Beruflich Dokumente
Kultur Dokumente
Guidelines
TR-110718
Final Report, December 1998
Effective December 6, 2006, this report has been made publicly available in accordance
with Section 734.3(b)(3) and published in accordance with Section 734.7 of the U.S. Export
Administration Regulations. As a result of this publication, this report is subject to only
copyright protection and does not require any license agreement from EPRI. This notice
supersedes the export control restrictions and any proprietary licensed material notices
embedded in the document prior to publication
EPRI 3412 Hillview Avenue, Palo Alto, California 94304 PO Box 10412, Palo Alto, California 94303 USA
800.313.3774 650.855.2121 askepri@epri.com www.epri.com
ORDERING INFORMATION
Requests for copies of this report should be directed to the EPRI Distribution Center, 207 Coggins Drive, P.O. Box
23205, Pleasant Hill, CA 94523, (925) 934-4212.
Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc.
EPRI. POWERING PROGRESS is a service mark of the Electric Power Research Institute, Inc.
Copyright 1998 Electric Power Research Institute, Inc. All rights reserved.
CITATIONS
This report was prepared by
Energy Technologies Enterprises Corp.
7722 Desdemona Court
McLean, Virginia 22102
Principal Investigator
S. Tavoulareas
This report describes research sponsored by EPRI.
The report is a corporate document that should be cited in the literature in the
following manner:
Power Plant Optimization Guidelines, EPRI, Palo Alto, CA: 1998. TR-110718.
iii
REPORT SUMMARY
During the last five years, new software products have become available that use
statistical analysis or neural network techniques to optimize power plant performance
based on multiple objectives. The growing impetus to reduce costs of NOx compliance
and electricity production has accelerated deployment of these applications in the
utility industry.
Background
Optimization software was being implemented in approximately 130 boilers as of
September 1998 (of these applications, more than half were one-time efforts; after an
optimized list of set points had been given to the plant, the optimization software was
removed). Reported NOx emission reduction has ranged from 5 to 40% and heat rate
reduction from 0.5 to 3%. In the combustion area, key objectives are reduction of NOx
emissions, heat rate, and unburned carbon. While some U.S. utilities have gained
experience with software tools, the majority of utilities need more information on how
to select the most appropriate software for their plant optimization programs.
Objectives
To determine the potential improvement in plant performance and emission
reduction through use of power plant optimization software.
To select the most appropriate optimization type among stand-alone, online/advisory, and closed-loop.
To evaluate various options by carrying out a cost-benefit analysis.
Approach
The Power Plant Optimization (PPO) Guidelines were developed with the expertise of
the contractor and EPRI, as well as advice from the Power Plant Optimization Interest
Group (consisting of EPRI members interested and involved in plant optimization). The
project teams approach was to lead utility planners through a five-step process, which
methodically would set appropriate optimization objectives, identify projected
performance improvements, and estimate costs and benefits of various optimization
types. To facilitate this analysis, the team included two Excel spreadsheets with the
program. A beta version of these Guidelines was distributed to 40 utility engineers,
many of whom provided feedback on how to improve them further. Version 1.0, which
accompanies this report, incorporates the comments of these utility users.
v
Results
Forty utility engineers have already used the beta version of these Guidelines. Many of
these engineers have used them to make decisions on how to structure optimization
programs and how to improve efficiency and cost-effectiveness of plant operations. Use
of optimization software is expected to grow significantly, and these Guidelines will
play a critical role in this growth. Two case studies carried out by Allegheny Power and
Northern State Power are included in this document and provide a good example of
how the Guidelines can be used.
EPRI Perspective
Power plant optimization software has become more widespread as a result of the
industrys attempt to find less expensive ways to comply with NOx emission
regulations. Optimization products are of strategic importance to the utility industry,
which soon will operate in a deregulated market in which key economic parameters
and operating objectives change continuously. Optimization tools will allow utilities to
adjust to such dynamic environments. EPRI has played and will continue to play a
leading role in this technical area by supporting utility demonstration of all
optimization tools, disseminating relevant information, and developing software
evaluation guidelines.
AP-110718
Interest Categories
Fossil steam plant performance optimization
Air emissions control
Emissions monitoring
Fossil steam plant O&M cost reduction
Keywords
Heat rate
Nitrogen oxides
Emission control
Performance testing
Computer applications
Computer applications
Cost Reduction
vi
select the most cost effective optimization type among: stand-alone1, on-line/
advisory, and closed-loop, and
low-cost options; it starts from a simple (quick) tuning and moves to more
expensive optimization and potential hardware changes (e.g. NOx control retrofit
options) only as needed or supported by sound economic justification;
Stand-alone type optimization projects were referred to as Off-line/One-time type projects during
the EPRI/ESEERCO Conference and in the conference proceedings (EPRI TR-108687).
vii
setting baseline performance after quick tuning to assess the real potential of
optimization; tuning and baseline performance are described in detail in EPRIs
NOx Emissions Testing and Optimization for Coal-fired Utility Boilers (TR105109) to which the reader is referred frequently;
identifying the type of optimization and the specific software package which is most
appropriate and cost-effective.
Box I Definitions
Closed-loop: fully integrated into the power plant controls; the optimization
system feeds control biases directly into the control system causing the field
equipment to change without human intervention.
viii
These guidelines consist of 5 basic steps (see the figure on the following page) which
are described in the first five sections of the report (one section devoted to each step).
x
Appendix A: Software Users Guide provides guidance on how to use the spreadsheets
which accompany this report on a floppy disk. Guidance is also included for the
applicable sections (Steps 3, 4, and the case studies in Appendix B).
Appendix B: Case Studies provides two examples which illustrate how the guidelines
may be applied to specific power plants.
Appendix C: Formulas provides all the formulas used for estimating the costs and
benefits of power plant optimization in the guidelines and spreadsheets.
The material provided in this report includes both general guidance and step-by-step
instructions to evaluate the various options for a specific power plant under
consideration. In steps 3 and 4, detailed instructions (highlighted in shaded boxes)
provide for using these spreadsheet tools (Tables 3-1 and 4-1).
ix
Step 1
Tune unit &
Establish baseline performance
Step 2
Establish clear needs
& objectives
Consider optimization in
combination with other
options such as:
Step 3
Determine optimization potential/
Can optimization alone meet
established objectives?
YES
Step 4
Identify the most cost-effective
type of optimization
Step 5
Select the best optimization
product for your application
x
NO
CONTENTS
1 STEP ONE: TUNE UNIT AND ESTABLISH BASELINE PERFORMANCE ........................ 1-1
A. Perform Diagnostic Testing ............................................................................................ 1-2
B. Perform Quick Tuning..................................................................................................... 1-5
What is Quick Tuning? .................................................................................................... 1-5
C. Establish Baseline Performance..................................................................................... 1-7
xiii
4 STEP FOUR: IDENTIFY THE MOST COST-EFFECTIVE TYPE OF OPTIMIZATION ........ 4-1
Types of Optimizations: ...................................................................................................... 4-2
A. Availability of DCS and DAS ....................................................................................... 4-2
B. Continuous vs. One-time or Periodic Optimization...................................................... 4-2
C. Performance Improvement Objectives........................................................................ 4-3
D. Cost-Benefit of Alternative Optimization Types .......................................................... 4-3
Background and Instructions for Completing Table 4-1 ...................................................... 4-7
COSTS ............................................................................................................................... 4-7
UP FRONT COSTS......................................................................................................... 4-7
Up Front License Fees ................................................................................................ 4-7
Additional Computer Software and Hardware ............................................................. 4-8
Installation and Calibration .......................................................................................... 4-8
Training ..................................................................................................................... 4-11
Power Plant Hardware Modifications......................................................................... 4-11
ANNUAL COSTS (O&M Including Recalibration).......................................................... 4-11
Annual Software License and Maintenance Fees ..................................................... 4-12
Technical Support ..................................................................................................... 4-12
OPTIMIZATION BENEFITS .............................................................................................. 4-14
UP FRONT (ONE TIME) BENEFITS ............................................................................. 4-14
NPV of Deferred Costs.............................................................................................. 4-14
NPV of Avoided Costs............................................................................................... 4-14
ANNUAL BENEFITS ..................................................................................................... 4-15
Annual Avoided Costs............................................................................................... 4-15
Fuel Cost Savings ..................................................................................................... 4-16
O&M Impacts ............................................................................................................ 4-17
Value of Change in Unit Availability .......................................................................... 4-18
Value of Change in Unit Output ................................................................................ 4-18
Benefits Due to Emission Reductions ....................................................................... 4-19
xiv
B CASE STUDIES..................................................................................................................B-1
Case Study 1: Allegheny Powers Armstrong 1 Unit ........................................................... B-1
Background.....................................................................................................................B-1
Armstrong Unit 1 ............................................................................................................. B-2
Evaluation of Alternative Power Plant Optimization Options ........................................... B-7
Case Study 2: Northern States Power Riverside 7 Unit .................................................... B-17
Background................................................................................................................... B-17
NSPs NOx Compliance Plans and Future Needs..................................................... B-17
Description of Riverside 7 ............................................................................................. B-18
Evaluation of Power Plant Optimization Software ......................................................... B-18
C FORMULAS ........................................................................................................................C-1
xv
LIST OF FIGURES
Figure 1-1 Impact of Auxiliary Air Register Settings on NOx Emissions and Heat Rate
at PEPCOs Potomac River Power Plant ........................................................................ 1-6
Figure 3-1 Effect of air biasing on NOx emissions--400 MW tangential-fired Boiler............... 3-5
Figure 3-2 BOOS results for a 365 MW single-wall-fired boiler.............................................. 3-6
Figure 3-3 Effect of Coal Quality on NOx Emissions Typical Uncontrolled ............................ 3-9
Figure 3-4 Hypothetical NOx Reduction with Balanced Combustion .................................... 3-9
Figure 3-5 Effect of Burner Tilt Position on NOx Emissions for Baseline (Uncontrolled)
and Low-NOx (Controlled) Operation--400 MW Tangential-Fired Boiler ....................... 3-11
Figure 3-6 Effect of Burner Tilt Position on NOx--105 MNW Tangential-Fired Boiler ........... 3-11
Figure 3-7 Effect of Varying the Ratio of Primary Air to Coal on NOx Emissions--105
MW Tangential-Fired Boiler........................................................................................... 3-13
Figure 3-8 Union Electric/Meramec Unit 4: Unburned Carbon vs. Coal Fineness, Coal
Flow Imbalance and Coal Blend.................................................................................... 3-20
Figure 3-9 PEPCos Potomac River 4/Effect of Mill Maintenance on LOI ........................... 3-21
Figure 3-10 Smith 2/Relationship Between LOI, NOx and Coal Fineness ........................... 3-22
Figure 3-11 Typical Particle Size Distribution with Static and Dynamic Classifiers .............. 3-23
Figure 3-12 Utility Experience with Combustion Tuning & Optimization NOx Reduction
Achieved ....................................................................................................................... 3-24
Figure 4-1 Optimization Cost Effectiveness ........................................................................... 4-3
Figure A-1 A Portion of Table 3-1 Before and After a Change is entered.............................. A-6
Figure A-2 A Portion of Table 3-1 Before and After a Second Input is Entered..................... A-7
Figure A-3 A Portion of Table 4-1 Applied to Boiler XYZ; Before Use and After a
Change is Entered .......................................................................................................... A-9
Figure A-4 A Portion of Table 4-1 Applied to Boiler XYZ; After a Second Group of
Changes is Entered....................................................................................................... A-11
Figure B-1 Armstrong Units 1 & 2 Boiler Arrangement .......................................................... B-4
Figure B-2 Burner and Mill Arrangement ............................................................................... B-5
Figure B-3 Armstrong Unit 1 NOx Emissions After IFS Burner Installation ............................ B-6
Figure B-4 Armstrong Unit 1 Boiler Efficiency After IFS Burner Installation ........................... B-6
Figure B-5 Alternative Coal Pipe Arrangements for Improved NOx Control......................... B-20
xvii
LIST OF TABLES
Table 1-1 References and Guidance Documents .................................................................. 1-2
Table 1-2 Comparison of Three Measurement Approaches .................................................. 1-3
Table 1-3 Suitability of Measurement Approaches for NOx Test Elements (Phases) ............ 1-4
Table 3-1 Project Classification to Determine Potential Performance Improvements
Due to Optimization......................................................................................................... 3-3
Table 3-2 Expected Performance Improvements................................................................. 3-23
Table 3-3 Previous Tuning Objective Safe Operation....................................................... 3-24
Table 3-4 Previous Tuning Objective Heat Rate Improvement ......................................... 3-24
Table 3-5 Previous Tuning Objective NOx Reduction...................................................... 3-25
Table 3-6 Previous Tuning Objective NOx Reduction & Heat Rate Improvement ............ 3-25
Table 4-1 Power Plant Optimization Cost-Benefit Analysis.................................................... 4-5
Table B-1 Allegheny's Experience with Power Plant Optimization........................................ B-2
Table B-2 Characteristics of Eastern Bituminous Coal Burned in Armstrong 1...................... B-3
Table B-3 Project Classification to Determine Potential Performance Improvements
Due to Optimization Armstrong 1/Advanced Control for Coal Flow Distribution .............. B-9
Table B-4 Expected Performance Improvements with Optimization (Case 1)...................... B-10
Table B-5 Project Classification to Determine Potential Performance Improvements
due to Optimization Armstrong 1/ No Coal Flow Distribution Control ............................ B-11
Table B-6 Expected Performance Improvements with Optimization (Case 2)...................... B-12
Table B-7 Power Plant Optimization Cost-Benefit Analysis--Armstrong 1 Advanced
Control for Coal Flow Distribution.................................................................................. B-13
Table B-8 Power Plant Optimization Cost-Benefit Analysis--Armstrong 1 Without Coal
Distribution Control........................................................................................................ B-15
Table B-9 Project Classification to Determine Potential Performance Improvements
Due to OptimizationRiverside 7 ................................................................................. B-23
Table B-10 Expected Performance Improvements with Optimization (Case 1).................... B-24
Table B-11 Power Plant Optimization Cost-Benefit AnalysisRiverside 7 .......................... B-25
xix
1
STEP ONE: TUNE UNIT AND ESTABLISH BASELINE
PERFORMANCE
Step 1
Tune unit &
Establish baseline performance
Step 2
Establish clear needs
& objectives
Consider optimization in
combination with other
options such as:
Step 3
Determine optimization potential/
Can optimization alone meet
established objectives?
