 Through 46
KV” ICEA, Carrolton, GA (2007).
28. K. Brown, “Performance of Medium Voltage Underground Medium Voltage Cable at
United States Nuclear Plants,” IEEE Insulated Conductors Committee Meeting, Scottsdale,
Arizona, pp. 886–909 (October/November 2005).
29. Hercules Chemical Company. Hercules Peroxides, Cross-Linking Agents and
Polymerization Catalysts. Bulletin ORC-1031, 2016 (1983).
30. R. L. Zapp and P. Hous, “Butyl and Chlorobutyl Rubber.” In Rubber Technology. 2nd ed.
edited by M. Morton, pp. 249–262. Van Nostrand Reinhold Publishing Co. (1973).
31. R. Frank, “Black EPR Insulated Medium Voltage Cables,” ICC Minutes, pp. 954-958
(2005).
11-2
10296123
References
32. C. Zuidema, “Evolution of EPR Compounds.” ICC Meeting, Education Forum on EPR,
pp. 920–953. (Fall 2005).
33. H. Hu, “Evaluation of Discharge Resistance of Solid Dielectric Power Cable Insulations,”
IEEE Electrical Insulation Magazine, Vol. 11, No. 2, pp. 18–22 (July/August 1995).
34. A. K. Hiranandani, “Condition Assessment of Butyl Insulated Power Cable Used in
Generating Stations,” Conference Record of 2004 IEEE International Symposium on
Electrical Insulation, pp. 450-453 (2004).
35. W. C. Smith, “Vulcanization of Butyl, Chlorobutyl and Bromobutyl Rubber.” In
Vulcanization of Elastomers, edited by G. J. Allimger and I. J. Srjothun, pp. 230–264.
Reinhold Publishing Co. (1964).
36. F. S. Conant, “The Effect of the State of Cure on Vulcanizated Properties.” . 78-125.
Vulcanization of Elastomers. In Vulcanization of Elastomers, edited by G. J. Allimger
and I. J. Srjothun, pp. 78–125. Reinhold Publishing Co. (1964).
37. L. A. Dissado and J. C. Fothergill. “Water Treeing Degradation.” In Electrical
Degradation and Breakdown in Polymers, IEE Materials and Devices Series 9,
G. C. Stevens, ed. United Kingdom: Peter Peregrinius Ltd. (1992).
38. M. Fenger, ”Failure Analysis—Feedwater Pump Motor Cable: Accurate Assessment for
Informed Decision Making,” Kinectrics, Inc., Transmission & Distribution Technologies,
publication T&D-CS-12.02.07.
http://www.kinectrics.com/en/solutions/CableFailureFeedwaterPump.html.
39. A. Charlesby. Atomic Radiation and Polymers. International Series of Monographs on
Radiation of Materials. Pergamon Press (1960).
40. A. Chapiro. Radiation Chemistry of Polymeric Systems, Sections IX and X. Interscience
Publishing, J. Wiley and Sons (1962).
41. J. Schiers, “Opportunities in Gamma Irradiation of Polymers.”
www.steritech.com.au/files/pdf/xlinkapr06.pdf.
42. G. Odian and B. S. Bernstein, “Radiation Cross-Linking of Polyisobutylene,” Journal of
Polymer Science. Polymer Letters, Vol. 2. p. 819 (1964).
43. Effects of Moisture on the Life of Power Plant Cables, Part 1: Medium Voltage Cable.
EPRI, Palo Alto, CA: 1994. TR-103834-P1.
44. Effects of Moisture on the Life of Power Plant Cables, Part 2: Low Voltage Cable. EPRI,
Palo Alto, CA: 1994. TR-103834-P2.
45. Medium Voltage Cables in Nuclear Plant Applications- State of Industry and Condition
Monitoring. EPRI, Palo Alto, CA: 2003. 1003664.
46. ICEA T-27-581/NEMA WC 53, “Standard Test Methods for Extruded Dielectric Power,
Control, Instrumentation, and Portable Cables for Test.” NEMA, Roslyn, VA (2005).
47. E. Walcott, W. S. Temple, and J. T. Smith III. 2005. “Accelerated Wet Testing of MV
EPR Insulated Cables,” presented at the Fall 2005 ICC Meeting, Education Forum on
EPR (2005).
11-3
10296123
References
48. Advanced Diagnostics: Life Estimation of Extruded Dielectric Cables. EPRI, Palo Alto,
CA: 1993. 1001726.
49. Advanced Diagnostics: Life Estimation of Extruded Dielectric Cables. EPRI, Palo Alto,
CA: 2005. 1011499.
50. C. Katz, “Performance Evaluation of EPR Underground Distribution Cable,” presented at
the IEEE Insulated Conductors Committee Meeting (Fall 2005).
51. IEEE 400-2001, “IEEE Guide for Field Testing and Evaluation of the Insulation of
Shielded Power Cable Systems.” IEEE, New York (2001).
52. In-Service Evaluation of Extruded Dielectric Distribution Cables. EPRI, Palo Alto, CA:
2002. 1009017.
53. Cable Polymer Aging and Condition Monitoring Research at Sandia National Laboratories
under the Nuclear Energy Plant Optimization (NEPO) Program. EPRI, Palo Alto, CA:
2005. 1011873.
54. Medium Voltage Cables in Nuclear Plant Applications—State of Industry and Condition
Monitoring. pp. 28–30. EPRI, Palo Alto, CA: 2003. 1003664.
55. IEEE 400.1-2007, “IEEE Guide for Field Testing of Laminated Dielectric, Shielded Power
Cable Systems Rated 5 kV and Above with High Direct Current Voltage.” IEEE, New
York (2007).
56. IEEE 400.2-2004, “IEEE Guide for Field Testing of Shielded Power Cable Systems Using
Very Low Frequency (VLF).” IEEE, New York (2004).
57. ICEA S-93-639/NEMA WC 74, “5–46 kV Shielded Power Cable for Use in the
Transmission and Distribution of Electric Energy.” National Electrical Manufacturers
Association, Rosslyn, VA (2006).
58. IEEE 383-1974, “IEEE Standard for Type Test of Class 1E Electric Cables, Field Splices,
and Connections for Nuclear Power Generating Stations.” IEEE, New York (1974).
59. IEEE 308-1978, “IEEE Standard Criteria for Class 1E Power Systems for Nuclear Power
Generating Stations.” IEEE, New York (1978).
60. IEEE 603-1998, “IEEE Standard Criteria for Safety Systems for Nuclear Power Generating
Stations.” IEEE, New York (1998).
61. IEEE 1185-1994, “IEEE Guide for Installation Methods for Generating Stations.” IEEE,
New York (1994).
62. ICEA S-97-682-2000, “Utility Shielded Power Cables Rated 5,000–46,000 Volts.” ICEA,
Carrolton, GA (2000).
63. ASTM D-2275-89. “Test Method for Voltage Endurance of Solid Electrical Insulating
Materials Subjected to Partial Discharges (Corona) on the Surface.” American Society for
Testing Materials, Conshohocken, PA.
64. ICEA T-31-610, “Test Method for Conducting Longitudinal Water Penetration Resistance
Tests On Blocked Conductors.” ICEA, Carrolton, GA (2006).
11-4
10296123
References
65. IEEE 383-2003, “IEEE Standard for Type Test of Class 1E Electric Cables and Field
Splices for Nuclear Power Generating Stations.” IEEE, New York (2003).
66. IEEE 1202, “Standard for Flame-Propagation Testing of Wire and Cable.” IEEE, New
York (2006).
67. ICEA T-24-380. “Guide for Partial Discharge Test Procedure.” ICEA, Carrolton, GA
(2007).
68. AEIC 5, “Specifications for Cross-Linked Polyethylene Insulated, Shielded Power Cables
Rated 5 through 46 kV.” 10th ed. Association of Edison Illuminating Companies (1994).
69. AEIC 6, “Specifications for Ethylene Propylene Rubber Insulated Shielded Power Cables
Rated 5 through 46 kV.” 1st ed. Association of Edison Illuminating Companies (1973).
70. ICEA S-19-81, “Rubber Insulated Wire and Cable for the Transmission and Distribution
of Electrical Energy.” ICEA, Carrolton, GA (1994). Withdrawn.
71. ICEA S-68-516/NEMA WC 8-1976, “Ethylene-Propylene-Rubber Insulated Wire and
Cable for the Transmission and Distribution of Electrical Energy.” Rev. 6. ICEA,
Carrolton, GA (pages dated January 30, 1983). Withdrawn.
72. ICEA S-66-524/NEMA WC 7, “Cross-Linked Thermosetting Polyethylene-Insulated Wire
and Cable for the Transmission and Distribution of Electrical Energy.” Second Edition.
ICEA, Carrolton, GA (1982). Withdrawn.
73. AEIC 5, “Specifications for Polyethylene and Cross-Linked Polyethylene Insulated
Shielded Power Cables Rated 2,001–35,000 Volts.” 2nd ed. Association of Edison
Illuminating Companies (1969).
74. AEIC CS5, “Specifications for Thermoplastic and Cross-Linked Polyethylene Insulated
Shielded Power Cables Rated 5 through 69 kV.” 6th ed. Association of Edison Illuminating
Companies (1979).
75. ICEA T-33-655, “Guide for Low Smoke, Halogen-Free (LSHF) Polymeric Cable Jackets.”
ICEA, Carrolton, GA (1994).
