United States
Environmental Protection
Agency
Industrial Environmental
Research Laboratory
Research Triangle Park IMC 2771
Research and Development
EPA-600/S7-84-080 Sept. 1984
&ERA Project Summary
Evaluation of Natural- and
Forced-Draft Staging Air Systems
for Nitric Oxide Reduction in
Refinery Process Heaters
R. C. Benson
Results of pilot-scale tests to evaluate
combustion modifications for emission
reduction and efficiency enhancement
on petroleum process heaters are re-
ported. Objectives were to determine
nitric oxide (NO) emission reductions,
thermal efficiency changes, long-term
performance, and cost of both natural-
and forced-draft staged-combustion-
air modifications. Forced-draft staged-
combustion-air modifications had been
shown to be the most promising
combustion modification in previous
pilot-scale tests. The test unit was a
vertical, cylindrical, natural-draft crude
heater, and the test fuels were natural
gas, refinery gas, and a combination of
No. 6 oil and refinery gas. The unit had a
16 MW heat input capacity and was
capable of a maximum throughput of
108 mVh of crude oil (rated input 96
mVh).
A natural-draft staging air system,
capable of providing 50 percent of the
staging air for 100 percent refinery gas
firing and crude charge rates of 55 - 90
percent rated capacity, reduced NO
emissions by about 50 percent. The
stack Oa was lowered from 4 percent
baseline to 2 percent low NO condition.
The efficiency gain with the natural-
draft staging air system was about 1.5
percent. Natural-draft staged combus-
tion air (4 percent stack O2) with 80
percent oil/20 percent gas firing
provided NO reduction of 30 percent
and an average efficiency gain of 0.6
percent. Lowered excess air with
staging (4 to 2 percent Oz) for the
80/20 mix provided an NO reduction of
about 60 percent and an average
efficiency gain of 3.6 percent, but
unacceptable CO emissions and smok-
ing problems. While firing 80 percent
oil/20 percent gas, the forced-draft
staging air system provided about 40
percent of the total combustion air and
an NO reduction of 40 percent (6
percent baseline stack Oa to 3 percent
low NO stack O2). The efficiency gain
with the forced-draft staging air system
was 5 percent. A 15-day evaluation of
the natural-draft staged air system
revealed no special operating difficul-
ties or process constraints. Costs of the
natural- and forced-draft staged air
systems are compared. The natural-
draft system (staging with lowered
excess air) has a cost of $0.03/lb NO
on 29.3 MW natural-draft process
heater, compared to $0.32/lb NO for a
forced-draft system.
This Project Summary was developed
by EPA's Industrial Environmental
Research Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering infor-
mation at back).
Introduction
The test work reported here includes
results of combustion modifications for
NO control on a refinery process heater.
The crude heater tested is the same one
described in EPA Contract 68-02-2645.
The heater is a natural-draft, vertical,
cylindrical crude heater containing six
-------
burners, capable of gas and/or oil fuel
firing. Of the advanced combustion
modification concepts considered for NO
control, staged air lances were considered
a feasible approach. The test program has
two objectives with respect to staged air
lances: (1) to evaluate the potential of
natural-draft staged air lances for reduc-
ing NO emissions and increasing thermal
efficiency; and (2) to evaluate the
performance of a staged air system on a
heater firing 100 percent residual oil. The
heater could not be run on 100 percent oil
as originally hoped. The original plan was
to have a temperature controlled valve in
the oil line to automatically adjust oil and
gas flow rates to satisfy thermal input
requirements. Since the valve was never
installed, the crude oil outlet temperature
was automatically controlled by gas flow
and manually controlled by oil flow. For
safety and control, the plant could not run
the heater solely on manual control. The
staged air system has 24 ports (4 per
burner) through which either natural-
or forced-draft staged air may pass.