YES
z Hardware modifications
Step 4
Identify the most cost-effective
type of optimization
Step 5
Select the best optimization
product for your application
1-1
Unit performance deteriorates with time and requires periodic retuning and
optimization. Before optimization is carried out, tuning is recommended because:
x
Sources/Support Material
identify site-specific constraints and practices which may preclude the unit from
achieving better performance, and
1-2
identify potential improvements which can be achieved with minimum effort and
budget.
there are indications of non-uniform air and coal flow distribution in the coal pipes
and flue gas in backpass ducts, which may require multiple gas sampling and coal
pipe measurements, and/or
past operating experience and maintenance records suggest high corrosion rate
and/or malfunctioning equipment.
Table 1-2
Comparison of Three Measurement Approaches
Measurement Approach
Element
Type 1
Type 2
Type 3
Scope
Emissions characterization
with comprehensive
combustion diagnostic and
performance testing
Instrumentation
and sampling and
analytical
procedures
See Table 2 of
TR-105109
Applicability of
Results
Emissions
measurement
precision
r25%
r10-15%
r5-10%
Relative Cost
Low
Moderate
High
Source: TR-105109
1-3
Type 1
Type 2
Type 3
X
X
X
X
Parametric Testing
(preliminary and detailed)
Combustion Tuning
Source: TR-105109
measurement of excess O2 before and after the air heater to estimate air in-leakage
through the boiler casing and the air heater,
All components which may affect measurements used for performance assessment
should be checked. Examples of components to check include:
x
Air registers: Make sure that they are in good operating condition and set properly
(based on operating guidelines).
1-4
Dampers: Air and gas dampers (the latter in case of split backpass) and associated
drives may not be in good operating condition, adversely affecting combustion,
NOx generation, and general unit performance. Tempering air and overfire air
dampers may have similar effects.
Burner and overfire air adjustment mechanisms: Burner tilts and yaws in T-fired boilers
may not be in good operating condition or may require tuning. Similarly, overfire
air adjustment mechanisms in wall-fired boilers may not be operating properly.
Certain measurements are inherently variable (e.g., feeder speed of pulverizers) and
the appropriate average should be used. However, there may be cases in which key
measurements are not repeatable and plant optimization may not be feasible or
practical. Such case has been observed in a couple of power plants which did not have
well-calibrated instrumentation.
Adjust air registers: Very often the registers are set-up to maximize combustion
efficiency (reduce unburned carbon). Sometimes they are in a sub-optimum setting
either because there is a problem with their controls (such as: broken drives and not
properly operating position sensor) or because they have not been tuned recently.
Resetting of the air registers is simple and can improve both combustion efficiency
and NOx emissions. The latter can be achieved if the air register settings contribute
to air staging (creating air-lean flame by distributing more air to the outside of the
flame and towards the top burner elevations). Tuning of Potomac Electric Power
Co.s (PEPCO) Potomac River power plant (see Figure 1-1) is a good example of
how the auxiliary air register settings affect NOx emissions and heat rate. If NOx
reduction is a key operating objective, setting the air registers as shown in Figure 11 could be a good starting point.
Burner Elevation
Original
Setting
(Baseline)
Aux Air
Fuel Air
Air Flow
Burner Elevation
0
O2 (%)
5
NOx (lb/MBtu)
Tuned for
NOx
Minimization
Air Flow
Source: PEPCo
O2 (%)
NOx (lb/MBtu)
Figure 1-1
Impact of Auxiliary Air Register Settings on NOx Emissions and Heat Rate at
PEPCOs Potomac River Power Plant
1-6
Adjust damper set-points: Changes in damper settings may improve air or gas nonuniformities and enhance combustion.
Assess low excess air operation: Lowering excess air is desirable because it reduces
NOx emissions and improves boiler efficiency. However, below a certain excess air
level, unburned carbon and CO emissions start increasing and counter-balance the
benefits of low excess air operation. It is therefore desirable to identify the optimum
excess air level and operate as close to this level as possible. The optimum excess
air level can be identified by reducing the excess air and monitoring CO emissions
and unburned carbon.
Assess combustion staging: Creating fuel-rich and fuel-lean combustion zones in the
boiler has been proven to reduce NOx emissions. Within certain limits, such
changes do not adversely affect combustion efficiency. Techniques which have been
well documented in the literature (EPRIs TR-105109 provides good guidance on
combustion staging) include:
air staging (see above example from PEPCos Potomac River plant),
Improve pulverizer performance: Air-to-fuel ratio and quick tuning of the pulverizer
(e.g., adjustment of spring tension and outlet temperature) can improve coal
fineness, combustion efficiency, and general plant performance. Balancing of air
and coal flows is very important, especially for wall-fired boilers, but it may not be
part of quick tuning because it requires more extensive effort to assess the level of
balancing needed and to implement the necessary modifications (e.g., coal pipe
orificing).
Changes made under quick tuning may have adverse impacts on unburned carbon, CO
emissions, slagging, fouling corrosion, unit generating capacity, rate of load change and
unit heat rate. More information regarding assessment of adverse impacts of tuning
(combustion tuning trade-offs) is provided in Section 7 of TR-105109.
1-7
Depending on the dispatch profile (cycling duty) of the unit, baseline tests may involve
testing only at full load or at various loads. Detailed guidance on how to plan and
implement baseline testing is provided in Section 4 of EPRIs NOx Emissions Testing
and Optimization for Coal-fired Utility Boilers (TR-105109)..
1-8
2
STEP TWO: ESTABLISH CLEAR NEEDS AND
OBJECTIVES
Step 1
Tune unit &
Establish baseline performance
Step 2
Establish clear needs
& objectives
Consider optimization in
combination with other
options such as:
Step 3
Determine optimization potential/
Can optimization alone meet
established objectives?
YES
z Hardware modifications
Step 4
Identify the most cost-effective
type of optimization
Step 5
Select the best optimization
product for your application
2-1
What are the unit, plant, and system NOx reduction requirements?
Lower heat rate: This objective is usually driven by economics; there are no specific
requirements, but targets may be set based on the historical records and the
assessment of potential heat rate improvement. Depending on how well the unit is
tuned, heat rate improvements due to optimization may range from 0.5 to 1.5
percentage points at full load or up to 5 percentage points at low loads.
Are there any unburned carbon requirements, such as maximum UBC (or
LOI), to sell the flyash?
Are there any other adverse impacts from high UBC such as reduced
collection efficiency of the ESP?
If there are no specific requirements to reduce or keep UBC to below a certain level,
economics should dictate the optimum level.
x
Lower opacity
Lower CO emissions:
2-2
Requirements and constraints associated with outlet steam temperatures: Are boiler outlet
steam temperatures below design levels? If yes, is it required to increase these
temperature for safe operation? If operating safety is not jeopardized, determine the
relationship between steam temperature (both superheat and reheat) and improved
heat rate (% improvement in heat rate per degree steam temperature increase).
Annual average NOx of the unit to comply with CAAA requirements; for example:
340 ppm (0.45 lbs/MBtu) for T-fired and 375 ppm (0.50 lbs/MBtu) for wall-fired
boilers burning coal,
NOx cap at plant and/or system level; for example: specified numbers of tons per
year from a multi-unit power plant or a power system,
UBC (or LOI) below a certain level for selling the flyash, if flyash disposal is an
available option, keeping UBC below a certain level may not be a firm requirement,
Lower NOx emissions than required by regulations to allow more expensive units
in the system to operate at higher NOx levels (NOx averaging or trading); therefore,
reducing system compliance costs,
firm requirements,
Reduce annual average NOx emissions from the present baseline level of 390 ppm
(0.52 lbs/MBtu) to 320 ppm (0.43 lbs/MBtu) without adverse operating impacts on
heat rate, LOI, CO, and opacity. NOx reductions beyond this target are not of any
value to the utility at this point in time.
Reduce annual average NOx emissions from the present baseline level of 390 ppm
(0.52 lbs/MBtu) to 320 ppm (0.43 lbs/MBtu). Further NOx reduction and heat rate
improvement are desirable to the extent that these improvements are economic.
Also, adverse O&M impacts are acceptable provided that they result in lower
overall production cost and do not jeopardized plant safety. (Notes: value of
additional NOx reduction to
the system is worth 400-500 $/ton of NOx removed; marginal cost of power is 8
cents/kWh).
2-4
Reduce heat rate while keeping annual average NOx emissions below 360 ppm
(0.48 lbs/MBtu). Additional NOx reductions are not of any value to the utility at
this point in time.
2-5
3
STEP THREE: DETERMINE OPTIMIZATION
POTENTIAL/CAN OPTIMIZATION ALONE MEET
ESTABLISHED OBJECTIVES?
Step 1
Tune unit &
Establish baseline performance
Step 2
Establish clear needs
& objectives
Consider optimization in
combination with other
options such as:
Step 3
Determine optimization potential/
Can optimization alone meet
established objectives?
YES
NO
z Hardware modifications
Step 4
Identify the most cost-effective
type of optimization
Step 5
Select the best optimization
product for your application
3-1
If optimization is not adequate to meet the established objectives, it is still a very costeffective option and should be evaluated in combination with other options such as low
NOx burners, SNCR, or SCR.