76. N. Ahmed and N. Srinivas, “In situ Partial Discharge Detection in Power Plant Cables.”
Proceedings for the Twenty-Sixth Water Reactor Safety Information Meeting. Vol. 3.
U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research,
pp. 289-301 (1998).
77. G. C. Montinari, “On Line Partial Discharge Diagnosis in Power Cables,” 2009 IEEE
Electrical Insulation Conference, Montreal, Canada, pp. 210–215 (May 31–June 3, 2009).
78. Cable Polymer Aging Database, Version 2. EPRI, Palo Alto, CA: 2005. 1011874.
79. Initial Acceptance Criteria Concepts and Data for Assessing Longevity of Low-Voltage
Cable Insulations and Jackets. EPRI, Palo Alto, CA, and U.S. Department of Energy,
Washington, DC: 2005. 1008211.
80. AEIC CS8-07, “Specification for Extruded Dielectric Shielded Power Cables Rated 5
through 46 kV.” 3rd ed. Association of Edison Illuminating Companies (2007).
11-5
10296123
References
81. Edison Electric Institute (EEI), “Cable Testing.” In Underground Systems Reference Book.
pp. 11-2 to 11-6. New York: Edison Electric Institute (1957).
82. N. Srinivas and B. S. Bernstein, “Effect of DC Testing on Aged XLPE Insulated Cables
with Splices,” presented at Jicable 91, Paris, France (June 1991).
83. Bruce S. Bernstein, “Cable Testing: Can We Do Better?” IEEE Electrical Insulation
Magazine, pp. 33-38 (July/August 1994).
84. G. Bahder, C. Katz, G. S. Eager, E. Leber, S. M. Chalmers, W. H. Jones, and W. H.
Mangrum, “Life Expectancy of Cross-Linked Polyethylene Insulated Cables Rated 15 to
35 kV.” IEEE Transactions on Power Apparatus and Systems, Vol. PAS-100, pp. 1581–
1590 (April 1981).
85. R. Bach, W. Kalkner, and H. Oldehoff, “Verlustfactormessungen bei 0.1 Hz an
betriebsgealterten PE/VPE-Kabelanlagen,” Elektrizitätswirtschaft, Jg. 92, Heft 17/18,
pp. 1076–1080 (1993).
86. S. Hvidsten, E. Ildstad, B. Holmgren, and P. Werelius, “Correlation Between AC
Breakdown Strength and Low Frequency Dielectric Loss of Water Tree Aged XLPE
Cables.” IEEE Transactions on Power Delivery, Vol. 13, No. 1, p. 40–45 (January 1998).
87. Testing and Diagnostics of Medium Voltage Cables. BAUR Prüf- und Messtechnik GmbH,
Sulz, Austria: 1996. PHG-TD Report 820-002.
88. Advanced Diagnostics: Estimation of Life of Extruded Cables. EPRI, Palo Alto, CA: 2005.
1001727.
89. P. Werelius, P. Tharning, R. Eriksson, B. Holmgren, and U. Gafvert, “Dielectric
Spectroscopy for Diagnosis of Water Tree Deterioration in XLPE Cables.” IEEE
Transactions on Dielecrics and Electrical Insulation, Vol.8, No.1. pp. 27–42 (February
2001).
90. Guide for Non-Destructive Diagnosis of Extruded Cable Systems. EPRI, Palo Alto, CA:
2005. 1001731.
91. IEEE 400.3, “IEEE Guide for Partial Discharge Testing of Shielded Power Cable Systems
in a Field Environment.” IEEE, New York (1994).
92. High Voltage, Low Frequency (0.1 Hz) Testing of Power Cable. EPRI, Palo Alto, CA:
1998. TR-110813.
93. N. Ahmed and N. Srinivas, “On-Line Partial Discharge Detection in Cables.” IEEE
Transactions on Dielectric and Electric Insulation. Vol. 5, No. 2 (1998).
94. N. Ahmed and N. Srinivas, “On-Line Partial Discharge Diagnostic System in Power Cable
Systems,” presented at the IEEE T&D Conference, Atlanta, GA (2001).
95. N. Srinivas, B. S. Bernstein, O. Morel, and D. Bogden, “Condition Assessment of
Distribution and Transmission Class Voltage Cable Systems,” presented at the IEEE T&D
Conference New York (2003).
96. D. Gross, “Signal Transmission and Calibration of On-Line Partial Discharge
Measurements,” presented at the IPCDAM Conference (June 2003).
11-6
10296123
References
11-7
10296123
10296123
A
MEDIUM-VOLTAGE CABLE FAILURES AND FIELD
EXPERIENCE
A.1 Introduction
This appendix describes failure and field experience for medium-voltage cables. The purpose is
to describe events and the lessons learned from them. These failure and field experience
descriptions were provided by utilities; however, the plant and owner’s name are not included.
This appendix is divided into five sections: this introduction; failures under dry conditions;
failures related to wet conditions; failures of terminations and splices; and a summary of lessons
learned. A number of failures under dry conditions are presented here. Dry failures occur less
frequently than wet failures (approximately 25% as often). The dry failures occur randomly from
a number of different causes and, unlike water-related failures, have no common stress
enhancement. Essentially, each dry failure event provides an interesting insight into potential
failure mechanisms and issues to avoid.
A-1
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-1
External Cable Condition at the Location of the Fault
Figure A-2
Condition of the Shield After the Jacket and Burn Hole Were Removed from the Fault
A-2
10296123
Medium-Voltage Cable Failures and Field Experience
The copper tape shield was found to have three longitudinal depressions around approximately
one-third of the circumference of the cable. A small burn hole in the tape shield occurred at the
location of the insulation failure. Figure A-3 shows the damage that occurred to the conductor
of the cable.
Figure A-3
Damage to the Conductor from the Fault
The exact cause of the failure was not identified. The cable was repaired by removing the
damaged insulation, tapering the ends of the insulation, and forming a taped repair equivalent
of a splice directly over the conductor.
A-3
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-4
Inner Surface of the Insulation with Severe Indentation from the Semiconducting Tape
To characterize the cause of the softening of the insulation, a number of chemical and physical
tests were performed on the insulation and jacket. For the insulation, the tests were the
following:
Fourier transform infrared analysis
Elongation at break
Tensile strength
Gel content
Shore hardness
For the jacket, the tests were the following:
Solubility
Elongation at break
Tensile strength
Plasticizer content
The Fourier transform infrared analysis confirmed that the insulation material was butyl rubber.
The solubility test confirmed that the jacket was PVC. The tensile strength tests of the more
thermally damaged insulation specimens ranged from 2.47 to 63.8 psi (0.17 to 0.44 MPa), and
the elongation values ranged from 53% to 159%. The less damaged specimen had a tensile
strength of approximately 145 psi (1 MPa) and an elongation at break of greater than 250%.
A-4
10296123
Medium-Voltage Cable Failures and Field Experience
However, all these values were significantly lower than the Insulated Cable Engineers
Association (ICEA) standard minimum tensile strength for new butyl rubber of 597.6 psi
(4.12 MPa) and minimum elongation at break of 350%.
The gel content of the less damaged specimen was 68%, whereas that of the two more
deteriorated specimens ranged from 0% to 13.5%. The expected gel content for unaged butyl
rubber insulation should be between 80% to 95%.
The Shore “A” hardness for the less damaged cable was 54.4, whereas the other two specimens
were too soft to measure. The evaluation estimated an operating temperature of approximately
176°F to 185°F (80°C to 85°C) over its 35-year life.
Evaluation of the jacket indicated that the high temperature was not from exterior conditions,
because the PVC had not released oily plasticizers or hydrochloric acid, which would have
corroded the shield. At 200%, the jacket had twice the minimum required elongation at break
and plasticizer levels (24% to 25%) that were within the normal bound (20% to 25%).
The conclusion was that the softening of the butyl rubber was due to elevated temperature from
ohmic heating of the conductor. The air temperature in the vicinity of the failure was not
elevated, and there were no adverse localized thermal environments in the area. Thermal
exposure of butyl rubber induces an oxidative chain scission reaction that leads to softening of
the rubber.
The cause of the ohmic heating was an unbalanced magnetic circuit that induced higher currents
on individual conductors in the circuit with several conductors per phase. This condition
occurred when conductors were not transposed along the length of the circuit at the time of
installation. This led to some individual conductors having low currents with respect to their
ampacity, whereas others had high currents. The softening of the material allowed the conductor
shield to rise up into the insulation, leading to high localized stress and eventual breakdown of
the insulation.
The insulation softening that occurred in this case is unique to sulfur-cured butyl rubbers; it
would not occur for EPR or cross-linked polyethylene (XLPE) insulation.
A-5
10296123
Medium-Voltage Cable Failures and Field Experience
The fault occurred at the base of a Kellems grip used to support the cable at a vertical drop in
elevation. Dissection of the cable at the faulted area indicated that the zinc shield tape had
crystallized and cracked, causing a separation in the shield and resulting in partial discharge (PD)
and arcing along the interface between the gap in the shield. Figure A-5 shows a section of good
shield, as well as the deteriorated shield. PD and tracking occurred at the gap in the shield around
the circumference of the cable, damaging the jacket and the insulation.
Figure A-5
Zinc Shield (Top, Good Condition; Bottom, Crystallized and Burned from Tracking)
Figure A-6 shows the tracking path and the fault site along the interior of the chlorosulfonated
polyethylene (CSPE) jacket. The figure shows that the tracking path followed the path of the gap
in the shield, as evidenced by the shield marks on the inner surface of the jacket.
