United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 2771
Research and Development
EPA-600/S7-82-057a Mar. 198
Project Summary
Hitachi Zosen NOX Flue Gas
Treatment Process:
Volume 1. Pilot Plant
Evaluation
Shingo Tanaka and Richard Wiener
The EPA/Hitachi Zosen pilot plant
project — evaluating selective cataly-
tic reduction (SCR) of NOx on a coal-
fired source — operated for a year and
a half. The objective of the contract, to
operate at 90% NOx removal over a
90-day period, was exceeded. TheO.5
MW pilot plant was installed on a
slipstream from Georgia Power Co.'s
Plant Mitchell near Albany, GA.
A newly developed catalyst, NOXNON
600, was successfully applied and
demonstrated 90% NOx removal
efficiency for over 5600 hours. Tests
to operate the catalyst outside of
design specifications showed good
tolerance to adverse situations.
The pilot plant 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 may be
usable as a NOx control option;
however, some technical concerns
remain before the technology can be
considered commercially available
and demonstrated for coal-fired sources
in the U.S.
This Project Summary was devel-
oped by EPA's Industrial Environmen-
tal Research Laboratory, Research
Triangle Park, NC, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
As part of the effort to assess
technology for control of nitrogen oxide
(NOx) emissions, EPA 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 gives
details of the Hitachi Zosen process,
designed to limit NOx 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% reduction in NOx
emissions.
Strict air pollution laws in Japan led to
the construction of several full-scale
systems for the removal of NOx from
flue gas. Hitachi Zosen took the lead in
this area with the construction of the
first large commercial unit in 1974.
Hitachi Zosen now has nine commercial
plants in operation.
To further the goal of controlling
stationary source NOx emissions, EPA
sought to enhance the reliability and
effectiveness of technology to reduce
these emissions. One aspect of EPA's
involvement includes sponsoring pro-
grams designed to demonstrate this
technology. Because combustion mod-
ifications can achieve only limited
-------
reduction in NOx emissions, some
emphasis has been placed on developing
FGT processes. In particular, SCR
technology appears to be a very promis-
ing method of reducing stationary
source NOx emissions by over 90%.
Consequently, the EPA has acted to
demonstrate some of the more advanced
SCR systems.
EPA initiated programs to demonstrate
two SCR processes on a 0.5 MW scale.
The processes are:
The Shell/UOP Simultaneous SOx/
NOx Removal System.
The Hitachi Zosen NOx Removal
System.
The EPA-sponsored programs demon-
strated 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
One method of removing NOx from
flue gas is using SCR. Ammonia is the
most practical reductant because it
reacts selectively and quantitatively
with NOx to produce innocuous nitro-
gen 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 tempera-
ture can be lowered to a more practical
range of 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 removal of NOx 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
600.
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 NOx in the flue gas to
form small quantities of nitrogen gas
and water vapor, both of which are
normal constituents of the atmosphere
and are environmentally acceptable.
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 NOx. This gas enters the
reactor, passes over the catalyst, and
then reactions proceed. The flue gas
then passes through the usual air
heater, paniculate control, SOz control,
fan, and stack. The ammonia for the
reaction is vaporized with 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, or
economizer bypass.
Chemistry
The exact relationships, between
ammonia and NOx are not completely
understood. However, certain reactions
are probably involved which may or may
not include oxygen. Virtually all of the
NOx in combustion gas is present as NO,
so the following equations are only for
NO reactions:
(1)
4NH3+6NO = 5Nz+6HzO (2)
If NOz is also present, the following
reactions may also occur:
4NH3 + 2NOz+02 = 3N2+ 6HzO{3)
8NH3 + 6NOz = 7Nz + 12HzO(4)
2NH3+NO + NO2 = 2N2 + 3HzO(5)
The reactions of ammonia with NOx
over the catalyst occur below 300°C.
