United States
                   Environmental Protection
                   Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
                   Research and Development
EPA/600/S7-86/026  Sept. 1986
f/EPA         Project  Summary
                   Prototype  Evaluation  of
                   Commercial  Second
                   Generation  Low-N0x  Burner
                   Performance and  Sulfur
                   Capture
                   R. A. Lisauskas and D. C. Itse
                    Pilot scale combustion tests were
                  conducted on a Riley Stoker second
                  generation Iow-N0x burner combined
                  with dry sorbent  injection for sulfur
                  dioxide (SO2) control. The burner de-
                  sign is based on the distributed mixing
                  concept. Combustion tests were con-
                  ducted at 100 x 106 Btu/hr (29 MW) in
                  EPA's Large Watertube Simulator
                  (LWS)  test furnace. Results were ob-
                  tained for three different U.S. coals and
                  two sorbents.
                    Nitrogen oxides  (NOX) were reduced
                  by up to 60% with this advanced burner
                  design. SO2 reductions of 50% at a Ca/S
                  ratio of 2 were obtained with hydrated
                  lime (Ca(OH)2). Highest reductions
                  were achieved when the hydroxide was
                  injected through tertiary air ports on
                  the periphery of the burner. When lime-
                  stone was used as the sorbent, SO2
                  capture was on the order of 35% at a
                  Ca/S of 2.
                    In order to aid the  scale-up of the
                  pilot scale results to utility and indus-
                  trial boilers, two commercial Riley
                  burners were also tested at two differ-
                  ent scales (100 and 50 x 106 Btu/hr). A
                  furnace heat release parameter was
                  used to extrapolate pilot scale NOX
                  emissions to operating field boilers. In
                  addition, the Riley burner test results
                  are compared with data from other
                  burners also tested in the LWS test fa-
                  cility.
                    This Project Summary was devel-
                  oped by EPA's Air and Energy Engineer-
ing Research Laboratory, Research Tri-
angle 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 and Objectives
  In recent years, the U.S. has turned
increasingly to coal to meet its energy
needs. Emission regulations for coal-
fired industrial and utility boilers are di-
rected toward  limitations on nitrogen
oxides (NOX), sulfur dioxide (S02), and
particulate matter. NOX and S02 are
believed to be two of the major acid-
forming precursor gases of acid precipi-
tation. Although there is still consider-
able scientific debate  over the
relationship between emissions and
acid deposition, the control of S02 from
power plants is the major focus of pro-
posed acid rain regulation strategies.
  As part of its effort to develop emis-
sion controls for industrial  and  utility
steam boilers, the U.S. EPA is develop-
ing Limestone Injection into Multistage
Burners (LIMB) technology as a poten-
tial low cost control technology for both
NOX and S02. The program described
here is one of several prototype-scale
test programs sponsored by the EPA to
evaluate the sulfur capture potential of
Iow-N0x burners combined with the in-
jection of conventional sorbents. Under
this program, five Riley Stoker burners
were tested in the EPA's Large Water-

-------
tube Simulator (LWS) furnace operated
by the Energy and Environmental Re-
search Corporation (EERC)  at their El
Toro, California, test facilities.
  The objectives of this program were
to (1) characterize NOX emissions for
each burner, (2) evaluate each burner
for combined NOX/S02 control with sor-
bent injection under acceptable operat-
ing conditions, (3) extrapolate the per-
formance of the burners to  field
conditions, and (4) compare the results
with other burner testing.

  The test  program was conducted in
two  phases. Phase 1 was performed on
a single  100 x 106 Btu/hr Riley Stoker
Distributed Mixing Burner (DMB). The
Riley DMB is a second generation com-
mercial scale low-NOx coal burner. Low-
NOX operation was first established for
the DMB during baseline tests. All of the
adjustable burner variables were inves-
tigated to achieve Iow-N0x emissions, a
wide operating range, acceptable flame
characteristics, and  combustion  effi-
ciency. Following these initial tests the
S02  reduction potential of  the burner
was  evaluated. Calcium-based sorbents
were injected through two different
burner passages: (1)the coal nozzle,
and  (2) the tertiary air ports. During the
Phase 1 testing, performance data were
obtained for three different fuels and
two  sorbents.

