!2005 Snell Infrared 800-636-9820 www.snellinfrared.

com
Best Practices for using Infrared Thermography for
Condition Monitoring of Oil-filled Utility Assets
John Snell
Snell Infrared
PO Box 6
Montpelier, VT 05601-0006
800-636-9820
jsnell@snellinfrared.com
www.snellinfrared.com


Abstract:
Infrared thermography is being used more and more often as a tool to monitor the condition of utility
assets, including transformers and other oil-filled devices. When conditions are right and the
thermographer is qualified, the results are remarkable: locating problems and potential problems long
before they fail. This enables system owners to manage further diagnosis or repairs in a timely, cost-
effective manner. This paper will discuss the current best practices for using thermography for
condition monitoring of utility assets, guidelines for success and an overview of mistakes that are
commonly made.

Keywords:
ASNT, condition monitoring electrical, utilities, emissivity, infrared, thermography, transformers

Introduction:
We have all grown dependent on power being delivered reliably and
continuously whenever we need or want it. Many utility system
assets are both aging and being loaded past anything previously
imagined. Replacements, especially transformers, are months or
years in construction and delivery. The cost of a failure in the grid
can, as a result, have costly, even devastating, consequences.

Utilities have long practiced preventive maintenance (PM), but
economic pressure on all line items in a budget has not spared
maintenance where cost reductions have been considerable. The use
of condition monitoring technologies, such as dissolved gas-oil
analysis (DGA), airborne ultrasound, and infrared thermography,
have not only helped reduce unnecessary PMs, but have also offered
greater assurance that assets are performing as they should.

What is infrared thermography?
This remarkable technology utilizes electronic imaging cameras that
detect infrared radiation in much the same way that conventional
camcorders see visible light. Both detect forms of electromagnetic
energy that are radiated by all objects above absolute zero. Once
radiated, that energy is also reflected by many types of surfaces and
!2005 Snell Infrared 800-636-9820 www.snellinfrared.com
transmitted by a few. Infrared radiation is defined as waves that are 2-14 micrometers in length and,
thus, is longer than visible light. As surfaces increase in temperature they emit more infrared radiation.
Surfaces emit with different efficiencies, termed their emissivity; unpainted metals are inefficient and
have low emissivity values. In addition, low emissivity surfaces are typically highly reflective of the
radiant energy of their surroundings.

Special lenses focus the infrared energy on a multi-element detector array producing an electrical
response that is converted into a visible image, or thermogram, portraying the thermal patterns. Todays
cameras can see temperature difference as small as 0.5C on a high-emissivity surface at 30C. Some
infrared cameras are specially designed and calibrated so that detected radiation can be converted into a
radiometric temperature measurement. When properly used, these systems can consistently yield non-
contact measurements accurate to 2 C or 2% of the measurement range. Unfortunately, it is
extremely difficult to make accurate, repeatable measurements of unpainted metal surfaces under field
conditions.

Condition monitoring of utility assets:
In order for thermal images of utility assets to have diagnostic value, it is necessary to (1) know how the
asset is constructed and function, (2) to understand how it fails and what the thermal signature of that
failure is, and (3) to understand how system and ambient conditions will affect the thermal signature.

The vast majority of failures in utility assets result from either (1) abnormal high-resistance heating at a
point of electrical contact or (2) overheating of the asset as a whole due to a failure of the cooling
system. When conditions are right, a thermal signature precedes these failures. The value of
thermography, then, is being able to locate these problems in advance of failure. While the surface
temperature is important, many influences affect surface temperature other than the severity of the
problem. Once a suspect component or problem area is located, additional tests may prove more
valuable to identify the exact nature or the condition or to determine how advanced it is.

We will look at what assets can be monitored with thermography and at the conditions required to do so
successfully. We will also discuss the mistakes that are most commonly made in the process.

What to inspect:
Utility assets are a diverse group of complex devices. For the purposes of discussion we can look at
these more generically and divide them into several groups that include oil-filled transformers, oil-filled
devices, surge protection, disconnects, lines and connectors. More and more assets are also gas-
insulated; this creates unique difficulties for thermographers that are beyond the discussion of this paper.

