SEFA
                                United States
                                Environmental Protection
                                Agency
                                 Industrial Environmental Research
                                 Laboratory
                                 Research Triangle Park NC 27711
                                Research and Development
                                 EPA-600/S7-82-021  Sept. 1982
Project  Summary
                                Application of  Advanced
                                Combustion  Modifications to
                                Industrial Process
                                Subscale Test  Results
                                S. C. Hunter, W. A. Carter, R. J. Tidona, and H. J. Buening
                                  Results of subscale tests to evaluate
                                combustion modifications for emission
                                control on petroleum process heaters,
                                cement kilns, and steel furnaces are
                                reported. The objective was to assess
                                applicability, NO> emissions reduc-
                                tions, and cost effectiveness of several
                                modifications and to select the most
                                promising for pilot scale tests. Sub-
                                scale  process heater baseline  NO,
                                emissions were about 55 ng/J firing
                                natural gas at 2.9 MW heat input. NO,
                                was reduced by 67% with staged
                                combustion (SC) and by 63% with flue
                                gas recirculation (FGR). Firing No. 6
                                oil, baseline NO, of 160 ng/J was re-
                                duced by 51% with SC and by  39%
                                with FGR. SC was selected for  pilot
                                scale  tests. Subscale cement kiln
                                baseline  NO, emissions were 30-60
                                ng/J firing natural gas at about 80 kW
                                heat input. Fly ash, kiln dust water,
                                and sulfur were injected separately to
                                evaluate the NOX reduction potential.
                                Fly ash injection reduced NO, emis-
                                sions by 28%, while the other inject-
                                ants reduced NO, by 12-20%. Further
                                work on a larger scale is planned prior
                                to selecting modifications for pilot
                                scale tests. For the subscale steel fur-
                                nace, baseline NO, emissions of 115
                                ng/J firing natural gas at 0.6 MWheat
                                input were reduced by 88% with  FGR
                                and by 47% with water injection. Fir-
                                ing No. 2 oil, baseline NO, emissions
                                of 160 ng/J were reduced by 77%
                                with FGR and by 89% with steam in-
                                jection.
                                  This Project Summary was dove/oped
                                 by SPA's Industrial Environmental
                                 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 report documents the subscale
                                 activities of a program whose objective
                                 was to develop advanced combustion
                                 modification concepts requiring relatively
                                 minor hardware modifications that
                                 could be used by  operators and/or
                                 manufacturers of selected industrial
                                 process equipment to control emissions.
                                 The development was to be performed
                                 for equipment in which the modifications
                                 would be most widely applicable and of
                                 the most significance in mitigating the
                                 impact of stationary source emissions
                                 on the environment.  Modifications were
                                 sought which could  be readily adopted
                                 by fuel-burning-equipment manufac-
                                 turers. The path to  this goal included
                                 concept definitions, economic and
                                 technical assessment, subscale per-
                                 formance evaluation tests, cost/benefit
                                 analysis, full scale equipment modifica-
                                 tion or retrofit, full  scale performance
                                 evaluation tests, and preparation of
                                 final reports and instructional guidelines.
                                  Subscale testing  is a necessity for
                                 such process categories as petroleum
                                 process heaters where equipment