  Under  Phase 2, two commercial Riley
burner designs,  the Flare and  Con-
trolled Combustion Venturi (CCV)*
burners,  were evaluated in the  LWS.
The  Flare burner is a  conventional high
turbulence  burner, while the CCV is a
first  generation Iow-N0x burner. Both
the Flare and CCV designs were tested
at scales of 50 x 106 and 100 x 106 Btu/
hr. These commercial  burner tests at
different  experimental  scales provided
a  link between the  LWS results  with
other pilot  scale furnaces and with ac-
tual  field operation. The CCV burner
was  also tested under staged combus-
tion  conditions in a DMB configuration.
As in the DMB design, staging air was
supplied by four tertiary air ports on the
periphery of the burner.

  In  addition to burner characterization
and  NOX emissions tests,  limited sor-
bent injection tests were  conducted
with the Flare and CCV burners to evalu-
ate the effect of burner design on SO2
reduction potential. The multistage CCV
burner was evaluated on three different
sorbents.
Description of Experimental
Systems

Test Burners

Riley Stoker DMB
  The Riley Stoker Distributed Mixing
burner (DMB), shown  in Figure 1,  is
based on design criteria developed  in
previous U.S. EPA studies. It is a dual
register burner with secondary air en-
tering through separate concentric air
passages surrounding the coal nozzle.
Swirl is imparted through adjustable ra-
dial vanes at the entrance of each pas-
sage. Although these two air passages
would normally be  incorporated in a
common windbox, they were fed sepa-
rately for  this test program. This al-
lowed independent flow measurement
of each secondary air stream.
  The burner was equipped with four
tertiary air ports around the periphery of
the  burner for staged combustion.
These tertiary air ports are smaller and
somewhat nearer the burner in the Riley
Stoker design than  prescribed in the
EPA's DMB criteria.  These design
changes increased flame stability under
deeply staged combustion conditions.
The tertiary air ports were also
equipped  with inserts to further  in-
crease the tertiary air velocity and im-
prove downstream air/fuel mixing.
  In addition, the burner incorporates
the venturi coal nozzle design devel-
oped for the Riley  Stoker  Controlled
Combustion Venturi (CCV) burner. Both
coal spreader and nozzle setback posi-
tion were adjustable in the test burner.
        Tertiary
        Air Port
      Venturi
      Coal Nozzle
        Coal
        Spreader
CCV Burner
  The CCV burner, shown in Figure 2,
was  developed for retrofit into existing
coal  fired boilers. Secondary air is sup-
plied through a single annular flow pas-
sage and register for swirl control. The
burner employs a  four bladed spreader
and  venturi coal  nozzle. NOX  is con-
trolled  through controlled air/fuel mix-
ing. The coal spreader imparts swirl to
the primary coal air stream and divides
the stream into fuel-rich and -lean layers
before  mixing with the secondary  air.
Tests were conducted on both a conical
and  straight cylindrical spreader body
design.
  The CCV burner was also tested in a
multistage, or distributed mixing
burner, configuration. Staging air was
supplied by four tertiary air ports on the
periphery  of the burner, as in the DMB
design. The CCV burner was tested at
two  different sizes: 100 x 106 and
50 x 106 Btu/hr. The objective of these
tests was  to  investigate the effects of
heat input on NOX emissions.

Flare  Burner
  The  Flare burner, also illustrated in
Figure 2, is designed to  produce a
rapidly mixed, intense stable  flame.
This burner utilizes secondary air swirl
control and a multivane coal spreader
which  promotes  rapid mixing  of  the
coal stream with the secondary air. This
burner produces  high NOX emissions,
1.0-1.2 Ib  NOX/106 Btu, and  low carbon
loss in the flyash.  The Flare burner was
also tested  at 100 x 106  and 50 x 106
Btu/hr.
                                                          Coal/Primary Air
                                 Secondary Air
 Figure 1.   Riley Stoker Distributed Mixing Burner.

-------
                                           Adjustable
                                           Swirl Control
   Flare Burner
   Coal Nozzle
   and Spreader
                    Coal/Primary Air
                                    Secondary
                                       Air
   CCV Burner
   Coal Nozzle
   and Spreader
 Figure 2.   Riley Stoker CCV* and Flare burners (*Protected by U.S. Patent No. 4.479.442).
Test Facility

LWS
  All of the testing under this contract
was conducted in the EPA Large Water-
tube Simulator (LWS) furnace, designed
to simulate furnace conditions in utility
boilers. The furnace is fired with a single
burner mounted on the front wall. The
furnace gas exit is at the top on the rear
wall. With  this configuration,  the gas
flow pattern is similar to that in a wall-
fired boiler. The outer surface of the fur-
nace is cooled by water sprays and is
open to the atmosphere. In order to
maintain thermal similarity with  coal
fired boilers, the flame zone of the fur-
nace is refractory lined. The  furnace
wall near the burner, however, is uncov-
ered to ensure that the wall temperature
near the burner will be cool as in field
operation.

