EPA
600/7
82-057A
PB83-113S29
Hitachi Zoson NOx FJue Gas Treatment Process
Volume 1. Pilot Plant Evaluation^
nitechi Shipbuilding and Engineering Co. Ltd.
Tokyo (Japc,n)
Prepared for
Industrial Environmental Research Lab.
Research Triangle Park, NC
Sep 82
Semsnorcs
tss&fsS Sssfes
-------
PfGII'i
f.nsT
T>P!Wr, GUIDE SHEET
LINE OF
TEXT
EPA-600/7-82-057a
September 1982
?nn-,pr.o
PB83-113829
L'VI.'O.W
Hir*' y, 1
Hitachi Zosen NOx Piue Gas Treatment
Process; Vol. 1. Pilot Plant Evaluation
by
CVs
i ^
! SHINGO TANAKA
t, —Hitachi Shipbuilding & Engineering Co.LtcU- jy*,
| "* j Palaceside Building
| 1-1, Hitotsubashi, 1-Chome
• ' Chiyoda-ku, Tokyo, Japan
RICHARD WIENER
| Contract No. 68-02-2675
EPA Project Officer:' J. DAVID MOBLEY
I ¦
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
i ,
Prepared for
U.S. Environmental Protection Agency
.Office of Research and Development
Washington, D.C. 20460
--j riw\r-!0>.'
BOTTOM OF
Ot'Tr or
NATIONAL TECHNICAL
. - INFORMATION SERVICE -
Bs*coucro sr
¦ -C11; TAIi i 5
YAM) IU.U-
LPA-287 (C.n i
I0-7G)
US DCPAftTUtRT Of COKMUCE
SffitKG/iUO. VA W1SI
-------
TECHNICAL REPORT DATA
(Pleetf read /turnerions on the rtrtnt btfort completing)
1 AEPORTNO 2.
EPA- 600/7-82-057a
3 RECIPIENT S ACCESSION NO.
PR8 3 1 1 382 9
« title anosubtitle Hitachi Zosen NOx Flue Gas
Treatment Process? Volume 1. Pilot Plant
Evaluation
5 REPORT OATE
SeDtember 1982
6. PERFORMING ORGANIZATION CODE
7 AUTHORIS)
Shingo Tanaka and Richard Wiener
8 PERFORMING ORGANIZATION REPORT NO
a performing organization name and aodress
Hitachi Shipbuilding & Engineering Co., Ltd.
1-1, Hitotsubashi, 1-Chome
Chiyoda-Ku, Tokyo, Japan
10 PROGRAM ELEMENT NO
11 CONTRACT/GRANT NO
'68-02-2675
12 SPONSORING AGENCY NAME AND AOORESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13 TYPE OF REPORT ANO PERIOD COVEREO
Final? 5/78-2/82
14 SPONSORING AGENCY COOE
EPA/600/13
is supplementary notes IERL-RTP project officer is J. David Mobley, Mail Drop
61, 919/541-2578.
is AesTRACTiphg report gives results of a pilot plant evaluation of the Hi-
tachi Zosen NOx flue gas treatment process. The project—evaluating sel-
ective catalytic reduction (SCR) of NOx on a coal-fired source—operated
for 1-1/2 years. A newly developed catalyst, NOXNON 600, was success-
fully applied at the 0.5 MW pilot plant on a slipstream from George Po-
wer Co.'s Plant Mitchell near Albany, GA: it demonstrated 90% NOx remo-
val efficiency for over 5600 hours. Tests to operate the catalyst out-
side of design specifications showed good tolerance to adverse situa-
tions. The project was the first demonstration and evaluation of NOx SCR
technology on a coal-fired source in the U.S. Project results indicate
that the process is a viable NOx control technology option; however,
some technical concerns remain before the technology can be considered
commercially available and demonstrated for coal-fired sources in the
U.S.
\
COLOR ILL JSTRAT.O;'S REPRODUCED
• IN BLACK AND WHITE
,7. KEY MORDS ANO DOCUMENT ANALYSIS ,
a DESCRIPTORS
b IDENTIFIERS/OPEN; ENOEO TERMS
c COS ATI held/Group
Pollution Coal
Nitrogen Oxides
Catalysis
Ammonia
Flue Gases
Evaluation
Pollution Control
Stationary Sources
Hitachi Zosen Pro-
cess
Selective Catalytic
Reduction
13B 21D
07B
07D
21B
14G
19 DISTRIBUTION STATEMENT
Release to Public
19 SECURITY CLASS (ThU KtporfJ
Unc'^sified
21 NO OF PAGES
246
JO SECURITY CLASS (This p*s*t
Unclassified
22 PPICe
CPA Form 2220*1 (!•?))
-------
NOTICE
This doo .ncn* has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade namas
or commercial products does not constitute endorse-
ment or recommendation for use.
ii
-------
TYPING GUIDE SHEET
CF'.Tri'
01 r„(-,F
FfPsr
Li'.E OF
tOf OF
, \"A
i i A
. ' K 1
— I P".'~
...J ' "! i-\ ''
-------
J I
T-_ —^ CONTENTS
I i I
'AESTRACT ! iii
I
| LIST OF FIGURES ! vi
, LIST OF TABLES ! vii
I EXECUTIVE SUMMARY Lx
j . J
1 - INTRODUCTION 1-1
-V*. . - -
THE PROBLEM OP NO CONTROL 1-1
METHODS TO CONTRofc NO EMISSIONS FROM
STATIONARY SOURCES 1-2
DEVELOPMENT HISTORY 1 1-4
SELECTIVE CATALYTIC REDUCTION DEMONSTRATION
PROGRAM 1-5
2 - PROCESS DESCRIPTION
2-1
OVERALL PROCESS SCHEME 2-2
APPLICATION OF PROCESS TO BOILERS 2-13
CATALYST DESCRIPTION ' 2-15
REACTOR DESIGN 1 2-18
I
I
3 - DEVELOPMENT HISTORY ' ...3-1
; |
4 - DEMONSTRATION PROGRAM 4-1
j SCOPE OF WORK ' 4-2
SCHEDULE : 4-3
j DESCRIPTION OF THE TEST PLANT 4-4
! CATALYST SPECIFICATIONS .. 4-7
\ CONTROL SYSTEM ' I...' 4-9
j HOST SITE : ! 4-10
, DATA COLLECTION 4-11
j PROGRAM MANAGEMENT ' 4-13
i EXPECTED RESULTS 1 4-14
! RELATING THE RESULTS TO A FULL-SCALE
OPERATION 4-15
-r
; L 1
? f iv
L'AFirn. 0 4*1 CO)
I ' V i l J w V 1 I 1 I
TYPIiJG GUIDL SHEET
-------
BEC.IN
rinsr
,'rnl Contents (Continued)
C!.M1 • !'! U'
r,vs
'OP 01
<\(irA
5-
pp-
- NOXNON 500 TESTS 6-1
sf C i IC. . j
PRELIMINARY OPERATIONS . 5-1
FIRST CHARGE : 5-2
SECOND CHARGE . 5-6
CATALYST EVALUATION . 5-10
i i
6 - NOXNON 600 TESTS [ .6-1
I j
INTRODOCTION .6-1
CATALYST LIFE TEST I .6-3
CATALYST PERFORMANCE TESTS .6-13
TRANSIENT TESTS L .6-27
-ef- -REGENERATION OF CATALYST .6-40
CHANGES IN PRESSURE DROP IN RELATION TO
NO REMOVAL EFFICIENCY .6~50
CATALYST INSPECTIONS .6-56
EVALUATION OF NOXNON 600 CATALYST .6-60
I J
7 - CONCLUSIONS .7-1
j
NO REMOVAL EFFICIENCY j. .7-1
AMMONIA SLIPPAGE .7-1
PRESSURE DROP ACROSS CATALYST BED .7-2
CONTROL SYSTEM :. .7-3
FLY ASH PROBLEMS .7-3
REGENERATION OF CATALYST ... .7-3
APPENDICES
I
A
b
C
D
E
F
G
H
EQUIPMENT DESCRIPTION ...
INSTRUMENTS AND CONTROLS
AMMONIA FEED MEASUREMENT
WET AMMONIA ANALYSIS ....
CORRECTION OF NH,/NO MOLE RATIO
ABBREVIATIONS ......V...'
CONVERSION FACTORS
LOG CHARTS - NOXNON 600 TESTS
A-l
B-l
C-l
D-l
E-l
F-J.
G-l
H-l
',1\ *
. s» ip.i
EPA ro n* Zi'jCA »<*
'MHl/IOJ.t* u. I < ' •
—I
.. -i . . ) 'i ,
1 . » T' u
TYPU.C GLJlDC SKCCT
-------
!r _
V"
( L-'l t(( O-
Trj? or
FIGURES
NUMBER
2-1.
! 2-2.
i
2-3
! * J*
i 2-4.
1-2-5.
2-6.
4-1.
4-2.
5-1.
5-2.
5-3.
5-4.
I
j 5-5.
j 5-6.
! 5-7.
j 5-8.
j 5-9.
! 5-10.
i
I
! 6-1.
i
I 6-2.
i
! 6-3.
| PAGE
Typical Flow Sheet of Commercial
System 2-3
Formation of Ammonium Sulfate
(Low Temperature Range) 2-11
Formation of Ammonium Sulfate
(High Temperature Range) 2-12
Typical Reactor Arrangement 2-14
Configuration of NOXNON 500 or 600
Series Catalyst 2-16
Typical Reactor Design ' (Elevation) .2-19
i
Process Flow Diagram of Pilot Plant 4-5
Photograph of the Pilot Plant 4-8
Operating Data with the First Charge of
NOXNON 500 j 5-3
Operating Data with the Second Charge of
NOXNON 500 5-7
Location of Test Pieces b-12
lest Equipment for Determining
Catalyst Activity 5-13
Evaluation of Used NOXNON 500
(First Charge) 5-14
Evaluation of Used NOXNON 500
(Second Charge) 5-17
Pore Size Distribution of NOXNON 500 5-20
Effect of Thermal Processing in Fly Ash 5-25
Microphotographs of Fly Ash 5-26
Photographs of Fly Ash After Thermal
Processing 5-27
\
Operating Data with NOXNON 600
(5,620 hours) 6-6
Pressure Drop Across Catalyst Layer
(Actual/Calculated) 6-7
Effect of Flue Gas Flow Rate on NO
Removal Efficiency 6-14
i
T
vi
f, .. t n'r! 7yj) ' : t
> i v i o. r i m » e> ¦%
TV,', tr.'im. sml'lt
-------
QcOIN
FIKST
U OF
tcxt
H £r>£
^igu
res
ttr6r-4..
6-5.
6-6.
6-7.
6-8.
6*9.
e^io ."
6-11.
6-12.
6-13.
6-14.
6-15.
6-16.
6-17.
5-18.
6-19.
6-20.
6-21.
6-22.
NUMBER
3-1
4-1
4-2
4-3
:5-l ""
5-2
6-1
6-2
6-3
I
I -r-
CENTt-n OF
PAGE
(Continued) |
i I
— 80 Percent Removal-Test .....6-17.
Effect of NHj/NOj. Ratio on NOx Removal
Efficiency and NH3 Slippage (3200 -
4500 hours) 4 .6-18
Effect of NHj/NOj. on NOx Removal
Efficiency and NHo Slippage (4 600 -
5600 hours) i 6-19
Effect of Reaction Temperature on
NOx Removal Efficiency 6-21
Effect of NOy Concentration on N0X
Removal Efficiency J 6-23
Effect of SOx Concentration on NOx
Removal Efficiency ....6-24
Typical SO2 to SOj Conversion 6-26
Plant Load Excursions 6-29
Emergency Shut-Off of Ammonia Feed 6-31
Cold Start-up Test ...J 6-33
Boiler Shut-Dovm and St-art-Up 6-35
Sudden Load Change . .. J 6-37
Operation Without SootlBlower 6-39
Influence of Water Washing on
Catalytic Activity {Virgin Catalyst) ....6-42
Influence of Water Washing on Catalytic
Activity (Used Catalyst) 6-43
Abrasion Test J 6-44
NOx Removal Efficiency;After Water
Washing 6-48
Effect of Flow Rate on Idpact/dPcalc 6-52
Effect of Pressure Drop on NOx
Removal Efficiency 4 6-55
top or
jA. If.'.uCF
-* - -\I1lA
1
TABLES
PAGE
LIST OF COMMERCIAL PLANTS BUILT BY
HITACHI ZOSEN .3-2
TYPICAL FLUE GAS COMPOSITION .4-11
TYPICAL COAL ANALYSIS I .4-11
J
DATA COLLECTED I .4-12
CHEMICAL ANALYSIS OF FLY ASH 5-22
FOOLING FACTOR OF FLY ASH ' .5-23
ANALYSIS OF SAMPLES OF; WASH WATER .6-45
SAMPLES OF WASH SOLUTION .6-47
ANALYSIS OF USED CATALYST .6-49
CPA Form ZTjO-4 (4 3D)
trm VtOU'jLY CIN lOHMilJI
vil
PAGt
J yy : it
\*;
* f t;
• i tom or
ACi AitEA.
N'SION
ADI IS
r u u*
'MATJONS
TYPING GUIDE SHEET
-------
BrGiN
Fi'ST
LWE OF
ir-.i
HERil SB-
h'PING GsJIDE SMEFf
cr\'T>«
0f PAfiE
0=?ED
EAD
CGif:
tC'lO\S
r.PE S! -
TOP Cf;
.-AGE
A'-iLA
JlCI® OWLE DGEMENTS_
i
This propect could not have been accomplished without the
enthusiastic support of Georgi* Power Company. The site that was
provided at Plant Mitchell proved to be ideal. The cooperation
of the Georgia Power Company personnel both at Atlanta and
i J
Albany is gratefully acknowledged. Mr. W. E. Ehrenspercer,
Senior Vice President, and his staff were particularly supportive
jlfr addition,-Mr. Harry ~Mahe~ra¥, Mr"." Cam "Daniel/and Mr 7 Jim"
Lightfoot went out of their way to be helpful. The constructive
j 1
advice and comments from Dr. W. B. Harrison, Mr. Randall Rush,
and Mr. David Burford of Southern Company Services was ?lso
appreciated. J
EPA also;wishes to acknowledge the financial support to
jthis project of the U. S. Department of Energy. Without their
esources, this project could not have been completed as planned
'he DOE project monitors were Mr!. Bill Fedarko, Mr. Ed Trexler,
and Mr. John Williams. They are with DOE's Fossil Energy Office,
V
fPA ^0*7 IC11« t
» t «'t»)
L_
i
__ viii
' t L
union or
, 'V.AGE "'v a
Ourpip?
—| n!'.,rt-Air-\!
i'O'I TAGI i S
Nj-ANO HI
, TRA7i0.h.S
-------
BEGIN
FiHa"
U.Mt OF
T7XT
rii'RI: JS-
TYPIMG GUIDE SHEET
C L l* '. *
Ci :'-C:
D-'OfPi'D
I ir AD
EEGI'°'I
f ir. "HNS
H LPE rgf
1
vjv o'
V*',E
i
ULC !• '
la',: 1 ini.L
or ir.xr r»-
: EXECUTIVE SUMMARY - - —
i
i i
1 INTRODUCTION
I
: I
As part of the effort to assess technology for
control of NOx emissions, EPA has sponsored the design,
construction and testing of a pilot-scale unit {0.5 MW
equivalent) which demonstrated the operation of Hitachi
^.Zosen's. process on flue gas_from a coal-fired boiler.-
This report presents details of the Hitachi Zosen process,
designed to limit NO^ emissions from coal-fired steam
generators, and results of the demonstration program.
This Flue Gas Treatment (FGT) process uses selective
catalytic reduction (SCR) of NOx with ammonia which can
achieve over 90 percent reduction in N0x emissions.
Strict air pollution laws in Japan led to the con-
struction of several full-scale systems for the removal of
N°x from flue gas. Hitachi Zosen took the lead in this area
with the construction of the first large commercial unit in
1974. !
To further the goal of controlling stationary source NO
emissions, EPA has sought to enhance the reliability and
effectiveness of technology to reduce these emissions. One
aspect of EPA's involvement includes sponsoring programs
designed Lo demonstrate this technology. Because combustion
modifications are capable of achieving only limited reduction
in N0X emissions, some emphasis has been placed on developing
flue gas treatment processes. In particular, SCR technology
appears to be a very promising method of reducing stationary
i
T
1
7
t
ix
U'A if,.,
('t
-------
source N0x emissions by over 90 percent. Consequently, the
EPA has acted to demonstrate some of the more advanced SCR
8ysterns.
EPA initiated programs to demonstrate two SCR processes
on ar. .0.5 MW scale. The processes are:
1) The Shell-OOP Simultaneous SOx/NOx Removal System
2) The Hitachi Zosen NOx Removal System
The EPA sponsored programs demonstrated these processes on
flue gas from coal-fired boilers. The demonstration programs
were expected to answer many of'the questions which remain
concerning the application of SCR technology. In addition,
these programs should provide an improved basis for estimating
the costs of applying SCR technology.
PROCESS DESCRIPTION
A method of removing nitrogen oxides (FOx) from
flue gas is through the use of selective catalytic reduction.
Ammonia is the nost practical reductant for this purpose be-
cause it reacts selectively and quantitatively with NOx to
produce innocuous nitrogen and water. Ammonia is available,
relatively inexpensive, safe to handle, and easy to store.
Ammonia will react with NOx without a catalyst in a
narrow temperature range at around 1000°C. By using a suitable
catalyst the required temperature can be lowered to a more
practical level of between 300-450°C.
Carrier-based catalysts have been developed and used by
Hitachi Zosen in several plants in Japan. These are pellets
of alumina, silica, titania, or other materials"into which the'
catalyst is impregnated.
With dust-containing gases, carrier-based catalysts are
not practical because the bed of catalyst will be plugged by
the particulate matter in a short time. However, parallel-flow
honeycomb catalysts have been developed and applied to the
X
rvpi tn Cu'ur. f hi.lt
-------
removal of NO^ from dusty gases. These catalysts permit lower -
operating costs because of their low pressure drops.
The honeycomb catalysts developed by Hitachi Zosen have
been extensively tested in several pilot plants and were used
in this test program. They are termed NOXNON 500 and NOXNON 60
Overall Process Scheme
The process consists of the injection of a small amount
of ammonia into the flue gas and passage of the flue gas over
a catalyst. The ammonia reacts almost entirely with the
nitrogen oxides in the flue gas to form small quantities of
nitrogen gas and water vapor, both of which are normal con-
stituents of the atmosphere and are environmentally accepta-
ble .
Ammonia is injected into the flue gas from a boiler
between the economizer and the air preheater (see Figure 1).
At this point the gas temperature is about 400°C which is
suitable for the catalytic reduction of N0X. This gas enters
the reactor, passes over the catalyst, and then reactions
proceed. The flue gas then passes through the usual air
heater, particulate control, SC>2 control fan and stack. The
ammonia for the reaction is vaporized rfith steam and is di-
luted with air (or steam) before injection into the duct. If
the flue gas temperature is too low for optimum removal effi-
ciencies, additional heat can be added by auxiliary burners,
or economizer bypass.
Chemistry —
The exact relations between ammonia and nitrogen oxide3
are not completely understood. However, certain reactions are
probably involved which may or may not include oxygen. Vir-
tually all of the NOx in combustion gas is present as NCT so
xi
-------
TEMPERATURE
STEAM
CR AIR *"]
"I RECORDER/
CONTROLLER
NOX
RECORDER
SOLENOIO
VALVE
VENT-
PRESSURE
RECOROER
CATALYST
TEMPERATURE
RECORDER
NOX
RECORDER
BOILER
REACTOR
AIR
AIR
PREHEATER
TO SOX AND PARTICULATE
CONTROL SYSTEMS
AIR
FLOW INDICATOR/
^ CONTROLLER ^
SOLENOID)
VALVE |
AMMONIA-
BOILER
LOAD SIGNAL
COMPRESSED
AIR "
FLOW INDICATOR/
CONTROLLER
Figure 1. Typical commercial system flow diagram.
xii
-------
UcGW
rmsT
l "ZO
Tl VI M
hem:
or
1-1. E
;_the following equations are only for NO reactions:
4 NH3 ~ 4 NO + 02
4 NH3 +' 6 NO
4 N2 + 6 H20 (1)
5 N2 + 6 fi20 (2)
Or ' »
t - " i,\ T5->;
" 7 J: ¦ :
If N02 is also present, the following equations represent
reactions which may also occur:
(3)
(4)
4 NH3 + 2 N02 + 02
3 N2 + 6 H20
8 NH3 + 6 N02
2 NH, ~ NO + NO.
7 N2 + 12 H20
2 N2 + 3 H20
(5)
The reactions of ammonia with N0X over the catalyst occur
below 300°C. Without a catalyst the reaction will only occur
in a narrow temperature range of 950-1000°C. Below this
temperature the reaction rate is very low.
Factors Affecting the Catalytic Reduction of N0X —
; '
In any chemical reaction, there are factors which can
influence the rate and extent of the reaction. These factors
include the reaction temperature, concentrations of reactant
species, and other parameters specific to the reaction system.
The following discussion briefly examines the major influences
on the catalytic reduction of N0x by ammonia.
The most critical variables which affect the degree of
NO removal are the mole ratio of ammonia to NO , the flue gas
a X
flow rate, and the reactor temperature.
Ammonia Emmissions — !
: • i
i
A small amount of ammonia will invariably pass through •
j the reactor and exit with the flue gas due to incomplete j
j reactions between the NO and NH-. Ammonia slippage (unreacted
t X J
i ammonia) is of concern and efforts are normally made to control '
i
j slippage to levels of 10 to 20 ppm. A high slippage of ammonia
j can be considered a pollutanL and, in some cases in Japan, '
j agreements with local governments have set allowable levels at
i V xiii
( > I
J For T
I1 i 1 V '1 l ' * t. . t i \ t ii 1 .
I V.v •• G GU! 0!1 i>HL FT
-------
uteri
•i
around 10 ppm. Another reason ^for the concern with ammonia
slippage is that flue gas desulfurization following the de-
I i
nitrification system could absorb any ammonia in the flue gas
and a build-up of ammonium salts can occur. Bleeding of the
ir.ammonium salts from the system can cause water pollution
| control problems in certain cases. The degree of slippage is
i a function of the NH3/NOx mole ratio, the area velocity, and
the temperature at which the reaction occurs. For commercial
application to coal-fired utility boilers, ammonia slippage
would be expected to be less than 5 ppm at an NOx removal
efficiency of 80%, and less than 10 ppm at an NOx removal
efficiency of 90%. I
i oL' or
-\
Ammonium Sulfate/Bisulfate —
The presence of sulfur trioxide (SO^) in the flue gas can
lead to a reaction with ammonia' to form ammonium sulfate and
i
ammonium bisulfate. When burning heavy fuel oil, approxi-
mately 2-4 percent of the sulfur oxides in the flue gas are
present as the trioxide. For coal combustion, S03 accounts
for approximately one percent of the total SOx. The reaction
of ammonia with this sulfur trioxide will not occur above
approximately 300°C (572°F). At very low concentrations a
temperature of around 200°C (392°F) might be sufficient to
avoid this formation. |
i
To avoid such deposits it is necessary that the tem-
perature in the reactor be maintained at a minimum of 320°C
(608°F) as long as ammonia is injected into the reactor.
The formation of bisulfate can be minimized to a large
extent by maintaining a low ammonia level in the flue gas
exiting the reactor. The presence of fly ash also reduces the
problem both by scouring the deposics off of surfaces and by
providing surfaces on which the deposits will form instead of
forming on metallic surfaces. 1
I
xiv
V
*
» * .-n / ViO 4 «¦
1 • ~.» l1 * i v i [ i r , r .i»i a
7VPl C Mor : H*' LT
-------
"• or
i •' r or
f«.*T
M > _
IM
^ M
A schematic flow diagram of the HZ process as applied to
I a boiler is shown in Figure 1. j Flue gas leaving the economizer
^,~at a temperature of about 390 °C (734°F) is first mixed with
j ammonia in quantities needed to meet NO removal requirements.
I x
; The gas is then passed through the fixed bed catalyst reactor.
| No dust removal is required prior to the catalyst because of
I the non-clogging design of the catalyst geometry. Nitrogen
j oxides are reduced by ammonia to innocuous nitrogen and water.
i The denitrified gas then resumes its passage through the
j normal boiler train: air preheater, dust collection, etc.
j Ammonia required for the process is first vaporized with
| steam and is then diluted with either air or steam to aid in
the distribution of ammonia into the flue gas.
I i
The reactor is supplied with retractable soot blowers in
j which blasts of steam or hot air are occasionally applied to
j the catalyst bed to remove deposits of dust which may adhere
j to the surface of the catalyst.1
Catalyst Description
A particularly effective physical design of the catalyst
structure has been developed by, Hitachi Zosen. This structure j
1 i
is honeycomb shaped as shown in Figure 2. Overall, the catalyst
has the following characteristics:
(1) The structure is a thin plate honeycomb.
(2) Due to substantially reduced pressure drop across
the catalyst layer, operating power costs are much
lower than with conventional catalysts.
(3) A straight gas flow path prevents dcfst clogging.
(4) It is applicable for gases with high1 S02 concen-
TCM'Or
¦lACf.
*
trations.
V-
i
XV
-1,
I
V.
_ I •
I'
rvprjr, GUIDL" SHEET
-------
Figure 2. Configuration of NOXNON 500 or 600 series catalyst.
xvi
-------
c&.r-EPOF
p^lc
1 { I lU'.'j
Mt
L_ As a result, in treating high-temperature gases with high j
SO and dust'concentrations such as coal-fired boiler flue 1
X j |
gas, the NO^'removal reactor can be installed immediately
downstream of the econonizer. In addition, dust elimination
c^or other pretreatnent is unecessary and, with the low pressure
drop, operating costs are low. j
The honeycomb catalysts are referred to by Hitachi Zosen
as NOXNON 500 or NOXNON 600 Series. These catalysts are manu-
i l
factured in the form of plates and are fabricated so as to
I ]
form parallel flow gas passagesjwhich provide excellent con-
tact with the flue gas with minimum impingement of fly ash on
the catalyst[ surface. J
The catalyst plates are arranged in a steel frame box
- . — — - —^ - - - - - - _ - - - 5*
supported byi retainers. A standard module is 1.0 meters long,
1.0 meters wide, and 0.5 meters( deep.
The activation of the catalyst follows after the corru-
gated catalyst assembly is made. In the NOXNON 500 catalyst,
thin stainless steel plates are^ used.
A newer, development is NOXNON 600 catalyst which, instead
of plates, uses a stainless steel wire mesh as a base metal to
qive mechanical strength upon which catalytic components are
I 1
cemented. The NOXHON 600 is considerably lighter in weight
and contains' more active material for a given volume of cata-
lyst. [ 1
The active components of the catalyst consists of vanadium!
and titanium compounds. Other components are added to increase J
resistance to fly ash abrasion..
The catalyst is designed with an expected life of approxi-
mately two years. Regeneration of the catalyst is not needed
during the planned catalyst life. At the end of the useful
life of the catalyst it would be removed from the reactor
vessel and scrapped. The catalyst does not contain any haz-
ardous materials and can be disposed of by recovering the
metal in the catalyst or by disposing of the spent catalyst as
top of
IMAGfc
\
A
I
Y
V
xvii
fPA F nrm '«• P0|
(. » % rik
rvPH'JG GUIDE SHFET
-------
OFC.kJ
t insT
Li';' O-
if XT
f'Erit
CfMTt'l 01
P,iDt
| , '.Ml —
11'. -'O
lindustrial waste,
disposal process.
I
!
Reactor Design
No special precautions are required in the
p.--
Reactor1 vessels housing the catalyst are of carbon steel
construction' and are divided into chambers each having its own
fly ash hopper (see Figure 3). The reactor is oriented so
i I
that flue gas flows downward. Flue gas usually enters at the
top of the reactor at one side,' passes through the gas dis-
i 1
tribution grid and catalyst , ar.2 exits the onoosite side
! i
of the reactor at the bottom. Fly ash that drops out of the
flue gas stream is collected in' the bottom fly ash hoppers and
- - -v-f *¦ — - - i
is disposed of periodically. |
i
DEMONSTRATION PROGRAM
9 1
I
A potential market for the denitrification of
power plant flue gases may be at coal-fired stations.
There has been limited test work done by Hitachi Zosen on flue
i 1
gas from coal-combustion. The EPA-sponsored demonstration
program at Pliant Mitchell of the Georgia Power Company pro-
vided an excellent site for testing of the Hitachi Zosen
process. The power plant burns' typically medium sulfur coal
i I
with relatively high ash levels'. Adverse effects, if any, on
the catalyst: by this flue gas could be readily evaluated.
The demonstration plant was highly instrumented and
provided much useful data to supplement data collection from
prevlous~work in Japan. Several parameters were evaluated and
their effects on removal efficiencies and ammonia utilization
were checked. The data were collected and evaluated to es-
tablish valid characterizations of the procesii. Long term
tests were conducted to provide information on the aging
(¦—
si-
TCP or
IMAGE
-a'K-V
w
J
xviii
* t y
A r i»i r, 'J CO I
LViotj * l. i i n i i tJirr.t
\ ".t ¦
<. A *.r>
* Am i n
) II ! • '>
~ Fh
TYWMG CUIDF SHfeET
-------
GAS DISTRIBUTION
GRIDS
REACTOR INLET
CATALYST
AMMONIA FEED
CONNECTIONS
SOOT BLOWER
HOPPERS
TREATED FLUE GAS TO
BOILER PREHEATER
Figure 3. Typical Reactor Arrangement.
xix
-------
•¦"¦rv. r
I A
tendencies of the catalyst when applied to flue gas from
American coal burning boilers. .Additionally, the data
collected during these tests were useful in establishing
operating and capital cost requirements for commercial
"installations. . .
Hitachi Zosen, with Chemico Air Pollution Control
Corporation, Division of Envirotech Corporation (CAPCC) as
their major subcontractor, provided the test plant and
operated the unit. This included detailed engineering,
procurement, fabrication, transportation, erection, test
operation and continuous demonstration operation of the
pilot plant. The work was performed in four phases:
_ Phase I (engineering) started with the basic design
package which was designed by Hitachi Zosen in Japan. Based
on these designs, CAPCC prepared a detailed design.
Phase II included startup, debugging, and parametric
tests. j
Following the successful completion of the system
optimization' tests, Phase III, the pilot plant was to be
continuously operated (24 hours/day, 7 days/week) for at
least three months as Phase IV. The period of continuous
operation was to include no less than 75 days of cumulative
operation in compliance with contract objectives and guaran-
tees. !
An additional series of tests were conducted as an
addition to the original scope of work. Phase V wa3 run to
examine the response of the SCR system's performance to
transient operating conditions.
[
Schedule
The period of performance for completion of the work
related to the pilot plant was originally eighteen (18)
months. The award of the contract wa« May 1978. The design
xx
i , i <
-------
TQ\ t
'*"_9nd specification period, Phase I, took approximately three
I ¦
months. Phase II required approximately nine months for
procurement, fabrication, transportation and erection.
These schedules were as originally expected. However,
f
r7 startup and debugging required about two months and so the
1 first tests were not started until August 1, 1979. This
1 first charge of catalyst showed less than optimum results
J after some four months of operation and it was decided to
; replace it in December of 1979. After some four additional
j months of operation, once again', the NOXNON 500 catalyst
| began to provide less than expected results and a decision
j was made to install a third charge of catalyst.
The third charge was a new type: NOXNON 600. This was
i installed April 18, 1980 and was utilized for some nine
I '
! months until the plant was shut' down on February 2, 1981.
! ; ;
Description of the Test Plant '
i . 1
I i
| The following is a description of the pilot unit in-
I
j eluding summaries of major items of equipment (see Figure
I 4): I
I
j Flue gas to be used as input to the pilot plant was
J
' drawn from the boiler duct downstream of the economizer and
; introduced to the reactor through a fourteeninch diameter
i pipeline. i
! An electric heater was provided in the piping between
: the flues and the reactor. This flue gas heater was used to
• control flue gas temperature to the reactor at the temperature
* i
< required. ,
Gaseous ammonia was injected to the gas stream after
1
the heater and before the reactor. The gas then flowed down
through the reactor in contact with the catalyst. The NOx
in the gas reacted with the ammonia to form small quantities
1 of gaseous nitrogen and water, which remained in.the gas
,wstream.
:• xxi
-------
EXIST
FLUE
GAS HEATER
CZJ
AMMONIA SUPPLY
TANK
T
n
J
t700 NmVHr
1057 SCFM
500 PPM NOi
REACTOR
143 Lb/Hr NHj
(AT IOMRI
NH3
S92
AIR
NOx/SOj GENERATOR
SOOT 3U0WER
TO Fl'JE
GAS OUCT
DUST SEPARATOR
2
BLOWER
A'R vR
STEAM
fly ash to
SOOT BLWik
G£S (itATER
VACUUM l.!NE
Figure 4. Process Flow Diagram of Pilot Plant.
-------
UCGlN
FJfjT
M'.r or
Tf.\ r 5
Hi if.
a m*7Kry O*
PAi ^
w'H „.^»0
HI
i ' I *
- (. i ON j
i •> b
tc!> or
A^I-SACC
sz; ()c tA
; Dust settling in the reactor was kept within acceptable i
f ** l
bounds through the use of a soot blowing apparatus which used J
¦ 1 I
either steam or air. j j
A blower was located downstream of the reactor to over- j
^come the flue gas pressure drop through the pilot plant. A !
cyclone dust separator waB supplied ahead of the boiler to J
prevent erosion of the blower by dust. i
Since it was planned to test the performance of NO,
SO.
x' 2'
c these substances
SOjf and since only fixed concentrations o;
were available from the Albany, Ga. plant boiler operation,
S02 was purchased for addition to the flue gas, and a N0X, SO2
and SO^ generating unit was provided to manufacture these
materials for addition to the flue gas. This permitted testing
at NO..
x'
the boiler.
SO2 and SO^
levels higher than those available from
The operating conditions which were controlled in the
pilot unit are as follows:
1. Flue gas flow rate.
i ,
Temperature of flue gas.
The amount of charged ammonia.
2.
3.
4.
Soot blowing gas pressure and temperature,
1 1
Host Site i |
l ;
<
!
. The pilot plant was located at the Unit #3 of Plant
Mitchell, Georgia Power Co., Albany, Georgia. This unit has a
pulverized coalfired Combustion Engineering boiler which was
initially operated April 18, 1964 with a 125 MW nameplate
rating. The boiler has a rated steam capacity of 1,075,000
lb/hr of steam to the turbine at 1800 psig. The unit is
tangentially fired with CE standard tilting burners for steam
\
temperature control. The pilot plant was located outdoors of
the Unit #3. 1
—I
I*-
I - _
xxiii
L'PA ) oi"> .*'0« '1 fH
' 1 v* 'C*» ' s' 1 r * r ;
n «(.
TYPING GUIDC OHfrET
-------
Data "Collection
The pilot plant operation was closely monitored through
i
j the use of an array of instruments and analyzers. Continuous
(-analyzers of the latest design provided measurements of the
9f~
N0X levels both entering and leaving the reactor.
Much cf the data at the pilot plant was collected in a
data logger which frequently scanned the instruments and
stored the data. Visual display and paper tape display were
provided.
Relating the Results to a Full Scale Operation
I
I
L,. The test unit included a section of catalyst which had
j the same depth as that to be used in a fullscale plant. The
I results could therefore be directly extrapolated to a large
plant assuming that the velocity, mole ratio, and temperature
remain the same and the gas distribution through the catalyst
j bed was properly designed.
i Ammonia consumption could also be determined from the
i
test results because it is calculated as a mole ratio of
ammonia to NO^ in the flue gas.
NOXNON 600 TESTS
Introduction
«
i
I
I
j Primarily because of clogging of the catalyst channels by
j fly ash due to narrow clearances the initial charge and the
! second charge of catalyst did not achieve their expected
j performances. The installation and testing of NOXNON 600
i catalyst having wider channels was proposed by Hitachi Zosen
J and accepted by EPA.
! NOXNON 600 is produced from thin stainless steel wire
r'
, ,j
y f xxiv
-------
^Cxaesh as a base metal to give mechanical strength to which
catalytic components are cemented. This technique has been
proven in applications on oilfired and coalfired combustion
flue gas in Japan.
Catalyst Performance
Starting from April 22, 1980, the demonstration operation
with NOXNON 600 ran for more than nine months until February 2,
1981. The operation with combustion flue gas from the Dnit #3
Boiler of Georgia Power Company, Plant Mitchell was for 5,620
hours. The program was terminated due to scheduled maintenance
of the power plant requiring moving of the pilot plant.
The pilot plant program required achieving NOx removal
efficiency of more than 90 percent continuously for a period
of more than three months. Afterwards, the project scope was
extended and transient tests were included in the scope of the
contract along with an extension of the operating period.
Catalyst life tests were run to confirm the expected cat-
alyst life. Prom April 22, 1980, the pilot plant was operated
maintaining NOx removal efficiency of more than 90 percent until
the end of October. After October, a nominal 80 percent N0X
removal was accepted in order to decrease ammonia slip as far
as possible.
Following the catalyst life test, further testing was
carried oat to determine the effects of transient conditions
on the catalyst and to provide an extended operating time so
that at least 5,000 hours of operation could be obtained to
evaluate the long term effectiveness of the NOXNON 600.