Oil-filled transformers produce heat in normal operation. Failure to disperse this heat adequately results
in premature failure. Failures can also occur when connections points themselves overheat due to
abnormal high-resistance heating. When these are internal to the transformer, they can be very difficult
or impossible to detect; DGA would be a far more diagnostic tool.

Transformer tank: Baseline thermal signatures of all sides of the tank may have some limited value
for trending changes over time. Two problems exist: first, changes to the thermal signature that are
indicative will often be extremely subtle because of the massive nature of the tank. Second, it is difficult
!2005 Snell Infrared 800-636-9820 www.snellinfrared.com
to see all surfaces on most transformers and those that can be seen are subject to considerable influences
in heat transfer that are difficult to characterize.

That said, it is still typically a simple matter to gather and store baseline images of each side of the
transformer under known ambient conditions and at maximum loading. These can be compared to
updated images on an annual basis or as needed. While there may be some merit in comparing the
signatures of similar transformers, transformers more often than not have unique signatures.

Cooling systems: Inspection of cooling systems can have
immense returns, even if the simple technique is often
unappreciated and undervalued. A normal pattern for a
convective system is warmer fluid at the top and cooler at the
bottom with an even thermal gradient between the two. A set
of thermal images of the cooling system can clearly show
anomalous patterns that may not be indicated by an oil-
temperature gauge. The most common anomalies show cool
tubes related to lack of oil circulation; this is most often
caused by a low oil condition, either normal or abnormal.
Lack of circulation has also been associated with out-of-level
pads, inverted riser tubes, closed valves, and blockage in the
tubes or headers. Reduced oil circulation during summer peak
conditions will result in a transformer that overheats and that
can dramatically reduce its life cycle.

Cooling fans and oil circulation pumps, if present, can also be inspected. Fans in particular are often
neglected in routine PM. Both should be place into operation for approximately fifteen minutes prior to
an inspection. Fans can show several signatures; when operating normally, the fan motor will appear
warm. When the motor has failed, it will be cold by comparison. Fan and pump motors that are seized or
that have failing bearings (or pump seals) will appear warmer than normal. It may even be possible to
locate fans or pumps that are running backwards or that are valved off. Fans and pumps represent the
Achilles Heel of the transformer: they are designed to operate only when worst-case conditions exist
so that, if they fail, serious repercussions, i.e. heat-related damage or failures, are almost guaranteed.

Cooling system inspections should be scheduled prior to summer peak and, as necessary, during summer
peak to insure all is operating as designed. Inspections after any repairs or maintenance work can
document running conditions and minimize problems such as inadvertently leaving a valve closed or a
fan running backwards.

Bushings: Both high- and low-side bushings can be profitably inspected
with thermography. When under load, anomalous conditions can be found on
the line-side connections, as well as internal connections, both in the bushing
head and in the connections to the coils. A direct view of the line-side
connections is necessary and simple comparisons among the phases should
be made. A normal signature is at or close to ambient air temperature. Any
unexplained, anomalous thermal signature or rise over ambient/another phase
is considered indicative and warrants further investigation.
!2005 Snell Infrared 800-636-9820 www.snellinfrared.com

Internal faults will be indicated by much more subtle signatures due to the large thermal gradient that
exists between the connection itself and the observed surfaces. While a fairly direct view of the bushing
head is usually possible, it is not possible to see the internal connections to the coil directly; here a
signature, when warm enough, will show up at the base of the bushing as heat is conducted up the
bushing stud itself from the coil connection. Even to be able to gain a view of this part of the
transformer, especially large ones, while standing on the ground can be challenging.

Tap changers: Most, but not all, tap changer compartments run at the same temperature as the tank or
slightly cooler. Abnormally overheated tap connections under oil will often generate enough heat to
cause the entire changer compartment to overheat. A tap changer that appears warmer than the tank,
unless normal by design, is cause for further investigation. Unfortunately, the faulty tap may not be
energized at the time of the inspection finding a problem is not assured. Regular DGA is still of
immense value.

Connectors: A normally operating connector will typically operate at the same temperature as the line
or slightly cooler. Connectors with abnormal resistance will heat up compared to those that are normal.
The often-massive connectors used on many utility assets may dissipate heat effectively enough that the
thermal signature will be difficult to detect until damage is significant. Evidence of this is readily seen
by damage to internal contact surfaces while external surfaces appear fine.