Figure A-6
Tracking Path and Failure Location on Interior of Jacket
A-6
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-7 shows the outer surface of the insulation, illustrating the tracking path and the
ultimate failure location on the insulation.
Figure A-7
Tracking Path and Failure Site on Insulation
Figure A-8 shows the damage to the conductor of the cable from the fault. It is not clear why
zinc was used as the shield in this cable. It is known that a copper production strike occurred in
the early 1970s, which might have affected copper cost and availability at the time that this cable
was produced. Most plants reporting tape metal shields indicate that copper or tinned copper
was used.
Figure A-8
Damage to Conductor from Fault
A-7
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-9
Transformer Connection of One of the Cables
A-8
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-10
Circuit Breaker Connection of One of the Cables
Figure A-11 shows the configuration of the cables in the manhole adjacent to the transformer. An
A, B, and C phase passes through each duct to balance the magnetic currents in the underground
section of the circuit. The configuration in the manhole converts from a 3 vertical × 4 horizontal
matrix to a 4 vertical × 3 horizontal matrix, so that removal of just a few deteriorated cables
would require replacing many cables.
Figure A-11
Cable Configuration in the Manhole Adjacent to the Transformer
A-9
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-11 indicates that water could have affected some circuits in the lower ducts, leading to
the expectation that cables in the upper ducts would not be affected because of the indication that
water had not reached them. As indicated by Figure A-12, however, cables in the upper ducts
failed under test or exhibited test results showing significant deterioration. (Boxes in the figure
indicate the locations of deteriorated cables.)
Figure A-12
Location of Failed, Faulted, and At-Risk Cables Within Duct Leading to Building
The locations of the deteriorated cables indicated that the water mark in the manhole was not
indicative of the locations of deteriorated cable. Moreover, the removal of cables from the
manhole would have required removing and repulling nearly every cable in the manhole, a
complex and nearly impossible job, given the outdoor temperatures at the time of the event. The
plant owner decided to run temporary cables instead of trying to pull new cables into the ducts
under adverse temperature conditions and to allow the unit to return to service more quickly.
A review of the complexity of the cable layout inside the plant indicated that replacement of the
entire lengths of the circuits would require approximately 3 months time and approximately
18,000 ft (5500 m) of cable. The plant had 7500 ft (2300 m) of spare cable on hand. A decision
was made to cut the circuits just inside the building wall and to splice temporary cables in place
between the transformer and the building wall. Tan δ measurements were used to determine the
A-10
10296123
Medium-Voltage Cable Failures and Field Experience
cables that had failed and to confirm that the in-building segments of these cables were in
acceptable condition. PD testing was attempted, but the corrosion of the shield and the high
attenuation of the butyl rubber insulation precluded detecting a reflection of even a high-level
calibration signal. This indicated that, even had PD been present, it would not have been
detectible from the breaker cubicle test location.
The cause of the failures was traced to water-related degradation. The ducts drained toward the
building, causing water to leak into the building in winter. The entry point into the building is in
an unheated area, allowing ice to form. To eliminate this condition, the ducts had been sealed in
the 1990s (see Figure A-13). Figure A-13 also shows water draining from a conductor of cable
B-5, the cable that failed during the event.
Figure A-13
Duct Sealing and Water Leaking from Failed Conductor B-5
Figure A-14 shows the as-found condition of cables A-5, B-5, and C-5 when they were removed
from the duct. During the original failure of the B-5 cable, the copper shield vaporized, causing
the voltage to seek an alternate ground (2 A.M. repetitive flashovers). Subsequently, the jacket
failed on the A-5 cable, allowing a flashover path to the shield of the A-5 cable. The 2 P.M.
repetitive flashovers are likely to have been to the A phase shield. Fortunately, the A phase
insulation withstood the condition, and a phase-to-phase fault did not occur. The original failure
of the B-5 cable insulation caused burnout of the copper shield, and the failure propagated to the
shield of the A-5 cable. The C-5 cable was not damaged.
A-11
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-14
Cables A-5, B-5, and C-5
Figure A-15 shows a close-up of the damage to the B-5 cable. The copper tape shield has
vaporized from the entire vicinity of the fault and is not visible at the edges of the hole in the
jacket. The fault evaporated 3/8 in. (1 cm) of the conductor.
Figure A-15
Close-Up of B-5 Cable Failure
A-12
10296123
Medium-Voltage Cable Failures and Field Experience
The faulted cable and adjacent phases were sent to a forensics laboratory, where all of the strands
of the conductor were found to be corroded in the vicinity of the fault, indicating that the
conductor had filled with water (see Figure A-16). B5B is the conductor of the cable that failed.
A5 is the conductor that was adjacent; it is slightly corroded. C5 is the adjacent conductor that
had good insulation, and B2 is from another cable with good insulation.
Figure A-16
Comparison of Conductor Corrosion
The insulation was found to have a high moisture content, as well. Examination of the insulation
showed channels of low insulation resistance between the surface of the insulation and the
conductor. The insulation resistances of these areas were 1/1000th of the resistances of the
surrounding insulation. These areas were marked, the shield was reformed, and withstand tests
were performed. The breakdowns of the insulation occurred at these channels of low insulation
resistance (see Figure A-17). The black tape and aluminum foil constitute the shield that was
applied to allow breakdown testing after the low-resistance channel was identified.
Figure A-17
Marked Low-Resistance Channel in the Insulation with Breakdown Hole in Center
A-13
10296123
Medium-Voltage Cable Failures and Field Experience
Wafers were cut from some of the low-resistance channels and the wafers were heated in water
to allow identification of degradation. Swelling and fissures were identified in the low-resistance
zones (see Figure A-18).
Figure A-18
Micrograph of Insulation Wall at Low-Resistance Channel, Showing Swelling and Fissures
The conclusion was that the failure was caused by water-induced degradation of the energized
cable’s insulation. The corrosion of the entire conductor indicated a long-term presence. The
water could have accumulated during construction or through permeation through the cable wall
and condensation at the lowest and coldest point on the cable. The failure occurred within a short
distance of the point at which the temperature of the cable would have been below freezing for
most of the winter. The load on these cables during normal operation is low, so ohmic heating
would not have tended to warm the cable and reduce condensation.
Temporary cables were run on the surface of the ground to replace the underground section of
the cables. Figure A-19 shows the temporary routing, and Figure A-20 shows the aboveground,
permanent replacement that was implemented during the next outage of the second unit at
the plant.
A-14
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-19
Temporary, Aboveground Cable System
Figure A-20
Permanent Aerial Structure Routing on Second Unit
A-15
10296123
Medium-Voltage Cable Failures and Field Experience
A-16
10296123
Medium-Voltage Cable Failures and Field Experience
to have sections of loose or failed jacket, and the third phase (likely A) was found to have its
jacket intact. As the cable was being removed from the duct, it was cut into random lengths from
3 to 5 feet in length. The electricians noted that water was dripping from between the strands of
the conductors when the sections were cut. The utility’s evaluation concluded that the failures
were due to water-related insulation degradation.
Up to the point of the failure, the utility had considered insulation resistance testing acceptable
for evaluating the condition of cables. However, as a result of the investigation, the utility
determined that insulation resistance testing is not a discriminating indicator of condition of
medium-voltage cables and that their acceptance criteria of a minimum of 5 MΩ was wholly
inadequate. The cables that failed had insulation resistances of 30 MΩ before the event, whereas
the other cables had 2 GΩ or higher. At the time of the first event, the alternate feed was found to
have low insulation resistance values. The cables were cut in the manhole, the section with the
low insulation resistance was replaced, and the cables with 10–18 MΩ insulation resistances
were reused. However, 6 hours after reenergization, the reused section B and C phase failed,
indicating that the 5 MΩ criterion was unsatisfactory.
Sections of the cable were provided to the Electric Power Research Institute (EPRI) for further
testing. The 60-Hz tan δ versus breakdown results for this cable are provided in Appendix C.1,
60-Hz Tan δ Data from 1972 Black Okoguard Insulation. The evaluation identified that the
neoprene jacket was loose in a number of cases and determined that low insulation resistance
channels had developed in the EPR, similar to those described for the butyl rubber cable (see
Section A.3.1, Failure of a 38-Year-Old Butyl Rubber Cable Due to Water-Induced
Degradation). The full results of the evaluation are provided in the EPRI report Failure
Mechanism Assessment of Medium Voltage Ethylene Propylene Rubber Cables (1015070) [101].
A-17
10296123
Medium-Voltage Cable Failures and Field Experience
Dimensional checks indicated that all the cable subcomponents were within specification, with
the exception of the PVC jacket, which was slightly thinner than expected. The jacket thickness
was determined to have no effect on the failure.
The section that had failed was evaluated in detail. The fault did not puncture the jacket, and the
jacket had to be removed to allow location of the actual fault. Figure A-21 shows the puncture
and damage caused by the vaporization of the shield during the fault.
Figure A-21
Damage to the Copper Shield at the Damage Site
Figure A-22 shows the cross section of the insulation at the location of the fault. The fault tube
shows the carbonization of the tube from the fault and the fault burn-in process that was used to
locate the fault. The inspection of the wafers of insulation at the fault shows no evidence of a
large water tree at the failure site. However, the fault did burn away an extensive area of the
insulation. No sign of an electrical tree was identified.