Sorbent Injection  System
  The sorbent  injection system  was de-
signed to feed into two locations: (1) the
primary air/coal stream after the pulver-
izer, and (2) the tertiary air  ports. Sor-
bent feed rate was controlled by a screw
feeder. The sorbent was  entrained by
the compressed air and conveyed to the
injection location. Sorbent was added
to the primary  air stream through a sin-
gle nozzle downstream of the pulver-
izer. For injection through  tertiary air
ports, the  sorbent was split into four
streams and was injected through noz-
zles located on the axis of each tertiary
air port.

Fuels and Sorbents
  Three coals were used during the test
program: Utah, Indiana, and Illinois.
Utah coal was chosen as the baseline
low (0.7%) sulfur coal  because it had
been used previously in Iow-N0x burner
tests supported by the EPA in the LWS.
Indiana coal has  also been  used  as a
medium (2.5%) sulfur base coal in other
EPA funded sorbent injection tests. The
high (3.5%) sulfur Illinois coal was se-
lected  in order to relate the tests to
other pilot- and full-scale burner devel-
opment tests conducted  by Riley
Stoker. All three coals were used during
the DMB tests. However, only the Utah
and Illinois coals were used for the CCV
and Flare burner tests.

  Three sorbents  were evaluated dur-
ing the program: limestone  (Vicron
45-3), hydrated lime, and dolomite. The
sorbents were purchased, pre-milled in
50 or 100 Ib (22.5  or 45 kg) sacks.
  Limestone and  hydrated lime were
used as sorbents  during the  DMB test.
Dolomite was tested  during the  CCV
and  Flare burner tests  in addition to
limestone and hydrated  lime.

Test Plan
  The test program was divided  into
two  phases:  (1)the DMB tests,  and
(2) the commercial burner tests. During
Phase 1, burner performance and S02
reduction potential were evaluated for
the Riley DMB. Burner performance in-
cludes NOX  emissions, CO emissions,
carbon content in the flyash, and flame
length.  Initially,  burner performance
was evaluated based on: coal spreader
position, coal nozzle position, swirl reg-
ister position, degree of staging, tertiary
air port  size, and coal spreader design.
  Following  these tests, the S02 reduc-
tion potential of burner sorbent injec-
tion was investigated at the optimum
burner configuration. Two burners were
evaluated in  Phase 2: the Riley CCV and
Flare burners. Field test data exist for
these commercial burner designs on a
variety of furnace sizes and configura-
tions. The prime objective of these tests
was to  vary furnace thermal environ-
ment  with a constant burner design.
These tests along with data contributed
by Riley Stoker provided a basis for pro-
jecting  NOX  emissions in  the LWS to
utility boilers.
   Sorbent  injection tests were  also
conducted in Phase 2 to evaluate the
SO2 reduction potential of LIMB  with
Riley's  present  commercial burner
products.
Test Results

A/OX Emissions

  NOX emissions for the 100 x 106 Btu/
hr test burner configurations are sum-
marized in Figure 3 for Utah coal. Test
results are  presented  as a function of
excess air for the final DMB design con-
figuration. The lowest NOX emissions
were achieved with the DMB, ranging
from 240 ppm firing Utah coal to 290
ppm firing Indiana coal at 20% excess
air. In comparison, the Flare burner pro-
duced NOX emissions of 675 to 700 ppm
at 20% excess air. This level of  emis-
sions is typical of boilers designed prior
to the New Source Performance Stand-
ards (NSPS), and is  representative of
uncontrolled NOX emissions.
  NOX emissions  from the  DMB were
less sensitive to excess air than from the
other burners. All of these data were
taken at  a burner zone stoichiometry
(SRB) of  0.7. CO emissions from the
DMB were only 20 to 40 ppm. The car-
bon content of the flyash, a measure of
combustion efficiency, was 6.2 to 10.1%
based on loss on ignition  (LOI). NOX
emissions from the Flare burner were
sensitive to  excess air, increasing 130
ppm for a 10% increase in excess air. CO

