| Page
Index Introduction
to the Measurement Wavelength
Effects: Near-IR, Mid-IR and Long IR Attenuation
by Absorption & Scattering Vs Wavelength Observations
and Open Questions References
and Links to Vendor Pages
Introduction
to the Measurement
In basic terms, the presence of slag on the surface of a stream
of liquid steel or liquid iron (pretty hot stuff~ 1500°-1650°C), can be
seen due to a difference in brightness between the two materials even though they
are both at about the same temperature. This fact has been used for many, many
years by operators of steel melters to detect when slag is coming out of a melting
furnace or ladle. Until 20 or so years ago, it was about the only way one could
see when it was time to change the furnace pouring angle to limit the amount of
slag poured into the next vessel. The original detector was, and in many plants
still is, the human eye. Liquid steel is very, bright because of the high temperature.
Thus, the person observing the stream looked through a special filter which reduced
the level of visible radiation reaching one's eyes. The
reason for the difference in brightness between the two materials, liquid slag
and liquid steel or liquid slag and liquid iron or liquid slag and just about
any liquid metal, is because liquid slag and liquid metals have different spectral
emissivities (Note: if you are unsure what the term spectral emissivity means,
visit the E-Trail to learn something about it).
The spectral
emissivity of slag is reasonably constant over the thermal infrared portion of
the Electromagnetic Spectrum, and has a larger value than that of the steel. Furthermore,
the spectral emissivity of liquid steel drops in value at longer and longer wavelengths.
Over the same wavelength range, say from a wavelength of about 0.5 micrometers
(microns) to a wavelength of about 15 micrometers, the spectral emissivity for
steel will reduce from about 0.5 to about 0.05. The value for slag, on the other
hand, remains nearly constant at about 0.8 to 0.9, depending upon its composition.
Wavelength
effects-Peak Wavelenth of Emitted Thermal Radiation Due
to the nonlinearity of the emission of radiant thermal power over the above wavelength
passband (0.5 to 15 micons), as described mathematically by The Planck's Law equation,
the peak in the Planck curve of Radiance vs Wavelength, occurs at about 1.5 microns.
Wavelength
effects-Near IR Instruments
measuring at wavelengths shorter than the wavelength of peak emission will have
a very non-linear response to temperature versus a linear response to radiance
(brightness). This fact allows a thermal imager operating at short wavelengths
to discriminate between slag and steel many times easier than a human eye, even
under conditions of serious attenuation in the sighting path. So, basically, if
a human can see the slag, a short wavelength thermal imager can see it even more
easily. Further, since most Silicon CID and CCD detector arrays (the detector
used in several instruments) operate at about a 0.9 micron peak sensitivity in
the Near IR, they may see better than the human eye if the dust in the sight path
is composed of particles in the sub 1 micron size range because of reduced scattering
effects. Instruments
operating at or near the wavelength of peak emission wavelength (~1.5 microns),
also Near IR, will still have a very non-linear response, but will have the possible
added advantage of even lower sensitivity to a major component of dust, if any,
in the sub 1 micron size range. Any Near IR Thermal Imager has
several other,very significant advantages over the alternative devices described
below. The final choice of an instrument is not usually based on just one or two
advantages, however, but an overall assement of the pros and cons of a given situation.
So, rather than making any recommendations, we have tried to lay out all the parameters
and related theory and tradeoffs below to inform a user of all the factors so
they can compare all the parameters and arrive at a conclusion that suits their
need. Wavelength
effects-Mid IR Instruments
operating in the MW -IR at about 3-5 microns, are on the long wavelength side
of the Planck curve. They have a reduced sensitivity to the apparent temperature
difference between the steel and slag, but there is an increased difference in
the perceived temperatures because of the reduced spectral emissivity value of
the steel. The steel will look colder in this wavelength range than in the Near-IR.
This improves "contrast" for an imager but this advantage is somewhat
offset by an increases sensitivity to attenuation due to any sight path attenuation.