Failing connections on the grounding system may also have detectable thermal signatures. Whether it is
the connection point itself or a failing ground mat itself, current flow in such situations will often
produce enough heat to be seen in the image.

Surge protection: Surge protection in and around a substation can vary widely by design. Most
commonly it is a high-resistance path connected to ground. Normally surge protection will operate at
ambient air temperatures because there is no current flow. A breakdown of the resister will result in a
small leakage of current to ground at all times, even when a surge is not present; this, in turn, will cause
the arrestor to heat continuously, typically in small sections defined by its structure and at only a few
degrees above ambient. Despite this seemingly insignificant signature, arrestor failure may be imminent;
because failure is often catastrophic and dangerous to any nearby personnel, these early warnings should
not be ignored.

Insulators: Normally, insulators operate at or near ambient
air temperatures. The only other normal thermal signature
shown by an insulator is one caused by solar absorption; here
darker insulators will often be warmer than lighter colored
ones. Cracked or dirty insulators may indicate a subtle
thermal signature due to the slight current leakage that
occurs over the resulting high-resistance pathway to ground.
These signatures can change or disappear with changes in
ambient conditions or as the insulator is washed by rain, so
the lack of a signature is not positive proof that a problem
might not exist at another time.

*> 108 .1F
*< 6 5 . 5 F
70. 0
75. 0
80. 0
85. 0
90. 0
95. 0
100. 0
105. 0
!2005 Snell Infrared 800-636-9820 www.snellinfrared.com
Associated switches and disconnects: High-resistance heating problems are common in all types of
switching and disconnect mechanisms. A normal device will operate at or near ambient air temperatures.
A simple comparison over the device and/or among phases will quickly reveal problems. It must be
noted, however, that in some cases where a device can be feed from two sides, such as a yoke
disconnect, the cooler side may represent the actual problem. In such cases one side may be warm
because current is shunted to that side, away from a high resistance point of contact on the other side.
Often a visual inspection or resistance test will reveal this, but one should not jump to the conclusion
that the hot spot is the problem.

Optimally, all switch/disconnect devices should be inspected after they are put back into operation. Even
slight heating (approximately 200F) over time can result in annealing damage that will be exacerbated
quickly into a device that has significant damage to contact surfaces or, in the worst case, has welded
itself together.

Conditions for inspection:
For the most part thermography cameras see surfaces; the radiant energy seen is a combination of both
energy emitted by the surfacethis tells us something about its temperatureand energy that is
reflected by the surface, which tells us nothing about its temperature. Unpainted metals have a low
efficiency of emission while at the same time having a high reflectivity. This means they dont reveal
much about their true thermal nature and mask over the little information that is there with other
information. Ideally we can correct the readings in radiometric infrared cameras for both emissivity and
background reflected temperature; unfortunately, and this despite what the literature and the suppliers
would have us believe, for all but the most corroded of metals, measurement error is unacceptably large.

The reality of looking at unpainted metal surfaces in most utility assets is that we will have someeven
if subtleindication of abnormal heating if other conditions are conducive. Every effort must be made,
then, to have the right conditions. The accuracy and repeatability of radiometric measurements is also
very low, even if temperature differences (phase to phase or rise over ambient) are measured rather than
actual temperatures. When the emissivity of a surface is below approximately 0.6, measurements, even
if corrected, are not recommended. This flies in the face of what most of the industry considers a best
practice.

Given this, we must recognize that even small temperature increases may indicate severe heating. When
conditions change, the signature may fall below the threshold of detection. What changes are
significant?

System load: As load increases heat output rises at the square of that increase; a tripling load, not
uncommon in a distribution substation, will result in a nine-fold increase of heat being produced. If the
system is inspected when loads are light, some anomalies will have thermal signatures below the
threshold of detection and others, though detectable, will be cooler than they will be when loads increase
at later, peak periods.