A-18
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-22
Cross Section of Insulation at the Fault Tube, Showing Carbonization of the Wall of the
Fault Tube
Sections of the insulation surrounding the fault were subjected to a hot oil bath, which turns the
insulation clear for visual examination. Figure A-23 shows evidence of a large, single water tree.
Figure A-24 shows a cross section of the 61-mil (1.5-mm) water tree after dying.
A-19
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-23
Large Water Tree Viewed Through Hot Oil Bath
Figure A-24
Cross Section of the Water Tree Shown in Figure A-23
A-20
10296123
Medium-Voltage Cable Failures and Field Experience
The water tree extends 70% through the wall of the insulation at the site of a small conductor
shield protrusion. The protrusion and the contaminant causing it are shown in Figure A-25.
Figure A-25
Embedded Particle at Base of Water Tree Shown in Figures A-23 and A-24
Similar assessments were performed on the other cable segments that had been taken to
breakdown during the laboratory testing. The examination at the failure sites showed low to
medium density of water trees, with a maximum tree length of 70% to 80% of wall. The overall
conclusion was that the most likely cause of failure was a large water tree growing from a
conductor shield defect that was present at the time of installation.
The results indicate that the insulation system was not highly treed but that a few large trees had
occurred at sites of defects from manufacture in the early 1970s, the period before manufacturers
made significant efforts to improve the cleanliness of the preparation and extrusion process.
A-21
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-26
As-Found Condition of Terminations
A-22
10296123
Medium-Voltage Cable Failures and Field Experience
Figure A-27
Burnthrough of Stress Relief Adaptor
Figure A-28
Burned-Through Insulation Found After Removal of the Stress Relief Adaptor
The failed termination and the non-failed C-phase termination were sent to a cable forensics
laboratory for evaluation. The laboratory found that the Elastimold adaptor was meant for a cable
with a diameter of 1.220 to 1.375 in. (31 to 35 mm) and the cable insulation diameter was
1.16 in. (29.5 mm). It appeared that a molded termination meant for a standard configuration was
used on an Anaconda UniShield cable having a compact design with a much smaller diameter.
The central part of the adapter should mate tightly with the cable insulation surface. Radial
pressure must be exerted on the cable insulation surface by the adapter to quench any PD at the
interface of the cable insulation and the stress relief adapter. Instead of the adaptor exerting
radial pressure on the cable insulation, it was loose. Accordingly, after the silicone grease dried
out, PD eroded both the adapter and the insulation, leading to the failure.
A-23
10296123
Medium-Voltage Cable Failures and Field Experience
Although the C-phase termination showed no outward signs of degradation, interior inspection
showed that the silicone grease beneath the stress relief assembly was highly discolored and
experiencing PD, as well.
The lesson learned from this failure is that the Elastimold and other types of molded terminations
(as well as cold-shrink and heat-shrink systems) must be sized properly to mate with the cable
insulation and shielding system. Similar terminations throughout the plant were assessed to
confirm that the correct size stress relief adapter had been used. If this type of problem is
suspected, handheld PD detection devices can be passed near the terminations to determine
whether PD is present.
Repairs were made by replacing 10-ft (3.0-m) sections of the field cable and applying splices and
new terminations.
A-24
10296123
Medium-Voltage Cable Failures and Field Experience
A-25
10296123
Medium-Voltage Cable Failures and Field Experience
A-26
10296123
B
RESULTS OF NUCLEAR ENERGY INSTITUTE SURVEY
AND NUCLEAR REGULATORY COMMISSION
GENERIC LETTER 2007-01
This section provides the results from a Nuclear Energy Institute (NEI) survey on underground
medium-voltage cables that was performed in 2005 and described in NEI 06-05, “Underground
Medium Voltage Cable” [11]. It also summarizes the utility responses to the Nuclear Regulatory
Commission (NRC) Generic Letter 2007-01, “Inaccessible or Underground Power Cable
Failures That Disable Accident Mitigation Systems or Cause Plant Transients” [4]. Since the
time of the survey, a few more medium-voltage cables have failed. These are not included in
the data. The results of the Generic Letter response and the NEI survey correlate well, with only
one exception, which is described in Section B.3, Results of Utility Responses to Generic
Letter 2007-01.
B-1
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Cable functions
Cable conductor shield and insulation shield attributes
If failures had occurred, information about the failure root cause and cable replacement
Most survey data were collected in February and March 2005. However, some plants conducting
refueling outages submitted data later.
B.2.1 Contributors
A total of 81 units (51 plants) provided information to the general survey questions. Of those
units, 74 units (47 plants) provided information on originally installed cables, failures, and
replacement cables.
B-2
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Table B-1 lists the 5-kV rated insulation types by number of units reporting the material.
Table B-1
Originally Installed 5-kV Insulation Types
Figure B-1 shows the distribution of manufacturers of black EPR insulated cables; Okonite and
Anaconda were the dominant manufacturers. The brown EPR is manufactured only by Kerite,
and the pink EPR was manufactured by either Anaconda (9 units reporting use) or Okonite
(21 units reporting use). One unit reported having both Anaconda and Okonite pink EPR.
Figure B-1
Manufacturers of 5-kV Cable
A total of 24 units reported having 8-kV rated cable in use on systems operating at 6.9 kV. Plants
with 6.9-kV systems tended to be constructed in the late 1970s and beyond, so that a lower
percentage of the plants had black EPR insulations for 8-kV rated cables than for the 5-kV rated
population. In addition, 6.9-kV systems were not adopted by most utilities, even in later plants.
Some later plants remained with 4.16-kV systems for the bulk of the plant and used 13.8-kV
systems to supply large loads such as the circulating water pumps. Table B-2 provides a
distribution of the insulation types in use in 8-kV cables.
B-3
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Table B-2
Distribution of 8-kV Cable Insulation Materials
A total of 40 units reported having 15-kV rated cables operating at 13.2-kV to13.8-kV. These
cables are used for distribution to large loads and feeds to 13–to-4 kV transformers. Table B-3
provides the distribution of the insulation types in use.
Table B-3
Distribution of 15-kV Cable Insulation Materials
A total of 20 units reported use of cables rated at 25 kV and 35 kV. The 35-kV cables were used
in 34.5-kV operating systems associated with off-site power feeds. The 25-kV cables were used
in 22-kV circuits between the generator output and auxiliary transformers.
Rather than having extruded polymer insulation, some cables are insulated by oil-impregnated
paper with an overall sheath of lead. Only a few plants reported using paper-insulated lead-
covered cable, and even at those plants, extruded polymer insulation was used on the bulk of the
installed cables. Table B-4 provides the distribution of the insulation materials on these cables.
Table B-4
Distribution of 25-kV to 35-kV Cable Insulation Materials
B-4
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
B.2.4 Shielding
At greater than 5 kV, cables are typically manufactured with insulation shields; however,
5-kV EPR cables can be manufactured and purchased with or without shields.
At 5-kV and greater, XLPE cables have insulation shields. Excluding the general-service cables,
22 units (30%) reported having a total of more than 271 circuits with nonshielded EPR cables.
Two plants reported having nonshielded cables, but they did not indicate the quantity. These
cables were used in safety, fire protection, operationally important, station black-out, and off-site
feed cables.
The lack of a shield on the EPR cables is not a reliability issue. Circuits without shields represent
an electrical testing issue. Electrical testing at high voltage requires a uniform ground plane. An
insulation shield provides such a ground plane. Circuits without a shield would not have a
uniform ground plane, and available electrical testing is unlikely to provide useful results.
For the cables rated above 5 kV, nearly all the cables were reported to have an insulation shield.
A few entries indicated that the respondent did not know; 42 units reported having 5-kV cables
with shields. This group with shields included all the XLPE cables and a portion of the EPR
cables.
Most shields involve a helically wrapped flat copper tape over a semiconducting extruded layer.
Earlier-style cables used a cotton or polymer semiconducting tape instead of the extruded layer.
Anaconda UniShield cables have 6 neutral wires located in the semiconducting extruded shield
and jacket.
B-5
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
5
Number of Units
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
Age of Units with No Failures
Figure B-2
Age Distribution of Units with No Failures
B-6
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Table B-5
Number of Failures per Plant Reporting Failures
A parallel evaluation of related data from the Institute of Nuclear Power Operations’ Equipment
Performance Information Exchange (EPIX) and Nuclear Plant Reliability Data System (NPRDS)
indicates that the failures reported by the 74 units responding to the NEI survey appear to be all
the medium-voltage, wet cable failures that occurred before 2006.
Figure B-3 shows the age distribution at time of failure for all cable types. The age distribution
at time of failure is quite broad, with at little as 5 years and as many as 30 years of service
before failure.
5
Number of Failures
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Years to Failure
Figure B-3
Age Distribution of Wet Cable Failures for All Insulation Types
B-7
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Figure B-4 shows the distribution of XLPE cable failures in relation to all failures. The total
distribution of failures are shown as gray bars, and the XLPE failures are shown in color. Twelve
XLPE failures occurred at 4 plants. Eight of the failures occurred at one plant in a specific type
of filled XLPE that was used only at that plant. These 8 failures represent 16% of the total wet
underground cable failures.
5
Number of Failures
4
Filled XLPE
Others
All Failures
3
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Age at Failure
Figure B-4
Age Distribution of Wet Cross-Linked Polyethylene Cable Failures
Figure B-5 shows the distribution of EPR failures in relation to all failures. Black EPR is the
dominant, original insulation in the industry (88% of units reported usage). Pink EPR was
installed in units that became commercial in the 1980s and beyond; it has become the dominant
replacement insulation. The failure of the pink EPR at five years was determined to be related to
a manufacturing flaw, causing a large contaminant at the failure site. The three pink EPR failures
at 10 years occurred at one plant; however, the cause of failure was not given.