Without a catalyst the reaction will only
occur in a narrow temperature range of
Temperature
r /VOK , Recorder/
Recorder \ Controller
Steam
or Air
Solenoid
Valve-
Vent"
Flow Indicator/ Solenoid
Temperature
Recorder
To S0» and Particulate
Control Systems
Ammonia
Boiler
Load Signal
Compressed
Air
Flow Indicator/
Controller
Figure 1. Typical commercial system flow diagram.
-------
950-1000°C. Below this temperature
the reaction rate is very low.
Factors Affecting the Catalytic
Reduction of NOx
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 NOx by ammonia.
The most critical variables which
affect the degree of NOx removal are the
mole ratio of ammonia to NOx, the flue
gas flow rate, and the reactor tempera-
ture.
Ammonia Emissions
A small amount of ammonia will
invariably pass through the reactor and
exit with the flue gas due to incomplete
reactions between the NOx and NH3.
Ammonia slippage (unreacted ammonia)
is of concern, and efforts are normally
made to control slippageto levels of 10-
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. Another
reason for the concern with ammonia
slippage is that flue gas desulfurization
following the denitrification system
could absorb any ammonia in the flue
gas and ammonium salts can build up.
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
NHs/NOx mole ratio, the area velocity,
and the temperature at which the
reaction occurs. For commercial appli-
cation 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%.
Ammonium Sulfate/Bisulfate
The presence of sulfur trioxide (SOs)
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% of the SOX in the flue gas are present
as SO3. For coal combustion, S03
accounts for approximately 1 % of the
total SOx. The reaction of ammonia with
this SO3 will not occur above about
300°C (572°F). At very low concentra-
tions, a temperature of around 200°C
(392°F) might be sufficient to avoid this
formation.
To avoid such deposits it is necessary
that the temperature 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 maintain-
ing a low ammonia level in the flue gas
exiting the reactor. The presence of fly
ash also reduces the problem both by
scouring the deposits from surfaces and
by providing surfaces on which the
deposits will form instead of on metallic
surfaces.
Application of Process to
Boilers
A schematic flow diagram of the
process as applied to a boiler is shown in
Figure 1. Flue gas leaving the economizer
at a temperature of about 390°C(734°F)
is first mixed with ammonia in quantities
needed to meet NOx removal require-
ments. 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 geometry. NOx is 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 then
diluted with either air or steam to aid in
the distribution of ammonia intothe flue
gas.
The reactor is supplied with retract-
able 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.
Catalyst Description
A particularly effective physical
design of the catalyst structure has
been developed by Hitachi Zosen. This
structure is honeycomb shaped as
shown in Figure 2. Overall, the catalyst
has four characteristics:
• The structure is a thin plate
honeycomb.
• Due to substantially reduced
pressure drop across the catalyst
layer, operating power costs are
much lower than with convention-
al catalysts.
A straight gas flow path prevents
dust clogging.
It is applicable for gases with high
SO2 concentrations.
Figure 2. Configuration of NOXNON
500 or 600 series catalyst.
As a result, in treating high-tempera-
ture gases with high SOx and dust
concentrations (such as coal-fired boi ler
flue gas), the NOx removal reactor can
be installed immediately downstream of
the economizer. In addition, dust
elimination or other pretreatment is
unnecessary and, with the low pressure
drop, operating costs are low.
The honeycomb catalysts are referred
to by Hitachi Zosen as NOXNON 500 or
NOXNON 600. These catalysts are
manufactured as plates and are fabri-
cated 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 catalyst plates are arranged in a
steel frame box supported by retainers.
A standard module is 1 meter long, 1
meter wide, and 0.5 meters deep. The
catalyst is activated after the corrugated
catalyst assembly is made. In the
NOXNON 500 catalyst, thin stainless
steel plates are used. The newer
NOXNON 600 catalyst, instead of
plates, uses a stainless steel wire mesh
as a base metal to give mechanical
strength upon which catalytic compo-
nents are cemented. The NOXNON 600
is considerably lighter in weight and
contains more active material for a
given volume of catalyst.
-------
The active components of the catalyst
consist of vanadium and titanium
compounds. Other components are
added to increase resistance to fly ash
abrasion.
The catalyst has an expected life of
about 2 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.