While some standards suggest that 40% of design load is an acceptable condition, an analysis of
potential fault current suggests it is the bare minimum and, especially when loads will greatly increase
beyond that, inspections are better done under worst case conditions whenever possible. Some in the
industry feel it is possible to predict temperatures at a future load based on the present measurement; this
!2005 Snell Infrared 800-636-9820 www.snellinfrared.com
methodology is highly unreliable and not recommended except in limited circumstances with significant
input from knowledgeable heat transfer engineers. It is clear, regardless, that as loads will increase so
will deterioration of an anomalous asset.

Convection/wind: The most significant of the limitations to using infrared thermography is typically
convective cooling by the wind. Convective cooling results in two issues; first, connections that are
early in the failure cycle, i.e. only slightly warmer than normal, will often be cooled below the threshold
of detection. Second, anomalies that are found are being cooled, often significantly, and will heat up
when the wind lessens.

Many thermographers have used several commonly accepted rules
of thumb to attempt to correct for wind. Unfortunately, these simple
measures are inadequate. Low temperature problems will still not be
made visible, and the complexities of convective heat transfer do not
always lend themselves to simple corrections for understanding more
advanced problems.

Wind speeds as small as 1-5mph can have a significant cooling effect
on a thermal signature. Wind speeds above 5 mph can, depending on
other conditions, reduce the temperature difference between the
abnormal component and ambient to a few degrees or less.
Inspections when winds exceed 15 mph should be avoided; when
they must be conducted in these adverse conditions, attention should
be paid to any and all findings, as they will increase in temperature
when the wind speed is reduced. The two images of a set of OCBs
(left) were taken under similar loading and ambient air conditions; in
the top image the wind was blowing at 15mph while in the lower
image it had dropped to 2mph.

Classic heat transfer methodology may be useful to create a thermal model to explore the range of
temperatures expected with changes in convection. For instance IEEE Standard 738 does this for bare
overhead conductors by taking into account the appropriate thermodynamic and fluid equations for
taking wind speed into account, as well as natural convection and radiant cooling effects. It is not
possible, however, to apply this same model to all assets under the widely varying situations
encountered in the field.

Ambient conditions: The impact of changes in ambient may also be significant and difficult to
quantify. The primary factors, aside from wind, are changes in air temperature, the impact of solar
heating, and the presence of precipitation.

The exact impact of changing air temperatures is difficult to predict, Clearly, an increase in air
temperature will result in an increase in the measured temperature of a component. Thus, because
warmer summer air has less cooling capacity, a high resistance connection will become hotter during hot
months. A 70-100F difference in air temperatures between the summer and winter extremes is not
unusual. During cold weather, problems that are only slightly warm may be cooled below a point where
detection is reliable or even possible.
!2005 Snell Infrared 800-636-9820 www.snellinfrared.com

Solar heating of component, especially those with a high absorptivity of the suns energy (such as aged
conductors and dark insulators/bushings), will mask over small thermal differences. Late afternoon
inspections during the summer months are particularly problematic and should be avoided if possible.

Precipitation in any form, whether snow, rain or fog, will often result in evaporative coolingespecially
of the abnormally warm component where there is enough heat energy to power evaporation.
Evaporative cooling can result in temperatures falling below ambient air temperature. Again, anomalies
that are only slightly warm may be cooled below a point where they can be detected.

Other oil-filled assets:
Many of the same techniques and problems apply to other oil-filled utility assets, such as oil-filled
circuit breakers, voltage regulators, and re-closers. In these cases, however, heat is not normally
generated in the same degree as is in transformers. The tanks, thus, will run much closer to ambient air
temperature. When any unexplained deviation in tank temperatures among the three phases occurs, it is
cause for concern and, often, immediate action to determine the root cause. The reason for this is that the
abnormal heating, even if slight, is caused by a small high-resistance heat source (an internal
connection), that must be very hot in order to change the temperature of an entire tank of oil.

As with transformers, bushing connections can also exhibit heating either internally and externally, the
former being much more difficult to detect and, typically, more serious to rectify. DGA testing can be
very useful to better understand many internal faults that are located with thermography.

Common problems with many inspection methodologies:
Far too many of the inspections being conducted today fail to achieve best practices. The most common
issues are as follows:

A focus on radiometric temperature measurements: As has been shown, the surface temperature
represented in a thermal signature can vary widely with a number of variables, most of which are
difficult to characterize and quantify. Most practitioners fail to pay attention to the influences of ambient
and system conditions. As a result, many problems are not detected while others that are detected, are
misdiagnosed.

