Merits and Plant Specific Experience with Bench-Scale SCR

Catalyst Testing

Paper #59
Presented at the Power Plant Pollutant and Effluent Control MEGA Symposium: Best
Practices and Future Trends
August 20-23, 2018
Baltimore, MD

Marc Harton, Richard Storm, Innovative Combustion Technologies, Inc., Pelham, AL;
Jason Bookout, Southern Company Services, Birmingham, AL

ABSTRACT
This paper summarizes the merits and plant specific experience with selective catalytic reduction
(SCR) catalyst bench scale testing. SCR systems are capable of 70-90% NOx removal and they
have been adopted as the most effective and reliable control of NOx emissions, but they are not
without costs. EGUs have transitioned from deployment of SCRs to long term management of
operation, maintenance and total cost of ownership. All catalyst deactivates over time and
catalyst life predictions are reasonably accurate when coal properties and operating conditions
are reasonably constant. Optimal catalyst management is of paramount importance when firing
multiple coals or blends and varied boiler operating conditions can slow or accelerate catalyst
deactivation causing uncertain deactivation rates. A methodical and proper catalyst sampling and
testing program can manage or avoid costs incurred by premature catalyst replacement, provide
precise estimations of catalyst life, evaluate impact of poisoning and fouling species, and
optimize catalyst replacement strategies by incorporating regenerated and-or new catalyst or
installation of additional layers.

INTRODUCTION
NOx is defined as nitrogen oxides (primarily NO and NO2) generated during high temperature
combustion of fossil fuels. Once emitted to the atmosphere, NOx is involved in several reactions
that generate secondary pollutants such as ground level ozone/smog, acid rain, and particulate
matter; therefore NOx levels from utility power plants are regulated by the EPA.1 Methods for
controlling NOx emissions include tuning combustion practices, staged combustion/low NOx
burners, and selective non-catalytic reduction (SNCR); however, the most widely accepted and
most efficient method of NOx control is selective catalytic reduction (SCR), which can generate
NOx reduction efficiencies greater than 90%. The general principle of SCR is the reduction of
NOx with ammonia (NH3) to inert nitrogen gas (N2) and water vapor (H2O) across a titania-
vanadia base catalyst.

1
 SCR catalyst will display reduced NOx reduction efficiency or deactivate over time due to
interactions with various products of combustion in the resultant flue gas. The specific
combustion product species include both solid, non-combusted fly ash material and gaseous trace
elements. Specifically, the most common flue gas species that affect SCR catalyst performance
are arsenic, alkali metals (i.e. Na, K), calcium, and phosphorous. These species interact with the
catalyst to either blind or chemically alter catalyst active sites thereby reducing catalyst deNOx
performance. Additionally, large particle ash (LPA) can settle within the reactors effectively
blocking flue gas from passing through large portions of the total catalyst volume which reduces
the reactor efficiency. Reductions in SCR operating efficiency can lead to emissions compliance
issues as well as operational or maintenance challenges downstream of the SCR due to high NH3
slip and the formation of ammonium bisulfate (ABS) salts. Table 1 below shows minimum,
maximum, and average coal composition values for the three most common coal ranks used in
US power plants.

Table 1 –Coal Quality and Catalyst Poison Composition for Three Different Coal Ranks2
Lignite Subbituminous Bituminous
Species Units Min Max Avg Min Max Avg Min Max Avg
HHV BTU/lb 3,790 7,985 6,000 4,701 11,080 8,726 6,207 15,588 11,782
Ash % 4.0 51.1 17.8 2.3 47.8 15.2 0.9 48.6 12.0
Moisture % 7.8 52.5 38.8 5.2 32.6 19.4 0.4 26.5 6.0
Sulfur (S) % 0.1 6.4 1.0 0.1 19.2 1.0 0.1 20.9 2.4
Arsenic (As) ppm 0.2 615 10 0.2 250 7 0.1 2,200 26
Sodium (Na) ppm 50 12,200 3,273 5 11,400 1,420 6 4,300 446
Potassium (K) ppm 51 10,000 1,211 15 13,500 1,351 15 20,300 1,913
Calcium (Ca) ppm 321 32,200 13,355 190 49,100 7,917 21 131,000 3,457
Phosphorous (P) ppm 15 2,320 539 5 3,820 478 1 8,070 270

