United States
Environmental Protection
Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S7-86/034 Feb. 1987
Project Summary
Evaluation of Primary Air
Vitiation for Nitric Oxide
Reduction in a Rotary
Cement Kiln
R. C. Benson and S. C. Hunter
Results of pilot-scale tests to evaluate
combustion modifications for nitric oxide
(NO) reduction and cement product quality
in a long-dry-process cement kiln are re-
ported, firing pulverized coal. The kiln is
rated at 11.35 kg/s (1080 tons/day) of ce-
ment with a thermal input rate of 70.3 MW
(240 x 106 Btu/hr). Of the combustion
modifications evaluated in previous stud-
ies, vitiation of the primary air with inert
gas (nitrogen) was considered to be the
best alternative for NO reduction on a
pilot-scale kiln.
As-found emissions (17 days of data,
89.75 hours) were about 3.6 kg (8 Ib) NO
as N02 per 907 kg (1 ton) of clinker. The
uncertainty of the as-found mean mass
emissions is estimated to be 10 percent.
Lowered excess air (LEA) decreased NO
volumetric and mass emissions by about
15-20 percent. The percent reduction was
established from a reference baseline of
1050 ppm (corrected to 3 percent O2)
and a kiln exit oxygen of 1.82 percent. The
NO volumetric emissions were lowered to
880 ppm (corrected to 3 percent O2)
with a kiln exit oxygen of 0.7 percent.
NOX reductions during the short term N2
injection tests were about 25-30 percent
with no adverse effects on product quality.
Because of the interacting feed chemistry/
kiln operations, it was not possible to
clearly isolate the effect of N2 injection.
Longer term tests with improved process
stability would help to evaluate the primary
air vitiation.
This Project Summary was developed
by £PA's Air and Energy Engineering
Research Laboratory, Research Triangle
Park, NC, to announce key findings of the
research project that is fully documented
in a separate report of the same title (see
Project Report ordering information at
back).
Introduction
This program is a follow-on study to
build on the results of the programs
reported in EPA-600/7-83-045 and EPA-
600/7-84-075. The objective of the earlier
effort was to determine the effect of
burner parameters on near-flame nitrogen
oxide (NOX) levels for natural gas and coal
fuels. This laboratory effort determined
that combustion air preheat, fuel injection
velocity, and oxygen (02) content of the
primary combustion air stream have first-
order effects on NOX levels. These param-
eters were then selected for implementa-
tion on a subscale kiln. Of these three
combustion modifications, the effect of
carrier (primary) air 02 concentration on
NOX emissions was most significant.
The effect of carrier air O2 content on
NOX emissions showed a strong influence
of carrier gas O2 concentration on NOX
emissions. NOX emissions were reduced
37 percent by lowering the carrier gas O2
to 12.5 percent. While NOX reduction was
significant, the subscale clinker quality
was not comparable to production clinker,
but primary air vitiation did not seem to
affect clinker quality substantially from the
baseline condition. Acceptable production
clinker is about 0.5-1.5 percent free lime.
The results of the combustion laboratory
tests were corroborated by the subscale
test results.
-------
These test results indicated that dilution
of the primary air stream with an inert gas
offered the highest probability of NOX
reduction by combustion modification on
a full-scale kiln.
Present Approach
The objective of this program was to im-
plement the results of previous work on
a full-scale cement kiln to investigate the
effect of combustion modification on
product quality.
This test program was conducted at a
cement plant in California. The cement
company was interested in such a test
program because emission limit goals of
1.4 kg (3.1 Ib) NOX/907 kg (1 ton) clinker
are being considered by the local air quality
management district. Data with inert gas
injection would be useful input in deter-
mining the feasibility of the emission limit.
Three alternatives were considered for
partial inerting of the primary air stream:
• A 9.8 kW (600 hp) boiler supplying
flue gas.
• Flue gas recirculation from the kiln
stack.
• Nitrogen (N2) injection.
Alternatives were analyzed for initial and
operating costs, operability, and flexibility.
N2 injection into the primary air stream
was the lowest cost alternative for a very
short term test. If successful, flue gas
recirculation would be more cost effective
for long term operation.