-------
operators are naturally  reluctant to
cooperate in a modification test program
until the principle has been demonstrated
on  a smaller scale. Subscale tests of
three industrial combustion devices
were conducted following  a nationwide
survey of industrial combustion equip-
ment. The survey ranked these devices
according to  NOX emissions and heat
input capacity. The equipment tested in
this program was chosen based  on
these rankings  and the modifications
tested were  determined  by process
constraints and cost considerations as
well as the potential for NO*reduction.
  The final report covers the test site
survey,  combustion modification con-
cept definition,  subscale  test results,
and cost analyses. The subscale test
sites included a process heater, rotary
kiln, and a steel furnace. Cost analyses
were performed for the process heater
and the steel furnace. Because the test
results  at the kiln were not deemed
representative of a full scale unit, a cost
analysis was not  performed for the
device.
Past Work
  The present program is a follow-on
study intended to build upon the results
of the program reported in EPA-600/7-
79-015a (Ref. 1). The objective of that
earlier effort was  to investigate the
effectiveness and applicability of  com-
bustion modifications  involving only
operating variable changes as means of
                             improvement in thermal efficiency and
                             for emissions control in industrial
                             combustion equipment.
                              The program scope provided for tests
                             on  22  industrial combustion  devices
                             representative of kilns, process furnaces,
                             boilers,  stationary engines, and gas
                             turbines in industrial  use.  Emissions
                             measured included  NO,  NOX, S02,
                             SOs,  CO, COz, Oz,  gaseous hydrocar-
                             bons, and where possible, particulates,
                             particle size distribution, smoke number,
                             and opacity. Combustion modifications
                             evaluated, where possible, included
                             lowered  excess air, staged combustion,
                             reduced air preheat, and burner register
                             adjustments. No hardware modifications
                             were attempted,  however. All experi-
                             ments involved only operating changes.

                             Summary of Results of
                             Present Program
                              The initial task of the present program
                             was to review existing source inventories
                             and update them where possible to
                             more clearly  define those processes
                             where controls would be of maximum
                             benefit. The review of source emission
                             data provided a relative ranking of each
                             candidate process.
                              The equipment recommended for
                             testing  included:  (1)  natural draft
                             process heaters, (2) forced draft process
                             heaters, (3)  cement  kilns, (4) steel
                             soaking  pits and  reheat furnaces, (5)
                             glass  container furnaces,  and (6)
                                                                               woodbark boilers. The  NO, emissions
                                                                               and  other characteristics  of  these
                                                                               industrial processes are presented in
                                                                               the final  report.

                                                                               Subscale Process Heater
                                                                                 Tests were conducted to evaluate the
                                                                               effect of  combustion modifications on
                                                                               emissions from a natural draft process
                                                                               heater. The reduction in NOX emissions
                                                                               and  the  change  in efficiency  were
                                                                               evaluated for these modifications: (1)
                                                                               lowered excess air, (2) staged combus-
                                                                               tion air, (3) Iow-N0x burners (tertiary air
                                                                               injection  and recirculating tile designs),
                                                                               (4) flue gas recirculation,  (5) steam
                                                                               injection, and (6) altered fuel injection
                                                                               geometry. The tests were conducted
                                                                               with  natural gas and No.  6 oil.  Only
                                                                               burner baseline  measurements were
                                                                               made with No.  2 oil. Fuel samples were
                                                                               taken for  all tests and the analyses were
                                                                               obtained from an independent laboratory.
                                                                                 Baseline  measurements were made
                                                                               prior to the implementation of combus-
                                                                               tion modifications with the burner firing
                                                                               natural gas. No. 6 oil, and No. 2 oil.
                                                                                 Table  1  shows that the  largest
                                                                               percentage reductions in NOX occurred
                                                                               with  staged combustion air (SCA) or flue
                                                                               gas  recirculation  (FGR) techniques
                                                                               when compared to conventional burner
                                                                               designs MA-16 and DBA-16 (a conven-
                                                                               tional burner differing only in tile design
                                                                               from the  MA-16  burner).  With SCA,
                                                                               these reductions seem to be a relatively
Table 1.
Summary of /VOX Reduction and Efficiency Change as a Function of Combustion Modification Technique for Natural
Gas and No. 6 Oil for Natural Draft Burners
                Fuel
                                                  Natural Gas
                                                                                No. 6 Oil
Average Baseline NO*
    MA-16
    DBA-16
                         ppm, dry @ 3% Oz
                                107
                                131
                                                               ng/J
                                                               54.6
                                                               66.8
ppm, dry @ 3% Oz
       285
ng/J
 16O
Combustion Modification Technique      NO* Reduction      Efficiency Change
                                                                               /VOX Reduction
                                                                                       Efficiency Change
Lowered Excess Air
Staged Combustion Air
    Floor Lances, Normal Oz
    Floor Lances, Low Oz
    Central Cylinder, Normal Oza
    Central Cylinder, Low Oz
Tertiary Air Burner, Lowest /VOX
Configuration (relative to average
baseline NO, for the MA-16)
Flue Gas Recirculation
    Normal Oz
    Low Oz
Steam Injection,  Normal Oz
Altered Fuel Injection Geometry
    Normal O2a
    Low Oz
                                27

