|
BRAKEGLEN
LIMITED
REFRACTORY CONSULTANTS
WHAT IS NOT ALWAYS KNOWN BY USERS
OF
CERAMIC FIBRE, INSULATING
FIREBRICK &
HIGH EMISSIVITY COATINGS
This report endeavours to show that things are not
always as they seem in the world of refractories. It is specific to
ethylene furnaces but the comments will also have a bearing on other
classes of high temperature process plant.
It submits that the move in recent years toward
ceramic fibre linings is fundamentally wrong.
It explains why plants with brick linings frequently
suffer high maintenance costs, and why casing plates often run
significantly hotter than design.
It destroys a major myth which has inhibited
technological advancement and clarifies some aspects of the use of
high emissivity coatings.
It questions the reliability and accuracy of
manufacturer's published data.
CERAMIC FIBRE:
Many new ethylene furnaces have been lined with
ceramic fibre in recent years. Serious questions can be raised about
the wisdom of this decision.
The main advantage of ceramic fibre to industry is
it's low thermal mass characteristic.
This allows intermittent and batch type furnaces to
operate on a reduced cycle time. Ethylene furnaces are continuously
operated under steady state conditions and here the sole advantage of
ceramic fibre is cost. Ceramic fibre linings are cheaper, quicker and
easier to install (and repair) than brick. All else about use of the
material is either questionable, a nonsense, or a danger.
A questionable advantage of ceramic fibre is that it
is a better thermal insulator. Undoubtedly it is, but the benefit of
reduced heat loss on a furnace which loses only 0.02 of total heat
fired through the main fabric is at best marginal. Furthermore, one
is asked to believe that heat saved at 90 deg C can be deducted from
heat supplied at 1200 deg C.
The First Law of Thermodynamics allows this
extravagance in heat balance calculations. The Second Law forbids it.
A 10% reduction in radiant wall heat loss equates, at absolute best,
to 0.0008 of fuel fired. Heat loss savings from improved insulation
pass unnoticed at the gas meter.
Also questionable is the claim that less supporting
steelwork is required with ceramic fibre. This is true in terms of
subsidiary support steel but not necessarily true for the main
structure. Wind load is an important, and sometimes dominating,
stress factor. A 90% reduction in the weight of the internal lining
system can occasion the need for more external bracing.
A nonsense claim is that ceramic fibre is unaffected
by water. This is true in terms of chemistry but not true in terms of
structural stability. Ceramic fibre will absorb water as a sponge.
The density increase can, and has, torn complete fibre systems from
their fixings. One must guard against significant amounts of water
entering a furnace during overhauls.
Improved temperature uniformity is also a nonsense
claim. It is only ever true when a new fibre system has replaced an
old and battle weary brick lining. On a new for new comparison there
is little difference. If anything, a brick's higher thermal mass and
higher surface emissivity should give it the edge.
Low thermal mass brings us into the danger area. This
characteristic of ceramic fibre is a potential hazard on ethylene
furnaces. Rapid cool down following a trip can lead to thermal shock
of the coils. The claim is that ceramic fibre is resistant to thermal
shock. Rarely advertised is the fact that it can cause thermal shock
elsewhere. Some ethylene plant operators have abandoned the use of
ceramic fibre for this reason alone.
Another danger is increased fuel use and/or a fall-off
in production. Ceramic fibre has the lowest emissivity of all the
commercially available furnace lining materials. It causes a
significant reduction in wall to coil radiation which in turn reduces
overall firebox efficiency. The material should only ever be used as
back-up insulation, for expansion joints, or for coil seals. Its use
as the primary hotface lining material on ethylene furnaces is
fundamentally wrong.
A word of caution to those plants already using
ceramic fibre. Trying to increase wall to coil radiation by coating
the ceramic fibre with a high emissivity coating is not without risk.
Coated fibre will cool down about 10% faster than uncoated fibre in
the event of a trip.
INSULATING FIREBRICK:
The vast majority of existing ethylene furnaces are
lined with insulating firebrick. Insulating firebricks were first
developed in the 1940's. Fifty years on there remain several
important aspects of the brick which are not generally appreciated by users.
Insulating firebricks are manufactured by an extrusion
process which gives a preferential bias to both the internal grain
structure and the thermal conductivity. Thermal resistance varies
with the construction orientation of a brick. Manufacturer's
published data inevitably quotes k-values in the most favourable
orientation ie, the brick built "stretcher" fashion (laid
sideways on). Built "header" fashion (laid endways on), the
construction orientation used on many ethylene furnaces, the thermal
conductivity is about 8% higher than published data.