Reactor Design
Reactor vessels housing the catalyst
are carbon steel and are divided into
chambers each having its own fly ash
hopper (see Figure 3). Flue gas flows
downward in the reactor. Flue gas
usually enters at the top of the reactorat
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.
Demonstration Program
A potential market for the denitrifica-
tion of power plant flue gases may be at
coal-fired stations. Limited test work
has been done by Hitachi Zosen on flue
gas from coal combustion. The EPA-
sponsored demonstration at Georgia
Power Co.'s Plant Mitchell was an
excellent site for testing 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 param-
eters were evaluated, and their effects
on removal efficiencies and ammonia
utilization were checked. The data were
collected and evaluated to establish
valid characterizations of the process.
Long-term tests provided 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
Ammonia Feed
Connections
Gas Distribution
Grids
Catalyst
Soot Blower
Treated Flue Gas to
Boiler Preheater
Figure 3. Typical reactor arrangement.
Hoppers
capital cost requirements for commercial
installations.
Hitachi Zosen, with Chemico Air
Pollution Control Corp. (CAPCC), a
Division of Envirotech Corp., as their
major subcontractor, provided the test
plant and operated the unit. This
included detailed engineering, procure-
ment, fabrication, transportation, erec-
tion, test operation, and continuous
demonstration operation of the pilot
plant. The work was in four phases:
Phase I (engineering) started with the
basic design package which was de-
signed by Hitachi Zosen in Japan. Based
on these designs, CAPCC prepared a de-
tailed design.
Phase II included start-up, debugging,
and parametric tests.
Following the successful completion
of the system optimization test. Phase
III, the pilot plant was to be operated
continuously (24 hours/day, 7 days/
week) for at least 3 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.
Additional 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 perform-
ance to transient operating conditions.
Schedule
The period of performance for comple-
tion of the work related to the pilot plant
was originally 18 months. The award of
the contract was May 1978. The design
and specification period, Phase I, took
approximately 3 months. Phase II re-
quired approximately 9 months for
procurement, fabrication, transportation,
and erection. These schedules were as
originally expected. However, since start-
up and debugging required about 2
months, the first tests were not started
until August 1,1979. This first charge of
catalyst showed less than optimum
results after some 4 months of operation,
and it was decided to replace it irr
December 1979. After some 4 additional
months of operation, the NOXNON 500
catalyst again began to provide less than
expected results, and a decision was
made to install a third charge of catalyst.
The third charge was the new NOXNON
600. This was installed on April 18,
1980, and was used for some 9 months
until the plant was shut down on
February 2, 1981.
Description of the Test Plant
The following description of the pilot
unit includes summaries of major items
of equipment (see Figure 4):
4
-------
Flue
Gas Heater
Ammonia Supply
Tank
NH3
1.43 Ib/hr NH3
fat 10 MR)
Reactor
J \ Soot Blower
SOi
/VO./SO, Generator
Figure 4. Process flow diagram of pilot 'plant.
Soot Blower
Gas Heater
Fly Ash to
Vacuum Line
Flue gas to be used as input to the
pilot plant was drawn from the boiler
duct downstream of the economizer and
introduced to the reactor through a 14-
in. diameter* pipeline.
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.
Gaseous ammonia was injected to
the gas stream after 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 of gaseous nitrogen
and water, which remained in the gas
stream.
Dust settling in the reactor was kept
within acceptable bounds, using a soot
blowing apparatus with either steam or
air.
A blower, downstream of the reactor,
overcame the flue gas pressure drop
through the pilot plant. A cyclone dust
separator, ahead of the boiler, prevented
erosion of the blower by dust.
Since it was planned to test the
performance of NOx, SOz, and SOs, and
since only fixed concentrations of these
('(Readers more accustomed to metric units are
asked to use the conversion factors at the end of
this summary.
substances were available from the
Albany, GA, plant boiler operation, SOg
was purchased for addition to the flue
gas, and a NO,, SO2, and SO3generating
unit was provided to manufacture these
materials for addition to the flue gas.