At various times tests were run to determine the catalyst
efficiency.
During the operating time with flue gas tests were run
including catalyst life test, catalyst performance test, and
transient tests. Controllability and reliability of the entire
XXV
1 \ f ' \t . Ij.'l'l! ;• l
-------
cSjCiN \.v; i Or
cihSr -c • :aca
^system was aiso evaluated at the same time. }
I I 1
Operating conditions were varied for the performance j
tests and transient tests. Therefore, when the activity of !
TLX »
H i.
the catalyst^ was '5e evaluated the operating conditions were
P^set at certain consistent levels each time. These conditions-
; were a flue gas flow rate of 1000 1200 SCFM, a reactor
i
i temperature of 700° 720°F, and a NBo/N0 mole ratio of 1.0.
I J x
i Under these conditions the NOx removal efficiency was measured
i over a period of several hours to determine the condition of
J the catalyst!
i At the beginning of the operation : 90 - 94% removal
j At the end of August '80 ,(2500 hours) : 90 - 94%
(4000 hours) : 90 - 92%
(4420 hours) : 90 - 91.5%
I November 1980
j"" Before transient tests 12/80
: During transient tests 1/5/81'(5000 hours) : 90 - 91% 1
I 1
I After regeneration 1/27/81 (5500 hours) : 91 - 94.5%
' . '"i t !
; Operatinq Variables !
1 I ;
I 1 i !
: Mole Ratio ;
j During the test operation the pilot plant was operated at
; a selected mole ratio. The control system was designed to
! automatically provide this mole-ratio. This was accomplished
j by using the flue gas flow rate signal and the inlet NOx
; concentration analysis to determine the quantity of NOx in the
j inlet stream. From the inlet N0X quantity and the selected
i mole ratio the required ammonia was automatically calculated.
' This signal was then relayed to the ammonia control system
j which set the ammonia control valve setting to provide the
j required ammonia flow. '
j Figure 5 is a mole ratio curve which shows that a removal
; of 80% requires a mole ratio of about 0.85 while a 90% removal
1 requires a mole ratio of 1.0.
r=r-
Y'~ \
^ ^ XXVI
: P . f. ,-i ,V> 1 riw
I / L « ! A ' V r
T ^ .GCv 'Ot
-------
100
•
u
60
120
o
uj
04
HN)/H0n HOLE RATIO t~l
Co 1100 SCFM 700* 10 *F
SYMBOL < £» 1100 SCFM 700£l0 'f
[a 1500 SCFM 710 t 5 'F
Figure 5. Effect of NH3/NOx Ratio on NOx Removal Efficiency
and NH^ Slippage.
-------
bLCIU
r ihsi
L'f'E Gf ,
Tr\T t"-riue Gaa Flotf Rate —
c. NT":r-i 01
i'«» L
rop of
A, IV.AGL"
-T Atlf A
The NOXNON 600 catalyst for the pilot plant waa designed to
operate at 1057 SCFM which is equivalent to an Area Velocity
i n j
s-iTfA.V.) of 9.6 Nm /m -hr. However, operations at a much higher"*""1,
i flow rate of'1650 SCFM (an A.V. of 15) provided the desired j
90 percent NO removal efficiency.
i X
Therefore, it was anti-
cipated that]between the designed flow rate of 1057 SCFM and
the normal operating flow rate of 1500 S«-FM the NOx would be
unchanged. As seen in Figure 6 the flow rate had little or no
effect on the removal through the program.
Influence of SO Concentration on NO Removal Efficiency —
, - — ----- — x - - - - — — — •?*»,
!
I i
Concentrations of SO^ measured in the pilot plant deviated
widely between approximately 500 ppm and 1,500 ppm depending on
the variations of sulfur in the'coal.
Through'the whole operating period, NOx removal efficiency
was not affected by SOx concentrations in the flue gas in this
relatively wide range.
Operating Temperature —
In the pilot plant, the operating temperature was varied
between 640°F (338°C) and 780°F (415°C), and N0X removal effi-
ciency was not affected in this temperature range. Optimum
operating temperatures obtained from fundamental experiments
in the laboratory are between 572°F (300°C) and 750°F (400°C).
N0X Concentration —
On August 24, 1980, the NOx Generator w^a operated to in-
crease the concentration of N0x at the inlet of the reactor, and
j the influence of NO concentration was investigated,
< ( «T I
,
XXViii
¦>
)!\
t PA l-onn ?3 jO '4 A W »
TYmt? GUIDE SMELT
-------
100
H-
X
»!
90
eo
£ 70
REACTION TEMP. . APPROX 700 T
NHj/NOj RmiiO • APPROX 10
SYMBOL OPERATING PER 100
/o ?00 ~ I250H
V • 2150 — J750H
1000 1100 1200 1300 1400 1500
FLUE GAS FLOW RATE I SCFM I
1600 1700
Figure 6. Effect of Flue Gas Flow Rate on NOx Removal Efficiency.
-------
*4-_ Between,400 ppm and 900 ppra of inlet NO concentrations, _j
r * •
the concentration of NO„ did not influence NO removal efficien—
X * i
cy. This data is consistent with fundamental data.
Ammonia Slippage —
Since a continuous ammonia analyzer was not available
during the operation of the pilot plant, ammonia had to be
measured by wet analysis.
Figure 5 shows ammonia slip when mole ratio was varied at
operating times of 3200 hours to 4500 hours. General con-
clusions from these tests:
- Flue gas flow rate had little effect on ammonia slip-
page.
- Ammonia slippage seemed to increase slightly with
increased operating time.
- Ammonia slippage was apparent even at low mole ratios,
for example at a mole ratio of 0.6.
The reason for this higher than expected slippage may have
been due to clogging by fly ash and fibers of asbestos yarn used
to seal a clearance between the catalyst box and reactor. These
fibers along with fly ash between the first and second catalyst
layers probably reduced the effective catalyst surface and ad-
versely affected the apparent catalyst activity resulting in
relatively high ammonia slip. Even a slight reactivity loss
while producing little loss in N0X removal efficiency, can
cause a significant increase in ammonia emissions.
Oxidation of SO2 to SO2 —
On July 23, 24 and 25, 1980 operating conditions were main-
tained at a constant level so that S03 measurements could be
obtained at the inlet and outlet of the reactor.
The results indicated an average ox^dacion rate of 1.8%.
XXX
-------
BEGIN
r »hST
L»\l t ^
11 *T B+*
Hlm!
CCNTC?> Or
PAG.
T~
lErom prior experimental results;, it was expected that at these--
conditions the ratio would be somewhat lower: about 1.0 to 1.5%.
Transient Tests
After the NOx removal efficiency of more than 90 percent
was demonstrated in the continuous run of three months as re-
quired in the contract, a decision was made to extend the scope
of the contract. This was done' to supplement the originally
planned operation of the pilot plant to further establish the
suitability and reliability of Hitachi Zosen's catalyst for
i
commercial operations on coal-fired combustion flue gas. The
following transient tests were performed:
TOP Of
IMAC.C
AUf A
— Eft-
Emergency Shut-off of Ammonia Feed —
' I
1. !
The EPA pilot plant was provided with a trip system for
ammonia supply. The purpose of' the trip system was to shut off
the ammonia feed when temperature in the reactor decreased to
600°F and allow it to introduce^ ammonia into the system when
temperature in the reactor increased and returns to 600°F. The
purpose of this test was to confirm the reliability of the
I
trip system.,
The results of this test proved that:
- Automatic shut-off and supply of ammonia operated
smoothly. |
~ N0x removal efficiency and pressure drop at a tempera-
ture of 700°F was constant and unchanged through the
three repeated tests.
Cold Start-up —
A commercial boiler normally starts into operation after a
I long shut down with the reactor and ductwork filled with ambient1
J
¦ ' I " '¦ ; V
? i xxxi I
1
typi;:g gui^l shecT
-------
TOP .If
. • - , - .. .. - .
**T.air. When operation commences flue gas would be introduced _.j
into the system as the temperature rises and the flue gas could
be relatively cool for a time when in contact with the catalyst.
| There is a risk of formation and deposition of sulfuric acid !
,Jjmist, ammonium sulfate and ammonium bisulfate. Previous expe-
< rience proved that sulfuric acid mist does not deteriorate the
! catalyst, and ammonium sulfate and ammonium bisulfate can be
I removed when the temperature rises. The purpose of this test
j was to prove that cold start-up with flue gas does not cause '
j any trouble to the N0X removal reaction. j
On December 27, 1980, the blower was started and flue gas
| I
j wa6 introduced into the system directly. The system was j
| heated up with flue gas and put into operation. !
I Twelve hours after commencing the start-up, operating
conditions became stable at the same conditions before cold j
start-up. The results proved that cold start-up does not in- !
fluence catalytic performances. '
j
i
j Boiler Shut-down and Start-up — I
|
From time to time a power plant boiler shuts down and |
starts up and an NOx removal system must follow such transition :
periods. The purpose of this test was to confirm the con- j
trollability of the NO removal system during the shut-down and
X I
start-up of the boiler. When the host boiler was shut-down for
maintenance, this test was executed with the N0X removal system '
| shutting down and starting up along with the boiler with no
| purging of the reactor. From the results, the system proved
{ that it could withstand the transient period of shut-down and
j start-up of the boiler. j
i 1 i
! j
i Sudden Load Change —
!
i
| The boiler for a power plant may occasionally change load
I
f f xxxii
-------
Lauddenly complying with variations of power consumption. The J
N0X removal system should follow these sudden boiler load !
changes. The following two levels of operating conditions were •
1 adopted as representative load levels: I
high load: 1,300 scfin, 700°F, 0.8 mole ratio
low load : 900 scfm, 610°F, 0.8 mole ratio
The above conditions were altered once every two hours !
i
j and continued for 24 hours. There seemed to have been no
: I
serious problem caused by these sudden load changes. ;
' '
! [ i
Sootbiower Requirements — | j
I ! !
! \
From the beginning of pilot plant operation with the third ¦
'' - • *
| charged catalyst, NOXNON 600, the sootbiower was operated three .
| times a day, one cycle at a time, in order to prevent clogging '
by fly ash. ,This frequency seemed to be the minimum to pre- j
vent clogging; however, there was no experience in operating the
pilot plant without the sootbiower. Thus, a trial was made to
operate without the sootbiower. i
At a flow rate of 1,100 scfm, 700°F, 0.8 mole ratio, soot-
biower operation was halted and,the progress of pressure drop
' i
I increase was, observed. Pressure drop appeared unstable and
increased gradually from 1.20 to 1.25 inches H-0 to 1.35 to !
i
1.40 inches 1^0 in approximately 28 hours, and seemed to con- j
tinue to increase. The sootbiower was then restarted at this i
i
| time. From this test it was concluded that operation of the 1
} sootbiower was necessary. i ;
! ' j
, Changes in Pressure Drop in Relation to N0„ Removal Efficiency
I X 1 1 ^
I i i
i
In the treatment of coal-fired combustion flue gas, it is
| realized that adhesion and clogging caused by fly ash, along
! with the formation and deposition of ammonium sulfate and
f •
I bisulfate, cause problems. In NO removal systems, these '
i — X
" \
' f ^ _ xxxiii ,
-------
d'u,
. ; . - -
^problems are first noticed as an increase in pressure drop. J
¦ i
| Therefore, changes in pressure drop were carefully studied |
I 1 !
| during the operation ot the EPA pilot plant. I
j Since operating conditions varied from time to time in t
'^"accordance with the needs of the pilot plant operation and the !
I scope of the contract, changes of pressure drop were evaluated
; by a ratio of measured value to a calculated value. j
I Several experiments were carried out, and the tendency for
i pressure drop changes are summarized below.
•
| - When the pilot plant was shut down for some reason the
j pressure drop generally increased after the next start-
j up. ' The increased pressure drop usually continued at .
i ' i
i this high level although the sootblower was operated
. ^ . i • - >
I three timas a day. However, the previous pressure drop
J was restored after a few days of continuous operation,
j - When the flue gas flow rate was changed, the pressure
! drop increased at times. For example, in November 1980
| after some twenty days of steady operation the flow rate
I was changed from 1300 SCFM to 1500 SCFM and after two
days, at this higher flow rate the pressure drop suddenly
increased and did not return to its original level for j
several days. j
- Whenever Georgia Power Unit #3 Boiler operated the |
economizer sootblower repeatedly, the pressure drop I
across the catalyst increased. |
- On December 18, 1980 the water tubes of the boiler were
washed with pressurized water to remove slag on the tube
t surface While the Tsoiler was In operation. This abnor- .
mal maintenance work caused a serious increase in pre3-
; sure drop. \ i
' * 1
, - Operation of the sootblower seems to be necessary for
coal-fired combustion flue gas. The pilot plant was
J operated for twenty-eight hours without the sootblower
and the pressure drop increased sharply. After the
xxxiv
7 YH! ':i 'I' hI !
-------
sootblower was returned to operation it took three days._|
to restore the pressure drop. j
i ' 1
J
Regeneration of Catalyst j
, i
E+- , _ _ .. _ '
When the catalyst was washed with warm water the N0x re-
moval efficienay was restored to its initial efficiency. The
pressure drop also seems to have been partially restored.
i
However, the catalyst was partially clogged with asbestos J
fibers which is not a typical condition. The effectiveness of ;
water washing for pressure drop restoration, therefore, is un- J
certain. Also, the waste wash solution contains dissolved
metals from the fly ash which would require water treatment be- :
fore disposal. " |
j ! j
Abrasion of Catalyst 1 i
Among the four blocks of catalyst, only some of the flat
] plates in the top block showed a loss of catalyst by abrasion. J
This was probably due to the vibration of the loose plates j
caused by the impact of the flue gas flow. There was no ab- j
rasion in the other blocks. A commercial system would incorpo- ,
[
rate a tighter catalyst structure to avoid such vibration. j
! ' i
Clogging
I
Anticipated clogging of the catalyst by fly ash was pre-
vented by operation of the sootblower three times a day. Re-
J ducing this frequency of operation may be possible but it was
j not tested. Apart from the clogging anticipated by fly ash, as-
j bestos fibers unfortunately led to plugging of about 30 to 35
j percent of the catalyst passages between the top catalyst blocks
I and the second catalyst blocks. This was caused by asbestos rope
i used to seal passages between the catalyst box and the reactor
shell. This asbestos vould not be used in a commercial plant.
: A .
v
¥ xxxv
-------
^CONCLUSIONS T
ri ,'i11.1" ior or
The contract objectives were met and exceeded. N0x removal
efficiency of more than 90% was demonstrated during an operating
^period of approximately 5000 hours at the designed capacity of""'
0.5 MW equivalent. Following this period, transient tests were
I 1
run. These demonstrated that the performance was not adversely J
affected by such conditions as sudden boiler load changes, cold '
start-ups, low boiler loads, or by boiler shut-downs and start- 1
i '
ups. The pilot plant operation was terminated after 5620 hours I
i . !
of cumulative operating time only because the host boiler had to
undergo major modifications necessitating that the pilot plant !
be moved [ j
During the operation of the pilot plant, catalysts had to j
be replaced twice. The first and second charge of catalyst had |
relatively narrow clearances between the catalyst plates and were
clogged by fly ash from the flue gas resulting in increased pres-
sure drops through the catalyst beds. There was also a slight j
decrease in the apparent catalyst activity, although the true !
' i
catalyst activity did not decrease. These catalysts, which were,1
NOXNON 500, had been tested in pilot plants in Japan for applica-
tion to coal-fired combustion flue gases from boilers prior to j
application to the EPA pilot plant. However, clogging caused by,
fly ash was not experienced even though no sootblower was install
led or operated in those pilot plants. j
There were obviously significant differences in the charac-;
l
ter and composition of the fly ash in the U.S. as compared to j
that in Japan. [
Nevertheless, the tests with NOXNON 600 were highly success-
ful. This catalyst has somewhat wider clearances between the j
catalyst plates helping to avoid fly ash clogging problems. 1
Fly ash contained in flue gas varies in its characteristics'
j and behavior relative to its clogging tendency in catalyst beds.;
j This depends on the source and composition of the coal. At !
i
* ¦ ' ' Y
I hA r 'I
i J'A form 4 t l f
; i ^ • i'>i; i ^ i a t * > n
\ !'!• . .".iiMl
-------
CLM Cr* i *\-
nAi.f
r
' —I
^present, qualitative measurements to estimate the tendencies to. 4
agglomeration and cohesion of fly ash are available through •
chemical analysis and thermal processing tests. However, more ;
adequate and accurate methods to estimate the behavior of fly !
jjTash in catalyst beds will be required. It is expected that this
method when further amplified will be useful in selecting ade-
quate linear' velocity to prevent clogging and abrasion caused by
fly ash, and to determine the necessity and operating conditions
of the soot blower. Improvement of catalyst configuration to '
prevent clogging by fly ash is also expected.
Testing, for catalyst regeneration by washing was examined
just before the pilot plant was dismantled. The results were
very encouraging. The regenerated catalyst exhibited properties
of a virgin catalyst. However, due to the limited time avail-
able, potential problems related to the catalyst regeneration I
were not clarified. For example, the method of drying the ;
catalyst and the reactor after regeneration without encountering
corrosion or fly ash clinging, and the treatment of waste wash- ,
ing solution should be investigated before commercial appli-
cation. This are^. should be studied further since the cost of
1 1
this technology could be substantially reduced if the catalyst ,
life could be extended by in situ regeneration techniques. !
Based on tests with NOXNON 600 both at the EPA pilot plant 1
and in Japanese pilot plants, 90% N0X removal could be expected
at a NK.j/NOx mole ratio of 0.92 - 1.0. Although ammonia i
slippage from the EPA pilot plant measured about 40 ppm under '
i these conditions, ammonia slippage would be expected to be •
I less than 10 ppm based on Japanese tests. The differential ;
j was probably due "to~the asbestos clogging in the EPA pilot I
i 1
I plant. For 80% NO removal the required mole ratio- would be
I X » I
| 0.82 - 0.85 with ammonia slippage of 5 ppm or less, again based
i on tests in Japan. '
t ~ !
I The pressure drop with NOXNON 600 catalyst in the pilot
| plant and expected commercially is only between 1.0 and 1.4
top cr
IflACl
At-'lr-,
I
xxxv a
'i "I ^IDE SHLC I
-------
^ ^ C * Nf
r-'F-3T
u : O'
"•L<1
lO< •;•
' " '* . .. .... >*"
linches HjO. ^Such low pressure drops required very little power :
consumption resulting in low operating costs. j
The controllability of the pilot plant wa3 satisfactory. 1
However, the control system for commercial plants would differ
^slightly from the pilot plant control system due mainly to the
problems in measuring flue gas flow. The reliability of the
system was very good as was verified by the high onstream factor
achieved. A desirable addition to the control system would be a
continuous analyzer to monitor slippage of ammonia. A desirable
addition to the control systems would be a continuous analyzer
to monitor slippage of ammonia.' j
The N0y removal efficiency of this SCR system in commercial
applications would be expected to be the same as that experienced
in the EPA pilot plant provided that the size and configuration
of the catalyst, superficial linear velocity of the flue gas
across the catalyst, and the temperature of flue gas are the ;
same. Therefore, the data is directly applicable for scaling of
commercial systems. J
The pilot plant project was the first demonstration aw§
evaluation of N0x selective catalytic reduction technology on a
coal-fired source in the U.S. The project results indicate that
i 1
the process may be useable as a control option: however,
some technical concerns remain before the technology can be
considered commercially available ar.d demonstrated for coal-fired
sources in the U.S. '
pi~
XXXV111
T > I'i' (i Ci..!! !l- Si il L I
-------
SECTION 1
INTRODUCTION
Nitrogen oxides in the atmosphere have been determined to
have adverse effects on human health and welfare. The Clean Air
Act of 1970 and subsequent amendments to this act requires the
reduction in the emission rates of nitrogen oxides along with
sulfur oxides, particulates, carbon monoxide, and hydrocarbons.
A projection indicates that, while emission rates of the other
four pollutants will stabilize in the near future, the nitrogen
oxide (NOx) emissions will continue to increase and by 1990 will
be 50 percent higher than in 1975. Because of this the EPA has
placed special emphasis on the development and demonstration of
technology to reduce NO emissions.
As part of the effort to advance the technology for control
of NOx emissions, EPA has sponsored the design, construction and
testing of a pilot-scale unit (0.5 MW equivalent) which demon-
strated the operation of Hitachi Zosen's process on flue gas
frcm a coal-fired boiler. This report presents details of the
Hitachi Zosen process, designed to limit N0x emissions from
coal-fired steam generators, and results of the demonstration
program.
TKE PROBLEM OF NC CONTROL,
9C
Over 95 percent of man-made NO emissions aresult from com-
bustion process. Nitrogen and oxygen can combihe to form several
nitrogen oxide compounds. However, for most combustion pro-
cesses, the only nitrogen oxide compounds produced in significant
concentrations are nitrogen oxide (NO) and nitrogen dioxide (N02).
1-1
-------
TVPING GUIDE SHEET
cr \'i ek
¦ :C
BEGIN
FihSJ
LiUe 0*
7e\t _ __ __ _
n:«[ .t-iypically, NO^represents 90 to 95 pcrcsnt of the NO emissions
I ** r X
from combustion. But this NO gradually oxidizes to N02 once it
n o £S eKpOSed to oxygen in the atmosphere.
? " ^ ^ \
Sources of man-made NOx can be classified as either mobile
TO.1' 0i:
.MAfcc
-A yirA
S_C
(-¦"•A- 5
or stationary. Each of these sources contributes approximately-
|50 percent to the annual N0x emissions in the U.S. Mobiie
^sources primarily consist of automobiles and trucks while the
majority of stationary source emissions results from utility
and industrial steam generators.
METHODS TO CONTROL NOx EMISSIONS FROM STATIONARY SOURCES
i NO is formed in the combustion process by one of two
L
routes". It can result from combination of fuel-bound nitrogen
with oxygen, or it can result from the thermal fixation of
I
.atmospheric nitrogen and oxygen in the conioustion zone. In
i
both cases the end result is primarily NO. This is because
the residence time in most combustion units is too short for
a significant amount of NO2 to form.
1 ; -7 1
The formation of N0x from
'fuel-bound nitrogen is relatively insensitive to combustion
temperature while NO formed by thermal fixation depends primar-
I, *
,ily on the reaction temperature. Another significant factor
which influences the formation of NO^ is the amount of trace
oxygen available for reaction.
J The contribution of the two mechanisms to total NO^ emis-
sions varies widely depending on the type of fuel and the com-
bustor in use. For example, N0x emissions from a natural gas
fired boiler result almost exclusively from thermal fixation.
up to 80^percent of the ,NQX emissions from a coal fired
boiler can result from oxidation of fuel-bound nitrogen,
i Because of the way N0x is formed, there are three methods
of controlling N0X emissions. These are pre-combustion, com-
bustion, and post-combustion control. Pre-combustion control
is designed to limit the oxidation of fuel-bound nitrogen by
iu: u
i 12*-
i M
/ (i^ 1
1-2
- f>
, ; "v: >j-
1 ::v 1 «i!L-
~K'T\pc
DIM IO'\'
: i-1?: 1 M!, I!
Ili.mS
, (ri'A7
(4 ,*C
-------
C -v -.1
" " . . „
\sfremoving the.nitrogen from the fuel prior to combustion. Cur- ]
'rently, this control technique is primarily limited to switching
from a high to a low-nitrogen fuel. Techniques are under devel-
opment for removing nitrogen from fuel, but success has been
limited.
Combustion control, or combustion modification, is a second
technique which limits the formation of NO^ by adjusting temper-
ature and the availability of free oxygen in the combustion zone.
:Combustion modification techniques have been demonstrated in the
, past and to a large extent they are considered available tech-
nology. In fact, it was these developed techniques which per-
mitted utility boilers to meet New Source Performance Standards
(NSPS) for N0x emissions.
Employing conventional combustion modifications, NO^ emis-
sions can be reduced up to 50 percent with some loss of boiler
efficiency. And recently, more sophisticated techniques employ-
ing burner designs which carefully control temperature and the
amount of oxygen in the flame have demonstrated somewhat higher
reduction in NO^ emissions with little or no loss in boiler
efficiency. Combustion modifications are relatively simple and
inexpensive. ' Their primary limitation is the relatively low
degree of NOx removal which can be achieved. Although this
level of removal has been able to comply with current NSPS, com-
bustion modifications alone may not be adequate control technique
for complying with future NO^ emission standards.
A third method for limiting stationary source N0x emissions
is post-combustion control, known as flue gas treatment (FGT).
FGT is a general classification for a variety of processes which
remove NO^ from the products of combustion. This technology was
first developed in Japan and it has reached a relatively advanced
state for use on gas and oil fired boilers. However, it remains
to be demonstrated in the United States.. -Also,, veary little work
has been done with respect to applying FGT to flue gas from coal
fired boilers.
1-3
-------
rtCGic;
m >5r
uif.i g;
TL .<1
11— fl (
to
" ov
CL" -TE . v
il""
A
i
I There ace many different flue gas treatment processes, but__j
in general, they are characterized by high NOx control efficien-
cies. However, this additional removal is attainable only with '
a significant increase in costs. Typically, greater than 80 |
^'percent reduction in NOx emissions is possible with an FGT pro-"-'
cess as compared to a maximum of about 50 percent with combus- 1
tion modifications. i
t i
I One of the most advanced, promising, and cost jffective FGT
{processes is selective catalytic reduction (SCR) which can
achieve over 90 percent reduction in NO emissions.
! !
I In summary, there are three methods available to control
( '
jNOx emissions. Pre-combustion control is primarily limited to
tswitching the type of fuel burned in order to reduce the quantity
'of fuel bound nitrogen and thus, NO emissions. Combustion i "
f
modification involves careful control of the temperature and the'
.oxygen available in the combustion zone. This method of control
is relatively well developed technology, and it is relatively
linexpensive. However, the NOx removal efficiency associated with
.combustion modifications is limited to about 50 percent. Flue
gas treatmsnt is the third method of controlling NOx emissions.
•N0X removal efficiencies greater than 80 percent are possible .
and SCR processes are capable of iver 90 percent reduction in J
N0X emissioi. i. But, FGT processes J^ave not been developed in
this country to the extent combustion modifications have. In I
addition, the costs associated with FGT appear significantly !
higher than those of combustion modifications. '
I
DEVELOPMENTAL HISTORY |
I
I
Strict air pollution laws in Japi-n led to the construction '
of several full-3cale systems for the removal fit NO " from flue '
i -XI
igas. Hitachi Zosei. took the lead in this area', with the construc-
tion of the first large commercial unit in 1974. Hitachi Zosen •
1 CM' Oc
fMAGL
A
now has nine commercial plants in operation.
1-4
" '»ii;
Alt*
w J
' ? .1
^ u ' «
A u'A
H PING C.UinEZ SHEET
-------
3=f_ Early in 1978 Chemico Air Pollution Control Corporation
(CAPCC)(a division of Envirotech) acquired the North American
:license for the Hitachi Zosen technology for the process to
I remove HO^ from flue gas by selective catalytic reduction using
J.amntonia. CAPCC is a major supplier of wet scrubbers and other
flue gas desulfurization systems designed to control air pollu-
tion emissions from electric utility boilers.
Hitachi Zosen began the development of catalysts in late
1959 with basic ret-arch and laboratory testing. The original
work was done with carrier-supported catalysts. The process was
3
tested in 1973 in a 10,000 Nm /hr (6000 scfm) pilot plant that
was used for the denitrification of flue gases from a boiler
plant. The data collected from an extensive series of tests
provided the basis for the first N0x removal plant which was for
Idemitsu Kosan.
In early 19"M a contract was signed between the Idemitsu
Kosan Company, one of the world's leading petroleum refineries,
and Hitachi Zosen for the construction of a NO removal plant
3 X
with the capacity of treating 350,000 Nm /hr (218,000 scfm) of
flue gas. The plant was the largest of its kind in the world.
Construction work commenced at the Chiba Refinery in May 1975
with test operations in November of the same year. The plant
operated successfully with removals as high as 95 percent. Fol-
lowing the start-up of the Idemitsu Kosan system other plants
were constructed.
SELECTIVE CATALYTIC REDUCTION DEMONSTRATION PROGRAM
To further the goal of controlling stationary source NOx
emissions, EPA has sought to enhance the reliability and effect-
iveness of technology to reduce these emissions. One aspect of
EPA's involvement includes sponsoring programs designed to dem-
onstrate this technology. Because combustion modifications r,i.e
capable of achieving only limited reductions in NO^ emissions,
-------
BEGIN
HHST
I INc CI"
i;:\t
HERE ^
TYPING G'JIUE SHEET
Ct;-7 ER
Or
IC3 Ur
IM "
Di;or°rD
Ht-At;
5*
: itRE. :
-jsome emphasis", has been placed on' developing flue gas treatment <
'processes. In particular, SCR technology appears to be a very j
ipromising method of reducing stationary source NO^ emissions by j
jover 90 percent. Consequently* the EPA has acted to demonstrate;
^some of the more advanced SCR systems. I
I EPA initiated programs to demonstrate two SCP processes on •
! 3
Jan 0.5 MW scale. The processes are: J
I 1) ' me Shell-UOP Simultaneous SO /NO Removal System
I I XX r
j , 2) The Hitachi Zosen NOx Removal System j
,The EPA sponsored programs demonstrated these processes on flue
jgas from coal-fired boilers. The demonstration programs were
iexpected to answer many of the questions which remain concerning!
j >
[the application of SCR technology. In addition, these programs •
[should provide an improved basis for estimating the costs of j
I * I
applying SCR technology.
) t.
ULGIM
L-vS>1 1
or ;l\t n
i
17
>>? (Cl'l J
I WGi
j.
1-6
i
....... .
mC L fV\ ..
! * - ,
* ' ; 0
l'\ \ '
¦i1 r
:
-------
SECTION 2
PROCESS DESCRIPTION
The best method of removing nitrogen oxides (N£>x) from flue
gas is through the use of selective catalytic reduction. Many
reducing agents can be used for this purpose including hydrogen,
carbon monoxide, hydrogen sulfide, ammonia, methane, and other
hydrocarbons. However, ammonia is the most practical reductant
for this purpose because it reacts selectively and quantitatively
with NO^ to produce innocuous nitrogen and water. Ammonia is
available, relatively inexpensive, safe to handle, and easy to
store.
Ammonia will react with NO without a catalyst in a narrow
temperature range at around 1000 C. By using a suitable catalyst
the required temperature can be lowered to a more practical level
of between 30Q-450°C. Several different catalysts have been
tested for this purpose. Platimum compounds are the most effi-
cient catalysts to promote the selective reduction of nitrogen
oxides. However, they cannot be applied to most flue gases due
to the poisoning of the catalyst by sulfur oxides and ocher
elements or compounds.
Carrier-based catalysts have been developed and used by
Hitachi Zosen in several plants in Japan. These are pellets of
alumina, silica, titania, or ot"her materials into which the cat-
alyst is impregnated. Alumina is a very suitable-material but it
has been found to react with sulfur oxides, thlis decreasing the
available surface area and catalyst activity. Other carrier
materials, however, have been found to be less affected by SO^
and these have been successfully applied in commercial systems.
With dust-containing gases, carrier-based catalysts are not
2-1
-------
,"l CI*.
! ""•:T, _ . . . . >
> r ^practical because the bed of catalyst will be plugged by the
jparticulate matter in a short time. However, parallel-flow
jhoneycomb catalysts have been developed and applied to the
(removal of NO from dusty gases. The catalysts permit lower
i
'operating costs because of their low pressure drops,
j The honeycomb catalysts developed by Hitachi Zosen have
|been extensively tested in several pilot plants and were used
jin this test program. They are termed NOXNON 500 and NOXNON 600.
I
I
OVERALL PROCESS SCHEME
; The process consists of the injection of a small amount of
ammonia into the flue gas and passage of the flue gas over a
catalyst. The ammonia reacts almost entirely with the nitrogen
joxides in the flue gas to form small quantities of nitrogen gas
and water vapor, both of which are usual constituents of the
atmosphere and are environmentally acceptable.
Figure 2-1 illustrates a typical flow sheet of a commercial
system. Ammonia j.s injected into the flue gas from a boiler
between the economizer and the air preheater. At this point the
igas temperature is about 400°C which is suitable for the cata-
lytic reduction of This gas enters the reactor, passes
;over the catalyst, and then reactions proceed. The flue gas then
ipasses through the usual air heater, precipitator, S02 control
•system, fan and stack. The ammonia for the reaction is vaporized
jwith steam and is diluted with air (or steam) before injection
into the duct. If the flue gas temperature is too low for
^optimum removal efficiencies, additional heat can be added by
'auxiliary burners.
2-2
-------
TEMPERATURE
"1 RECORDER/
STEAM
OR AIR "*¦¦]
TEMPERATURE
RECORDER
NOX
RECORDER
I CONTROLLER
L.
SOLENOID
VALVE —
VEHT-s
PRESSURE
RECORDER
CATALYST
NOX
RECORDER
BOILER
REACTOR
w
PREHEATER
FLOW INDICATOR/
SOLENOID
VALVE
TO SOxANO PARTICULATE
CONTROL SYSTEMS
BOILER
LOAD SIGNAL'
COMPRESSED
AIR
FLOW INDICATOR/
CONTROLLER
Figure 2 1. Typical commercial system flow diagram
2-3
-------
^Chemistry
i"C" 'jc
<*G[
I ;
The exact relations between ammonia and nitrogen oxides are
not completely understood. However, certain reactions are prob-
ably involved which may or may not include oxygen. Virtually all
of the N0x in combustion gas is present as NO so the following
equations are only for NO reactions:
4 NH3 + 4 NO + 02 = 4 N2 + 6 H20 (1)
4 NH3 + 6 NO = 5 N2 + 6 H20 (2)
If N02 is also present, the following equations represent
reactions which may also occur:
4 NH3 + 2 N02 + 02 = 3 N2 + 6 H20 (3)
8 NH3 + 6 N02 = 7 N2 + 12 H20 (4)
2 NH3 + NO + N02 = 2 Nj +3 H20 (5)
At high temperatuiss (above 1000°C), additional reactions
can take place as illustrated by equations 6, 7 and 3:
4 NH3 + 3 02 = 2 N2 + 6 H20 (6)
4 NH3 + 5 02 = 4 NO + 6 H20 (7)
4 NH3 + 4 02 = 2 N20 + 6 H20 (8)
The reactions of ammonia with NO over the catalyst occur
o ^
below 300 C. Without a catalyst the reaction will only occur in
a narrow temperature range of 950-1000°C. Below this temperature
the reaction rate is very low while at higher temperatures reac-
tions (6), (7) and (8) readily occur producing more N0x than
there was originally.
With the presence of sulfur oxides in the flue gas, certain
undesirable reactions can also take place. These are the reic-
;tions between ammonia and sulfur trioxide to form ammonium sul-
fate or ammonium bisulfat'e.
2 nh3 + so3 + h2o = (nh4)2so4
nh3 + so3 + h2o = nh4hso4
A discussion of the reactions which form ammonium sulfate/
bisulfate is presented later.
2-4
-------
Tl ¦ i ^jkinetics of the Reaction
- r c "¦¦ 11 l— . ¦ i ¦ ——¦ ¦ ¦
It appears from the results of laboratory and pilot plant
i
•tests that the reduction of nitrogen oxide with ammonia is sim-
'V ^ilar to a first order reaction. With sufficient ammonia present
the rate of disappearance of nitrogen oxide is proportional to
the concentration of nitrogen oxide. In practical terms, at a
fixed ges velocity and a fixed volume of catalyst the percent
removal of NC>x (at any one ratio of ammonia to NC>x) would be the
same at any N0x concentration. This can be shown as follows:
A first order reaction can be expressed mathematically
as: dC = . where t = residence time
dt = C = concentration
k = a constant
Integrating this expression under the conditions t = 0
and the initial concentration as CQ, the result is:
If t is a constant at a fixed velocity of the gas and a
fixed volume of catalyst, then C/CQ (or the fraction removal)
would also be a constant. This has been substantiated by exper-
imentation.
Regeneration of the Catalyst
The catalyst is designed with an expected life of approxi-
mately two years. Regeneration of the catalyst is not needed
during the planned catalyst life. At the end of the useful life
of the catalyst it would be removed from the reactor vessel and
scrapped. The catalyst does not contain any hazardous materials
,and can be disposed of by recovering the metal in the catalyst
or by disposing of the spent catalyst as industrial waste. No
special precautions are required in the disposal process.
2-5
-------
C. ^ ; 1 »i r
^Factors Affecting the Catalytic Reduction of NQ^
In any chemical reaction, there are factcrs which can
influence the rate and extent of the reaction. These factors
include the reaction temperature, concentrations of reactant
species, and other parameters specific to the reaction system.
The following discussion briefly examines the major influences
"on the catalytic reduction of NO by ammonia.
The most critical variables which affect the degree of NO^
iremoval are the mole ratio of ammonia to N0x, the flue gas flow
rate, and the recctor temperature.
¦Mole Ratio —
This variable is defined as the ratio of moles of ammonia.
1
fed into the reactor to the moles of NO in the flue gas to be
treated. Typically a ratio of 0.9 to 1.0 is required for a NOx
removal of 90 percent or higher. Ammonia will remove one mole
of NOx per mole of ammonia up to a mole ratio of 0.7. Above
this the removal becomes less efficient and the removal of more
than 90 percent NO^ requires a lot more ammonia for the final
quantities of NOx to be reacted.
