microsoft word - section 6

7/27/2019 Microsoft Word - Section 6

http://slidepdf.com/reader/full/microsoft-word-section-6 1/18

Level I Course Manual – Section 6Publ No 1 560 009 C

1

6. Measurement TechniquesWhat will you learn

• Atmospheric and environmental considerations • System considerations • Optimizing measurement factors and system performance

By having the camera look at very exact blackbody radiators we can establishthis relation for the camera, store that relation in the memory and use it for thecalculation of the temperature of the object. That relation has a mathematical

form, which is derived from Planck’s law.

Atmospheric transmission

Another important parameter is the transmission through the atmosphere. As hasbeen said before, the atmosphere absorbs a bit of the radiation from the object.That absorption depends on the distance, so the object distance is an importantparameter.

These pictures show a tube carrying a warm substance. The objective with theimage has been to search for heat leaks. However, we can also see that thetemperature is falling along the tube. This might depend on the distance. If we put incorrect distance values we will find that the original temperature difference of 0.8ºCin the left picture will be totally compensated for in the picture to the right. And asthe attenuation in the atmosphere depends on the distance but also on its relativehumidity, these two factors have to be considered for correct measurement.




2

Ambient temperature

Consider the picture above. The emissivity of the wall is set to 0.92, an appropriatevalue for the wall material. It shows a badly insulated house. The temperature of thesurroundings and of the atmosphere is set to 0ºC. Suppose now, that the ambienttemperature is incorrectly set to 20ºC. In the spot we measure 2.2ºC.

This drastic change can be explained. The camera always measures the totalradiation. We have set the parameter ‘ambient temperature’ to 20ºC. Hence, mostof the radiation is considered to be reflected in the object, and the rest, which in thiscase is very small, is considered to come from the wall, the calculated temperatureof which will now be –8.2ºC, instead of the correct +2.2ºC. The difference is quiterampant. So the correct setting of the ambient temperature is of great importance.

SP01*:




3

Optimising the infrared image

So far we have seen a number of infrared images, shown with various colour scalesor greyscales. Consider the images below.

The pictures show three fuses with different temperatures. In fact only the one tothe right is clearly visible. It is also the hottest. And the situation is the sameindependent of the colour scale, so-called iron scale to the left and rainbow to theright. Below two greyscales are displayed, the left with white hot, the right with blackhot.

The selection of preferred colour or grey scale is mainly a matter of taste. It is notpossible to objectively state that one scale is ‘better’ than the other. However, as amatter of fact, details are more easily distinguished in the greyscale, black hot,image than in the others. Probably in this particular case but not necessarily as arule.

40

60

40

60

40

60

40

60




4

Selecting the temperature range

The AGEMA 500 cameras can register thermal information from a large

temperature range, typically from –10º to +90ºC, before the detector starts beingsaturated. The images above show the range from 21.8º to 61ºC, because thereare no other temperatures present in the image. This will, however, only showdetails of the hottest fuse. We might also be interested in the other two fuses. Seethe two pictures below.

In the image to the left the colour scale is chosen so that it is distributed over atemperature range between 28.1º and 42.6ºC. By doing this the middle fuse will beenhanced. The hottest fuse, the one to the right, is now saturated, i.e. the greycolour indicates that those temperatures are higher than 42.6º, which is the upper

limit of the selected scale.

Similarly, in the picture to the right, the temperature scale has been set to enhancethe coldest of the three fuses. Consequently the other two fuses appear saturated.

28.1°

42.6°

30

35

40

24.2°C

31.8°C

25

30




5

The isotherm

‘Isotherm’ means ‘same temperature’. It is a measurement function, whichhighlights areas of the same

temperature. See the image above.

In the result table we can also see for top temperatures for ISO 01 and ISO 02, and

their widths. The ISO 01 highlights all temperatures between 8.3º and 5.1º, where5.1º is 8.3 – 3.2 = 5.1, i.e. the top temperature minus the width. This can of coursebe done with any colour or grey scale.

If we use a colour scale with only 10 colours, we will in fact get 10 isotherms in theimage. That scale makes it sometimes very easy to see the temperature distributionover an object. See the image below.

The temperature scale is now selected so, that each colour is an isotherm of thewidth of 50ºC. The white colour – the 11th colour – shows temperatures above 900º. Actually the maximum temperature in the image is 937ºC.