Basic thermodynamics: Even without an advanced degree in heat transfer, a simple understanding of
the relationship between the source of heating (most often internal to some degree) and the surface we
see in the thermal image (most often external) is essential. Many factors affect exactly what the surface
temperature will be at any given moment. The precise relationship between that moving target and the
internal heating source is often just as complex, especially when the thermal gradient is very large as is
the case with any oil-filled device.

A thermal image alone is not sufficient to make sense of the radiometric data. Ambient influences, such
as wind and air temperature, should be carefully measuredat the location of the anomalyand
recorded along with the thermal data. Similarly, all relevant system influences, such as load and circuit
configuration, should also be noted and recorded.

!2005 Snell Infrared 800-636-9820 www.snellinfrared.com
All this points to two things often ignored by thermographers. First, we must work with conditions such
that, if there is an anomaly, we will be able to see it. Less than optimum conditions will often result in
many important anomalies being at or just above the threshold of detection, i.e. with a very low rise over
ambient or similar assets. Second, the radiometric temperature of an anomaly is not a reliable indicator
of its progress toward failure.

Thermal gradient: Thermal energy seen by the system radiates from the components surface. The
heat of high resistance almost always is being generated at some point internal to the surface. There
exists a thermal gradient between the hottest spot inside the splice and the surface being viewed. This
gradient can be very large, on the order of hundreds of degrees. In fact, some splices are designed
specifically to dissipate heat. Some research has attempted to estimate what impact the thermal gradient
may have
4
; it would be useful to accurately model this specifically for typical splice components.

The relationship between temperature and electrical resistance: A widely unappreciated, yet
important, relationship exists between electrical resistance and temperature. As an abnormally high-
resistance component begins to heat up, resistance increases which, in turn, results in additional heating
and, as a consequence, in a further increase in temperature. In other words, a bad situation gets worse,
often quickly! An increase in component temperature of 10C results in a 4% increase in resistance. This
accelerating deterioration is one of the reasons why components often fail shortly after they are first
detected or when they finally reach a runaway stage.

Current shunting: Also commonly overlooked are situations
where a warm component or asset may be the normal one, while the
cooler one may be have a higher resistance. The most basic example
is the barber pole effect often seen on multi-strand conductors.
This occurs when several strands are broken or isolated electrically
by corrosion, the remainder of the strands must care the entire load
and, thus, appear warmer. Current is shunted away from cool
strands that, because of their higher resistance, do not function
properly. A similar situation can often be seen to exist on many
yoke-type disconnects (left).

On a larger, and much more significant scale, are any double fed circuits, such as a ring-bus, where high
resistance on one side will cause current to shunt to the other side of the circuit; this additional current,
in turn, will cause any high-resistance anomalies, even if only slightly so, to heat. If the circuit is opened
to repair these false positives, all current will flow through the side of the circuit with high resistance,
often with disastrous results. Extreme care must be given to analyzing any problems found on a double-
fed circuit and, due to the potential complexities and consequences of a failure, qualified engineering
support is probably more often warranted than not.

Emissivity and reflected background: A best practice for radiometric temperature measurement
suggests that none be made until an anomaly is located. Even then, attention must be given to low
emissivity surfaces as they will fail to emit strongly and may also offer confusing reflections. When a
suspect asset is located, a careful evaluation should be made of the emissivity of the surface(s) in
question. Emissivity correction value look-up tables can be helpful, but, in the end, the qualified
thermographer will need to adjust many of these values to fit reality. Many components will exhibit a
!2005 Snell Infrared 800-636-9820 www.snellinfrared.com
fairly high-degree of cavity radiation despite their otherwise low-emissivity characteristics. Once a
suitable value has been determined, an appropriate background reflected temperature compensation
value must also be determined. This is typically a fairly simple matter for a qualified thermographer,
even if not without pitfalls.