While the thermal conductivity of an insulating brick
is more or less linear with temperature, the thermal gradient through
the brick is not. The common practise is to only take the k-value at
mean temperature into consideration. Simplifying the temperature
gradient to a straight line then results in the tip temperatures of
shelf plates and tie-back rods being understated.
Thermal conductivity is also affected by furnace
atmosphere. Insulating firebricks are extremely porous and typically
have an apparent porosity (connected open pore structure) of 70%.
Reducing conditions are frequently experienced around burners and in
the lower reaches of a radiant firebox. With the thermal
conductivities of air, methane and hydrogen being 0.09, 0.29, and
0.61 W/mK respectively at 1200 deg C, the effect of a reducing
atmosphere on thermal resistance is dramatic and self-evident. Note:
Brick porosity has been quoted to illustrate the point but it is the
permeability of the brick (not necessarily related to porosity and
rarely published) that is the true measure of it's vulnerability in
this situation.
The shrinkage resistance and hot strength
characteristics of insulating firebrick also diminish rapidly in a
reducing atmosphere.
The expansion of brickwork is often assumed to be
uniform in all directions. This does not happen in practise. Bricks
do not expand upward against gravity as much as they do sideways.
Upward expansion is frequently only half that of sideways expansion.
Also, published data on expansion is inevitably overstated as it is
for the brick in a fully soaked temperature environment. The end
result is usually excessive expansion allowance. While this may be
seen as erring on the safe side, it is not conducive to a gas tight lining.
A misunderstanding often arises over brick selection
due to the Classification Temperature system used for insulating
firebrick. Published data typically states 2300 Grade, 2600 Grade (23
Grade and 26 Grade to ASTM C155), and so on. It is often assumed that
these figures refer to the maximum or safe operating temperature in
Fahrenheit. This is not the case. Historically the figures related to
the temperature at which a brick would shrink by 1.5% under it's own
weight. This criteria being a long assumed acceptable level for all
industrial applications. Improvements in the manufacture of
insulating firebrick over the years mean that the shrinkage figures
are now considerably less but the Classification Temperatures remain
and have no bearing on the safe working temperature.
A reduction occurs to the properties of refractories
with time at temperature. Engineers are familiar with the creep
phenomena of metals at temperature but far less so with the decline
of refractory properties with time in service.
Disappointments with refractory life, high maintenance
costs, and excessively hot casing plates can usually be traced
to a lack of appreciation of some of the above mentioned points.
HIGH EMISSIVITY COATINGS:
The introduction mentioned destroying a major myth
which has inhibited technological advancement and it relates to the
use of high emissivity coatings.
Refractory linings have long been considered to be
docile static structures with their primary function being to inhibit
the escape of heat to atmosphere. This is a myth. The refractory
walls in a radiant firebox are a dynamic entity. In reality a working
component. They are akin to a large heat exchanger capturing heat
before it escapes to the bridgewall area and radiating the heat so
captured to the coils. Wall to coil radiation is the predominant mode
of heat transfer in an ethylene furnace firebox.
This fact, and the scope it offers for improvements in
radiant firebox efficiency, has passed largely unrecognised for
decades. Text books are partly to blame and have no doubt misled the
industry. Who would look for plant improvement from refractory
emissivity enhancement when advised that, "Refractory bricks
have an emissivity of 0.9", and "Furnace enclosures act as
a black body"?
The end result is that the vast majority of ethylene
furnaces, regardless of their assumed high overall thermal
efficiency, are firing somewhere between 6% and 10% of fuel unnecessarily.
With the interest in high emissivity coatings now
gathering pace some misunderstandings and misconceptions are already apparent.
The first concerns the colour of high emissivity
coatings. Some vendors make a major selling point of having a black,
or near black, coating and misuse the term "black body"
coating. Visible colour is not the dominant factor in the infrared
region of the spectrum. Strange things happen in the quantum world of
electromagnetic radiation; some white radiator paints have a higher
emissivity than black paint at room temperature and human skin has an
emissivity of 0.9 in the infrared! The preoccupation with a visibly
black coating has often led to the use of metallic oxide pigments in
the formulation with their attendant health, safety, and final
disposal risks.
It is also common practise for vendors to quote a
blanket emissivity figure, such as 0.9, for a particular coating.
Most refractory materials have an emissivity of about 0.9 at room
temperature. At elevated temperature the emissivity becomes strongly
wavelength dependent; varying accordingly and sometimes dramatically.