Estimating the potential performance improvement requires consideration of sitespecific factors and requirements, as well as the performance improvement potential of
the various types of optimization software. Table 3-1 provides a structure through
which the user may determine the potential performance improvement achieved
through optimization by considering site-specific factors (see printout of Table 3-1 on
the next page and the text Box 3-1 which describes the software version of Table 3-1.
Box 3-1 Software Guidance on Table 3-1
For system requirements, installation and loading of the Excel files included on
the floppy disk, see Appendix A. The Excel file Table 3-1.xls provided on the
floppy disk should be used to carry out Step 3 of the guidelines. A print out of
the table appears in these guidelines on page 3-3. The user should answer the
questions about the technical circumstances of the boiler in question by entering
the appropriate score from 1 to 3 based on the instructions given in this section.
The table then automatically calculates a Total Score between 100 and 300
which indicates the degree to which optimization can improve boiler
performance. Tables 3-2 and 3-2a to 3-2d on pages 37-39 of the guidelines (they
are not included on the disk) provide the performance improvements that can
be expected based on the score generated in Table 3-1.
The main criteria for estimating the level of performance improvement are grouped
into the following four categories:
1. time since last tuning and its objective (weighting: 30%),
3-2
Table 3-1
Project Classification to Determine Potential Performance Improvements Due to Optimization
3-3
3-4
Tuning Objective?
x
Enter 1 if the goal was to ensure safe operation (stable flame and
complete combustion),
Enter 4 if the goal was to reduce both heat rate and NOx emissions.
450
0.6
Baseline
300
0.4
Biased Firing
(m inim um fuel air,
m axim um auxiliary
air to top elevations)
0.2
150
0.0
N O x E m issions (pp m )
The capability to bias fuel flow relative to the baseline setting presents the potential to
reduce NOx emissions by up to 20-25%. More fuel is typically directed to the lower
burner elevations than the upper elevations with the possibility of no fuel in the top
burner. Figure 3-1 shows the NOx reduction achieved at Kansas Power & Lights
Lawrence #5 unit due to biased firing
0
0
E xcess O 2 (% )
Source: Kansas P ower & Light Co, Law rence U nit 5; E PR I Report TR -102906
Figure 3-1
Effect of air biasing on NOx emissions--400 MW tangential-fired Boiler
3-5
When there is adequate fuel supply system capacity, burners out of service (BOOS)
may be used to reduce NOx emissions. Figure 3-2 shows the impact of BOOS on NOx
emissions in a 365 MW front wall-fired boiler.
1.0
750
0.8
600
0.6
450
0.0
0.0
Top pulverizer
Bottom pulverizer
Figure 3-2
BOOS results for a 365 MW single-wall-fired boiler
Fuel biasing may have an impact on combustion efficiency, because it affects the fuel
distribution along the height of the combustion zone and the residence time of the fuel
particles in the furnace.
3-6
Enter 1 for no biasing; i.e. if the fuel feed system (or mills in the case of
coal-fired power plants) are operating at or near maximum capacity,
Enter 2 if some fuel biasing is available, but not enough to have one
mill or row of burners out of service, or
Enter 3 if there is enough fuel feed system capacity to have one mill or
row of burners out of service at full load.
3-7
Enter 1 for no biasing; this is the case in which the air supply system
(fans) are operating at or near maximum capacity,
Enter 2 if air biasing is possible and can increase or decrease air flow
rate through each air compartment by up to 20% relative to baseline
conditions, or
1 .4
1 .2
900
S u bb itu m in o u s
750
1 .0
600
0 .8
L ig nite
0 .6
450
0 .4
300
0
0 .0
0
E xc e s s O 2 (% )
S o urc e: E E R
Figure 3-3
Effect of Coal Quality on NOx Emissions Typical Uncontrolled
3-8
N O x E m is s io ns (p p m )
N O x E m is s io ns (L b /M B tu )
1 .6
NOx
Avg. O2-Unbal.
UBC
Avg. O2-Bal.
NOx Reduction
with Bal. Comb.
Unbalanced
Balanced
Excess O 2
Note: Points on curves represent average operating conditions for each burner mill group (assumes 3 mills total).
Source: EPRI/TR-105109
Figure 3-4
Hypothetical NOx Reduction with Balanced Combustion
The maximum excess O2 and the level of excess O2 at which combustion efficiency
starts decreasing define an operating range. The wider this range, the higher the
potential for performance improvements.
The excess O2 operating range changes with load and operating conditions. If this
operating range has not been established, it is recommended that it be established, at
least at full load, by reducing the excess O 2 until CO emissions and/or unburned carbon
start increasing significantly. For such measurements, a CO monitor and LOI analysis
are required.
3-9
Enter 1 if the baseline excess O2 is within 0.5 percentage point from the
minimum excess O2,
Enter 2 if the baseline excess O2 is within 0.5 and 1.5 percentage point
from the minimum excess O2, or
600
U n co n tro lled
0 .6 0
450
0 .4 0
300
C o n tro lled
0 .0 0
-1 5
0
-1 0
-5
+5
+10
+15
B u r n e r T ilt A n g le (d e g )
S o u r c e : U ta h P o w e r & L ig h t C o m p a n y , H u n te r U n it 2
3-10
+20
+25
N O x E m is s io n s (p p m )
N O x E m is s io n s (L b /M B tu )
0 .8 0
Enter 2 if the burner tilts are operational, but over a limited range
(e.g., -15 to +10 degrees), or
Enter 3 if the burner tilts are operational over the full operating range,
typically -30 to +30 degrees.
0.70
525
0.68
510
0.66
495
0.64
480
0.62
465
0.60
-20
-10
+10
455
+20
Figure 3-6
Effect of Burner Tilt Position on NOx--105 MNW Tangential-Fired Boiler
3-11
Air dampers which control the air flow rate to the four corners of T-fired boilers or
to front versus rear wall of opposed wall-fired boilers,
Enter 3 if the dampers can be adjusted over the full operating range.
3-12
Enter 3 if there is more flexibility with the PA/fuel ratio, for example
PA fuel ratio may change within the range of 1.5 to 2.5.
0.45
337.5
0.40
300
0.35
262.5
0.30
70
80
90
100
225
110
Figure 3-7
Effect of Varying the Ratio of Primary Air to Coal on NOx Emissions--105 MW
Tangential-Fired Boiler
3-13
air registers,
air sleeve dampers and coal nozzle axial position (in some wall-fired boilers),
yaw of burners and/or overfire air ports (in the case of T-fired boilers), and
relationship between burner tilt and overfire air tilt (in the case of T-fired boilers); in
most cases, the overfire air tilt angle is set based on the burner tilt angle; the original
setting could be changed or the overfire air tilt may be decoupled from the burner
tilt.
These parameters are usually set during combustion system tuning which is carried out
by the burner vendor. While readjustment of these parameters may yield performance
improvements, it is not always easy to do so. Constraints may relate to:
3-14
physical layout of the combustion system which makes it difficult to adjust burner
settings, and
3-15
Burner Settings
Indicate the level of flexibility to adjust the burner settings:
x
3-16
Enter 3 if there are more than one piece of equipment that can be
adjusted.
Hardware Modifications
This category includes hardware modifications which are not considered major in
terms of amount of investment, but can improve plant performance significantly.
Examples of such modifications include:
x
addition of dampers,
3-17
Note: there is no Enter 2 option in this and several of the following boxes
3-18
Box 3-2 PEPCo Modifies the Air Distribution System for NOx Optimization
Installation of splitter dampers, tuning vanes and windbox compartmentalization at PEPCos Potomac
River 4 balanced the air distribution from the front to the rear corners and resulted in a more uniform
air distribution in each burner. While it was not possible to assess the NOx reduction achieved with
these modifications (because other modifications were made at the same time followed by optimization),
they were a contributing factor to the overall NOx reduction achieved which reached 30-35%.
Before Modifications
To Front
Corners
To Rear
Corners
Furnace
Hot Air from
Air Heater
After Modifications
Splitter Damper
To Front
Corners
To Rear
Corners
Turning Vane
Windbox Compartmentalization
Furnace
Source: PEPCO
3-19
beneficial impacts of improved fuel balancing, as well as higher coal fineness at Union
Electrics Meramec #4 unit (a front wall-fired 360 MW boiler retrofitted with B&Ws
XCL burners). Improving the coal balance from 10% to 5% from the uniform flow
distribution reduces unburned carbon in the flyash by 2-3 percentage points.
25
+/- 10% Fuel Balance
99.0% 50 Mesh
70% 200 Mesh
% Carbon in Flysash
20
15
+/- 5% Fuel Balance
10
5
+/- 5% Fuel Balance
0
0
25
50
75
100
Figure 3-8
Union Electric/Meramec Unit 4: Unburned Carbon vs. Coal Fineness, Coal Flow
Imbalance and Coal Blend
program, or
x Enter 3 if coal pipe orificing will be included in the optimization
program.
3-20
Mill Modifications
Mill modifications, including installation of exhausters, riffle distributors, and dynamic
classifiers, may result in better coal fineness and coal size distribution control, which in
turn improves combustion efficiency and plant performance.
Coal fineness and size distribution have a significant effect on the combustion process,
and control of size distribution can be used to reduce LOI. Generally, higher coal
fineness results in better coal combustion. Coarse coal particles (usually defined to be
larger than 150 microns [remaining on 100 mesh]) have been shown to be the primary
source of unburned carbon. The percentage of coal remaining on 50 mesh (300 micron)
is increasingly used as the key indicator of adequate coal size distribution.
LOI (%)
Figure 3-9 shows the effect of improved size distribution on LOI at PEPCos Potomac
River #4, a 100 MW T-fired boiler without combustion NOx controls. This
improvement was the result of pulverizer maintenance and tuning. As this figure
shows, the reduction of larger particles (percentage remaining on 50 mesh [300 micron]
was reduced from 3% to 0.5%, resulting in lower LOI even though the percentage
through 200 mesh (75 micron) did not change significantly. Furthermore, improved
coal size distribution allows operation at lower excess O2, which reduces NOx
emissions.
Before
Plus 50 mesh
2 to 3%
Minus 200 mesh 74%
20
18
16
14
12
10
8
6
4
2
0
After
0.5%
73%
Economizer O2 (%)
Source: PEPCo
Figure 3-9
PEPCos Potomac River 4/Effect of Mill Maintenance on LOI
The beneficial effect of higher coal fineness and improved size distribution on both LOI
and NOx emissions is also demonstrated by Figure 3-10 which shows the test results
3-21
from Gulf Powers Smith #2 (a 180 MW T-fired boiler) retrofitted with ABB/CEs
LNCFS III.
increa
sing e
xce
300
ss oxy
0.3
0.2
gen
225
150
low fineness
medium fineness
0.1
75
high fineness
0
0.0
2.0
4.0
6.0
8.0
10.0
12.0
0.4
0
14.0
LOI (%)
Source: Southern Company Services
Figure 3-10
Smith 2/Relationship Between LOI, NOx and Coal Fineness
As coal fineness improves, the same NOx emission level can be achieved with lower
LOI and excess O2. For example, 225 ppm (0.30 lbs/MBtu) NOx level can be achieved
with low fineness2 coal resulting in 10-11% LOI or with high fineness coal resulting in
3% LOI. Such relationships are useful in optimizing unit performance to satisfy its
operating objectives (e.g., NOx requirements and LOI constraints).
One of the options used to improve and control coal size distribution is dynamic
classifiers which achieve a better particle size distribution than static classifiers, (see
Figure 3-11). Additional advantages of dynamic classifiers include:
x
improved flexibility which is particularly important when the coal quality varies,
and
The low and high fineness represent the minimum and maximum fineness achievable with the
pulverizers available at Smith 2; different pulverizers may achieve wider range of finenesses.