Key items from the above table include how the different poisons vary between the different
ranks and the variability, particularly in the trace elements, within each coal rank. The paragraph
above listed the primary flue gas constituents responsible for SCR catalyst deactivation;
however, these species do not impact the catalyst in the same manner. Generally, there are two
main mechanisms for catalyst deactivation: catalyst pore structure blinding or fouling by ash
particles and calcium, and active site poisoning by arsenic, alkali metals, and phosphorous. As
shown in Table 1, both lignite and subbituminous coals tend to be higher in ash, calcium, and
phosphorous content which leads to pore structure fouling/blinding being the primary
deactivation mechanisms for lignite and subbituminous coal-fired applications; while,
bituminous coals tend to be higher in arsenic and alkali metals which leads to active site
poisoning being the primary deactivation mechanism. These different deactivation mechanisms
affect the catalyst deactivation trends to different degrees, with lignite and subbituminous coal-
fired applications generally displaying more rapid catalyst deactivation trends than bituminous
coal-fired applications. It is also important to note the inherent variability in coal quality, not just
between different coal ranks, but also between different coals within the same rank. While major
swings in coal quality are required to have a significant impact on unit operations, relatively
small changes in trace elements such as a 1,000 ppm (0.1%) change in arsenic content can
significantly alter SCR catalyst deactivation trends.

2
SCR owners and operators must have a detailed catalyst management plan in place in order to
ensure the SCR system always operates at the desired efficiencies without incurring undue
capital expense. As part of the catalyst management plan, SCR owners/operators must model
catalyst deactivation rates in order to predict when catalyst replacements are required. These
models can be developed by using either catalyst original equipment manufacturer (OEM)
general predictive models or actual empirical operating and catalyst performance data. Basing
catalyst management decisions on the OEM general predictive models is the simpler option as
the SCR owner/operator does not need to provide any information beyond their design
specifications of their SCR system; however, these boilerplate predictive models are not tailored
to plant specific operating conditions and therefore have an inherent amount of uncertainty
within them. To account for this uncertainty, these OEM models tend to conservatively estimate
catalyst deactivation rates which leads to premature catalyst replacements and unnecessarily
spending plant resources. The second option is implementing a methodical and comprehensive
catalyst management strategy that incorporates empirical plant operating data and catalyst
performance test data to develop true predictive models for catalyst deactivation. While this
option is more rigorous and time intensive, the result is a predictive model uniquely tailored to
each specific SCR system. Using the most current SCR operating data is important as most SCR
systems are operated under different conditions and/or the type of fuel used is different than
when the systems were originally designed. Additionally, conducting annual or semiannual
bench-scale catalyst performance tests precisely determines actual catalyst deactivation rates
over time. It is important to regularly document the actual deactivation trends as slight changes
in the fuel composition or blend as well as varied boiler operating parameters can either reduce
or extend catalyst lifetime by multiple years.

As mentioned above, one of the most important aspects of any catalyst management program is
regular bench-scale testing of the SCR catalyst. The catalyst material itself is one of the more
expensive consumable elements in the SCR system, with all-in replacement costs on the order of
$2MM per catalyst layer. Bench-scale SCR catalyst testing is the most effective way to evaluate
the performance of this expensive asset isolated from any outside operational variables. The data
generated from bench-scale catalyst testing can be used to verify the initial quality of the
material received, calculate maximum deNOx reduction potential of the SCR system, and project
remaining usable lifetime of the catalyst before a replacement is necessary.

The current landscape of utility power plant operation requires the units to be more flexible in
terms of generation output and fuels burned, which leads to additional challenges when
managing an SCR system. Most SCR systems were designed several years ago when utility
power plants were more regularly operated at a base load. Estimating catalyst deactivation rates
and lifetimes under these stable, consistent flue gas conditions was fairly straightforward as the
deactivation rates were predictable. With the change in the operating landscape, current flue gas
conditions at the SCR inlet are much different than the design case. Additionally, any changes in
the fuel being fired, including fuel blending, can alter the SCR inlet conditions as well as the
amount of catalyst poisoning and fouling species present in the flue gas. As changes to the flue
gas and coal ash composition occur due to altering fuels or fuel blends, accurately predicting
catalyst lifetime becomes much more complex. These deviations from the original reactor design
can lead to increased or decreased usable catalyst lifetimes and must be accounted for in the
catalyst management plan to minimize all operating and capital expenses associated with SCR

3
operation. The catalyst management plans used at the Southern Company facilities Gorgas Steam
Plant, Plant Scherer, and Crist Generating Plant are highlighted in this paper to demonstrate how
bench-scale catalyst testing can be utilized in a catalyst management plan to optimize catalyst
replacement schedules and update catalyst deactivation trends following a major operational
change.

