This research is focused on better understanding the relationship between conditions that can cause heating, the resulting heating that may be identified by IR inspection, and the expected remaining life of the heated connectors. The results of this research will inform field practices and infrared thermography.
This section describes a multi-year project investigating the performance of low voltage (LV) cable accessories. The research has included investigation of the causes of heating in LV connectors and sought data to develop representative secondary connector load versus temperature characteristics. Research included the development of a testing facility at EPRI’s Lenox Laboratory to enable testing of LV connectors under various loading histories and measure the steady state thermal response to develop load characteristics, including the time to heat and temperatures measured with embedded thermocouples and Infrared (IR) imaging.
The research has focused on a one-way to one-way 500 MCM copper (Cu) to 500 MCM Cu crimp type connector used in a low voltage secondary network application. With the help of Con Edison, EPRI prepared 26 different connections for study, including samples with various intentional defects. The connectors were linked in series and connected to a programmable current power supply which enabled EPRI to control and vary the current to simulate various loading and heating conditions. Thermocouples were attached to each of the connections, enabling monitoring and recording of temperatures.
This research seeks to better understand the relationship between loading and temperature based on connector defect type and intentional workmanship variations. The results of this research will inform workmanship methods and provide information for improved interpretation of results from infrared thermographic inspections of low voltage connectors.
Update video describing low voltage connector test facility at EPRI’s Lenox Laboratory.
Background, research objectives, and approach of the LV connector accessory project
Introduction
Utility companies that own and operate urban underground distribution
systems face multiple challenges. Installing, assessing and maintaining
underground infrastructure is inherently challenging due to limited
accessibility and the inescapable environmental conditions.
Additionally, installed underground distribution infrastructure is
aging, and much of the early installed plant is reaching the end of its
useful life. Utility managers are charged with managing costs, improving
system reliability and resiliency, increasing power throughput, and
ensuring the health and safety of workers and the public. In addition,
significant changes to electric distribution systems are underway,
including the application new load types, new distributed generation
sources, new underground equipment and materials, and automation
technologies. At the same time, many utilities are facing loss of
institutional knowledge. EPRI’s Underground Assets research project
focuses on providing utilities the knowledge to acquire, optimize, and
maintain underground distribution infrastructure.
One challenge faced by operators of low voltage secondary network
systems is assessing and assuring the health of low voltage (LV)
connectors (splices). Utilities use low voltage connecters in urban
underground manholes and vaults to join low voltage cable sections,
using either a crimp type, or a bolted type connection. These
connections will experience varied levels of loading throughout their
service life and may be subjected to high levels of loading when a
network is operating in a contingency (N-1 or N-2, for example)
condition. Overheating of secondary cables and connectors in network
secondary systems can lead to failures of either or both - a not
uncommon failure scenario [1].
Overheating of secondary cables and connectors, and subsequent failures,
may be due to various causes, including workmanship issues, component
deficiencies, contamination at cable-connector interfaces and design
errors. To assess the ongoing ‘health’ of low voltage connectors, many
companies perform periodic inspections utilizing infrared imaging
thermography (IR) to identify connectors that are ‘hot’, thus candidates
for replacement. However, the relationship between loading, temperature,
and expected remaining service life is not well understood. Further, the
industry lacks information regarding the implications of various
workmanship deficiencies on eventual connector heating and service life
expectancy.
Figure 1:
Overheated secondary connector – IR thermography image
Figure 2:
Overheated secondary connector from Figure 1, as-removed from service
In 2019, EPRI launched research into the causes of heating in LV cable
connectors and began gathering data that will be used to understand
fundamental secondary connector load versus temperature characteristics.
Research included examination of field-aged and overheated low voltage
cable and cable accessory materials, and development of a dedicated
testing facility at EPRI’s Lenox Laboratory. This facility will enable
stressing LV cable connectors under various static and dynamic load
conditions while measuring the steady state thermal response to
understand thermal behavior, including the heating rate and temperatures
measured through embedded thermocouples and IR imaging.
Project Objectives
The overall goal of this research is to better understand the
relationship between current loading and temperature rise based on LV
connector crimp practices and cable conditions to inform workmanship
methods and to inform infrared thermographic inspection.
This multi-year project seeks answers to the following research
questions:
What is a typical secondary splice load vs. temperature
characteristic?
How is a ‘high’ (non-recoverable) temperature defined for a
secondary splice?
Do hot secondary cable connections continue to become hotter over
time under constant load?
How much time is required for an overheated connection to thermally
destroy the overlying insulation and/or the insulation over the
attached cables, considering historic and present materials?
What are common causes of hot spots in LV connectors?
What are the implications of workmanship to LV connector
performance?
Do some splice installation tools better mitigate workmanship
errors?
Currently, the project is focused on analysis of a one-way to one-way
500 MCM Cu to 500 MCM Cu crimp type connections used in low voltage
secondary network applications.
Research Approach
The EPRI research approach consists of two parallel efforts:
Analysis of aged, overheated or failed splices (connectors) of the
type selected above.
Comparison of healthy (normal, properly built) and unhealthy
(abnormal, improperly built) splices of the type selected above, by
subjecting to varying loading and heating over time, and monitoring
performance.