3-22
100
% Passing
90
80
Static
Dynamic
70
60
60
100
140
180
220
260
300
Sieve (micron)
Source: EnTEC
Figure 3-11
Typical Particle Size Distribution with Static and Dynamic Classifiers
Mill Modifications
x Enter 1 if mill modifications will not be included in the optimization
program, or
x Enter 3 if mill modifications will be included in the optimization
program.
Total Score
% NOx Reduction
Group A
100-160
5.0 - 15.0
Heat Rate
Improvement
0.00 - 0.75
Group B
160-240
15.0 - 30.0
0.50 - 1.25
Group C
240-300
25.0 - 40.0
1.00 - 1.50
3-23
45
Group C
40
Group B
% NOx Reduction
35
30
25
Group A
20
15
10
5
0
Increasing Operating Flexibility
Worsening Operating Condition of Existing Equipment
Figure 3-12
Utility Experience with Combustion Tuning & Optimization NOx Reduction
Achieved
If the objective of the previous tuning was safe operation, use the upper end of the
projected performance for both NOx and heat rate, as shown in Table 3-2a.
Table 3-3
Previous Tuning Objective Safe Operation
Project Classification
Group A
NOx
10.0 - 15.0
Heat Rate
0.25 - 0.50
Group B
Group C
22.5 - 30.0
32.5 - 40.0
0.75 - 1.25
1.25 - 1.50
If the objective of the previous tuning was heat rate improvement, use the upper
end of the projected NOx reduction, but the lower end of the heat rate
improvement, as shown in Table 3-2b.
Table 3-4
Previous Tuning Objective Heat Rate Improvement
Project Classification
NOx
Heat Rate
Group A
10.0 - 15.0
0.0 - 0.40
Group B
22.5 - 30.0
0.50 - 0.75
Group C
32.5 - 40.0
1.00 - 1.25
3-24
If the objective of the previous tuning was NOx reduction, use the lower end of
the projected NOx reduction range, but the upper end of the heat rate improvement,
as shown in Table 3-2c, and
Table 3-5
Previous Tuning Objective NOx Reduction
Project Classification
Group A
Group B
Group C
x
NOx
5.0 - 10.0
15.0 - 22.5
25.0 - 32.5
Heat Rate
0.25 - 0.50
0.75 - 1.25
1.25 - 1.50
If the objective of the previous tuning was combined NOx reduction and heat rate
improvement, use the lower end of both NOx reduction and heat rate
improvement, as shown in Table 3-2d.
Table 3-6
Previous Tuning Objective NOx Reduction & Heat Rate Improvement
Project Classification
Group A
Group B
NOx
5.0 - 10.0
15.0 - 22.5
Heat Rate
0.0 - 0.40
0.50 - 0.75
Group C
25.0 - 32.5
1.00 - 1.25
The user is encourage to override the above performance predictions, if he/she has a
better understanding of expected performance improvements. For example, prior
efforts to improve heat rate may have proven that the maximum heat rate reduction
may be limited to a specific level which is outside the ranges provided in Table 3-2.
LOI, CO emissions, and opacity can be kept at present levels or improved slightly
depending on the level of improvements being sought for NOx and heat rate; the
higher the NOx and heat rate improvements sought, the lower the possibility for
significant LOI, CO and opacity improvements.
Compare potential performance improvements to project objectives and decide if
plant optimization is a feasible and suitable option.
If optimization seems to meet objectives, proceed to Step 4. If it does not, hardware
modifications should be considered. For NOx Control Retrofit Options, see the
following EPRI publications: 1) Retrofit NOx Controls for Coal-Fired Utility Boilers 1996 Update Addendum (TR-102906-Addendum), and 2) Retrofit NOx Control
Guidelines for Gas- and Oil-Fired Boilers (TR-102413).
3-25
4
STEP FOUR: IDENTIFY THE MOST COST-EFFECTIVE
TYPE OF OPTIMIZATION
Step 1
Tune unit &
Establish baseline performance
Step 2
Establish clear needs
& objectives
Consider optimization in
combination with other
options such as:
Step 3
Determine optimization potential/
Can optimization alone meet
established objectives?
YES
z Hardware modifications
Step 4
Identify the most cost-effective
type of optimization
Step 5
Select the best optimization
product for your application
4-1
Types of Optimizations:
x
On-line/advisory
Closed-loop
The selection of the most appropriate and cost-effective optimization type depends on
the following:
x
Availability of Digital Control Systems (DCS) and Data Acquisition Systems (DAS),
4-2
Closed-Loop
On-Line/
Advisory
% Performance
Improvement
(e.g., % NOx
Reduction)
Stand-Alone
Quick tuning
to new operating mode
Costs ($)
Figure 4-1
Optimization Cost Effectiveness
4-3
For system requirements, installation and loading of the Excel files included on the
floppy disk, see Appendix A. The Excel file Table 4-1.xls provided on the floppy
disk should be used in Step 4 of the guidelines to carry out a cost-benefit analysis of
alternative optimization types. A print out of the Excel spreadsheet appears on pages
4-5 to 4-6. The printed version is for illustration purposes and uses default values
which are based on a typical 500 MW coal-fired boiler. Step 4 of the guidelines walks
the user through the key questions which need to be addressed to adjust the default
values in the electronic version of Table 4-1 to better reflect their own technical
requirements and financial assumptions.
In the Excel file, values appearing in a blue font are user-defined inputs which the
spreadsheet uses to automatically calculate other parameters and cost-benefit
measures. The calculated values and the text appear in black font and they are locked.
To assure consistency, inputs which appear more than once in the spreadsheet are
linked so that the user inputs the value only once (where it appears in blue). The
spreadsheet begins with Up Front Costs and Annual Costs, continues with Up Front
(one time) Benefits and Annual Benefits, and at the bottom provides Cost-Benefit
Analysis Measures.
The user can examine the formulas used throughout the tables in several different
ways. The user can read a description of a formula in English by moving the Excel
cursor on to cells in the left most column to pop up the memo function. Only cells
with a small red spot in the upper right hand side of their cells contain memos. These
formula descriptions also are all listed in the manual as Appendix C.
The values that are used as defaults within Table 4-1 fall within a range that has been
found to be typical based on optimization experience to date. This experience and the
typical ranges are presented in this manual in shaded boxes. The general text in this
section provides background information on all the inputs while the text in the shaded
boxes provides step-by-step instructions.
4-4
Table 4-1
Power Plant Optimization Cost-Benefit Analysis
COSTS
Optimization Types
Stand-alone
O n-line/advisory
Closed-loop
$30,000
$130,000
$170,000
$5,000
$0
$5,000
$0
$0
$5,000
$0
$5,000
$0
$0
$5,000
$0
$5,000
$0
$0
$33,000
$14,600
12
$800
$5,000
$18,400
25
$320
2.3
$0
$44,040
$14,600
12
$800
$5,000
$29,440
40
$320
2.3
$0
$50,120
$17,000
15
$800
$5,000
$33,120
45
$320
2.3
$0
Training ($)
$0
$0
$0
$0
$0
$0
$68,000
$179,040
$225,120
$5,000
$13,000
$17,000
$31,720
$17,000
15.0
$800
$5,000
$14,720
20
$320
2.3
$24,040
$13,000
10.0
$800
$5,000
$11,040
15
$320
2.3
$20,360
$13,000
10.0
$800
$5,000
$7,360
10
$320
2.3
$36,720
$37,040
$37,360
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
$0
$0
$0
$0
$130,165
9,000
0.40
1.27
500
65.0
$244,059
9,000
0.75
1.27
500
65.0
$406,765
9,000
1.25
1.27
500
65.0
Up Front Costs
Up Front Software License Fees ($)
Additional Computer Software and Hardware ($)
Com puter so ftw are
Com puter hardw are
Instrum entation and co ntrols (I&C)
Data acquisition system (DAS)
BENEFITS (Losses)
Up Front (one time) Benefits
NPV of Deferred Costs ($)
Deferred costs
How long does optim izatio n d elay investment? (in month s)
Annual Benefits
Annual Avoided Costs ($/yr)
Fuel Cost Savings ($/yr)
Baseline heat rate (Btu/kWh)
Unit heat rate im provem ent (% Pts)
Fu el cost ($/MBtu)
Unit output (MW)
Capacity factor (%)
4-5
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
0
0
$0
0
0
$0
0
0
$0
0
0.00
65
0
$0
0
0.00
65
0
$0
0
0.00
65
0
NOx Emissions
Avoided fines/penalties ($/yr)
Annual value of NOx emission credits ($/yr)
-Value of NOx emission credits ($/ton removed)
NOx emission credits (tons/yr)
-NOx reference level for credits (lbs/MBtu)
-NOx emission reduction below reference level (%)
-Unit output (MW )
-Unit baseline heat rate (Btu/kWh)
-Capacity factor (%)
$0
$0
0
320.3
0.500
5.0
500
9,000
65
$0
$0
0
640.6
0.500
10.0
500
9,000
65
$0
$0
0
960.9
0.500
15.0
500
9,000
65
SO2 Emissions
Avoided fines/penalties ($/yr)
Annual value of SO2 emission credits ($/yr)
-Value of SO2 emission credits ($/ton remov ed)
SO2 emissions reduction (tons/yr)
-Baseline SO2 emissions (lbs/MBtu)
-Baseline unit heat rate (Btu/kW h)
-Unit heat rate improvement (%)
-Unit output (MW )
-Capacity factor (%)
$0
$6,150
100
61.50
1.20
9,000
0.40
500
65
$0
$11,530
100
115.30
1.20
9,000
0.75
500
65
$0
$19,217
100
192.17
1.20
9,000
1.25
500
65
CO Emissions
Avoided fines/penalties ($/yr)
$0
$0
$0
Opacity
Avoided fines/penalties ($/yr)
$0
$0
$0
$0
0
10,686
208.52
9,000
0.40
500
65
$0
0
20,036
208.52
9,000
0.75
500
65
$0
0
33,393
208.52
9,000
1.25
500
65
$136,314
0.50
$255,589
0.75
$425,982
0.90
$68,157
$191,692
$383,384
5.0
4.9
5.0
4.9
5.0
4.9
$31,437
$136,484
$13,600
$154,652
$671,417
$35,808
$346,024
$1,502,254
$45,024
$68,484
26
26.0
1.0
$492,377
66
13.9
2.8
$1,277,134
134
7.8
5.7
Waterwall corrosion
Operating flexibility
CO2 Emissions
Annual value of CO2 emission credits ($/yr)
-Value of CO2 credits ($/ton remov ed)
CO2 emissions reduction (tons/yr)
-Baseline CO2 emissions (lbs/MBtu)
-Baseline unit heat rate (Btu/kWh)
-Unit heat rate improv ement (%)
-Unit output (MW)
-Capacity factor (%)
FINANCIAL ASSUMPTIONS
Book life of the project (default 5 years)
Discount rate (%) (default 4.9%)
INTERMEDIARY FIGURES
Net annual benefits ($)
PV of future net annual benefits ($)
Annual depreciation (straightline)
4-6
UP FRONT COSTS
Up Front License Fees
Up front license fees may apply to a unit, a multi-unit power plant, or power system in
which the optimization software is used and may include other items such as:
x
training.
Stand-alone: $30,000
On-line/advisory: $130,000
Closed-loop: $170,000
Vendors have quoted up front fees for stand-alone systems as low as $15,000 and
as high as $60,000 per boiler. Similarly, on-line/advisory and closed-loop
systems have been quoted as low as $100,000 and as high as $250,000 per boiler.
In most cases, these costs are for turn-key systems which include software license,
computer hardware, outside technical support, and training. In such cases, avoid
entering the costs twice; enter the cost once (e.g., as Up Front License Fees).
Annual licensing and/or maintenance fees, which are usually in the range of
10% of the Up front fee, are recorded under Annual O&M Costs (see further
explanation below). The buyer should clarify with the optimization software
vendors what are the proposed prices and what is included.