Table 1 summarizes the significant
results obtained during the pilot-scale
test program. Natural-draft staged com-
bustion air lowered NO emissions by
about 45 percent, while low excess air
decreased NO emissions by 20 percent
with 100 percent gas firing. The stack O2
was maintained at about 4 percent for the
staged air test and lowered from 4 to 2
percent for the low excess air test. In a
short term test, the combination of
natural-draft staged combustion air and
lowered excess air reduced NO emissions
by about 50 percent (4.1 to 2.6 percent
stack Oa) while decreasing gas consump-
tion about 1 -2 percent.
Forced-draft staged combustion air
with 80 percent oil/20 percent gas firing
(6 percent stack O2) lowered NO emis-
sions by about the same amount (20-25
percent) as low excess air with no staging
(6 to 4 percent 02). Lowered excess air
together with staged combustion air
reduced NO emissions by 40 percent (6 to
3 percent 02) on natural-draft vertical
cylindrical heater firing 80 percent oil/20
percent gas.
In addition to the testing described, a
15-day evaluation of the natural-draft
staged air system was made. Cost
analyses of the natural- and forced-draft
staged air systems were performed to
compare the two systems.
Test Heater Description
The test unit is a natural-draft, vertical,
cylindrical crude oil process heater, used
to supply a heated charge to a crude oil
distillation column. A maximum load of
108 mVh (16,250 bbl/d) could be sent
through the heater in two passes. The
rated capacity of the heater is 91.8 m3/h
(14,000 bbl/day).
The maximum firing rate of the heater
is 16.1 MW thermal input (55 x 106
Btu/hr). It is fired by six John Zink DBA-
22 natural-draft burners. The burners are
combination gas/oil burners rated at a
maximum of 2.68 MW(9.14x 106Btu/hr)
each with a turndown ratio of 3:1.
Although combination gas/oil burners
are used, some gas is always fired
because the unit is base loaded on oil fuel
and an automatic temperature controller
adjusts the gas fuel flow to maintain crude
oil outlet temperature.
Staged Air Systems
The forced draft system consisted of 24
vertical 316L stainless steel pipes of 3.18
cm (1-1/4 in.) diameter arranged four-
per-burner, 90 deg. apart. A 45-deg.
elbow on each pipe provided better
mixing across the flame. A fan supplied
air to the lances through a manifold and
flexible tubing. The lances could be varied
in height up to 1.2m (4 ft) from the floor of
the heater. Extensions for the lances
allowed staging heights up to 2.4 m (8 ft)
for oil firing tests.
For the natural-draft system, holes
were drilled through the heater floor so
that one end of the 4-in. pipe would be
flush against the heater floor and the
other end (threaded for pipe caps) would
protrude a few inches below the heater
floor. For the natural-draft staged air
tests, the pipe caps were removed and 1 -
1/4 in. lances, 3-in. lances, or the 4-in.
ports were used.
Emissions Test
Instrumentation
All emission measurement instruments
(see Table 2) were carried in a 12.8 x 2.4
m (42 x 8 ft) mobile laboratory trailer. The
Table 2. Emission Measurement Instrumentation
Species
Carbon Monoxide
Oxygen
Carbon Dioxide
Nitrogen Oxides
Particulates
Sulfur Dioxide
Particle Sizing
Smoke Spot
Opacity
Sulfur Oxides
Manufacturer
Beckman Instuments
Teledyne
Beckman Instruments
Thermo Electron Co.
Andersen Samplers. Inc.
DuPont Instruments
Andersen Samplers, Inc.
Bacharach
Measurement Method
IR Spectrometer
Polarographic
IR Spectrometer
Chemiluminescent
EPA Method 5 Train
UV Spectrometer
Cascade Impactor
ASTM 21 56-65
EPA Method 9
Goksoyr-Ross
Model
No.
865
326A
864
10A
EPA
400
Mark III
RCC
Table 1. Summary of Combustion Modification Tests on a Pilot-Scale Process Heater
Heat
Input
/WM/T
14.2
12.8
12.9
15.6
14.4
10.7
9.4
13.6
15.2
Fuel
fief.
Ref.
Ref.
Ref.
Ref.