Analysis of aged, overheated, or failed connectors
EPRI will perform analysis of samples of secondary connections (500 MCM
Cu – 500 MCM Cu crimp connections) that are removed because of
identified hot spots through IR, or because of failure. These samples
will be accompanied by information associated with the removed
connectors, including:
Device material specifications
IR images
Other data that shows temperature
Any information about the age of the splice
Any information about the operating environment of the connector,
including loading history, and contingency loads,
Samples will include the housings / coverings / insulation
associated with the failed connectors and cables
Since 2019, EPRI received 11 connectors for analysis as described below.
Comparison of the healthy (normal, properly built) and unhealthy
(abnormal, improperly built) splices
With the help of Con Edison, EPRI prepared 26 different connections for
study, including samples with various known defects. The range of
defects was based on a number of probable adverse conditions in the
field, such as constraints imposed by cable bends and limited free
lengths, as well as work practice challenges known to the Con Edison
crews. Fundamental considerations were also applied, for example, in
joining service-aged cables with non-pretreated conductors. The
connector and adjoined cables were linked in series and energized with a
current power supply that enables EPRI to control and vary the current
in a programmed manner. Thermocouples were directly attached to each of
the connections before the Con Edison standard heat-shrinkable
insulating sleeve was applied. These sensors enable continuous
monitoring and recording of temperatures. In addition, IR thermography
is being used to record temperatures, while enabling comparisons to the
actual connector heating as measured by the thermocouples.
Figure 3:
Preparation of LV connectors for study
Figure 4:
LV crimp connector, 500 Cu MCM, using Con Edison tool
EPRI will ’load-age’ tested units by applying a load / temperature cycle
to simulate the effects of aging and to better understand the time
required for a connection to increase in temperature, given its
installation type (workmanship associated with the preparation of the
connector), and operational parameters.
After aging the units (both the ’normal’ and ‘abnormal’ splices), EPRI
will re-examine the temperature / load characteristic, including the
time to heat and measured temperatures from both a thermocouple and IR
camera, to determine if aging has an impact on the load / temperature
characteristics. Testing will include applying a constant load to an
aged splice to determine if the splice will become hotter over time,
given a constant load.
The load-temperature characteristics gathered during these tests will
inform the load needed to cause cable insulation damage associated with
‘healthy’ and ’non-healthy’ connectors.
The results of this research will inform workmanship methods and provide
information for better interpreting results from infrared thermography
scans of low voltage connectors.
2 - Overheated Connector Examination
Examination of overheated crimp connectors
Introduction
Low voltage (LV) connector heating can be driven by a number of factors
including high levels of loading and component anomalies, such as a bad,
and thus more resistive or reduced-ampacity, connection.
Secondary network systems are designed to carry varying levels of load
(current) in normal modes of operation and in contingency conditions,
where one or more of the medium voltage feeders that supplies a
secondary network grid is out of service. By design, secondary cables
and components are sized so that the levels of loading and load cycle
anticipated in contingency conditions are manageable and within the
emergency loading limits specified for the components. Many utility
companies who operate network systems have developed specifications for
600V secondary cables of different types used in network mains that
include normal and contingency ampacity levels. Con Edison, for example,
provides detailed ampacity tables for various cable types, conductor
sizes, load curves, duct bank loading, based on whether the cables are
being operated in a normal, first (N-1), or secondary (N-2) contingency.
For one particular anticipated winter load cycle curve, for example, 500
MCM Cu cable with a dual layer low smoke zero halogen jacket has a
normal ampacity rating of 670A, a first contingency rating of 715A, and
a second contingency rating of 750A [2].
In practice, situations may occur where these ampacity limits are
exceeded. The network could fall into a contingency situation beyond
N-2. The secondary system could have ‘open mains’, sections of secondary
cables which comprise the mains that are out of service because of cable
limiters which may have blown but are unidentified. This may result in
higher than anticipated current flows in the remaining, energized cable
sections. Finally, anomalies in the components that comprise the
secondary itself may cause ‘hot spots’, areas within the secondary
system which are prone to high heating, such as a poorly assembled LV
connections.
Sample Description
A sample of an overheated LV connector and cable section was provided to
EPRI for examination by Con Edison. The sample provided is a one-way to
one-way 500 MCM Cu crimp connector of the type being studied in this
research. Con Edison has experienced periodic failures of connectors of
this type in their low voltage secondary network system, where these
connectors are used to join 500 MCM Cu cables that comprise their
network secondary mains.
Con Edison identified the sample through a proactive IR inspection
program, aimed at identifying and remedying hot spots before they lead
to failure. Con Edison has established a standard policy for performing
IR on secondary distribution cables and splices. Within that policy, a
“hot spot” is defined as “An isolated segment of cable or connection
operating at a temperature greater than the rest of the cable length. If
the maximum temperature is greater than 55°C (130°F) and 15°C (24°F)
above the temperature of the cable at either duct edge, a hot spot
condition is declared” [3].
Con Edison has established threshold values based on ANSI/NETA ATS-2009 Table 100.18, Thermographic Survey,
Suggested Actions Based on Temperature Rise.
The sample provided experienced high levels of heating as shown by the
IR images shown in Figures 1, 2 and 3 below. Note in Figure 2,
that cracking of the cable jacket and insulation can be clearly seen in
the IR image. Con Edison subsequently removed the overheated sample and
provided it to EPRI for analysis.