-------
     7000
      800
   0  600
   I
   of
   O
   .s
   I
400
   i
      200
        Burner, Spreader

           Flare
           CCV, Straight
           CCV. Cone
           CCV. Straight, S/?B = 0.7
           DMB.
                       /./              1.2

                           Overall Stoichiometric Ratio (SR-r)
                                                  1.3
                                                                 1.4
Figure 3.    NOt emissions from Riley Stoker burners firing Utah coal in the L WS.
emissions were less than 60 ppm while
the carbon  content of the flyash was
only 3.9 to 5.5%.
  The Riley CCV burner was tested in
two coal nozzle configurations while fir-
ing the  Utah coal:  an expanding cone
coal spreader and a straight center body
coal spreader. Testing during the origi-
nal CCV burner development program
showed that  spreader design can
change  NOX emissions by a factor of
two.
  The cone spreader  design was origi-
nally developed for the CCV burner and
has been installed  on three utility fur-
naces. Testing this spreader provided
data for extrapolating NOX emissions
from the LWS to the field. The straight
spreader was developed for  staged
combustion with the CCV burner.
  The performance of the Riley DMB in
the LWS is compared with various com-
mercial  burners in  Table 1. NOX emis-
                                sions from Utah coal for the DMB were
                                over 400 ppm lower than the  Flare
                                burner and 70 ppm lower than the CCV
                                burner  equipped with the cone
                                spreader. The flame length for the DMB
                                was the same as the baseline CCV
                                burner and longer than the Flare burner
                                flame. The combustion efficiency for
                                the unstaged CCV burner can be im-
                                proved with  the straight spreader de-
sign at the expense of NOX. Otherwise,
the combustion  efficiency  for each
burner design was comparable.
  The DMB could also be operated over
an excess air and  load range similar to
the Flare and CCV burners. Comparison
of the  performance indicates that the
DMB should  be retrofittable to units cur-
rently equipped with the Flare or CCV
burners.

Sulfur Capture
  Sulfur capture with the DMB is shown
in Figure 4 as percent S02 capture, for
the three coals, two sorbents, and three
injection locations. At a calcium to sul-
fur molar ratio (Ca/S) equal to 2, sulfur
capture ranged from an average of 32%
with limestone to 50% capture with hy-
drated  lime for each coal.
  Limestone with slightly more effec-
tive as a sorbent when injected through
the coal nozzle (35% capture) than when
injected through the tertiary air ports
(32% capture). Injection of  hydrated
lime through the  coal nozzle was not
evaluated in detail because it performed
poorly, achieving 27% SO2 capture dur-
ing screening tests.
  Coal composition, or more  specifi-
cally sulfur content,  was not a major
variable in sulfur capture with the DMB.
Figure 4 reveals no measurable effect of
coal composition on the degree of S02
reduction with any of the sorbent/injec-
tion combinations. Other researchers,
however, have found that coal composi-
tion can have a major effect on sulfur
capture. The sulfation reaction is
thought to be driven in part by the con-
centration of sulfur species in the flue
gas.
  The most favorable window for sulfur
capture has  been identified as 870 to
1230°C. At above 1230°C, sorbents tend
to dead burn, and  chemical equilibrium
prohibits SO2 adsorption if SO2 is less
than 2500 ppm. At below 870°C, SO2 ad-
sorption is too slow to  be significant.
Gas temperatures measured at the exit
                                 Table 1.   Performance of 100 x 106 Btu/Hr Riley Burners Firing Utah Coal in the LWS

Spreader
Minimum Excess Air, %
Minimum Load,
% Capacity
Flame Length, ft (m)
Carbon Utilization, %
Nox @ 0% O2, ppm
Flare
Flare
8
75
14-16 (4.3-4.9)
99.3
686
CCV
Unstaged
Cone
10
60
20 (6. 1)
97.9
304
Straight
10
70
15 (4.6)
99.1
465
CCV
SRg=0.7
Straight
17
20-21 (6.1-6.4)
99.1
274
DMB
SRB=0.7
Straight
10
60
20(6.1)
98.9
234

-------
     70
     60
     50
     40
  §
  r
     20
     JO
Coal Nozzle Injection
   of Limestone
           \       \      r