Flue Gas Flow Rate —
The gas velocity through the reactor is an important para-
meter. At lower velocities the contact time with the catalyst is
increased and better removals are obtained. Operating at higher
velocities will, of course, allow the system to use less catalyst
and reduce the costs. Higher velocities will, however, increase
the pressure drop and will also lead to increased erosion of the
catalyst surface particularly when significant quantities of fly
ash are present in the gas. (The degree of erosion is very
slight and the catalyst is non-toxic and, therefore, there would
2-6
-------
24be no air pollution problems.) The increased erosiion will 1
decrease the life of the catalyst. Hitachi Zosen has performed
Ja great deal of research work to develop catalysts which are :
{resistant to erosion and they have established optimum gas '
i
Velocities for obtaining high removal effiencies with low
!pressure drops and long catalyst life. Tests at a wider range
iof flows performed in Japan indicate a definite drop off of
efficiency at very high flow rates. Very high flow rates are
|not recommended because of increased pressure drops.
The concept of space velocity is normally applied to cat-
alysts which are granular, cylindrical, ring, extrudate, and so
!cn. However, the concept of space velocity is not useful in
designing for the use of corrugated catalysts because of this
unique structure. Instead of space velocity, Hitachi Zosen uses
area velocity (A.V.), which is defined a3 flue gas volume flow
rate per unit of apparent catalyst surface:
Nm3/m2~hr
i
The particular A.V. used in the Hitachi Zosen design depends
upon the percent removal of N0X required, the gas temperature,
and other factors.
<
Temperature —
The flue gas temperature in the reactor is an important
: variable. Normally the reactor temperature is held above 320"C
(508°F) to avoid the possibility of ammonium sulfate/bisulfate
.formation and the reactivity of the catalyst is very high at
3 2
this level. At area velocities of about 10 Nm /m -hr tempera-
tures above 300°C have little effect on the percent NOx removal
lefficiency. In effect, the normal expected operating temperature
ranges {above 300°C) have essentially no influence over the
removal efficiency. > •
-------
^Sulfur Oxide Concentration —
I
i
The presence of SO has essentially no effect on NO„
I X X
•removal since the catalyse has been formulated to tolerate the
^presence of these compounds.
i i
;Oxygen Concentration —
i
i A certain amount of oxygen in the flue gas is necessary to
promote NOx reduction. With flue gas containing 2 percent or
more oxygen, there is little effect of oxygen on the NO removal.
I x
'However, at oxygen levels below one percent the removal is
adversely affected.
Moisture Content of Flue Gas —
1 Moisture in the flue gas tends to decrease some of the cat-
alytic activity. However, this decrease is small and level3 off
above 5 percent moisture in the flue gas.
I
NOx Concentration —
i
The performance of the catalyst is not affected by the usual
,inlet NOx concentrations expected from a boiler provided ammonia
is injected in stoichiometric quantities.
:Ammonia Emissions
i
I
! A small amount of ammonia will invariably pass through the
;reactor and exit with the flue gas due to incomplete reactions
i
between the N0X and NH^. Ammonia slippage (unreacted ammonia)
;is of concern and efforts are normally made to control slippage
i
|to levels of 10 to 20 ppm. A high slippage of ammonia can be
considered a pollutant and, in some cases in Japan, agreements
-with local governments have set allowable levels at around 10 ppm
-------
-o OS
oic-\ c Ar; 11 •'
^1V,^ Another"reason for the concern with ammonia slippage is that flue
igas desulfurization following the denitrification system could
absorb any ammonia in the f)ue gas and a build-up of ammonium
salts can occur. Bleeding of the ammonium salts from the system
can cause water pollution control problems in certain cases.
The degree of slippage is a function of the NH3/NC>X mole ratio*
;the area velocity, and the temperature at which the reaction
'occurs. For commercial application to coal-fired utility boil-
ers, ammonia slippage would be expected to be less than 5 ppm
'at an N0X removal efficiency of 80%, and less than 10 ppm at
'an NOx re.aoval efficiency of 90%.
I
Ammonium Su1fate/Bisulfate
The presence of sulfur trioxide (S03) in the flue gas can
lead to a reaction with ammonia to form ammonium sulfate and
ammonium bisulfate. When burning heavy fuel oil, approximately
2-4 percent of the sulfur oxides in the flue gas are present as
the trioxide. For coal combustion, S03 accounts for approxi-
mately one percent of the total SOx- The reaction of ammonia
with SO3 will not occur above approximately 300°C (572°F). At
very low concentrations a temperature of around 200°C (392°F)
might be sufficient to avoid this formation. Figures 2-2 and
,2-3 show the relationship between the ammonia concentration,
sulfur trioxide concentration and temperature of the flue gas
below which the reaction to form ammonium sulfate or ammonium
bisulfate will occur.
The condition in which ammonium sulfate and bisulfate are
formed depends on the equilibrium in the reactions described
below:
2 nh3 + so3 + h2o = (nh4)2so4 (1)
nh3 + so3 + h2o = nh4hso^ (2)
(NH4)2S04 + S03 + H20 = 2 NH4HS04 (3)
NH4HS04 + NH3 = (NH4)2S04 (4)
2-9
TYr . ' r
-------
. t I '-it sjt"
r i *'V _ __
1 jsC Figures 2-2 and 2-3 show phase diagrams of the above reac- ^
jtion equilibrium when the vapor pressure is atmospheric and the
moisture content is ten percent. Solid lines shown as A and B,
and dotted lines C and D indicate the equilibrium between equa-
tions (1), (2), (3) and (4), respectively.
-¦r
- i The area surrounded by the solid line A and B, the vertical
axis, and the horizontal axis represents the condition at which
(NH4)2S04 and NH4HS04 cannot be formed.
i In the area surrounded by the solid line A, the dotted line
C, and the vertical axis where the concentration of ammonia is
higher than the concentration of line A, only (NH4)2S04 is pro-
duced according to equation (1). In the area to the right of
solid line B and below dotted line D only NH^KSO^ is formed
according to equation (2).
! In the area to the right side of dotced line C and the upper
side of dotted line D, both (NH4)2S04 and NH4HS04, produced by
equations (3) and (4), coexist.
j For instance, with a concentration of 10 ppm NH, and 50 ppra
\ ^
S03, ammonium bisulfate (NH4HSC>4) will form if the temperature
drops to around 230°C (446°F). (See Figure 2-3).
Double salts of ammonium sulfate and ammonium bisulfate can
also form upon the cooling of flue gas. The flue gas temperature
leaving the air preheater is in the range of 150-200#C (302° -
392CF) and deposits can, therefore, form. The deposits are often
double salts of ferrous sulfate and ammonium bisulfite. The
ferrous ions are the results of the corrosion of steel by
ammonium bisulfate which is very corrosive.
' It is also known that certain deposits can form on catalyst
surfaces at 300°C (572°F) or higher even though the gas concen-
tration is relatively low in SO^ and NH^. This condition could
be the result of localized high concentrations on the surface of
the catalyst due to the oxidation of sulfur dioxide and the
absorption of ammonia.
To avoid such deposits Hitachi Zosen recommends that the
temperature in the reactor be maintained at a minimum of 320°C
2-10
-------
s
a.
a.
392 *F
(NH4)2S04l , ,7s
HNH4I2 SO4
!NH4HS0«
Z
o
I—
Cfc
374 *F
H*
Z
Ui
u
z
o
o
356 *F
X
z
GAS
338 "F
001
Q05
001
SO3 CONCENTRATION (PPM)
Figure 2-2. Formation of Ammonium Sulfate
(Low Temperature Range)
2-11
-------
1000
536 r
500
518 *F
500 *F
100
z
o
482 °F
Ui
o
z
o
o
464 °F
K>
X
446 *F
z
— 0
428 "F
392 *F
10 50 100
S03 CONCENTRATION (PPM)
500 1000
Figure 2-3. Formation of Ammonium Sulfate
(High Temperature Range)
2-12
-------
jl(608 F) as long as ammonia is injected into the reactor.
j The formation of bisulfate can be minimized to a large
¦extent by maintaining a low ammonia level in the flue gas exit-
'ing the reactor. The presence of fly ash also reduces the prob-
lem both by scouring the deposits off of surfaces and by provid-
ing surfaces on which the deposits will form instead of forming
,on metallic surfaces. If preheater deposits do form, sootblowing
is effective for removal, as is occasional water washing.
APPLICATION CF PROCESS TO BOILERS
A schematic flow diagram of the HZ process as applied to a
boiler is shown in Figure 2-4. Flue g-is leaving the economizer
at a temperature of about 390°C (734°F) is first mixed with
ammonia in quantities needed to meet N0x removal requirements.
The gas is then passed through the fixed bed catalyst reactor.
No dust removal is required prior to the catalyst because of the
non-clogging design of the catalyst geowetry. Nitrogen oxides
are reduced by ammonia to innocuous nitrogen and water. The
denitrified gas then resumes its passage through the normal
boiler train: air preheater, dust collection, etc.
Ammonia required for the process is first vaporized with
steam and is then diluted with either air or steam to aid in the
distribution of ammonia into the flue gas. The diluted ammonia
is fed into the duct through a network of nozzles or perforated
pipes at some distance before the reactor to allow complete dif-
fusion into the flue gas.
The reactor is supplied with retractable soot blowers in
which blasts of steam or hot air are occasionally applied to the
catalyst bed to remove deposits of dust which may adhere to the
surface of the catalyst.
The N0x removal efficiency depends mainly upon the gas
temperature, the ratio of ammonia to N0x, and the gas velocity.
Removals of 80 percent or higher can be readily attained under
2-13
-------
GAS DISTRIBUTION
^ GRIDS
REACTOR INLET
CATALYST
AMMONIA FEED
CONNECTIONS
SOOT BLOWER
HOPPERS
TREATED FLUE GAS TO
BOILER PREHtATER
Figure 2-4. Typical Reactor
2-14
Arrangement
-------
s^proper conditions of operation, j —j
Minimum design temperature for the system is around 320°C
(603°F). At temperatures lower than this ammonium sulfate or
jbisulfate produced by the reaction between ammonia and S03 pay
i ,
^crystallize on the catalyst surface resulting in a decrease of
[catalyst activity due to adhesion on the catalyst activated
jsurface and/or blockage of the catalyst layer.
! The gas temperature exiting the economizer is normally
!
higher than 320°C. In other cases additional heat may be
[required. This additional heat can be attained by an auxiliary
{burner which provides high temperature flue gas into the flue
Jgas leaving the economizer. An economizer bypass is also pos-
sible. Most of this additional supplied heat would be recovered
!in the air preheater.
i t '
j On the other hand, at temperatures higher than approximately
!450°C (842°F), while the efficiency of the catalyst for NOx
removal is not affected, the activity of the catalyst for NOx
formation by oxidation of ammonia may become appreciable. This
.activity, objectionable from the standpoint of NOx removal,
appears at 450-470°C (842-878°F). Although slight at such temp-
eratures, it gradually increases with further increase in temper-
ature. Because of the undesirable effects discussed above, the
'optimum reaction temperature range is 330-400°C (626-752°F).
CATALYST DESCRIPTION
I
A particularly effective physical design of the catalyst
structure has been developed by Hitachi Zosen. This structure
i
is honeycomb shaped as shown in Figure 2-5. Overall, the
.catalyst has the following characteristics:
(1) The structure is a thin plate honeycomb,
j \
! (2) Due to substantially reduced pressure! drop across the
| catalyst layer, operating power costs are much lower
than with conventional catalyses.
-------
Figure 2-5 Configuration of NOXNON 500 or 600 series catalyst.
2-16
-------
(3) A straight gas flow path prevents dust clogging,
i (4) It is applicable for gases with high S02 concen-
; trations. j
i ;
' As a result, in treating high-temperature gases with high
,"SOx aid dust concentrations such as coal-fired boiler flue gas,
,the N0x removal reactor can be installed immediately downstream
jof the economizer. In addition, dust elimination or other pre-
|treatment is unnecessary and, with the low pressure drop, oper-
ating costs are low. ,
! The honeycomb "atalysts are referred to by Hitachi Zosen as
I -
.NOXNON 500 or NOXNON 600 5eries. These catalysts are manufac-
tured in the form of plates and are fabricated so as to form
.parallel flow gas passages which provide excellent contact with
the flue gas with minimum impingement of fly ash on the catalyst
surface. The gas passages are created by thin plates folded in
•a plicated pattern and inserted between flat plates that act to
separate one folded plate from adjacent folded plates. The
resulting passages are then oriented in the gas stream to be
parallel to the direction of gas flow. The problem of erosion
by dust is minimized by the selection of the proper linear vel-
ocity of flue gas over the catalyst layer. This has been demon-
strated at a coal-fired power plant in Japan where tests have
been run for several thousand hours without significant deter-
ioration of the catalyst.
I (
t
Catalyst Preparation
The catalyst plates are arranged in a steel frame box sup-
ported by retainers,. A standard module is 1.0 meter long, 1.0
meter wide, and 0.5 meter deep.
; The activation of the catalyst follows after the corrugated
catalyst assembly is made. In the NOXNON 500 catalyst, thin
stainless steel plates are used. The surface is first converted
to an aluminum alloy which is then treated with an aluminum dis-
solving solution rendering the surface layer porous. The steel
*
-------
bFl'IN! TOP 0?
nesr ¦ ,>vj. 1
L'" s^aurface may tlhen be immsrsed in a solution containing active con^
Ht Rt
ponents which adhere to the porous surface layer. A permanent 1
bond is obtained by further proprietary treatment. !
I
A newer development is NOXNON 600 catalyst which, instead of
;plates, uses a stainless steel wire mesh as a base metal to give '
[mechanical strength upon which catalytic components are cemented.
!The NOXNON 600 is considerably lighter in weight and contains
I '
[more active material for e given volume of catalyst. !
j The active components of the catalyst consist of vanadium
land titanium compounds. Other components are added to increase
Resistance to fly ash abrasion. \
! ' i i
1 ' i
jREACTOR DESIGN ;
- - J " ' "
! Reactor vessels housing the catalyst (see Figure 2-6) are
j
iof carbon steel construction and are divided into chambers each
having its own fly ash hopper. The reactor is oriented so that
'flue gas flows downward. This minimizes the buildup of fly ash
on the reactor internals. Flue gas usually enters at the top of
the reactor at one side, passes through the gas distribution grid
and catalyst bed, and exits the opposite side of the reactor at
the bottom. Fly ash that drops out of the flue gas stream is '
collected in the bottom fly ash hoppers and is disposed of
periodically. >
' It is important to provide good gas distribution that
I i
smooths and evenly distributes the flue gas flow as it enters
,the reactor. The catalyst cells thereby get equal exposure to
the flue gas and gas channeling is decreased. Channeling could
'lower NO removals and decrease ammonia utilization. Fly ash
** I
erosion could also result at higher gas velocities through a
•channeled area. J , j
j Proper gas distribution is provided by ensuring, for one
thing, that the gas velocity entering the inlet duct is evenly
I
distributed. Even gas distribution at the inlet duct will
¦depend upon the design of the ductwork upstream and turning
4 '
i f 2rl8
rvr,"'C;'JU)C-c i-it'L-T
-------
GAS IN
GRIDS
GAS OUT
FLY ASH HOPPERS
Figure 2-6. Elevation view of typical reactor design
2-19
-------
tl/t0 ®4vanes may be /required at bends in the ductwork. Additional
rl£Fil [ ~
;design factors for even distribution are incorporated in the
!reactor design. The slope of the oblique top plate is one
iof the critical parameters. Specially designed grids above
..the catalyst bed are also very important. Other factors are
,, - !the spacing between the bottom of the catalyst layer and the
jtop of the outlet duct, and the width to length ratio of the
reactor. These parameters have been established through
extensive experimental efforts.
! Space above the catalyst bed is also required for the
jinstallation of soot blowers and for monorails required for the
!removal of the catalyst blocks.
i
1 !
p - - - - - | -
2-20
\ D- • (J -*
-------
,47 ~ SECTION 3 " ,
• , - ,
I
| •
I ' development history
I '
!
j Hitachi Zosen began the development of selective catalysts
:for N0x removal in late 1969 with basic research and laboratory
[testing. These were carrier-supported catalysts. By 1974, after
|extensive pilot plant studies, the catalysts were considered to
be commercial. The Green Chemical Company, a wholly-owned sub-
sidiary company, was then formed in June 1974 to manufacture
'these catalysts and started operation in 1975.
' In early 1974 a contract was signed between the Idemitsu
^Kosan Company, Ltd., one of the world's leading petroleum refin-
'eries in Japan, and Hitachi Zosen for the construction of an N0x
:removal plant with a capacity of treating 350,000 NM3/H (218,000
¦SCFM) of flue gases. The plant was the first of its kind and
isize in the world. Construction work commenced at the Chiba
Refinery in May, 1975 with test operation in November the same
iyear. Nux removal efficiencies of 95 percent were demonstrated,
j Following the successful start-up of the Idemitsu-Kosan
I system, other plants soon went into operation including a
!440,000 NM3/H (274,000 SCFM) unit at a petrochemical plant, and
'two plants at s»-°el manufacturing facilities. All or these have
'operated succ^ x-ully. (See Table 3-1 for a listing of commer-
cial plants built by Hitachi Zosen).
i
A particularly effective physical design of. the--catalyst
t
structure has been developed by Hitachi Zosen. This structure
is of a metallic corrugated shape.
1 One of the primary applications for Hitachi-Zosen's corrug-
;ated NOXNON catalyst is the treatment of high temperature gases
*with high SO and high dust concentration such as coal-fired
1 Y jrjc c? 1J1 J'. r
-------
1
2
3
4
5
6
7
8
9
TABLE 3-1. LIST OF COMMERCIAL PLANTS BUILT BY HITACHI ZOSEN
Customer
Treating
Cspdcity/
Nm /hr
Flue Gas Source
Process
Osaka Gas Co.,
Sakai
5 3,000 LNG or naptha-fired Ammonia
Furnace Reduction
Dakai Engineering, 5/000
Chiba
LPG-fired furnace
Ammonia
Reduction
Idemitsu Kosan,
Chiba
350,000 CO boiler and gas
-fired heater
Ammonia
Reduction
Shin-Daikyowa
Petrochemical,
Yokkaichi
Fuel oil-fired
440,000 boiler with wet
-type desulfuri-
zation
Ammonia
Reduction
Hitachi Sosen,
Osaka
6,000 Gas-fired annealing Ammonia
furnace Reduction
Toshin Steel Mill,
Hime j i
70,900 Kerosene-fired
steel heating
furnace
Ammonia
Reduction
Kawasaki Steel,
Chiba
762,000 Iron ore sintering Ammonia
plant with wet-type Reduction
desulfurization
Nippon Satetsu,
Himej i
10,000 Fuel oil-fired
steel heating
furnace
Ammonia
Reduction
Maruzen Oil,
Sakai
150,0C0
Fuel oil-fired
boiler
Ammonia
-------
^boiler flue gases. Due to the catalyst's non-clogging feature,
the NO removal system reactor can be installed directly behind
the economizer. Thus, expensive flue gas pre-treatment for dust
removal is not required.
Development work by Hitachi Zosen has been conducted at
various bench-scale and prototype pilot plants to ensure the
successful application of N0x removal process technology for
the treatment of dirty gases exhausted from power plants (coal-
fired and high sulfur-containing heavy oil or residual oil
fired), iron ore sintering plants, cement kilns, and other sim-
ilar sources.
One of the most important of the pilot plant operations has
been at the Electric Power Development Company, Ltd. (EPDC) power
plant at the Isogo Station in Yokohama, Japan. This is a colla-
borative effort between EPDC and Hitac.nl Zosen under subsidy of
the Japanese government. This has resulted in the collection
of a great deal of valuable data. Important developments, part-
icularly in methods of gas distribution, have been achieved.
There were three reactors installed, each treating about
3
200 Nm /hr of gas, plus several smaller reactors for abrasion
testing. These reactors operated essentially continually for
several years.
NOXNON 600 catalyst was tested at this facility and some of
the runs were over 6000 hours and were halted only because of
the shut-down of the boiler for scheduled maintenance. As an
example, one test was run for about 6300 hours at the end of
which the targeted removal of 80 percent was still being attained.
Ammonia slippage was very low as was the conversion of SO2 to
SO^. These tests established the effectiveness of NOXNON 600
catalyst, its resistance to abrasion, its long-term reactivity,
and the low tendency to convert SO2 to SO^-
3-3
-------
SECTION 4
DEMONSTRATION PROGRAM
The success of Hitachi Zosen in the design and construction
of several full-scale denitrification plants in Japan has prompt-
ed the EPA to fund further research on flue gas from coal-fired
boilers. This work was initiated to accurately define the cost,
process performance, complexity, reliability, and the process
impact on the operations of power plant equipment.
The principal potential market for the denitrification of
power plant flue gases will probably be at coal-fired stations.
There has been limited test work done by Hitachi Zosen on flue
gas from coal combustion. The EPA-sponsored demonstration pro-
gram at Plant Mitchell of the Georgia Power Company provided an
excellent site for testing of the Hitachi Zosen process. The
power plant burns typically medium sulfur coal with relatively
high ash levels. Adverse effects, if any, on the catalyst by
this flue gas could be readily evaluated.
The demonstration plant was highly instrumented and provided
much useful data to supplement data collection from previous work
in Japan. Several parameters were evaluated and their effects on
removal efficiencies and ammonia utilization were checked. The
data was collected and evaluated to establish valid characteriz-
ations of the process.
Long term tests were conducted to provide information on the
aging tendencies of the catalyst when applied to flue gas from
American coal burning boilers. Additionally, the data collected
during these tests were useful in establishing operating and
.capital cost requirements for commercial installations.
-------
TV.JI:\'G C'JIut SHLlT
- • , fl-
ii
~SCOPE"~"OF WORK ' ]
i- * ; ~1
I i ;
. Hitachi Zosen, with CAPCC as their major subcontractor, j
provided the test plant and operated the unit. This included
'detailed engineering, procurement, fabrication, transportation;
erection, test operation and continuous demonstration operation "
of the pilot plant. The work was performed in four phases:
Phase I (engineering) started with the basic design package
which was prepared by Hitachi Zosen in Japan. Based on these
I
designs, CAPCC prepared a detailed design. This was sufficient
for the issuing of requisitions for all equipment, vessels,
instruments, electricals, piping, insulation, and a construction'
subcontract. During Phase I, equipment with unusually long lead
times were identified so that procurement of these items could
be initiated during Phase I. Also included in Phase I were a
detailed caoital and operating cost estimate including off-sites
and interfaces with the host boiler and spare parts. At the end
of Phase I, detailed reports were submitted. These consisted of
a Process Design Manual, containing all the drawings, specifica-
tions, and engineering documents generated, and a cost estimate
for the project. |
Phase II included the procurement and construction parts of
the pilot unit. All components were procured, fabrication of
equipment and vessels was completed, the pilot plant was erected,
mechanical acceptability was demonstrated, spare parts and other
supplies were obtained, startup and operating personnel were
selected and trained, and all arrangements for the purchase of
materials and utilities was completed. In addition, an Operating
Manual was prepared.
Phase III included startup, debugging, and parametric tests.
Parameters and conditions were to be varied so as to optimize
the plant performance. The ability of the system to respond to
variations in inlet conditions was tested during this phase.
Following the successful completion of the system optimiza-
4-2
-------
^;1:ion tests, the pilot plant was to be continuously operated (24
hours/day, 7 days/week) for at least three months as Phase IV.
The period of continuous operation was to include no less than
75 days of cumulative operation in compliance with contract
objectives and guarantees.
An additional series of tests were conducted as an addition
to the original scope of work. Phase V was run to examine the
response of the SCR System's performance to transient operating
conditions. The object was to quantify any changes in catalyst
activity due to transient conditions. The transient conditions
tested would be similar to those which would be experienced by
an SCR systen operated in conjunction with a utility boiler,
including: start-up and shutdown, temperature variations, and
flow rate variations. In addition, tests were to be conducted
to examine the effects of changing the reactor sootblowing
frequency on reactor performance. Further, catalyst regeneration
procedures were to be evaluated.
SCHEDULE
The period of performance for completion of the work relat-
ed to the pilot plant was originally eighteen (18) months. The
award of the contract was May 1978. The design and specification
period, Phc*e I, took approximately three months. Phase II re-
quired approximately nine months for procurement, fabrication,
transportation and erection. These schedules were as original-
ly expected. However, start-up and debugging required about two
months and so the first tests were not started until August 1,
1979. This first charge of catalyst, however, showed less than
optimum results after some four months of operation and it was
decided to replace it in December of 1979. After some four ad-
ditional months of operation, the second batch' of NOXNON 500
catalyst began to provide less than expected results and a
decision was made to install a third charge of catalyst.
-------
Hi- T^e third charge was a new type: NOXNON 600. This
installed April 18, 1980 and was utilized for some nine
i
|until the plant was shut down on February 2, 1981.
f
Milestone Dates
Contract Awarded
i
Phase I Work Plan Submitted
Last Purchase Order Placed
Delivery of First Piece of Equipment
Mechanical Completion
First Catalyst Installed
Second Catalyst Installed
Third Catalyst Installed
Plant Shut Down Permanently
DESCRIPTION OF THE TEST PLANT
The following is a description of the pilot unit including
summaries of major items of equipment (see Figure 4-1):
Flue gas to be used as input to the pilot plant was drawn
from the boiler duct downstream of the economizer and intrcduced
to the reactor through a fourteen-inch diameter pipeline. The
boiler flue gas was withdrawn through three points from the duct
to insure that the pilot plant gas is representative of that in
the flues, particularly with respect to fly ash (dust) content
and particulate size distribution.
The pipeline was insulated and was provided with sampling
nozzles for continuous analysis of nitrogen oxides, sulfur
dioxide and oxygen.
An electric heater was provided in the piping between the
flues and the reactor. This flue gas heater was used to control
flue gas temperature to the reactor at the temperatures required.
The flue gas heater was also used to warm up the catalyst, equip-
ment, and piping to prevent corrosion by the sulfur trioxide
4-4
was
months
05/26/78
09/15/78
12/18/78
01/24/79
06/15/79
08/01/79
12/10/79
04/18/80
02/02/81
-------
1700 NrnVHr
1057 SCFM
500 PPM NOx
FLUE
GAS HEATER
EXIST
FLUE
GAS
DUCT
REACTOR
SOOT 8LOWER "
143 Lb/Hr NHj
TO FlUE
GAS DUCT
DUST SEPARATOR
AMMONIA SUPPLY
TANK
BLOWER
AIR OR
STEAM
S02
AIR
SOOT BLOWER
GAS HEATER
FLY ASH TO
NOx/SOx GENERATOR
VACUUM LINE
Figure 4-1. Process Flow Diagram of Pilot Plant
-------
A '
>->present in the flue gas when starting up from a cold state. At. .•
temperatures below the dew point of sulfur trioxide this corros-
ive compound would condense on the surface of catalyst, equipment
and piping if there were no provisions for warm-up with heated
. ai~. Finally, the flue gas heater was used for purging flue gas
from the catalyst layer, equipment, and piping with hot air
prior to the long-term shutdown of the pilot plant. This was
required to prevent the condensation of sulfur trioxide upon
cooling of the equipment.
Gaseous ammonia was injected into the gas stream after the
heater and before the reactor. The gas then flowed down through
the reactor in contact with the catalyst. The N0X in the gas
reacted with the ammonia to form small quantities of gaseous
nitrogen and water, which remained in the gas stream.
Fly ash from the flue gas settled in the reactor and Viould
have tended to block flow through the reactor and blind off and
partly inactivate the catalyst. These effects of dust settling
in the reactor were kept within acceptable bounds through the use
of a soot blowing apparatus which used either steam or air.
A blower was located downstream of the reactor to overcome
the flue gas pressure drop through the pilot plant. A cyclone
dust separator was supplied ahead of the blower to prevent eros-
ion of the blower by dust. The dust was collected in a hopper
below the separator. Dust (ash) collected in the hoppers was
periodically discharged into the boiler plant vacuum system.
The soot blowing gas heater heated compressed air or steam
used for the soot blowing apparatus. Since it was necessary to
perform soot blowing during continuous operation without shutting
down the system or reducing the flow rate of gas to be denitri-
fied, air or steam was heated up to the denitrificatio.n reaction
temperature.
The ammonia supply consisted of an ammonia tank with connec-
tion hoses, valves, safety relief valves, excess flow valves,
pressure gauge, and percentage liquid full gauge. The ammonia
¦ supply tank was filled from a delivery trailer when the ammonia
4-6
r- .-'in:; oii'C. jru-1
-------
--...v vr ii
** supply ran low. The ammonia was stored as a liquid at approxi-
j **"
jmatelv ambient temperature. The ammonia was supplied to the
i
process in a gaseous state and a small electrically heated vapor-
izer was supplied with the tank.
Since it was planned to test the performance of N0X» SC^/
SO^r and since only fixed concentrations of these substances were
available from the Albany, Ga. plant boiler operation, SC^ was
purchased for addition to the flue gas, and a N0X, SC>2 and SO^
generating unit was provided to manufacture these materials for
addition to the flue gas. This permitted testing at NC>X, SC>2
and SO^ levels higher than those available from the boiler.
The NOx/SC>x unit consisted cf an SC^ storage facility, an
air blower, two small converters, piping and instrumentation and
related control systems. One converter, for the oxidation of
ammonia, contained platinum gauze catalyst. The other converter,
containing vanadium catalyst, was for production of sulfur tri-
oxide from sulfur dioxide.
Compressed air was required for oxidation of NH^ and SC^.
The air was mixed with ammonia or sulfur dioxide upstream of the
converters. The air pressure provided the pressure necessary to
inject the produced N0X and S03 into the pilot plant flue gas
3tream. The unit was arranged in one compact skid-mounted pack-
age including a control panel. This unit was fabricated in Japan
and transported to the United States.
The operating conditions which were controlled in the pilot
unit are as follows:
1. Flue gas flow rate.
2. Temperature of flue gas.
3. The amount of charged ammonia and its pressure.
4. Soot blowing gas pressure and temperature.
CATALYST SPECIFICATIONS
The first two charges of catalyst (as described later in
Sections 5 and 6) were replaced. These were both NOXNON 500
4-7
-------
a:> < 7.
V. "
I
. X-' ¦ '• 4
V" » • . i
' f ' '• '
Figure 4-2. Photograph of the pilot plant
4-8
-------
„¦&'
^series. The .third charge of catalyst, which was used until the
jend of the test program, was NOXNON 600. Specifications of
'these catalyst changes are shown below:
Type
Size
First Charge
Catalyst
NOXNON 500
Second Charge
Catalyst
NOXNON 500
Third Charge
Catalyst
NOXNON 600
Length
270 mm
250 mm
525 mm
Width
540 mm
500 mm
525 mm
Height
1500 mm
1500 mm
2130 mm
Actual Catalyst
) Volume
0.437 M3
0.375 M3
0.587 M3
'Surface Area
245 M2
218 M2
217 M2
Pitch
8 mm
8 mm
12.6 mm
Void
0.713
0.746
0.838
Linear Velocity*
1.62 Nm/sec
1.89 Nm/sec
1.71 Nm/sec
Area Velocity*
6.94
7.80
7.83
Space Velocity*
3,890
4,580
2,900
* Flue gas flow rate at 1,057 scfm (1,700 Nm /h)
CONTROL SYSTEM
In the control system of the EPA pilot plant, the system
was operated under a set NH^/NOx mole ratio. The flue gas flow
rate and incoming NOx concentration were measured and the signal
was multiplied to provide the mass flow of NOx entering into the
reactor. This quantity was then converted to the set mole ratio
to determine the amount of ammonia required.
This system showed reliable controllability in the pilot
plant. However, the control system for commercial plants would
differ slightly from the pilot plant control system.
Due to the very large flow rate, the direct measurement of
flue gas may not be practical, and some other measurements which
are proporional to the boiler load, for example, the amount of
fuel burned in the boiler, the steam generated, or the power
4-9
-------
Bs{generated are available as representing the plant load instead
jof the direct measurement of flue gas flow rate.
i The product of the above signal and the concentration of
^N0X at the inlet of reactor would be used as a control signal to
, regulate the selected NH^/MOjj mole ratio or the selected N0X
jremoval efficiency. This feed forward system provides a fairly
jaccurate control of ammonia flow rate. Along with this the
other feedback systems which measure the outlet concentration of
•N0X, or possibly the NH^ emission, can provide signals to fine
tune the ammonia feed rate.
HOST SITE
The pilot plant was located at the Unit #3 of Plant Mitchell,
Georgia Power Co., Albany, Georgia. This unit has a pulverized
coal-fired Combustion Engineering boiler which was initially
operated April 18, 1964 with a 125 MW nameplate rating. The
boiler has a rated steam capacity of 1,075,000 lb/hr of steam to
the turbine at 1800 psig. The unit is tangentially fired with
CE standard tilting burners for steam temperature control. The
pilot plant was located outdoors of the Unit #3.
Georgia Power Co. is headquartered in Atlanta and serves
about a million customers in Georgia. Plant Mitchell has three
units, two of which have nameplate ratings of 22,5 MW while the
Number 3 is rated at 125 MW.
The boiler normally burns coal from Kentucky or other mid-
south coals with sulfur contents below 2 percent. The coal
composition can vary widely. Typical flue gas compositions to
the pilot plant are shown in Table 4-1 below.
4-10
-------
A TABLE 4-1 TYPICAL FLPE GAS COMPOSITION
Component
Concentration
N2
74.8%
°2
3-7%
CO2
13.8%
h2o
8.6%
NOx
500 ppm
so2
400 ppm
S03
5 ppm
! Gas temperatures were 550-650°F and the dust content of the
gas was typically 6.6 grains/SCFD. The pressure at the inlet
duct to the pilot plant was about 6 inches H20.
TABLE 4-2 TYPICAL COAL ANALYSES
AS BURNEO
COAL ANALYSES
DATE
MOISTURE (%)
ASH (%)
SULFUR (%)
BTU/lb
05/08/80
5.50
11.92
1.26
12,368
05/19/80
4.28
12.98
1.40
12,414
07/11/80
4.92
10.58
1.04
12,564
08/05/80
1
4.11
11.87
1.32
12,549
08/11/80
4.60
13.37
1.26
12,250
10/09/80
4.32
10.33
1.22
12,840
10/21/80
3.72
9.43
1.19
13,059
11/10/80
4.05
10.62
1.13
12,835
11/17/89
3.48
10/92
1.18
12,874
12/09/80
3.99
12.38
1.11
12,544
01/07/81
4.12
11.63
1.24
12,709
DATA COLLECTION
The pilot plant operation was closely monitored through the
use of an array of instruments and analyzers. Continuous anal-
yzers of the latest design provided measurements of the NO
X
4-11
-------
"level both entering and leaving the reactor. —,
I In addition, similar analyzers were to be used for measuring
'the ammonia slippage past the reactor. These analyzers were man-
ufactured by Thermal Electron Corporation and use the principle
of chemiluminescence. The ammonia was to be measured by split-
ting a sample of flue gas into two portions, converting the
ammonia in one of the portions to N0X, analyzing N0X levels in
each of the portions, and the difference in the analyses would
be a measure of the ammonia level.
The measurements that were recorded by CAPCC personnel
during the plant operations are shown in Table 4-3.
TABLE 4-3 DATA COLLECTED
Parameter
Temperature
Flue Gas Heater Inlet
Reactor Inlet
Reactor Outlet
Soot Blower Inlet
Pressure
Pressure drop across reactor
Gas Flow Rate
Flue Gas Flow Rate
NH3 Flow Rate
Analytical Measurements
Inlet and Outlet N0X
Inlet or Outlet SO2
Outlet NH^
Inlet O2
Other Recorded Data
Percent N0X Removal
Mole ratio NH.j/NOx
Method of Measurement
Duplex thermocouple
Duplex thermocouple
Duplex thermocouple
Duplex thermocouple
Pressure difference
Venturi flow meter
Rotameter with magnetic
float
Chemiluminescence
Pulsed Fluorescent
Chemiluminescence
»
Paramagnetic
Calculated automatically
Calculated automatically
4-12
-------
In addition to the data collected by CAPCC personnel, out-
side stack testiny consultants were employed to take air pollu-
tion measurements of S02» N^x» particulates, and HC1. This was
done to check the analytical instruments and to obtain data for
which no analyzers are installed.
Much of the data at the pilot plant was collected in a data
logger which frequently scanned the instruments and stored the
data. Visual display and paper tape display were provided. Zt
was originally intended that the data logger would store the
data on magnetic tapes in cassettes. The cassettes were to be
sent to CAPCC's New York office to be transmitted through a
terminal to a computer. This data would then be processed by
electronic data processing methods to evaluate various factors
and then stored for future use. However, the magnetic tape
recorded was found to be incompatible with the equipment in New
York and the tapes could not be used.
PROGRAM MANAGEMENT
For the overall management of the project, Mr. Shingo
Tanaka was appointed project manager by Hitachi Zosen. Mt,
Tanaka directed the progress of the work and coordinated with
EPA and with CAPCC. The Project Director, Mr. H. In^br:. 'ai
located in Tokyo and supplied overall direction, sup Tin^eivUno
the efforts of those department managers concerneJ wit'i. r&seatch,
development, design, procurement, erection, and comir-jrcio 1
aspects of Hitachi Zosen's N0X removal process.