Figure 1:
IR Image of heated LV connector
Figure 2:
IR image of heated cable near LV connector
Figure 3:
IR image of heated LV connector
Figure 4:
Overheated LV connector and cable section sample
Visual Examination
A visual examination of the sample corroborated the IR finding that
cable section and connector were subjected to very high heating. The
remaining cable insulation and jacket was cracked and partially
disintegrated over the conductor surface. There was no remaining
evidence of the heat-shrinkable insulation tube, indicating that it had
burned away. The conductor was found heavily oxidized, as evidenced by
the red surface color (cuprous oxide). This is shown in Figure 5.
Figure 5:
Overheated LV network cable with red (cuprous) copper oxide on surface
One theory posited by Con Edison and being explored by EPRI is that the
overheating of the cable / cable connector due to loading may have
caused the shrink tube sealant material to melt, and that molten
material from the inside of the heated shrink tube may have migrated
into open spaces within the connector, including between the cable
strands and between the outer cable strands and the connector barrel.
The addition of this material may have increased the resistance of the
connector assembly, thus increasing the heating rate and decreasing the
corresponding ampacity. The heat-shrinkable tube includes an outer layer
of insulation and an inner layer of a rubbery material (i.e., ‘mastic’
or ‘adhesive’) that melts then solidifies during installation to form a
water-tight seal. Overheating of the materials from which the shrink
tubes are made may also have led to the formation of reactive gaseous
decomposition products which, when they come into contact with the
copper cable strands and connector barrel, may also result in reduced
conductivity (increased resistance) of the connector through formation
of oxides or chlorides, for example.
Any change in performance of the connector (change in resistance) may
not be linearly related to temperature and time. There are organic
antioxidant compounds in the cable insulation and jacket materials that
may inhibit a change in conductivity due to overheating – at least for a
period of time. After the antioxidant has been consumed through heating,
then the change in performance will accelerate. The temperature and time
relationship depends on the chemistry of the insulation and jacket, as
well as the type and concentration of the antioxidant or antioxidants.
To better understand the performance of both the heat-shrinkable tube
materials and of the cable insulation/jacket when subjected to heating,
EPRI will perform a series of thermal exposure tests and corresponding
chemical analyses.
Initial tests, to establish some baseline indications of performance,
were performed at the Lenox Lab, with a Con Edison representative in
attendance. EPRI sectioned a new sample of the Con Edison LV cable
(General Cable 500 MCM Cu with EAM insulation and LSZH jacket), into
roughly one-inch long sections (see Figure 6). Each section was
sequentially placed into a furnace for 45 minutes at a predetermined
temperature to observe changes to the cable conductors, insulation, and
jacket. In addition, EPRI sectioned the shrink tube material into small
strips, roughly two inches long, and ¼ inch wide. These were also placed
in the furnace to observe changes at various temperatures. To conduct
these tests in a single day, coarse temperature increments were selected
based on fundamental materials considerations, such as the oxidation
onset temperature for copper. The exposure temperatures and
corresponding cable response are summarized in Table 1.
Figure 6:
Cable sections prepared for thermal exposure testing
Table 1:
Summary of short-term thermal stress tests
Sample ID
Temp., °F
Temp., °C
Observations
Cable and Jacket
257
125
No visible changes
Heat-Shrink Tube
257
125
Slight softening of adhesive layer
Cable and Jacket
302
150
No visible changes
Heat-Shrink Tube
302
150
Continued softening of adhesive layer, outer sleeve not shrinking
Cable and Jacket
400
204
Initial oxidation of conductor strands; jacket and insulation not affected
Heat-Shrink Tube
400
204
Liquified mastic layer and outer sleeve shrinkage complete
Cable and Jacket
500
260
Initial indication of thermal damage to jacket layer
Heat-Shrink Tube
500
260
Tube fully recovered, liquified mastic layer
Cable and Jacket
600
316
Insulation and jacket layer decomposing and releasing smoke at end of 45 minutes exposure, evidence of liquification and boiling of insulation at conductor interface
Heat-Shrink Tube
600
316
Immediate combustion of heat-shrinkable tube and mastic
Cable and Jacket
700
371
Complete destruction including pyrolized ash
Heat-Shrink Tube
700
371
N/A
Photographs that document some of the thermal exposure tests are
presented in Figure 7, Figure 8, and Figure 9. EPRI will continue
this testing, subjecting cable samples to heating at smaller temperature
increments for longer periods of time.
Figure 7:
Cable and heat-shrinkable sleeve following heating at 257°F/125°C
Figure 8:
Cable condition following thermal exposure at 316°C/600°F
Figure 9:
Cable condition following heat exposure at 700°F/371°C
In total, EPRI examined 11 connectors that were returned from the field.
The table below provides a description and the associated findings for
each of these components.
Conductor strand missing; consequential inadequate crimp force
H
splice with large and small diameter Cu conductors
Two strands missing from 7 conductor wire; crimp pressure non-uniform and asymmetric due to distorted geometry
I
small Cu to large Cu splice
Over-crimping damage of conductor
J
2 Al x 1 Cu Y-splice
singe crimps applied - two recommended by manufacturer and utility work practice
K
welding cable with crimped ring terminal
Ring connector bore drilled oversize from 0.3875 in. to 0.661 in.