B-8
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
5
Number of Failures
4
Black EPR
Red EPR
3 All Failures
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Age at Failure (years)
Figure B-5
Age Distribution of Wet Ethylene-Propylene Rubber Cable Failures
B-9
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Figure B-6 shows the distribution of butyl rubber failures in relation to all failures. Although the
number of butyl rubber failures appears small, only 4 of the reporting units indicated use of butyl
rubber cables.
5
Number of Failures
4
Butyl Failures
All Failures
3
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Years to Failure
Figure B-6
Age Distribution of Butyl Rubber Cable Failures
B-10
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Figure B-7 shows the age of the cables at the time of their failure. The general trend in the age at
failure is increasing. The trend in the number of failures at a given age is relatively stable, but it
shows a slight increase with time.
35
30
Age at Failure/Number of Failures
25
20 Age at Failure
# of Failures
Linear (Age at Failure)
15 Linear (# of Failures)
10
0
1975 1980 1985 1990 1995 2000 2005 2010
Year of Failure
Figure B-7
Age of Wet Cable at Failure Versus Year of Failure
B-11
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Figure B-8 shows the age distribution at failure versus the year of failure for EPR cables. The
trends for both black and pink EPR are parallel with time to failure increasing. The population of
black EPR cables is larger than that of pink EPR cables. The first installations of pink EPR
cables occurred approximately 10 years after installation of black EPR cables began.
35
30
Age at Failure/Number of Failures
25
Black EPR
Pink EPR
20 # Black EPR Failures
# Pink EPR Failures
Linear (# Pink EPR Failures)
15
Linear (# Black EPR Failures)
Linear (Black EPR)
0
1985 1990 1995 2000 2005 2010
Year of Failure
Figure B-8
Age at Time of Failure for Wet Ethylene-Propylene Rubber Cables
B-12
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Figure B-9 shows the age distribution at failure versus the year of failure for XLPE cables. The
six failures through 1985 are of one unique, filled XLPE insulation system at one plant. Failures
after that point are a mix of the filled XLPE cables at that one plant and nonfilled XLPE cables
at several plants.
30
25
Age at Failure/Number of Failures
20
15 Age at Failure
10
0
1975 1980 1985 1990 1995 2000 2005
Year of Failure
Figure B-9
Age at Time of Failure for Wet Cross-Linked Polyethylene Cables
B-13
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Figure B-10
Inaccessible Cable Failures by Plant, from Responses to Generic Letter 2007-01 [4]
The six plants with five or more failures account for 73% of the 121 total failures, and the
remaining 19 plants with one to three failures account for 27%. This distribution is heavily
skewed by the plant with 43 events associated with physical damage to the cables. If these
failures are ignored, 78 failures remain. The six plants with five or more failures account for
56% of the failures.
Figure B-11 shows the number of failures by year. The 1993 number is an outlier associated with
rodent problem at Browns Ferry. This chart appears to indicate that the number of failures per
year is increasing; however, the chart is not normalized to reflect the number of units in service
in the United States.
B-14
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Figure B-11
Number of Failures of Wet Cable by Year
Figure B-12 shows the failures with respect to the age of the cable. This chart indicates that the
failure rate is relatively constant and shows no peaking with respect to age of the wet cable. The
large spike at year 19 is once again associated with the rodent-related events at Browns Ferry
in 1993.
Figure B-12
Wet Medium-Voltage Cable Failures Versus Age at Time of Failure
B-15
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Figure B-13 provides an analysis of the 68 failures for which insulation type was listed in the
response. The data indicate an increasing trend in black EPR failures with length of service and a
stable trend with time for XLPE. The data seem to indicate a decreasing failure rate for pink
UniShield type cables. The three pink EPR failures seem to be “infant mortality” type failures.
Too few data points exist for the brown nonshielded EPR and butyl rubber cables to identify
a trend.
5
XLPE
Black EPR
Pink Uni-Shield
Pink EPR
4
Brown Unshielded EPR
Butyl Rubber
Number of Failures
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
Age at Failure
Figure B-13
Age of Cable at Time of Failure By Insulation and Design
B-16
10296123
Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01
Table B-6
Summary of Assessment and Testing Responses from Generic Letter 2007-01
Number of Percentage
GL 2007-01 Medium-Voltage Cable Assessment and Testing Responses Plants of 65
Committed Plants
Manhole Inspection 25 38.5
Manhole Pumps and Dewatering 20 30.8
Very-Low-Frequency Tan δ 5 7.7
Very-Low-Frequency High Potential 3 4.6
On-Line Partial Discharge Testing 3 4.6
Power Factor Testing 2 3.1
DC High Potential (Mostly of Motor with Cable; One Cable Separately) 9 13.8
Insulation Resistance Test with Load 47 72.3
Polarization Index with Load 20 30.8
B-17
10296123
10296123
C
TAN δ DATA FOR ETHYLENE-PROPYLENE RUBBER
AND BUTYL RUBBER INSULATED CABLES
This section provides tan δ data from in-plant and laboratory tests for ethylene-propylene rubber
(EPR) and butyl rubber. The EPRI report Aging Management Program Guidance for Medium
Voltage Cable Systems for Nuclear Power Plant (1020805) provides assessment criteria based on
the data in this section and expert opinion [5].
Because butyl rubber and EPRs are compounds, variations in normal tan δ can be expected
between manufacturers’ formulations and between generations. Some variations (~20%) have
been noted within a particular manufacturer’s insulation of the same generation. In addition,
circuit configurations can affect criteria. One plant that has nonshielded cable for dry segments
of cable in series with shielded cables in wet locations has noted much higher tan δ results for
normal condition.
The following subsections evaluate data from laboratory and field tests to help identify criteria
for retaining and replacing EPR and butyl rubber cable based on tan δ results. The approximate
range of normal and abnormal results can be identified. The data are provided for specific
manufacturers’ formulations and vintages. The data might not apply to other manufacturer’s
formulations or vintages of cable.
C-1
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
Figure C-1
AC Breakdown Strength Versus 60-Hz Tan δ Results
The line in Figure C-1 is a regression curve obtained by the formula y = c + a*exp(-bx) with R2
(coefficient of determination, equal to 0.83), c = 0.26, a = 56.4, and b = 0.69.
The graph demonstrates the predicted breakdown strength of the sample up to 9 × 2.4 kV, based
on dissipation factor measured at the cable operating voltage of 2.4 kV. After the field faults
were removed from the segments, the lowest laboratory breakdown voltage was four times the
operating voltage for the cables. The dry result was obtained by drying a cable segment in a
C-2
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
203°F (95°C) circulating air oven for 5 days. The data indicate that, for a 60-Hz measurement,
a tan δ of 40 × 10-3 the breakdown strength of the insulation is 4 × V0 (9.6 kV), which cable
experts consider extremely weakened insulation (~70 V/mil [2.7 kV/mm]). With a tan δ of
4 × 10-3, the breakdown strength is about 8.6 × V0 (20.6 kV), which cable experts consider
highly weakened (147 V/mil [5.8 kV/mm]). Industry experts consider that a value of 200 V/mil
(7.9 kV/mm) is indicative of end of life for an XLPE cable. The dry value for the black EPR
indicates that the ultimate breakdown voltage for the material was in excess of 55 kV or 394
V/mil (15.5 kV/mm). New cable would likely have a breakdown strength of twice this value.
The failure had been cut out of the cable before these tests were performed, and the sections with
elevated tan δ and low breakdown voltages were determined to have channels of localized low
insulation resistance through the depth of the insulation. The data indicate that there is an inverse
correlation between tan δ and breakdown voltage. These tests were performed with a 60-Hz test
set because of the short lengths involved. The results are not directly comparable to results from
VLF testing, which would likely have values on the order of twice the 60-Hz values.
Figure C-2 shows the 60-Hz tan δ results for the same test specimens by voltage step. Segments
2A and 2B resulted when the fault was cut from the original specimen 2. The removal of the
fault left two segments with reasonably low tan δ measurements and reasonable stability between
voltage steps. Specimens 3 and 4 had high initial tan δ measures at 0.25 V0 and unstable readings
at 0.5 V0, preventing measurements at higher voltages. Specimen 5 had somewhat high tan δ at
0.25 V0 and 0.5 V0 and had a step change increase at 0.75 V0, with instability at 1V0. These three
specimens were found to have low resistance channels between the conductor and shield where
there were small areas having 1/10 or less of the surrounding insulation resistance. When
specimen 4 was evaluated, an additional failure was identified. When it was cut out, the
remaining segment had tan δ measurements between 5.9 × 10-3 and 5.3 × 10-3 through all the
voltage steps. The data for all voltage steps are shown in Figure C-2 because many of the cable
subsections were degraded to the point that testing at 1V0 and 2V0 was not possible. Most other
data in this report are given in terms of 1V0 and 2V0 and the differential between the two.