Kiln System Description
The kiln is a long dry rotary cement kiln
149 m (490 ft) long and 4.1 m (13.5 ft) in-
side diameter. Kiln rated production is
11.35 kg/s (1080 tons per day). The ther-
mal input rate is about 70.3 MW (240 x
106 Btu/hr). The kiln is direct fired with a
Southern Utah bituminous coal at about
0.105 kg/s (10 tons/hr maximum).
N2 Injection
Gaseous N2 was injected at the coal
mill inlet, downstream of the tempering air
dampers. The N2 injected into the primary
air duct was maintained at about the same
temperature as the primary air stream.
Data Acquisition Methods
The kiln system is instrumented for
emissions data along with temperatures,
pressures, and controls for the coal mill
system, clinker cooler box, the kiln, and
downstream components (e.g., multiclone
and baghouse). Feed input and clinker
product output are also monitored.
Gaseous emissions were measured
using analytical instruments and equip-
ment contained in a government-furnished
mobile instrumentation laboratory. The
laboratory is equipped with analytical in-
struments to continuously measure con-
centrations of NO, NO2, CO, CO2, and 02.
For the as-found, low excess air and N2
injection test series, only NO was meas-
ured with CO, CO2, and O2. Measure-
ment of total NOX (NO + N02) requires a
heated sample line and involves a differen-
tial measurement of NO and total NOX
(the difference is NO2). Conditions in the
kiln were not sufficiently steady to allow
accurate differential determination of
NO2. The sampling system extracted gas
from the exit of the kiln, where raw feed
is injected.
An On-line Emissions Monitor Program
was written for the Apple II Plus 48K com-
puter to provide automatic digital data
acquisition for real time monitoring of
emissions data. The computer was pro-
grammed to scan every 5 seconds and
print the gaseous emissions data averaged
over 1, 3, or 5 minutes. The program pro-
vides a continuous visual display of time
plots for any two of the gaseous species
(e.g., NO, 02) and records the data printed
out on magnetic disk. Strip chart recorders
are used to verify the data recorded on
magnetic disk and fill in any gaps from the
computer, if they occur.
O2 levels in the primary air were sam-
pled during the N2 injection tests using a
portable Teledyne Model 320A 02 ana-
lyzer. The sample port for O2 measure-
ment was in the 41.6-cm (16 in.) diameter
burner line about 15m (50 ft) from the
burner tip. During the N2 tests, carrier gas
O2 was monitored every 1-2 minutes.
Data Analysis Methods
Daily kiln exit gaseous emissions nor-
mally analyzed were NO, CO, CO2, and
O2. The 02, C02, and CO data are re-
ported on a dry volume basis, and the NO
data were reported on a dry volume basis
corrected to 3 percent 02. The dry stack
gas data were statistically analyzed for the
arithmetic mean, standard deviation, co-
efficient of variation, and 95 percent con-
fidence level.
The volumetric NO concentration was
converted to pounds per ton of clinker and
reported as N02. The plant measure-
ments used for clinker rate are:
• Clinker Weigh Belt Scale.
• Kiln Feeder Speed (strip chart).
• Kiln Feeder Revolutions Counter.
• Kiln Speed.
Instantaneous clinker rate can also be cal-
culated from a formula for million Btu per
ton clinker based on gaseous emissions
data (CO2, O2), heating value, ultimate
analysis, and ignition loss of the raw feed.
Once the million Btu per ton of clinker i
calculated, the fuel rate and heating value1
are used to calculate the clinker rate. This
method was found to be less reliable than
using the plant instruments.
As-Found Test Series
Seventeen days of as-found data were
collected to assess process variability.
Twelve coal samples, 67 clinker samples,
30 raw feed samples, 2 multiclone dust
samples, and 2 combined feed samples
(raw feed plus multiclone dust) were col-
lected. A multiclone dust sample was
analyzed for particle size and chemical
composition.
The coefficient of variation for the
volumetric NO emissions has been plotted
on a daily basis. The coefficient of varia-
tion ranges from a low of 14.5 percent to
a high of 57.8 percent. A large dispersion
from the mean (time weighted average
coefficient of variation of 31 percent)
shows that daily cement kiln emissions
(volume basis) are extremely variable. The
distribution of emissions factors has been
plotted on a pound NO (as NO2) per ton
of clinker basis. The mean emissions fac-
tor for the 90 hours of testing was 3.6 kg
(8.0 Ib) NO (as N02) per 907 kg (1 ton) of
clinker. The uncertainty of the mass emis-
sions factor is about 10 percent based on
the input data and measurements needed
to calculate pounds NOX per ton of
clinker.