Mr. Richard Wiener represented CAPCC on this project and
acted as Project Coordinator. He was responsible for (.he con-
tract between Hitachi Zosen and CAPCC and for liaiocn wi>:h EPA,
Georgia Power, and Hitachi Zosen.
The Project Engineer for CAPCC was Mr. Rafat Morcos who was
responsible for all in-house activities (technical and commer-
cial) during the engineering, procurement, and erection stages
of the project. The Field Engineer was Mr. Atef Demian.
4-13
-------
The Project Officer for EPA was Mr. J. David Mobley who was
responsible for the entire project. Working with EPA was the
Radian Corp. of Austin, Texas providing technical support and
reviewing all aspects of the project.
EXPECTED RESULTS
Tlte purpose of the program was to demonstrate the applica-
tion of the catalytic process to coal-fired boilers for 90 per-
cent reduction of NOx emissions. It was expected that the pro-
gram would demonstrate the applicability and efficiency of the
process for the removal of high percentages of nitrogen oxide
from flue gas typical of that produced from boilers burning O.S.
coal. The program tested several parameters so that when the
tests are completed it would be possible to predict the removal
efficiencies under various conditions of operation.
The energy requirements could be evaluated by determining
the pressure drops across the catalyst reactor over a period of
time. Unexpected increases in pressure drop would of course add
to the energy needs of the process. Sootblowing capabilities
could establish the effectiveness of this approach to the main-
tenance of low pressure drops.
Materials of construction, although not a problem in this
process, were evaluated to determine if unusual corrosion
effects were found. Most of the equipment was of carbon steel
construction which was suitable for this process. High temp-
erature operation could cause problems with some materials and
this was to be studied.
The analytical system for NOx, SC^, and NH^ was of consid-
erable interest and the effectiveness, accuracy, and reliability
of these instruments could be applied to commercial applications.
Chemiluminescence analysis, if successfully applied here, would
be used in full-scale plants. The SC^ analyzer (Pulsed fluor-
escent) would also be followed closely for future use.
The control system was essentially the same as that which
4-14
-------
would be used for commercial plants. The normal control method
would use a feed-forward control in which the flue gas flow and
N0X concentration in the feed will be used to set the ammonia
feed. The success or failure of the control methods could be
applied to commercial designs.
Additional items which were tested during the demonstration
program are the instruments and controls and the handling by data
logger. Certain variables such as "percent N0X removal" and
"mole ratio" were calculated and recorded. All of these systems
were evaluated for commercial applications.
RELATING THE RESULTS TO A FULL SCALE OPERATION
The test unit included a section of catalyst which had the
same depth as that to be used ^n a full-scale plant. The results
could therefore be directly extrapolated to a large plant assum-
ing that the velocity, mole ratio, and temperature remain the
same and the gas distribution through the catalyst bed was prop-
erly designed. From these pilot plant results the gas velocity
required to produce a certain reinoval efficiency could be deter-
mined and the amount of cataJyst required could, therefore, be
calculated.
Ammonia consumption could also be determined from the test
results because it is calculated as a mole ratio of ammonia to
N0X in the flue gas. Ammonia slippage would also be expected to
be the san:? in a commercial unit.
Energy consumption is a function of the pressure drop
through the catalyst bed and through the reactor inlet, reactor
outlet, dampers, ducts, etc. This test work could only be used
to establish the expected pressure drop through the catalyst bed.
A commercial unit would produce different pressure drops depend-
ing upon the plant arrangement. However, these pressure drops
can normally be calculated and the major pressure drop cf inter-
est is the one through the bed of catalyst.
The results from the pilot plant should be close to that
4-15
-------
¦expected in a,commercial unit except ror variations in the type
of ash and the design of the soot blowers. The expected life
of the catalyst is about two years. The testing, however, could
only extend about six months. Extrapolation of any measurable
decrease in reactivity would give an indication of the expected
catalyst life.
4-16
-------
SECTION 5
NOXNON 500 TESTS
PRELIMINARY OPERATIONS
Oil June 4, 1979 construction of the pilot plant was essen-
tially completed. Checking of the piping and instruments was
then begun by CAPCC and HZ personnel. Orientation lectures for
the training of newly hired operators was started a week later
and trial practice runs were provided for the operators starting
on June 18th. i
Debugging of the plant required several weeks and the plant
was not ready for operation until the end of July. Many of the
problems were related to the electrical work. This included
errors in the wiring installations, faulty connections in the
motor control center, and mistakes in the connections to the
control panel. In addition to the electrical system, several
adjustments were required for the analyzer system to make it
operable. Another problem was with thi blower motor which burn-
ed out and required replacing.
Before the NOXNON 500 catalyst was loaded into the reactor
the whole flue gas system was operated with circulating air.
This was for training of the operators, for calibration and
adjustment of the controllers, for testing of the Blower and Flue
Gas Heater, for checking of the sample lines to the analyzer, and
for general testing of the whole system.
On July 16, 1979 continuous operation
-------
^ammonia supply system, NO /SO generator, and ash removal equip-
* x
ment. ,
j Although calibration and adjustment of the analyzers and
controllers were not completed, it was decided to install the
catalyst into the reactor on August 2nd, which was approximately
one month later than originally scheduled.
FIRST CHARGE
On August 2, 1979 the first charge of NOXNON 500 was loaded
into the reactor and the testing commenced. Testing continued
until the beginning of December providing 2168 hours of operating
time on flue gas and 15 4 hours on hot air.
Overall data with the first charge of NOXNON 500 is shown in
Figure 5-1. When operations were initiated with flue gas at the
beginning of August, 1979, the removal efficiency was better than
90 percent at design conditions at 1.0 mole ratio. The removal
seemed to stabilize at approximately 92-93 percent with a 1.5
inches t^O pressure drop across the reactor. This efficiency
was attained until the middle of September.
On September I4th, after about 900 hours of operating time
the pressure drop suddenly increased and the removal of NOx
dropped sharply. The next day the plant was shut down and the
top surface of the catalyst layer was inspected from the man-
hole of the reactor. About half of its cross-section was found
to be covered with a thick layer of fly ash. It appeared that
the fly ash had built up on the side of the reactor in front of
the manhole and had then slid down for some reason covering a
large part of the catalyst. It was decided to close up the
reactor again and operate the soot blower for several cycles
with superheated steam. This reduced the pressure drop from
1.85 inches to the normal 1.4. Removal efficiency also increased
significantly.
It should be explained that up to this time the sootblower
had not been operated. This was based on experience at the
5-2
-------
CATALYST CATALYST
INSPECTION REPLACEMENT
100
CD
o
05
AUG / 79
SEP
OCT
NOV
DEC /79
Figure 5-1.
REACTION TEMPERATURE , APPROX. 700*F
FLUE GAS FLOW RATE , 1057 SCFM
Operating Data with the First Charge of NOXNON 500.
-------
• 'iVI ' _ _jA '• "
'¦<" Mlsogo Power Station pilot plant in Japan. Similar reactors
h < c r I
; there operated on fly ash loaded flue gas for long periods of
j '
itime without the need for sootblowing. Clogging by fly ash
i '
idid not occur. The sootblower was installed at the EPA pilot
, plant only for the purpose of emergency use. It was not plan-
ned to be run regularly. However, differences in the fly ash
^characteristics were obvious and the decision was made to
| operate the sootblower at least once a day from this date.
! By the middle of October (operating time 1170 hours) the
pressure drop appeared to be gradually increasing. The pressure
i
drop continued to increase and on October 30th there was a rapid
lincrease to 3.2 inches f^O. The sootblower was operated repeat-
edly so as to continue the operation. This measure brought a
stable pressure drop of approximately 2.4 inches H^O, still some-
what high. However, the N0X removal efficiency was not restored.
On November 9th the reactor and ductwork was dismantled to
find the reasons for the low removal efficiency. It was sur-
I
prising to discover that the ductwork leading to the reactor was
heavily built up with deposits of fly ash. About three-quarters
of the elbow before the reactor was plugged. The reason for the
fly ash deposit was the expansion of the gas in the transition
piece which was located in the horizontal. The lowered velocity
permitted the particulates to drop out at that point. Deposits
were also found at the sootblower opening and at the manhole
shelf.
! The catalyst assembly could only be cleaned by lifting it
out of the reactor and setting it on the ground. The upper sur-
face of the catalyst looked to be full of fly ash. Most of the
openings were full. An air hose was used to clean out most of
the fly ash. Further quantities of fly a3h were removed by
poking rods into the openings.
The ductwork was modified so that the transition piece was
located in a vertical position above the reactor. Baffle plates
'were installed at the manhole and at the sootblower opening to
prevent fly ash deposits at those locations. The catalyst was
5-4
1 GU'.Ds. 1
-------
Hreinstalled and the system was bolted together and put back into
l "
|operation on November 2 3rd.
I The initial results indicated better NO removals. However,
I X
,after about a half-day the removal dropped below 90 percent
although the pressure drop was substantially lower than before
cleaning of the catalyst.
i
On November 30th, the system was shut down for an inspection
of the catalyst through the manhole. The surface looked clean
and the changes in ductwork, along with the baffles at the man-
hole and at the sootblower, had been effective in avoiding piles
of fly ash. However, inspection of catalyst by insertion of rods
:indicated that about 30 to 35 percent of the openings were plug-
ged again.
The catalyst in place was rodded out as far as possible and
compressed air was also used to try to clean it. After this
effort, the system was once more operated to evalute the NOx
removal efficiency by varying the mole ratio. NOx removals were
found to be mostly in the 80-85 percent range even with sootblow-
ing three times per day.
A decision was made to replace the catalyst as it appeared
that the activity had decreased to the point where the expected
90 percent removal could not be attained. The catalyst had been
in service some 2500 hours. The reason for the activity decline
was not certain. However, it appears that there were two major
reasons. One was the buildup of the fly ash within the layers
of the catalyst caused in part by the inappropriate duct design
thus blocking off ^ portion of the surface. The second major
cause was a definite detrition of active material from the
metallic substrate. Possibly 10 to 20 percent of the catalyst
may have been lost.
On December 10, 1979 the original catalyst load was replaced
by the spare catalyst which had been stored at the site. The
removed catalyst was inspected and tests were run in an effort
to determine the cause of the loss of activity. This is discus-
sed in a later section, "Catalyst Evaluation."
-------
SECOND CHARGE
j The second charge of catalyst was loaded into the reactor on
December 10, 1979. This charge was the same as the original ex-
•cept that it was in six blocks instead of four. The total vol-
ume, composition, and pitch were the same.
The system was started up again on the evening of December
10th and removals of over 90 percent were readily attained.
Figure 5-2 shows overall performance with the second charge of
NOXNON 500 catalyst. Based on the experience of the first charge,
the sootblower was operated three times a day from the beginning.
This frequency seemed the minimum required to prevent clogging by
fly ash. The system was operated until March 3, 1980 with a
stable 90-92 percent NOx removal and a pressure drop of less than
two inches f^O.
Typically, over a two day period of February 25th and 26th
the NOx removal averaged 93.1 percent. The pressure drop across
the catalyst bed was 1.72 inches I^O. (These results are all at
a flue gas rate of 1057 SCFM, a reactor inlet temperature of
710°F, and a mole ratio of 1.0).
On March 4th the pilot plant was shut down so that the flue
gas heaters could be replaced. Several of the elements had burn-
ed out. The shut down and start up followed the normal procedure.
When the system went back into operation, it was found that the
operating results were significantly different. The following
data are for the indicated days showing the average of percent
NOx removal and pressure drop:
3/5 91.8% 2.43 inches H^O
3/6 92.8 2.25
3/7 91.4 2.45
3/9 89.8 2.59
There was a small but significant drop in removal but a
relatively large increase in pressure dro^-. All of the instru-
ments and controls were thoroughly checked and found to be sat-
5-6
-------
CATALYST
catalyst INSPECTION INSPECTION ano cleaning
100
rtr
—1^3#
u_
u.
Ui
: APPROX 700 'F
FLUE GAS FLOW RATE • APPftOX 1,057 SCFM
REACTION TEMP
REPLACEMENT OF ELEMENTS
IN FLUE CAS HEATER
uj
ce
o
z
o
o
N
X
5 10
| 2.0
o
UJ
cc
Q.
DEC /79
JAN /80
FEB
MAR
APR /SO
Figure 5-2. Operating Data with the Second Charge of NOXNON 500.
-------
isfactory. The cause of the pressure drop increase was of
concern.
It was decided to inspect the surface of the catalyst as
there was a possiblility that the shut down may have caused an
influx of fly ash onto the catalyst. The system was shut down
again by proper procedures, the manhole was opened for inspec-
tion, the inspection was made, the manhole was closed and the
system restarted. No unusual accumulations of f]y ash were seen
on the top of the catalyst bed.
Surprisingly, when the system was restarted the removals
decreased substantially and the pressure drop went up even higher.
3/10 84.0% 3.25 inches H20
3/11 85.6 3.22
No cause for this drop in performance could be found and
once more all the instrument systems were checked and found to
be working well.
On March 14th the system was closed down again, the catalyst
was removed, each of 'the six blocks were cleaned with compressed
air, and then replaced. The pressure drop went down but the
removals did not improve. It appeared that there was a perman-
ent loss of catalyst activity.
3/15 36.2% 2.12 inches H^O
3/16 86.6 2.09
3/17 87.0 2.00
3/18 86.2 2.04
On March 21st the boiler was shut down for two weeks of
maintenance. It was decided to circulate hot air through the
reactor so as to avoid cooling it down. For two weeks the temp-
erature was maintained at 400°F and then increased to 550°F two
days prior to the reintroduction of flue gas. On April 9th the
pilot plant was restarted and was run until the next day when
the boiler was forced to go down due to a tube leak.
The results were even worst this time. A six hour period
on April 11th indicated an average removal of 78.2 percent and
5-8
-------
¦a pressure drop of 2.17 inches. (At a mole ratio of 1.01, 3056
SCFM, 724°F). For some reason the removal had decreased from
the previous 86 percent even though no flue gas had been run
through it and a cooling and reheating cycle had not been ex-
perienced .
There was no apparent cause found for this degeneration.
However, it seemed obvious that the catalyst was no longer
acceptable. The catalyst was then removed and a new charge of
catalyst was installed on the week of April 14th. The catalyst
had been in operation with flue gas for 218 4 hours.
When the Type 500 catalyst which had been removed from the
reactor was inspected it was found to be in relatively good
physical shape. For an inspection, several plates were stripped
off of several of the blocks. (There were six blocks piled in
three layers). There was certainly a quantity of fly ash depos-
ited within the internal layers, but when this was brushed off
or the plate was lightly tapped, the surface appeared mechanic-
ally sound. Except for the top two or three inches of the upper
layer, there was no indication of erosion. Also, there was no
sign of delamination, cracking, or peeling of the active naterial.
The upper few inches was worn down to the stainless base due to
the action of the fly ash or soot blower. But <.he remainder of
the five feet length of catalyst was in excellent shape.
The cause of the loss of activity could have been due to
the accumulation of fly ash or to a coating of the surface by
some foreign material or to some other reason. Samples were
prepared and sent to Japan for examination along with samples of
fly ash. See "Catalyst Evaluation" for the results of these
tests.
A thorough investigation of the operating data during the
shut downs and start ups was done and it indicated that all
proper procedures were followed. There was probably no chance
of sulfur trioxide condensing during any of these times because
when flue gas was present the temperatures were always above
400°F, which is considerably above the dewpoint.
5-9
-------
r*:vrr
-------
_ ^ _ ^
RrjCatalytic activity-- ; 1
I Test pieces for catalytic activity examination were trken
t
|from the top catalyst blocks. Figure 5-3 describes the location
of "est pieces cut out of the top catalyst block. Figure 5-4
. shows a schematic explanation of the testing apparatus.
Experimental Conditions—The following conditions were
maintained in the testing apparatus:
Area Velocity : 20 Nm3/m2.h
Size of Test Pieces : 20 mm x 50 mm
Composition of mixed gas (volume on wet basis)
NO
390 ppm
nh3
390 ppm
so2
225 ppm
°2
3.6%
CO 2
10.8%
h2o
10.0%
N2
balance
Experimental Data—Catalytic activity of the test pieces
were examined along with three virgin test pieces which were
produced at the same time as the first charged catalyst and kept
in storage by Hitachi Zosen.
Catalytic activity of the unwashed used catalyst was tested
at first with the test pieces. Afterwards they were tested after
washing with water at 40°C (104°F) in order to remove adhered rly
ash from the activated surface, and then compared with the test
pieces of virgin catalyst.
Results—Figure 5-5 shows the results of catalytic activity
testings.
- Catalytic performance of the upper section of used
catalyst which was abraded by fly ash or by soot blowing
with steam had decreased considerably.
- Catalytic activity of the lower section which had not
5-11
-------
o
GAS FLOW
ABRAOEO
TEST PIECE
ABRAOEO
125-30 T.)
UN-ABRADED
TEST PIECE
CATALYST PLATE
Figure 5-3. Location of Test Pieces.
5-12
-------
HEATING TRACE
REACTOR I 30mm & I
6'
ELECTRIC HEATER
WATER BATH
CONDENSER
VCW
CONDENSER
kCW (
FLOW
METER
EXHAUST
CW
COOLER
COOLER
EXHAUST
O.TAIN
Nj-0j«C0j Nj^OjNfHIO Nj*NHj
Figure 5-4. Test Equipment for Determining Catalyst Activity.
-------
Symbol
CAMITST / DESCRIPTION
VIRGIN CATALYST
UN-ABRaOEO UfTtR WASHING I
UN-A8BADED I BEFORE WASHING)
AURdOED
100
<3
>
O
3
•u
•K
O
r
250
300
350
REACTION TEMPERATE (V.I
TEST CONDITIONS
TEST PIECE SIZE 20™" 140mm (PLATEI
A V 20 NmJ/m* H
GAS COMPOSITION I WET BASIS I
NO 190 PPM
NHj 190 PPM
S02 225 PPM
0? 3 6 */•
COi 10 8 V.
HjO CO T.
BALANCE
NHj /NO MOLE RATIO = t 0
Figure 5-5. Evaluation of Used NOXNON 500 (First Charge).
-------
abraded but was masked by adhered fly ash also decreased.
- Testing of the lower section after water washing showed
a slight decrease of catalytic performance in comparison
with the virgin catalyst. However, it was concluded that
the catalytic performance of the lower and unabraded
section still retained its original activity after the
adhered fly ash was removed from the surface.
Inspection of the Second Charge of Catalyst
T"ue second charge of catalyst were taken out of the reactor
and disassembled on the ground on April 11, 1980. There were six
blocks of catalyst. Two at the top, two in the middle, and two
at the bottom.
Abrasion—
Abrasion was observed mainly at the upper part of the top
catalyst blocks and slightly lower side of the top block. There
appeared to be no abrasion in the middle and bottom blocks. The
ratio of abraded surface area to the total surface area was
estimated to be about 10 percent. Abrasion seemed more severe
under the sootblower discharge nozzles.
Plugging—
Inspections were first made of the six blocks of catalyst
prior to removal of the catalyst from their containers. The
upper surfaces varied in the degree of apparent plugging. The
top catalyst appeared relatively clean with only about 5 to 7
percent of the channels plugged. The middle blocks, however,
were significantly plugged, estimated at 40 to 50 percent of the
total number of channels. Also, about 25 to 30 percent of the
channels in the bottom blocks appeared plugged1.
Three bundles of catalyst were taken out of the containers,
one each from the top, middle, and bottom. These were dis-
assembled for inspection. A substantial amount of adhered and
5-15
- f ..'I.j' L I
-------
accumulated fly ash was found in a rrajority of the channels
between the catalyst plates, especially in the middle bundle,
although narrow paths for flue gas still remained.
Catalytic Activity—
The test pieces for catalytic activity examination were cut
out of the top and bottom catalyst bundles respectively. The
size of test pieces, testing apparatus, and the experimental
conditions were the same as those used for the first charge of
catalyst.
Experimental Data—Catalytic activity of the test pieces
taken from the top and bottom catalyst bundles were tested
before and after washing with water and shewn in Figure 5-6
and was compared with the fresh catalyst.
Results—Catalyst activity before water washing was consid-
erably lower than the original. However, with water washing the
activity was restored almost to the original activity.
Investigation of the Deterioration of Catalyst
The application of NOXNON 500 to coal-fired flue gas was
examined at the pilot plants of EPDC*3 Isogo Power Station and
at the Georgia Power Company's Plant Mitchell. Radical differ-
ences between the two plants in respect to clogging by fly ash
were observed.
At the EPDC pilot plant, the flue gas source had been div-
ided into two streams. One stream was pre-treated through an
electrostatic precipitator or cyclone dust separator, producing
a low-fly-ash-containing flue gas. The other stream was not
pretreated, so that high-fly-ash-containing flue gas was sent
directly into the catalytic reactors. There were not soot blow-
ers operating in either case.
At the EPDC pilot plant, clogging by fly ash did not occur.
5-16
-------
l/l
I
100
_ 00
— 60
> 40
jYhbol
CATALYST / OE SCRIP HON
VIRGIN CATALYST
TOP CATALYST I AFTER WASHING I
TOP CATALYST I BEFORE WASHING I
BOTTOM CATALYST (AFTER WASHING)
BOTTOM CATALYST (BEFORE WASHINGI
2
20
250 JOO 350 400
REACTIOH TEMPERATURE ( T.)
450
TEST CONDITIONS
TEST PIECE SIZE - 20mm 150™" (PLATE I
i V 20 Hn^/m? H
GAS COMPOSITION (WET BASIS I
NO 390 PPM
NHj 390 PPM
SO* 225 PPM
< Oi 3 6 1.
COj 10 8 V.
HjO 10 0 'A
N, BALANCE
NH3/NO MOLE RATIO = 10
Figure 5-6. Evaluation of Used NOXNON 500 {Second Charge).
-------
The low-fly-ash-loaded flue gas produced an increase in pressure
drop, but no clogging. The high-fly-ash-containing flue gas also
did not cause clogging and there was no increase in pressure drop.
Pressure drop was stable and constant during more than 8,000
hours of operation.
However, at the EPA pilot plant NOx removal efficiency de-
creased and pressure drop increased with both the first charge
and the second charge of catalyst after about 2200 hours of
operation. A major factor for these undesirable results was the
character and the behavior of the fly ash which was obviously
different from that of the fly ash at the EPDC pilot plant.
In order to investigate the cause of the decrease in NOx
removal efficiency and increase in pressure drop, the following
studies were conducted.
Measurement of Particle Size Distribution of Fly Ash—
Samples—The following samples were collected:
Sample
A
B
C
D
E
F
G
Date Location
09/15/79 upper surface of catalyst bed
11/08/79 top catalyst block
11/08/79 bottom catalyst block
11/08/79 upper surface of catalyst bed
04/17/80 top catalyst block
04/17/80 middle catalyst block
04/17/80 bottom catalyst block
Method—Particulate size distribution was measured by
electron microscope.
5-18
-------
Results-
Sample
Median
Diameter
(Micron)
Rosin-Rammler's
Index No.*
Log Standard
Geometric
Deviation
A
B
C
D
E
F
G
13
21
15
16
18
19
18
3.2
3.0
3. 3
3.3
3.1
3. 4
3.2
1.6
1.7
1.8
2.0
1.8
1.6
1.6
* The above Rosin-Rammler*s distribution is based on ANSI/ASTM
E 20-68 and DIN 66145.
The results showed no significant difference in particle
size at the various sampling points.
Pore Size Distribution of Catalyst—
Pore size distribution of NOXNON 500 catalyst was measured
by a Mercury Penetration Method and is shown in Figure 5-7.
The results indicate that the pore size of the catalyst i3
very small, measured in angstrom units, compared to the particle
size of the fly ash, measured in microns. There could be no
possibility of the pores of the catalyst being plugged by fly
ash.
Chemical Analysis of F-l-y Ash—
Methods—100 grams of fly ash were pulverized to less than
100 mesh, dried at 120°C for two hours, and one gram of this was
weighed accurately.
This one gram of fly ash was mixed with concentrated HC1
and HNOj and evaporated to dryness, heated with HC1 solution,
dissolved and filtered.
Al, Fe, v, Ti, Mg, Ca, Na, K and Li wore then determined by
an atomic absorption spectrophotometer from the filtrate. 10%
5-19
-------
0)
>
N
o
>
w
>
9
a
3
w
(0 I pml
11 IB)
10*
PORC RADIUS |t I
Figure 5-7. Pore Size Distribution of NOXNON 500.
-------
of BaCl2 solution was added to a definite amount of filtrate,
and SO^ group was determined by measuring the weight of precip-
itated BaSO^.
The residue wis ignited at 800°C and SiC>2 in the residue
was vaporized with HF, and a decrease in weight was determined
as the amount of SiC^.
Non-volatile residue was melted with 1<2S207, dissolved into
water, and the amount of Al, Fe, V, Ti, Mg, Ca and Li was deter-
mined by an atomic absorption spectrophotometer and added to the
amount included in the filtrate.
Ammonia was extracted with hot water and determined by an
Indophenol process.
Results—In order to investigate the characteristics of
Georgia Power fly ash, a comparison was made of the chemical
analysis of the Georgia Power fly ash and a typical fly ash
produced from a Japanese utility company.
Table 5-1 describes the chemical analysis of the fly ash
collected from Georgia Power and a Japanese utility. Rf shown
at the bottom of the table was cited from "K.H. Haller: Design
of Large Coal Fired Steam Generators, Babcock & Wilcox Co.,
Technical Paper BR-1082, 1-8 (1977)," and shows the fouling
tendency of fly ash in relationship to the composition of the fly
ash as described in Table 5-2.
The analytical results in Table 5-1 show no particular
reason for the causes of the shorter catalyst life in the EPA
pilot plant. From the fouling factor evaluation (R^) it would
seem that Japanese fly ash would have more of a tendency to foul
than the Georgia Power fly ash. The only significant analytical
difference appears to be the potassium level which is about ten
times higher in the Georgia Power Co. fly ash. However, it is
not known whether this would be a factor that would increase
fouling tendencies. A3 neither the chemical analysis nor the
fouling factor seemed to offer an explanation for the shorter
catalyst life, further investigations were required.
5-21
-------
T*BL£ 5-1
CSXIGAt AKALTSIS CP T1* JsStl fvt. I]
cmncLA rr> n gxpAW plv/t nnqini*
QTT£
SVPUD
Oct/70
9/15/79
ii/a/79
U/VJ9
11/8/79
4/17/BO
4/17/80
4/17/80
LOCAJICW
SWUD
ESP
TCv Of
f "alyat
Uwa*
Catalyst
Datum
CauUyBt
Ttp of
Citalyat
Lfl/Vr
*ftjp
Catalyst
HiddU
Calalyst
Bottom
Catalyoi
rujc g\s
high rjf K3i uy^rc fuz c*s
SiOj
A1>0
SO. 5
53.1
52.0
51. •
51.7
46.0
44.0
41.0
29.1
26.2
26.3
26.)
as. 7
29.3
27.0
26.8
re2°!
VJ°S
11.Q
9.0
1G.2
9 4
9.7
10.0
10.7
».«7
0.04
0.0*
0.04
0 11
0.04
0.056
0.071
0.161
TlOj
2.3
1.8
i a
2.0
i.a
2.84
2.64
3.34
tli-0
0.06?
o.u
0.14
0.16
0.14
0.162
0.162
0.189
V
0.241
0.96
0 96
1.75
0.95
i.ib
1.43
1.45
UO
0.034
0.079
C 069
0 073
0.056
0.022
0.024
0 026
HP
0.36
0.78
0 91
0 76
0.01
0.75
0.90
o a:
Ca0
0.161
0.01
0 60
0 07
0.29
0 38
0.15
0 15
5.1
1 2
7.4
5 2
2.0
3.4
6.0
6.5
,J,4
nil
0.005
0.012
0.006
0.0L9
0.006
0.011
0.014
TOTAL
9?. 30
94.13
95.43
97.23
91 21
93 0
91.3
90.3
Rf
0.26
0.06
o.u
0.37
0.20
0.18
0.36
Q.3J
TYPICAL JAPAffl&C F1.Y f£ll
Low
Fly Aflh
U3W
Fly Ash
nigh
riy toh
Itigh
riy /nti
Low
riy fKsh
Low
fly Mh
44.)
45.1
46.0
47.8
41.7
44.5
21.4
31.5
22.3
22.3
20.3
22.5
5.0
5.2
6.3
6.4
12.0
5.7
0.071
0.079
0 024
0.024
0.034
0.055
1.37
1.44
1 22
l.W
1.34
1.2?
0.96
0.90
0 36
0.28
0.35
0.59
o oas
0.082
0.162
0.165
0.191
0.246
0 027
0 025
0.010
0.011
0 011
0.017
1.G9
1.76
2.r J
1 9 C
2 03
i.«e
0,80
1 05
6.34
6 86
4.55
4.72
10 8
r.i
3 2
1 9
4.3
5.3
0 297
0.044
0 010
0 OS
0.020
0 013
86 8
06.3
B6.e
B9 0
B7.3
86.8
0.46
0.10
0.23
0.14
0.43
0,34
-------
TABLE 5-2 FOULING FACTOR OF FLY ASH
Bituminous Ash
Defined as (CaO + MgO) < Fe203
CaO + MgO + Fe203 + Na20 + K20
Fouling Factor (Rf) = si02 + A1203 + Tlo2
Rf Fouling Tendency
below 0.2 low
0.2 - 0.5 medium
0.5 - 1.0 high
above 1.0 severe
Lignitic Ash
Defined as (CaO + MgO)> Fe203
Na20 (%) Fouling Tendency
below 3 medium
3-6 high
above 6 severe
5-23
-------
Influence on Catalyst Activity Caused by Fly Ash
In order to evaluate the influence on catalyst activity by
fly ash a test piece of catalyst (50 mm x 20 mm) was put into fly
ash and kept at a temperature of 380°C (716°F) for approximately
500 hours. The fly ash had been collected at Georgia Power Com-
pany, Plant Mitchell on September 15, 1979 from the upper surface
of catalyst layer, and chemical analysis data is shown in Table
5-1.
200 milliliters of this fly ash was put into a steel cylin-
der and the test piece was buried in the fly ash and kept at
380°C for 500 hours in an electric furnace. This method, though
simple, demonstrates that when any component which is poisonous
to catalyst activity is present in fly ash, the catalyst is
deteriorated by this te3t.
The catalytic activity of the test piece after this test is
shown in Figure 5-8 along with its initial activity. The results
proved that there was no component in the fly ash which deterior-
ated catalytic performance of NOXNON 500.
Observation of Fly Ash by Spectroelectronic Microscope—
By observing fly ash with a spectroelectronic microscope,
Georgia Power fly ash was found to contain some porous substance
which may have been an unburned hydrocarbon or unburned pulver-
ized coal. Japanese fly ash does not contain these porous
substances.
Shown on Figure 5-9 are microphotograpl.s of Georgia Power
fly ash at magnifications of 2700 and 900. They suggest that
some high boiling point compound condensed and acted as an
adhesive leading to agglomeration or conglutination of the
particles of fly ash.
Thermal Processing of Fly Ash—
The above microscopic examination of the fly ash indicated
the possibility of a compound found in Georgia Power fly ash
which was a high boiling point compound. However, the mechanism
5-24
-------
SYMBOL TEST PIECES PROCESSED
O VIRGIN CATALYST
~Z 500 HRS ON STREAM
1001
60 -
40 -
20 -
TEST CONDITIONS
A V , 24 NmJ/m2 H
GAS COMPOSITION (WET BASIS)
NO
220 PPM
NHj
220 PPM
02
5 7.
CO?
(0 7.
HZ0
10 v.
N2
BALANCE
NH3/N0
M R = 10
PROCESSING TEMP : 380 *C
FLY ASH FROM G
300 350 400
REACTION TEMPERATURE <7.1
Figure 5-8. Effect of Thermal Processing in Fly Ash.
-------
FLY ASH FROM ESP (x900) FLY ASH FROM ESP (x2700)
FLY ASH FROM TOP CATALYST
(x900)
FLY ASH FROM TOP CATALYST
(x2700)
Figure 5-9. Microphotographs of Fly Ash
5-26
-------
s^of agglomeration was not clear. Therefore, the behavior of fly
aBh mixed with coal at the operating temperature was investiga-
ted .
Since the temperature and time are functions which influence
adhesion and agglomeration of fly ash, the following qualitative
tests were run using Georgia Power fly ash as well as EPDC fly
ash to investigate cohesive characteristics of Georgia Power fly
ash.
Fly ash from Georgia Power and EPDC was screened on a 70
mesh screen, introduced into lidded steel cylinders of 30 milli-
meters diameter, and processed at the conditions listed below in
an electric furnace.
PRETREATMENT
Step 1 screened through 70 mesh
sieve.
Step 2 after Step 1 is finished,
again screened through
70 mesh.
Step 3 after Step 2 is finished,
screened through 70 mesh
sieve and mixed with 5%
active carbon.
TEMPERATURE
380 °C
380 °C
380°C
380 °C
TIME
100 hours
150 hours
150 hours
200 hours
Step 4 after Step 3 is finished,
screened through 70 mesh
sieve and mixed with 5%
pulverized coal from EPDC.
Figure 5-10 show photographs of fly ash after these
thermal processing steps were completed.
The photographs from Step 1 at the top show that Georgia
Power fly ash produced large size agglomerates which have the
same diameter as the internal diameter of the steel cylinders
used for the thermal processing after Step 1 processing, whereas
EPDC fly ash did not agglomerate.
The photographs from Step 2 show that Georgia Power fly ash
again produced agglomerates although the size was smaller. EPDC
fly ash did not agglomerate.
From the Step 3 photographs it was observed that 5% active
5-27
i s iV.i ti'J'"1" ' ' 'i
-------
tn
I
Kj
00
Step 1 Processing
Y}y Ash of GPC
Screened by 70 mesh sieve.
3 80 °C x IOC Hours
-ic-y*
' * * • • ,-5
' * ' ' ' Kf£ ft
... ¦.
2 Processing (GPC)
Step
After Step 1, screened
again by 70 mesh.
380 °C x 150 Hours
Figure 5-10. Photographs of Fly Ash
W'\
•UH". . « •' • > .' -
after Thermal Processing.
-------
Ul
I
10
VD
[Step 4 Processing tEPDC)
j After Step 3, screened
, again and mixed with
pulverized coal.
380°C x 200 Hours
** -Step 3 Processing (GPC)
¦ ---fi&msv&tnTJui jiwj» .
After Step 2, screened
again and mixed with
active carbon.
3 80 °C x 150 Hours
After Step
again and mixed with
active carbon.
380 °C x 150 Hours
After Step 3, screened,
again and mixed with k..
pulverized coal.
380°C x 200 Hours
Figure 5-10 (continued). Photographs of Fly Ash after Thermal Processing
-------
carbon mixed with fly ash seemed not to promote the cohesion of
fly ash. However/ a small amount of granular agglomerates were
found in Georgia Power fly ash.
The bottom two photographs showed proof that Step 4 process-
ing produced large size agglomerates for Georgia Power fly ash-,
whereas there were very small size agglomerates in EPDC fly ash.
From the above quantitiative testings,
- Georgia Power fly ash appears more cohesive than EPDC
fly ash.
- Onburned pulverized coal or hydrocarbons derived from
coal appear to seriously increase the cohesion of fly
ash at the operating temperature of SCR.
Discussion
NOXNON 500 catalyst was replaced twice due to an increase
in pressure drop and a decrease in NOx removal efficiency after
operations of approximately 2,100 hours and 2,300 hours.
Phenomena observed were:
Catalytic activity did not deteriorate—
After the fly ash which adhered to the surface and was plug-
ging the channels of the catalyst was removed, the catalyst it-
self retained its original catalytic performance.
Fly ash produced in the Unit No. 3 boiler does not include
components poisonous to the catalyst.
Abrasion was observed at the upper part of the catalyst-
It is unknown whether the abrasion was caused by ash-cut or
condensate in the steam used for the soot blower. However, the
ratio of abraded surface to the total catalyst surface was not
high, probably less than 10 to 15 percent. The abrasion, of
course, contributed to the decrease in NOx removal efficiency,
but this was not a major factor compared with fly ash plugging.
5-30
-------
-Accumulation of fly ash on the catalyst surface—
Since this accumulation was caused mainly by electrostatic
forces (Van der Waals Force), it could be removed easily by the
soot blower operation. During catalyst cleaning a tremendous
amount of fly ash was removed by blowing with compressed air.
Chemical analysis data of fly ash—
These did not lead to any definite explanation for the plug-
ging of the catalyst channels by fly ash. However, fly ash pro-
duced in the Plant Mitchell Boiler No. 3 does agglomerate at the
reactor operating temperatures. Inspection of the catalyst,
which was disassembled at the site, indicated that the agglomer-
ated fly ash caused plugging of the channels between the catalyst
layers. The mfichanism of agglomeration is not clear. However,
since the agglomeration leads to adhesion of fly ash in the cat-
alyst channels, fly ash adhering to the catalyst surface should
be removed as soon as possible by soot blowing.
Soot Blower Operations—
It was concluded that the soot blower should be operated for
the treatment of Appalachian-produced coal-fired combustion flue
gas. The minimum operation frequency was determined from exper-
ience.
However, a problem still remains. There seems to be a
possibility that abrasion of catalyst was caused by condensate
which existed in steam pipes between the nozzles of soot blower
and the block valve when the soot blower was not operated. Also,
this condensate may cause agglomeration of fly ash leading to
plugging of the catalyst. Therefore, steam for the soot blower
was repaced with air to prevent possible abrasion caused by con-
densate.