BACKGROUND
Bench-scale SCR catalyst testing is the industry standard for evaluating SCR catalyst
performance. The two testing protocols/methods that provide the standards for this testing are the
VGB Guideline for the Testing of DENOX Catalysts, VGB-R 302 and the EPRI Protocol for
Laboratory Testing SCR Catalyst Samples: 2nd Edition. These methods lay out the specific
standards for the catalyst performance, physical property, and chemical analysis testing. There
are many different testing methodologies laid out in these standards, but the three tests most
widely utilized by SCR owners and operators in their catalyst management programs are deNOx
activity, SO2-to-SO3 conversion, and chemical composition analysis. The deNOx activity and
SO2-to-SO3 conversion tests are conducted at a bench-scale testing facility, while the chemical
composition analysis is performed in a typical lab setting.

The goal of a bench-scale SCR catalyst testing facility is to generate a flue gas stream that
closely matches the SCR inlet flue gas composition and characteristics of the full-scale system.
The key flue gas conditions controlled during the testing are temperature, flowrate, oxygen (O2),
NOx, sulfur dioxide (SO2), and moisture. Most facilities burn either propane or natural gas to
generate the flue gas and then dose the flue gas with pertinent gas species to match the full-scale
composition.

Activity is a measure of a catalyst sample’s ability to reduce NOx and is a function of NOx
reduction across the catalyst sample, flue gas flow rate, and total catalyst surface area exposed to
flue gas. Activity is calculated by Equations 1-3 below:

𝑘 = −𝐴𝑉 ln(1 − 𝜂𝑁𝑂𝑥 )

Equation 1
Where:
𝑘 is the NOx activity in m/h
𝜂𝑁𝑂𝑥 is the NOx reduction across the catalyst sample
𝐴𝑉 is the area velocity in m/h

𝜂𝑁𝑂𝑥 is determined by the following equation:

𝐶𝑁𝑂𝑥 ,𝑖𝑛 − 𝐶𝑁𝑂𝑥 ,𝑜𝑢𝑡

𝜂𝑁𝑂𝑥 =
𝐶𝑁𝑂𝑥 ,𝑖𝑛
Equation 2
Where:
𝐶𝑁𝑂𝑥 ,𝑖𝑛 is the concentration of NOx in ppmv at the catalyst inlet
𝐶𝑁𝑂𝑥 ,𝑜𝑢𝑡 is the concentration of NOx in ppmv at the catalyst outlet

4
Area Velocity, 𝐴𝑉, is determined by the following equation:

̇
𝑉𝑓𝑔
𝐴𝑉 =
𝐴𝑔𝑠
Equation 3
Where:
̇ is the volumetric wet flue gas flow in Nm3/h at Normal (Std) temperature and pressure (0 °C,
𝑉𝑓𝑔
1 atm)
𝐴𝑔𝑠 is the total geometric surface area of the catalyst in m2 exposed to flue gas flow

The activity values measured during bench-scale testing are then used to estimate catalyst
deactivation trends over time and to project remaining usable catalyst lifetime. In order to make
these analyses, a few additional pieces of information are required. For estimating the
deactivation trend, the fresh catalyst activity at the time of installation, or initial activity (k0), and
the total hours of flue gas exposure at the time of extraction from the SCR are required. The
accuracy of this additional data is of utmost importance and minor biases in this data can have
major impacts on the quality of the catalyst management plan.

Generally, catalyst deactivation trends are plotted as relative activity over time, where relative
activity (k/k0) is the measured activity at any given time of flue gas exposure (k) over the
measured initial activity (k0). The figure below is an example plot of relative activity over time.

Figure 1 – Plot of Relative Activity Data Over Time with Estimated Deactivation Curves
1.1

1.0

0.9

0.8
k/k0

0.7
A-L1 B-L1
0.6 A-L2 B-L2
A-L3 B-L3
0.5
A-L4 B-L4
0.4
0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000
Operating time, h
In the plot, the individual markers represent the measured relative activity values from each
separate reactor-layer using the measured initial activity of the material and the measured
activity after a given amount of flue gas exposure. The solid lines represent the average catalyst
deactivation rates observed from the bench-scale testing. Typical deactivation rates for a variety
of different fuels/applications are shown in the table below:

5
 Table 2 – Typical Catalyst Relative Activity at 16,000 h Flue Gas Exposure for Different
Fuels/Applications
Application Relative Activity, k/k0
Bituminous 0.70-0.75
PRB 0.60-0.65
Lignite 0.50-0.55
60/40 PRB/Bituminous Blend 0.70-0.75
Low Dust 0.80-0.85

Along with the SCR design data and normal operating conditions, the measured activity values
are used to calculate the individual layer and total reactor potential. The slopes of the trend lines
are used to project catalyst and reactor lifetimes.