80%
80%
80%
80%
Gas
Gas
Gas
Gas
Gas
Oil/20%
Oil/20%
Oil/20%
Oil/20%
Gas
Gas
Gas
Gas
Baseline
Oz.%
4
4.2
3.6
4
4.1
3.8
4.1
5.7
6.1
Combus.
Mod.
Oz.%
3.9
3.8
3.4
2.0
2.6
3.8
2.0
4.1
3
Baseline NO
ng/J
67
73
70
78
78
135
167
140
176
ppm dry
at 3% Oz
131
142
138
152
152
265
328
344
345
Percent NO Reduction
Staging from Baseline
Air %
40
33
8
b
45
41
__b
b
39
43
21
12
21
46
25
28
13
37
Change in Fuel
Consumption, % Combustion Modification3
-1.1
0
+1.8
-1.3
-1.4
-4.5
-2.4
-2.5
-6.0
ND (SCAj 4" ports
ND fSCAJ 3" lances
ND fSCA) 1-1/4" lances
LEA
ND ISCA + LEA) 4" ports
FD (SCA)
LEA (4% + 2% Oz)
LEA (5.7% + 4.1% Oz)
FD (SCA + LEA)
6% + 3% O2
aFD = forced draft, LEA - lowered excess air, ND = natural draft, and SCA = staged combustion air.
b/V0 staging.
-------
gaseous species measurements were
made with analyzers located in the trailer.
Baseline Conditions
Baseline conditions were 4 percent
stack Oa for 100 percent gas firing and 4 -
6 percent stack Os for 80 percent oil/20
percent gas firing. The 4 percent stack 02
for 100 percent gas firing is considered to
be a normal operating level by plant
personnel. The 4 percent stack 02
baseline for 80 percent oil/20 percent
gas firing was also considered to be a
viable starting point by plant pesonnel
and also would permit direct comparison
of NO emission data previously collected
at the 50 percent oil/50 percent gas firing
condition (Ref. EPA-600/7-83-022). Base-
line stack Oz of 4 percent was initially
achieved for several test days under
various thermal input rates and crude
charge rates. However, some test series
at the low and medium loads from
February 22, 1983, onward could not be
run at 4 percent baseline Oz without
making extreme stack damper and
register adjustments. It is not clear why
the 4 percent baseline Oz was more diffi-
cult to achieve on some test days.
Reasons may include the crude inlet
temperature change (which alters the fuel
input to maintain constant crude outlet
temperature) or the fuel composition ratio
(80 percent oil/20 percent gas) which
was not burning effectively at lower stack
02 levels. These baseline conditions
could be achieved solely with stack
damper and secondary air register
adjustments (i.e., no combustion modifi-
cations). The 4 percent stack Oz for 100
percent gas firing and 6 percent stack 02
for the 80/20 oil/gas mixture was
achievable under all process rates
studied (55-96 percent full load). The
baseline conditions were established to
determine the extent of NO reduction to 2
percent stack Ojon 100 percent gas firing
and typically 3 percent O2 on an 80/20
oil/gas mixture.
Results
Natural-Draft Staged Air
Three staged air injection configura-
tions were tested using 4-in. ports, 3-in.
pipe, and 1-1/4 in. pipe, respectively.
Load was maintained nominally at 80
percent rated capacity and stack 02 of 4
percent for testing with the three
configurations. The 4-in. ports provided
the most staging air: about 40 percent of
the total combustion air.
The 4-in. ports achieved the greatest NO
reduction. A reduction of about 45-50
percent was achieved from baseline
conditions (SCA only).
The burner equivalence ratio,
$B= (air/fuel)bumer
(air/fuellstoictuometric
was about 0.72 with the secondary air
registers 10 percent open. At this
minimum $>& NO emissions decreased
47 percent below the baseline of 131 ppm
dry at 3 percent 02 (67 ng/J). The 4-in.
ports are flush with the heater floor;
whereas, the pipes could be inserted
through the ports to introduce the staging
air higher above the heater floor.