5-31
-------
SECTION 6
NOXNON 600 TESTS
INTRODUCTION
Primarily because of clogging of the catalyst channels by
fly ash due to narrow clearances and partially due to abrasion
by fly ash and by soot blowing, the initial charge and the second
charge of catalyst did not achieve their expected performances.
The installation and testing of NOXNON 600 catalyst which has
wider channels was proposed by Hitachi Zosen. This plan was
accepted by EPA.
The development of NOXNON 600 catalyst is based on the con-
cept that a highly active catalyst which consists of thin plates
should be the most effective. Therefore, NOXNON 600 is produced
from thin stainless steel wire mesh as a base metal to give mech-
anical strength to which catalytic components are cemented. The
thickness of the catalyst is approximately 0.8 millimeters and
the principle catalytic components are vanadium and titanium
oxide.
NOXNON 600 has been proven in applications on oil-fired and
coal-fired combustion flue gas in Japan.
Before NOXNON 600 was installed in the reactor, the reactor
had to be rebuilt because the length of the NOXNON 600 catalyst
bed was longer than the one of NOXNON 500. The soot-blower was
»
changed to use hot air instead of superheated;steam.
Starting from April 22, 1980, the demonstration operation
with NOXNON 600 continued for more than nine months until Feb-
ruary 2, 1981. The operation with combustion flue gas from the
6-1
-------
Unit #3 Boiler of Georgia Power Company, Plant Mitchell exceeded
5,600 hours.
The pilot plant program required achieving NC>x removal
efficiency of more than 90 percent continuously for a period of
more than three months. Afterwards, the project scope was extend-
ed and transient tests v/ere included in the scope of the Contract
along with an extension of the operating period.
Tests with NOXNON 600 were, in general, as follows:
Catalyst life tests were run to confirm the expected cat-
alyst life. From April 22, 1980, the pilot plant was operated
maintaining NOx removal efficiency of more than 90 percent until
the end of October. After October, a nominal 80 percent NO^
removal was accepted in order to decrease ammonia slip as far as
possible. However, from time to time, operating conditions such
as flue gas flow rate, temperature, and mole ratio of ammonia to
NO were varied to evaluate catalyst performance,
x
Following the catalyst life test, further testing of the
Hitachi Zosen catalyst was needed to establish its suitability
for commercial applications on coal-fired boilers. The purpose
of this testing was, first, to determine the effects of trans-
ient conditions on the catalyst; second, to provide an extended
operating time so that at least 5,000 hours of operation could
be obtained to evaluate the long term effectiveness of the
NOXNON 600. During this testing period, catalytic performance
was evaluated by varying the flow rate, temperature and mole ratio.
Transient tests consisted of power plant load excursions, emer-
gency shut-off of ammonia feed, cold start-up, boiler shut down
and start-up, sudden load changes, operation without soot blower,
soot blower operation with ambient air and regeneration of cata-
lyst.
These tests established the applicability of Hitachi Zosen
catalyst NOXNON 600 to coal-fired boilers in the United States.
6-2
-------
CATALYST LIFB TEST
Operating Period
Operating hours with NOXNON 600 were,
operation with flue gas :
operation with air :
shut down r
5,620 hours
920 hours
320 hours
6,860 hours
TOTAL
Operation with air was done with air heated by the flue gas
heater to maintain the temperature in the reactor so that the
time for restartup could be reduced and to avoid thermal expan-
sion and contraction of the catalyst. Ammonia was not injected.
Interruptions in flue gas testing were caused approximately
80 percent by maintenance of Unit S3 Boiler of Georgia Power Com-
pany, and 20 percent by troubles with the soot blower and anal-
yzers inside the pilot plant.
Approximately 25 percent of complete shut downs were caused
by problems with the Blower, the Flue Gas Heater and the rotary
valve of the Dust Separator; 40 percent was due to catalyst
inspection and holidays; and 20 percent was for troubles at the
Unit #3 Boiler.
Timetable
Major milestones of the pilot plant are described below.
4/14/80 — 4/16/80 Reconstruction of reactor and ducting.
4/17/80 Catalyst charge.
4/22/80 Flue gas introduction into the system.
Due to trouble of Unit S3 Boiler, flue
ga3 was replaced with air on the same
day.
6-3
-------
4/24/80 — 4/26/80 Operation with flue gas. Due to boiler
trouble, again replaced with air.
5/8/80 — 5/14/80 Operation with flue gas at 1,200 scfm.
5/14/80 Catalyst inspection. Packing added to
clearance between reactor wall and
catalyst.
5/16/80 — 5/26/80 Phase IV Test with 1,500 scfm.
5/24/80 ~ 5/25/80 Flow rate test.
5/27/80 — 6/27/80 Operation at 1,500 scfm.
(6/5/80 -- 6/8/80) Stack test by Radian Corporation.
6/27/80 Heater bundle of Flue Gas Heater
replaced.
The second catalyst inspection.
7/3/80 -- 7/4/80 Flow rate test.
7/7/80 — 7/9/80 mole ratio test.
7/8/80 Flue gas flow rate changed from
1,500 scfm to 1,300 scfm.
7/20/80 — 8/8/80 Stack test by Radian Corporation.
8/4/80 — 8/5/80 SO2 concentration test.
8/6/80 ~ 8/9/80 Mole ratio test.
8/24/80 NOx level test.
8/27/80 Maintenance of soo* blower.
9/12/80 — 9/13/80 Mole ratio test.
9/21/80 — 9/26/80 Shut down due to troubles with Blower
and rotary valve.
10/28/80—10/30/80 Flow rate test.
11/7/80 Temperature test.
11/8/80 Flow rate was changed from 1,300 to
1,100 scfm.
11/9/80—11/17/80 Mole ratio test.
11/18/80 Mole ratio test.
11/20/80—11/25/80 80 percent removal test.
11/24/80 Flow rate was changed to 1,300 scfm.
11/26/80—11/28/80 Thanksgiving day holidays.
6-4
ouiot ;;hff- r
-------
12/5/80 -
12/8/80 -
12/10/80
12/11/80 ¦
12/18/80
12/19/80 ¦
• 12/6/80
• 12/9/80
- 12/17/80
- 12/22/80
12/24/80 — 12/25/80
12/26/80
12/27/80
1/1/81 and 1/4/81
1/5/81 — 1/6/81
1/7/81
1/9/81
1/15/81
1/26/81
1/30/81 — 1/31/81
1/31/81 — 2/1/81
2/2/81
Mole ratio test.
80 percent removal test.
Emergency shut-off of ammonia feed.
Power plant load excursions.
Mole ratio test
Operation with air due to increase
in pressure drop.
Christmas holiday.
Catalyst inspection.
Cold start up.
Boiler shut down and start-up.
Sudden load change.
Flow rate was changed to 1,100 scfm.
Operation without soot blower.
Mole ratio test.
Regeneration of catalyst.
Mole ratio test.
Flow rate test.
Operations ended.
Operations
The catalytic performance of NOXNON 600 was investigated for
the factors of N0x removal efficiency, mole ratio of ammonia to
NOx, operating temperature, and pressure drop across the catalyst
layer as summarized in Figure 6-1.
Furthermore, the relationship between actual pressure drop
and calculated pressure drop across the catalyst layer is expres-
sed in Figure 6-2.
Operating data used in Figure 6-1 are, in most cases, the
average of four measurements between 12:00 to 13:00. When some
special experiments were examined during this period and the
average seemed unrepresentative, the data were deleted, or other-
wise, collected from another period when the operating conditions
6-5
-------
lOOp
„
to
«
K >
8-
JV
IOC
5S
g-tK
- to
|e lJ
II ,s
.
I
. t, **e*a^*4 wA»**sy
U 20
MU
-J—r
Ju*l
» *3
-ah
10 »
r 2
10 8®
01 g«
JO 30
JEP
O
4
cn
100
90
.J ¦"
K
.w
70
a
J53
8-
?0C
S!
w
(* -
u
<1 •
!Sb
Si*
ig
Z&
03
, WK"»
-j^—
0,^7
-Sv&o
-a> « -" '
•in
-so *
H ^ SV,« * •
• . ¦ H«
¦;t; . r^r
4 g ft pflQ a
—S*gc*rtflt^ ft wa.^— yog *f
a!i *- «¦¦-** i"*"'-•**
* 't^ .
-fe «J » ' lo ' JO ' JS 13 ^10 x) IS » ' 3 ' i6 M~
«cr ht,v pec /{} jsn .-ei rti
1 2
ij ge
o» i«
Figure 6-1. Operating Data with NOXNON 600 (5,620 hours).
-------
10
i 0
cmaust
INSPECTION
°0°o
*10.
L
two SCFM
CO>W>*n'y-^^pTrrfcQ->^_rrrfPc~^W^i
oo-
{MM£!1
^ Ta^CP;
APR /80
20
*¦
200 S.FM
hat/bo
20
-fr
JUN/80
20
FLO * RATE
TEST
JUt/60
JO
10
20
Z 0
10
1100 SCFM
ST"
/iV/rwvv/f^o^
^4*4 a
blower
trouble
OcocAi^-..^ Oaa
I ^WA'oo,
AUG'60
I I I
50"
20
30'
SEP/SO
j , L
20
Tfe"
OCT/BO
20
2 0
10
FM I:
aoa.E(t vttSHEO
BTWIIER
T7
'«ie:
O
NllOO SCFM
I
THl'.tS GlVt?oQcfc0coc
/flOO SCFM '
20
yo
soor BLOrfER
^HBISTMtS fVK M j.gTESr
GEC /SO IAN /8I
j i I i U 1 1 .—
CiTiLrST
RfSfHERaTlOS
FEB/81
10
20
30'
10
20
30'
FLOW RATE TEST
—3^—
10
V catalyst inspection
• LESS TKAN 1000 SCFM
o 1100 SCFM
* 1200 •
A 1300 •
Q 1500 ¦
} HOT AIR OPERATION
K shut down
Figure 6-2. Pressure Drop across Catalyst Layer (Actual/Ca]culated).
-------
s^were more stable.
A. Period from 0 hours to 200 hours (4/22/80 to 5/14/80).
The operation of NOXNON 600 with combustion flue gas began
on April 22, 1980. However, due to problems at the Unit #3
Boiler, the pilot plant was forced to watch and wait for flue
gas operation while maintaining the operating temperature in
the reactor with heated air. Therefore, actual start-up with
flue gas began on May 8, 1980.
During this period, standard operating conditions were,
During the first week of operation N0X removal efficiency
was 88 percent to 93 percent which was lower than the expected
efficiency. At the same time, pressure drop was almost the same
as calculated pressure drop (0.77 inches HjO). However, due to
the very low pressure drop, non-uniformity of flue gas distribu-
tion was suspected, and the catalyst was inspected from the man-
hole of the reactor on May 14, 1980. The results of the observa-
tion will be described later in "Catalyst Inspections On Site,"
and summarized as follows:
(1) The seal of the clearance between the reactor wall and
the catalyst appeared insufficient, and additional
asbestos yarn was supplemented.
(2) From the appearance of the adhered fly ash over the
reactor wall, some non-uniformity of flue gas distrib-
ution was assumed.
Subsequently, after this observation of the inside of the
reactor, the flue gas flow rate was increased to 1,500 scfm
although this flow rate corresponds to 0.7 MW which is 40 percent
larger than the designed capacity of 0.5 KW equivalent. This
Flow rate
Temperature
NH^/NOx mole ratio
1,200 scfm
700°F
1.0
6-8
-------
¦increased flue gas flow rate would provide better gas distribu-
tion.
B. Period from 200 hours to 1,320 hours (5/14/80 to 7/7/80).
Dntil 7/7/80, the pilot plant was operated at the flue gas
flow rate of 1,500 scfm. NOx removal efficiency was between 90
percent to 94 percent and seemed stable. During this period it
was found that the NOx removal efficiency appeared to increase
slightly when the flow rate was increased. This tendency was
noticed especially in the performance test run in the middle of
May.
This tendency was possibly caused by:
- Increase in the flue gas flow rate decreasing the
deposits of fly ash on the catalyst surface.
- Higher superficial linear velocity contributing to the
uniformity of flue gas flow distribution due to higher
pressure drop.
Meanwhile, the pressure drop started at 1.0 to 1.1 inches
H2O at the beginning of the operation, and increased very grad-
ually up to 1.3 inches H^O. From this experience, the pressure
drop seemed to be more stable at the flow rate of 1,300 scfm, and
cacalyst inspection on June 27, 1980 showed that abrasion of cat-
alyst seemed to be in progress although very slow. Flue gas flow
rate was changed to 1,300 scfm, equivalent to 0.6 MW.
C. Period from 1,320 hours to 4,000 hours (7/8/80 to 11/9/80).
During this period, catalytic performance seemed stable.
At the end of August when the operating time was 2,500 hours,
performances were:
flow rate : 1,300 scfm
operating temperature : 700° to 720°F
NH^/NOx mole ratio : 1.0
6-9
1 \ ~ \ \ \ ij l m 1 r
-------
N0X removal efficiency : 90 to 94 percent
Pressure drop : l.io to 1.30 inches H20
At the end of September, for the purpose of confirming cat-
alytic performance at more than 90 percent NOx removal at the
design condition, a performance test was run at the flow rate of
1,100 scfm. More than 90 percent NO^ removal was proven through
the performance test. However, from the end of August, from time
to time, pressure drop across the catalyst layer fluctuated as
did the NO^ removal efficiency. This tendency was especially
observed when the Unit #3 Boiler operated the soot blower at its
economizer repeatedly and frequently. When the pressure drop
increased and the NOx removal decreased due to the frequent soot
blower operation at the economizer, the unstable pressure drop
and NOx removal efficiency were restored after a few days by
operating at a lower gas flow rate: 1,100 scfm.
Between October 13, 1980 and November 8, 1980 (3,340 hours
to 3,960 hours), mole ratio of ammonia to NOx was decreased to
approximate?y 0.9 to obtain less ammonia slip.
On November 8 and 9, 1980, catalytic performance of more
than 90 percent was again confirmed at the designed capacity,
1,100 dcfm. At this time, operating time had been 4,000 hours.
NOx removal efficiency at a flow rate of 1,300 scfm and mole
ratio of 1.0 was approximately 89 percent.
D. Period from 4,000 hours to 4,230 hours (11/10/80 to 11/21/80).
At the design condition, a mole ratio test and a temperature
test were run to check the catalyst activity prior to entering
into transient tests. Afterwards, an 80 percent test was run.
Flue gas flow rate
Operating temperature
NH^/N0X mole ratio
NOx removal efficiency
1,300 scfm
700° to 720°F
0.88 to 0.94
8 3 to 88 percent
6-10
-------
*jThe purpose of the 80 percent test was to reduce ammonia slip as
far as possible. The target of N0x removal was decreased to 80
percent from 90 percent, and NII^/NO mole ratio was set so as to
maintain the NC>x removal at slightly higher than 80 percent,
operating conditions were:
flue gas flow rate
temperature
The
NH^/NOx mole ratio
1,100 scfm
700°F
approximately 0.85
,E. Period from 4,230 hours to 4,420 hours (11/24/80 to 12/4/80)
During this period the operation was continued at the fol-
lowing conditions:
flue gas flow rate
temperature
NH3/NOx mole ratio
NOx removal
1,300 scfm
700°F
approximately 0.8
slightly greater than 80 percent
The purpose of this test was to investigate operations with
less ammonia slip and 80 percent NO^ removal at the flow rate of
1,300 scfm.
After Thanksgiving Day, temperature and pressure drop fluc-
tuated widely resulting in changes in N0x removal. These fluctu-
ations were not caused by any problems in Unit #3 Boiler.
F. Period from 4,420 hours to 5,030 hours (12/5/80 to 1/7/81).
To evaluate the catalyst performance while transient tests
were being executed, a performance test was run on 12/5/80 and
12/6/80. The result was:
flue gas flow rate
temperature
NH^/NOx mole ratio
NOx removal
1,100 scfm
700°F
0.99 to 1.02
90 to 91.5 percent
On 12/7/80, transient tests commenced. The results of the
6-11
-------
>*"
^transient tests are described later in "Transient Tests." While
the transient tests were being run, operating conditions were
isharply varied, and N0X removal efficiency was difficult to
estimate, however, at the conditions of:
flue gas flow rate : 1,100 scfm
temperature : approximately 700°F
Nh3/N°x mole ratio : approximately 1.0
NOx removal efficiencies were:
12/5/80 and 12/6/80 : 90 to 91.5 percent
12/18/80 : 89 to 91 percent
1/5/81 : 91 percent
From 12/11/80, while "Power Plant Load Excursions" test was
run, the pressure drop gradually increased, and on 12/18/80,
when Georgia Power Company had trouble with leakage in water
tubes of Unit #3 Boiler, the pressure drop increased rapidly and
seriously. In order to reduce the pressure drop, the pilot plant
was operated at 1,100 scfm and 1,300 scfm, and air operation was
also tried. However, higher and unstable pressure drop continued
until 1/7/81.
G. Period from 5,0 30 hours to 5,470 hours (1/7/81 to 1/26/81).
Transient tests continued during this period, and the pilot
plant was shut down on 1/26/81 for regeneration of catalyst.
H. Period from 5,470 hours to 5,620 hours (1/27/81 to 2/2/81).
After the regeneration of catalyst, the catalyst performance
was examined. NOx removal efficiency after the regeneration of
catalyst was the same as the initial efficiency, and the pressure
drop decreased and seemed stable.
6-12
-------
CATALYST PERFORMANCE TESTS
While catalyst life test and transient tests were being run,
catalyst performance was evaluated by varying operating condi-
tions .
The purpose of Oie catalyst performance tests were (1) to
evaluate the catalyst performance from time to time over the
elapsed operating time, and (2) to investigate influences of
operating variables on the catalyst performance. The perform-
ance tests were examined approximately once every 1,000 hours
of operation.
Flue Gas Flow Rate
In the middle of May, the first catalytic performance test
was performed. As shown in Figure 6-3, when the flue gas flow
rate increased, the NO^ removal efficiency increased slightly.
This rather contrary tendency seems peculiar to high fly-ash-
loaded combustion flue gas from a coal-fired boiler. When the
flow rate is somewhat low, fly ash clinging to the catalyst sur-
face results in partial masking of the catalyst surface and, to
some extent, to non-uniformity of flue gas flow through the
catalyst channels. When the flow rate is increased, deposited
fly ash is blown off the catalyst increasing the apparent NO
removal efficiency. (At the same time, the fact that should be
noted is pressure drop across the catalyst layer of NOXNON 600
was lower than that of NOXNON 500 due to wider channels between
catalyst layers). From these results, the flue gas flow rate in
the catalyst life test and the transient tests were increased to
1,300 scfm and 1,500 scfm although designed flow rate was 1,057
scf m.
After approximately 1,000 hours of operating time elapsed,
the tendency to increased NOx removal with an increase in the
flue gas flow rate disappeared, and NOx removal decreased when
6-13
-------
100
I
90
80
70
°o
C~
°&r
REACTION TEMP APPROX 700 *F
NHj/NO* RATIO APPfiOX 10
SYMBOL OPERATING PERlOO
/o 200 - I250H
\ o 2150 ~ 3 7 SOM
1000 1100 1200 1300 1^00 1500 1600 1700
FLUE GAS FLOW RATE I SCFM I
Figure 6-3. Effect of Flue Gas Flow Rate on NOx Removal
Efficiency.
-------
rsfthe flow rate of flue gas increased. This is as normally seen
jin pilot plants applied to oil-fired combustion flue gas. It
'was assumed that with the elapsed operating time fly ash deposit-
ed on the catalyst surface, and this thin fly ash layer stabil-
ized in equilibrium with the flue gas flow, especially at the
corners of the triangular channels, and resulted in a decrease
of N0X removal efficiency.
Mole Ratio
During the test operation the pilot plant was operated at a
selected mole ratio. The control system was designed to auto-
matically provide this mole ratio. This was accomplished by
using the flue gas flow rate signal and the inlet N0X concentra-
tion analysis to determine the quantity of N0X in the inlet
stream. From the inlet N0X quantity and the selected mole ratio
the required ammonia was automatically calculated. This signal
was then relayed to the ammonia control system which set the
ammonia control valve setting to provide the required ammonia
flow.
The actual mole ratio tended to vary somewhat because of
fluctuations in N0X levels, flue gas flow rates, flue gas
pressures, etc. Efforts were constantly made to be certain of
the accuracy of these measurements (see Appendix) but it is
believed that there nay have been a consistent error.
The fraction N0X removal is often higher than the mole ratio
especially at lower levels. Fundamental experiments, however,
have proven that this is not possible. In the presence of oxygen
the NOx removal cannot be greater than the equivalent amount of
ammonia available. The reaction between NO and NH^ is an equi-
molar reaction.
Although all of the instruments were checked and rechecked
and calibrated several times, there is a suspicion that the
ammonia flow meter may have been in error, particularly at low
flows. It may have been reading too low. The ammonia flow rate
6-15
-------
was very low (about 0.5 acfm) and the very narrow clearance
between the rotor and the tapered tube could readily catch small
particles such as dust, rust, and even moisture causing inaccu-
rate readings. This may have happened despite the installation
of very fine micron filters before the flowmeter.
Because of the questions regarding the ammonia flowmeter
the data in tne later tests were recalculated to obtain mole
ratio curves (Figure 6-4, 6-5 and 6-6). The ammonia feed was
calculated by adding together the ammonia slippage and the
ammonia theoretically required to react with the NOx removed.
The mole ratio was then determined by dividing the calculated
moles of ammonia by the moles of inlet NOx« (The NOx was assumed
to be all NO).
Figure 6-4 shows the results of the test at an 80% removal.
Figures 6-5 and 6-6 are mole ratio curves taken over different
operating periods. The mole ratio was calculated by the methods
described above. From these results it can be seen that flue
gas flow rates had no measurable effect on NOx removal. A NO^
removal of 80% required a mole ratio of about 0.85 while a 90%
removal requires a role ratio of 1.0.
Ammonia Slippage
Since a continuous ammonia analyzer was not available during
the operation of the pilot plant, ammonia had to be measured by
wet analysis. The method was refined to the point where by
August 15, 1980 the data was considered reliable. The accuracy
of the method was checked by Radian Corp. and found to be sat-
isfactory. Data relative to ammonia slippage was collected after
that time.
Figures 6-4, 6-5 and 6-6 show ammonia slip under various
conditions. Figure b-4 shows ammonia slip measured during the
80 percent removal test. Figure3 6-5 and 6-6 show ammonia slip
when mole ratio was varied at operating times of 3200 hours to
4500 hours and 4600 hours to 5600 hours respectively. General
j
6-16
-------
Figure 6-4 80 porcent Removal Test
conoiTion ;
REACTION TEMP . 704 ~ 718 t
INLET NOi * 36 7 -4 J) FPU
NO REMOVAL EfF : 80 0 - 80 » y.
OPERATING PERIOD . 4200 - 4 1
SYMBOL f * 80 * N°" REM0WL
I o nh} slippage
-------
100
>-
u
z
120
UJ
a:
*
o
06
HfJj/HOx MOLE RATIO (-(
(O 1100 SCFM 700*10 *F
symbol <& uoo scfm 7ooho*f
to 1500 SCFM 710 t 5 *F
Figure 6-5. Effect of NH^/NC^ Ratio on NOx Removal
Efficiency and NH3 Slippage. (3200-4500 hrs)
-------
ol 1 " 1 1 1 1 1 0
o 02 04 06 Oft 10 II
MNj/N0» HOLE R 4H0 I - 1
REACTION TEMP 700 - 10 'f
fo I 100 SCFM
SYMBOL I® 11 00 SCFM (AFTER REGENERATION I
la 1300 SCFM
Figure 6-6. Effect of Ratio on NOx Removal
Efficiency and NH^ Slippage. (4600-5600 hrs)
-------
conclusions from these tests:
- Flue gas flow rate had little effect on ammonia slippage.
Ammonia slippage seemed to increase slightly with
increased operating time.
- Ammonia slippage was apparent even at low mole ratios,
for example at a mole ratio of 0.6.
The reason for this higher than expected slippage may have
been due to clogging by fibers of asbestos yarn used to seal a
clearance between the catalyst box and reactor (see "Catalyst
Inspections"). These fibers along with fly ash between the first
and second catalyst layers may have reduced the effective cat-
alyst surface and adversely affected the apparent catalyst acti-
vity resulting in relatively high ammonia slip.
Operating Temperature
Th3 influence of operating temperature on NOx removal effi-
ciency is shown in Figure 6-7, "Reaction Temperature vs. N0X
Reinoval Efficiency." In the pilot plant, the operating tempera-
ture was varied between 640°F (338°C) and 780°F (415°C), and
NOx removal efficiency was not affected in this temperature
range. Optimum operating temperatures obtained from fundamental
experiments in the laboratory are between 572°F <300°CJ and 750°F
(400eC}.
The results obtained in the EPA pilot plant are fairly con-
sistent with the fundamental experimental data. At the beginning
of operations, NOx removal efficiency did not vary between 640°F
and 780°F, and after approximately 4,100 hours of operation, N0x
removal at a temperature higher than 730°F seemed slightly lower.
However, the difference was very small. It would seem, there-
fore, that between 640°F and 780°F, the effect of temperature on
removal efficiency is almost negligible.
6-20
-------
(FLOW RATE : 1500 SCFM
Operating perigo : 200- 500 h
/FLOW RATE : 1100 SCFM
\ OPERATING PERIOO : ABOUT 4100 H
MOLE RATIO : 10
NOx CONC : 450 ~ 500 PPM
_i_
600 650 700 750 800
REACTION TEMP. (°F1
Figure 6-7 Effect of Reaction Temperature on NOx Removal Efficiency.
-------
N0|t Concentration
On August 2 4, 1980, the NOx Generator was operated to in-
crease the concentration of NOx at the inlet of the reactor, and
the influence of NOx concentration was investigated. Operating
conditions were:
flue gas flow rate
operating temperature
NH^/NOx mole ratio
NOx concentration
1,320 scfm (average)
700 °F
1.0
773 - 906 ppm
NOx removal efficiency during this test was compared with
that obtained on August 22, 1980 after approximately 2,400 hours
of operation. Results are shown in Figure 6-8.
Between 400 ppm and 900 ppm of inlet N0X concentrations, the
concentration of NO does not influence NO removal efficiency.
X X
This data is consistent with fundamental data.
Influence of SO.. Concentration on NO„ Removal Efficiency
K K
It has been well established through fundamental experiments
and other pilot plant's data in Japan that catalyst activities are
not affected by SOx concentration. In operation at the EPA pilot
plant, even though the load of Unit #3 Boiler had been relatively
stable at approximately 135 MW, concentrations of SO2 measured in
the pilot plant deviated widely between approximately 500 ppm and
1,500 ppm depending on the variations of sulfur in the coal.
Therefore, various data under the same operating conditions ex-
cept for S02 concentrations were selected and summarized in Fig-
ure 6-9. This showed that through the whole operating period,
NOx removal efficiency was not affected by SOj concentrations in
the flue gas between the relatively wide range of 500 ppm and"
1,500 ppm.
b-22
-------
FLOW RATE : 1,300 SCFM
REACIiw'.- TEMP • 700 °F
NH3/NOx RATIO : 1.0
Operating period : about 2500 h
200 400 600 800
INLET NO* CONCENTRATION (PPM)
1000
Figure 6-8 Effect of NOx Concentration on
NOx Removal Efficiency.
6-23
-------
100
>- 90
u
ao
<
>
o
O 70
-0——O-
o 6/8,9/80
1500SCFM, 710-720°F , NHj/NOx- 10
A B/20 ,21,22,24,26 / 80
1300 SCFM, 700~710°F, NH3/N0*-I0
o 9/10, 11/60
1300 5CFM, 650-670°F,NH3/N0x^ 10
0 1/7,15/81
1100 SCFM, 710 °F , NHj/NOx- 0 8
500 1000 1500
SO* CONCENTRATION (PPM)
Figure 5-9 Effect of SOx Concentration on
NO^ Removal Efficiency.
6-24
-------
Oxidation of SO^ to SO3
On July 23, 24 and 25, 1900 operating conditions were main-
tained at a constant level so that SO^ measurements could be
obtained at the inlet and outlet of the reactor. The following
were the conditions of operation. The SO2 and SO^ measurements
were made by Radian.
Date
7/23/80
7/24/80
7/25/80
Time
1630
1650
1625
Flue Gas Flow Rate (scfm)
1320
1320
1320
Temperature °F (°C)
715 ( 380)
715 ( 380)
716 ( 380)
Inlet SO2 - ppm
820
579
679
Inlet NO - ppm
473
514
468
Oxygen - %
5.7
5.8
5.5
NH3/NOx 11,016 ratio
0.915
0.931
0.91
N0X removed
91.2
92.4
91.9
Inlet SO^ - ppm
11.0
7.5
6.6
Outlet SO^ - ppm
25.3
14.8
21.9
S02 to SO3 conversion - %
1.7
1.3
2.3
These results indicate an average oxidation rate of 1.8%.
It was expected that at these conditions the ratio would be
lower: about 1.0 to 1.5%. Those expectations were based on
experimental laboratory data and on pilot plant data from Japan.
Experimental data results are shown in Figure 6-10. At the
catalytic composition as used for the EPA pilot plant, the per-
cent oxidation would be expected to be 0.5%. However, this ex-
perimental data was performed at a somewhat lower temperature
(350°C vs. 3808C) and a higher Area Velocity (12.1 vs. 9.6).
Data from a coal-fired pilot plant in Japan indicated an
oxidation rate of between 0.89 and 1.11% at the same temperature
and area velocity as used for the EPA pilot plant tests.
Of course, measurements of SO^ are quite difficult and
require a high degree of technical skill and experience. Even
6-25
-------
TEMPERATURE : G60°F(350°C)
EPA PILOT PLANT
100
•<
>
o
2
UJ
o:
M
o
z
I
I
a
o*
Z
o
<
Q
•X
o
o
to
0.2
CATALYTIC COMPOSITION (% ACTIVE MATERIAL)
Figure 6-10 Typical S02 to S03 Conversion.
-------
with that, the presence of fly ash always influences the analyt-
ical results. Although there were no questions about the tech-
niques followed in obtaining this data, the expected scattering
of data would require many points to provide a reasonable stat-
istical confidence in the results. Unfortunately, only three
points were obtained due to the difficulties and the limitations
of funds. Nevertheless, the data was relatively consistent and
the oxidation rate was close to what would have been expected.
The results of these tests showed an oxidation rate which
was relatively low. However, these rates can be reduced even
further by varying the catalyst composition. The need for very
low oxidation would depend on the particular application of the
process.
TRANSIENT TESTS
After the N0x removal efficiency of more than 90 percent was
demonstrated in the continuous run of three months as required in
the contract, a decision was made to extend the scope of the con-
tract. This was done to supplement the originally planned opera-
tion of the pilot plant to further establish the suitability and
reliability of Hitachi Zosen's catalyst for commercial operations
on coal-fired combustion flue gas. For one thing, the continuous
long term run would be extended to reach at least 5,000 hours of
operation, if possible. Also, a series of tests were planned to
determine the effects on the catalyst and the system of transient
conditions which could be anticipated in commercial operations.
A3 Georgia Power Company planned modifications of its Unit
#3 Boiler starting from February, the testing could not be
extended beyond that period since the pilot plant would have to
be removed to enable the modifications to be made.
The following transient tests were performed:
6-27
-------
Power Plant Load Excursiors
A commercial power plant varies boiler load aS required by
power consumption. The change in boiler load is relatively rapid
which can cause sharp variations in flue gas flow rate and temp-
erature. The purpose of this test is to confirm the control-
lability of the system and the reliability of the catalyst for
the transient conditions of load excursions. Two ranges of con-
ditions were tested. One was at "full load" which is at high
flue gas flow rate and high temperature, and the other was at
"low load" with low flow rate and low temperature.
Operating conditions during daytime (9:00 am to 5:00 pm):
flue gas flow rate: 800 scfm (approx. 1,300 Nm3/h)
temperature : 610°F (320°C)
NOx removal : 80 percent
Operating conditions during nighttime (5:00 pm to 9:00 am):
flue gas flow rate: 1,300 scfm (approx. 2,000 Nm3/h)
temperature : 700°F (370°C)
NOx removal : 80 percent
The above conditions were repeated 6 times from December 11,
1980 to December 18, 1980. Transition from one condition to
another was done rapidly.
Before and after the tests, catalytic performances were con-
firmed at the following conditions:
flue gas flow rate: 1,100 scfm (1,700 Nm3/h)
temperature : 7006F (370°C)
NH3/NOx mole ratio: 0.8 and 1.0
Figure 6-11 shows the results of this test. During these
repeated tests, NO^ removal efficiency at a given mole ratio
showed no change. Pressure drop at the low load did not change,
however, pressure drop at the high load increased gradually from
1.03 inches to 1.16 inches f^O.
Usually pressure drop increased due to soot blowing applied
6-28
-------
SYMBOL
TEST CONDITION
—0
1100 SCFM, 700V 60 V. REMOVED
—o—
800 SCFM, 6l0"f BD'A REMOVED
Figure 6-11. Test of System Performance under
Plant Load Excursions'.
6-29
-------
to the economizer in Unit #3 Boiler, therefore, this increased
pressure drop did not seem to be related to the variation in
operating conditions. Catalytic performances before and after
this test were the same: 83.6 percent at mole ratio 0.8 and 90.3
percent at mole ratio 1.0. These results proved tliat the system
and catalyst can withstand sudden load changes in the boiler.
Emergency Shut-off of Ammonia Feed
The EPA pilot plant was provided with a trip system for
ammonia supply. The purpose of the trip system was to shut off
the ammonia feed when temperature in the reactor decreased to
600°F (315°C)/ and allow it to introduce ammonia into the system
when temperature in the reactor increased and returns to 600°F.
The purpose of this test was to confirm the reliability of the
trip system.
On December 10, 1980, the following tests were performed.
While the pilot plant was being operated at the conditions of
1,300 scfm, 700°F, and 0.8 mole ratio, the temperature was allow-
ed to decrease from 700°F to 550°F. During this change, the
automatic shut-off of ammonia feed at a temperature of 600°F was
confirmed.
Afterwards, the temperature in the reactor was allowed to
increase from 550°F to 700 °F and it was confirmed that ammonia
feed was started and introduced automatically into the system
when the temperature of 600°F was reached. This experimental
work was repeated three tinws in order to prove the repeatability
of this system. The experimental data is shown in Figure 6-12.
The results of this proved that:
- Automatic shut-off and supply of ammonia operated
smoothly.
NOx removal efficiency and pressure drop at a temperature
of 700°F was constant and unchanged through the three
repealed tests.
6-30
-------
too
90
0.9
08
800
700
600
500
17-05
15.05
13 05
11:05
9.05
NHj OFF
t O o
NHj ON
NHj OFF
NHj ON
NHj OFF
OO
oo
OO
Oo
Oo
oo
oo
TIME
Figure 6-12. Test of System Performance with
Emergency Shut-off of Ammonia Feed.
6-31
-------
Cold Startup
The EPA pilot plant was provided with an air intake valve
and air circulation piping line for startup and shut down in
order to prevent the formation and deposit of sulfuric acid mist,
ammonium sulfate, and ammonium bisulfate. During the pilot plant
operation, the catalyst was routinely heated up with hot air
prior to the introduction of flue gas. However, a commercial
boiler normally starts into operation after a long shut down with
the reactor and ductwork filled with ambient air. When operation
commences flue gas would be introduced into the system as the
temperature rises and the flue gas could be relatively cool for
a time when in contact with the catalyst. There is a risk of
formation and deposition of sulfuric acid mist, ammonium sulfate
and ammonium bisulfate. Previous experiences proved that sulfur-
ic acid mist does not deteriorate the catalyst, and ammonium sul-
fate and ammonium bisulfate can be removed when the temperature
rises. The purpose of this test was to prove that cold start-up
with flue gas does not cause any trouble to the N0X removal
reaction.
Before the pilot plant°was shut down for Christmas holidays
on December 24, 1980, operating conditions were 1,100 scfm, 700°F,
1.1 to 1.2 inches HjO, and 80 percent N0y removal with 0.8 mole
ratio. When the plant was shut down, the whole system was purged
with air. On December 27, 1980, the blower was started and flue
gas was introduced into the system directly. The system was
heated up with flue gas and put into operation.
Operating esfr^fctfcefwa ¦be'fGHee -and 'after the cold start-up are
shown in Figure 6-13. After the start-up, relatively higher pres-
sure drop and lower NO removal efficiency in comparison with
*
previous operating conditions continued for'^approximately a half
day, and these were restored as time elapsed. These phenomena
are normally observed when the pilot plant is shut down and
started up. After a half-day after commencing the start-up,
operating conditions became stable at the same conditions before
6-32
-------
100
_J
90
(X.
60
o
X
70
pilot plant down
—cooc£_
DEC 24
e ooam
DEC 27
8 OOa M
DEC 28
OFC 29
8 OOAM
g 09
5 ~ 0 8
txP"
o c
c^po^^So^ccrp
cP°o00o
I200r
g a tupO
w uoo —Sttr_Q_
* o
O i/>
10001-J-
iOLQ. STifll- UP.
UJ Q ' 4
3-
K S 12
—cfp,
Oco
a 10
800
£ 700
I 600
*-
soo
-CQcOOQl
0EC 2*
e oo am
-epeorfiiyro-
o o
°o
DEC 27
8 00AM
DEC 28
out
DEC 29
8.00AM
Figure 6-13. Test of System Performance with
Cold Start-up.