4-7
Computer Software x
Computer Hardware x
Instrumentation and controls-related components; in most cases, especially in standalone optimization, there is no need for special instrumentation or controls.
However, some instrumentation, such as CO monitor and LOI measurement device,
may be considered essential depending on the optimization objective. CO monitor is
essential in most cases. A LOI monitor may be necessary, especially in cases where
LOI is expected to be affected significantly or where LOI is already a concern.
Default Input: $0
Default Input: $0
Data acquisition-related software and hardware; in most cases, the existing DAS is
adequate, or the optimization software includes the required DAS capabilities.
However, the adequacy of the DAS-optimization system interface should be
confirmed and if additional components are needed, they should be included in the
cost-benefit analysis.
Default Input: $0
4-8
Experience has shown that initial calibration takes 3-4 calendar months. Typical
level of effort is:
Outside Technical Support: is usually included in the fixed fee charged for the
optimization software. However, it is important to clarify how much support is
provided (in terms of man-days in the field and support by phone or e-mail). Twothree weeks of on-site support seem to be an appropriate level at this point in the
development of the software. In the future, these requirements may decrease. At
present, if the vendor claims that installation/ calibration can be done in less than the
average time (see Table 4-1), it should document this with data other utility sites.
Reduced technical support by the vendor may compromise the objectives of the
optimization program and/or may increase the work required by in-house staff.
Outside Technical Support - Defaults Inputs:
x
x
x
Stand-alone:
On-line/advisory:
Closed-loop:
12 man-days
12 man-days
15 man-days
Typical average daily rates for outside technical support are $800/day, and
miscellaneous expenses for travel, report writing, etc. over the time period
required typically add up to $5,000.
4-9
In-house Technical Support: is critical for the success of the project, but very often is
overlooked in the evaluation of the various options. Activities which should be
included are:
x
Based on the above assessment regarding the duration of the installation and
calibration of optimization software, the following estimates should be considered:
In-house Technical Support - Default Inputs:
x
x
x
Default Input: $0
4-10
define clearly (along with the optimization vendor) the tests which
result in non-economic operating conditions,
determine their impact on key parameters, especially unit output (MW),
heat rate, O&M costs, reliability, and emission rates,
estimate the time the unit will be in each non-economic dispatch, and
estimate the total cost due to non-economic dispatch and insert value of
result in Table 4-1.
Training
Training is usually included in the up front software license fee. The importance of
training relates to the plant owners strategy regarding long-term utilization of the
software, and the general plant O&M management approach. If the plant owner relies
more on outsourcing, limited training is needed (e.g., less than one week). However,
adequate outside technical support should be provided under the annual O&M costs.
Many plant owners prefer to build in-house expertise to use the software, in which
case, more training may be needed (e.g., two weeks).
In terms of normal operation of the optimization software, the need for training is the
highest for stand-alone and the lowest for closed-loop applications. Stand-alone
systems require well-trained staff, because of the need to repeat the optimization
frequently, if the desirable performance improvement is to be sustained. Online/advisory optimization requires similar training because the plant staff
(performance engineers and plant operators) needs to be well-aware of how to use the
optimization system because it decides whether or not to implement the optimum
settings. Closed-loop systems require minimum training (relative to the other two
options), they require more in-depth training for interfacing the optimization software
with the control system.
Training -
Default Input: $0
Assume that training costs are included in the up front license fee.
Power Plant Hardware Modifications
If hardware modifications are included in the optimization project (see Step 3), the
capital costs associated with such modifications should be included. Because the
objective of most optimization projects is to use optimization software to avoid the need
for hardware modifications or reduce their scope, the default value of this entry is zero.
Power Plant Hardware Modifications
Default Input: $0
4-11
It is important to clarify both the price and the scope of the annual license/AF fee.
Annual Software License Fees - Default Inputs:
Annual license fees are assumed to be:
x
x
Technical Support
On-going technical support and maintenance depends on:
x
Stand-alone:
15 man-days
On-line/advisory:
10 man-days
Closed-loop:
10 man-days
Average daily rate for outside technical support: $800/day
Miscellaneous expenses: $5,000 for travel, report writing, etc.
Provided that the optimization system works reliably and does not require major
reconfiguration and maintenance, closed-loop systems require less on-going technical
support than stand-alone or on-line/advisory systems. However, if the software
experiences problems or resetting of the control loops is required, technical support
from both in-house staff and outside organizations may be needed. Whether in-house
staff will be used or not depends on the level of training provided, their availability,
and preferences of the plant owner.
In-house Technical Support Default Inputs:
x
x
x
x
x
Stand-alone:
20 man-days
On-line/advisory: 15 man-days
Closed-loop:
10 man-days
Average daily rate for in-house technical support: $320/day
Cost multiplier for overhead: 2.3
It should be noted that all annual O&M costs associated with optimization software are
treated as fixed O&M.
4-13
OPTIMIZATION BENEFITS
This section describes the assumptions that are used to determine the benefits of an
optimization project. Similar to the costs, the benefits are broken down into two
categories. Up front benefits consist of the benefits from deferring or avoiding the need
to install capital equipment (usually NOx control equipment). All the other benefits are
derived from more efficient plant operation and are thus counted annually. Annual
benefits include fuel cost savings, O&M impacts, improved unit availability, increased
unit output, and emission reductions.
Default Input: $0
Default Input: $0
ANNUAL BENEFITS
Table 4-1 provides a step-by-step process to determine the annual benefits due to
optimization. First the user is guided to provide inputs based on which the annual
benefits (ideal annual benefits reflecting the initial performance improvements) due
to optimization are estimated. Then, a diminishing effectiveness factor is applied to
take into account potential degradation of performance with time, after initial
optimization has been performed. These items are found at the bottom of the benefits
section of the spreadsheet, just before financial assumptions. Such degradation is real
and varies with the optimization type.
Diminishing effectiveness factor
Definition: percentage of the initial benefits which will actually be realized over time
given the gradual deviation between actual plant performance and software predictions.
Given industry experience, we estimate that the effectiveness of stand-alone
optimization will achieve 40-70% of their initial benefits. This is mainly due to
changes in the operating characteristics of the plant, including coal quality, plant
operators preferences, equipment degradation, etc. On-line/advisory
optimization is expected to maintain a higher level (60-90%) of the initial
benefits, while closed-loop will maintain an even higher level (80-95%). Default
inputs for the diminishing effectiveness factor used in Table 4-1 are as follows:
Default Inputs:
x
x
x
Stand-alone:
On-line/advisory:
Closed-loop:
0.50
0.75
0.90
increase in annual O&M costs due to a low NOx burner and overfire air
retrofit
difference in annual O&M costs between (a) an SCR system which is designed
to achieve a certain NOx reduction and (b) a combination of optimization
software plus SCR system; the latter SCR has lower NOx emission at its inlet
and therefore has smaller catalyst and lower consumption of ammonia.
Default Input: $0
Excess O2,
CO emissions,
Other unit operating variables (e.g., overfire air, combustion air distribution and
burners out of service) which may impact stack and steam temperatures,
During the actual optimization, the impact of each of these parameters on heat rate
needs to be assessed in detail. For the purposes of these guidelines, the results of Step 3
are adequate to assess the cost-effectiveness of the different optimization options.
The baseline unit heat rate and the unit heat rate improvement used in the above
formula should be based on a weighted average which reflects unit dispatch profile.
The heat rate improvement range provided above is typical of coal-fired baseloaded
units. In case of intermediate load or cycling units, the heat rate improvement may be
higher (up to 4-5%).
4-16
Unit heat rate improvements vary with optimization type and the specific
operating conditions of each unit. The user should input the heat rate
improvement predicted during Step 3 of these guidelines (see Tables 3-2a
through 3-2d) or should override it if s/he has a better estimate (based on sitespecific considerations). From the range developed in Step 3 (e.g., heat rate
improvement: 0.40 - 1.25%), stand-alone optimization is expected to achieve heat
rate improvement closer to the lower end of the range (e.g., 0.40%), online/advisory in the middle (e.g., 0.75%), and closed-loop at the upper end of the
range (e.g., 1.25%). Default inputs used in Table 4-1:
x
x
x
Stand-alone:
On-line/advisory:
Closed-loop:
O&M Impacts
In addition to impacts which directly affect the unit heat rate (which have been
included in the previous input), there may be others impacts (both positive and
negative) on operation and maintenance of the unit including:
x
unit operating flexibility such as rate of load change and limited operating range of
certain control variables.
Waterwall corrosion may occur especially in cases where the optimization results in
lower (than baseline) excess O2 and/or staged combustion. However, there is not yet
adequate information to quantify the impacts due to waterwall corrosion. EPRI is
monitoring the impacts of staged combustion on waterwall corrosion and is expected to
release guidelines on how to assess such impacts. EPRI is carrying our research on
waterwall corrosion and will continue to provide the latest information in this field.
Default Fuel cost used $1.27/MBtu for Illinois #6 coal delivered to plants in the East/West Central US.
Source: EPRI Technical Assessment Guide (TAG), P-6587L, Vol. 1, 1989.
4-17
Similarly, quantification of the benefits or costs due to reduced rate of load change or
limited operating flexibility is site-specific and not easy to generalize.
Considering that no such impacts have been documented in plants which have used
optimization software, we assume that these impacts are negligible.
O&M Impacts
Default Input: $0
0
0
4-18
0
0
65
0
Avoided Fines (Penalties); such fines may be imposed if the unit does not comply with
environmental regulations,
Emission Credits; these are allowances which are accumulated as a result of emission
reductions.
For all pollutants, but especially for NOx, the emission reduction requirements should
be clarified to avoid double-counting. For example, if optimization helps the power
plant achieve compliance, either an avoided fine or a deferred/avoided cost should be
recorded, not both.
NOx Emissions
Avoided Fines/Penalties: NOx emissions are usually limited below a certain level (e.g.,
under CAAA requirements). While not always the case, penalties may be imposed by
Federal EPA, State agencies or local authorities if compliance is not achieved. If
combustion optimization achieves compliance and avoids penalties, they should be
included in Table 4-1 as benefits.
Avoided Fines/Penalties ($/yr) -
Default Input: $0
Annual Value of NOx Emission Credits: Credits may be generated by reducing NOx
emissions below a certain level, usually the compliance level for the unit as dictated by
unit, plant or power system NOx requirements.
4-19
0
0.50
5-15%
500
9,000
65
NOx reference level could be either the baseline NOx emissions or a NOx emission
level below which NOx credits are accumulated. If the user knows the value of NOx
($/ton) and wants to estimate the total value due to the NOx reduction achieved by the
optimization project, s/he should use the baseline NOx emissions. Alternatively, NOx
credits may be accumulated after a certain level of NOx has been achieved. For T-fired
units, such NOx level may be 320 ppm (0.43 lbs/MBtu), which satisfies a presumed
compliance level of 340 ppm (0.45 lbs/MBtu) and provides some safety margin.
NOx emission reduction due to optimization has been estimated in Step 3 of these
guidelines. The following clarifications should be made:
x
The % NOx reduction entered into the spreadsheet should be in reference to NOx
reference level which could be different from the baseline NOx emissions,
SO2 Emissions
Avoided Fines/Penalties: SO2 emissions are limited to a certain level either on a unitby-unit basis or through a system-wide maximum. Penalties which may be imposed by
Federal EPA, State agencies, or local authorities are site specific. Usually, combustion
optimization helps reduce SO2 emissions through improvements in plant efficiency and
heat rate. Because such improvements are usually no more than 1-2% for full load and
less than 5% at low loads, optimization is not expected to reduce SO2 emissions
significantly enough to be the main reason for avoiding penalties. However, if such
penalties are avoided through a combination of options with plant optimization being
one of them, a portion of the avoided penalties should be included in Table 4-1.