Since the 4-in. ports had provided the
most staging and NO reduction of the
diameters tried, further testing with load
variations and Oz variations were con-
ducted with the 4-in. ports. The three
loads tested were 55, 80, and 90 percent
rated capacity with Oz variation from
about 2 - 6 percent stack Oz. Under these
conditions, the CO concentration levels
measured were minimal, and flame
impingement was not a problem.
Figure 1 shows the effect of staging
(decreased burner equivalence ratio) on
NO emissions. The crude charge rate for
this test series was 52.7 rnVhr (8000
bbl/day), 55 percent of rated capacity. For
each Oz level (2,4 percent), the burner
equivalence ratio was lowered in incre-
ments by closing the stack damper and
closing down the burner air registers to
50,33, or 10 percent open. The minimum
$B for any 02 level was that associated
with the registers 10 percent open. At 4
percent Oz (SCA only), the minimum 4>e
obtained was 0.63 (48 percent of total air
is staged air) which reduced NO emissions
48 percent below the baseline of 152
200
180
160
140
120
$ 100
to
I
E' 80
o'
60
40
20 h
1
Fuels: Absorber Gas
Process Rate: 13.1 kg/s (8000 bbl/d)
4" Ports, Natural-Draft Staged Combustion Air
(10/25-41
(10/25-7R)
(10/25-7)
(10.25/8) |
I (10/25-5)
• Baseline.
™ no staging
(all combustion air
through burners}
Stack Oa
approximately 4%
D2% Oa (Staging + Low Excess Air)
NO = 11.34 + (124.3 X&B)
r2 = 0.98
(~\ 4%0z (Staging)
V-X NO = 3.60 + (123.9 x <*>B)
r2 = 0.97
I
0.2 0.4 0.6 0.8 1.0
a. Burner Equivalence Ratio
1.2
1.4
1.6
Figure 1.
NO emissions at two different stack Oz concentrations as a function of burner
equivalence ratio.
-------
ppm, dry at 3 percent 02. At 2.3 percent
O2 (SCA + LEA from a stack 02 percent of
3.7), the minimum $B obtained was 0.58
(51 percent of total air is staged air) which
reduced NO emissions 51 percent below
the baseline.
Percent changes in fuel consumption
with NO reduction were calculated to
show the extent of fuel savings with
staged air combustion modification,
lowered excess air, and staged air
combined with lowered excess air.
Staged combustion air at 4 percent and 2
percent O2 (no lowered excess air)
showed consistent NO reduction rates 45
- 50 percent. Average fuel savings are on
the order of 0.5 - 1 percent. The fuel
savings become more significant (on the
order of 1 - 2 percent) with staged air
combined with lowered excess air.
Forced-Draft Staged Air
System
Tests were initially conducted with the
forced-draft staged-combustion-air sys-
tem to assess the levels of NO reduction
attainable at various staging heights and
burner equivalence ratios. Only 1-1/4 in.
lances were tested with the forced-draft
system. The optimum stage height for NO
reduction and minimal CO and smoke
was determined to be 4 ft with the lance
tips oriented toward the burner center-
line.
Figure 2 shows NO emissions as a
function of burner equivalence ratio for
two O2 levels. The staging height for
these tests was 1.2 m (4 ft) and the crude
charge rate was 76 m3/h (11,500
bbl/day). The $B was decreased in incre-
ments to its minimum value determined
by limitations of the staged air combus-
tion fan. At 4 percent O2, the minimum <|>B
(maximum staging) obtained was 0.71
which decreased NO emissions 24 per-
cent below the baseline of 265 ppm NO,
dry at 3 percent stack 02. About 40 per-
cent of the total combustion air was injec-
ted through the 24 1-1/4 in. lances for
the 4 percent stack 02 tests, while 29 per-
cent of the total combustion air was in-
jected through the lances for the 6 per-
cent stack 02 tests. The lower percent
staged air for the 6 percent stack O2 tests
is due in part to the higher fuel heat input
rate that day.
The efficiency gains were significant
with 80 percent oil/20 percent gas firing
and forced-draft staged combustion air.