Catalyst potential (P) is another measure of the catalyst’s ability to reduce NOx and is calculated
by Equation 4 below.

𝑘
𝑃=
𝐴𝑉
Equation 4
Where:
𝑘 is the deNOx activity in m/h as determined by laboratory testing
𝐴𝑉 is the area velocity in m/h (see Equation 5)

For the potential calculations, the area velocity (𝐴𝑉) is calculated differently than was described
above for calculating bench-scale activity (Equations 1-3). In this case, the AV will be calculated
using the design data for the installed catalyst rather than from the measurements of the bench-
scale sample. This calculation is shown in Equation 5 below.

̇
𝑉𝑓𝑔
𝐴𝑉 =
𝑉𝑜𝑙𝐶𝑎𝑡 𝐴𝑠𝑝𝑒𝑐
Equation 5
Where:
̇ is the total flue gas flow through the reactor in Nm3/h
𝑉𝑓𝑔
𝑉𝑜𝑙𝐶𝑎𝑡 is the catalyst volume for the given layer in m3
𝐴𝑠𝑝𝑒𝑐 is the specific surface area for the given layer in m2/m3

Ideally the two calculated AVs will be the same; however, due to slight differences in the
specific bench-scale sample, damage to the sample, bench testing only one sub-layer, etc. the two
values may differ.

Reactor Potential (PTot) refers to the total catalyst potential existing in the SCR reactor at a given
time. Reactor potential is the sum of the catalyst potentials from each layer of the SCR system as
shown below in Equation 6.

6
 𝑘 𝑘 𝑘
𝑃𝑇𝑜𝑡 = ] + ] +⋯ ]
𝐴𝑉 𝐿𝑎𝑦𝑒𝑟 1 𝐴𝑉 𝐿𝑎𝑦𝑒𝑟 2 𝐴𝑉 𝐿𝑎𝑦𝑒𝑟 𝑛
Equation 6

Once the total reactor potential is calculated from the bench-scale data, the catalyst deactivation
trends noted earlier can be used to project the decrease in individual layer and total reactor
potentials over time and estimate when the total potential will cross below the minimum
potential threshold. Minimum potential (Pmin) is the theoretical minimum threshold to achieve a
NOx reduction performance requirement while maintaining ammonia slip below a specified limit.
Pmin is a function of the required NOx reduction, the NH3/NOx injection ratio, and mal-
distribution in NH3-NOx mixing. As long as the total installed reactor potential, PTot, is greater
than Pmin, the required NOx reduction performance can be achieved while maintaining ammonia
slip below the specified limit.

From the calculated reactor and minimum potential, the observed deactivation trends over time,
and basic reactor design data, a “saw-tooth” curve can be generated. An example plot is shown
below in Figure 2.

Figure 2 – Example “Saw-tooth” Curve of Historical and Projected Reactor Potential

9.0

8.0

7.0

6.0
Potential

5.0

4.0

3.0

2.0

1.0

0.0
0 20,000 40,000 60,000 80,000 100,000 120,000
Operating Time, h
Actual Sampling Trend Projected
Layer 1 Layer 2 Layer 3
Layer 4 Trend-10% Plugging Pmin 90% deNOx

In Figure 2, the black diamond markers are total measured potential values determined from the
bench-scale activity testing. The black curve represents the potential trend to the current testing
period. The “saw-tooth” peaks in this curve represent catalyst layer replacements as replacing an

7
older layer with new or regenerated catalyst increases the total reactor potential. Potential
contributions from the individual layers are shown by the dark red, dark blue, green, and purple
curves. The horizontal light blue line is the minimum potential threshold to achieve the required
deNOx performance. The orange curve is the projected total potential with further flue gas
exposure based on the observed deactivation trends. The dashed red curve represents the
projected potential considering 10% lost due to blockage and erosion, which is a common design
standard.