6-33
-------
cold start-up. The results proved that cold start-up does not
influence catalytic performances.
Boiler Shut Down and Start-up
From time to time a power plant boiler shuts down and starts
up and an N0X removal system must follow such transition periods.
The purpose of this test was to confirm the controllability of
the N0X removal system during the shut-down and start-up of the
boiler, and the ability of the catalyst to withstand these trans-
ient operating conditions.
When the host boiler was shut-down for maintenance from Jan-
uary 1 to January 4, 1981, this test was executed. As shown in
Figure 6-14, when the fan of the host boiler was switched off,
the blower of the N0X removal system was switched off at the same
time on January 1, 1981. When the fan was switched on to purge
the furnace and ductwork, the blower was again switched on, and
then switched off following the operation of the fan. Before the
start-up on January 4, 1981, the whole system was left as it is.
On 1/4/81 the fan of the host boiler was started up and the
blower of the N0X removal system was also started at the same
time. The Flue Gas Heater in the pilot plant was not switched
on, and cool flue gas was introduced into the system as tempera-
ture of flue gas in the boiler furnace rose. Thus, the pilot
plant was filled with the same flue gas as the boiler during the
transient period of boiler shut-down and start-up. During this
test, ammonia was automatically cut off and fed at 600°F.
After the operating conditions became stable, catalytic per-
formance was examined through a mole ratio test.
The result was:
flue gas flow rate
temperature
N0x removal efficiency
1,100 scfm
700 °F
at 0.8 of mole ratio ° 83%
at 1.0 gf .mPle ratio =¦ 91%
6-34
-------
ck suavro shui •
do«« wooovai
or ') »on(«
BOttft
|CK D*ii«<0 IhC •HI*
loul or K DOO.C*
I
Inunc im( toac*
IvilH «*((•
[oil rum or axra
|rm orr| | ow |
1, J |— —J
i i '
5"
1 ftn ow |
X
T
| iu«»«( oi | I wu« r®(0 win cou
^=r~
COala
O'CMttOn
* - 1100
PllOf I s
Pv**f 1000
.0QOo©®OO©OG^
HW' V\
OO *
p oOOO
©OOOq®oOOO©OOo
Hfrf CTf
t
,
an i
1 00
JM 4
1 00
ua J
1 00
tyt
I
oj
en
100
Z 500
•. «0Q
m
~ too
0
I >
a -io
g* *
iSr 05
o O O o o
-OOOOOOOO
»9«
,0O®C
°o
1 - 1 1
ooooo
o°ooooo0oo
° o
OOOO
oo o 0
oo
e
o
e
o
0
~0
-
OOOO
10
. ouritr mnrn* ikofiteif
o
O !
0
rut
1 0
OOOO
01
Ooo°®
O&o
ot
.
jt* i
Jtn j
• 00
1 00
1 00
Figure 6-14. Test of System Performance with Boiler Shutdown
and Start-up.
-------
Pressure drop across the catalyst layer was 1.15 inches 1^0,
and there was no change through this test.
From these results, the system proved that it could withstand
the transient period of ohut-down and start-up of the boiler.
Sudden Load Change
The boiler for a power plant may occasionally change load
suddenly complying with variations of power consumption. The N0X
removal system should follow these sudden boiler load changes.
The following two levels of operating conditions were
adopted as representative load levels:
high load: 1,300 scfm, 70G"F, 0.8 mole ratio
low load : 900 scfm, 610°F, 0.8 mole ratio
The above conditions ware altered once every two hours and
continued for 24 hours from 11:00 AM on January 6, 1981. After
this test was completed, catalytic performance was examined by a
mole ratio test. The development of this test is shown in Fig-
ure 6-15.
The results showed a gradual increase in pressure drop and a
slight decrease in NO^ removal efficiency. Catalytic performance
checked by a mole ratio test also showed a slight decrease in NOx
removal efficiency:
81 percent at 0.8 mole ratio
87 percent at 1.0 mole ratio
Possibly, fly ash accumulated in ducting at the upstream of
the reactor during low load operations and was blown off by the
increased flue gas flow when the high load was reinstated. This
fly ash settled on the catalyst resulting in slighly lower NOx
removal and slightly higher pressure drop. Experience in this
pilot plant has shown that normal operating results are usually
restored after a few days when this occurs. Therefore, there
seems to have been no serious problem caused by these sudden load
changes.
6-36
-------
I 2 i 4 5 6 7 6 9 10 II 12 13 14
C Y C L E S
'
x —
5 £
o _ 1000
800
I 2 3 4 5 6 7 6 9 10 11 12 13 14
CYCLES
SYMBOL
TEST CONDITIONS
—O—
1300 SCFM, 700 *F
«
900 SCFM, 610 *F
a
H00 SCFM, 700 *F
Figure 6-15. Test of System'-Performance
with Sudden Load Change.
6-37
-------
Sootblower Requirements
From the beginning of pilot plant operation with the third
charged catalyst, NOXNON 600, the sootblower was operated three
times a day, one cycle at a time, in order to prevent clogging by
fly ash. This frequency seemed to be the minimum to prevent clog-
ging, however, there was no experience in operating the pilot
plant without the sootblower. Thus, a trial was mads to operate
without the sootblower on January 11, 1981. At a flow rate of
1,100 scfm, 700°F, 0.8 mole ratio, operation without sootblower
was tried and pressure drop increase was observed. (See Figure
6-16). The pressure drop appeared unstable and increased grad-
ually from 1.20 to 1.25 inches H^O to 1.35 to 1.40 inches f^O in
approximately 28 hours, and seemed to continue to increase. The
sootblower had to be operated at this time. From this observa-
tion it was concluded that operation of the sootblower was
necessary.
From the beginning of the pilot plant operations with the
third charged catalyst, the sootblower had been operated with air
heated by the sootblower heater in order to prevent sudden shrink-
ing and damage of catalyst due to cold air. The temperature of
the air was 650°F. On January 12, 1981, after the 28 hours pilot
plant operation without the sootblower, unstable and increasing
pressure drop was being experienced, and it was decided to oper-
ate the sootblower. However, due to problems in the sootblower
heater, heated air was not available. Thus, the sootblower was
operated with ambient air as a necessity. One travel of the
sootblower requires two minutes, and the temperature at the out-
let of reactor decreased by only 5° to 10°F. There was no change
in NOx removal efficiency and pressure drop after this sootblower
operation. Therefore, from this time to the end of the pilot
plant operation, ambient air was used for the sootblower and it
is possible that it can also be used in commercial systems.
6-38
-------
6
5
4
3
2
0
TEST CONOITION
IIOO SCFM, 700 *F
9
6
0
2
6
8
10
14
16 18 20 22 24 26 28
HOURS OF OPERATION WITHOUT SOOT BLOWER
Figure 6-16. Test of System Performance without Soot Blower.
-------
REGENERATION OF CATALYST
Experience with NOXNON 500 catalyst indicated that washing
with warm water could remove ammonium sulfate and ammonium bisul-
fate deposited on the surface along with fly ash which had adhered
to the catalyst surface. This resulted in some degree of regen-
eration of deteriorated catalyst. Since the catalytic composition
of NOXNON 600 is the same as that of NOXNON 500 the same effect
could be expected in applying this procedure to NOXNON 600. Fur-
thermore, preliminary tests were run in Hitachi Zosen's laboratory
to confirm that the regeneration of NOXNON 600 by washing with
warm water is applicable. Therefore, this test was run on site.
Preliminary Tests
Preliminary tests were tried at Hitachi Zosen's laboratory
prior to the test on site. These tests evaluated the following:
The influence of water washing on catalyst activity
- The chemical analysis of waste wash water.
- An abrasion test of the catalyst after water washing.
The following samples of catalyst were tested:
- Virgin catalyst which was produced for the EPA pilot
plant and was stored at Hitachi Zosen's laboratory.
- Used catalyst from the EPDC pilot plant.
- Virgin catalyst from EPDC pilot plant (for chemical
analysis of waste wash water only).
Due to the presence of sulfate and bisulfate, the pH of the
water would be expected to become somewhat acidic (pH 2 to 3)
when the used catalyst was washed by water. Therefore, instead
of using ordinary water, a dilute (.001 N) sulfuric acid solu-
tion having a pH of about 3 was used.
- Amount of wash solution: 10 liters/sq. meter of catalyst.
- Washing times: 0, 5, 10, 30, 60, 120 minutes.
6-40
-------
- Washing procedure: Samples of catalyst were soaked in
the solution.
- Drying: at 110°C (230°F) overnight
- Analysis of solution: filtered and analyzed by atomic
absorption spectrophotometry.
- Abrasion tests: Catalyst samples were tested by placing
on vibrating screen with an alumina ball.
The results of these tests showed:
Water soaking of the virgin catalyst for less than 60
minutes caused only a slight decrease in removal effi-
ciency. However, soaking for 120 minutes caused a sig-
nificant decrease in NOx removal. (See Figure 6-17).
- Water soaking of used catalyst also showed the tendency
to a decrease in NOx removal efficiency depending on the
soaking time. (See Figure 6-18).
- Chemical analysis of waste wash solution (see Table 6-1)
indicates a definite increase in leached vanadium with
increased soaking time.
- The abrasion tests (see Figure 6-19) indicated that the
strength of the catalyst did not change after water soak-
ing.
In conclusion, water washing with warm water appears to be
applicable provided that the washing times are kept relatively
short. Since water washing is obviously effective in removing
sulfates and bisulfates along with adhered fly ash on the cat-
alyst surface, it appeared possible that the catalyst could be
regenerated to some extent by this procedure.
Site Tests
In the preliminary tests samples of catalyst were soaked in
washing solutions. However, soaking in water is impractical for
commercial systems. Therefore, instead of soaking in wash water
6-41
-------
SYMBOL
WASHING TIME
Cffin 3
o
- NONE
' NO
3
-------
100
200
400
100
150
REACTION TEMP C*C]
SYMBOL
WASHING TIME
C mm 3
O NOME
4 5
A 10
O JO
Tfr - 60
O - »20
GAS
COMPOSITION
NO
310
PPM
HKj
340
PPM
so,
250
PPM
0|
4
%
CO,
12 2
%
H]0
>0
X
11
BALANCE
AV
20 Nm/H
Figure 6-18
Influence of Water Washing o.a.-
Catalytic Activity (Used Catalyst)
6-43
-------
c\
I
£»
.tfc
U
UJ
CL
to
UJ
cn
in
o
to
VIRGIN
120 MIN
a
USED
120 MIN
Figure 6-13 Abrasion Test,
-------
TABLE 6-1. ANALYSIS OF SAMPLES OF WASH WATER
Catalyst
Washing Time
Na
K
Fe
V
A1
Ti
(minutes)
5
na
na
na
3.6
3.8
none
Virgin
10
na
na
na
3.5
2.7
none
Catalyst
30
na
na
na
6. 2
2.9
none
60
na
na
na
7.0
3.1
none
120
na
na
na
9.0
2.0
none
5
9.6
5.0
1.2
3.9
35. 3
none
10
10.0
5.0
1.2
4.4
41.1
none
Catalyst
30
10.6
5.6
1.6
5.6
44.8
none
60
11.5
5.6
1.0
5.6
59.5
none
120
o
o
5.6
0.3
6. 3
41.8
none
NOTE: na =» not analyzed
6-45
-------
the catalyst was washed with warm water delivered from a water
hose. The manhole at the upper side of the reactor and the 14"
elbow at the bottom of the reactor were removed. Water at 104°F
was directed from the 1" hose back and forth across the catalyst
surface as evenly as possible. The water flow was 2.7 gpm and
this continued for two hours.
Samples of the waste wash water were taken every 15 minutes
and the pH of the samples were measured right after sampling.
Chemical analyses were performed later.
After the water washing the manhole and elbow were reinstal-
led, the biower was restarted, and the catalyst was dried with
air at 600 scfm and 230 °F overnight.
The analysis of the samples of wash water are shown in
Table 6-2.
The regeneration procedure was evaluated by running a mole
ratio test at 1,100 scfm and 700°F. The results were compared to
an earlier mole ratio test performed before the catalyst was
washed. The improved results can be seen from curves on Figure
6-23.
Before the washing test a sample of catalyst was taken for
analysis. After the plant was dismantled other samples were
taken for analysis. These analyses were performed by Hitachi
Zosen in Japan.
After the regeneration the activity of the catalyst appears
to have been restored to its original condition as indicated by
mole ratio tests. (See Figure 6-20). This would indicate that
it may be practical to restore catalyst activity by water washing.
Table 6-3 (Analysis of Used Catalyst) shows the results from
Hitachi Zosen's laboratory. Except for vanadijm, the elements
and groups shown in this table are not included in the catalyst
and aie, therefore, derived from the fly ash artd appear to be
removed from the catalyst to some extent by water washing.
The vanadium content of tl.e catalyst decreased with water
washing. The actual amount of lost vanadium cannot be estimated
precisely because the catalyst is a solid and uniform sampling
6-46
-------
TABLE 6-2. SAMPLES OF WASH SOLUTION
Time pH Color
Sample of of of Chemical Analysis (mg/1)
No.
Sample
Sample
Sample
Na
K
Fe
V
NH .
4
f04
1
13:35
2.0
Dark Green
46
360
6
470
190
2860
2
13:40
1.8
Dark Green
185
1140
38
780
74
7580
3
13:50
1.8
Dark Green
175
1160
45
650
74
72S0
4
14:05
2.1
Green
83
383
9.2
490
29
1060
5
14:20
2.4
Green
29
188
2.4
220
5.0
2850
6
14:35
2.4
Green/Gray
47
306
24
140
6.2
1910
7
14:50
2.7
Gray/Blue
19
125
1.0
80
8.4
650
8
15:05
2.8
Gray/Yellow
13
88
0.6
65
11.0
280
9
15:20
3.1
Gray/Blue
11
63
0.4
45
5.0
190
10
15:30
3.9
Yellow/Green
6.5
35
0.4
27
3.6
10
-------
too
80
60
40
20
Figure
-o-
AFTER WASHING, U00 SCFM, 700 *F
-o-
BEFORE WASHING, ItOO SCFM, 700 *F
0PERATING PERIOD ' 5200 - 5600 H
—' 1 1 —i_ i i_
04 06 0 8 1.0 12 u
NHj / NO* MOLE RATIO (-1
NOx Removal Efficiency after Water Washing.
-------
TABLE 6-3 - ANALYSIS OF USED CATALYST
(wt %)
Sample
Na
K
S
NH .
4
Decreased V *
Before Washing
.09
.32
CO
•
.008
0
After Washing
Top Block**
.07
.18
. 07
.002
4.1
Top Block
.06
.13
.17
.002
00
•
(N
Second Block
.06
.12
.13
.002
10.6
Third Block
.09
.22
.20
.002
2.3
Bottom Block
.07
.12
.18
.002
•
00
* Percent of decreased vanadium as compared to amount of vanadium
in sample before washing.
** This test piece was from the used catalyst immediately after
watei washing. Other test pieces were taken from the used
catalyst after three days of operation following the regen-
eration of catalyst.
6-49
-------
is difficult. Also, fly ash interferes with this analysis.
Nevertheless, though the loss was small, repeated water washing
is not recommended.
The waste wash solution includes various elements which
leached from the fly ash as shown in Table 6-2 (Samples of Wash
Solution), and some of these metals are toxic. Water treatment
before disposal would be required.
CHANGES IN PRESSORE DROP IN RELATION TO NO^ REMOVAL EFFICIENCY
In the treatment of coal-fired combustion flue gas, it is
well-known that the adhesion and clogging caused by fly ash,
along with the formation and deposition of ammonium sulfate and
bisulfate, cause problems. In N0X removal systems, these prob-
lems are first noticed as an increase in pressure drop. There-
fore, changes in pressure drop were carefully studied during the
operation of the EPA pilot plant.
Since operating conditions varied from time to time in
accordance with the needs of the pilot plant operation and the
scope of the Contract, changes of pressure drop were evaluated
by a ratio of measured value to a calculated value.
Calculation of Pressure Drop
The following equation has been established by Hitachi Zosen
for calculation of pressure drop on NOXNON 500 catalyst and
NOXNON 600 catalyst: 2
dPcl . +
e e y
whereas,
dPcal = pressure drop (r.imAq)
n = number of catalyst layers (-)
e = void fraction (-)
f = friction coefficient (-)
C a constant (-)
6-50
-------
L 3 length of catalyst layer (m)
D = hydraulic diameter (m)
e 3
rg = density of gas (kg/w)
v^ » superficial gas velocity (m/s)
g = acceleration of gravity (in/s^)
The above equation has proven to be accurate in the measure-
ment of pressure drops from various pilot plants.
Data from the EPA pilot plant obtained in flow rate tests
performed in May 1980 were evaluated according to this equation.
These data were plotted in Figure 6-21 and clearly show the
pressure drop variation through changes in flue gas flow rate,
temperature, and other operating conditions. Actual measured
pressure drop was slightly higher than calculated. However, in
the range of 1000 to 1500 SCFM the flue gas flow rate did not
influence the pressure drop variation.
Pressure Drop Observations
On May 14, 1980, at the first catalyst inspection, addition-
al asbestos yarn was added to ensure the seal between the reactor
shell and catalyst box. As a result, dPact/dPcal increased from
0.9 to 1.0 - 1.1 and stable operation continued at this level un-
til the second catelyst inspection on June 27, 1980.
From this date, several experiments were carried out, and
the tendency for pressure drop changes are summarized below.
- When the pilot plant was shut down for some reason the
pressure drop generally increased after the next startup.
The increased pressure drop usually continued at this
high level although the sootblower was operated three
times a day. However, from time to time, the previous
pressure drop was restored after a few. days "of continu-
ous operation.
— When the flue gas flow rate was changed, the pressure
drop increased at times. For example, ii. November 1980
6-51
-------
I.I
Q.
-o
I 0
a
o_
0 9
&
0
A
o 5 / 22 , 24/80
a
REACTION TEMP :
710 °F
* 5/24/80
1
REACTION TEMP :
' "
780 °F
i "
1000 1100 1200 1 300 1400 1500
FLUE GAS FLOW RATE {SCFM )
Figure 6-21 Effect of Flow Rate on dPact/dPcalc.
-------
after some twenty days of steady operation the flow rate
was changed from 1300 SCFM to 1500 SCFM and after two
days at this higher flow rate the pressure drop suddenly
increased and did not return to its original level for
several days.
- Whenever Georgia Power Unit #3 Boiler operated the
economizer soot blower repeatedly, the pressure drop
across the catalyst increased.
- On December 18, 1980 the water tubes of the boiler were
washed with pressurized water to remove slag on the tube
surfaces while the boiler was in operation. This abnorm-
al maintenance work caused a serious increase in pressure
drop.
- Operation of the soot blower seems to be necessary for
coal-fired combustion flue gas. On January 10-11, 1981
the pilot plant was operated for approximately 28 hours
without the sootblower. The pressure drop increased and
after the sootblower was returned to operation it took
three days to restore the pressure drop.
As described above, the pressure drop across the catalyst
bed often drifted upward for a while and then gradually decreased.
Overall, however, the operation was relatively stable except when
the boiler tubes were water washed in an abnormal manner on Dec-
ember 12, 1980 causing a permanent increase in pressure drop.
The washing obviously released a large amount of accumulated fly
ash from the tubes and some of this material obviously entered
the reactor. This unusual slug of material could have led to
this pressure drop increase.
Relationship between Pressure Drop and NOjc Removal Efficiency
Relationship between pressure drop and N0x removal effic-
iency observed before and after the first and second catalyst
inspections are as follows:
6-53
-------
flow rate pressure drop N0X removal
date
(scfm)
(inches Ho0)
05/15/80
1,500
1.1 - 1.14
91.5 - 95.0
06/24/80
1,500
1.08 - 1.11
89.5 - 92.0
06/26/80
the first
catalyst inspection
(1,200 hours)
06/29/80
1,500
1.19 - 1.26
89.5 - 93.0
07/04/80
1,500
1.23 - 1.32
91.5 - 94.0
08/22/80
1,300
1.08 - 1.15
90.5 - 94.5
08/24/80
1, 300
1.05 - 1.09
90.5 - 94.5
08/27/80
the second
catalyst inspection
i (2,500 hours)
08/27/80
1, 300
1.27 - 1.32
88.5 - 91.5
08/28/80
1,300
1.15 - 1.29
89.5 - 92.5
After approximately 1,2C0 hours of operation, even though the
pressure drop increased, N0X removal was not influenced. How-
ever, after 2,500 hours operation, N0X removal decreased with an
increase in pressure drop.
When the pilot plant was shut down and started up, the pres-
sure drop generally increased, and these increases were stepwise.
The increase in pressure drop seems to have resulted from an in-
crease in flue gas flow rate which was caused by partial clogging
of flue gas passages by adhesion of fly ash.
The relationship between pressure drop and N0X removal
efficiency is shown in Figure 6-22. According to this data,
N0X removal efficiency decreased along with an increase in pres-
sure drop.
Operating data after regeneration of catalyst shows the
same NOx removal efficiency as the initial efficiency. There was
a definite decrease in pressure drop after regeneration, how-
ever, it was still about twenty percent above the initial level.
Before regeneration it was about forty percent ?Jbove the initial
level.
Although the surface of the catalyst was cleaned by water
washing, clogging by asbestos yarn found between catalyst blocks
increased the pressure drop.
6-54
-------
1 FLOW RATE ! 100 SCFM
REACTION TEMP 700 °F
NHj/NOj RATIO 10
SYMBOL / OPERATING PERIOD
[° APPROX 450 H
• O 5,000~5,500H la WITHOUT SOOT BLOWER)
lo APPROX 5,500 H ( AFTER REGENERATION )
1 i
05 10 15
PRESSURE DROP ACROSS CATALYST BED I IN H20 )
Figure 6-22. Effect of Pressure Drop on NOx
Removal Efficiency.
6-55
-------
During the final stages of the pilot plant operation runs
were made on two occasions without the use of the sootblower.
These trials produced a rapid increase in pressure drop. It was
concluded that the sootblower was required to maintain stable
operation of the system.
CATALYST INSPECTIONS
Inspections
During the operation of pilot plant, the catalyst was
inspected from time to time while the pilot plant was shut down.
05/14/80 The manhole was opened and the top surface of
catalyst was observed through the manhole. The
reason for this inspection was a relatively low
N0X removal efficiency and low pressure drop
across catalyst.
Additional packing for sealing clearance between
the catalyst and reactor wall was added.
06/27/80 Element of Flue Gas Heater was broken and replaced.
During this shut down the catalyst was inspected.
08/27/80 The manhole of the reactor was opened to repair
the Soot Blower. The catalyst was observed at this
time.
11/15/80 The pressure drop fluctuated and increased, and the
top surface of the catalyst was inspected from the
manhole prior to starting mole ratio test. Nothing
unusual was found at the top of catalyst.
12/26/80 The pilot plant was shut down for the Christtras
holiday, and the catalyst was observed prior to the
start-up.
01/26/81 Before regeneration of catalyst, the catalyst was
observed and a few plates were sampled from the top
layer.
6-56
-------
02/0 2/81 The operation of pilot plant was terminated due to
modification of Unit #3 Boiler of Georgia Power
Company. The catalyst box was placed on the ground
and all the catalyst plates were dismantled, inspec-
ted and sampled. Samples of catalyst were air
mailed to Hitachi Zosen's Technical Research Insti-
tute, Osaka, Japan.
Observations
The first inspection on May 14, 1980—
On May 14, 1980 the pilot plant operation was shut down and
the catalyst was observed from the manhole.
The catalyst had been installed in the reactor on April 22,
1980, however, due to repeated leakage from water tubes of Unit
#3 Boiler, introduction of flue gas was stopped three times, and
air was introduced into the system during these shut downs.
Therefore, operating time with flue gas was only 18 5 hours.
During the operation of 185 hours, NOx removal seemed lower than
designed efficiency.
Operating conditions before the shut down were:
flue gas flow rate : approx. 1,200 scfm (1,930 Nm /h)
temperature
NO concentration
SO2 concentration
pressure drop
700°F (370°C)
approximately 480 - 510 ppm
600 - 1,000 ppm
0.76 - 0.79 inches H20
(19 - 20 mm Aq)
NO removal efficiency : approximately 91%
Pressure drop was almost the same as the calculated value,
however, N0X removal of 91% was lower than expeqted because
designed NOx removal at the beginning of the operation was 95%,
and it seemed that there was no allowance to continue the opera-
tion for more than six months or one year. For this reason, it
6-57
-------
was decided to shut down and observe the catalyst from the man-
hole.
When the catalyst was observed it was seen that some of the
corrugated plates had moved upward a few inches. None of the
flat plates had moved. However, the upward displacement of these
plates did not affect the efficiency of the system.
Some catalyst plates at the soot blower side were loose, and
a few plates could be moved by hand. Bypassing of flue gas
through these loose plates could be expected. These loose plates
were tightened with added asbestos yarn.
The movement of the corrugated plates was caused by vibra-
tion of catalyst plates. It is required that in commercial
applications a method of tightening and fixing in place the cat-
alyst plates needs to be devised.
From the traces of fly ash adhering to the inside of the
reactor shell, it was assumed that there might be partial flue
gas flow through the clearance between the catalyst box and the
reactor shell. Additional asbestos yarn was added to make sure
of the seal.
No abrasion caused by fly ash or soot blowing was found.
Other Catalyst Observations from the Manhole—
From time to time the catalyst was observed when the opera-
tion was shut down for some reasons. Appearance of the catalyst
was almost the same each time.
Clogging and masking by fly ash was not observed. The
operation and frequency of soot blower seemed suitable.
- Wearing away of catalyst components from the stainless
wiremesh proceeded gradually through each inspection.
However, the extent of the wear could not be observed
until the catalyst was disassembled on 2/2/81. In
general, abrasion seemed not serious.
From the traces of fly ash on the inner surface of the
reactor shell, the flue gas flow distribution seemed
6-58
-------
acceptable.
- The upwurd displacement of some of the catalyst plates
did not continue after the first inspection. It quickly
reached a stable configuration.
Catalyst incpection after the operation terminated—
On February 2, 1981 the pilot plant operation was completed.
The catalyst was removed from the reactor, placed on the ground,
disassembled, and all of the catalyst plates were inspected. The
catalyst layers consisted of four blocks. The clearances in
between were carefully observed.
- Abrasion
Abrasion was observed at only the flat plates in the
top block. These appeared relatively abraded and a por-
tion of the catalyst material had been completely worn
away. Corrugated plates were not abraded. All of the
plates in the other blocks were unabraded and looked
like new catalyst.
- Clogging and Adhesion by Fly Ash
The catalyst surface and the passages between the
catalyst plates appeared to be clean. There was no
fly ash adhering to and clogging the activated catalyst
surface in the passages.
However, in the clearances between the top block and
the second block, fibers of asbestos yarn along with
fly ash were found. The upper surface of the second
block seemed to be blocked as much as 15 to 20 percent.
In the clearance between the second and third block fine
fly ash was found on the catalyst supports, piled in the
clearance area, and blocking about 15% of the passages
of the third block.
6-59
-------
Discussion
In summary, approximately 30 to 35 percent blockage caused
by fly ash was found in the clearance areas between the blocks of
catalyst. Passageways between the plates of catalyst, however,
were clean. The blockage could affect NOx removal efficiency
and pressure drop during operations. However, this did not pre-
vent a successful operation of the project.
EVALUATION OF NOXNON 600 CATALYST
NO]t Removal Efficiency
Operation of the pilot plant with NOXNON 600 catalyst began
on April 22, 1980 a..d terminated on February 2, 1981 due to
scheduled maintenance of the boiler. The operating time with
flue gas was 5,620 hours and tests examined during this period
included catalyst life test, catalyst performance test, and
transient tests. Controllability and reliability of the entire
system was also evaluated at the same time.
Operating conditions were varied for the performance tests
and transient tests. Therefore, when the activity of the cat-
alyst was to be evaluated the operating conditions were set at
certain consistent levels each time. These conditions were a
flue gas flow rate of 1000 - 1200 SCFM, a reactor temperature of
700° - 720°F, and a NH^/NO^ mole ratio of 1.0. Under these con-
ditions the NOx removal efficiency was measured over a period of
several hours to determine the condition of the catalyst.
At the beginning of the operation : 90-9 4%
At the end of August '80 (2500 hours): 90 -.,.94%
November 1980 (4000 hours^: 90 - 92%
Before transient tests 12/80 (4420 hoursj: 90 - 91.5%
During transient tests 1/5/81 (5000 hours); 90 - 91%
During the transient tests, the "Sudden Load Change" test,
6-60
-------
and the "Operation Without Soot Blower" caused an increase in
the pressure drop and a decrease in the NOx removal efficiency.
Experience has shown that these deviations would be corrected if
the pilot plant was operated under stable conditions with period-
ic operation of the sootblower. However, due to the limited
period of time available, the testing program had to proceed and
there was no opportunity for the catalyst restoration by the
usual method. Before the regeneration of the catalyst the N0X
removal efficiency had slipped to 88 percent.
After the regeneration of the catalyst the efficiency was
restored to 91.0 - 9 4.5 percent which was equivalent to the
activity when initially tested.
Although approximately 30 to 35 percent of the cross-section
area of the catalyst was found to be clogged by fibers of asbes-
tos yarn along with fly ash (see "Catalyst Inspections on Site")
satisfactory NOx removal efficiency was achieved throughout the
pilot plant operational period of over 5,000 hours.
It must be emphasized that the problems with the asbestos
yarn were peculiar to this pilot plant and would not be expected
in commercial applications.
Flue Gas Flow Rate
The NOXNON 600 catalyst for the pilot plant was designed to
operate at an Area Velocity (A.V.) of 9.6 (1057 SCFM). However,
operations at a much higher flow rate, at an A.V. of 15 (1650
SCFM) provided the desired 90 percent NC>X removal efficiency.
Therefore, it was anticipated that between the designed flow
rate of 1057 SCFM and the normal operating flow rate of 1500 SCFM
the N0x removal would be unchanged. As 3een in Figure 6-3, the-
flow rate had little or no effect on the removal through the
program.
6-61
-------
Mole Ratio
Mole ratio of ammonia to N0X is an important variable
affecting N0X removal efficiency. One mole of ammonia reacts
with at lease one nole of N0X and for a 90 percent removal a
typical required mole ratio would be 0.9 to 1.0.
Ammonia Slippage
Since a continuous ammonia analyzer was not available during
the operation of the pilot plant, ammonia had to be measured per-
iodically by wet analysis. The method was not considered reli-
able until after August 15, 1980 and all data relative to ammonia
leakage was collected after that time. By then the catalyst was
probably partially clogged by asbestos fibers and fly ash and
this caused a higher than expected ammonia slippage. <\t the flue
gas flow rate of 1100 SCFM and an operating temperature of 700°F,
the following results were collected in the operating period
between"3200 hours and 4500 hours.
Hole Ratio No]t Removal (%) Ammonia Slippage (PPli)
Operating Temperature
Between 640°F and 780°F the reactor operating temperature
had no effect on N0X removal efficiency.
NO[t Concentration in the Flue Gas
NOx concentration does not affect NOx removal efficiency
within the typical boiler range of between 300 and 1,000 ppm.
6-62
0.6
0.7
0.8
0.9
1.0
60
70
78
86
90
0
4 - ID
20 - 30
35 - 40
50 - 60
-------
S0H Concentration in the Flue Gas
So„ concentration does not affect NO removal efficiency.
X X
Power Plant Load Changes
The system can follow boiler load excursions in commercial
systems without problems. This was proven by cycling the system
three times between 0.4 MW equivalent and 0.95 MW equivalent
without any consequent effect on catalyst activity.
Emergency Shut-off of Ammonia Feed
Automatic shut-off of ammonia is required to avoid build-up
of deposits at low temperatures. This system was incorporated
in the pilot plant and proved to be reliable through several
cycles of temperature excursions.
Cold Start-up
The system exhibited no problems when it was started up with
cold flue gas without prior heating of the system with hot air.
This proved that in a commercial system no auxiliary heating is
required for startups.
Boiler Shutdown and Startup
The system was allowed to shut down ar.d start up with the
boiler with no air purging ot the system and without auxiliary
heating. The system withstood this transient period without
difficulty.
Sudden Load Change
Six cycles of sodden load changes were tried. This increas-
6-6 3
-------
ed the pressure drop somewhat and the NO removal efficiency
decreased slightly. However, experience has shown that normal
conditions are usually restored after awhile during normal plant
operations. Therefore, it is probable that sudden load changes
will not cause serious problems.
Operation Without Sootblower
This brought an obvious decrease in NOx removal efficiency
and an increase in pressure drop. Therefore, periodic operation
of the sootblower is imperative. The pilot plant normally oper-
ated the sootblower three times each day and this frequency seems
adequate.
Regeneration of Catalyst
When the catalyst was washed with warm water the N0x removal
efficiency was restored to its initial efficiency. The pressure
drop also seems to have been partially restored, however, this
system was partially clogged with asbestos fibers which is not
a typical condition. The effectiveness of water washing for
pressure drop restoration, therefore, is uncertain. Also, the
waste wash solution contains dissolved metals from the fly ash
which would require water treatment before disposal.
Abrasion of Catalyst
Among the four blocks of catalyst, only some of the flat
plates in the top block showed a loss of catalyst by abrasion.
This was probably due to the vibration of the loose plates caused
by the flue gas flow. There was no abrasion in the other blocks.
A commercial system should incorporate a tighter catalyst struc-
ture to avoid such vibration.
6-64
-------
Clogging
Anticipated clogging of the catalyst by fly ash was prevent-
ed by operation of the sootblower three times a day. Reducing
this frequency of operation may be possible but it was not tested.
Apart from the clogging anticipated by fly ash, asbestos fibers
unfortunately led to plugging of about 30 to 35 percent of the
catalyst passages between the top catalyst blocks and the second
catalyst blocks. This was caused by asbestos rope used to seal
passages between the catalyst box and the reactor shell.
6-65
-------
SECTION 7
CONCLUSIONS
NOx REMOVAL EFFICIENCY
NOy removal efficiency of more than 90% was demonstrated
during the operating period of approximately 5,000 hours (from
April 1S80 to the beginning of January 1981) at the designed
capacity of 0.5 MW equivalent.
Following the above period, transient tests were run and
the operation of the pilot plant was terminated after 5,620 hours
of cumulative operating time. Performance was not adversely
affected by sudden boiler load changes, cold startups, sudden
boiler shutdowns and start-ups, and low boiler load operations.
Therefore, it was expected that NOx removal of more than 90%
would have continued longer, for example, more than one or two
years as was demonstrated in other pilot plant operations in
Japan.
AMMONIA SLIPPAGE
Ammonia slippage i this pilot plant was higher than expect-
ed as shown below:
Mole Ratio NO}[ Removal (%) Ammonia Slippage (ppm)
0.6 60 0
0.7 70 4-10
0.8 78 20 - 30
0.9 86 35 - 40
1.0 90 50 - 60
-------
These results probably resulted from the decrease in
apparent surface area of catalyst caused by clogging of the
clearances of catalyst plates with fibers of asbestos yarn
along with fly ash as described in "Catalyst Inspections."
Recent representative ammonia emissions with NOXNON 600
catalyst for coal-fired combustion flue gas demonstrated at a
pilot plant in Japan are, as referred to in Figure E-2 and E-3
in "Appendix E,° less than 5 ppm at the NOx removal efficiency
of more than 80% and less than 10 ppm at the NOx removal effi-
ciency of more than 90%.
A commercial system would be expected to experience low
ammonia slippage similar to that found in tests in Japan. The
catalyst in the EPA pilot plant was obviously clogged by asbestos
and fly ash to an extent that a significant decrease in avail-
able surface area was found. Such a problem could not occur in
a commercial plant as asbestos would not be used for this purpose.
NOXNON 600 catalyst operating under typical conditions in the
presence of only fly ash would remain unclogged as shown by the
experience in Japanese pilot plant operations. Therefore, the
ammonia slippage data from the EPA pilot plant is not considered
valid and the Japanese experience seems more reliable.
PRESSURE DROP ACROSS CATALYST BED
The pressure drop in commercial applications would be ex-
pected to be the same as that experienced in the EPA pilot plant
provided that the size and configuration of catalyst, superficial
linear velocity of the flue gas across the catalyst, and the
tewperature of flue gas are the same.
The pressure drop with NOXNON 600 catalyst in the EPA pilot
plant was between 1.0 and 1.4 inches H20. Such low pressure
drops reduce power consumption resulting in lower operating
cost.
7-2
-------
CONTROL SYSTEM
The analyzers re-quired frequent backflushing and cleaning
of sampling probes and sampling tubes, the regular replacement
of filters in the sample gas conditioner, and also frequent
instrument air purging of the impulse lines of the controllers
and indicators. However, in general, controllability and reli-
ability of the pilot plant operation was satisfactory.
A desirable addition to the control system would be a
continuous ammonia analyzer to measure slippage from the system.
FLY ASH PROBLEMS
Fly ash contained in flue gas varies in its characteristics
and behavior relative to its clogging tendency in catalyst beds.