This permitted testing at NOx, S02, and
80s levels higher than those available
from the boiler.
The operating conditions controlled in
the pilot unit are:
• Flue gas flow rate.
• Flue gas temperature.
• The amount of charged ammonia.
• Soot blowing gas pressure and
temperature.
Host Site
The pilot plant was at Unit 3 of
Georgia Power Co.'s Plant Mitchell,
Albany, GA. This unit has a pulverized-
coal-fired Combustion Engineering
boiler, initially operated on April 18,
1964, with a 125 MWnameplate rating.
The boiler has a rated steam capacity of
1,075,000 Ib/hr of steam to the turbine
at 1800 psig. The unit is tangentially
fired with CE standard tilting bu rners for
steam temperature control. The pilot
plant was located outside Unit 3.
Data Collection
The pilot plant operation was closely
monitored by an array of instruments
and analyzers. Continuous analyzers of
the latest design measured the NOx
levels both entering and leaving the
reactor.
Much of the data at the pilot plant
were collected in a data logger which
frequently scanned the instruments and
stored the data. Visual and paper tape
displays were provided.
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 in a full-scale plant. The
results could therefore be directly
extrapolated to a large plant assuming
that the velocity, mole ratio, and
temperature remained the same and the
gas distribution through the catalyst
bed was properly designed.
Ammonia consumption could also be
determined from the test results because
it is calculated as a mole ratio of
ammonia to NOX in the flue gas.
NOXNON 600 Tests
Introduction
Primarily because of clogging of the
catalyst channels by fly ash due to
narrow clearances, the initial and
second charges of catalyst did not
perform as expected. The installation
and testing of NOXNON 600 catalyst
having wider channels was proposed by
Hitachi Zosen and accepted by EPA.
NOXNON 600 is produced from thin
stainless steel wire mesh (as a base
metal to give mechanical strength) to
which catalytic components are ce-
mented. This technique has been
proven in applications on oil- and coal-
fired combustion flue gas in Japan.
Catalyst Performance
Starting from April 22, 1980, the
demonstration operation with NOXNON
600 ran for more than 9 months until
February 2, 1981. The operation with
combustion flue gas from the Unit 3
Boiler of Georgia Power Co.'s Plant
Mitchell was for 5620 hours. The
program was terminated due to sched-
uled maintenance of the power plant
requiring moving of the pilot plant.
The pilot plant program required
achieving NOx removal efficiency of
more than 90% continuously for more
than 3 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 catalyst life. From April 22,
1980, the pilot plant was operated
maintaining NOx removal efficiency of
more than 90% until the end of October.
After October, a nominal 80% NOx
removal was accepted to decrease
ammonia slip as far as possible.
Following the catalyst life test, further
testing was carried out to determine the
effects of transient conditions on the
catalyst and to provide an extended
operating time so that at least 5000
hours of operation could be obtained to
evaluate the long-term effectiveness of
the NOXNON 600.
At various times, tests were run to
determine catalyst efficiency.
During the operating time with flue
gas, tests were run including catalyst
life tests, catalyst performance tests,
and transient tests. Controllability and
reliability of the entire system were also
evaluated at the same time.
Operating conditions were varied for
the performance and transient tests.
Therefore, when the activity of the
catalyst was to be evaluated, 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 NH3/NOx mole ratio of 1.0. Under
these conditions, the NOx removal ef-
ficiency was measured over a period of
several hours to determine the condi-
tion of the catalyst:
%Removal
At the beginning
of the operation 90 - 94
At the end of 8/80(2500 hours) 90 - 94
11780 (4000 hours) 90 - 92
Before transient
tests 12/80 (4420 hours) 90 - 91.5
During transient
tests 1/5/81 (5OOO hours) 90 - 91
After regeneration
1/27/81 (5500 hours) 91-94.5
Operating Variables
Mole Ratio
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
by using the flue gas flow rate signal
and the inlet NOx concentration analysis
to determine the quantity of NOx in
the inlet stream. From the inlet NOx
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.