                                46
                                67
                                31
                                59
                                30
                                59
                                63
                                33

                                31
                                44
                                                               +4.7

                                                               +0.7
                                                               +2.6
                                                                0.0
                                                               +3.4
                                                               •2.0
                                                               +4.7
                                                               +4.9
                                                                0.0
                                                               +3.4
       10

       35
       51
       42
       31
       39
+0.1

-0.7
-0.4
 0.0
 -2.6
+2.0
 ' NO* reduction is relative to average baseline NOX for the DBA-16.

-------
strong function of excess air; whereas,
with FGR they are a rather weak function
of excess air. With natural gas fuel, all
modifications (except the tertiary air
burner)  appeared to  increase furnace
efficiency. With  No.  6 oil, efficiency
decreased slightly with SCA and de-
creased with FGR, but increased when
FGR was coupled with low excess air.
  The cost effectiveness (CE)  of the
most effective NO« reduction techniques
is graphed versus heater size in Figures
1 and 2. As with other modifications,
the CE ratio decreases as the unit size
increases. Costs  are based  on 1978
dollars.

Subscale Rotary  Cement Kiln
  KVB completed a series of tests on a
subscale cement kiln. The cement kiln,
at a major cement industry association
                          facility, had a 13 cm (5 in.) ID, 30 cm (12
                          in.) OD, and was 4.6 m (15 ft) long. The
                          maximum kiln feed  rate was  only
                          0.0015 kg/s (12 Ib/hr), and the unit had
                          no air preheat capability.
                            All tests were conducted with natural
                          gas fuel. The combustion modifications
                          tested were: (1) sulfur addition either
                          with the fuel or with the feed, (2) water
                          injection  at  the burner, (3)  kiln  dust
                          injection at the  burner, and (4) fly ash
                          injection at the burner. The effects of
                          these modifications  on gaseous emis-
                          sions, kiln operating conditions (tem-
                          perature), and  clinker  quality were
                          studied.
                            Table 2 summarizes the NO, reduc-
                          tions obtained with  each of these
                          techniques. Essentially, the injection of
                          these materials had  little effect on
                          clinker quality. Excess air changes had
  10000 r
   100O
 o
 is
 -g
 $
 o
 I
 o
o
     100
      10
               I    I  I  I 1 Illl
                                I  I  I  11 ill
                                   Flue gas recirculation
Staged combustion
 air - floor lances
                      Staged combust/on air-
                          central cylinder
  I    III
ill
I    I   I  I I I III
                                 10
                                 Heater size, mw
                                             100
Figure 1.  Estimated cost as a function of heater size for three combustion
          modifications to natural draft process heaters firing natural gas only.
more significant effects on the clinker
with the clinker noticeably degraded at
oxygen contents of 0.5% or less.
  It  is important to note  that  the
baseline  NOX levels  observed  for the
subscale kiln « 100 ppm, dry at 3% 02)
were far lower than any observed by
KVB on full-scale kilns. The most likely
explanation is that ambient air was used
in all of the subscale tests. In an actual
kiln, air preheater temperatures of 144
K (1600°F) are not uncommon. In
addition, the high  surface-to-volume
ratio in the small kiln may have resulted
in greater heat losses from the flame
zone, thus lowering NO* production.
Also, the high gas-to-solids ratio limited
the effect of kiln feed nitrogen on the
NO, emissions.