Clinker in the kiln is normally burned at
the plant by maintaining a constant fuel
rate and varying the kiln speed to hold the
burning zone temperature at the desired
level. When the material temperature is
higher than the desired level, the kiln speed
is increased to decrease the front-end
temperature. Maintaining stable kiln con-
ditions is difficult. The strip chart record-
ings for NO and kiln speed have been
redrawn onto one figure (Figure 1) to show
how closely the NO tracks the kiln speed
changes. Also, the strip chart recordings
for NO and material temperature have
been redrawn onto another figure (Figure
2) to show how NO and material tempera-
ture coincide for underburning and over-
burning conditions. A strong correlation of
NO with material temperature is evident.
Clinker analysis results for the as-found
clinker samples showed the standard devi-
ation of tricalcium silicate to be about 6.6
percent. The variability was traceable to
homogenizing silo problems and rock mix
difficulties.
During another test segment, 49 con-
secutive hours of as-found data were
analyzed for 3- and 24-hour rolling aver-
ages. The 3-hour rolling averages ranged
-------
0600 0800 1000 1200
Time, Hour of Day
Figure 1. Time trace of NO and kiln speed.
1400
1600
£ (1650)
.C 3000
' (1550)
^ 2800
2600
^ 2400
| (1200)
a 2200
I
(1100)
2000
-i 1 i 1 r
Material
Temperature
\
/vo
0600
OSOO
Figure 2.
1000 7200
Time, Hour of Day
Time trace of NO and material temperature.
1400
1600
2500
:2000
c
7500
7000
500
0
I
i
from 2.1 to 4.55 kg (4.62 to 10.04 \b) NO
(as NO2) per 907 kg (1 ton) of clinker and
the 24-hour rolling averages ranged from
2.99 to 3.82 kg (6.60 to 8.42 Ib) NO (as
NO2) per 907 kg (1 ton) of clinker.
The main conclusion from the as-found
tests was that the wide variation in NO,
the strong correlation of NO with material
temperature, and the difficulty in maintain-
ing stable operation of the kiln make con-
trolling NO very difficult. On the other
hand, if means can be found to stabilize
the process using NO as a key control var-
iable, product quality would be more
uniform and controlling NO would be
enhanced.
Initial N2 Injection Results
Evaluation of the primary air vitiation
system was based on tests conducted on
three days (November 15 and 17, 1983,
and April 17-18, 1984). Between the two
November tests and the April test, low ex-
cess air tests were conducted.
The objective of the November 15 N2
injection tests was to shake down the
system to ensure that nominal flow rates
of up to 0.80 m3/s (100,000 scfh) would
not adversely affect mill operation or
burner stability. The mill outlet average 02
concentration was about 17 percent. NO
emissions decreased during the N2 injec-
tion tests. The extent of NO reduction with
N2 injection was more difficult to quantify
because kiln speed changes affect NO
emissions. Clinker analysis shows that the
tricalcium silicate and dicalcium silicate for
baseline and N2 injections showed very
little variation. The difference of the
averages between baseline and N2 injec-
tion was 0.40 for tricalcium silicate and
0.09 for dicalcium silicate. The objective
for the November 15 tests was achieved;
N2 could be injected with no detrimental
effect on coal mill operation and no ob-
vious effect on flame stability. The results
with regard to the effect of N2 injection
on NO2 emissions were inconclusive at
17 percent carrier air 02.
A longer term N2 injection test (about
3-1/2 hours) at higher N2 flow rate was
performed on November 17. Plotting the
emissions data for NO and kiln exit 02 as
a function of time shows a distinct de-
crease in NO. But also apparent is a de-
crease in kiln O2 that closely follows the
NO decrease. At the lowest NO point, CO
spikes of 1000 ppm or more were ob-
served. It was not clear from these tests
if the NO decreased as a result of N2 in-
jection or because the 02 decreased. Free
lime and x-ray diffraction analyses did not
indicate any deterioration of product qua-
lity during the injection tests.
Low Excess Air Tests
Initial tests with short-term N2 injection
(1-3 hr) indicated that the N2 injection
caused a reduction in excess air as the
result of flow balance changes in the kiln.