This depends on the source and composition of the coal. At
present, qualitative measurements to estimate the tendencies to
agglomeration and cohesion of fly ash are available through
cheirical analysis and thermal processing tests. However, more
adequate and accurate methods to estimate the behavior of fly
ash in catalyst beds will be required. It is expected that this
method when further amplified will be useful in selecting ade-
quate linear velocity to prevent abrasion caused by fly ash, and
to determine the necessity and operating conditions of the soot
blower. Improvement of catalyst configuration to prevent clog-
ging by fly ash is also expected.
REGENERATION OF CATALYST
Testing for catalyst regeneration was examined immediately
before the pilot plant was dismantled, and the results were.very
encouraging. The regenerated catalyst exhibited-properties of a
virgin catalyst.
However, due to the limited time available, problems relat-
ed to the catalyst regeneration had not been clarified. For ex-
7-3
-------
ample, the method of drying the catalyst and the reactor after
regeneration without encountering corrosion or fly ash clinging,
and the treatment of waste washing solution should be investi-
gated before commercial application. This area should be studied
further since the cost of this technology could be substantially
reduced if the catalyst life could be extended by in situ regen-
eration techniques.
7-4
-------
APPENDIX A
EQUIPMENT DESCRIPTION
FLUE GAS INTAKE SYSTEM
Flue gas to be used as input to the pilot plant was drawn
from the boiler flues downstream of the economizer and intro-
duced to the Flue Gas Heater. Multiple withdrawal points from
the flues were used to ensure that the pilot plant gas was
representative of that in the flues, particularly with respect
to fly ash (dust) content and particle size distribution.
REACTOR
The gas flows down through the reactor in contact with the
catalyst and the N0X content of the gas stream was reduced by
reaction with the ammonia to form small quantities of gaseous
nitrogen and water which remain in the gas stream.
The operating conditions in the reactor were specified so as
to simulate the operating conditions of boilers in power plants.
Fly ash from the flue gas can settle out in the reactor and
tend to block flow through the reactor and to blind it off and
so partly inactivate the catalyst. These effects of dust set-
tling in the reactor were kept within acceptable bounds through
the use of soot blowing apparatus. This was demonstrated by
the soot blower (which will be described in a later section).
The reactor size was 10* high, 23Jj" square, with a 16"
transition at the bottom. The sootblower, described below,
was attached to the reactor. The catalyst is described in
Section 4 of this report.
A-l
-------
FLUE GAS HEATER
The flue gas heater was designed to heat 1600 SCFM of flue
gas from 550°F to 770°F.
The heater consisted of one unit containing 3 08 tubular
elements connected into six subcircuits. The six subcircuits
were divided into two groups, each of three circuits. One of
the groups was on on-off control and the other was on automatic
temperature control.
The entire heater was rated at 162 KW at 460 V, 3 phase.
The elements were contained in a weatherproof box of
welded steel construction with a gasketed cover. A thermocouple
which senses gas temperature was mounted on the back plate.
SOOT BLOWER
This was a retractable soot blower designed for side wall
entry complete with all controls, supports, drive mechanism and
control panel. The main lance was equipped with a removable
cross T-head tube 3 inches in diameter 21H inches long and
having 11 x 3/16" diameter holes for gas injection in a vertical
downward direction.
Lance traverse was 30-inches at an approximate speed of
24 inches per minute. Controls were provided for continuous
extension and reaction of lance throughout the pre-set operating
cycle.
The lance seal was gas-tight to prevent air leakage into
the duct.
Blowing medium could be steam or air at 130 psig and a
temperature of 660°F. A maximum design temperature and pressure
were 840°F and 150 psig. The soot blower and auxiliaries were
suitable for outdoor operation.
The blower was furnished with blowing medium inlet valve
with an outside adjustable blowing pressure control.
The operation was semi-automatic. Operation was initiated
A-2
-------
by a pushbutton. All subsequent operations were automatic.
Initial operating cycle was set for 2.5 minutes. Upon completion
of the cycle the lance would retract, the unit would shut down
and be ready for re-start. All components such as motor, limit
switches, etc. were wired to terminal strips in a junction box
for connection. The unit was driven by an electric motor.
HEATER FOR SOOT BLOWER
Service Conditions
A stream of 400 SCFrt of air entering at 130 psig and ambient
temperature was to be heated to 660°F ± 20 °F. The same heater
was also required to heat an alternate flow of 1760 lb/hr of
steam entering at 130 psig and 360-500°F to 600°F ± 20"F.
Description
One circulating heater rated at 96 KW (48 KWbundle) at
460 V, containing two bundles each of 15 General Electric Calrod
stainless sheathed tubular hairpin elements, was installed.
The pressure drop was below 0.1 psig at stated flow and temp-
erature. The heater was mounted and wired together with the
control enclosure on a structural steel base.
The control assembly was in a weather-proof enameled steel
vented enclosure with all components mounted and wired.
DUST SEPARATOR SYSTEM
This consisted of a vertical standing cyclone having an
inlet and outlet flanged connection, motor driven rotary valve
to discharge collected particulates, and solids receiving hopper,
all mounted on a fabricated steel supporting structure.
The system was designed to remove particulates larger than
30 microns at a design flow with a minimum efficiency of 95%.
A-3
-------
Collected particulates were discharged into a receiving hopper
by means of a motor-driven rotary valve. The hopper was designed
to discharge solids into a vacuum removal system.
BLOWER
Performance and Characteristics
ACFM
S.P.
BHP
RPM
Weight
Operating Temp.
Tip Speed
Rotation Discharge
3,700
21" hot, 50" cold
17.4 hot, 41.4 cold
3550
2230 lb.
770°F hot, 70°F cold
26062 Ft/Min.
CCW-UB
Coolant Water
Max. Pressure 80 psig
Normal Pressure 15 psig
Flow Rate 0.7 GPM
Motor Specification
25 HP
460 Volt
3600 RPM
AMMONIA SUPPLY UNIT
Ammonia Storage Tank
1000 gallon water capacity having a design working
pressure of 250 psig.
A-4
-------
Instrumentation and Controls
Relief Valves —
One d al relief valve assembly consisting of one three-way
dual control valve and two Safety Valves having a relieving cap-
acity of 1500 CFM air per minute at 250 psig. The dual control
valve permitted the shutting off of one relief valve at a time
leaving one protecting the tank at all times. Each relief valve
had full discharge capability for the size of the tank.
Valves and Fittings —
Tank and openings were fitted with excess flow check valves
between the openings and the shut—off valve with the exception
of the relief valve assembly.
Vaporiser Assembly
Immersion Heater with one external temperature cut-out
switch.
A-5
-------
APPENDIX B
INSTRUMENTS AND CONTROLS
CONTINUOUS ANALYZERS
Description of Equipment
Flue gas samples were drawn continually from the ducts before
and after the reactor. Comparisons between the inlet and outlet
samples were made to determine the removal of NO^. The inlet port
was located downstream of the flue gas heater and upstream of the
ammonia feed point. The outlet was located in the duct downstream
of the cyclone and before the fan.
The inlet port and outlet port of the reactor had sample
probes (sintered 20 micron stainless steel) thai' removed large
particles. The probes were connected via heated Ixnes to the
sample gas conditioners where the sample stream was again filter-
ed (5 microns) and diluted with dry air to reduce the dew point
of the sample stream.
The sample gas from the pick off point was conducted to the
Sample Gas Conditioner via electrically heated sample tubing.
The heated tube was teflon and was controlled at a temperature
of 170°F. Heating maintained the sample temperature above the
dewpoint and prevented alteration of the gas ccnstituents.
Sample Gas Conditioner —
The Sample Gas Conditioner was a module which prepared a flue
gas sample for analysis; the gas was filtered and the dewpoint was
B-l
-------
lowered by dilution. The module contained a network of control
flow orifices which accurately diluted the sample for NO/NO^ and
S02 analysis. This network was contained within a heated chamber
which was controlled at a temperature above the sample gas dew-
point.
Control Module —
Periodically the stack monito.-ing sequence vas interrupted
to check the operation of the analy . first the analyzer
was purged with zero gas, which was «.hert L->I lowed by a calibra-
tion gas (supplied in coi'i.iressea gas cylindtrs) purge. The
sample lines were periodically purgec with a short surge of com-
pressed air in order to clear the lines ar.d the stack ptobe
filter of particulates.
Model 10A Chemilunr. escent 'lO/NO^ Analyze r (inlet San; le) —
For measurement of NO concentrations, the gas sample to oe
analyzed is blended with 0^ (ozone) in a »< action chamber. The
resultant chemiluminescence is monitored chrough an optical fil-
ter by a h'.gh sensitivity photoinultipl'Vr ^»itioned at ci t' und
of the chamber. The filter/photoniUi. J cowresponds
to light in '."irrow avoler.glh banc unions to this N0/0, *-rac-
j
Lion. i Joi- paramct'.'.s can be adjust,.:d in such a way that the
output, from the photo?ul' iplier tube is linc^; ly proportional to
ths NO concentration.
To measure NO concentratiops (i.e., NO plus NO-), the
x 2
sample gas flow is first diverged through a r^-to-NO converter.
Model 14D Nil^ Analyzer (Outlet Sample) --
This analyzer is •¦he same as the NO^ analyser described
above except for the following changes:
B-2
-------
The TECO Model 14D has rnach higher sensitivity. This higher
sensitivity is required because it measured the outlet stream
which contained low levels of NH, and NO,,. The Model 14D also
utilized a dual chamber design which was to continuously monitor
the NHj concentration.
The chemiluminescent N0X analyzer was converted to a NH^
analyzer by adding a molybdenum converter to the inlet of one of
the reaction chambers (resulting in the measurement of NO and
NO2) and using a stainless steel converter with the other reac-
tion chamber (resulting in the analysis of NO, NO2 and NH^). The
analyzer read the output signal from both chambers and automatic-
ally subtracts the two signals, and the resultant difference is
the NHj concentration.
Pulsed Fluorescent S02 Analyzer —
The SO2 Analyzer within the system used Thermo Electron's
pulsed Fluorescent technique. Within the Analyzer a high-inten-
sity ultraviolet lamp is pulsed to excite SO2 molecules, which
fluoresce in the blue region of the light spectrum, between
1900 and 2300 angstroms. A characteristic fluorescent light is
emitted as the excited molecules returned to the ground state.
This specific illumination is passed through a narrow band
filter and impinges upon the sensitive surface of a photomultip-
lier tube. The emitted light is proportional to the concentra-
tion of SO2 molecules in the sample. The amplified and condi-
tioned signal provides a meter reading and an electronic analog
signal for recorder output.
Operating Problems
Dry Ammonia Analysis —
The analytical system was originally intended to provide
B-3
-------
con' lnuous aimnonxd redUHt^a ah uie a. iuc ay be present to NO
because the analyzer can only detect NO. The design provided
a molybdenum converter which would convert only NO2 bat not
The second converter, the stainless steel unit, converts both
the NO2 and the NH^. The gas sample would be split between two
reactor chambers: one equipped with a molybdenum converter, the
other with a stainless steel converter. The analyzer could read
the output from both chambers and could automatically subtract
the two signals, the resultant difference being the ammonia con-
centration.
The problem which finally defeated the system was the loss
of ammonia in the sample line and probably in the probe, if
ammonia was introduced directly into the analyzer the ammonia
could be detected. However, samples from the duct lost its
ammonia content upon passage through the teflon tubing (heated
to 170°F). The atnmonia probably reacted with sulfuric acid mist
deposited in the tubing. Possibly there was also a reaction with
S02« Whatever, the system could not be used to detect ammonia.
Problems with Flue Gas Sample Probes —
The analytical system for the determination of N0X had sev-
eral unexplained problems for some time. Major efforts to deter-
mine their causes were undertaken. It appears that the causes
were catalytic material which had become imbedded in the sample
probe.
The first indication of something amiss occurred when it was
found that the outlet N0X reading (after the reactor) was affect-
ed by the rate of sampling flow through the sample line. This
effect was noticed when the oxygen analyzer pump was turned on or
off. The oxygen analyzer draws the sample from the same sample
B-4
-------
probe and line as the NO^ analyzer. There is a tee in the line
before the NO gas conditioning system and the gas is drawn
X
throv.h one leg of the tee to the oxygen analyzer. When the
oxygen pump was turned on the amount of sample drawn from the
outlet duct through the probe and sample line was doubled. When
this happened the apparent N0x reading dropped substantially.
Concerted efforts to determine the cause of this phenomenon
were unsuccessful. The analyzer gave accurate results when the
standard gas was fed through the sample line with or without the
oxygen pump in operation. There were no leaks found in the
system and the same event occurred after washing of the sample
line.
The next unexplained event occurred when the solenoid valves
on the backflush (blowback) lines were replaced. Every 15 min-
utes the sample lines are blown back for a few seconds with high
pressure instrument air to remove dust from the probes. It was
found that the solenoid valves which supply the air had orifices
much smaller than specified. The amount of air supplied for blow-
back was lower than desired. Larger solenoid valves were obtained
and installed. These provided considerably more air.
When these new solenoid valves were tested two differences
in the operation were discovered. One, the oxygen pump did not
appear to affect the outlet NO^ readings. The second and prob-
ably more significant change was that tho NO^ concentration in
the outlet dropped substantially to an unusually low level.
Again, a series of tests were run to determine whether the ana-
lyzer was operating correctly and all indicated that it was.
No explanation could be found. Eventually, after a few weeks
the oxygen pump began to affect the readings but the overall
outlet NO levels remained low.
x
Some weeks later, the next event took place. The outlet
sample probe was replaced by a new one and the apparent NO^ out-
let level rose sharply from 40 PPM to 62 PPM ur\der- the existing
operating conditions. Replacing the original probe restored
B-5
-------
the original condition. The new probe was temporarily replaced
by another new probe which was found to provide identical results.
Apparently, the original probe was reducing the NC>x level for
some reason.
The original outlet probe had been installed on the inlet
line to replace the original inlet probe which had developed a
crack.
One inter*--"-.ing experiment was then tried. Anmonia was
introduced before the inlet probe to see if there was an effect
on the NO^ reading at the inlet probe. Theoretically it should
not because no catalyst is present. However, when ammonia was
sent into the line upstream of the inlet probe the apparent NO
X
i*s the inlet dropped from 580 PPM to 140 PPM. This test was run
for about twenty minutes. It was apparent that a reaction be-
tween ammonia and NO^ took place and could only have occurred
within the probe itself. The old probe obviously had catalyst
imbedded within it!
The catalytic material within the probe must have been picked
up over a period of time because of the erosion of some of this
material from the reactor. Some of the fine particles of active
material were probably imbedded in the sintered stainless steel
of the probe.
After this the probes were often replaced temporarily to
make certain that the permanent probes were functioning properly.
This condition of probes may exist in other installations and
should be considered when experimental or full-scale operating
plants are evaluated. Such probes, if contaminated with catalyst,
will produce artificially high removal indications.
VENTURI FLOW TUBE
Flue qas flow was measured by means of a Vickery-Simms
Venturi tube equipped with pressure taps. From the reduced
pressure at the venturi throat the flow rate was measured. The
B-6
-------
Venturi was about 60" long with a throat diameter of 7 3/8".
The unit was installed in a 14" pipe duct with flanges.
DATA LOGGER
Data Acquisition System (Logger) for scanning and logging
analog inputs.
The system included internal visual display and recording
of all data and a cassette tape recorder for peripheral output
recording. The data logger had the following specifications:
(1) Enclosure: Portable for desk top with dust cover
(2) Input Terminals: Screw Type
(3) Inputs: 12 linear, 4-20 MA (includes 2 spare points),
6 non-linear direct thermocouple ISA Type K (includes
2 spare points)
(4) Fixed Data Entry: For data identification associated
with a test
(5) Day Calendar
(6) Serial Output
(7) Power Supply: 120 V, 60 Hz
Cassette Recorder Specification:
(1) Enclosure: Desk top mounting
(2) Recording Media: Magnetic tape cassette
(3) Storage Capacity: It is compatible with T.I. 733 ASR
(4) Power Supply: 120 V, 60 Hz
B-7
-------
APPENDIX C
AMMONIA FEED MEASUREMENT
The feed of ammonia into the process was a major problem and
was still subject to question after the program was completed.
The maj'or problem was to accurately measure a continuous, very
low flow of gaseous ammonia. The flow was only about 0.5 SCFM.
The flowmeter originally selected for this installation finally
proved to be the best one to be found. This was a unit provided
by Ramapo Instrument Co. It was a rotometer in which a differ-
ential transformer was attached to the float. The position of
the float was precisely measured by the differential transformer
which develops an output voltage dependent upon its position
which is proportional to the flow of fluid.
This unit caused a considerable degree of trouble for some
time. Most problems were due to plugging of the opening with
bits of dirt. A very fine filter was installed just before the
meter which helped to avoid most of the plugging. The installa-
tion of damping fluid for some reason helped to avoid further
problems.
The unit was calibrated frequently using a dry test meter.
The ammonia was bypassed after the control valve through the
meter for an accurately determined period of time and the volume
of gas was determined. This dry test meter was then often check-
ed by a standardized unit with air. At one time a cylinder of
ammonia was used as the source of ammonia for most of a day. The
cylinder was weighed before and after the test and the .results
indicated the meter to be very accurate.
Other meters were tried at times because <>f the early dif-
ficulties with the Ramapo. One was a mass flow meter which worked
C-l
-------
extremely well for a day or so but which eventually became in-
operable. After several such trials it was abandoned. It seemed
that ammonia was leaking into the electronics causing corrosion.
Another unit tried wa9 a turbine meter which after installation
and failure the vendor admitted to its inapplicability.
C-2
-------
APPENDIX D
WET AMMONIA ANALYSIS
Because of the failure of the dry ammonia analytical system,
ammonia slippage had to be determined by wet methods. There was
much trial and error experimentation before a satisfactory pro-
cedure could be attained. The problem was to obtain an accurate
volume of flue gas (which was at 700°F, and which contained quan-
tities of fly ash and sulfur oxides) to absorb all of the ammonia
present in this gas in a suitable absorbent, and to analyze the
absorbent for its ammonia content.
The sampling probe proved to be a big problem because of the
possibility of acid mist deposition at temperatures below 300°F
with consequent reaction with ammonia in the gas. After many
suggestions and trials a very simple procedure was adopted. The
probe consisted of a 3/8" SS tube inserted through a 3" nozzle
into the duct. The tube passed through a drilled hole in a 3"
cap which sealed the nozzle. The tube was welded to the cap to
prevent leakage. A short length of teflon tubing was then used
to connect the stainless steel tube to the sample impinger train.
After each analysis, the stainless probe and the teflon tube were
rinsed with distilled water and this water was added to the liquid
in the impinger train. In this way no special precautions were
required to avoid condensation in the sample lines.
The ammonia was absorbed in a standard impinger train set
in a tray of ice. Sulfuric acid was used as an absorbent. The
gas was drawn through the system by a vacuum pump and the volume
was measured with a dry test meter.
The absorbant was analyzed by a Nessler method using Hach
D-l
-------
equipment and chemicals. The sample had to be neutralized before
analysis.
This procedure was checked by Radian and by Hitachi Zosen
and was considered to be reliable.
D-2
-------
APPENDIX E
CORRECTION OF NH3/NOx MOLE RATIO
The operation of the EPA pilot plant which was to demon-
strate the performance of Hitachi Zosen's N0X removal process
for the application to coal-fired utility boilers was successful-
ly completed on February 2, 1981. However, there seemed to be a
few questionable cperating data which were inconsistent with the
results of fundamental experiments obtained in the laboratory of
Hitachi Zosen s Technical Research Institute and experimental
results collected at various pilot plants in Japan. One of them
is the high N0x removal efficiency when the mole ratio of NH^ to
N0X is lower than approximately 0.8 to 0.9, and the other is
the high ammonia slip even though the mole ratio is fairly low.
In the presence of oxygen, the amount of ammonia for reduc-
ing N0x is equimolecular to the quantity of NOx in the flue gas.
This is proven from fundamental data and from results collected
at pilot plants. Figure E-l shows the results of fundamental ex-
periments obtained in the laboratory. Figure E-2 and Figure E-3
are experimental results collected at a pilot plant applied to
flue gas exhausted from a coal-fired utility boiler in Japan. In
this pilot plant, the flue gas source from the boiler was divided
into two streams. One of them was directly introduced into the
NOx removal reactor as a high fly-ash-loaded flue gas. The other
wa9 introduced into a reactor after the fly ash was removed by a
high temperature electrostatic precipitator as a low fly-ash-loaded
flue gas. Figure E-4 and Figure E-5 show the results obtained in a
pilot plant which is applied to oil-fired boilers in Japan.
All of these data indicate that the catalytic reduction of
E-l
-------
NO with ammonia is an equimolecular reaction between NH„ and NO .
x 3 x
When the EPA pilot plant was operated, all the analyzers and flow
meters which measure the mole ratio and ammonia slip were frequent-
ly and carefully calibrated and adjusted.
It is well understood that all flow meters and analyzers
involve some errors. However, because of frequent and careful
calibrations and adjustments, direct readings of flow meters and
analyzers including the results of wet ammonia analysis were con-
sidered to be fairly accurate.
On the other hand, the concept of the equimolecular reaction
between ammonia and nitrogen oxide which had been proven through
fundamental experiments through various pilot plant operations is
also an indisputable theoretical principle.
From the above circumstances, an attempt was made to correct
the NH^/NO^ mole ratio based on the assumption that the selective
catalytic reduction of NO with NH^ is an .equimolecular reaction.
After August 1980, ammonia emission could be measured by wet
analysis, and this wet analysis also made the correction of NH^/
NOx mole ratio possible based on an ammonia balance. The correc-
tion was made as described below:
The amount of ammonia injected into the reactor was deter-
mined as the total of an equimolecular amount of ammonia to remove
N0x and the amount of ammpnia emitted from-th'e reactor. The
amount of NO^ introduced into the reactor was not changed. Thus,
the mole ratio of ammonia to NO was corrected.
The basis of this correction in which the NH^ flow meter
was deemed inaccurate was brought about from the very precise
and susceptible configuration of the rotometer, especially, very
narrow clearance between the rotor and the tapered tube which
-ceroid result in erroneous flow rate measuring, particularly, at
the very low flow rate. The phenomena is not anticipated in
commercial systems because of the large flow rate of ammonia.
However, when the small flow rate is applied to pilot plant oper-
ations, the likelihood of an erroneous measurement seems to re-
quire taking this into consideration. This correction is also
E-2
-------
supported by the fact that higher NO^ removal efficiencies were
observed when the mole ratios were relatively low.
As a result of this correction, for example, the effect of
NH^/NOx mole ratio on NOx removal efficiency and ammonia slippage
after 3 200 hours operation, which is described in Section 6,
"Evaluation of NOXNON 600 Catalyst," was revised as follows:
Mole Ratio NO^ Removal Ammonia Slippage
Direct Reading
Revised
(%)
(ppm)
0.5
0.6
60
0
0.6
0.7
70
4 -
10
o
•
00
00
•
o
78
20 -
30
0.9
0.9
86
35 -
40
1.0
1.0
90
50 -
60
In this correction, it should be emphasized that the NOx
removal efficiencies and ammonia emissions were not changed, and
only traversed along with the corresponding mole ratios.
E-3
-------
too
©
e 60
o
in
ec
UJ
o
u
o
20
0
0 0.2 04 06 a8 10 12 14
NHj/NO HOLE RATIO t~)
SOURCE SYNTHETIC GAS I FROM CYLINDERS I
TEST CONDITION
A V : 20.4 HtriJrfi H
GAS COMPOSITION ¦
rN0
500 PPM
SO;
250 PPM
0?
6 X
C02
10 %
HjO
10 %
N!
BALANCE
Figure E-l. Effect of NH^/NO Ratio on NO Conversion
Rate.
E-4
0~ 200 t
J I | I I I I
-------
too
tu
o
li-
lt.
Ui
s
LU
E
X
O
z
0
Q6
as
10
12
nh3/no, MOLE RATIO (-)
SOURCE : COAL - FIRED UTILITY BOILER
(HIGH FLY ASH LOADING GAS)
TEST CONDITION
rFLY ASH CONCENTRATION
A V
REACTION TEMPERATURE
APPROX 14 g/NmJ
(5 0 gran/scFD)
8 9 NmVmZ-H
350 X (662 *F)
Figure E-2. Effect of NH-/NO Ratio on NO
3 x
Reraoyal Efficiency.
(Reported at "The U.S.-Japan Exchange of NO Control
Technical Information Conference" by Hitachi Zosen
on May 27, 1981)
S-5
-------
20
40
10
04 06 08 10
MH3/N0x MOLE RATIO (-)
1 2
SOURCE ' COAL - FIRED UTILITY BOILER
( LOW FLY ASH LOADING GAS J
TEST CONDITION
rFLY ASH CONCENTRATION
A V
,REACTION TEMPERATURE
0 1-02 9/Nit3
(Q04 - 0 08 groin/scFD)
10 2 NmVm2 H
350 *C (662 *F)
Figure E-3. Effect of NH,/NO Ratio on NO
J X X
Removal Efficiency.
(Reported at "The U.S.-Japan Exchange of NO„ Control
Technical Information Conference" by Hitachi Zosen
on May 27, 1981)
-------
SYMBOL
TEMPERATURE
O
375 *C (707 °F)
A
356 *C (673 *F)
P
339 *C (641' *F J
_l I I I 1—
<14 oe 08 10 1.2
NHj/NO* MOLE RATIO (-)
8 50~8 64 Nm3/m2-H
SOURCE : OIL - FIRED STEAM GENERATOR
TEST CONDITION
A V
GAS COMPOSITION
'NOx
SO*
02
CO2
H2O
N2
.DUST
(SOj : 7 PPM)
97-140 PPM
32 PPM
56 %
105 %
115 7.
BALANCE
25 — 45 mq/Nm3 (Dry 80ns)
Figure E-4. Effect of NH^/NC^ Ratio on NC>x
Removal Efficiency.
(Normal Flow Rate)
E-7
-------
100
u
z
UJ
u
3
o
s
80
60
40
SYMBOL
TEMPERATURE
O
375 'C (707 *F)
A
356 t (673 *FI
~
339 *C (642 *F)
04 0.6 Q8 10
NHj/NOx MOLE RATIO (-)
1 2
SOURCE : OIL - FIREO STEAM GENERATOR
TEST CONDITION
A. V
GAS COMPOSITION
'NOx
SOx
02
CO2
HjO
N2
OUST
If 9 ~ 12 1 NmVmZ-H
97~ 140 PPM
32 PPM (SO3 .
56 V.
tO 5 V.
7 PPM J
114 V.
BALANCE
25~45 mQ/Nm3 (Dry Basis)
Figure E-5.
Effect of NH,/NO Ratio on NO
3 x x
Removal Efficiency.
(High Flow Rate)
E-8
-------
APPENDIX F
ABBREVIATIONS
abs
absolute
hr
acf
actual cubic feet
in.
Btu
British thermal unit
kg
°C
degrees Celsius
kW
cm
centimeter
1
dia
diameter
lb
ESP
electrostatic precipitator
m
°F
degrees Farenheit
mg
FGD
flue gas desulfurization
min
FGT
flue gas treatment
mm
ft
feet
MW
ft/sec
feet per second
Nm3
g
gram
ppm
gal
gal Ion
psi
gpm
gallons per minute
SCR
gr
grain
hp
horsepower
sec
scf
hoar
inch
kilogram
kilowatt
liter
pound
meter
mi lligram
minute
millimeter
megawatt
normal cubic meter
parts per million
pounds per square inch
selective catalytic
reduction
second
standard cubic feet
F-l
-------
APPENDIX G
CONVERSION FACTORS
To convert from
To
Multiply by
British thermal unit
degrees Fahrenheit-32
feet
cubic feet
feet per minute
cubic feet per minute
galIons
gallons per minute
grains per standard
cubic foot
pounds
standard cubic feet
per minute (60°F)
tons (short)
gram-calories
degrees Centigrade
centimeters
cubic meters
centimeters per second
cubic meters per second
liters
liters per second
grams per normal cubic
meter
kilograms
normal cubic meters per
hour (0°C)
metric tons
252
0.5555
30. 48
0.02832
0.508
0.000472
3.785
0.06308
2.280
0.4536
1.607
0.90718
G-l
-------
fX~'
r
e
row iTLD?
e = g«* (
if fi
f»
i
tf
L
5s
&
1
s
$
*
9
?
4
»¦
1
5
*•
8
r,
to
£
«L
Oi
S
r
P
1
ta>
rf
5
*:
s
•
E
c:
r
<
s
t
mJ
Wt
f
*
t
I
1
«
Sf
h-
1
\
i
tu
<
o
t
&
•*
3
B
N
X)
c
c
e
4-
2
1
APPENDIX H
LOG CHARTS -- NOXNON 600 TESTS
I
i
i
¦ --i
(.***>
S"
• *1NCD «0 y
Jcaa 3«nna9J . J
t
I
*
I
W 13Tfe
8 S » *
iwcwa *on-
WHU I
ft * tf It "fl*
11V* wumynifl*
"WC? *w t 'la'J
TYPirjf? i nJI'.'E r.;irf T
-------
—1
-<
c.
rnm
frPt SttUTPOWt* CeCAUSfl <* U^^C,
tt*1 *y*Tln« QWM^p pa&ir^ ^ PWt?!%
WS^wv^8!-^
tartto*
L '' IL U
(ttk/ I
CPA PROTECT.
**,J.giPaa ue..
M (5-2- '80~ ff-ll- 'S
hati
fnr-
td<:
,r*'t,
,T
I-l
ttii)
»(!7
.Ji-.ep
_M
ISO
6-1
liv
; vp
c9"»l
IT
It
i,
HO
*7
aK
itt*>
W'»
*!¦—1
>-ra
<»;»2
,W9
¦Jw
.Tia
s-n
f™"
t**.
l£ii.
JUL.
JffiL
A"*
•Tin
-£t-
•£»&
t ! .u: I, j _!_¦ i... :i
jJl[ J lljjj :.I' ! i !..:
3 1
* J c»
1
|-
i 1 ¦
1
-
1
"i
Tm
1
, i i '
• t ,
-: i i-i I
isdtKAMD noy bTf
- nv*r ¦ »**™ . i~ ~ p' r--?— r i - * rr" ¦»¦ '"p1 t—
to aotd recca £waM»)
¦rfcB-O t*U»)
-------
u 8
in* r .
to, ¦ t- A&flLLfll
¦D-srmn
rsKi;
( pp mtlCO
mRMM- L/r^C> OUTLZT TXXtp
m / *
Mw )
jmOUfH HOH-BftTA UWSJ
MP OUTLET ANALittR
IWO-
t* IUO
I50C-
SCPrl
j.0 CWLK)
czn.li)
-1EPA fROJECT
-- ?
H, ¦ V- JLwVLLVILUl
* H (»-•!- *ffO~ *
-XI- *6
SOOT OtLW
VAT a
(wsttO
Vj r»
IU7I5) (IthS)
"T
(1*3.ZSJ
Y»
(Zto 11)
Vh
{««.«)
*r
U3t.«J
*ii._ Hi r»»
DATS
HHf
*
f-fl
Ml
*»»
m
II P
?*v
» «
Vo
»-!»
I*1
* ¦
Oio
JK?
>"»"
i»t u
V3
*?
iijy
1P»
u
»-»
r rr
r n
IK)
n
I">
-
J1''
? »n
V
ft
9-9
1 JT
m
V?
41
o-ts
"IW
ft
>a*
Of!
hi
JH6
\ "
l.*7
1T7
*1
it o»
TT7
%
}TO
JVM
lis
Tf
?*!!
1 IT
ri
ttf
V-
vr>
T"
'¦p
Tff
ti
W
1 ST
1 V,
ir»
n
x «J
? "»
ft
m
44
M
1 *
' ? >
n»
ft
">H
177
a
4#F
r-n
P»
n
9-»
l«T
ra
V
M
ft
IS
y-ii
|1T
C Fi
T"
tt
'tr'
>f »
fro
n
jfv
z> q
TTf
«*
NH» StlfPA&fc
JEJrraJ
fm
i-*-1
i»
5-3- 1
0
1-J-J
M
I-S-*
Ct
S-3-5
9
-------
<
C":
r
j",
cr.
dut ua* xm
'(ppa-SO
AMJtY«?fcr wrwt »**
10—
a too
CMW )
FUMBAfC TOT
•ijL. lL i
i-; ¦ - I -14:
L'_ u^li _,j Iff <'
sc*n
5,0 TO-*)
EP^ project
Jg&m LPCr CWJtT m
t»-ZZ->0O~C-3l- •BO)
HHJ
SIWA4C
*>*t KdwO.
ttc-e
DATE
P7H
I3HE
'* I***
9-n
i*T
«
?1
V?
n
H V
0*
M-
)*»
P »
n
i«
T-*1
Trfl
It
¦F14
1H
44
4 U
V7
l»T
l •
J2.
IT
1»
)»
1*
f19
4>
JM
y jj
|6f
u
M
I"
rev
ft*
^0
1**
FlH
H
1®
0 19
fM
«#
]>•*».
We %«:»
(IMU) IStff-li)
*»
(Hj.nr
*~
;w.n)
li»
m-vi
%t '/»i
{•tl.U)
Vjo
V>\ Kl:f0
(304.UX U0S.1t>
1
. o
-------
¦' ' j-r '
<
"C
c.
r~
C
EgA-EBgJECT
CHART
s
MTLET MAlfVlrt* SfJTW Ml r*D»lMS>
PATE
Li*. J**
ro-jj*
LZTSZ7'.^
r-\T. -;r
J.-L JJ.JJ., <**
a.n£l}.tK.LS4_
IkOT -
ttD0~
S|
J ,J , ' ,lL ifcfl ^
| , .| '1 -f. 4 •sr;.1
£T< ~ 1
]> ^ILL L"_li- ^^4-1
Li-J L _J
Ultsiwt 5T
-TTfu^^jKljs '
iEiJ Li
mrf s'-oo V/ Vi ; ^ ; v% ^ ^ % ¦—# vlS:t)9-
Ofoua) (609.11, <13* U) tOi.U) (tmu) («*») (P8.W) (ttJ.?*) (W-U) (»»%» (324.1t) (|*8.1l)
-------
r "TFrnniT^-j
cj too >
CPA PTOJCCT
H*fiQO LD6 CH*J?T
(6-11- '80 ~ t-ZO
too
90 -
i
• •
;«*>-
1 Bl t>~
;«¦
»>
x
o
lw
r-T-pT
^ 1?
v r
4J
I -4-1
j
_t__ i
I«VIOk
* t—)
IX#
+ 0-*
ft*
i
poo
*
. » too ¦
x ;
P>
I WO —
tJ
1 t wo
1 t» \tOO -
looO
3
2
SCFM
. 2 COO
5 woo-
t
a 0,
(rr>j
avJ'
;W-U
•V*W-*" V*"W
Mtc Mi tjo VI t
(Meun) (812.121
Mi VH
c«Mti ttiem im.u)
Vi* Vn y?® y(i Vio
(*••!»). ftlMll (ItMt.U) ('•~«¦!*) fim.Uj (CJ8.TJ)
r+®
9
M
V
§
*.•
<5
-*•
0*A)
PATC
HHt
itimH
VMT
TIrtE
rm
Tirtt
T
nut
4-ii
1 JT
« 01
l"
¦>
tfr4
4ft
)"
» 71
i-ll
1 **
li.n
«e>
41
r-^i
lU
d (6
j"
/rn
1 79
tP#
#
n
«=
fc-i»
i"
i"
9ifr
»•*
» n
T*>
4-i*
i"
*: *>
nc_
1»*
~«0
!»'
u oi
)¦
-
-
fc-l»
1"
i« -H
1"
¦ftf
3"
i*,y\
V*
40
t-A
1"
Mh
r?,»
W
n
t —
*u»
»*•
0*4
t*0
i-n
1"
l*>
ik:n
n>
t*
W
~f
J"
33 «-!
41
(-9
I*1
i~
L
«•
«
L-
c3
t-
49
1<*
fio
4|
)t»
9 9)
np
S
i-t
I<1
*•
1*0
V
y»
14 W
Vf
J"
0119
l-a
J'L
*¦"
_*L
m
P*
i^»l
HO
4 O
«9
-------
It - I
J i
EPA PROJECT
NO M
S " Jfltfg-tJfcJattRT CD
U-ii--fo~ t-jo - fo)
ftTt JSrtlT. BU*H_a 3-AM.lR.
BoT , ftHWfc fT> KWBIH&.
-
**rtAL **rr oo*H
(%)
X
I
-3
-hi
¦ il l llitl ,.L »1«—> 4-
. too-
it *«
a *
I I
\-W-
-V*
-U — t«-s.
^ . -
ri>
i
_ i
"1 TT
¦TfWlMT .VkTT BttlH'.
Huusr at HrnTrn mntu .
. IMHAlfD MttAtMltiT TMffCTIOM
I I ! I
.iJLJJ
LJ ' 1 ^Xi.
^'u
.gj 1. i
*,1
:Ti»
( H«
i ; i ,i r ,¦ • 1 ! 1 1 ¦
i 1— j--^* r- J-' * » j •* »- t
FCTIlf
*.*•» fc» V
dtrt.Ui 0'i*.t*"» CiiW.ir) ciiw-u)
««»
HlfTAM
>nt um
ITC.L*
PM1
TWt
l»»i
r*<*
•F
i»*
nt
•M
f
%
*•
+*
rH
}—
• •*
W?