Figure 5 is a mole ratio curve which
shows that 80% removal requires a
mole ratio of about 0.85, while 90%
removal requires a mole ratio of 1.0.
Flue Gas Flow Rate
The NOXNON 600 catalyst for the
pilot plant was designed to operate at
1057 scfm, equivalent to an area
velocity (A.V.) of 9.6 Nm3/m2-hr.
However, operating at a much higher
flow rate of 1650 scfm (an A.V. of 15)
provided the desired 90% NOx 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 NOx would
be unchanged. As seen in Figure 6, the
flow rate had little or no effect on NOx
removal.
Influence of SOx Concentration
on NOx Removal Efficiency
Concentrations of SOx 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) without
affecting NOx removal efficiency. Opti-
mum operating temperatures from
fundamental experiments in the labora-
tory are between 572°F (300°C) and
750°F (400°C).
NOx Concentration
On August 24,1980, the NOxgenera-
tor was operated to increase the
100
80
I"
20
120
.1
40
0.2 0.4 0.6 0.8 1.0
HN3/NOi Mole Ratio
1.2
1.4
! 01100 scfm 700±10°F
* 1300 scfm 700±10°F
01500 scfm 710± 5°f
Figure 5. Effect of NHs/NO* ratio on /VO« removal efficiency and NH3 slippage.
-------
concentration of NOx at the inlet of the
reactor, and the influence of NOx
concentration was investigated.
Between 400 ppm and 900 ppm of
inlet NOx concentrations, the concen-
tration of NOx did not influence NOx
removal efficiency. This data is consis-
tent 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 slippage
when mole ratio was varied at operating
times of 3200 hours to 4500 hours.
General conclusions from these tests
include:'
• 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; e.g., at
0.6.
The reason for this higher than
expected slippage may have been
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 adversely affected the
apparent catalyst activity resulting in
relatively high ammonia slippage. Even
a slight reactivity loss, while producing
little loss in NOx removal efficiency, can
significantly increase ammonia emis-
sions.
Oxidation of SO2 to SO3
On July 23-25, 1980, operating
conditions were maintained constant so
that S03 measurements could be
obtained at the inlet and outlet of the
reactor.
The results indicated an average
oxidation rate of 1.8%.
Prior experimental results indicated
that, at these conditions, the ratio
should have been somewhat lower: 1.0
- 1.5%.
Transient Tests
After the NOx removal efficiency of
more than 90% was demonstrated in
the continuous run of 3 months as
required 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
100
90
u
.§
.S
2
o
I
i
80\-
70
Reaction Temp.: Approx. 7OO°F
NHs/NO* Ratio: Approx. 1.0
Symbol Operating Period
fo 200
,• 2/50
1250hr
3750 hr
WOO 1100 1200 1300 1400 1500 1600 1700
Flue Gas Flow Rate, scfm
Figure 6. Effect of flue gas flow rate on /VO« removal efficiency.
Zosen's catalyst for commercial opera-
tions on coal-fired combustion flue gas.
Five transient tests were performed.
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 tempera-
ture in the reactor decreased to 600°F
and allow it to introduce ammonia into
the system when temperature in the
reactor increased and returned to
600°F. The purpose of this test was to
confirm the reliability of the trip system.
The results of this test proved that:
• Automatic shut-off and supply of am-
monia operated smoothly.
• NOx removal efficiency and pressure
drop at a temperature of 700°F was
constant and unchanged through the
three repeated tests.
Cold Start-up
A commercial boiler normally starts
operating after a long shutdown with
the reactor and ductwork filled with
ambient air When operation com-
mences, flue gas is introduced into the
system as the temperature rises and the
flue gas can 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
experience proved that sulfuric acid
mist does not deteriorate the catalyst,
and ammonium sulfate 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 affect the NOx removal
reaction.
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.
Twelve hours after commencing the
start-up, operating conditions became
stable at the same conditions before
cold start-up The results proved that
cold start-up does not influence catalytic
performance.
Boiler Shutdown and Start-up
From time to time a power plant boiler
shuts down and starts up, and an NOx
removal system must follow such
transitions.