C-3
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
90
80
70
60
60 hz Tan Delta (E-3)
0.6 kv
50 1.2 kV
2.4 kV
3.6 kV
40 4.8 kV
Diff 2Vo-Vo
30
20
10
0
1 2 3 4 5 6 7 8 9 10 11
Specimen
Figure C-2
60-Hz Tan δ Results by Test Voltage
The data in Figure C-1 show that there is an inverse correlation of tan δ with breakdown
strength. Elevated tan δ indicates that the insulation is weakened and susceptible to in-service
failure. Although Figure C-2 was developed in an attempt to show that significant increases in
tan δ are indicative of deterioration, only specimen 5 tends to show this condition. However,
specimens 3 and 4 partially support this concept, in that the tan δ measurement became unstable
at the second voltage step.
C-4
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
4
Tan Delta (E-3)
1.1 kV
3 3.3 kV
Diff 1.1 to 3.3 kV
0
1 2 3
Figure C-3
Very-Low-Frequency Tan δ Result for a 5-kV Okonite Black Ethylene-Propylene Rubber
Cable
C-5
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
12
10
8
Tan Delta (E-3)
8 kV
6 16 kV
Diff 8 - 16 kV
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Figure C-4
Very-Low-Frequency Tan δ Results for 15-kV Okonite Black Ethylene-Propylene Rubber
Cable
C-6
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
3.5
2.5
2
Tan Delta (E-3)
8 kV
16 kV
8-16 kV Diff
1.5
0.5
0
A11 2 A2 3 B1 C14 C2
5
Specimen
Figure C-5
60-Hz Tan δ for Anaconda UniShield Pink Ethylene-Propylene Rubber
5
Tan Delta (E-3)
8 kV
4 16 kV
8-16 kV Diff
0
A1 1 A2 2 B1 3 C14 C25
Specimen
Figure C-6
Laboratory-Measured Very-Low-Frequency Tan δ for Anaconda Pink Ethylene-Propylene
Rubber UniShield
C-7
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
Figure C-7 shows field VLF tan δ results for 15-kV Anaconda UniShield cables. This chart
shows good through moderate condition results. The results for two dry cables (six individual
phases) are shown at the left of the plot. The remainder of the cables were wet for 50% to 57% of
their length. The order of this plot is by the value of the initial measurement at 8 kV. The dry
results and many of the wet cable results correlate well with the results shown in Figure C-6 for
the laboratory tests on the specimens after the damaged sections were removed. (Each cluster of
two voltages and a difference represents the test of one cable phase.)
30
25
20
Tan Delta (E-3)
8-kV
15 16-kV
Diff 16-8 kV
10
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57
Figure C-7
Tan δ Results for Anaconda Pink UniShield Cables (Ordered by 8-kV result)
C-8
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
Figure C-8 shows the same field VLF tan δ measurements arranged in the order of the difference
between the 2V0 (16-kV) and 1V0 (8-kV) measurements. When ordered by this difference, the
results tend to also be arranged in the order of most elevated 1V0 to 2 V0 result. (Each cluster
represents the test of one cable phase.)
30
25
8-kV
16-kV
Diff 16-8 kV
20
Tan Delta (E-3)
15
10
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57
Figure C-8
Tan δ Results for Anaconda Pink UniShield Cables (Ordered by Difference in 16 kV and
8kV Results)
C-9
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
Figure C-9 shows the field data in order of the value at 2V0. In this figure, entries with 2V0
values less than 10 × 10-3 have been omitted, and a number of entries have been added for cables
having much higher tan δ measurements. Some of these tests with higher measurements were
stopped before the 1.5 V0 or 2.0 V0 measurement was taken, because the results indicated
significant problems and possible failure during the test. (Each cluster of four voltages and a
difference represents the test of one cable phase.)
140
120
4-kV
8-kV
12-kV
16-kV
100
Diff 16-8 kV
80
Tan Delta (E-3)
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
Figure C-9
Tan δ Results for Anaconda Pink UniShield Cables (Showing Moderate to High Results)
Figure C-10 shows tests that occurred approximately one year apart for four Anaconda
UniShield cables. The cables were tested in either the winter or the spring. Cable C-3 (arbitrary
circuit number for this report) shows acceptable tan δ values with little difference between the
measurements taken at 1V0 and 2V0. The A and C phases of cable C-1 show stable and good
results from year to year. The B phase has elevated results at 2V0 and a differential that is nearly
equal to the measurement at 1V0 in 2007. In 2008, the values for the B phase have dropped
somewhat, which could be the effect of temperatures or bulk wetting of the cable at the time of
the test. For cable 2, the tests one year apart indicate that the A and B phases have nearly
identical behavior. However, the C phase has a more significant change at 2V0 in 2008 than in
2007 and the difference between 1V0 and 2V0 has nearly doubled, indicating a worsening
leakage current. Finally, the data for cable C4 show that all three phases have improved in a
similar manner from the 2006 tests to the 2007 tests, indicating again that the temperature might
have been lower or the circuits might have been drier during the 2007 tests. Given that these
cables are showing uniform behavior between the phases, it is likely that the degradation is not
worsening significantly and has remained stable.
C-10
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
25
20
8-kV
16-kV
Diff 16-8 kV
15
Tan Delta (E-3)
10
0
7
7
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
2
2
-2
-2
-2
-2
-2
-2
-2
A-
B-
A-
B-
A-
B-
A-
B-
A-
B-
A-
B-
A-
B-
A-
B-
C
C
C
C
1-
1-
2-
2-
2-
3-
4-
4-
4-
4-
1-
1-
2-
3-
3-
3-
1-
1-
2-
2-
3-
3-
4-
C
C
C
C
Figure C-10
Tan δ Measurements for Anaconda UniShield Cables Taken One Year Apart
The data shown in Figure C-10 indicate that comparing results between phases and between
years for multiple phases is important to determine whether apparent changes are the result of
environmental conditions at the time of the test or are actual changes in the condition of the cable
insulation.
C-11
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
18
16
14
12
2 kV
Tan Delta (E-3)
10
4kV
6 kV
8 kV
8
Diff 4 - 8 kV
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Figure C-11
Tan δ Measurements for Anaconda UniShield with Black Insulation
C-12
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
1000
800
600
500
400
300
60 Hz Dissipation Factor at Cable Operating Voltage, tan (E-3
200
100
80
60
50
40
30
20
10
8
6
5
4
1
0 2 4 6 8 10 12 14 16 18 20
Figure C-12
Tan δ Versus Breakdown Strength for Okonex Butyl Rubber
C-13
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
Figure C-13 compares the 60 Hz tan δ for black Okonite EPR as shown in Figure C-1 to that for
Okonite butyl rubber from Figure C-12. The characteristic shape of the plots is essentially
identical for the two materials such that as the breakdown strength drops below 10 times
operating voltage, the tan δ value is beginning to increase. For the butyl and black EPR rubber
specimens in the assessments, the 10 times operating voltage value is 24,000 Vac. Due to
insulation thickness differences the voltage stress for the two materials was significantly
different. For the butyl rubber, it was 120 V/mil (4.7 kV/mm) and for the black EPR, it was
141 V/mil (5.5 kV/mm). At a breakdown strength of eight times operating voltage (19,200 Vac),
the tan δ for both materials has doubled (2X for EPR, 2.4X for butyl) from its lowest value. As
the breakdown strength decreases to four times operating voltage, the tan δ has increased by a
factor of 16 for the black EPR and by 24 for the butyl rubber. The comparisons show that the
lossier butyl rubber will have characteristically higher 60 Hz tan δ and have a larger change than
the black EPR as it deteriorates.
1000
Butyl
Black EPR
100
Tan Delta (x 10-3)
10
1
2 4 6 8 10 12 14 16 18 20 22 24
ac Breakdown Voltage/Operating Voltage
Figure C-13
Comparison of 60-Hz Tan δ Versus Breakdown Strength of Okonite Butyl Rubber to Black
Ethylene-Propylene Rubber
C-14
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
Figure C-14 shows the same data as Figure C-13 but in terms of volts per mil. The figure shows
that tan δ does not change for Okonite black EPR and butyl until the breakdown strength has
decreased below 200 to 240 V/mil (7.9 to 9.5 kV/mm). For both materials, tan δ is slightly
greater than 10 × 10-3 when the ac breakdown voltage has dropped to 100 V/mil (3.9 kV/mm).
When tan δ has increased to 30 × 10-3, the breakdown strength has dropped to 75 V/mil
(2.9 kV/mm).
1000
100
Tan Delta (E-3)
Butyl
Black EPR
10
1
0 50 100 150 200 250 300 350 400
ac Breakdown (V/mil)
Figure C-14
AC Breakdown Voltage (in V/mil) for Okonite Butyl and Black Ethylene-Propylene Rubber
Insulations
The EPRI report Advanced Diagnostics and Life Estimation of Extruded Dielectric Cables
(1001727) also describes partial discharge (PD) testing performed on specimens in the laboratory
[88]. These butyl rubber cables had high natural attenuation due to the characteristics of the butyl
rubber and had corroded but intact helically wrapped copper tape shields. When the full length of
one of the cables with severe water-related degradation was tested, PD could not be detected
above background. However, an appreciable PD (50–150 pC, depending on test voltage) could
be detected on the full length of the second cable . As both of the cables were cut into shorter
sections following breakdowns, PD was identifiable in each section with elevated tan δ. The
result of the PD measurements indicates that PD can be present in severely deteriorated butyl
rubber cable that is near failure; however, the combination of high attenuation of the butyl rubber
and shield corrosion that causes the tape shield to act as an inductor and further attenuate the
signal might make detection of PD difficult to impossible.