On ethylene furnaces it is only the refractory emissivity in the 1.5
to 6 micron wavelength region of the infrared spectrum, and to a
lesser extent in the 6 to 14 micron region, that has any real bearing
on net heat transfer. Refractory emissivities quoted without
reference to wavelength are not only misleading but meaningless.
Having said this, a serious problem arises in the
measurement of emissivity at discrete wavelengths. No International
Standards exist. There is no approved test apparatus and no approved
test method. No guidance notes exist as to the correct interpretation
of results. Test houses and authorities on the subject are often
critical of each other's approach to measurement. As things stand
today it is a minefield through which coating vendors are free to
stroll without fear of explosion.
Another problem arises when trying to predict a change
in heat balance from an increase in refractory emissivity. Most plant
operators have access to a mathematical model of their furnaces.
Aside from a handful of universities and research establishments, no
models are programmed with the spectrally selective emission
characteristics of hot gas and the dramatic change in refractory
emissivity with wavelength. Consequently, the majority of
mathematical models in present use cannot accept the phenomena at
work let alone predict the outcome. This presents a problem for those
who do not like to go where the computer cannot go.
A final word of caution. The technology may not be
suitable for all plants. An increase in radiant firebox efficiency
inevitably results in a reduced mass flow of cooler gases to the
convection section. The use of a high emissivity coating may not be
appropriate on plants which run lean on steam.
THERMOPHYSICAL PROPERTIES:
"The published data is typical and presented in
good faith. It does not form a specification and no warranty is
implied". These, or similar words, appear without exception on
published data throughout the refractories and high temperature
insulation industry. Why?
Refractory products, castables, bricks, cements,
mortars, fibres, etc are manufactured for the most part from
naturally occurring clays and minerals which are subject to
considerable variation from country to country and quarry to quarry.
Small variations, particularly with the unwanted inclusions, can have
a marked effect on the final properties. The crushing, grinding,
blending, mixing, pressing, drying, firing, and so on, all have a
bearing on the final outcome. A not unreasonable comparison lay with
baking a cake when the ingredients are subject to variation. The
skill of the chef comes into play. Refractory manufacture is not all
science. Some art is present. So is compromise. Rarely, if ever, can
a given property be improved except at the direct expense of other properties.
So, with a variable product and unwarranted published
data what tolerances should be applied to the figures which do appear
in print?
Before trying to apply sensible percentages,
consideration also has to be given to the accuracy and consistency of
existing test methods. Surveys often show surprisingly large
variations. A recent example is not unusual. Seven accredited test
houses were sent a sample of 135kg/cu metre density mineral wool cut
from the same "control" piece of board. The measurements
for thermal conductivity gave a spread of +15% -15% about the average
in the temperature range 100 to 500 deg C! Bear in mind that this is
a relatively low temperature survey. The higher the temperature the
more difficult the accurate measurement of conductivity becomes.
The problem is not limited to the measurement of
thermal conductivity. It exists to some degree with all the
thermophysical properties at high temperature - thermal diffusivity,
permeability, refractoriness under load, crushing strength, modulus
of rupture, reversible thermal expansion, permanent linear change,
and so on.
Competitive pressures frequently force designers close
to the line on safety margins. When the line is unwarranted and
ill-defined then real dangers exist. For those who wish to sail in
safer water and uphold high engineering standards, it is recommended
that all published data on refractory materials, regardless of the
particular property concerned, be subject to at least a 20% downward adjustment.
CONCLUDING COMMENT:
It is perhaps not by mere chance that the word
"refractory" has a shared meaning with that which is
perverse and unruly. Refractory technology certainly differs from
other engineering disciplines in that it is not an exact science.
Engineers can design with confidence in steel. The
application of reliable and consistent physical property data to
established formula readily gives accurate values for stress, strain,
sheer, deflection, creep, yield point, and so on. The refractory
designer not only uses different language for the same phenomena but
only has unwarranted property data to go on. Furthermore, the
published data is rarely suitable for use in standard formula.
The shortfall in "science" with refractory
technology can only be made up by experience and a sympathetic
understanding of the materials concerned.
Copyright © John G Clements 1999
THE AUTHOR: John G Clements is Managing Director of
Brakeglen Ltd (Refractory Consultants), Helston, Cornwall, England.
His 40 years experience in the refractories industry includes 25
years with M H Detrick Company in the UK where he also served as
Managing Director. Mr Clements is a Chartered Mechanical Enginneer, a
member of the Institute of Energy, and a founder member of the
Thermophysical Properties Awareness Club at the National Physical Laboratory.
|