The point at which the orange, or red dashed curve if considering catalyst plugging, crosses the
light blue Pmin line provides a projection of remaining currently installed catalyst lifetime based
on the testing data. This projection is a key part of any catalyst management program in regard to
scheduling catalyst replacements at the appropriate time, allowing each plant to get the optimal
usage from their catalyst investment without risking any emissions compliance related issues.

Other than catalyst activity and potential determination, the other two primary tests key to
catalyst management plans are SO2 – SO3 conversion and chemical composition analysis. In
addition to the desirable deNOx reaction, SCR catalyst also facilitates an undesirable oxidation of
SO2 into SO3. Formation of SO3 in the flue gas is undesirable due to its reactivity with other
species in the flue gas. SO3 reacts with moisture to create corrosive sulfuric acid and reacts with
unreacted NH3 leaving the SCR system (or NH3 slip) to form sticky ammonium bisulfate salts
which can foul downstream equipment such as air heaters. SO2 – SO3 conversion by the catalyst
tends to increase with increased catalyst age. While this characteristic is not typically the primary
concern when making catalyst management decisions, it is important to track to minimize any
negative impacts on downstream equipment.

Chemical composition analysis is used to evaluate accumulation of catalyst fouling and

poisoning species over time. There are two primary mechanisms that lead to the degradation of
catalyst performance. The first is pore fouling by large macro ash particles that effectively plug
the catalyst pores reducing the total surface area available for facilitating the deNOx reaction.
The second mechanism is active site poisoning where flue gas species bind to the catalyst active
sites preventing the deNOx reaction from occurring. The most common poisoning species seen in
coal-fired applications is arsenic. The primary mechanism causing catalyst deactivation is
directly dependent on the type of fuel being burned. For example, bituminous coals tend to be
higher in arsenic leading to arsenic poisoning typically being the primary deactivation
mechanism in bituminous coal fired application; however, PRB coal is low in arsenic but has a
much higher ash content which leads to pore fouling typically being the primary deactivation
mechanism in PRB coal fired applications. Knowing the primary deactivation mechanisms is
important to any catalyst management program as in some cases minor changes in operating
conditions can decrease the impact of these poisoning and fouling species. An example of such
an operational change is treatment of the coal with limestone prior to being introduced to the
boiler in an effort to reduce the amount of gaseous arsenic at the SCR inlet. These catalyst
chemical composition results and deactivation mechanisms are of utmost importance when
making a fuel switch or blending different fuels as small changes can create major changes in the
flue gas chemistry at the SCR inlet leading to changes in catalyst deactivation rates.
These catalyst performance tests provide invaluable data to be used in tracking catalyst
performance over time and in developing a catalyst management plan unique to each facility.

8
That said, there are several minor considerations that can also have a major impact on the quality
of the test data and analysis. These considerations include:

1. Following proper catalyst sampling, handling, and shipping procedures to protect the
integrity of the samples to be tested
2. Using a measured value for initial catalyst activity rather than the manufacturer guarantee
3. Planning a catalyst sampling event immediately following a fuel switch or a major
change in operating conditions
4. Maintaining accurate records of the flue gas exposure ages of each installed catalyst layer
5. Selecting bench-scale test conditions that match up with actual operating conditions
rather than SCR design data

Proper catalyst sampling, handling, and shipping procedures are of utmost importance to
ensuring quality bench-scale testing results. Any deviations from proper procedures can impact
sample integrity by either physically damaging the samples or contaminating the samples, which
will lead to the extracted sample not being representative of the installed catalyst and
unrepresentative bench-scale test results. Catalyst material must be handled with care to be sure
the material is not torn, bent, or broken upon extraction from the reactor and should be carefully
packed for shipping to the testing facility. Catalyst openings and notches must be preserved to
ensure accurate representation of the flow distribution through the test sample.

It is important to obtain a measured value for any newly installed catalyst material not only to
verify the quality of the material provided, but also to have an exact initial value as the basis for
evaluating catalyst deactivation trends over time. Catalyst manufacturer guarantee values
typically will underestimate the actual installed activity by 1-3 m/h. This underestimation will
lead to a subsequent underestimation of catalyst deactivation trends as shown in the figures
below.