The purpose of this test was to
confirm the controllability of the NOX
removal system during the shutdown
-------
and start-upof the boiler. When the host
boiler was shut down for maintenance,
this test was executed with the NOx
removal system shutting down and
starting up along with the boiler with no
purging of the reactor. The results
proved that the system could withstand
the transient shutdown and start-up of
the boiler.
Sudden Load Change
A power plant boiler may occasionally
change load suddenly, complying with
variations of power consumption. The
NOx removal" system should follow
these sudden boiler load changes. Two
levels of operating conditions were
adopted as representative load levels:
• High load: 1,300 scfm, 700°F, 0.8
mole ratio.
• Low load: 900 scfm, 610°F, 0.8
mole ratio.
The above conditions were altered
every 2 hours and continued for 24
hours. No serious problem seemed to be
caused by these sudden load changes.
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, to prevent clogging by fly ash.
This frequency seemed to be the mini-
mum to prevent clogging; however,
there was no experience in operating
the pilot plant without the sootblower.
Thus, a trial was made to operate with-
out the sootblower.
At a flow rate of 1,100 scfm, 700°F,
and 0.8 mole ratio, sootblower operation
was halted and the progress of pressure
drop increase was observed. Pressure
drop appeared unstable (increasing
gradually from 1.20 to 1.25 in. H20 to
1.35 to 1.40 in. H2O in approximately 28
hours) and seemed to continue to
increase. The sootblower was then
restarted. This test indicated that
operation of the sootblower was neces-
sary.
Changes in Pressure Drop in
Relation to A/Ox Removal
Efficiency
In the treatment of coal-fired combus-
tion flue gas, it is realized that adhesion
and clogging caused by fly ash, along
with the formation and deposition of
ammonium sulfate and bisulfate, cause
problems. In NOx removal systems,
these problems are first noticed as an
increase in pressure drop. Therefore,
changes in pressure drop were carefully
studied during the operation of the EPA
pilot plant.
Since operating conditions varied
from time totime 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.
Several experiments were carried
out. The tendency for pressure drop
changes is 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 oper-
ated three times a day. However, the
previous pressure drop was restored
after a few days of continuous
operation.
• When the flue gas flow rate was
changed, the pressure drop increased
at times. For example, in November
1980 after some 20 days of steady
operation the flow rate was changed
from 1300 to 1500 scfm and after 2
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's Unit 3
boiler operated the economizer
sootblower 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 surface while the boiler was
in operation. This abnormal mainte-
nance work caused a serious increase
in pressure drop.
• Operation of the sootblower seems to
be necessary for coal-fired combus-
tion flue gas. The pilot plant was
operated for 28 hours without the
sootblower, and the pressure drop
increased sharply. After the soot-
blower was returned to operation, it
took 3 days to restore the pressure
drop.
Regeneration of Catalyst
When the catalyst was washed with
warm water, NOx removal efficiency
was restored to its initial efficiency. The
pressure drop also seems to have been
partially restored. However, the catalyst
was partially clogged with asbestos
fibers, which is not typical. The effec-
tiveness of water washing for pressure
drop restoration, therefore, is u ncerta i n.
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 impact of the flue gas flow. There
was no abrasion in the other blocks. A
commercial system would incorporate a
tighter catalyst structure to avoid such
vibration.
Clogging
Anticipated clogging of the catalyst by
fly ash was prevented by operating 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 30-35% 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. This asbestos
would not be used in a commercial
plant.
Conclusions
The contract objectives were ex-
ceeded. NOx removal efficiency of more
than 90% was demonstrated during an
operating period of about 5000 hours at
the designed capacity of 0.5 MW
equivalent. Following this period,
transient tests were run. These demon-
strated that the performance was not
adversely affected by such conditions as
sudden boiler load changes, cold start-
ups, low boiler loads, or by boiler
shutdowns and start-ups. The pilot
plant operation was terminated after
5620 hours of cumulative operating
time only because the host boiler had to
undergo major modifications necessita-
ting that the pilot plant be moved.
