Note: Descriptions are shown in the official language in which they were submitted.
~7~
METHOD FOR REMOVING CARBONACEOUS DEPOSITS
FROM HEAT TREATING FURNACES
TECHNICAL FIELD
This invention pertains to the removal of carbona-
'ceous deposits such as soot from heat-treating furnaces
used to heat treat ferrous metals under an atmosphere
containing reactive carbon.
S BACKGROUND O~ THE PRIOR ART
The metallurgical industry employs heat-treating
furnaces for a variety of purposes. Under certain
heat-treating conditions deposits of carbon may form in
the furnace. For example, in gas carburization an
atmosphere containing carbon donating (reactive carbon)
constituents is employed to transfer carbon to the
surface of steel, thereby causing a high carbon surface
layer to be formed which increases the surface hardness
of the part after rapid cooling (e.g. quenching). The
presence of these carbon donors in the atmosphere may
~'
also lead to unwanted deposits of el~mental carbon
being formed at various points in the furnace.
~ owder metallurgy involves the sintering of objects
produced by compression of powdered metals. To insure
S the development of adequate density, and release of the
ohject from the mold, these powders contain a lubricant
W}liCh is usually a carbon-containing solid, such as a
metal soap. During heating in the furnace, the lubri-
cant decomposes releasing volatile materials which form
sooty deposits in the furnace.
Deposi-tion of soot or other carbonaceous material
in the furnace is undesirable. It ~ay obstruct the
flow of gases through the furnace, interfere with
effective heat transfer, and in some cases, may react
with high temperature alloy components of the furnace
reducing their strength and durability. It is necessary
periodically to remove carbon from heat-treating furnaces.
An effective way to achieve this end is to burn the
carbon out with air or with air diluted with an inert
gas such as nitrogen.
However, since the reaction of oxygen with carbon
to form carbon dioxide is highly exothermic (96,000
calories (96 kcal) per gram mole at 1700F.) great care
must be exercised to avoid overheating which can cause
severe damage to the furnace or its internal components.
The effects of sooting and of the catastrophic melting
of an alloy radiant heating tube during soot buxnout is
shown and discussed in an article entitled, "Under-
standing Conditions that Affect Performance of Heat
Resisting Alloys`', Parts I and II appearing in the
March and April 1979 editions of Industrial Heating,
Vol. XLVI, No. 3, pp. 8-11 and Vol. XLVI, No. 4, pp.
44-47. It is customary to control the rate of carbon
~urnout by lowering the furnace temperature, and by
using a gas containing only a low concentration of
J L~
oxygen so the heat may be removed as it is produced.
However, this slow process may require many hours for
completion, and during this time the furnace is not
available for useful work.
BRIEF 5UMMARY OF THE INVENTION
We have discovered means which rapidly remove
carbon deposits by reaction with a gaseous cleaning
agent (burnout gas) without creating excessive tempera-
tures (hot spots) at any point in the furnace. It has
been found that if the furnace is maintained at the
temperature normally employed for heat~treating, e.g.,
from about 1500F. to about 1900F. (816C to 1038C~,
an atmosphere containiny substantial quantities of
carbon dioxide will rapidly and completely convert any
deposits oE carbon to gaseous carbon monoxide. Further,
since the reaction of carbon and carbon dioxide to form
carbon monoxide is endothermic, (40 kcal being absorbed
per gram mole of carbon removed), the region in which
carbon removal takes place is actually cooler than the
remainder of the furnace, and heat must be supplied.
Upon completion of the carbon removal, which may be
monitored by observing the concentration of carbon
monoxide in the exit gases, the ~urnace may be put back
on stream in a short time simply by flushing with the
normal heat-treating atmosphere to which a small amount
of a carbon dioxide scavanger, such as natural gas, has
been added.
Water may also be employed for safe removal of
carbon since it too reacts rapidly with carbon in a
process which requires 32 kcal per gram mole of carbon
removed. Water is more difficult to remove from a
furnace than is carbon dioxide, and, therefore, additional
time will be lost before ~he furnace is completely
prepared to go back into heat-treating service. Further-
more, when water is employed as a cleaning agent, itusually must be diluted with an inert carrier gas such
as nitrogen to avoid condensation in cooler parts of
the system such as inlet ducts, vent lines and stacks~
Carbon dioxide may be employed in admixture with
an inert carrier, or if desired, may be used undiluted,
in which case the most rapid clean out of the furnace
will be attained.
DETA I LED DE S CR I PT I ON OF THE I NVENT I ON
-
Normally furnaces containing heavy deposits of
carbonaceous material (e.g. soot) are burned out with
lQ air or diluted air, for example, 10% air in nitrogen.