Combustion Modifications to
a Steel Furnace
  The average baseline  NO, emission
for a steel furnace burner firing natural
gas and No. 2 oil is given in Table 3. The
maximum NO reductions obtained for
each modification  are summarized in
Table 4.
  The cost effectiveness of  these
modifications is plotted in Figures 3 and
4 against heater size for No. 2 oil fuel
and  natural gas, respectively. Flue gas
recirculation (FGR) with  heat recovery
capability has the  lowest cost  per Mg
N0« reduction for  heaters larger than
13.5 MW (46 x 106 Btu/hr). For natural
gas  fuel, FGR  with heat  recovery
becomes less expensive than water
injection  for heater sizes greater than
8.5 MW (29 x 106 Btu/hr), and it is less
expensive  than steam  injection at
heater sizes in excess of  60 MW (205 x
106 Btu/hr) as shown in Figure 4. Costs
for the steel furnace  modifications are
in 198O dollars.

Conclusions and
Recommendations
  From the data  obtained  and  the
analyses  made during this program, the
following conclusions and recommen-
dations may be made.
  1.  For a subscale natural draft
process heater, baseline NOX levels for
two  standard burners were 54.6-67.0
ng/J firing natural gas.  One standard
burner was found to emit  150 ng/J
firing No. 6 oil and  63 ng/J firing No. 2
oil.
  2.   Two low-NOx burner designs had
baseline  NO, emissions of  47.1-53.0
ng/J firing natural gas. Thus, the mean
NOx emission level  from these burners
was  about 18% lower than the mean

-------
  10000
    iooo
 §

 1

 o
 O
     ;oo
      ;o
i    \
                          i  i
\  i  i
                                    Flue gas recirculation
                                                          Normal 02
  Staged air -
  floor lances
                          Tertiary air burner
                                                          Low Oz
I    t  I t \ IUI
    i  i tint
                                  w
                                 Heater size, mw
                                            100
Figure 2.  Estimated cost as a function of heater size for two combustion
          modifications and for changeover to tertiary air burner in natural draft
          process heaters firing No. 6 fuel oil only.
Table 2.    Maximum NO* Reductions
           for Four Combustion Mod-
           ifications to  a Research
           Cement Kiln
Combustion
Modification
Sulfur Injection
Water Injection
Kiln Dust Injection
Fly Ash Injection
Maximum /VOX
Reduction, %
12-20
14
14
28
value for the two standard burners.
Firing No. 6  oil, one  low-NOx burner
design produced 149 ng/J, a reduction
of 7% below the standard burner. The
reduction of NOX due to the low-NOx
burner when firing No. 2 oil was only 2%
below the standard burner baseline.
                          3.  Combustion modification tech-
                        niques were effective in reducing NOx
                        emissions on a subscale process heater
                        firing either natural  gas or No. 6 oil.
                        Staged  combustion air (accomplished
                        by lances through the heater floor and
                        coupled with lowered excess air) was
                        the most effective technique, followed
                        by flue gas recirculation at either nor-
                        mal or reduced excess air.  When  prop-
                        erly adjusted and under reduced excess
                        air conditions, a low-NOx(tertiary air de-
                        sign) burner was also shown to be effec-
                        tive in lowering NO* emissions  firing
                        both  natural gas and No. 6 oil fuel.
                        Lowered excess air alone was not very
                        effective in reducing the NOx concentra-
                        tion when firing No. 6 oil.
                          4.  Modifications which  worked well
                        firing gas fuel but which were not tried
firing  oil because of  time or test
equipment limitations  included staged
combustion air using a central cylinder
above  the  primary air zone, steam
injection, and altered fuel  injection
geometry. Each of these modifications
reduced NOX emissions  by more than
30% below baseline and some may be
applicable to oil firing as well as gas
firing.
  5. Staged combustion air by means
of floor lances reduced the  NO» at a
normal operating excess air  level by
46% below baseline (54.6 ng/J) firing
natural gas fuel. At lowered 02 levels
the reduction was  as much as 67%
(natural gas fuel). At normal Oa condi-
tions the NO« reduction firing No. 6 oil
was 35% below baseline (160 ng/J),
and at reduced Qz,  the reduction
reached 51%. This staged air technique
was also the most cost-effective tech-
nique based on the data available. Costs
are predicted to be roughly $700/Mg of
NOx reduction for small heaters  (2.9
MW and below) firing  gas, dropping to
only $39/Mg of NOX reduction for large
heaters (147  MW and  above) firing oil.
  6. Cost calculations did not include
annual fuel costs or savings due to the
combustion  modifications because of
the unrealistic efficiency changes  that
were observed on the small-scale
heater. With staged  combustion air,
however, no large effect on efficiency is
foreseen except in cases where excess
air can be reduced. Then, efficiency can
be  expected to increase with  the
application  of staged air.
  7.  In all tests, the N0« levels
observed at  the pilot kiln were well
below those previously reported for full-
scale kilns indicating that the subscate
unit was not truly representative  of a
production kiln. Itisthoughtthattheuse
of ambient temperature combustion air
as well as high gas-to-solids ratio in the
small kiln  were the  primary factors
contributing to low-NOx emissions.
  8. Although the absolute NOx levels
may not have accurately reflected those
present in a full-scale kiln, the trends in
emission levels which  resulted from the
injection of sulfur, water, kiln dust, and
fly ash should still provide a basis for
determining the relative effectiveness
of each  injected material in reducing
NOx. With that in mind,  it was discovered
that fly  ash injection was  the  most
effective of the above  in reducing NOX
emissions. At normal operating Oa NOX
emissions were reduced by 2Q% of the
baseline value when  fly  ash was
injected  into the flame zone. Clinker
                                  4