Detailed examinations of the overheated connectors revealed the
following concerns. All of these provide direction for improved
practices in the field.
Incorrect crimp barrel and conductor size coordination
Inadequate cleaning of conductors prior to crimping
Inconsistent crimping practices, including depth control and number
applied
Problems with interconnecting copper and aluminum conductors
3 - LV Connector Test Setup
LV connector preparation and setup for testing
Overview
For 2019, the research focused on a ‘one-way’ to ‘one-way’ compression
type tin-plated high conductivity copper connector used join 500 MCM
secondary cables (see Figure 1). This connector, custom-produced for
Con Edison, is a common size and type used in their low voltage
secondary network system and is very similar to other
commercially-available crimp connectors. The design includes a straight
barrel, without a center stop or any embossed or printed labels. The
connector is 4 inches (10.16 cm) long, has an outside diameter of 1.06
in (2.69 cm), and an inside diameter of 0.834 in (2.12 cm).
Con Edison has experienced some in service failures of these connections
caused by overheating due to a variety of factors including abnormally
high loading and anomalies or defects in the preparation of the
connection due to factors such as contamination and variations in
workmanship.
Figure 1:
Low voltage copper crimp connector
Con Edison performs infrared thermographic scans of its in-service LV
connectors to identify hot spots and candidates for replacement. One of
the goals of this research is to better understand the relationship
between loading and temperature both to inform workmanship associated
with LV connectors and response decision-making based on IR findings.
The test set up includes 26 different connectors prepared for study,
linked in series and connected to a current supply which enables EPRI to
control and vary the current in a programmed manner. These connectors
were built with Con Edison’s assistance with various known defects,
described below, as the research will seek to identify and distinguish
load versus temperature profiles over time associated with various
defect types. Depending on the findings, it may be possible to leverage
learnings from the research to predict both the likely type of defect,
and expected remaining life based on an analysis of results from IR
scans.
Preparation of Samples for Study
The connections for study were prepared by a Con Edison field crew at a
Con Edison facility in the Bronx. As one of the goals of the research is
to compare performance of connectors built with various defects, EPRI
prepared a draft listing of defects targeted for study. This draft was
reviewed with Con Edison Engineering personnel and field crew members
and, based on their feedback, modified so that the chosen defects for
study were those which are realistically representative of defects that
may exist in the field.
All connectors were assembled onto the Con Edison standard stranded
copper 500 MCM, FR-EAM insulated 600 V network cable. The connectors
are, as described above, ‘one-way’ to ‘one-way’ compression type,
manufactured with tin-plated high conductivity copper.
All crimps were made with a hydraulic, die-less hand-held tool, with a
line-powered hydraulic pump. This crimping system is custom-manufactured
for Con Edison and has been the standard tool for approximately forty
(40) years. The device includes a flat-faced anvil and an opposing Vee
block support. Crimping ‘shots’ are completed using visual cues only,
with the press stroke controlled by the companion crew member who
receives direction from the tool operator who determines when the crimp
shot is complete. The same tool set includes a separate head for cable
cutting, driven from the same hydraulic pump. Two people are required to
run either the crimping tool or cutter, with one dedicated to pump
control. The term ‘shot’ refers to a single crimp impression. The crimp
is an indent style, with two normally applied over one side of the
connector, co-aligned, and spaced uniformly between the center and end
of the barrel.
Figure 2:
Crimping tool hydraulic pump
Figure 3:
Handheld crimping tool head
Figure 4:
Handheld crimping tool, ready to crimp
Figure 5:
Handheld crimping tool, applying a crimp “shot”
Figure 6:
Completed crimps on one side of connector barrel
The samples were assembled with the crimp defect under study applied to
one end of the connector only, crimped to the end of an approximately 3
ft (91 cm) long section of cable. This resulted in various subassemblies
which were later joined together in a test loop at an EPRI laboratory
using a controlled depth concentric crimping tool. It is noted that Con
Edison has begun implementing battery-powered hand-held crimping tools
that include a controlled crimp depth operation. Evaluation of these
tools by Con Edison is ongoing.
Figure 7:
Sub-assembled cable sections with crimp defects
Table 1describes the various crimping defects prepared for study. Two
or more samples were prepared with each defect type. Three samples were
prepared for defect Type G, which includes crimps made over
contaminated, corroded, or oxidized conductor strands. The cable used
for the Type G defects was obtained from a scrap container at the Con
Edison Bronx facility. An internal marker tape was partially recovered.
This indicates a manufacturing date of 1961. The manufacturer’s name is
incomplete due to where the cable was sectioned. This will be recovered
at a later date when the corresponding cable is removed from testing.
Finally, it is noted that defect Type M is a properly-prepared crimp
connection. The 3 ft. (91 cm) cable lengths prepared at Con Edison were
later interconnected at the EPRI site in Lenox, MA using a
battery-powered, handheld Burndy concentric crimping tool.
Table 1:
Connector defects for evaluation
Defect Code
Defect Description
How Implemented
Notes
Example
A
Dual under-crimp
Crimp duration reduced (pressure cannot be adjusted).
Simulates poor field practice.