Avoided Fines/Penalties ($/yr) -
4-20
Default Input: $0
Annual Value of SO2 Emission Credits: Such credits can be generated by reducing SO 2
emissions below a certain level, usually the compliance level for the unit dictated by
unit, plant or power system SO2 requirements.
Annual Value of SO2 Emissions Credits ($/yr) - Default Inputs:
x
x
x
x
x
x
CO Emissions
CO emissions are usually regulated to 100 ppm. Most optimization projects are carried
out in such a way that this limit is not exceeded. Where CO emissions are above the
allowable limit, optimization could be used to reduce them. In this case, the project
should be credited for avoiding fines.
Avoided Fines/Penalties ($/yr) -
Default Input: $0
Opacity
Similarly to CO emissions, opacity is regulated by local and federal standards.
Limiting opacity below the allowable limit is usually a constraint in power plant
optimization and therefore no deterioration is expected as a result of optimization. In
cases where the opacity does not meet the regulatory limit, the optimization project
should take credit for opacity reduction, including potential avoidance for fines.
Avoided Fines/Penalties ($/yr) -
Default Input: $0
CO2 Emissions
Presently, there are no regulations limiting the release of CO2 emissions. While it is
unlikely that such regulations imposing fines would be enacted in the foreseeable
future, it is possible that CO2 emission reduction may have a value ($/ton). In this case,
power plant optimization which results in heat rate improvement should take into
account reduced CO2 emissions as an additional benefit. Most international
organizations (e.g the UN Inter-governmental Panel on Climate Change) have
suggested the values ranging from 20 to 100 $/ton of CO2 removed.
4-21
x
x
0
208.521
9,000
0.4 - 1.25
500
65
The Baseline CO2 emissions default 208.52 reflects the following assumptions: Illinois #6 coal
(see EPRI TAG for coal analysis and heating value) and 20% excess air.
Return on Investment (%): ROI = Net annual benefits (Annual benefits - Annual
costs) - annual depreciation / Up front costs
ROI does not take in to account the time value of money or cash flow profile of the
project. It also does not take into account the overall scale of a project relative to
other projects.
4-22
4-23
5
STEP FIVE: SELECT THE BEST OPTIMIZATION
PRODUCT FOR YOUR APPLICATION
Step 1
Tune unit &
Establish baseline performance
Step 2
Establish clear needs
& objectives
Consider optimization in
combination with other
options such as:
Step 3
Determine optimization potential/
Can optimization alone meet
established objectives?
YES
z Hardware modifications
Step 4
Identify the most cost-effective
type of optimization
Step 5
Select the best optimization
product for your application
5-1
number of power plants which will utilize optimization software: the larger the number of
plants and therefore the level of commitment, the more need for a rigorous
assessment, and
the utilitys approach regarding procurement of products and services: some utilities may
prefer establishing a close working relationship with the software vendors in which
case informal evaluation and negotiations of contractual or joint venture terms
dictate the final selection of an optimization tool. In other cases, utilities may prefer
to thoroughly evaluate every optimization tool under consideration.
Independently of the level of detail, there are a number of key factors which should be
considered in the evaluation. The purpose of this section is to identify these factors and
provide general guidance on how to select the best optimization software for a given
project.
Key Factors
Key factors which should be considered in evaluating optimization products include:
x
In most cases, the first three factors determine the final selection. The fourth factor
(benefits) is very important too. However, it is often difficult to estimate the difference
5-2
in benefits between the various software. The few rigorous evaluations performed so
far have concluded that the expected benefits from various software for each type of
optimization are within r 5-10%, which is well within the level of accuracy of the
analysis. Any larger differences were based on benefits claimed by the vendors, but
not confirmed with actual field experience.
Box 5-1 (see next page) provides an example of a detailed software specification and
evaluation for neural network-based software which was carried out at Tennessee
Valley Authoritys (TVA) for the Kingston power plant.
5-3
Box 5-1 TVAs Experience in Selecting a Neural Network-based Optimization Software for
Kingston #9
In June 1997, TVA selected a neural network-based optimization software for its
Kingston #9 coal-fired unit. The objectives of this project were to:
x
build in-house expertise needed to carry out similar projects throughout the
TVA power system (tentatively 59 units have been identified for
optimization).
Because of the level of commitment and the potential benefits, TVA decided that the
optimization software selection deserved detailed specification and evaluation. TVA
took the following steps:
A. Assessment of experience base of commercial optimization software
An industry survey was carried out to familiarize the utility staff with the state of the
technology and industrys experience. This was carried out mainly through requests to
software suppliers for information on their products and the industry sites using them.
Based on this information, TVA narrowed down the list of qualified bidders to four and
invited them to TVAs offices for presentations and detailed discussions. Finally, TVA
visited selected utility sites which use such optimization software and held extensive
meetings with plant personnel. These meetings were considered to be particularly
useful.
B. Detailed specification document prepared and sent to the vendors
The specification included what the TVA team considered to be the best features of
commercially available software. These features were identified in the previous step
and included: sensor validation, model retraining, financial predictor, and contribution
analysis. Furthermore, the specification provided a description of the Kingston #9 unit,
the vendor scope of work, DCS/communication requirements, optimization software
capabilities, technical support requirements, implementation schedule, licensing
agreement preferences, and evaluation criteria.
C. Bid evaluation and vendor selection
Key evaluation criteria included: power plant optimization experience, neural network
experience, neural network support capabilities, training, documentation and overall
costs. TVA requested the prices of various options in order to assess the cost of
implementing optimization throughout the power system (enterprise license), at a
5-4
TVA provided plant data for process model building prior to delivery,
the key factors which contributed to the final selection were: proven field
experience of the software, experienced team, and costs,
the evaluation process, especially the blind test, was particularly useful in
assessing the experience of the suppliers and their ability to support the
application, and
the performance and expected benefits of the evaluated products was too
close to call.
5-5
Request from the vendors to provide a list of the sites which have used their
software, as well as key contacts,
Call selected utilities (at least two for each software being considered) and asked
them about their overall experience,
select the sites which resemble more closely the application you are
considering; for example, if you plan to install a closed-loop optimization
system, select a site which utilizes optimization software in a closed-loop
mode over one that is on-line/advisory; however, if you have not decided
which operating mode is best for your application, it may be worth
visiting both closed-loop and on-line/advisory sites to assess their
advantages and disadvantages,
during the plant visit, it is more important to spend time with key plant
personnel rather than tour the site; especially in the case of closed-loop
systems where there is not much available to see. In most cases,
discussion with plant personnel provides invaluable insights into the
process of planning and implementing optimization systems.
key components of the optimization software system, especially data acquisition, data
validation, neural networks or statistical analysis modules, and optimization
algorithms.
Most importantly, the vendors should understand the units flexibility and capabilities
to establish realistic goals for the optimization program. Considering that some of the
power plant optimization vendors are small companies with limited track records, it
may be necessary to assess their financial depth and business viability.
A first assessment of the vendors experience can be done by asking them to submit
information on prior projects, as well as resumes of key personnel who will be involved
in an optimization project. A more thorough assessment requires direct (person-toperson) interaction. Opportunities to do so, include:
5-6
a working meeting between the utility and the optimization vendor to review the
requirements of the project, identify the key issues associated with the project, and
discuss the potential approach which could be followed,
Computer hardware,
If these items are not included in a fixed price fee provided by the vendor,
they should be added to the total project cost estimate (Table 4-1).
x
Computer hardware, I&C and DAS: to the extent that additional computer
hardware, instrumentation, controls (e.g., new DCS) and data acquisition
system are needed, the user should make sure that there is consistency in
assumptions between software; for example, if the utility decides to add an
on-line LOI analyzer, its cost should be added to all software being
evaluated.
5-7
Installation & Calibration: In addition to the cost of outside technical support, which
is usually provided by the vendors as a fixed price, the utility should ask vendors to
estimate and document the duration of the installation and package calibration. For
example, a software package which requires a separate calibration at various loads,
will need more time to be calibrated than a model which treats load as a variable.
Particular attention should be paid to the data provided by the vendors in case they
differ significantly from the default values in Table 4-1.
Training: Most optimization software packages are user friendly in terms of normal
utilization. Therefore, training needs and costs provided by the various vendors
should be similar. If not similar, it is worth clarifying further the differences. The
utility should consider asking the vendor to provide the costs of training as a
separate cost item. This is particularly important if the utility plans to build inhouse capability to do future optimization projects, in which case it may need more
training.
Annual O&M Costs: The vendors should be asked to clarify the following:
What does the annual fee include? Does it provide all new releases of
the software?
How often and long does it take to recalibrate the model? Is it done
automatically or does the software requires retraining similar to the
original calibration? Note:
some models learn on-line, but others need to be off-line to
be recalibrated,
some models require separate calibration at each load which
is more time-consuming,
The licensing fee and, in general, the contract with the vendor is negotiable and
depends on market conditions (level of competition, how badly the vendor wants the
job, potential future business as a result of the project under consideration, etc.).
Each vendor has some flexibility to reduce the price being offered (after all software
development expenses are sunk costs which the vendor hopes to recover), but a
significant percentage of the costs are labor costs associated with installation,
5-8
calibration, and technical support which can not be reduced significantly. In fact,
utilities should be careful to ensure that adequate level of technical support is included
in the vendors proposal; otherwise, they may have to increase the estimated level of
effort for in-house staff.
the limited experience available suggests that there is no significant difference in the
final outcome between various models of the same optimization type, and
make sure that the claimed benefits are due to the optimization software and not
due to other factors, for example:
Select a representative set of operating data covering at least 30-days; the more
complete the data-set, the better; typical number of variables: 100-400; data
provided in 5-10 minute averages reflecting steady-state (not transient) conditions,
State clearly the operating constraints and equipment limitations, so all vendors
abide by the same constraints,
Send the data set to vendors and ask them to provide the optimum settings of the
key control variables, the expected outcome (performance improvements), the
control variables in order of decreasing priority relative to the optimization
objective and the confidence level of the process models developed using the data.
While potentially useful, the blind test could be time-consuming for both the utility and
the vendors. If the plant does not have DCS, collecting such data may be a very timeconsuming activity. Also, evaluation of the results developed by the vendors requires
checking that all the operating constraints have been observed and the claimed
performance improvements can be confirmed by actual field experience.
Some vendors have designed service packages which are intended to assess the
performance improvement potential. The drawback for utilities is that they have to pay
for these assessments (usually priced at $10,000 to $30,000). However, in most cases
these costs are subtracted from future optimization projects. Furthermore, the vendors
may be willing to guarantee performance improvements. Such assessments may be
particularly suitable when a utility has to reduce the number of optimization products
it considers to two and needs more information on all aspects of the products (costs,
benefits, and experience of the vendors team) to make a decision.
5-10
A
SOFTWARE USERS GUIDE
System Requirements
Hardware Requirements
x
A graphics card compatible with Microsoft Windows version 3.0 or later, such as
IBM VGA, EGA, or Hercules graphics cards
At least 2 MB of RAM
Software Requirements
x
Operating system: MS-DOS version 3.1 or later and Microsoft Windows 95 or later
running in standard or 386 enhanced mode.
Installation
1. You must have Microsoft Excel version 7.0 or later installed on your computer to
run the PPO Guidelines tables.
2. The PPO Guidelines disk containing the tables has been checked for viruses with
McAfee VirusScan Version 3.1.9 Deluxe and has been found to be free of viruses.