During operation of the pilot plant,
catalysts had to be replaced twice. The
first and second charges of catalyst had
relatively narrow clearances between
the catalyst plates and were clogged by
fly ash from the flue gas resulting in
increased pressure drops through the
catalyst beds. There was also a slight
decrease in apparent catalyst activity,
although true catalyst activity did not
decrease. These catalysts, which were
NOXNON 500, had been tested in pilot
plants in Japan for application to coal-
fired combustion flue gases from
-------
boilers prior to application to the EPA
pilot plant. However, clogging caused by
fly ash was not experienced even
though no sootblower was installed or
operated in those pilot plants.
There were obviously significant
differences in the character and compo-
sition of the fly ash in the U.S. as
compared to that in Japan.
Nevertheless, the tests with NOXNON
600 were highly successful. This
catalyst has somewhat wider clearances
between the catalyst plates, helping to
avoid fly ash clogging 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 compo-
sition of the coal. At present, qualitative
measurements to estimate the tenden-
cies to 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
ash in catalyst beds will be required. It is
expected that this method when further
amplified will be useful in selecting
adequate linear velocity to prevent
clogging and abrasion caused by fly ash,
and to determine the necessity and
operating conditions of the sootblower.
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 available, potential problems
related to the catalyst regeneration
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
washing solution should be investigated
before commercial application. This
area should be studied further since the
cost of this technology could be sub-
stantially 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 and in
Japanese pilot plants, 90% NOX removal
could be expected at a NH3/NOx mole
ratio of 0.92 - 1.0. Although ammonia
slippage from the EPA pilot plant
measured about 40 ppm under these
conditions, ammonia slippage would be
expected to be less than 10 ppm based
on Japanese tests. The differential was
probably due to the asbestos clogging in
the EPA pilot plant. For 80% NOx
removal the required mole ratio would
be 0.82 - 0.85 with ammonia slippage of
5 ppm or less, again based on tests in
Japan.
The pressure drop with NOXNON 600
catalyst in the pilot plant and expected
commercially is only 1.0-1.4 in. HzO.
Such low pressure drops required very
little power consumption resulting in
low operating costs.
The controllability of the pilot plant
was satisfactory. 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 verified
by the high onstream factor achieved. A
desirable addition to the control system
would be a continuous analyzer to
monitor slippage of ammonia.
The NOx removal efficiency of this
SCR system in commercial applications
would be expected to be the same as
that experienced in the EPA pilot plant if
the size and configuration of the
catalyst, superficial linear velocity of the
flue gas across the catalyst, and the
temperature of flue gas were the same.
Therefore, the data is directly applicable
for scaling of commercial systems.
The pilot plant project was the first
demonstration and evaluation of NOx
SCR technology on a coal-fired source
in the U.S. The project results indicate
that the process may be usable as a NOx
control option; however, some technical
concerns remain before the technology
can be considered commercially avail-
able and demonstrated for coal-fired
sources in the U.S.
Metric Conversion
Readers more familiar with metric
units are asked to use the following
factors to convert certain non-metric
units used in this summary.
Non-metric Multiplied by Yields metric
°F 5/9(°F - 32) °C
ft3 28.3 1
in. 2.54 cm
in.2 6.45 cm2
Ib 0.45 kg
Shingo Tanaka is with Hitachi Shipbuilding & Engineering Co.. Ltd., Palaceside
Building, 1-1, Hitotsubashi, 1-Chome, Chiyoda-Ku, Tokyo, Japan and Richard
Wiener is with Chemico Air Pollution Control Corporation, New York, NY
10001.
J. David Mobley is the EPA Project Officer (see below).
The complete report, entitled "Hitachi Zosen NO, Flue Gas Treatment Process:
Volume 1. Pilot Plant Evaluation," (Order No. PB 83-113 829; Cost: $20.50,
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park. NC 27711
-------
3
O
5
a
K
a
a
-o m
?! 11
(*J O O O
00 •< S. 3
ai o 3
3 (D
------- |