2 samples prepared
B
Crimp placement error Type 1
One crimp made at edge of barrel over conductor, other crimp adjacent, but not overlapping. Both crimps offset, but normally spaced and at normal depth.
Simulates poor field practice.
2 samples prepared
C
Crimp placement error Type 2
The two crimps overlapping. Each crimp made to normal depth.
Simulates poor field practice.
2 samples prepared
D
Over-crimp combination
First crimp at cable end excessively deep; second crimp at normal depth. Both crimps positioned normally.
Simulates poor field practice.
2 samples prepared
E
Conductor strands partially excluded
Both crimps at normal depth and position; two conductor strands directed outside of connector.
Simulates poor field practice where 'stray' conductor strands may not be directed into barrel, with condition not visible due to position against vault wall. This may also simulate condition where conductor strands have been nicked and broken away during cable preparation.
2 samples prepared
F
Improper cable cutback
Cable jacket cutback length too short, resulting in incomplete insertion into barrel. This results in 'inboard' crimp incompletely engaging or missing conductor end. Both crimps approximately normal.
Simulates poor field practice
2 samples prepared
G
Crimps over contaminated, corroded, or oxidized conductor strands
Field-aged lengths of cable harvested from scrap dumpster at ConEdison. Three samples prepared with both crimps properly positioned at normal depth. Corroded conductor strands subsequently confirmed. Partially recovered market tape confirms 1961 vintage
Simulates common field experience where new and aged cables are joined.
3 samples prepared
H
Contaminated conductor strands
Prepared cable ends were immersed into a soil-water slurry then dried before crimping
Simulates field condition where the prepared cable end may be dropped in a wet vault before crimping.
2 samples prepared
I,J,K
Categories reserved for additional experiments using same cable and connectors
N/A
N/A
L
Oblique conductor cut
Cable cut at oblique angle. Both crimps properly positioned, at normal depth.
Simulates poor field practice
2 samples prepared
M
Normal assembly
Both crimps properly positioned, at normal depth
2 samples prepared
N
Single crimp
Single normal depth crimp placed at mid-length of barrel half.
Simulates poor field practice. If made in this manner, insufficient space remains for second crimp.
2 samples prepared
O
Separator tape incompletely removed
Separator tape was only partially removed and left in place on one side only. Both crimps positioned normally, at normal depth.
Simulates possible field error where tape may be present on side of cable against vault wall and hidden from view
2 samples prepared
Figure 8:
Completed crimped connections
Test Set Up – Lenox Laboratory
Overview
EPRI will ’load age’ the various connector samples, by applying a
long-term continuously-repeated load cycle, to correlate temperature
with loading, and to better understand the connector temperature rise
kinetics, as related to the variety of assembly parameters and
pre-existing cable conditions addressed in Table 1.
To conduct these tests, EPRI connected the test samples in a series loop
with subsequent connection to a programmable current supply. This
current supply can be programmed to vary the loading to simulate utility
load cycles, and to vary the heating experienced by the test loop to
simulate and accelerate aging of the connections. In order to understand
the temperatures experienced in each of the connectors, EPRI applied
thermocouples to every connection, thus enabling continuous monitoring
and recording of the temperatures. EPRI developed data acquisition
programming to support recording temperatures from each of the 26 defect
connectors and two ambient temperature connectors at programmed
intervals. In addition, EPRI mounted an IR camera to periodically record
temperatures and IR images. This information will be used to correlate
outside temperatures as determined through an IR camera with internal
temperatures experienced in the connectors, as ascertained by the
thermocouples. Finally, EPRI will periodically record more detailed
imagery of each connector from a handheld, higher resolution IR camera
to better understand the specific heating characteristics of LV
connectors as they age.
Completing the Test Loop
With the support of Con Edison, EPRI had prepared 26 different samples
assembled with various crimp defects for study to be connected into a
test loop. To do so, each 3 ft. cable section was connected together to
create a series loop. Figure 9 illustrates the final test loop
configuration. The letters correspond to the Defect Code identified in
Table 1 above. The numbers refer to the specific sample. For example,
Connector A-1 refers to the first sample of defect Type A, the dual
under-crimped connection. Defect Code A-2 refers to the second sample
with defect Type A.
Figure 9:
Test loop schematic
Figure 10:
Test loop construction – in progress
To prepare the test loop, EPRI connected each of the three-foot cable
samples together. Each of these samples was prepared with the connector
on the one end of the sample, with one side of the connector barrel
crimped to the cable with the known defect and the other side of the
connector barrel open. On the other end of the sample, the cable
insulation was removed to expose bare conductor which would be joined to
the adjacent sample. For consistency, and to eliminate potential human
factor influences, EPRI joined the cable samples with a controlled
depth, concentric Burndy compression type crimping tool, using the
specific die identified for 500 MCM copper cable. This compression style
crimp differs from the lateral indentation style crimp used by Con
Edison. These differences are illustrated in Figure 12 and Figure 13.
Figure 11:
Burndy crimping tool and close-up view of die
Figure 12:
Lateral indentation style crimp, side view and view into the Barrel
Figure 13:
Concentric style crimp, side view and view into the barrel
Figure 14:
Illustration of test sample components
Figure 15:
Joining two samples together
Figure 16:
Test sample joined to linking cable section
Figure 17 shows the test loop under construction. This photograph
shows the assembled loop segments prior to thermocouple installation,
heating of the shrink tubes, and the installation of heat sinks.