Due to some differences in Excel 7.0 and Excel 97, two sets of the Excel tables are
provided on the disk, each set in its own folder (subdirectory). The user should select
and install the appropriate version of the Excel tables. In addition to being saved into
separate folders, the Excel 97 files can be differentiated from those in Excel 7.0 by an
E97 added to the name of the Excel file, e.g. Table 3-1E97 is Excel 97 while Table 3-1
is Excel 7.0. To optimize computer performance using the tables, copy the appropriate
tables (e.g. Table 3-1.xls, Table 4-1.xls, Table B-3.xls, Table B-5.xls, Table B-7.xls, Table
A-1
B-8.xls, Table B-9.xls, Table B-11.xls) from the disk (drive A:\) to your hard (usually
drive C:\). This can be accomplished by using Microsoft Windows Explorer:
x
open the folder containing the files saved in the version of Excel installed on
your computer,
select all files, click Ctrl-A (all the files will be highlighted)
click Edit
Go to the new folder you have created on drive C:\ and double click on it
Click Edit
Click Paste
Close the Excel files if you opened them from disk initially and relaunch them as
needed from the copies that now reside on your hard drive.
A-2
To help the user avoid mistakes, all the tables have built in warnings that appear when
a questionable input is entered. When an input outside a reasonable range for a given
cell is entered, a message appears warning the user (In Table 4-1 and the case study
tables based on 4-1 the message Input outside typical range appears in the first
empty cell below the entered input. In Table 3-1 and the case study tables based on it
the message Bad Input appears to the right of the input cell). Likewise, when an
alphanumeric is accidentally entered instead of a numeric value, the tables warn the
user. In Table 4-1 the message Input Value, not Text appears in the first empty cell
below the most recently entered input. In Table 3-1 the message Input Value appears
to the right of the input cell. These warning messages refer to the last input that has
been changed. For either type of error, the tables will not compute final results, costbenefit analysis results in Table 4-1 and Total Score for Table 3-1. The ranges that have
been programmed into the table are broad and should accommodate all plant
optimization scenarios. If for some reason, the user wishes to use inputs outside the
range limitations, he/she should contact EPRI Project Manager and ask for the table
password which is needed to unprotect the table and change the range limitations.
The following provides an overview of each Excel file:
Table 3-1
This table should be used during Step 3 of the guidelines. A print out of this table
appears on page 3-3. Based on site-specific considerations and the guidance provided
in the PPO manual, the user enters the appropriate score from 1 to 3. Using these input
scores, Table 3-1 calculates a Total Score between 100 and 300. Based on the Total
Score, a proposed project will fall into one of three categories (A, B, or C) which
indicates the potential range of improved boiler performance that can be achieved from
boiler optimization. Tables 3-2 and 3-2a to 3-2d on pages 3-24 and 3-25 of the manual
(they are not included on the disk) provide the performance improvements that can be
expected based on the Total Score from Tables 3-1.
Table 4-1
This table is a cost-benefit analysis of the potential boiler optimization project and
should be used during Step 4 of the Guidelines. A print out of the Excel file, appears
on pages 4-5 to 4-6 of the Guidelines. Step 4 of the manual walks the user through the
fine details related to refining the cost and benefit inputs to Table 4-1. Values that can
be changed appear in blue. To assure consistency, when the same value appears more
than once in a table, the value only appears in blue (i.e. is changeable) the first time it
appears. Subsequent appearances are written in black (i.e. are locked) and are linked to
the initial appearance.
A-3
Given the values entered, Table 4-1 automatically calculates the cost-benefit results (see
bottom of the table). The user can examine the formulas used throughout the table in
several different ways:
x
read the Excel math formulas by placing the Excel cursor directly on the
resulting cost-benefit values and looking at the Excel formula bar,
read formula descriptions which are also listed in Appendix C of the manual.
The following tables are based on Table 4-1 and finalize the analyses initiated in the
above options:
x
User Tutorial
Appendix B provides two case studies illustrating how utilities used the PPO
Guidelines to carry out evaluation of different optimization types. To familiarize the
user with how these utilities used the software tables and how the user can utilize the
tables to carry out his/her own evaluation of optimization options for a specific boiler,
this tutorial walks the user through a few changes in the two primary tables, 3-1 and
A-4
4-1. By continuing the same process of changing one input at a time from the default
value in these tables to a value that more accurately reflects the circumstances of the
specific boiler and the financial assumptions of the users company, the tables calculate
the costs and benefits of various options.
While the case studies in Appendix B go into detail regarding technical parameters and
how to determine specific table inputs, this tutorial focuses instead on the mechanical
steps of inputting this data into the tables. For any boiler, the user would first carry out
Step 1 of the Guidelines by undertaking a quick tuning of the boiler and establishing its
baseline performance. The user would then carry out Step 2 by setting optimization
objectives.
Step 3 is the first step in the Guidelines to use a software tool, Table 3-1. The user
would open this table from his/her hard drive (see installation above) and save it
under a different name, e.g. Table 3-1 Boiler XYZ (for example, in Case Study 1, the
table is saved as Table B-3 to easily identify it as the third table referenced in
Appendix B). Then, the user will modify one input at a time to reflect the requirements
and consideration of he specific boiler being evaluated.
Figure A-1 below shows the top and bottom parts of Table 3-1 (the jagged break in
vertical lines represents a jump in the table). The top box in Figure A-1, labeled
Before, shows Table 3-1 as it appears on the Guidelines disk. The bottom box in
Figure A-1, labeled After, shows how it should look after the user has saved it as
Table 3-1 Boiler XYZ and entered new inputs.
Breaking these changes down step-by-step, under the first topic Information on the
last tune up, the first category requiring an input is When was unit last tuned?
Reading from the shaded box presenting this question on page 3-4 of Guidelines
manual, the user will see this question as the first line requiring a score input in Table
3-1.
In this hypothetical case, boiler XYZ was tuned 18 months before which the Guidelines
indicate is a score of 2. In the table, since the default value that appears in blue in the
Score column is set at 1, the user should:
x
delete the other options listed in the cell the right in the Clarifications
column so that the table clearly identifies the specific circumstances of this
boiler.
A-5
In the bottom box of Figure A-1, we see that based solely on this change, the table has
recalculated the Total Score of boiler XYZ, changing from 100 to 130.
Before:
Criteria
Score Clarifications
TOTAL SCORE
1 Options: less than a year (enter 1); 1-2 years (enter 2); more than 2 years (enter 3)
1 Options: safe operation (enter 1); heat rate (enter 2); NOx (enter 3); heat rate & NOx (enter 4)
1 Options: no biasing (enter 1); some biasing, but no BOOS (enter 2); BOOS possible (enter 3)
After:
Criteria
Score Clarifications
2 18 months ago
1 Options: safe operation (enter 1); heat rate (enter 2); NOx (enter 3); heat rate & NOx (enter 4)
1 Options: no biasing (enter 1); some biasing, but no BOOS (enter 2); BOOS possible (enter 3)
TOTAL SCORE
Figure A-1
A Portion of Table 3-1 Before and After a Change is entered
The second category under the first topic is Tuning Objective/Tuned for: (see Figure
A-2 below). In the case of the boiler XYZ, let us assume that the objective was safe
operation which the shaded box on page 3-5 of the Guidelines indicates is a score of 1.
Since the default value in the Score column is already set at 1, the user does not need
to change the score. However, the user may delete the other options in the
Clarifications cell to the right of his/her input and add any comments they he/she
wishes pertaining to the tuning objective.
Continuing on with Step 3, the next topic in Table 3-1 is Unit Operating Flexibility (at
full load). The first category requiring an input is Fuel flow biasing capability, (e.g.
mill capacity). In the case of boiler XYZ, let us assume that some biasing is possible,
but there are no burners-out-of-service (BOOS). The Guidelines indicate this as a score
of 2. The user should:
A-6
As shown in the After box of Figure A-2, with this change, the table has
automatically recalculated the Total Score for the boiler XYZ, changing from 130 to 133.
Before:
Criteria
Score Clarifications
2 18 months ago
1 Options: no biasing (enter 1); some biasing, but no BOOS (enter 2); BOOS possible (enter 3)
TOTAL SCORE
1 Options: safe operation (enter 1); heat rate (enter 2); NOx (enter 3); heat rate & NOx (enter 4)
After:
Criteria
Score Clarifications
2 18 months ago
1 Safe operation
TOTAL SCORE
Figure A-2
A Portion of Table 3-1 Before and After a Second Input is Entered
By continuing on through all the topics and categories listed in Step 3 in this manner,
the user will tailor every row in the table to the specific boiler they are evaluating,
thereby completing the table. Based on the final Total Score, a proposed project will
fall into one of three categories (A, B, or C) which indicates the potential range of
improved boiler performance that can be achieved from boiler optimization.
In the hypothetical case of boiler XYZ, let us assume a final Total Score of 170 so that
we can see how the user would use this result from Table 3-1 to how he/she would use
Table 4-1. With a Total Score of 170, Table 3-2 on page 3-24 classifies the project as
Group B. Given the previous tuning objective of safe operation, Table 3-2a (on page
3-25) indicates the range of potential improvements in heat rate and NOx reduction.
The low end of the range can be reached using stand-alone optimization, the middle
end of the range by on-line/advisory, and the high end of the performance
improvement range can be met with closed-loop type optimization.
A-7
To determine the cost of reaching these improvements, the user should use Table 4-1 to
carry out a cost-benefit analysis. The cost for each type of optimization can then be
weighed against that projects estimated performance improvements. The mechanics of
using Table 4-1 are the same as for Table 3-1. The user should open the file from
his/her hard drive, save it under a different name, and then work step-by-step through
the shaded boxes in Step 4 of the Guidelines manual to tailor all the inputs to the boiler
under consideration and the financial assumptions of his/her company.
Figure A-3 below shows the top and bottom parts of Table 4-1. The first box in Figure
A-3, labeled Before, shows Table 4-1 as it appears on the Guidelines disk. The
second box in Figure A-3, labeled After, shows how it should look after the user has
saved it as Table 4-1 Boiler XYZ and entered new inputs.
Breaking these change down step-by-step, under the first topic Up Front Costs, the
first category requiring an input is Up Front Software License Fees. Reading from
the shaded box presenting this cost category on page 4-7 of Guidelines manual, the user
will see this category as the first line in Table 4-1 with a value that appears in blue, i.e.
it is changeable.
A-8
Before:
COSTS
Optimization Types
Stand-alone
On-line/advisory
Closed-loop
Up Front Costs
Up Front Software License Fees ($)
$30,000
$130,000
$170,000
$5,000
$0
$5,000
$5,000
$0
$5,000
$5,000
$0
$5,000
Computer software
Computer hardware
$68,484
26
26.0
1.0
$492,377
66
13.9
2.8
$1,277,134
134
7.8
5.7
After:
COSTS
Optimization Types
Stand-alone
On-line/advisory
Closed-loop
$20,000
$130,000
$170,000
$5,000
$0
$5,000
$5,000
$0
$5,000
$5,000
$0
$5,000
$78,484
34
22.1
1.4
$492,377
66
13.9
2.8
Up Front Costs
Up Front Software License Fees ($)
Additional Computer Software and Hardware ($)
Computer software
Computer hardware
$1,277,134
134
7.8
5.7
Figure A-3
A Portion of Table 4-1 Applied to Boiler XYZ; Before Use and After a Change is
Entered
For boiler XYZ, let us assume that the utility is considering application of optimization
to several boilers at the same time and that a optimization vendor that sells stand alone
optimization has offered a reduced up front software fee of $20,000 if applied to several
boilers. Since this amount is lower than default value in the Stand alone column, the
user should:
x
If the user has any comments to add to regarding this line item, the user can
input comments to the right of the optimization types columns (off the edge
of A-3).
A-9
In Figure A-3, all four measurements of project value listed under Cost-Benefit
Analysis Results have now been recalculated in the Stand-alone column to reflect this
improvement in price.
Continuing on with Step 4 (see Figure A-4), the next cost category is Additional
Computer Software and Hardware ($). In the case of boiler XYZ, let us assume that
the utility has an older computer system that is inadequate to deal with the processing
requirements of any of the optimization types and that a total of $10,000 is required to
purchase new computer equipment.