400

600

800




6

Measuring object emissivities

In most measurement cases we assume that we know the emissivity of the object –

and the other object parameters as well. We can then calculate the objecttemperature. If we work the measurement formula backwards, then, knowing theobject temperature we should be able to calculate its emissivity.

Hence, we have to create a situation where we know the temperature of the object,preferably the temperature should be the same for the whole object. It is absolutelynecessary that the object have another temperature than the surroundings. Usuallywe heat up the object to a temperature at least 10ºC above the surroundings. Elsethe physics do not work.

It is not so easy to give an object a perfectly even temperature. But it has to bedone when exact emissivities are wanted. We can then use a so-called equalization

box. See the photos below.

This box heats the object by means of circulating hot air, the temperature of whichis very well controlled. Once the object has reached the set temperature, the lid isdrawn off, and a thermogram is captured of the object. The camera or the softwarefor processing thermograms can be used to get an emissivity value.

Another method, which is much simpler, but still gives reasonably exact values of the emissivity, uses a known emissivity. The idea is to determine the temperature of the object the normal way. By adjusting the object so, that an area with the

unknown emissivity is very close to an area of known emissivity, so close that it canbe assumed

that they have the same temperature, the unknown emissivity can be calculated.See the picture on next page.




7

There are two areas put into the image. The average temperature in the areas hasbeen selected. They are close to each other. One is on the motor cover, which hasa high emissivity, 0.93. The other area is placed on a metal strip, probably withsome text on it. Let us assume that it has the same temperature as the cover justbeside it. By adapting the emissivity value of AR01 we find that it shows the sametemperature as AR02, our reference area, if the emissivity is selected to 0.64.

35.1°C

89.8°C

40

60

80

AR01*:

AR02*:




8

Use Low-Cost Materials to Increase Target Emissivity

Clean, unoxidized, bare metal surfaces have quite low emissivity. So low theyare difficult to measure with an infrared camera. To get good measurements,we need to increase the emissivity of these problematic targets. In predictivemaintenance (PdM) applications, there are many low emissivity targets,especially for electrical applications. The purpose of this tech-note is to givePdM thermographers some guidelines as what they can do to improvemeasurement capability for low emissivity targets.

As many of these targets are often electrically energized, one must alwaysmaintain good safety practices. Coat the targets only when they are notenergized. Use only an approved coating to ensure proper operation when

energized.

Most high quality electrical tape has an emissivity of 0.95. One must be careful,especially with short wavelength cameras (3-5 m), that the tape is opaque.Some vinyl tapes are thin enough to have some infrared transmittance, and areunacceptable for use as high emissivity coatings. Scotch Brand 88 blackvinyl electrical tape has an emissivity of 0.95 in both the short wavelength (3-5 m) and long wavelength (8-12 m) regions, and is recommended. Theexample below shows two cans with tape. The one on the left is filled with hotwater, the other is at ambient temperature. For the hot can, the temperature

read from the tape is 163 F, from the can is 74.3 F. The latter reading isessentially ambient temperature as the emissivity of the can is quite low. This isa classic example of the necessity of using a high emissivity application on alow emissivity target.

Most paints have an emissivity of about 0.9 to 0.95. Metallic based paints havelow emissivity and are not recommended. The color of the paint is not thesignificant variable in its infrared emissivity. The flatness of the paint is moreimportant than its color. Flat paints are preferred over glossy paints. Thicknessof the paint also plays a role. The coating must be thick enough to be opaque.Two coats usually suffice.

69.4°F

172.3°F

80

100

120

140

160

SP01: 163.1°F

SP02: 74.3°F




9

Tape is good for small areas. Paint is good for larger areas, but is a permanentcoating. For coatings over large areas that need to be removed, or where tape

is inappropriate, powders suspended in a slurry or spray form can work well.Dye penetrant developer and Dr. Scholl’s spray foot powder are twoexamples. Emissivities of these powders are in the 0.9 to 0.95 range providingthey are applied thick enough to be opaque. Other materials such as Aquadag have high emissivity, around 0.9.

Be sure you make the coating cover an area of sufficient size. Know your camera’s spot size ratio for measurement, and the minimum operating distanceyou can safely employ. For example, a camera with a 250:1 spot size ratio canmeasure a one-inch target at a maximum of 250 inches, or 20.8 feet.