Figure 17:
Test loop construction – in progress
To minimize the transfer of heat from one connector to another, EPRI
installed split type heat sinks midway between each of the connectors.
EPRI utilized split zinc anodes, used for marine propeller shafts, as
heat sinks[^1]. To apply these, a section of the cable insulation was
removed approximately midway between connectors and a piece of copper
braid was used between the heat sinks and the conductor itself to ensure
a tight connection and optimized heat transfer.
Figure 18:
Cable with insulation removed to apply heat sink
Figure 19:
Zinc Shaft Anodes used as heat sinks
Figure 20:
Test loop construction – in progress (Note installation of heat sinks)
Baseline Resistance Measurements
The electrical resistance through the connector-cable sample
combinations was measured at elevated current using a precision digital
low voltage ohmmeter (DLRO), manufactured by Megger Corporation. A 200 A
current was applied for all of these measurements. Three sequential
measurements were made for every sample tested. The electrical contacts
were rotated in place following each measurement in an effort to reduce
the effects of surface resistance due to oxide films or contamination,
for example. The measured values are presented in Table 2. Initial
measurements were made through a crimp connector and through a 3 ft (91
cm) length of the Con Edison cable. Measurements through a crimp barrel
indicated resistance values of 5.3 µΩ, 4.9 µΩ, and 5.4 µΩ, for an
average of 5.2 µΩ. Resistance measurements through the cable section
indicated values of 30.4 µΩ, 29.7 µΩ, and 28.4 µΩ, for an average of
29.5 µΩ. The latter values were considered higher than expected,
possibly due to a surface film on the conductor strands. This is not an
uncommon occurrence due to the presence of lubricant residues used for
drawing the copper wire through a die, as well as waxy processing aids
used in the rubber extrusion process. Since the surface of these cables
is representative of field conditions, the measurements were
subsequently made though assembled cable and connectors. Plastic
deformation during the crimping process is intended to displace
conductor surface films, including oxides and lubricants.
For reference, the resistance measured through a 3 ft (91 cm) length of
the same cable, with Burndy crimp connections applied to each end was
found to be 55.3 µΩ, 55.2 µΩ, and 55.1 µΩ, for an average value of 55.2
µΩ. The corresponding values obtained from crimps applied to each end of
an identical cable, using a ’normal’ Con Edison crimp connection are as
follows: 43.3 µΩ, 43.2 µΩ, and 43.2 µΩ, for an average value of 43.2 µΩ.
The Con Edison crimping practice appears to exhibit lower interfacial
resistance, however the crimping method does not engage all conductors
in a uniform manner, as shown when Figure 12 and Figure 13 are
compared. The concentric crimp is expected to exhibit greater
comparative ampacity as a result of having full engagement of the entire
conductor cross-section without gaps between adjacent conductor strands.
The lateral crimping method radially displaces the conductors that do
not lie immediately below the crimp impression.
Table 2:
Measured resistance through test connections
Connection ID
Res. Test 1, µΩ
Res, Test 2, µΩ
Res, Test 3, µΩ
Res. Ave., µΩ
A-1
63.2
62.9
63.0
63.0
B-1
61.4
61.3
61.3
61.3
B-2
60.9
60.9
60.8
60.9
C-1
62.0
62.0
62.0
62.0
C-2
64.9
64.8
64.8
64.8
D-2
61.6
61.7
61.6
61.6
E-1
62.3
61.8
62.0
62.0
E-2
63.3
63.2
63.2
63.2
F-1
63.4
63.4
63.4
63.4
F-2
63.9
63.8
63.8
63.8
G-2
62.9
63.1
63.1
63.0
G-3
81.8
81.8
81.7
81.8
H-1
61.2
61.4
61.2
61.2
H-2
63.5
63.1
63.5
63.4
L-1
58.8
58.7
58.7
58.7
L-2
57.2
57.1
57.1
57.1
M-1
62.1
62.3
62.5
62.3
M-2
62.1
62.1
62.1
62.1
N-1
62.9
62.9
63.0
62.9
N-2
62.3
62.3
62.1
62.3
O-1
69.8
69.8
69.7
69.8
O-2
65.7
65.8
65.8
65.8
Final Assembly
Upon completion of the joining of the cable samples and installation of
the heat sinks, the test loop was connected to the current supply, a
Magna-Power Electronics TS Series IV programmable DC power supply. This
current supply can be programmed to vary the current in the loop to
simulate load cycling. The current supply control circuit adjusts the
voltage as needed to maintain the desired current in the test loop.
Figure 21:
Magna-Power Electronics DC Power Supply
Figure 22:
Test loop construction – in progress. Connection to current supply
To complete the test loop, the thermocouples were installed, then the
heat-shrinkable tubes were installed and ‘recovered’ over the
connections. One thermocouple was applied to each of the crimp
connectors, on the back side of the barrel where the indent style crimps
were applied. Two additional thermocouples were added to the test
facility to provide ambient temperature readings, one in the front of
the test room, and the other at the rear. The thermocouples were
terminated into a data acquisition block, which, through programming,
enables EPRI to monitor and record temperatures from each.