The change in fuel consumption relative
to baseline when SCA was applied was
3.5 percent for 6 percent stack 02. Low-
ered excess air and forced-draft staged
combustion air reducesfuel consumption
by 6 percent.
Natural- Versus Forced-Draft
Staging
Figure 3 compares the effectiveness of
natural- with forced-draft staged air for
NO reduction only. In this instance, the
process rate was maintained at about
14,000 bbl/day and stack O2 maintained
at 4 percent for baseline and staging
conditions. The results show that, for
maximum staging (minimum 4>B), the 4-
in. ports (NDSCA) and 1-1/4 in. lances
(FDSCA) provide very comparable NO
reduction for about the same staging.
Note that the 1-1/4 in. lances of the
forced-draft system mixed the air and fuel
better, resulting in lower amounts of
smoke and CO compared with the 4-in.
ports of the natural-draft system.
Long Term Test (Natural-Draft
Staging with 3-in. Lances)
Figure 4 shows the baseline and
staging data of NO versus stack Oz.
Linear regression analyses were run for
both the baseline and staged-air data
points. The linear regression analysis
shows a very good correlation for the
staged air data (coefficient of determina-
tion, r2 = 0.9), while the coefficient of
determination is fairly good for the
baseline data (r2 = 0.78). Calculations
with the linear curve fit equations were
performed to determine the extent of NO
reduction with staging only (SCA) and
a combination of staging and low excess
air (SCE + LEA). The NO reduction from
baseline is about 15 percent with staging
500
450
400
350
6§" 300
250
t
a
200
750
700
50
Fuels- 80% #6 Oil, 20% Gas
Staging Height: 4 ft
Crude Charge Rate. J 1,500 bbl/day
(2/17-3J
(2/17-4)
(2/17-1,5)
v-' •• 12/14-1
(2/14-1)
(2/14-5)
(2/14-2)
(2/14-3,4)
^1 ^P Baseline, no staging
Q 4% 02
NO = 98.9 + (138.7 x B)
r* = 0.99
Q 6%O2
/VO= 157.5+ (135.8 xB)
I
I
0.4 0.8 1.2
cj>a. Burner Equivalence Ratio
1.6
2.0
Figure 2. NO emissions as a function of burner equivalence ratio for two stack Oi
concentrations.
-------
only, and about 28 percent with staging
and low excess air.
Cost Analysis of Staged
Combustion Air Applied to a
Natural-Draft Process Heater
Total annualized costs for 16.1, 29.3,
and 147 MW thermal input process
heaters were calculated for SCA, LEA,
and SCA + LEA. These total annualized
costs were then divided by the annual
N0< reduction potential (Table 3) to give a
dollar amount per metric ton of NOX
removed by the modification (Table 4).
The savings of natural-draft staged
combustion air on a natural-draft process
heater applied at normal stack 02 are
calculated as $714/Mg NOX reduction
for a 16.1 MW heater, increasing to
$964/Mg at 147 MW heat input. The
cost of natural-draft staged air combined
with lowered excess air on a natural-draft
crude heater is calculated as $336/Mg
NOx reduction for 16.1 MW unit and a
savings of $435/Mg NOX reduction for a
147 MW unit. The net cost at lower unit
heat rates with the combined modification
is due to an automatic stack draft
controller that permits operating the
heater at lowered Oa.
400
350
300
250
B
200
o
O
150
100
50
Fuels: 80% Oil/20% Gas
Crude Charge Rate: 14,000 bbl/day
Stack Oxygen: 4%
(2/28-4)
Baseline, no staging
LJ Forced Draft Staged Combustion Air
<1-14" lances)
NO = 128.3 + (143.2 x <*>„)
r' = 0.91
Natural Draft Staged Combustion Air
(4" ports)
I
0.3 0.6 0.9
a. Burner Equivalence Ratio
1.2
1.5
Figure 3. NO emissions as a function of burner equivalence ratio for natural- and forced-draft
staged air.