        Coal Type


        Indiana A ~ ~ ~~ ~ *

        Illinois  D	
                                                   Tertiary Air Port Injection
                                                       of Limestone
                                                Tertiary Air Port Injection
                                                   of Hydrated Lime

                                              j	1	I	L
                                        50      1      2      3      4     50

                                               Calcium to Sulfur Molar Ratio (Ca/S)
Figure 4.   SOz capture with the Riley Stoker DMB.
of the LWS with the DMB were in the
range of 900°C at low load and 930 to
1010°C at full load  (100 x 106  Btu/hr).
Peak temperatures as high as 1270°C
were measured within the flame at full
load. The effect of thermal environment
on S02 capture was investigated by re-
ducing the heat input to the  burner to
62 x 106 Btu/hr. As shown in Figure 5,
limestone injection S02 capture at low
load decreased for low Ca/S ratios, but
was comparable to the full load results
at Ca/S equal to 3.0. The opposite was
true for injection of hydrated lime: S02
capture at low load was comparable to
the full load SO2 capture, and  increased
less rapidly as Ca/S increased.
  Figure 6 compares sulfur capture for
all five of the Riley Stoker burners  using
limestone injected through the coal noz-
zle. At a Ca/S ratio of two, sulfur capture
ranged from 22 to 35%. The highest sul-
fur capture was with the CCV burner
staged to  SRB equal to 0.7. Sulfur cap-
ture with the DMB was slightly lower at
32%. The Flare burner achieved the low-
est sulfur capture.

  The effect of firing rate and burner
scale on sulfur capture for the unstaged
CCV burner with the cone spreader is
shown in Figure 7 for Illinois coal. Sulfur
capture using the 100 x 106 Btu/hr CCV
burner at full load is compared to that
for the 100 x 106 Btu/hr burner at part
load and the 50 x 106 Btu/hr CCV burner
at full  load. Sulfur  capture increased
50% from  28% capture to 42% capture
by firing the smaller burner in the LWS.
Sulfur capture was 38% at Ca/S equal to
2 for the 100 x 106  Btu/hr  fired at
58 x 106 Btu/hr.  This  may  be due to
higher peak temperatures  in the un-
staged burners.

Application to the Field
  Data gathered under this  program
were used in conjunction with data from
other combustion tests of Riley Stoker
burners to extrapolate LWS  NOX, emis-
sions to  operating  boilers. A Burner
Area Heat  Release  (BAHR)  parameter
was used to relate NOX emissions from
furnaces of various sizes and thermal
environments. BAHR ranks the relative
combustion intensity  of different fur-
naces, and is similar to parameters used
by  other  boiler manufacturers. The
BAHR is defined  as  the total gross fuel
input divided by the cooled surface in
the main flame zone.
  Full load NOX  emissions at 20% ex-
cess air are plotted  for various  burner
and furnace combinations in  Figure 8.
Data for the Flare burner and the CCV
burner with the cone spreader were
                                                             available from three test facilities of in-
                                                             terest and a number of field units.  In
                                                             addition to the EPA LWS, the test facili-
                                                             ties included the Riley Coal Burner Test
                                                             Facility (CBTF) and the EERC Medium
                                                             Tunnel (MT) furnace. These Flare and
                                                             CCV burner results were used as a basis
                                                             for extrapolating NOX emissions for the
                                                             other low-NOx burners. Unstaged CCV
                                                             burner data were available from several
                                                             utility boilers ranging from 360-400
                                                             MWe.
                                                               As shown in Figure 8, there is a linear
                                                             relationship between NOX and BAHR for
                                                             the Flare and CCV  burners. The slope  of
                                                             the line for NOX versus BAHR is nearly
                                                             identical for the two burners. Other low-
                                                             NOx burner performance has been ex-
                                                             trapolated parallel to  the CCV burner
                                                             with the cone  spreader correlation.
                                                             Using this thermal scaling criteria, DMB
                                                             NOX emissions would be about 400 ppm
                                                             in a large field  boiler  as compared  to
                                                             900 ppm for the Flare burner.