-------
 Table3.    Average Baseline /VO, Emission, Subscale Steel Furnace Burner
                   (Including All Baseline Tests at Location 4)
Fuel
NG
No. 2
ppm*
222
277
/VOx
ng/J
114.6
153.4
Number
of Tests
11
8
Coefficient*
of Variation
0.19
0.23
"ppm corrected to 3% O& dry.
 " Coefficient of variation =Std deviation
                           Mean
Table 4.    Summary of Significant Test Results. Subscale Steel Furnace Burner
Test
Number
4/3-11
4/4-13
4/3-12
4/7-2
4/8-10
Fuel
NG
NG
NG
No. 2
No. 2
Combustion
Modification
Water Injection
FGR
FGR + Water Inj.
Steam Injection
FGR
Firing
Rate
%Cap.
100
100
100
100
100
02
%
2.2
2.0
1.8
2.1
2.0
% Reduction in NO
NO From
ppm3 Nearest Baseline
98
38
24
24
57
47
88
87
89
77
"NO corrected to 3% Oa dry.
   1OOOO
   5000
I
I
I
 to
 CO
    1000-
     500
      10
               !    I  1 I Mill	1   I  I  MIIH	1   I  MUM
                                             Fuel: No. 2 oil

                                               Steam injection
                                            Q Water injection
                                               FGR (no heat recovery)
                                            O
                                               FGR (with heat recovery)
                                               ll\ll     1   \  \ \\\\
                             10 13.5         50    100