Figure 3 – Relative Activity (k/k0) Over Time with Guarantee Value as Initial Activity (k0)
1.1
A-L1 B-L1
1.0 A-L2 B-L2
A-L3 B-L3
Std Model
0.9
k/k0

0.8

0.7

0.6

0.5
0 5,000 10,000 15,000 20,000 25,000 30,000 35,000
Operating time, h

9
 Figure 4 – Relative Activity (k/k0) Over Time with Measured Value as Initial Activity (k0)
1.1
A-L1 B-L1
1.0 A-L2 B-L2
A-L3 B-L3
0.9 Std Model
k/k0

0.8

0.7

0.6

0.5
0 5,000 10,000 15,000 20,000 25,000 30,000 35,000
Operating time, h
In Figures 3 and 4, the same historical activity performance test data is used; however, in Figure
3 the guarantee value for initial activity is 3 m/h less than the measured value used in Figure 4,
which changes how the relative activities are calculated from each testing period. Comparing the
two figures it is clear that using the underestimated guarantee value will cause an
underestimation of the observed catalyst deactivation trend; ultimately this will lead to an
overestimation of remaining catalyst lifetime.

Along the same lines as conducting baseline testing of newly installed catalyst, it is important to
test all layers of installed catalyst immediately following a fuel switch or other change in
operating conditions. Following any major operational change, the historical catalyst
deactivation trends will likely no longer be an accurate representation of true catalyst
deactivation behavior. By conducting catalyst testing shortly after a major operational change, a
baseline performance of the catalyst is established and will allow for more accurate
determination of deactivation trends during future sampling events.

Furthermore, it is important to maintain accurate records of exposure hours for each layer of
catalyst in order to accurately estimate catalyst deactivation trends based on performance testing.
Overestimating the true catalyst exposure time will lead to an underestimation of catalyst
deactivation rates, while underestimating exposure time will lead to overestimating deactivation
rates. To illustrate this point, Figure 5 below shows a plot of relative activity over time with
accurate exposure hours while Figure 6 shows the relative activity plot with exposure hours
overestimated by 2,000 hours.

10
 Figure 5 – Relative Activity Over Time with Accurate Exposure Time
1.1
A-L1 B-L1
1.0 A-L2 B-L2
A-L3 B-L3
0.9 Std Model
k/k0

0.8

0.7

0.6

0.5
0 5,000 10,000 15,000 20,000 25,000 30,000 35,000
Operating time, h
Figure 6 – Relative Activity Over Time with Exposure Time Overestimated by 2,000 hours
1.1
A-L1 B-L1
1.0 A-L2 B-L2
A-L3 B-L3
0.9 Std Model
k/k0

0.8

0.7

0.6

0.5
0 5,000 10,000 15,000 20,000 25,000 30,000 35,000
Operating time, h

RESULTS AND DISCUSSION

The following discussion details the role of bench-scale testing results in the catalyst
management plans of several Southern Company plants to realize economic benefits and avoid
any emissions compliance issues after making changes to operating conditions. Currently, the
Southern Company coal fleet operates in a much different mode than when the SCR systems
were first brought online. The units are cycled from high to low loads rather than being operated
at a base load. Also, several plants are transitioning to firing different, cheaper fuels or blends of
different coals. These operational changes must be accounted for in the catalyst management

11
plans for each facility for optimal planning of outages and catalyst replacements. Continuing to
use historical data and management strategies leads to inefficient use of plant resources.

Prior to 2017, Southern Company used a single Subject Matter Expert (SME) to create all
catalyst management plans from the original SCR design data. Currently, Southern Company
uses a more rigorous catalyst management tool that uses actual Operational Information Systems
(OIS) data from each plant to develop a unique management strategy tailored to the specific
conditions of each SCR system. The more rigorous and time intensive management tool has
developed more accurate catalyst management plans that have deferred catalyst replacements
and the need for unit outages at several Southern Company plants. The three plants that will be
discussed in further detail are Gorgas Steam Plant, Plant Scherer, and Crist Generating Plant.

Gorgas Steam Plant, Unit 10, is a coal-fired Electric Generating Unit (EGU) located in Parrish,
AL with a nameplate generating capacity of 700 MW. Its SCR system was started in May 2002
and is a high-dust configuration. The bituminous coal burned at Gorgas is high in arsenic and
low in calcium. This leads to very rapid catalyst deactivation by arsenic poisoning. A plot of the
measured catalyst deactivation trends for each layer is shown in the figure below.