C-15
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
C-16
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
The laboratory VLF tan δ results for Anaconda pink EPR are consistent with healthy field
results. Unfortunately, no complete tan δ versus ac breakdown curve is available for the
Anaconda pink EPR. Figure C-8 indicates that increasing tan δ relates to deterioration, as does
the difference in tan δ when measured at 1V0 and 2 V0. From Figure C-8, there appears to be a
strong tendency of differential increases of tan δ with increasing test voltage for the Anaconda
cable. Figure C-9 indicates that, in some cases with a tan δ of above approximately 30 × 10-3, the
differential increases with increasing test voltage are so large that testing at 1.5V0 to 2V0 cannot
be safely accomplished and that investigation of the need to replace the cable will be necessary.
90
80
70
1.1 kV
2.2 kV
3.4 kV
60
4 kV
Tan Delta (E-3)
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
Figure C-15
Tan δ Results for Mixed Shielded and Nonshielded Black Okonite Ethylene-Propylene
Rubber Circuits
C-17
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
Figure C-16
Tan δ Results for Mixed Shielded and Nonshielded Black Okonite Ethylene-Propylene
Rubber Circuits
In Figure C-15, the cable tests represented by markers 37, 38, and 39 are a special case for two
reasons: 1) only 15% of the cable was shielded and 2) each phase had three conductors in
parallel. The large amount of nonshielded cable significantly elevates the tan δ. The further
elevation of the tan δ results could be caused by having the nonshielded cables between the test
set and the shielded sections. The effects of the length of the nonshielded sections on the tan δ is
described below.
In Figure C-16, the data for a three-phase cable are associated with x scale markers 4, 5, and 6.
The data for the A phase of the cable is shown at scale marker 4. This phase had elevated tan δ
results and a significant differential between 1V0 and 2 V0. The aged wet section of the cable was
replaced with modern Okonite pink EPR. Also in Figure C-16, the C phase (marker 27) of cable
having the x scale markers 25, 26, and 27 shows similar degradation. The utility plans to replace
the cable based on the tan δ measurements that have been observed.
C-18
10296123
Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables
Figure C-17 shows a plot of the average tan δ for the mixed circuits versus the percent of
nonshielded cable in the circuits. Although the correlation is not perfect, the tan δ is
proportionally related to the percent of the nonshielded cable in the circuit. This result is logical,
in that the nonshielded segments would have a lower capacitive current than the shielded
circuits. At the same time, there would still be leakage currents in the nonshielded circuits with
respect to the armor in the three-phase cable segments. The approximate relationship of tan δ to
the percent of nonshielded cable is interesting but not accurate enough to allow an adjustment so
that an absolute value of tan δ could be used to assess a cable. However, comparison of results
between phases and significant increases in tan δ with increasing test voltage obviously can
identify deteriorated insulation, even though the circuits have a mix of shielded and nonshielded
segments.
70
60
50
Average Tan Delta (E-3)
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
% Unshielded
Figure C-17
Average Tan δ Value Versus Percent of Nonshielded Cable Section
The testing of the circuits was by means of connecting the test set to the conductor and to the
station ground. It was not necessary to connect directly to the cable shield connection.
C-19
10296123
10296123
D
ADDITIONAL POLYMER MATERIALS INFORMATION
D-1
10296123
Additional Polymer Materials Information
Figure D-1
A Cross-Linked Elastomer
D-2
10296123
Additional Polymer Materials Information
Zinc Oxide 5
Calcined Clay 90
Sulfur 1.5
D-3
10296123
Additional Polymer Materials Information
The tensile strength for this composition was 930 psi (6.4 MPa); elongation, 290%; moisture
sorption, 10.9%; and dielectric strength, 695 V/mil (27.4 kV/mm). The curing agent system plays
a significant role in resistance to degradation under thermal stress. The sulfur cure can soften
with thermal aging, whereas the lead oxide cure tends to remain stable.
Polymers such as phenol-formaldehyde resins can be used as the cross-linking agent for butyl
rubber, although no specific notation for wire applications is included in the referenced literature.
What is relevant is that each method of inducing cross linking will impart a different nature to
the cross-linked regions, different additives in the final rubber, and potentially different
responses with aging.
Figure D-2
Ethylene-Propylene-Diene Monomer Terpolymer
Cable accessories (such as joints and terminations) are made of EPDM, and manufacturers’
literature designates them as EPDM.
D-4
10296123
Additional Polymer Materials Information
Amount (Relative
Component
Weight)
EPR Vistalon 404 80
Low-Density Polyethylene 20
Zinc Oxide 5
Coated Clay 110
Carbon Black 10
Flectol H 1.5
Lead Oxide 5
Wax 5
Silane 2
SR-350 1.5
Peroxide 7
Significant properties for this black, filled EPR are shown in Table D-4.
Table D-4
Properties of Black Ethylene-Propylene Rubber Compound Shown in Table D-3
Property Value
Tensile Strength 1210 psi (0.85 kgf/mm2)
Elongation 200%
Elongation Percent Retention After Seven Days at 300°F (149°C) 95%
D-5
10296123
Additional Polymer Materials Information
The moisture sorption is greater than an order of magnitude less than that for butyl rubber. Also,
the dielectric strength is higher, and the physical properties are excellent. A similar (but not
exactly the same) formulation for black power cable insulation from the 1970s was provided by
Exxon [26] using Vistalon 2504, an EPDM (see Table D-5).
Table D-5
Medium-Voltage Black Ethylene-Propylene Rubber Insulation Using Ethylene-Propylene-
Diene Monomer
Amount (Relative
Component
Weight)
EPDM Vistalon 2504 100
Carbon Black 10
Treated Clay 110
Antioxidant 1.5
Lead Oxide 5
Zinc Oxide 5
Process Oil 15
Paraffin 5
Silane 1
Peroxide 3.5
Formulation components can differ between suppliers. In addition, although the function of the
components in any EPR formulation is well known, specific individual components used by
individual suppliers can differ from one another.
D-6
10296123
Additional Polymer Materials Information
Table D-7 shows the effect of treatment of the clay on EPR properties and the limitations of
calcined clay with respect to electrical properties in EPR. The power factor, dielectric constant,
and water sorption characteristics all improve in the treated calcined clay system. This indicates
that the use of treated calcined clay has been a significant improvement in compounding EPR.
Table D-7
Influence of Clay Nature on Ethylene-Propylene Rubber Cable Properties: Calcined Versus
Coated Clay
Another factor in influencing elastomer properties is clay particle size—the smaller the average
size, the greater the polymer–inorganic component interfacial contact and the better the
anticipated properties. More even size distributions with smaller-sized particles have been
implemented in cable manufacture, as well, to improve electrical and mechanical properties.
D-7
10296123
10296123
E
OFF-LINE TESTS THAT ARE UNDER DEVELOPMENT
E.1 Introduction
Several additional diagnostic tools are being developed but are not commonly used. They are
summarized in this section. Additional information is available in the EPRI reports Advanced
Diagnostics and Life Estimation of Extruded Dielectric Cables (1013085) and Guide for Non-
Destructive Diagnosis of Extruded Cable Systems (1001731) [104, 90].
Figure E-1
Isothermal Return Current Plots for Cross-Linked Polyethylene Cable Having Undergone
Different Degrees of Aging
E-1
10296123
Off-Line Tests That Are Under Development
An advantage of this diagnostic method is that the important data are collected after the dc stress
has been removed; a disadvantage is that the return currents are extremely small.
A typical procedure applies 1000 Vdc for 30 minutes (to allow the current to saturate due to the
dc conductivity), grounding occurs for ~20 seconds, and then measurements are made for the
next 30 minutes while discharging is taking place. The use of relatively low dc voltage is
claimed to be unable to inject charges into the system. This process induces a polarization within
the insulation system. The depolarization current is considered to be due to trapped charge
release as well as polymer chain motion. However, the currents that are measured are likely not
only due to the insulation response to the applied dc, but also to the entire cable construction; for
example, shield-insulation interfaces, polymer-filler interaction, impurities such as water (in
addition to likely polar water treed regions), and cable length (joints further increase the
complexity). The analysis of the data generated requires the use of computer-based network
evaluation tools to estimate the aging status of the cable. Information about the cable
construction and length can be fed into the database, but not the obviously more subtle
information. Regardless, the interpretation is quite complex, and it has been noted that IRC can
be considered to be a superposition of different components exhibiting different relaxation times.
E-2
10296123
Off-Line Tests That Are Under Development
E-3
10296123
10296123
F
INSULATION RESISTANCE TEST MEASUREMENTS:
THEIR VALUE AND LIMITATIONS
F.1 Introduction
Very-low-frequency or ac withstand, tan δ, dielectric spectroscopy, and partial discharge (PD)
testing are the preferred test methods for extruded polymer testing, depending on the potential
problem being evaluated and the nature of the cable design. However, insulation resistance
testing is often used to assess cables due to its ease of application and ready availability.
Insulation resistance measurement is a valuable troubleshooting tool if used properly, and in
some cases, it will be a good indicator of severe degradation of a cable’s insulating system.
However, insulation resistance measurements must be used cautiously as a means of
understanding age-related degradation. Some plants have adopted inappropriately low values
of insulation resistance for acceptance criteria for return-to-service for medium-voltage cable.
In addition, this acceptance criterion does not take circuit length into consideration. This section
describes why higher values of insulation resistance are appropriate for cable insulation
assessment and provides cautions regarding sole dependence on insulation resistance for
assessment of medium-voltage cable. For nonshielded cable, insulation resistance is likely to be
the only test that is practical to use. Consideration of the cautions provided in this section will
improve the value of the test until an improved test is available.