Therefore, prior to proceeding with further
N2 injection, a 4-day test was conducted
to evaluate more closely the effect of
excess air on NO.
Test results indicated a significant effect
of kiln exit 02 on NO, but also showed a
strong effect of material temperature.
Lowered excess air (LEA) decreased NO
volumetric and mass emissions by about
15-20 percent. The percent reduction was
established from a reference baseline of
1050 ppm (corrected to 3 percent O2)
and a kiln exit oxygen of 1.82 percent. The
NO volumetric emissions were lowered to
880 ppm (corrected to 3 percent O2)
with a kiln exit oxygen of 0.7 percent. The
average calculated kilograms (pounds) NO
as N02 per 907 kg (1 ton) of clinker was
3.66 (8.07) at reference baseline condi-
tions and 3.18 (7.02) at low excess air con-
ditions: the resultant reduction was 16.5
percent. The average burning zone tem-
perature, as measured by an optical pyro-
meter, was 1358 °C (2477 °F) at reference
baseline conditions and 1349°C (2460 °F)
at LEA conditions. The effect of kiln feed
burnability factor on NO emissions was
not included in this analysis.
Final IM2 Injection Tests
The final N2 injection tests were per-
formed for about 8 hours on April 17 and
18. The objective of the tests was to ad-
just kiln operating conditions during N2
injection so that kiln exit O2 would remain
as constant as possible A constant O2
would factor out the effect of O2 which
has complicated the data analysis for
previous tests.
A time plot of pounds NO2/ton of
clinker, kiln exit O2, and mill exit O2
showed that the objective of holding kiln
02 constant was not completely met, but
the 02 was considerably steadier than in
previous tests. The plot also showed the
period of N2 injection: pounds N02/ton of
clinker was reduced from about 3.9 to 3.2
kg/907 kg (8.6 to 7.0 Ib/ton) with a pri-
mary air 02 at 13 percent. The percent
NO reduction was 19 percent. This degree
-------
of reduction cannot be attributed entirely
to the N2 injection. Changes in kiln speed,
coal flow, material temperature, feed com-
position, and other variables, in spite of the
operator's desire to maintain constant
conditions, complicate the data analysis.
The data were subject to a statistical time
series analysis to attempt to isolate the ef-
fects of the important variables. The data
show that the NO concentration change
per 1 percent change in mill outlet oxygen
ranges from 16 to 76 ppm NO decrease
per 1 percent mill outlet 02 decrease. The
actual effect depends on what is held con-
stant by the model. NO check was made
for the linearity of this effect.
High speed motion pictures were taken
of the flame near the tip of the burner at
2,000 frames per second. These films and
operator observations indicated a definite
lengthening and cooling of the flame. This
effect is normal for low NO operation on
other combustion devices, but would re-
quire further investigation over a longer
time to determine if this effect could be
acceptable for clinker quality and kiln life.
R. C. Benson and S. C. Hunter are with KVB, Inc., 18006.
Robert E. Hall is the EPA Project Officer (see below).
The complete report consists of three volumes, entitled "Evaluation of Primary
Air Vitiation for Nitric Oxide Reduction in a Rotary Cement Kiln:," (Set Order
No. PB 87-113 239/AS; Cost: $64.00)
"Volume 1. Technical Report," (Order No. PB 87-113 247/AS; Cost: $30.95)
"Volume 2. Data Supplement A, "(Order No. PB 87-113254/AS; Cost: $18.95)
"Volume3. Data Supplement B."(Order No. PB87-113262/AS; Cost: $24.95)
The above reports will be available only from: (costs subject to change)
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Analysis of raw feed and clinker sam-J
pies indicates that most of the clinke"
variability could be explained by feed com-
position variation. A time trace of the NO
volumetric emissions, free lime, and mate-
rial temperature for the nitrogen injection
tests showed that the free lime content
throughout the tests was less than 0.3
percent and alkali content remained less
than 0.6 percent. It was concluded that
the test did not result in deterioration of
clinker quality, but it was apparent that the
clinker was being overburned, probably
due to the kiln operator's concern for
avoiding an upset as the result of the test.
United States
Environmental Protection
Agency
Onter for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S7-86/034
0000329
PROTECTION
------- |