Figure 7 – Gorgas Unit 10 SCR Catalyst Deactivation Trends from Bench-Scale Test
Results
1.1
1.0 A-L1 B-L1
A-L2 B-L2
0.9 A-L3 B-L3
0.8 A-L4 B-L4
Std Model
0.7
0.6
k/k0

0.5
0.4
0.3
0.2
0.1
0.0
0 5,000 10,000 15,000 20,000 25,000 30,000
Operating time, h
In addition to the actual deactivation trends of each layer in the above figure, the black curve
shows a standard deactivation curve with relative activity (k/k0) at 0.7 after 16,000 hours of flue
gas exposure. The observed deactivation trends of the installed catalyst are all faster than could
be expected from the standard trends due to severe arsenic poisoning. Historically, catalyst
replacements have been required on an annual basis at Gorgas. This creates a challenge when
trying to develop a catalyst management plan that is not overly conservative, which wastes plant
resources by replacing catalyst before it has reached the end of its usable life, or overly
aggressive, which could lead to difficultly maintaining compliance with emissions regulations.

12
Through the use of actual operating data and regular bench-scale SCR catalyst testing to
regularly document the actual catalyst deactivation rates, the catalyst management plan
developed using the new tool generated significant savings for the plant. The original catalyst
management plan generated by the older tool called for a total of four catalyst layer replacements
and three outages between spring 2017 and fall 2018. Using the updated tool with bench-scale
test results and current operating data, the total number of catalyst replacements was reduced to
three layers with only two outages in the same time period. This change in the catalyst
management plan created a cost savings for the plant over $1.4MM. Table 2 below gives a more
detailed breakdown of the economic analysis and cost savings.

Table 3 – Economic Analysis of Revised Catalyst Management at Gorgas Unit 10

Description Cost
New Catalyst Layer $1,054,911
Catalyst Removal $156,800
Catalyst Installation $196,000
Total Savings $1,407,711

Plant Scherer is located near Juliette, GA and has four EGUs onsite each burning PRB coal with
a total nameplate generating capacity of 818 MW. SCR systems were started on all four units
between fall 2010 and fall 2013. The EGUs at Plant Scherer are unique in the Southern Company
fleet in that the SCRs are only operated on a seasonal basis rather than year-round. This limits
the amount of flue gas exposure time of the catalyst and allows the installed catalyst to remain in
the reactors for longer periods of time than most other SCRs in the fleet. It is important that the
catalyst management plan reflect this unique operating strategy when planning catalyst
replacements. By regularly updating catalyst deactivation trends from the bench-scale test results
and including the specific operating conditions in the decision-making, more aggressive catalyst
replacement schedules can be adopted which can lead to deferring costs until the next outage
period. The SCR system at Scherer Unit 1 will be the focus of this analysis. Figure 8 shows the
catalyst deactivation trends observed from the bench-scale testing at Scherer Unit 1.

13
 Figure 8 – Scherer Unit 1 SCR Catalyst Deactivation Trends from Bench-Scale Test
Results
1.1
A-L1 B-L1
A-L2 B-L2
1.0
A-L3 B-L3
Std Model
0.9
k/k0

0.8

0.7

0.6

0.5
0 5,000 10,000 15,000 20,000 25,000
Operating time, h
As shown in the figure, the measured catalyst deactivation rates for all the layers are slower than
the standard deactivation trend for a PRB coal-fired application (k/k0 of 0.60 at 16,000 h). This
coupled with the seasonal operation of the SCR system has allowed for the deferral of catalyst
replacement costs. The updated catalyst management plan was able to defer over $2MM in costs
from fall 2017 until fall 2020 as shown in the table below.

Table 4 – Economic Analysis of Revised Catalyst Management Plan at Plant Scherer Unit 1
Description Cost
New Catalyst Layer $1,556,749
Catalyst Removal $217,600
Catalyst Installation $272,000
Total Savings $2,046,349

Further analyzing the economic analyses from these two plants shows how insignificant the cost
of a testing program is relative to the savings realized from applying a more accurate catalyst
management strategy. The SCR systems at both Plants Gorgas and Scherer have split SCR
reactors with four layers of catalyst each; therefore, during each catalyst sampling event eight
total samples are extracted for catalyst performance testing. If it is then assumed that a
replacement of one catalyst layer can be deferred once every five years by instituting an annual
catalyst testing program, a rough estimate of return on investment in the catalyst testing program
can be calculated.

Table 5 – Return on Investment of Catalyst Testing Program Over Five Years

Catalyst Estimated 5-year Catalyst
Plant ROI
Replacement Cost Testing Program Cost
Gorgas Unit 10 $1,407,711 $240,000 587%
Scherer Unit 1 $2,046,349 $240,000 853%

14
As shown in Table 5, the cost of a regular catalyst testing program is orders of magnitude smaller
than the potential savings on catalyst replacement.