M
l«T
III
It
*A*
1«»
Iff
u
1"
••it
Itt
fc»
1"
rt
fill
im
1"
n»
rt
1"
r«
M
rr
»•»
l*»
-
-
-
%»
l«
fi«f
•f
J*
_
-
_
1 *»
—
—
-
H*
1"
flf
M
«•
-
—
"
-
v.
l>T
_
—
-
,HT
f«<*
!!•
it
7*1
!»
-
S»
1 n
i:;»
«•
l**»
tIJI
—r
1 *P
-
fci
l*T
1H
•if
U.P
H
JW
n;H
r»
n
%>
1"
r»i
n»
r*>
I*
*P
p
M
»»
in.
JS.
tiu»«)
-------
T j
*n
B^TIAL «.>tur POH.M MID,
a£c-3ttixjp»ni.±£Mil£fU.
r "Ti~n r :'7tit.[T]4TP
I ! ' ! I 'I, ' I '
iiii
i"^wi
-
i »-f j
! i.l' tuotatjjj."-HJiaSt j , rt
KHi/nn.
FUN KITE
U3E» T
JJji
HMMIMl
gP-
jfg ito9
%crn
PTTrrTTrTT
-» . S
;„!,
» \~ \> - ;
1
-I 1
t'iil
J. I
! ]•
J I I .' ¦_•
- h --j-+-M
,1 | '
® I
e 1
N
Hi
«a» IMf
Mmit |crcu
DATl
rwt
mi [tme
K
%
in
b;»
•»
I"
J* v«
x«*»
$Cp
%
r»t
lr.ti
|«i*i
}ir
[tsrt
H
,1"
»"•
|^i*1
A®
. jcvitf
%
i*L
&i [K0
t**
pf4
)**
jy;w
%
oilJ
tan
»**¦>
|»*«
t4-f
91 \vv*
/«*
IttWt
j»
jo:44i
%
t»t f
7«r
•jftj&iwt.
j^.rt
. b ir
i»r
rJ
¦4Arfi.it
i»
krt
K»
_ b;w
Jt
r
. _ B*«ty
i»i
3 IHO
3
* •¦
:f
: i 1 i
1 i-1 c
! !
1 1 I '
yv/w -y\fK
i !
MTT V, 140 VI
CAevK*)
V»
"v*
M II wi
•Mi rt»
3.0 <«»-#>
inn.n; (u+3-on mil. 01? (ijji*1) (IM9-«1) (i>«.*i) (*u.tf)
-------
, • H :7 TTT
tPA PTOJgcr
¦~rrr
PUBUM wiTH
OUTLCt AMLYS*
, p-p-r
'W-^'
MtattaM «(1M 6tiC$VL5Tvj
I eur-er ro AttifSl*.
ygwr.eg
iWj/wa
f— j—» • - • -r : , I - -J - . , * i • , [
."" i . "I ! r ."! T" "1 j ""i"5' "~' r"" r : . i . -• :
4*«gw
f nai |,1V
jLlljl-L: J_ "i1! J_
£fa loeo-l
_l jjjj I ; !"TTTrhf?'jl
. /^Vjj^J^vWw 1 liJ ¦
Hit
|-fiT n^j"~[^TT5"ir7
i*w )
I.' i'Lii ' M.I I L! ii I I_1 I-I a. .J. I
L. I j i_
.•OSKCC
P"i»]
DATc >f» f.OO TU jjt '/w VI» rn J|» rn r'» r««"w
U4II.01) (IW»J1) (IW.C1) fiwiol) tw.ol) (rajl^l) H5TV»f) Wf Wj
SLirrMrte
trcie
9ATC
mi
TWt
•t
t*T
f .a
id
V,
«>»
¦r >l
7fl
<»»
OiW
tn
iM
r»r
T»
$L
«•»
t.pi
r»
JBJ
t«t
w
l»T
r.ft-
w
Ml
fH
>**
o.tl
7l<
*v
1"
r«
Tfl
ir^
W
I"
frSI
fll
V?
!¦»
?«
.421L
nu»
A
Oi
l»1
mil
n»
%
I«w
1*
}*»
-
»*T
(•I
mr
m
*1
il
if in
m
s«
%
| *T
<1
ra!i»
a=?
Pi*Ut
JftJ
Wt 11
*i
1"
r;o»
n
?W»
»*»
H'SilB
\4o
l'T
C:u
l»
r»
i»:»
m
)W
-------
7" | i hit*
i - ¦ !
j
5
? s
ZOO
9
»C3
*>-
ta-
Se
4&
_Li_L
pro exert «i»t4
JbMAVfSfcK.1
•i—r
-h-
rrfy^
n7j j. i_.j..
ma
" 8
Nȣ
I ti
Et* PBOJgCT
H*600 IRS- CHART j
cn-zi- to-v i-jo- •
nu>Ken with >wn.e
nut HMttHV i
[ -j-r !
_L 1 J ¦ • ' - " ! L
-i
I
-I
- - r -t - ;— i , -' -. i -*¦ -
I i
A*
| r
i
i L ':i! Lj_
I-
j ! !. I j
I
I
j.
7 T
¦ i i
tob/wn,
1
< — )
it.
kc
Nil Cb4i
'fcw'k
1 . . • I i j_L !
Mil i I
Mm *4
JjtJjJL'LJj-Lj
(WW )
>^^»vC3Cr<
I I
-J u L
cw« H|#:»0 Hi H> V* %» V* v*
D«w*V) .Ciwoi) (U15.01) (ittv»l) 0TZJ.C") (PHI.^) Iirii-fl) CIMfol) (inioi) trni.f*) em.K) cib^.j*)
•wI 1 '¦».>* »' (WMJ
"Ssxsl
WH*
UlfTAfr*
ioor i
cvu,e
PATE
TWC
fTH
TW
i
fct
I'T
lilt
1 M
*
iwri
}tp
r
Mz
Tfc
|»T
41
l.ff
l*T '
Tf
JLft
u*
i
l;
ft
irifv
J
T
>«
t;l
Hv
t t
?:ft
fl
***
IN*
l4f
)n ,
*-f?
If
,l1lt
ft
***
n;r*
V
y»
IK-
7o»
ti^L
rsn
Ltt
if
Jw
fi*i
»»
l»T
liF
ff
V"
I7;ri
n
t**
t*
*»
1
jlT
l*L
rat
Wfc s
Ml
,TB
-------
CI
CPA—EflgJKT..; ..,.1
' 1 ¦' I M -
J,.|
rt—\r
i 80
s ®
£«i i°
3 f (a
IS>
K
I
10 W'A%!
i I , ! i t i . v i ' i • , ' r
-_LLLL
i
! ¦*' i-
I
!4 ! -
1- ! | '
^ , vWA 1
:j ¦¦¦Hi:'. J ^ 1
ULili
-
I HH I
•n,-
vuiPr^f
Imt aua
tnue
P*TE
rwt
rm
n*
T
f n
if
1"
KJt
ti.il
ji«_
JK»
•a
1"
«:*
W*
i"»
mn
TV»
?u>
ma
22L
i
IOO0
two
TTM.
VWV
! | | 1 i
rr~r
!•' I
W f*g!Art , I I ; |
W*" L.> '' ii j—-<—-wv^~
%! t-BO */,
-M
•
-*•
g?
- %o
- r
o ~
J-O
to-X)
l*w*0 imo-w m*-*) (H6I.W cncs-^) (Mm-?*) rzm-»4) («*«.*) (m^) (zi»o3<»)
-------
i-il+Ji...
^nwUi
MMXJm&BL."
\nJu vUi»i <
*i
ii
Ml.
!' CPA , PROJKT
" U* too
Iw.> (f-10 ''80" »-l
-j
n
C
EC
to
,:!"ll • i"t"'!' I i ! f r I r I I ¦ | 1~T~ r~i7m~X>vi pn
—i-4-L..-Ll i!,! _ t_ . I.. U\! L J u_' i.:
J il!T ! I i i . s » i i li I • i :I i : sii< l L .
i 1 ^ ; L
-4 |
- -4-i
. ; i i
—L. l I* "*'
t"T sagsi*'"T~i J'T: T7i !""¦* I. i -j-i.j-U
rriffTl-l Wi—i »I J -!-4--rHhtt;»
I
L- i -
-•w,
¦JjI.
5
1
n'
:i.
, ,r,...., .....
T ' ¦ ' " ' "I • 1 ' ' T " 1 -r . ••
:. "r"m
ito ¦
....1 I,.
i r"
! :; i ; -Tj —
~*>-
___r^
'"I
¦ — j
L. 1
"1
i i 1 * 1" j r,! '
i.i •
Q'li
3rwjrT
(rsf )
.rw. II TJ:
I i'¦ '-a
, ...J ' l_J
ui. i.iiI •! I i.
j, a i***/
i*re .. Viqv-oo k, v,L «s "*i v,» Vi .
Mmu) tiuor+i cinf-il) izumi) MUWi) 1 flffo.il). CWMH) rzjw.ii) cuw-il)
^rua
tftL*
OATt
TV<«
mi
i
TW
•r
B
p"
JLA
«•
J
r*
w
J
f
*£2,
Mr
J
*_
i"
t»r
w
J
z>»
j
rt.it
J
P*»
j
%.
i«
, n
3d
tut
3
V
at j
iftif |
nt
_a
t*9
I
h»
j
%
i"
441
lj£
w
j
I-W
nil
'it
j
»*»
i£2L
T»
j
*L
i •?
I
ZifL
*•
jt
(U
ji
1
.W®
JB
fc
En
3d
fit I
_a
li5L
lifl-
ttf
_*
i
*»
%,
;*T
I l*r>J
«•
J!
JSS
JU
lit*
iSRiri
Zj
_
%
jrr
"n
CULJ
iEa
31
l»:w
3
,
25j
I'iP
%r
lMJ
ISJg
SI
fUl*1
3
it
i5U
fur
JK
%
i2L
zm;
HZ
XI
JM
M
[0LIM
-------
fp
n
uww
^TLrU/i
i r: 11
i r
i r
. W»>«« *» I .
0*.tWfr NATfct Tfr*iM3t> fl»
( MW >
g WI
*» tffliwt mrnfoH.
i i Mi
v T
tra..<3)
IIM.I1)
fPA rttOJFL T
*T
«»t*> lAft fittET
(»-!#-»o~
\
K*T MM
cYcia
P*TC
Tf«
tue
ft
i»T
~rfl
1
ft
1*'
nn
n
Tifl
wo
t
l*»
ttA
1
*1
I"
°>m
».rt
mn
W
1
;«»
Ir*"
TP
Tt*
*
%1
1"
\
rjL
¦mf
*7
h*
1
I"
f"
It?
1
%
1"
V
1
T*»
rn
n*
1
l»
r«?
*p
j
%
l»T
tT>
*r
1
111
JO
fiJZ
JEL
ttl
£L«u
e.n
J
t
IT
•: it
*Tfc
1
J1*
riFp
t*
*T»
J«»
r;i
TF*
1
fc
1"
/"~
-
•
jsp
—
•
.
1"
—
—
lK®
IV,fit
ive
1
p;»
TP»
1
I'T
rsc
T?»
i
in
rr|s
rp
1
3"
v n
w
1
%
1"
f?o
*•
'•P
1"»
00
1
w
W
1
f»TE ,
(HPuO) (U1011)
cmi.ii)
M+UK) .rt)
V>« «:?•
i
-------
? " tco
5?
H« 400
lpf-l
m-
o
loo
-TT,
i! iK 1*7
r T n r
. . - •
• joaya. BBurttrcz ! , J
L
m
11
Nik
X
EPA rRgJtCT.
w»fcna in& cmpy
>¦ * ^W^—b 'v^ir+nr «'J *s<»
. ' ' ' — ¦»"¦* Hcau»f
t«-»- *®-«. »- "is;
S K
* Ul
§*
a
¦
<*>
SD-j
to
40
l ; i j.
f?** "" VvJ W* A^.NTj^Yv^V
-f
T
v^-
ire wrn «vjtfM3$p
-------
n
"i
[ < WO
Tifwire f>{» rm
* *
ftl TFSL
.fg&PH WITH IftW TTHPPRATUftB
gil f»
X.
EPA PSOTFCT
N*£ff> ICft OW?T _
(vs- *00 -w Jf<
. i :,J.' l :». t ; 1 MLLj-J. — •
-r [~1
i
!¦, Is. !¦
'. ! 1 = | '
St
3
Si
in*)
2000
—r
lOOO
0
PATE 1A « se
. _C3H.lf)
-fe—
(WW)
L™ **V
~*/W %t
tunw (»!«) <««.»*)
***
\t>
tti r?o
(Ml\t+1
v,t V*
»s
mmnse
SOOT HACR
OClA.
DATE
TOE
rm
T*E
*F
PH*
%
C
C1C7
u
•n^i
ip
I«e
^MI
jp
IT*?
u?
n
}»
m
H>
W
w
%
%
Is'
y?\
41P
V
i®
JO
in
AU>
Jk*_
HVA.
n.sn
i.n
J&L
t&Q
IfrC
JSL
_m.
<9
*>'R
W
l?.*»j MO
?p
fft
JP
nuffijswj
X
in
s;p»
?,"¦
"iVT
50
tf-iS
MC 44
%
|iT
1.51, MS
44
'• M
(¦it
H
SP5
rr.ft
W
%
l>T
r.H
to
12
,«>
tfl
3»
Mff
ff
%
1"
1.4*
iff
99
*"
kpf
w
S3
JO
rr*
%
J-?I_
1*1
iixa
JS.
¦»
ff.M
*10
(Vi
ft
J"
0 K
W
l^L
n
*
(>T
("ii*
¦n
VII
j CP
J-~»
5_*
• ill
osl.
Hi
L«-
H
%
lit
ttrr?
Mfl
tff
ta
1A
11^
ji»
__
3/fi
,4fii
ti,
Vpj
.. i L_v
-------
<
-o
C!
c:
."W-riT
tv-jf fr»m
HAiMTrKwi way mtw.i
w
HOBtEH tin
Lftfrf TfHUTBJTKIW
atxpio-rcy muT dbuk-
th» vmm Ana tnTtny vn urn
-I- I i j- u 1 »a r-rQ TW HMTT HHntl!
: ~ ~ I mm inrutm..
, V . . t
IbOO-r
i
(jo«i n) (jeu»)
BATE .
(Itca3»)
u
l-W
•
*9
<¦»
ar
9.0
>•»
(vaff)
.1
I r
.EES BPJECt
J
MV-ca LOflr OnBT tlfr
d-fs- 'bo — i-n- *»oj
JUF7AM
MOT atMtl'
crci«
OATE
T«*
fW
twe
•*
*1
l»T
jap
ia
r«
jn
~*r
-A,
f*
S"»
o-tc
fit"
1*
I"
VK
(ttff
tf
ln
J*»
_
- '
S.
1ST
_
•
-* 1
r?
•
-
$£2
i"
•
;o
-
3m
—
->
l»t
_
- ,
-
—
W
-
—
in
—
*°®
-
—
—
-
—
—
l*T
Vn
ilf
M
a;h
41*
f9
#*>
If
P?3
19 P
H
*»
l»T
T*
!7iFf
r?
)*9
flift
ten.
(J120.W) (ji<*.w) (Jiw.fo)
-------
z .Mfc
V" [ft
1*
4CO
>
I0£>-
0
too
T"11 r
-tr
—M»%4 IL,
I
! * | ! 1
1 I
"•Wl
§T
lo-
ts
*>
£0
I
W
i
'f ' 'I ¦;
¦\~T
i 1
i
¦t r
- I
: r (• -
i :
'vwu
,iL:.
JJ-U
I
i 1 ,f
11.
P".!
S
K
3
eoo-
£00.
«o
l&oo-
'~T", !
111
•Jj L_
irnr\
f9MU*B gtSt. I
mt/NO.
I r
¦ ?;
miTOAU ' SMUT D9WN BtOWtC VP
tatM&e c*wjep m t*c «j jo«-e*.
rut nor rwwT »v*tbh mas •tirm*?
.JKftrabhv*3»t«.W£r. -
V) r»
St tZoO
scrn
~r
.J. i -J-..
**, ft
~i~
| r ¦ ¦ ll>
j-
! V-u^i
¦ ¦+
TTT'"
4--}-- h-
. aw-T-
3 iewf
<5
55 T
.fain*}-;. o>tu.w>} tlniio) £Hik.$®) i»W-Ba)
.'»/*. .-.'"A .
I . 1
¦*xr-^
\---J
! . i i I ¦> )'
iLi Ilil. J L _i.
< Hvf
¦I #
¦ 1
¦ IW-D1 .WITH
* *
u n fir
* "
/If
w ^
;«^u>
I 1
! I
*vT
» '
~w
*5?
uoi»; ti»i uoj
II
XR.*c
a i
»
9.0
|
3H
L S.O
b —
i 1.4
(MM.*)
f.oo
,T.
fep* MmjgCT
N»^n tir. cmjrffl
(i-«~ *o -v io-*- "#f;
it, IWr urn
Kmvt jtitwc"
DATE
Tine
fHijn*
*F
Hi
l»T
'Ml
|i n
H ,
it >i
>m
ito
4»
J£L(
an
k»
Irrt
jp-
jf-
*9f
liT
}Xt
J"
n.if
nil
— it H
- tf.ll
o b».»
JUL
t?P
•H
1"
i?-f?
« Itia
it®
J"
|n:R
ftf
H
1"
h.n
fcK
M
'*
1"
I'M
rfl-f
t"
Ib.«
(< u
4t7
~o
%
1"
r»
MP
41
lm
!» n
i»
IbiA
%
i."
**»
ltutlttl.
\il S% 1*9
;««
—
*j MO
1-2
%
i»T
Tfl
U0
t?
j»
rrt
e*»
lz?:«| i*o
40
%
t1T
J.tiUs?
I**
K.W
JU5.
fid
h^j-
?o
flJ.W
%
1"
r ijJ ifp
ff
*<•
fruTirr
Hr
*•*
Ji/pj Hf.
P,
-------
-1
l;
C.
a
r~,
• ~ - I
A
* u WO-
* M>i
k. R ifiO
5,'
r;
"trr-Tpr**
Tf
f «! J
• (epdi
-J ~
sc
w
o . —
IQO -
& <
ljui i tL £
&
CI
Nt>*
s» ®
»»^(ipni
S 3 EPA PROTECT
Itw *
N*600 una cwaT @
CIO-"* - '80 ~ lo-i8-'ao)
'"NvH^.|-
s*
0w
w
I
J—
CO
l — —I ~ |f," TT-|1 ;
,-jrU* iM*i" 1 .
it. j —ii UL
i - - [ - r I I *~»
, U.-Ailn
11 i .. i-.M ( 1
Kzmu» .MuaajmME.u
L»f llAfrQ mi rxiMfP .. TK. » » »^«f
TT^ ««TTH *1* Cr»IATPt> WtTH UaT *l«.
fcl^lWf. IT-» KCPaiaiHtf.
Uilju
/•'W
**
ii
¦rao* aua^»j
il;.
JJj lL I
no
(HMO
1*00—
i;
-rr-r-r--,- -TST-gri, -|; '!!¦ !¦ if
I I I , -1 J.j'
Lu
a*
¦=* I0»
I" '
in imi T"
SCFrt
1
jeaxbazjott.
i—nil" r'i ,
1 . j idu.gj.JL tli-:Lfr £
t-I A
-H.S
£
TiliW
K^»
aipotite
iio» 2&13JT"
trite
PATE
r>«s
nk
Tt«
•P
m«r
i«
_
t V
fK
'Mf
rT?-
>30
—
(f v'
70
IS
w.
!»t
il.w
~P>,
.K2_
f o
_SS_
i^L
»R»
—
%
I*7
<*7
J"
izr.Mjifcf
<9fi
l«T
• :a>w
«g
«-
fS « «P
'i,
?OP
u
ii«
t.irrci
J3.
JL
Wt
%
1"
ff
auo.
BL&
I-f>
-JXJ
Lf
%.
1 *"*
SI
r«jMo
tf
H'J
7T
OialiUL
BUtiHE.
' n!«M
JB.
W
%
l»T
p-lt
*•"
JM
l» jH |*s
no
%
l»f
rttis
•»
• :t»
_tte.
If*4
_«_
W
I"
i
n:n
a«» in«
¦U'WlfJP
p>
**
!»t
ir-1*
uus
JIj
«;*3
a:il
•.f®
HA.
4?"
W
J6_
W
irV
Sl»
ftt?
em
/ie
-23.
.
a-.*iS
*_ 1.. .4 ...
55 0-
<"«Q -j
OATI '?A S-oo icy{0
(noun) (Jia^.w) U3U.OOJ
rV>
"•vfi ^5
(umw iuts.it)
~W~
msi-iBf (jMi.n)
-7^T
ttl ¦
W»*:ro"
M, 0®LS)
-------
-I
<
"3
c
c
FWOJECT 1
tftso Iter CHART (ft
*90 -v> 10. f»« £0
Mint
pwJT HBi AJEUflL'^.
•^ITipH'r
£ a
?«
«>
to -
9-
to-
<— >
S3
i
|«d|MV
I J&SLZilV
i',
¦—-( -M— -(-•
4ir««-
I .1,1 I.' 1
2M0-
|fT*>
PATS
CMUKS)
iy^
(3«3I.*F)
'VI I. W/2»
ijwi i»)
<»W «»¦
in? '
(*»i.i»; («i»-v) (*«•¦)
»yrt *«o
(siii-i*;
<•*3
Sa»T WXUVK
CTCLC
OATe
rwe
it*
<*
!,t
n:g
V
Tfl
*1
1*
¦;«?
U*7
«¦
gv
r f
It
HI
tt
'fe
l»"
|j.iy
Y;
1® jy
£79
90
c«
4*7
fO
tt/P
HTTP
(TO
K
i»r
fit
<9
%
ft"®
U»i
fi
)Kt>
|»»1
n
l»r
j».i#
|W
*5
'*1
rj»
|t>
fa
|u.h
(<•
u
'*»
llT ,
!1M
&®°
t»
|r.u
ffe
J«S>
Wt
ft*
M
%
11j
&
l*T
IX*
tVa.
.35:1*
*»lr*
l«.»9
~4#
i9S-
«a
.*]_
%
i<»
UK
H®
HI
»
JW>
'dm
u
%
l*T
lid
(f*
p
JO
(ft*
K7
-------
n ioc
tflTH
¦]l.,":"Tiji"t*T5'TP 'f ffl ¦
,7]-®
_ I
'l
aa yo
f" >
* -HO
r i.iuL ff
5 1 cp* PROJECT
N(>> V
¦ 2 N*^o h yt rmwr @
(WWJ8 - 11-1- 'go)1
- 1
¦^TTT
1 tl M
, 1 , , J" '"i" "t it
1 I 1 Ij:):i i il
i u m^No«
1 TS
I tjW(
lr ll'h N
5.
H t I » t _ !i 1 ' it _ » * 1
^ LljIJ
liinT
iw P V
n t»6 —
kVwJYW*s*ft
t> IZOO
5" .»* 1 new mft Tttl
fl,' 11 ' I' il iil |1
vVVw-
m»)
to
fJSH.t*)
<*I»VJ
core iwis 00
(ttouK*) (3UJ.1F)
)
(JUMf)
~Vl
(J»rt.1»J
«•
xirr»*e
K»T kWtK
it
i f>.
" 1?
**
*»
*>
?*»
a/w
¦ni
, 4T
i;»>
I"
*,XP
lI*L
4JB
tf?
w
1f
,,a'
U-^f
%
I'T
f«
1""
"Slws
U «C
t«
?"
>n
I4T
til |(«0
w
!< Hi (SD
p
21.1S. tto
*0
%
ill
1,a
UW»J
i»
JSL-
0
2"P
r>,t1
r"
m.ri'.H*
%
OiT
i>*
ifv
JL«
SI
iJ3
iaa
»!'
U9L.
4£_
jfd
¦fc
1"
7 XI
ULW* (1
lt,rt
S*t>
U:»
fc*>
*
lJT
T,»}| 1T9
'1*
l"
iVEjftB
JW_
4-Y
>t»
r.ffjisa
&
B-JL
JA
_i».
JJllW.
ff-tfT as_
Lkl
JH-
IflClf)
-------
"TT
- t- i
EPA PROJECT
4
01
i'
(S>
X
i
to
• «
:*
r
f«* )
/V-vf1-*.
-I
Tvi
rn [
. ".q-j—y w»T-yrtr« «-r T - ^ - t — - -j •< r - • - " -1~-> • -t -I
J ; . j i i !
I'll!, . i
" ' ; '« '¦
! ". H
i i ,
iWTMi. jmx M«
|KW(# NMUM
*TT»L
ww.« uno re»r«
vr
•>. i i. • ¦ .
- r-r--'— r "¦; mi 11
- r ----t-y
Uo
i '
LL
J.U Jr
. L J.J.i L: ....
' J
n
/ffrpTr1
' TTTr~TTTT
I I . I ! .!
n
frmnm
''^e l*?n ^/n ll^ia
>.0 C*W
UW»
sumht
MOT •UMCR
t*tU!
PATH
rm
IT*
T*tt
T
HU,
%
|>T
If'iT
J*
"n
nil?
V|
k II
fcr*
-»_
0 ?
ti
'*
«¦-»?
Hf
w
It IP
Wl
iH_
w
!>»
0~
111
69.
ShU
rit
*y>
0JL
-W-
-Ml.
H
~«
X
icintn
3«*
13_U
mti
0
CM
>'-H
X
%
I1!
Bii*
V
J<¥
ti;ii
li
—
3«»
l»T
It
2JS.
.ML
-M-
!*
Pitt
n
-
—
—
JW
—
—
—
%
l»T
ir.tc
u
-
—
l"»
n-tf
it
-
}«
—
%j
1"
—
—
—
7*
-
JV
—'""
~ I
—
%
1*7
—
— ¦
—
;mp
— ;
_
?*P
I.Vf
i**\
'&
%
lJT
1JB
iS
Vf
,i!»
,1*9
W
Lli.
ua
3
ti
rcassj
rarrl
FUFTBT
JA
P'H
JUL!
LM.
twre "Vs • » "A ''/« "Vn "Tii Itf5 "/m
(•~out*) IHtyit) (M«1.ls) t *»n.<*09-1» (t.ul.lf) (+*•;.») (~»!¦*
vii W**i
(**.n> (~!¦#. n>
-------
M) i Em PROJECT
»-
H'&X? UXt CHART
(li-if — To -v II-Z1- '
!, i : ;
jBLBIMA£Llk.
i^TOLl^rtf imi.
ipr»'
m^TiAL ,.>it ro^
cctiZLtiSi^M-cezii
pfHj
*Uf?Alj£
n&ts n •pt-iw*.
Uia^anr.,
C2£^£jc_k^JL>£J5'i&J^i,
C2B2aJSflLli
CYCLB
X** .JKA
>5^ jan _«_ latt i
jCLUraJqBUfiflfc
S£suoan!3-aai. I
JiUitfaUfc-
I**)
mui
tfilS
y^rH.v.'>»* -.Ah>;W i"
¦ I ,- I !¦ , I 4 I tnv3ti nViii ¦
LLii
™ m"» i
~» T3«f
i-lft
A^Jfca JS
IfiU
UL JJ
rni-
tit **
p.na
11- gre^A-yjir.
4- — -4- - •
VwiPtni*
IUXM-
u'
5° I0>0
JtT
Tt»m
^sr1 "A*
C*3t "/|0 S oo "/li "/» Til"
(Hcu(?i) Itl5g.l1) MIS/. 11) (WH) H)
<4flM.<1)_. JVUM
-------
p
i EPA PROJECT
N'fiPO ICfa CMRT
(n-ie- "«o^ ii-1- 'fa
|-^r Xw
HHI
*>irr«Me
crcte
KH>WO.
niA.ii!
LlLL
to u» «nn ai»
"! "i .v:> t" wfvuri 1
¦c
s:
siaju-£n. (1JI1.S4) rtrtS.W) IH«W) MUM*
-------
n
a
r,
i—1
"\
wi «mii i-wr nr-a
tt
! i
! I
I.
irtrr
*vwi/
\ "| i>*«tld*
f
n i
* 3
~«A CMlEtl.
*'wg N*tfQ Iflfr CHART fl
i*« ° (ii- f- »o ~ u- n- '»<
f*
(X)
TT ¦" ¦; HT| ":' ";
,fen m.
. *mnn
\ — ¦ tVv|y-"
. - (-
rt|Ctcw. ia* J®i
•. ' >>/***/, >v;
•T K I- • 1 1 ! 1 ' - I - ! - ! - I 1 - • i J 4 *-¦-%*
-a
ffH»y|«i
(-
)
¦m»»gnrv wjt ip»
HM> ^JBfTtaT •%-**
-Ji
jryrn rwir una muftiion* cnn
f- <
4 j
>v
1 :
rJ
! i
n.v-H^JW*
t i •
A
PSiR
Mfi'
U.iJ..<- !«, ,1 J»«l». iAi
£M>
B
tew
o
~vj—vVvn-*>*j>
rtt^T= Vt : nh #, tiS TB^ < aft," % <^-
»T«. TRTw—& nPr. &, :—*Z S5 ft) «*S afc——% WSS*
o«-«) iwiw (wi».»+) ttwvn) (WJ-ft) cwii-M) c**5*-s4j. tnsvH)
j®
J«J1
niP
-1£_
tf
UI
n,n
JH
jUA
*
in
jar
JIB
sir?
t*
«£
H-d
J®
%
1"
1.11
w
s»
iJrW
AAA
**?0
X
jlTL
I"
1M
f»'-T
J2-.
Ii
UL
F'f
A*
*w
l?T1
n;
%
i*T_
s?s
JA_
,U«
»(?»
r^i?
X&L
Alt.
n#
%
%
|H
7™
ui«
rt>*J
_U_
}l
jtaq
IS p
JC
I"1
i£J
7°
|) 0»
-------
h
i i
¦M-
•
IB:
try
i
""! ffi"
,-|-fe"H-
1 i
.. 1
. ...
. *
J
• • J > a
ir
i
-ul..-
, 1
¦Tir-
ij
l.l.
II
"iiiTTV;-
n I _ Et*A pRPircr.
tcti. K
? 5
twtrmz J>~vr towmi
acsur* jm nrr.rtMrm „¦»
'iill!;i:i
iJJ
c/clp
A-^v'Tt
- (—h
Mb/fWX
»Uy»^
-•ek. i¦; —7—va4
' .Li ¦ i'L.tUii'i-J..;,l^ Li libiU.
IffH 1
""itttr:
USZ^LSS^ L 1
&S t».»4.
TZ%5 'vil
1 ''UvjU J_>vA;^U
'Vi8 leo *3?!i
."r1"- . --r»o * w -ni ¦*/!©
Wo«t*) (*1Jl-.ff*) WtiH) IVK'Ml)
_ " 1^' *
C+0»Q9) (+*1+9)
Wi3.u;
-------
<
TTTTTTT;' ^fJPT77!
T-:- 'i : - -i - r -i~t
£ «.
X II
too-
§ *
V»
r-
>PV>
Trrrm
L Jj.iiJ I
j-LLLU
1 an »,a
mm; rnfiTTiw»i
ftiiiiiit
ittSS
.!±l±i.1L!.lJL1j
CT»T.U>f7PT »»-al
BUUHUII
i* -iP
KTFTT3
Li ' ! : "' ' ¦ ' (- | J; j 1 i I ( {
I • ,: r_ ¦ .j
1,1 1 1—- **o "<&', ' ¦*>, -'•'¦•—L
n
y j"
i L;.
pj
(fr«>
«.« im*)
ERAt PRfllECi.
m»*b ifts, QiAtrr <£)
" - iff
»«-»-'*> ~ i- « - 'ei >
\
U»
hmm
Mrr Mmii
(VlK
DATE
TtK
r?«
Tn*
*
nn
%
%
i«r
r*?
HI
>¦»
rff
fir
ti
15L
)»T
yf
j>9
.Uir
m.
«•*
%
| *T
un
ttt
*¦»
?D
irn
Ml
4*
%
l»T
pi*
If
r«s
M
~
•
X
Pi 10
m
H
_
«
*
t *T
jwt
J*
l»t
Jl*
>*»
*
irk
2*»
)W
%
| *T
ir.ifl
tf
t.n
1 m
fTfi
5»
irn
3 (fiULH-;
-------
c
n*
r*
m
*j> m rf
_MT^-H-"i
-iAUy^nru.: -/^~-i
' >vj'-\; ¦ \
h"" M
Ptactn ros-iim.
Kiuooeu iaffi
ritur.a n n
Ja.wm tr rerttaaj
j
t ow*«rm|.|
1 " lulv i
h mktU-
T
j ^'Llill I :r'L 1 -.It l^.tlra.Nn':
weswuf* »arg"-T
3St
;ig nrr*vi Mivruti
StCtT
t Kwwp fTir
'CjZZ,¦;
' I " ~ ""I t r 1 t ! ' --'-+! t t- i r I— j X* I. H
,<\ ' ' :-(•- ,|T!' ''flit 1 - J-f Muaz-fiai , "« I i I r" 1 «—*
rr—«-;—r -'t I ¦ - * r ~i 1 ! 1— -J-^-j -tubmimax m _ . -H-t ¦<—¦4-1 *® » *•*
JEEA PgftJgCf
J-1 UXr LMRT r?
tJO
«
J"
IL-J£
"11 (tf.ni
<»
w
l>i
I
jew
.
l«-*P i
-i2.
J3L
Sfi
taiRTSim
;WO
i -
_
3".
-
Xi
ia_
l«.o
252.
[».«
a»L
Li-is
r*
/»
1 »T
ln-lf
iS-
Li»
»1
t9
JW
L> tf
n
ft
'A
ILL
i»«
J3
W»J
)P.
pIp
M.«
tp
Li®.
M
jLU_
ZijlM
Ft,
;«P
a;v
[« *
n
K>
l».l1
41
l)» ,
i_LL
p;»J
1 i
If*
a»
. 1h;m
jL
^3
j*i ».»• Ya. . '/f
/.i
Ollt.01)
/it
(HJ1.fl)
Wj '/n-
(WMU. Oiat.U)
H» Vt6 '/U *'*3
f»um»j (Wi7.u;
(***)
-------
1
_ n
9 M'I
• 400
y. ¦ «e0
-I «
¦»' yiin_ II I
00 W*\f
Tjn "i 1 pnn~
M V"''
11
t 'i .41 til I .Ml). I
n-n-prr]
w
EPA hpject f
-y^?-
li/»i
MOttil
Miliaria
wte«si*t+.
— ¦ ¦«¦ — •' -
i
LL'JJQ1 ji
»r
Vat-,—
g
-
<»ml>J
WMV
tin *»•
m
aufffcM
*>ot idUl
true
pate
T>"«
rr«
iwt
IM
•Vi
• >t
f.»
n
J3
aii
rrM
Jl_
H
JO
*•
i"
f *f
M
*¦»
i»:«»
.«
3"
HI#
i.ij
Kl
1 »T
»
Jff
f*
if.Mj
ff
J"
mjufLm.
Jt
&
1"
f«»
u:CI
fi
i»
***] P#
IV®
M
p;it|
r;i
«A
*»
J?"
**«?
n
J*
Jfe
l"
• a
*1
r*
V
*•»
3*»
trts
wtf
4t
JO
Jfe
I'T
|W
to
1*9
ijw"
i/iff1
tJ
41
)£s
1*
ru
«
~?,
l"
ifiW
JR
n
3 rv
**W
N
v
Jk
\iT
r««
-
_
•?*»
-
—
*C («¦*)
L
-------
I ' 1
ti
O
¦ft
X
1*
( ppra
90
?G" 10
j. »—fWrt.-yr - • - - - - n A i-*q - [;
t- - - {J^i
/'V
'^/Aa^A
to
CO
«
*«
rsJI
«*¦
•5
»ti mr
4.
—v—-'
Ji.
20CJ
159®-
| > io»^
21 m~
1.
If ATE fg/^
r 1 4
ti;
SCFH
1 "»f
i •« u
id»o
a '
JL£T0j
CATV
(|UU*v)
& "-*7 *T
hi a w vn Yr\ y» ._
13W1.Z») (IMI.tt) (IMH) (3*1 J i4) (wn.Jf)
*r
!, n -jr^
r?-
epa project 1
M^ra LOG ck'AT
( 1-21- '•! ~ i- I- '«! )
rrrTTi,"1ii''r'iiiij,i
-u L) -III." L^L.
: 'i^}A
t_UijUilJ. til .il I L
i' p
'!!. iJ"
till. I.'. I,
• i-jL-' !"¦ «'
- - i»j- >r--hO mi
I ' f* %
^ I u. ' i ' - m- .. erg
- -+ irtt'""
.1 .L i I J . •<¦'¦ *'"> 1
PATC
Ya
KKl
juriwae
TPC
RV
ivr
TME *
&5L
rrpjjo.
£&.
diT jSr-^JK. £». » ft
jp]J.{P4ji?Jatfi H ,+t
& —i.
~1
|IL«l.M ^
- .It
z*?_.
.IB...
.1*1. II*1*.
»'Wi .~#.
jrwLwJ.*
j\.
*L
tt
UUZ Jll_ I5:Hj_«s 1»
15. Jiaa _u_ easUi. I «
Ittijtt. LP | M ! M
¦*•¦»! a? 1 <9
5X5235
—r*~p*T
~*r
,.! WW"
¦; i
-L.-
'
1
ill!;.:
M •
ar ^
«* |2
*
*0
-------
' ''* f
v 11 , 3
' [7^9 I'T* '-'- ^
•**=• ifi'iniiii
Ii»wr r
------- |