For higher temperature applications, use a high temperature paint such asengine or charcoal grill paint. Tapes and powders are limited in temperaturerange of application. For electrical systems, if the tape is melting, the problemis probably significant. So, you should not need a high temperature material for this application.

For most predictive maintenance applications, you don’t have to be overlyconcerned whether the emissivity of your dielectric coating is 0.9 or 0.95.Errors introduced by this uncertainty will be small relative to all the other measurement uncertainties you will encounter, such as backgroundtemperature, directness of the reading and so on.

Most dielectric (nonmetal) materials have emissivities near 0.9. Most polishedmetals have emissivities near 0.05 to 0.1. Tarnished, oxidized or otherwisecorroded metals have emissivities ranging from 0.3 to 0.9 depending on theamount of oxidation or corrosion. These are the difficult targets. Dielectrics arestraightforward: they have high emissivity. Metals are straightforward: youcannot make the measurement without improving the emissivity with a surfacecoating. Other difficult targets are semi-transparent materials, a subject beyondthe scope of this article.

For those cases where accurate values of emissivity are desired, or where you just don’t know the emissivity as in the case of oxidized metals, you canmeasure the bandpass emissivity with your infrared camera. The InfraredTraining Center teaches emissivity measurement in its predictive maintenancecourses. For highly accurate measurements you can send the sample to alaboratory for spectral measurement.

Table values of emissivity must be used with caution. Often it is not clear over what wavelength band the emissivity value is valid. And emissivities do changewith wavelength. Also surface condition, texture and shape play key roles in theemissivity of a material.




10

Here is one way to understand the effect of emissivity uncertainty onmeasurement accuracy: Suppose the uncertainty in target emissivity is ±0.05.

For an emissivity of 0.95, this represents about a 5% error (0.05/0.95). For amaterial such as shiny copper, emissivity 0.05, this represents a 100% error (0.05/0.05). These errors propagate into the temperature calculation,increasing the error in temperature reading. Our recommendation is to notattempt temperature measurement for target emissivities below about 0.5because of this effect. Coat the target with a high emissivity material.




11

Environmental Influences on IR Thermography SurveysBy

Robert MaddingAnd

Bernard R. Lyon, Jr.Infrared Training Center

North Billerica, MA

Abstract

Do I have to worry about environmental effects such as sun and wind whendoing an IR thermography survey? This is a recurring question bythermographers at all levels. The answer can be as complex as the

environment. Factors such as survey severity criteria guidelines, whether themeasurement is direct or indirect, equipment type and load and the severity of the environmental parameters all influence the thermographer’s evaluation of potential problems.

Introduction

Thermographers working outdoors on breezy days, or in areas with nearbycooling fans or blowers are faced with the challenge of the effects of convectiveheat transfer. We use cooling by convection all the time. It should be nosurprise to us then when the temperature rise of a hot spot is reduced by thewind or fans. In fact, maintenance often must use cooling fans on knownoverheated components in hopes of making it to the next outage.

Why then, do we often ignore the wind when doing thermographic surveys? If we believe in the power of convective cooling, we must realize this will affectour readings on hot spots. Most thermographers simply do not know just howimportant wind is in cooling down a hot spot. Also, we do not know how tocompensate for convective cooling effects. This paper gives some interestingdata using a simple experiment of blowing air on a hot spot simulated on a fuse

cutout.

The sun can also be a strong influence on outdoor thermographic surveys fromboth a reflective and warming standpoint . Reflective effects of solar have beenwidely discussed. Use of long-wave cameras (8-12 µm) is the optimal solutionfor solar reflection problems. With short-wave cameras (3-5 µm)thermographers have had good success by changing position with respect tothe target, surveying at night, and learning to interpret reflections.

Solar warming can be a more subtle effect, especially for hot spots that arethermally isolated from the surfaces the IR camera sees. For these “indirect”




12

targets, temperature rises of a few degrees Fahrenheit can indicate significantproblems. Transient solar loading can wipe out these small temperature rises

and you won’t see them. One utility found that great care must be taken whenperforming thermographic surveys on underground equipment that is heavilyelectrically insulated. The electrical insulation also serves well as a thermalinsulator, making these underground components indirect targets. Just a fewminutes exposure to sunlight made thermography impossible on theseunderground components. It is possible that by waiting long enough for thermalequilibrium, perhaps several hours, that the rise due to the internal problemwould be re-established “on top” of the solar loading. But most thermographersdon’t have that kind of time. It is simpler and quicker just to shield thecomponents from direct sunlight.