The reason for this is the generation of hot spots when
substantial amounts of soot are deposited on the radiant
tubes. Oxygen reacts with carbon to form either carbon
monoxide or carbon dioxide depending on the relative
amounts of oxygen and carbon. This reaction is exo-
thermic, by generating one gram mole of carbon monoxide
at 1700F about 28 kcal would be produced; by ~enerating
one gram mole of carbon dioxide at 1700F about 95 kcal
would be produced. On the other hand, the reaction
between carbon dioxide and carbon consumes 39 kcal per
gram mole and hot spots simply cannot occur. Therefore,
it was reasoned that pure carbon dioxide could be used
for removing the soot and a much faster operation would
result.
The following example illustrates the furnace
cleaning process of the present invention.
A heat-treating furnace having a volume of 7.5 cu.
ft. was intentionally sooted by passing a mixture of
nitrogen and propane through it for a period of 17 hrs.
at a temperature of 1~00F (927C). The flow rate of
nitrogen was 100 standard cubic feet per hour (SCFH3
and of propane was 2.5 SCFH. A total of about four
pounds of soot was deposited in the furnace. The
nitrogen propane mixture was then replaced by a stream
of carbon dioxide at a rate of 100 SC~1 whereupon the
temperature of the furnace dropped about 60F (33.3C~,
an indication of an endothermic process. The furnace
atmosphere was sarnpled and analyzed a-t periodic intervals
with the results as shown in the following Table I:
TABLE I
-
Time Furnace Atmosphere (By Volume~
Minute _ ~ C2 % C0
7 5.49 2.42 B5.B7
13 2.05 2.37 89.74
0.92 2.37 89.74
27 0.48 3.22 91.40
34 0.28 3.89 91.26
41 0.17 7.32 88.52
47 0.11 14.61 81.79
54 0.07 21.72 75.22
6~ 0.05 30.87 66.S0
, The concentration of hydrogen, which resulted from
- 20 cracking of the propane, drops rapidly as carbon dioxide
sweeps through the furnace. However, only a low concen-
tration of the latter appears at the exit of the furnace,
most of it being converted to carbon monoxide by reaction
with soot. This process occurs for about 40 minutes,
at which time an abrupt rise in carbon dioxide concen-
tration and a corresponding decline in carbon monoxide
is ohserved as complete burnout is approached.
From these data it is estimated that approximately
three pounds of carbon has been removed in a period of
about 60 minutes without production of a hot spot in
the furnace.
To remove the same quantity of carbon with 100
SCFH flow of pure air reguires 2.2. hours, and would
~ ê)~
result in generation of localized high temperatures at
the points where the soot burns. If the normal mixture
of 10% air and 90% nitrogen were employed at a flow
rate of 100 SCF~, 22 hours would be re~uired to achieve
5 the same result.
The carbon dioxide level in the burnou-t mixture is
preferentially 100% but could be varied between 10 and
100%. The flow of burnout gas can be as high as 300
SCFH or 40 v~lume changes. Since the gas volume is
doubled upon combining with carbon, a considerable
increase in gas pressure occurs in the furnace. The
upper limit of the carbon dioxide flow is determined by
the maximum flow the furnace is designed for. The
pr~ferred carbon dioxide flow is between 10 and 40
volume changes. At smaller flows, carbon dioxide would
be completely converted but the burnout would last
longer. The burnout temperature is limited to the
normal carburizing temperatures and will prefexentially
be between 1500 and 1700F (816C and 927C). At lower
temperatures burn out will be slower and less efficient.
Higher temperatures will not be a problem. The upper
limit is given by the maximum operating temperature of
the furnace.
In order to prepare the furnace for normal opera-
tion and xeduce the carbon dioxide level a mixture of100 SCFH nitrogen and 20 SCFH natural gas was fed to
the furnace. After 10 minutes the furnace atmosphere
contained 20% hydrogen, 0.05% carbon dioxide, 1.6%
methane and 12% carbon monoxide. This may be considered
a perfect atmosphere to start any nitrogen based carbur-
izing system.
The results indicate that the proposed method of
burning out a furnace results in production time savings
of 1.2 up to 21 hours. At the same time the risk of
tube burnout is completely eliminated.
Water vapor can be used as a component of the burn
out gas since water vapor has a heat consumption of 32
kcal/grc~m mole of carbon. The water in a nitrogen gas
4~i~
mixture should be present in an amount from, by volume,
0% to 20%. However, water in the liquid form compli-
cates metering it into the furnace and ~he furnace will
have a high dew point. Thus it reguires at least
several hours at temperature after cessation of ~ater
injection to dry out the furnace.