The final case-study to be discussed is Crist Generating Plant. Plant Crist is located in Pensacola,
FL and has two EGUs with SCR systems onsite, Units 6 and 7, which have nameplate generating
capacities of 320 and 550 MW, respectively. The SCR systems on these units were started in
spring 2012 and 2005, respectively. Both units fire bituminous coal and historically had used a
coal with a moderate arsenic level. In 2017, plant personnel decided to switch to a cheaper coal
with a much higher arsenic level. While it was expected that the catalyst would deactivate more
rapidly due to the increased arsenic content of the coal, the impact on catalyst deactivation trends
was expected to be fairly minimal; however, subsequent catalyst testing following the fuel switch
showed that the increased arsenic content in the coal had a significant impact on the catalyst
deactivation rates. Figures 9 and 10 below show the historical bench-scale testing results from
Crist Units 6 and 7.

Figure 9 – Crist Unit 6 SCR Catalyst Deactivation Trends from Bench-Scale Test Results
1.1

1.0

0.9

0.8
k/k0

0.7

0.6

0.5 Layer 1 Layer 2

Layer 3 Layer 4
0.4 Std Model

0.3
0 5,000 10,000 15,000 20,000 25,000 30,000 35,000
Operating time, h

15
Figure 10 – Crist Unit 7 SCR Catalyst Deactivation Trends from Bench-Scale Test Results
1.1
1.0 Layer 2
Layer 3
0.9
Layer 4
0.8 Std Model
0.7
0.6
k/k0

0.5
0.4
0.3
0.2
0.1
0.0
0 10,000 20,000 30,000 40,000 50,000
Operating time, h
As shown in the two figures above, there were noticeable step-changes in the catalyst
deactivation rates following the fuel switch. Additionally, the Layer 2 catalyst installed in Crist
Unit 7 showed a decrease in activity of nearly 80% after only one-year of operation. Further
backing up the bench-scale test data, the XRF chemical analysis results showed that the new
Layer 2 catalyst had accumulated ~40,000 ppm of arsenic during one-year of operation. While
this result was unexpected, due to the regular catalyst sampling that took place at Plant Crist, the
catalyst management plans were able to be updated to reflect the change in operating conditions
with the new high arsenic coal before any compliance related issues occurred.

SUMMARY
Bench-scale SCR catalyst testing is a vital component of any catalyst management program. The
performance data gathered from a regular catalyst testing program benefits SCR system owners
and operators to get the most value out of the catalyst and ultimately can lead to significant costs
savings to the plant by deferring catalyst replacements. From the Southern Company plants
discussed in this paper, revising the catalyst management programs for two units based on the
results of the rigorous catalyst testing program generated a total savings of over $3.4MM.

In order to extract the most value, and to generate accurate and reliable testing results, the
following five rules should be followed:

1. Proper catalyst sampling, handling, and shipping procedures

2. Using a measured value for initial catalyst activity rather than the manufacturer guarantee
3. Planning a catalyst sampling event immediately following a fuel switch or major change
in operating conditions
4. Maintaining accurate records of the flue gas exposure ages of each installed catalyst layer
5. Selecting bench-scale test conditions that reflect actual operating conditions rather than
SCR design data

16
Relatively small changes the composition of trace elements in the fuel, such as a 1,000 ppm
(0.1%) change in arsenic content, can significantly alter SCR catalyst deactivation trends. In the
case of Crist Units 6 and 7, noticeable step-changes in the catalyst deactivation rates were
detected following a change to a fuel with a higher arsenic content. Although this rapid
deactivation was unexpected, the results from the regular catalyst performance testing allowed
for the catalyst management plans to be updated preventing any operational or compliance
issues.

As the landscape of the power generation industry changes over time with fuel cost and unit
flexibility driving the market, bench-scale testing data is also invaluable to SCR system
owners/operators to provide empirical data to document any changes in catalyst performance.
SCR owners and operators must have a detailed catalyst management plan in place in order to
ensure the SCR system operates at the desired efficiencies without incurring any undue capital
expenses or compliance related issues.

REFERENCES
1. Kitto, J.B.; Stultz, S.C.; Steam its generation and use: Edition 41; 1992; pp 34-1 – 34-15.
2. USGS CoalQual Database. See https://ncrdspublic.er.usgs.gov/coalqual/ (accessed July
2018).

17