Two, additional important, but impossible questions: How hot is too hot? and How long will it
last? Unfortunately, little is know about the temperature/time relationship for failures caused by high-
resistance heating at a small contact surface. What we do know it that the process often proceeds slowly;
even if early heating does not appear severe, pitting and melting occurs at a micro-site inside the point of
contact. This, in turn, results in increase heating and resistance and oxidation. Annealing begins to occur
at fairly low temperatures (200F) over a fairly short period of time (30 days). Once the metals have lost
their temper, a run away condition, at which point more melting occurs, can result quickly.

It must be noted that the heating is concentrated on an area that is most often extremely small; this has
little to do with a temperature-based specification that typically accompanies all assets. The asset itself
will probably not be heated greatly, but temperatures at the localized, high-resistance hotspot will, given
time and the right conditions, certainly reach the melting point of the metal.

While it is not possible to answer either question, it can be guaranteed that high-resistance hot spots will
not get better without intervention, at least not permanently! In extreme circumstances heating may
cause re-welding of a hot connection that may, in turn, result in a temporary improvement and
reduction in resistance, heating and temperature. Clearly this condition would most often be considered
extremely dangerous and warrant immediate corrective action.

Limitations of the infrared system: While the utility asset and the surrounding environment present a
number of difficulties for thermographers, the infrared system itself has inherent limitations many of
which are poorly understood by thermographers. All optical systems, including infrared, have limits to
their abilities to resolve data. Thermographers must deal with both spatial and measurement resolution.

Spatial resolution is a function of the detector size and the optical
path of the system, including the lens. Longer lenses can improve
resolution, but result in a narrower field of view. Moving closer, if
an option, achieves the same results. Distance and object size are
the two variables. It is common that systems used for utility
inspections have spatial resolutions in the range of approximately
1.2mRad. This means that a typical component that is two inches
in diameter can be detected, if hot enough, from a maximum
distance of 138 feet. A one-inch component could be seen from
half that distance. The two images (left) clearly show how moving
closer allows us to see more detail; changing to a telephoto lens
would produce similar results.

The measurement resolution of an infrared system is less than its
spatial resolution, typically by a factor of between two and four. It
is, therefore, common to be able to see hot spots while still being
outside of the measurement resolution for the system.
!2005 Snell Infrared 800-636-9820 www.snellinfrared.com
Measurements taken be yond the limits of resolution typically appear less than they actually are because
the hot spot is averaged together with the cooler sky temperature that often makes up the rest of the field
of view. The apparent temperatures of problems are thus often underestimated. The only solution is to
work well within the limits of measurement resolution, either by being closer or using a lens or system
that has a more favorable resolution. Unfortunately most infrared system suppliers do not provide the
specification for measurement resolutions so it must be derived from practice.

One other system factor that should be considered is the waveband being detected. Most infrared
systems used for utility inspections detect radiation between 8-15 micrometers or the long wave band.
This band is relatively insensitive to false positives resulting from reflected solar energy or solar glint.
In the past systems that sensed in the midwave band, 2-5 micrometers, were common. Problems with
solar glint are much more common and this waveband is no longer widely used as a result.

Qualifying thermographers: Too many thermographers are not adequately qualified and the results
are, on the one hand, less than optimum returns on the investment or, on the other, dangerous errors and
omissions in their work.

Qualification of thermographers is best defined by the personnel qualification guidelines of the
American Society for Nondestructive Testing (ASNT) SNT-TC-1A. Qualification is based on training,
experience and testing. Clearly the skills required to perform high quality utility asset inspections are
considerable; some are general to the field of thermography while others require specialized knowledge
of the assets themselves.

Level I thermographers are qualified to gather data; certification entails a week of training and three
months experience, at a minimum. At Level II thermographers may also interpret the recorded data. To
be certified at Level II requires an additional week of training and up to nine months additional
experience. Training at both levels should follow the training outlines prescribed in SNT-TC-1A.
Performance-based testing, including a practical examination, should be done for general knowledge as
well as knowledge specific to utility applications.

Educate users of thermography services: Too often the expectations of users of thermography
inspection services are not based on the reality of heat transfer and radiation physics. Disappointment is
the only thing that can be assured! The popular press has done little to dispel the many industry myths
that contribute to this unfortunate situation.