Instruments working in this waveband will have even less sensitivity to scattering
by dust in the sub 1 micron waveband region, but can be severely affected by scattering
and direct blockage of radiance due to larger dust, smoke and macroscopic-sized
particles, just as the other two wavelength range devices would be. Wavelength
effects-Long IR Instruments
operating in the LW-IR region at about 8-12 and 8-14 microns, are far out on the
wing of the long wavelength part of the Planck curve. Their response is nearly
linear in terms of radiance and apparent temperature. The "contrast"
is further improved between steel and slag by the much reduced spectral emissivity
value of the steel. The steel will look very much colder than the slag. This 'advantage'
is even more significantly offset by a very much increased sensitivity in the
temperature difference to the effects of any attenuation in the sight path. In
these wavelength passbands, the effects of sight path attenuation due to scattering
from sub 1 micron sized particles is very small, but, as mentioned above, larger
particels can affect the attenaution significantly. Attenuation
by Absorption & Scattering Vs Wavelength
Clearly the effects
of absorption of radiation from the stream by absorption and scattering of IR
radiation in the sight path between the steel stream and the thermal imaging camera
is one of the parameters involved in choosing an instrument to perform the measurement.
It is not the only parameter, however, but about the only one that is not well
described in the literature and one that is still subject to many untested hypotheses
in the year 2001. The
problem exits because there is an atmosphere containing Water Vapor and much smoke
and dust. The latter is generated in the pouring of liquid steel into a vessel
like a refractory-lined ladle. Early in most pours, various solid, cold additives
and fluxes are also added to the ladle. Some portion of the additives and refractory
linings of the ladle are vaporized when the steel first arrives and some parts
are physically ejected upwards and out the ladle. The net result is a rapidly-generated,
billowing cloud of dust, smoke and possibly steam that rises up from the ladle,
surrounding and obscurring a view of the stream of steel-slag. The
question of absorption of emitted infrared by atmospheric components is a relatively
small issue since most of the instruments' measuring wavebands are chosen where
there is good, not necessarily perfect, transmission in the presence of nominal
amounts of Water Vapor and Carbon Dioxide. Given the very large magnitude of apparent
temperature difference between the steel and slag in all these passbands, an atmospheric
transmission correction or variation will make liitle difference to the overall
measurement. Attenuation
by scattering is a different matter since it is wavelength and particle size dependent.
Small particles scatter short wavelength radiation widely more than they scatter
long wavelength radiation. Thus, it is often observed thatLlong wave IR sees better
than Medium wave IR cameras which in turn can can see better in smoke and fog
than can shorter, Near IR cameras. That's smoke and fog, which are composed of
very small particles, one's that approach the size of the NearIR wavelengths,
about 1 micron or less. We
have searched the literature and have not found a characterization or estimate
of a typical dust and particle cloud in terms of the particles that make it up
and their size distributions in the case of the smoke and dust generated during
steel pouring into a ladle. Suffice it to say, if it were composed of a large
amount of submicron particles, it is very likely that, at least in the USA, the
US Government OHSA organization would be mandating personal protectives measures
against sub-micron sized particle inhalation hazards in steel-making plants. Since
OSHA does test these plants and since they don't have such a mandate, it is likely
that no suchvery small particle distribution exists, at least not in significant
quantities.
However, in most cases, this may be a moot point and not really significant in
terms of these measurements.
Observations and Open
Questions In
the vast majority of pours, the smoke and dust cloud disappears within a few tens
of seconds after the beginning of a pour. The view of the stream is clear and
unobstructed after that until the presence of slag occurs, usually near the end
of a pour. That means that in many plants, the sight path attenuation question
never arises. Does it occur in your plant or is the point moot? In
those plants where additives are placed into the ladle near the end of a pour,
the sight path is fully or partially obscurred during much of the pour. In these
cases, there may indeed be an advantage to the use of a MW or LW thermal imager,
but only if a significant micron or sub-micron particle size distribution exists.
If this turns out to be the case, then it should also be a clear indication that
a careful OSHA particle hazard test may be also warrented. Has OSHA tested your
plant atmosphere during additive additions? Might be worth a check. Is
it possible to modify an additive addition practice in order to improve the detection
of slag through use of a lower cost, superior imaging system? That, too is an
open question which has not been discussed in the literature.
| 