Figure 23:
Installation of thermocouple on connector prior to application of heat-shrink tube
Figure 24:
Data acquisition terminal block
Figure 25:
Test loop construction – nearing completion
In order to obtain periodic infrared images of the test loop, EPRI
permanently mounted an IR camera on the ceiling of the test facility.
This camera is tied to data acquisition software which enables detailed
recording of infrared images of the test loop on a predefined period. In
addition, EPRI plans to supplement this imagery with more detailed
thermal imaging of each connector using a higher resolution IR camera.
At the outset, EPRI plans to capture images from the permanently mounted
camera every 15 minutes and prepare more detailed images of each
connector in the test loop at weekly intervals.
Figure 26:
Positioning the permanently mounted IR camera
Figure 27:
Test loop construction-final layout
Data Acquisition / Test Approach
The load cycle used for testing will vary the loading between a higher
and a lower current level to replicate cyclical heating and cooling of
the connections as would be experienced in the field.
In total, seven load profiles were utilized over the course of 2+ years
of testing, to attempt to heat the connectors, and bring about internal
changes that would yield thermal runaway. These profiles are shown
graphically below:
Figure 28:
Load profiles used to age connectors
Over the course of the testing the load cycle was gradually made more
severe to encourage connector The use of each cycle over the 1,271
cycles completed over the course of the testing.
Figure 29:
Load profile used for each connector aging cycle (total cycles = 1,271)
EPRI controlled the current cycling through software mounted on a laptop
computer and connected to the current supply. Temperature information
from the thermocouples attached to the connectors was recorded through
the data acquisition software at 1-minute intervals.
Images from the ceiling-mounted IR camera were also captured every 15
minutes and stored on the computer hard drive and were later used in the
analysis.
4 - LV Connector Findings
Current findings on LV connectors
Findings
The installation defects included in this testing were expected to
reveal performance differences in terms of thermal behavior during
thermal cycling. These defects may be classified into a small number of
groups for analysis as follows:
The analysis on the thermal data was performed considering each
connector’s relative performance to the reference connector group over
all completed cycles.
Reference Connectors
Within the full loop, two reference connectors were included that were
used as performance benchmarks for assessing the population of
connectors with installation errors. The progressive increase in
operating temperature shown below occured as a result of the increasing
“severity” of the load cycle. Averaging these two connectors leads to
the “reference” connector that is used for the analysis
Reference connector maximum temperatures (M-1, M-2, and average) for all
load cycle profiles
Connectors with Improper Crimping Related Defects
Six connector pairs were prepared with defects related to improper
crimping. Such defects include crimp location, crimp depth, and number
of crimps.
The performance (maximum temperature during each cycle) of this group of
connectors relative to the “reference” connector over the completed
cycles appears below. To interpret this figure, the zero line denotes
when connectors match the temperature of the “reference” connector.
Temperatures that are above or below the zero line indicate the
connector ran hotter (greater than zero) or cooler (less than zero) than
the “reference” connector. The red regression lines are used to provide
a visual indicator of the overall trend in the data – increasing,
decreasing, or stable with respect to temperature. The slope of the
regression line may be used to quantify the performance gap between
these connectors and the “reference” connectors. Positive slope
indicates worsening performance while negative slope implies improving
thermal performance (more likely stable performance versus worsening
performance of the “reference” connector).
Relative thermal performance of connectors with improper
crimping-related installation defects
Taking the improperly crimped connector class as a whole, this group of
connectors consistently remain more or less within 5°C of the
“reference” connector. There are occasional thermal excursions which
occur during individual cycles, but the relatively narrow temperature
bandwidth is nonetheless surprising given the types and severity of the
defects present.
Connectors with Contaminated Conductor Defects
This class of connectors includes a number of commonly occurring issues
with contamination on the conductor. These issues include corrosion,
dirt, and cable components not fully removed during stripping / prep.
Using the “reference” connector comparison approach, the resulting
thermal behavior for these connectors appears below. Note that the scale
is substantially increased due to the elevated temperatures these
connectors achieved over the “reference” connectors. Connector G-3
experiences over 20°C increase in temperature over the “reference”
connector. This is the largest observed temperature difference to date
in this research.
Relative thermal performance of contaminated conductor-related
installation defects
Connectors with Conductor Insertion Defects
This class of connectors is focused on understanding the performance
difference when there are issues with conductor insertion. These issues
include not fully inserting the conductor into the connector that could
result from an improper cable cutback as well as not inserting all
conductor strands into the connector. In the latter case, two strands
were intentionally bent 90 degrees away from the conductor resulting in
a loss of conductor cross-sectional area. This conductor is specified as
using 61 strands and so the loss of two strands represents a loss of
3.2% in cross section. The relative thermal behavior for these
connectors is shown below:
Relative thermal performance of conductor insertion-related installation
defects
The regression lines indicate some degradation in performance in three
of four samples. These differences are larger than the improperly
crimped connector group but smaller than those observed in the
contaminated conductor samples. In two of the samples, the temperature
remained consistently below the “reference” connector.
Effect of Other Installation Defects
A final installation defect was tested during this program – an oblique
insulation cutback that did not prevent full insertion of the conductor
within the connector. The relative thermal behavior for these connectors
is shown below. As with the conductor insertion samples, the performance
was mixed but relatively stable as indicated by the slightly positive
sloping regression line. It is interesting to note that sample L-2 ran
considerably cooler than the “reference” connector throughout the
cycling. It is possible that this connector’s position near the entry
door contributed to this behavior.