For indirect targets that soak in sunlight, such as oil filled circuit breakers(OCBs), thermographers need to compare “apples to apples”. That is, be surewhen comparing OCBs that they are equally solar loaded and have been for some time. More work needs to be done in this area, but thermographers havehad success in documenting major problems indicated by small temperaturerises for equipment in full sunlight.

Wind Effects

We set up an experiment in our Infrared Training Center (itc) student laboratorythat allows the students to vary and measure wind speed blowing on asimulated hot spot on an actual fuse cutout. We recognized the possibility of deriving some good data from this laboratory. We can control wind speed from1.0 mph to over 30 mph. The figure below shows a picture of the setuptogether with an example thermogram.

59.0°F

329.9°F

100

150

200

250

300

SP01

SP02

Figure 1. Wind effects experimental setup.




13

A squirrel cage blower provides the wind onto a Type XS 14.4 KV 100 ampfuse cutout. The wind is aimed at the top of the cutout, nominally centered on

the knurled brass piece. We taped Scotch Brand 88 black vinyl electrical tapeto this piece to: A) increase the emissivity to 0.95 and B) to attach a type Kthermocouple.

Regulated, 18 VDC variable power supplies provide power to both the squirrelcage blower and the heat source mounted internally near the top of the fusecutout. We used a pocket wind meter to measure wind speed. We first heatedthe cutout without any wind, allowing 1 hour to attain thermal stability. We didexperiments with initial temperature rises varying from 130 Fº down to 45 Fº byvarying the power to the heat source. We then applied power to the squirrelcage blower to achieve various wind speeds ranging from 1.0 mph to 25 mph.

Temperatures were measured both with an IR camera (ThermaCAM SC1000)and a dual thermocouple setup. Figure 2 is a plot of our IR camera data for three different power settings together with natural log fits to the data.

The experiments show for several power inputs that the influence of wind isquite strong, even for low wind speeds. The temperature rise was cut in half with just a little over a 3 mph breeze! The stronger the wind, the cooler the hotspot, up to a point. As the curves show, the largest changes occur at lower wind speeds. Our data shows that between 50 and 55 mph, the wind hascooled the hot spot to ambient for the power levels we used.

Cooling by convection depends on many factors, not the least of which isshape. The size, shape, orientation to the wind and surrounding structures alleffect convective cooling. Whether the hot spot is emanating from a recessed

area in the component as is often the case with hinges, for example, could

Figure 2. Hot spot temperature rise vs. wind speed.




14

make a tremendous difference in interpretation. Such a region may be shieldedfrom the wind, and largely unaffected by it.

What does all this mean for thermographers? Here are our recommendationsfor dealing with wind, whether from natural sources or generated within your facility.

✦ Buy an anemometer. Pocket size units are quite accurate and cost about$100. Use it on your surveys. Note that getting the actual wind speed onthe hot spot can, and often will be difficult. You are NOT going to place ananemometer within inches of energized equipment! Try to get enoughmeasurements to give yourself confidence of the range of wind speeds thehot spot “sees”. Recognize the shape and orientation of the hot spot

component relative to any surrounding structures. These factors stronglyaffect wind effects.

✦ Within a facility, blowing air can affect measurement on components insidenormally closed cabinets. Opening the door can allow cooling air to enter.We have found some hot spots can be significantly cooled this way. If thereis air blowing on cabinet doors, we recommend shooting them just after opening, before cooling can take place.

✦ If possible, measure component temperatures on the leeward (downwind)side of the hot spot. There will be a temperature difference from thewindward to the leeward side of the hot component. Measuring out of thewind gets you closer to the no-wind condition.

✦ If you are using severity criteria, find out if they are for no wind or lightbreeze. If they are for no wind, even a slight breeze can throw you off by afactor of two on temperature rise.

✦ As you can see from the figure, the higher the T for a given wind speed,the higher the power dissipation in the hot spot. For a 100 amp current togenerate 30 watts of power, the resistance would be 3,000 micro-ohms This resistance level would be a real problem in medium to highvoltage circuitry.