-------
150
120
I
30
Crude Charge Rate: 11,500 bbl/day /18.9 kg/s)
100% Gas Firing
I
Baseline (3/31-4/1)
MO = S2.0 + (8.6xO2.%)
r* = 0.79
8 data points (21 hours total)
3" Lances -Staging Air (3/1 6-3/30)
6 data points (113.3 hours total)
/
4
Stack Oz,%, Dry
Figure 4, NO versus O2 under baseline and staging conditions.
-------
Table 3. Annual NO Reduction Potential of Combustion Modifications for Three Heater Sizes
Modification/Test Series
Natural-Draft Staged Combustion Air,
700% Gas Firing (4% QJ / 10/8/82
Natural-Draft Staged Combustion Air
+ Lowered Excess Air, 100% Gas
Firing (4% + 2% Oz) / 10/27/82
Lowered Excess Air, 700% Gas Firing
(4% + 2% Oi) / 10/27/82
Lowered Excess Air, 50% #6 Oil/50% Gas
Firing (4. 1% + 1.8% Oif
Lowered Excess Air, 80% #6 Oil/20% Gas
Firing (4.5% + 1.8% Oz) / 2/1 1/83
Baseline
NO (ng/J)
67
73
67
114
176
ng/J
Reduction
30
40
10
32
60
% Reduction
from Baseline
45
55
IS
28
34
»vu ni
76.7 MW
10.7
14.2
3.7
11.4
18.1
29.3 MW
19.4
25.9
6.8
20.7
33.0
147 MW
97.7
129.7
34.1
103.8
165.4
"Mg/yr = megagrams per year.
"Tidona. R.J. et al., "Refinery Process Heater /V0« Reductions Using Staged Combustion A ir Lances," EPA -600/7-83-022 (NTIS No. PB83-193946).
March 1983.
Table 4. Cost Effectiveness Ratio of Combustion Modifications Applied to Three Natural-Draft Process Heater Sizes
161MW 29.3 MW
147 MW
Modification
Annual Cost
Total Reduc. Effec.
Annual/zed Costs Poten. Ratio
$ (Savings) Mg/yr' S/Mg
Annual Cost
Total Reduc. Effec.
Annual/zed Costs Poten. Ratio
S (Savings) Mg/yr' S/Mg
Annual Cost
Total Reduc. Effect.
Annual/zed Costs Poten. Ratio
$ (Savings) Mg/yr" $/Mg
Natural-Draft Staged Combustion Air
10O% Gas Firing (4% Oz/
Natural-Draft Staged Combustion Air
17.642)
4.775
10.7
14.2
(714)
336
(15.593)
1.770
19.4
25.9
(804)
68
(94. 146)
(56.475)
97.7
129.7
(964)
(435)
Lowered Excess Air. 10O% Gas
Firing (4% + 2%
Lowered Excess Air, 100% Gas Firing
Firing (4% + 2% Oil
Lowered Excess Air. 8O% #6 Oil/20%
Gas Firing (4.5% + 1.8% OJ
Lowered Excess Air, 50% #6 Oil/50%
Gas Firing (4. 1% + 1.8 % O2
872
(30.516)
(18.914)
3.7
18.1
11.4
218
(1,686)
(1.659)
(3.755)
(60.798)
(39.652)
6.8
33.0
20.7
(543)
(1,842)
(1.911)
(68.263)
(354,201)
(248.270)
34.1
165.4
103.8
(1.997)
(2.141)
(2,392)
'Mg/yr - megagrams per year.
R. C. Benson is with KVB, Inc., Irvine, CA 92714.
Robert E. Hall is the EPA Project Officer (see below).
The complete report, entitled "Evaluation of Natural- and Forced-Draft Staging Air
Systems for Nitric Oxide Reduction in Refinery Process Heaters," consists of
two volumes:
"Volume I. Technical Report," (Order No. PB 84-229 640; Cost $20.50,
subject to change).
"Volume II. Data Supplement," (Order No. PB 84-229 657; Cost 17.50,
subject to change).
The above reports will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC27711
. S. GOVERNMENT PRINTING OFFICE: 1984/759-102/10699
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
------- |