A pre-inspection meeting can establish the requirements necessary for condition monitoring as well as
define the limits of the technology. If work must be done with less than ideal conditions, as is often the
case, expectations can be adjusted appropriately, and provisions made for completing the inspection at a
later date.

Unfortunately very few industry standards exist. ASTM E1934, Standard guide for examining electrical
and mechanical equipment with infrared thermography can provide some minimal guidance even if it is
not specific to utility assets; it is in the process of a major revision at this time. Some thermographers
use rules of thumb that masquerade as standards; some are even now incorporated into formal
standards despite the fact that they are not based in science. This situation is problematic and costly for
all. Until this changes, we can expect continued problems with achieving anticipated results.
!2005 Snell Infrared 800-636-9820 www.snellinfrared.com


A sample Best Practice
for the use of infrared thermography
for condition monitoring of utility assets

Inspections will be conducted with appropriate precaution, planning and attention to safely. If
circumstances change such that safety is an issue, the inspection will be curtailed immediately until
plans can be revised or circumstances change.

Inspections will be conducted at peak loading conditions with a wind no greater than 5-10 mph. Any and
all differences in temperature among phases or similar components are noted for follow-up to determine
the root cause and, as appropriate, any corrective actions that should be taken. When inspection must be
conducted with higher wind speeds, it must be recognized that some anomalies may not be detectable
and those detected may be hotter than they appear when winds are reduced.

Baseline inspections will be conducted for all newly installed assets. Whenever damage occurs, repairs
are performed or destructive testing done, follow-up monitoring will be done to assure the condition is
normal. Periodic monitoring of asset condition will be scheduled on a frequency based on risk and a
statistical analysis of current condition. If resources do not allow for optimum implementation, the
deviation shall be noted and overall maintenance investments adjusted to accommodate as required.

Thermographers will be qualified to use the infrared system and to understand the findings or work with
someone who can do so. Thermographers will work within their limits and have the support of a
qualified, knowledgeable engineering staff.

Radiometric temperatures will be measured when conditions allow, the correct emissivity and
background reflected temperature and working within the measurement resolution of the system being
used. If necessary, supplemental lenses will be employed to improve resolution. The IR system will be
checked regularly to ensure it is within calibration.

When possible, the emissivity of surfaces will be increased by applying high-emissivity targets, such as
a light colored paint or a grease, to improve measurement reliability. Temperatures will be reported in
three ways (1) corrected, (2) rise over similar component and (3) rise over ambient. Generally
measurements will be made using a small area measurement rather than a single spot.

All relevant system and environmental influences will be measured and noted, including current and
future loading, asset data, wind, ambient air temperature, precipitation, etc.

All components will be viewed directly if possible, and, even so, the thermal gradient of the component
will be taken into account when evaluating the thermal data. For components with no direct view, such
as an internal fault in any oil-filled device, any abnormal rise in temperature will suggest immediate
further diagnostics such as a visual inspection, airborne ultrasound or DGA testing.

All will agree that the primary purpose of an infrared inspection is to monitor the condition of all assets
and locate any that are not performing adequately. Thermography provides a snapshot picture in time;
!2005 Snell Infrared 800-636-9820 www.snellinfrared.com
while it may be possible to understand changes to thermal signatures, doing so accurately typically
requires a very sophisticated modeling and engineering study. Trending changes in temperature may
have limited value given that a number of variables, aside from degradation, can influence temperature.
Recommendations for repair will be based on all relevant factors rather than on temperature alone.
These may include such factors as criticality, history, current and future duty/load cycles, age, operating
environment, and other test data.

Conclusions:
Thermography is a powerful tool for monitoring the condition of many utility assets. When properly
used it can validate operating condition and help to locate and, in some cases, diagnose, anomalous
conditions. Thermography is often poorly used with the result being that detectable failures continue to
occur unnecessarily. Qualifying thermographers and educating customers will go a long way to
achieving success.

Author:
John Snell, president and founder of Snell Infrared, has been
teaching people to use this remarkable technology since 1983. He
was the first person in the world to receive an ASNT Level III
certificate in the thermal/infrared method and continues to be very
active professionally on numerous standards committees and at
conferences. To learn more about thermography and Snell Infrared
visit http://www.snellinfared.com or call 800-636-9820.