Where:
IR1min is the recommended minimum insulation resistance, in MΩ, at 104°F (40°C), of
the entire three-phase winding.
kV is the rated machine terminal-to-terminal voltage, in rms kV.
F-1
10296123
Insulation Resistance Test Measurements: Their Value and Limitations
If individual phases were to be tested separately and guard circuits were used for the phases not
under test, the minimum insulation resistance is three times this value.
IEEE Std. 43-2000 contains an additional minimum insulation resistance value of 100 MΩ for
form-wound windings built after 1970. For separately tested phases, the requirement is 300 MΩ.
Although the minimum insulation resistance of Equation F-1 was meant to apply only to wet or
deteriorated windings, the industry adopted it for general use and applied it to cables and the
combination of motors tested with their field cable. Insulation resistance testing of the motor
through the extruded polymer-insulated cable generally will not be a problem with respect to
understanding the condition of the motor winding insulation system. As long as the cable has not
degraded, the test will be dominated by the motor. However, a number of problems occur when
the condition of the cable is evaluated either with the motor or by itself using the criteria of
Equation F-1. The minimum insulation resistance of Equation F-1 is applicable to a varnished or
asphalt winding that was wet to determine if it was safe to re-energize or to expose to an elevated
voltage test. The leakage currents were not necessarily associated with a localized deterioration
but rather a generalized leakage over the entire surface of insulation the winding. The 100 MΩ
requirement for form-wound windings built after 1970 applies to insulation systems that have
epoxy and/or mica tape insulation systems that have better barriers to surface moisture and,
accordingly, much higher insulation resistance values, even when damp. Extruded medium-
voltage cable insulation systems do not behave like random-wound varnished or asphalt
windings. Wetting the jacket of a shielded cable will have no immediate effect on insulation
resistance. Although wetting the jacket of nonshielded cable might provide a return path for the
test, it will not cause a low insulation resistance. The criteria of Equation F-1 from IEEE Std. 43
definitely do not apply to extruded polymer insulation. The 100–300 MΩ value for post-1970
form-wound motors comes closer to being a valid minimum. However, the following are
also true:
The insulation resistance of a cable is indirectly proportional to its length. Accordingly,
minimum values must be corrected to a specific length, such as 1000 ft or 1 km. If the
minimum insulation resistance is 100 MΩ-1000 ft (30.5 MΩ-km), and the circuit length is
250 ft (76 m), the minimum insulation resistance is 400 MΩ. Similarly, if the length is
2000 ft (61 m), the minimum insulation resistance is 50 MΩ.
The introduction to IEEE Std. 43-2000 states “Whereas individual insulation resistance
measurements may be of questionable value, the carefully maintained record of periodic
measurements, accumulated over months and years of service, is of unquestioned value as a
measure of some aspects of the condition of the electrical insulation.” There are many
variables associated with insulation resistance testing of motors, including temperature and
humidity. The same is true for extruded polymer insulation of cables. Gross trending of the
results over time is useful; however, trying to evaluate small changes (factors of 2 to 10 for
high values) is inappropriate. General decreasing trends of decades are important.
F-2
10296123
Insulation Resistance Test Measurements: Their Value and Limitations
Insulation resistance testing is a global or bulk test. It cannot differentiate between an overall
decrease in the insulation resistance in the wall of the polymer and one or two severe,
localized, low insulation resistances. In addition, if a near-through-wall defect has a thin
layer of good insulation between it and the shield or conductor, the insulation resistance test
will not identify it.
Even when evaluating a failed cable, the insulation resistance can be tens of megohms or
more because the fault is likely to have blown out the metal shield, causing long surface
resistance paths between the exposed conductor and the remainder of metallic shield.
Insulation resistance measurement is a useful troubleshooting tool for evaluation of medium-
voltage cable. For extruded cable, values less than 100 MΩ-1000 ft (30.5 MΩ-km) are a strong
indication that a problem exists with the circuit and investigation is necessary. However, the
converse is not necessarily true. Reasonable confidence that the circuit is sound exists when
values of gigohms or higher are present when adjusted for length, but the test will not necessarily
detect voids and partial through-wall damage that could result in ultimate failure under ac
operating conditions. Some improvement in value of insulation resistance results are possible for
cable testing if the following are considered:
The insulation resistances of each phase of a cable should be similar to those of the other
phases. Decades of difference between phases indicates that a problem might exist with a
phase or phases.
Periodic measurements taken over the course of years should be similar. Variations are
expected due to differing conditions at the time of test, but the overall value should remain
high and relatively consistent.
Gross decreasing trends or major decreasing step changes between periodic measurements
are an indication of a problem and should be investigated further. A more sophisticated test is
recommended for shielded cable, such as ac withstand, tan δ, dielectric spectroscopy, or PD,
as applicable.
The insulation resistance measurement should smoothly increase or remain stable during the
period of the test. Instability in the measurement, such as oscillations or continuing decreases
in values, are indications of a breakdown phenomenon. Such oscillations might be present in
a single-phase cable that has failed and the shield continuity has been lost in the vicinity of
the fault. Returning such a cable to service has a high probability of resulting in a phase-to-
phase fault after a short period of operation.
F-3
10296123
Insulation Resistance Test Measurements: Their Value and Limitations
Where:
D is the diameter over the insulation.
d is the diameter under the insulation.
To allow a comparison of the new cable requirement to the minimum acceptance criteria, assume
a 500-ft (152-m) long cable with a 1/0 AWG (53.5 mm2) conductor and an 80-mil (2-mm) thick
600-V rated insulation. A representative cable of this type would have a 0.49 in. (12.4 mm)
diameter. The field acceptance limit would be 1.6 MΩ. The new cable value would equal
the following:
IR = 10,000 log10 (0.49 in./0.33 in.) MΩ * 1000 ft/500 ft =3440 MΩ
If the circuit was 2000 ft (610 m) long, the new cable value would be 860 MΩ. The length of
the circuit is a significant consideration related to expected insulation resistance values.
A newly installed cable should have a near-factory-value insulation resistance per 1000 ft
(305 m), which, in this example, would be 6.8 GΩ. Using 100 to 300 MΩ-1000 ft (30 to
100 MΩ-km) as an acceptance criteria for the continued use of this cable example results in
a value that is 23 to 68 times the allowable reduction from the original minimum insulation
resistance. Most cables, when manufactured, have insulation resistance values that greatly
exceed the minimum manufacturing requirements, making the difference between the as-
manufactured to the degraded condition even larger.
F.4 Conclusions
Insulation resistance is a valuable troubleshooting tool. However, it must be used cautiously as a
condition monitoring tool. The acceptance criterion of Equation F-1 (1 MΩ plus 1 MΩ per kV of
rating) is not adequate for extruded polymer insulation used in medium-voltage cable. A
minimum value of 100 MΩ-1000 ft (30 MΩ-km) is recommended to preclude severely degraded
cables from being returned to service. Even when this value is used, higher values of insulation
resistance might not be indicative of a sound insulation system. The cautions and techniques for
improving the value of the results described in Section F.2, Determining Minimum Insulation
Resistance Value and Improving Interpretation of Results, should be considered.
F-4
10296123
10296123
Export Control Restrictions The Electric Power Research Institute, Inc. (EPRI, www.epri.com)
Access to and use of EPRI Intellectual Property is granted with the spe- conducts research and development relating to the generation, delivery
cific understanding and requirement that responsibility for ensuring full and use of electricity for the benefit of the public. An independent,
compliance with all applicable U.S. and foreign export laws and regu- nonprofit organization, EPRI brings together its scientists and engineers
lations is being undertaken by you and your company. This includes as well as experts from academia and industry to help address
an obligation to ensure that any individual receiving access hereunder challenges in electricity, including reliability, efficiency, health, safety and
who is not a U.S. citizen or permanent U.S. resident is permitted access the environment. EPRI also provides technology, policy and economic
under applicable U.S. and foreign export laws and regulations. In the analyses to drive long-range research and development planning, and
event you are uncertain whether you or your company may lawfully supports research in emerging technologies. EPRI’s members represent
obtain access to this EPRI Intellectual Property, you acknowledge that it approximately 90 percent of the electricity generated and delivered in
is your obligation to consult with your company’s legal counsel to deter- the United States, and international participation extends to more than
mine whether this access is lawful. Although EPRI may make available 30 countries. EPRI’s principal offices and laboratories are located in
on a case-by-case basis an informal assessment of the applicable U.S. Palo Alto, Calif.; Charlotte, N.C.; Knoxville, Tenn.; and Lenox, Mass.
export classification for specific EPRI Intellectual Property, you and your
Together...Shaping the Future of Electricity
company acknowledge that this assessment is solely for informational
purposes and not for reliance purposes. You and your company ac-
knowledge that it is still the obligation of you and your company to make
your own assessment of the applicable U.S. export classification and
ensure compliance accordingly. You and your company understand and
acknowledge your obligations to make a prompt report to EPRI and the
appropriate authorities regarding any access to or use of EPRI Intellec-
tual Property hereunder that may be in violation of applicable U.S. or
foreign export laws or regulations.
Program:
Generation Maintenance Applications Center
© 2012 Electric Power Research Institute (EPRI), Inc. All rights reserved. Electric Power
Research Institute, EPRI, and TOGETHER...SHAPING THE FUTURE OF ELECTRICITY are
registered service marks of the Electric Power Research Institute, Inc.
1024044