                                                               During these pilot scale tests, sulfur
                                                             capture  resulting  from injecting lime-
                                                             stone with the coal  ranged  from 22  to
                                                             35%.  The best  results were achieved
                                                             with the DMB and CCV burner, straight
                                                             spreader. Injection of hydrated lime
                                                             through the tertiary air  ports in the DMB
                                                             produced 50% S02 capture at Ca/S
                                                             equal  to 2. Injection of sorbents through

-------
    70
    50
    40
 O
 O
 to
 §  30
 0.
    20
    10
                         Full Load
                         Hydrated Lime
                         Through Tertiaries
                                                 Full Load
                                                 Vicron With
                                                 Coal
                                        / ^^ Full Load
                                      *         Vicron Through
                  Tertiaries
 Data at 62 Percent of Full Load

    Vicron Through Tertiary Ports

A  Vicron With Coal

D  Hydrated Lime Through Tertiary Ports
 Figure 5.
                    12345

                            Calcium to Sulfur Molar Ratio (Ca/S)

            Effect of thermal environment on SOz capture with the Ftiley DMB firing Utah coal.
the tertiary air ports of the CCV-MS
burner only produced 35 to 40% S02 re-
duction at Ca/S equal to 2.

  The injection of sorbents in the flame
zone of boilers presents problems to be
addressed prior to application to the
field. The  increased solids  loading in
bottom ash and flyash will have to be
accommodated. A  more serious prob-
lem may be the reduction of ash melting
temperatures and  thus, the increased
tendency toward slagging. For the LIMB
process to become a viable  control al-
ternative, there must be additional
understanding of the sorbent injection
process, particle interaction,  and the
controlling mechanisms of sulfation.

-------
          60
          50
          40
     I
     O    30

          20
          10
Burner, Spreader
Flare
CCV, Straight
CCV, Cone
CCV, Straight. SRB = 0. 7
O '
A '
O •
V '
                                          2             3
                                 Calcium to Sulfur Molar Ratio (Ca/SJ
Figure 6.    SOi capture with Riley Stoker burners firing Utah coal.

-------
         60

         50
         40
         3°
         20
         10
                                         Burner Size
                                        A   SOx10eBtu/hr
                                        O  100xWeBtu/hr
                                        D  100x10*Btu/hrat
                                          .  Low Load    .
Figure 7.
01              2345
                    Calcium to Sulfur Molar Ratio (Ca/S)
Effect of thermal environment on SOz capture during coal nozzle injection of
limestone with an unstaged CCV burner firing Illinois coal.

-------
    1200
    1000
     800
to
I
g'
.o
.§
I
d
     600
     400
     200
            O Flare
          '  A CCV Burner, Straight Spreader
            V CCV Burner, Cone Spreader
          .  O CCV-MS Burner
                                                               Flare -
                                                              CCV
                                                             Straight
                                  Open Symbols—EPA 68-02-3912
                                  Solid Symbols—Riley Stoker
                                  Open/Solid Symbols—EPRIRP21S4
             LWS LWS MT  CBTF
               I   .   1 I ,   .  I,
                                         360 MW,  400 MW,
                                         ,   ,1  .   .   \   .   ,
                                                                       1.4
                                                                      1.2
                                                                       1.0
                                                                       0.8
                                                                       0.6
                                                                       0.4
                                                                       0.2
                                                                             3
                                                                             DO
.1
                                                                           I

                                                                           i
         0            100             200            300            400

             Burner Area Heat fie/ease, W3 Btu/(hr-f?) (x 3.153 x 10'* W/crrf)

Figure 8.   Projected /VOX emissions from Riley Stoker burners.
   R. A. Lisauskas andD. C. Itse are with Riley Stoker Corp., Worcester, MA 01606.
   Charles C. Masser is the EPA Project Officer (see below).
   The  complete report,  entitled "Prototype Evaluation  of Commercial Second
     Generation Low-NO* Burner Performance and Sulfur Capture," (Order No. PB
     86-220 407'/AS; Cost: $22.95, 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:
           Air and Energy Engineering Research Laboratory
           U.S. Environmental Protection Agency
           Research Triangle Park. NC 27711

-------

D =!
Q) O
ff«
-  Cfl

II'
  
                                                         (0 <
                                                         3 =;
                                                         o o
                                                         •< 3


                                                          CD

                                                         il
                                                         01 2
                                                         2. o

                                                         0 =
                                                         X
                                                         01
                                                         N>
                                                         O)
                                                         00
O  M
ris a
oo
  »r
     »
  OOB»
  x>»o
  **•-*
 t-t»
  o
  CK
      «»
                                                   'O
                                                   "T"

                                                   o

-------