                                Furnace size, mw
                                                                  500  1000
Figure 3.   Cost effectiveness as a function of furnace size for No. 2 fuel oil.
quality was actually slightly improved
during those tests. Injection of the other
materials reduced the NO* levels by 12-
20% with essentially no effect on pro-
duct quality.
  9.  Lowered excess air  « 1.5% Oa)
was not a practical NOX reduction
technique for the subscale kiln. Accom-
panying CO levels were high and clinker
quality was degraded. In general, it was
found, the  industry already maintains
the lowest practical oxygen levels in
most kilns (1.5-2.0% 02).
  10.  For the two fuels tested at the
subscale steel furnace, natural gas and
No. 2 oil, FGR  proved to  be the most
effective combustion modification.
Steam injection was effective firing No.
2 oil but was not tried firing natural gas
because of constraints  in the test
apparatus. Water injection  was less
effective firing natural gas than steam
injection was when firing oil.
   11.   FGR  of 20% applied to  the
subscale steel  furnace at normal Oa
levels reduced N0« emissions firing gas
fuel by 88% and firing No.  2 oil by 77%.
Cost analyses showed that FGR  is the
most attractive combustion modification
for larger furnaces (73.3 MW and up)  if
the waste heat of the recirculated flue
gas is recovered. However, FGR is not as
cost effective as  steam injection firing
No. 2 oil for heaters smaller than about
13.5 MW nor is it as cost effective as
water injection  firing natural gas for
heaters under 8.5 MW in size.
  Based on  the test  results on  the
subscale process heater,  KVB recom-
mends testing the staged air lances on a
pilot  scale process heater. The test
variables  should  include  injection
height and  pattern, the  burner  stoi-
chiometric  ratio,  the  excess air  level,
and system performance  as a function
of load. Natural gas, refinery gas,  and
residual  fuel oil  should  be  tested as
these are the fuels most commonly used
at refineries.
  Because of the very low NOX concen-
trations measured in the subscale
cement  kiln, KVB recommends further
subscale studies in its own laboratory. A
nonproducing nonrotating model should
be constructed and tests made in which
the level of air preheat is varied. Other
operating variables which should be
investigated include inside wall  tem-
perature, primary and secondary air
velocities,  and  primary  air vitiation
(variation of O2 in the  primary air
stream). Tests should be done for coal
firing and for natural gas firing.

-------
  Post-combustion-zone   mechanisms
of NOxformation as well as near-burner
mechanisms should be considered to
ensure that N Ox destroyed at the burner
end of the kiln will not be recreated at
the opposite end of the kiln. A full-scale
kiln should  be tested with the most
promising combustion  modification
techniques.
  Both steam injection and FGR should
be applied to a pilot-scale  steel furnace
based on the results of the subscale
tests reported here. Appropriate methods
to recover  the  waste  heat  of  the
recirculated flue gas stream  or  the
steam (such as  heat exchangers or
condensation heat recovery devices)
could  be used to minimize efficiency
losses due to these modifications.  The
possibility  of  doing this should be
investigated. Application of  steam
injection firing  gas fuel should be
investigated to verify that the technique
is effective for that fuel as well as for No.
2 oil.

Reference
1. Hunter, S. C. et al., "Application of
   Combustion Modifications to
   Industrial Combustion Equipment,"
   EPA-600/7-79-015a,  NTIS PB 294
   214, January 1979.
  /OOOO
   5000
 c
 o
   7000
I

    500
jj>  100
«fc
     50
I   I   |  I I 11II      I   I  TT I 11II     I   I   IT TTT
              I   III I Jill
                                              Fuel: Natural Gas
                                b
                                                 Steam injection
                                              U Water injection
                                              t \FGR (no heat recovery)
                                               jFGR (with heat recovery)-
                                  III!
I   I   I  I III
                       5I/O            50      100
                          8.5                60
                               Furnace size, mw
                                                   50O  1000
                                         Figure 4.  Cost effectiveness as a function of furnace size for natural gas fuel.
                                          S. C. Hunter, W. A. Carter, R. J. Tidona, and H. J. Buening are with KVB, Inc.,
                                            Irvine. CA 92714.
                                          Robert E. Hall is the EPA Project Officer (see below).
                                          The complete report, entitled "Application of Advanced Combustion Modifica-
                                            tions to Industrial Process Equipment: Subscale Test Results," (Order No.
                                            PB 82-239 310; Cost: $21.00, subject to change) will be available only from:
                                                  National Technical Information Service
                                                  5285 Port Royal Road
                                                  Springfield, VA 22161
                                                  Telephone:  703-487-4650
                                          The EPA Project Officer can be contacted at:
                                                  Industrial Environmental Research Laboratory
                                                  U.S. Environmental Protection Agency
                                                  Research Triangle Park, NC 27711
                                                                                        OUSGPO: 1982—559-092/0507

-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
        PS   0000329
        US  ENVIR  PROTECTION AGENCY
        KfcGION  5 LIBRARY
        230  S DEARBORN STREET
        CHICAGO IL  60604

-------