Relative thermal performance of oblique conductor cutback installation
defect
Regression Line Comparison
For each defect class, the individual samples were fitted with linear
regression lines to provide a visual and quantitative metric for
evaluating the performance over the course of the cycling. The slopes of
these regression lines may be considered as overall indicators of the
additional heating these mis-installed connectors experienced. In
addition to providing a consistent reference for the multiple load cycle
profiles, this also helps account for any environmental conditions that
occurred within the test area as a result of ambient temperature
fluctuations. Using this approach, it is reasonable to assume that the
temperature differences are primarily due to connector performance. An
analysis of the resulting regression fit slopes reveals interesting
behavior in this connector population.
The recorded slopes for each defect class can be displayed in boxplot
format – the black dots (●) represent the recorded slopes for the
individual connector samples within the same defect class while the
boxes show the spread in the slope data by encompassing the
25th to 75th percentiles (middle 50%) of the
recorded slope data. Boxes which appear small in height indicate the
slope data are highly consistent within the class and non-overlapping
boxes, in terms of the linear fit slope, imply the differences are
likely statistically significant. Of the four defect classes, only the
contaminated conductor class appears significantly different from the
rest of the sample population with a median slope of 0.04°C/cycle
(approximately 50 times greater than the second highest median slope).
Comparison of regression slopes for each defect class (boxes represent
middle 50% of data while the black dots (●) represent individual
connector slopes)
The remaining defect classes (conductor insertion, improper crimping,
and oblique cable cutback) do not appear to generate connector
performance differences at least within the testing completed during
this program.
Summary
The findings can be summarized as:
No connector experienced thermal runaway or altered its behavior
significantly from established analytical models
Analysis indicates that defects involving a contaminated conductor
are more likely to lead to elevated temperatures during operation
IR imaging is not as sensitive to the changes in connector thermal
performance when connectors are covered by an insulating sleeve
Identifying cycle maximum temperatures is not necessarily required
for using IR information when making connector-to-connector
comparisons – individual images do provide meaningful comparisons
5 - Future Research
What are our next options?
This multiple year research project sought to better understand how
field practices used to prepare low voltage (LV) connectors can impact
the relationship between loading and temperature over time and
contribute to connector aging and heating. Results from this research
are aimed at informing both field connector preparation practices and
methods for leveraging infrared thermography as a diagnostic tool in LV
systems.
In 2020, EPRI completed and began applying load to a LV connector test
set up comprised of connectors with various build defects. EPRI varied
the loading to try to understand the relationship between loading and
heating, and to bring about internal changes to the connectors to
simulate conditions found in field aged components. As described the
2020 update report, analysis of the data revealed that the connectors
under study performed normally and predictably, without indications of
thermal runaway or other anomalies. This suggested that the load cycle
utilized may not be sufficiently aggressive for effecting internal
changes to the connector / cable system, where temperatures can ‘run
away’ from normal expectations due to increased resistance or decreased
ampacity.
In 2021, EPRI revised the loading profile to increase the connector
maximum temperature excursion to attempt to bring about run-away
heating, or until such time as a failure or incipient failure causes
test interruption. No such behavior was observed during the additional
loading profile that was utilized. The testing did reveal some
performance differences between defect classes. In addition to the LV
connector loop aging, further research was performed to understand the
composition of the connectors used in the testing as well as forensic
analyses of failed or overheating components from field.
In 2022, EPRI completed its study of the specific low voltage connectors under test since 2020. Research included subjecting the components to high current and associated heating in an attempt to bring about premature failure. EPRI performed analysis of data from the testing by defect classes under examination, including defects associated with conductor insertion, contaminated conductor, improper crimping, and an oblique conductor cut. At the conclusion of the testing, the connector samples were dissected and analyzed revealing evidence of thermal damage to the underlying copper conductors that could indicate the start of failure due to thermal runaway. Results from these analyses can be found in report 3002024696.
Also in 2022, EPRI began the design and construction of a test bed for examining heating in medium voltage cable joints with a focus on producing results that can inform the use of infrared thermography as an underground cable systems diagnostic tool.
For 2023, EPRI will complete the construction of the test bed and begin the MV cable joint study. The study will continue into 2024.
Other potential future research avenues under consideration include:
Comparison of different crimping approaches, including radial and concentric crimping, to understand their effectiveness, consistency, and anticipated long term performance.
Expanded forensics analyses that would include failed or overheating
shear bolt connectors.
Examination of the implications of the use of fire proof tapes on UG cable system infrared thermography.
Expansion of the UG test bed to examine heating and IR in separable connectors.
Expansion of the UG test bed to continue examination of heating in LV components.
Standard Ampacity Ratings for 600 Volt A.C. Mains Cables Installed
Underground in Ducts, Consolidated Edison Company of New York, Inc,
New York, NY: Dec 2016. Specification EO-6040, Rev 10.
Infrared Imaging on Secondary Distribution Cables and Splices,
Consolidated Edison Company of New York, Inc, New York, NY:
May 2018. Specification EO-10370, Rev 0.