✦ We did all measurements under steady state conditions. Steady statemeans the heat capacity (thermal mass) of the component does not enter

into the physics of what is happening. If you are making measurements in avariable wind, or the wind just changed from high to low or vice versa, this isnon-steady state; you must consider the heat capacity of the component.This complicates matters considerably. High heat capacity components willbe slower to heat up after the wind dies down, and slower to cool downwhen the wind picks up.

✦ The authors do not represent the equations shown in Figure 2 to beapplicable to general thermography survey cases. There are manyvariables, not the least of which are the steady state condition criterion andothers mentioned above.




15

Solar Effects

The sun can help thermographers greatly for those transient heating/coolingapplications such as roof moisture surveys. For steady-state heat flowapplications, the sun can cause problems in measurement. The effect of solar reflection creating false indications or masking true hot spots has been widelydiscussed. In this paper, we will concentrate on the effects of solar loading onindirect measurements, particularly those underground components that the sunilluminates only when the thermographer opens a door or cover.

When opening normally closed compartments, you must be careful of environmental effects such as air flow mentioned above, and the sun if outside.Underground switchgear is normally heavily electrically insulated. A typicalcomponent, a high voltage elbow, is shown in the figure 3. A hot spot simulatedin this elbow with an internal temperature rise of 133 Fº, as measured bythermocouple, has an external hot spot temperature rise of only 17 Fº. Figure 4is a thermogram showing the surface thermal pattern associated with internalheating of this component. Heating was simulated with an internal sourceunder laboratory conditions. In this condition, a thermographer aware of indirectmeasurement criteria would easily determine a problem condition.

But what happens if we allow the part to be warmed by the sun? Figure 5

shows the experimental setup for artificially warming the elbow. We did notcalibrate the lamp to deliver exactly equivalent solar radiance to the elbow.Rather we wanted to show the effects of solar warming as the variation inambient solar radiance can be considerable.

Figure 3. 15.2/26.3 KV Underground Elbow




16

The lamp delivered more energy than would the sun. However, we haveobserved this effect under actual solar loading conditions.

To compare our results of solar warming of both a good and bad elbow, weadded a good elbow to the setup. This is not shown in Figure 5. Thethermograms in Figure 6 do show both elbows. The good elbow is at an angleand slightly above the bad elbow.

Mean

88.2

Mean

88.2

Mean

70.8

Mean

70.8

Diff. 1

17.5

Internal TC reads 204 F.*>90.5°F

*<65.3°F

70.0

75.0

80.0

85.0

90.0

Figure 4. Surface thermal pattern of elbow with Figure 5. Solar loading experimental setup.simulated internal problem, no solar loading.

Diff--Bad

25.4

Good Elbow

Bad Elbow

Diff--Good

29.7

*>90.5°F

*<65.3°F

70.0

75.0

80.0

85.0

90.0

Bad Elbow*<58.1°F

Diff--Bad

42.6

Good Elbow

Diff--Good

44.6

*>149.0°F

60.0

80.0

100.0

120.0

140.0

Figure 6. Thermograms showing solar loading effects on underground components. Right side thermogram has elbows shielded from sun after warming.




17

Results of our experiment are given in figure 6. The thermogram on the left was

obtained with the “sun” still shining on the elbow. There is considerable “glint”or reflection, as we were using a 3 m to 5 m bandpass camera. After warming, we shielded the elbows from the lamp. The right side thermogramshows the thermal patterns due to warming by the lamp and internal heating bythe simulated hot spot.

In both cases, we could not tell the good elbow from the bad elbow. Theheating by the lamp with or without the glint totally masked the problem. Thelamp was on only for a few minutes. Lamp intensity was greater than that of thesun, but our experience has shown it only takes a few minutes for actual solar effects to produce similar effects.

The bottom line is that for normally shaded components where the problems areindirect (thus low temperature rise), you should not let the sun shine on them.

Summary

The wind and sun can strongly affect surface thermal patterns. Asthermographers we must be aware of all the possibilities and take them intoaccount. Indirect measurements where the hot spot is thermally isolated fromthe surface viewed by the camera are more susceptible to wind and sun thandirect measurements. They have a much lower temperature rise and can bemasked more easily.

Attempting to quantify these effects can result in some degree of frustration.Even under controlled conditions there are many variables to consider. In the“real world” we cannot control much. Documenting electrical load, wind and sunconditions can go a long way to helping us trend problems over time.




18

microsoft word - section 6

Documents