In Figure A-4, we see that Table 4-1 provides a default value of $5,000 for sub-category
Computer hardware. Since the case of boiler XYZ requires a greater investment, the
user should:
x
By doing so, the user will note that the table automatically recalculates the total cost for
Additional Computer Software and Hardware ($), increasing it from $5,000 to
$10,000 (appears in bold which above the Computer hardware sub-category). Since
Computer hardware is the only sub-category with a value entered, the Additional
Computer Software and Hardware ($) total equals the total computer hardware entry.
The user must repeat these changes in the cells for computer hardware in the columns
for On-line/advisory and Closed-loop as well. In the electronic version of the
table, the user will note that the values in these cells appear in blue (i.e. are changeable)
and are not linked to each other. The table is designed in this way so that varying
requirements for different optimization types can be entered and cost-benefit results
calculated accurately for each. The user should:
x
repeat these steps in the Computer hardware cell row in the Closed-loop
column
As shown in the bottom box of Figure A-4, these changes have automatically been used
to recalculate the project value of the three optimization types. In this case, since the
A-10
cost of all three optimization types was increased from the default values, the project
values presented under Cost-Benefit Analysis Results have decreased. The projected
numbers for of Net Present Value, Return on Investment, and Benefit-Cost Ratio have
all dropped somewhat. The number for Payback Period has gone up, indicating a
longer period required for recouping the initial investment and consistent with a lower
project value.
Before:
COSTS
Optimization Types
Stand-alone
On-line/advisory
Closed-loop
$20,000
$130,000
$170,000
$5,000
$0
$5,000
$5,000
$0
$5,000
$5,000
$0
$5,000
$78,484
34
22.1
1.4
$492,377
66
13.9
2.8
Up Front Costs
Up Front Software License Fees ($)
Additional Computer Software and Hardware ($)
Computer software
Computer hardware
$1,277,134
134
7.8
5.7
After:
COSTS
Optimization Types
Stand-alone
On-line/advisory
Closed-loop
$20,000
$130,000
$170,000
$10,000
$0
$10,000
$10,000
$0
$10,000
$10,000
$0
$10,000
$73,484
30
24.0
1.2
$487,377
64
14.3
2.6
$1,272,134
130
8.0
5.5
Up Front Costs
Computer software
Computer hardware
Figure A-4
A Portion of Table 4-1 Applied to Boiler XYZ; After a Second Group of Changes is
Entered
By continuing on through all the categories and sub-categories of costs and benefits
listed in Step 4 in this manner, the user will tailor every row in the table to the specific
boiler they are evaluating, thereby completing the table. With an estimate of the
potential performance improvements of each optimization type and an estimate of the
value of each optimization type, the utility will be well prepared to make a decision
regarding optimization on their boiler units.
A-11
The user is advised to review the case studies in Appendix B to familiarize him/herself
with the Guidelines before applying them to his/her own project.
A-12
B
CASE STUDIES
Armstrong 1 and 2 retrofitted with Foster Wheeler's IFS burners and overfire
air
Harrison units 1, 2, and 3 retrofitted with Foster Wheeler's IFS burners and
overfire air
These projects were driven mainly by the need to comply with the requirements of
Phase I of the Clean Air Act Amendment (CAAA). However since 1995, Allegheny has
focused its efforts in assessing the potential for further NOx emission reductions and
performance improvements utilizing power plant optimization software. Such
improvements are important in fine-tuning its future NOx compliance plans (CAAA
Title I and Phase II/ Title IV) and increasing its competitiveness in the postderegulation era.
As of May 1998, Allegheny has utilized optimization software in four power stations.
As shown in Table B-1, the results have been very encouraging. Simultaneous NOx
reduction and heat rate improvement was achieved in all units by utilizing
optimization software in stand-alone mode.
B-1
Table B-1
Allegheny's Experience with Power Plant Optimization
Station Name
Burner Type
Front-wall
Optimization
Software Used
Ultramax
% NOx Reduction
Achieved
20
% Heat Rate
Improvement
0.5
Harrison 2
Hatfield 1-3
Cell burners
Ultramax
0.3
Ft. Martin 1
T-fired
Ultramax
17
1.5
Smith 4
T-fired
Boiler OP
NA (ongoing)
NA (ongoing)
the number of control variables in this boiler may be much higher that in the
units shown in Table B-1; this is mainly due to the fact that coal flow rate
may be monitored in each of the 12 coal pipes;
the potential for NOx emission trading requires increased flexibility and online optimization;
Such assessment was carried out using a beta version of the EPRI Power Plant
Optimization Guidelines. Based on the outcomes of this assessment, Allegheny is now
selecting specific software through competitive solicitation (note: this step is not
included in the PPO Guidelines).
Armstrong Unit 1
Armstrong 1 and 2 are front wall-fired boilers rated at 180 MWs each designed to burn
Eastern Bituminous coal (see Table B-2). They were originally designed and built by
Foster Wheeler in the mid-1950s, but were refurbished in the early 1990s (see Figure
B-1). Refurbishment included replacement of pressure parts and installation of Foster
Wheeler's IFS low NOx burners.
B-2
Table B-2
Characteristics of Eastern Bituminous Coal Burned in Armstrong 1
Grindability
55
7.80
Volatile matter
31.82
Fixed carbon
49.05
Ash
11.33
1283/1378
Ash
11.33
Sulfur
1.94
H2
5.22
66.54
H2O
7.80
N2
1.35
O2
5.82
KJ/Kg. (Btu/lb)
27,928 (12,007)
Each unit utilizes two ball mills (see Figure B-2) feeding 12 burners which are arranged
in three rows along the front wall of the boiler. Significant imbalance has been
measured among coal pipes resulting in increased LOI (Loss On Ignition ). In most
cases, LOI averages at 12-14%. Plant engineers estimate that LOI could be reduced to
the 5-6% range with better coal flow balancing and unit optimization. Coal flow
balancing is expected to be achieved through the utilization of an electrostatic or
microwave-based coal flow measurement technique and control of primary air flow
rate. Depending on the effectiveness of coal flow balancing and optimization, a coal
additive may be used, too.
B-3
Figure B-1
Armstrong Units 1 & 2 Boiler Arrangement
B-4
Burners
12
11
10
11
10
12
Coal
Conduits
Coal
Conduits
3 way distributor
Ball mill
pulverizer A
Ball mill
pulverizer B
Figure B-2
Burner and Mill Arrangement
After the IFS burner installation, Armstrong 1 operates as an intermediate load unit and
generates an average 300 ppm (0.40 lbs/MBtu) of NOx emissions; as Figure B-3 shows,
NOx as low as 225 ppm (0.30 lbs/MBtu) and as high as 360 ppm (0.48 lbs/MBtu) have
been measured. Such performance satisfies present CAAA requirements, but further
tightening of NOx regulations is likely. For this reason, Allegheny is interested to
assess the potential for NOx reduction at Armstrong to assess its NOx compliance
options including system averaging.
As Figure B-4 shows, boiler efficiency varies considerably suggesting that heat rate
improvement of the order of 1% is feasible. However, such improvement may have to
be balanced against potential LOI and NOx emission reductions.
B-5
375
0.5
300
0.40
0.35
0.30
N O x (ppm )
N O x (Lb/M btu)
0.45
225
0.25
0.20
150
50
150
100
200
Lo ad (G M W )
Boiler Efficiency
Boiler Efficiency
Figure B-3
Armstrong Unit 1 NOx Emissions After IFS Burner Installation
Load (G MW )
Figure B-4
Armstrong Unit 1 Boiler Efficiency After IFS Burner Installation
B-6
While NOx emission reduction was not the primary objective, it was recognized that
NOx trading might increase the importance of NOx reduction in the future. To reflect
this, the value of NOx emissions for Allegheny was determined at an average of $400
per ton. This was based on average compliance costs and the potential value of NOx if
a trading system is established.
Other parameters, which relate to the objectives and were set at this step include:
x
Incremental cost for replacement power: $20 per MWh; this was important
because Armstrong 1 is derated by 10 MWs due to opacity limits
approximately 30% of the time;
Value of SO2 emissions: $100 per ton; note: baseline SO2 emissions were
473 ppm (1.1 lbs/MBtu);
Case 1: with monitoring of coal flow in each coal pipe and control of
imbalance through modulation of the secondary air sleeve dampers, and
For each case, the electronic version of Table 3-1 of the Guidelines was used to
determine the potential performance improvement.
Case 1: Monitoring and controlling coal flow
Table B-3 is the result of tailoring Table 3-1 to the specific circumstances of Armstrong
1. The following bullets summarize the key observations and assumptions:
x
The combustion system was tuned by Foster Wheeler within the last year.
Tuning objective was NOx emission reduction
Some air flow biasing (more air to the upper elevation and lower air to the
lower elevation of burners) is possible
Re-orificing of coal pipes, monitoring of coal flow and control of primary air
are included in this option.
As shown in Table B-3, the Armstrong optimization project scores 168 points (out of 300
points) which Table 3-2 on page 3-24 of the Guidelines manual classifies as a Group B
project. Given the fact that the previous tuning (within the last year) focused on NOx
emission reduction, Table 3-2c on page 3-25 provides an even more refined estimate of
potential performance improvements
B-8
Table B-3
Project Classification to Determine Potential Performance Improvements Due to Optimization
Armstrong 1/Advanced Control for Coal Flow Distribution
Criteria
Information on the last tune up
- When was the unit last tuned?
- Tuning Objective/Tuned for:
Unit operating plexibility (at full load)
- Fuel flow biasing capability (e.g., mill capacity)
- Air flow biasing capability
- Excess O2 range (min-max)
- Burner tilts (for T-fired units)
- Air and Gas dampers
- PA/Fuel ratio
- Other control variables
Ability to change equipment settings
- Burner settings (e.g., registers, yaw, etc.)
- Pulverizer settings (e.g., spring tension; classifier; exit temp
- Other equipment settings
Hardware modifications
- Air distribution modifications
- Coal pipe orificing
- Mill Modifications
- Other
TOTAL SCORE (weighted average)
Score
Clarifications
B-9
NOx Reduction:
15.0-22.5%, and
Within these ranges, the projected improvements for the various types of optimization
are shown in Table B-4. (Note: it is assumed that closed-loop will achieve the upper
end of the projected range, the stand-alone will achieve the lower end and the online/advisory the average).
Table B-4
Expected Performance Improvements with Optimization (Case 1)
Optimization Type
Stand-alone
15.00
0.750
On-Line/Advisory
18.75
1.000
Closed-loop
22.50
1.250
Also, it was estimated that the stand-alone option could regain 5 out of the 10 MWs lost
due to opacity limitations, while on-line/advisory and closed-loop could regain all 10
MWs.
Case 2: No Coal Flow Monitoring and Control
Case 2 is different from Case 1 in that no re-orificing of coal pipes and monitoring of
coal flow are included in this option. As a result (see Table B-5), the Total Score is
lower (153 out of 300 points). Table 3-2 on page 3-24 classifies this optimization project
as Group A and Table 3-2c on page 3-25 provides an even more refined estimate of
potential performance improvements:
x
NOx Reduction:
5.0-10.0%, and
Within these ranges, the projected improvements for the various types of optimization
are shown in Table B-6
B-10
Table B-5
Project Classification to Determine Potential Performance Improvements due to Optimization
Armstrong 1/ No Coal Flow Distribution Control
Criteria
Information on the last tune up
- When was unit last tuned?
- Tuning Objective/Tuned for:
Unit operating flexibility (at full load)
- Fuel flow biasing capability (e.g., mill capacity)
- Air flow biasing capability
- Excess O2 range (min-max)
- Burner tilts (for T-fired units)
- Air and Gas dampers
- PA/Fuel ratio
- Other control variables
Ability to change equipment settings
- Burner settings (e.g., registers, yaw, etc.)
- Pulverizer settings (e.g., spring tension; classifier; exit temp
- Other equipment settings
Hardware modifications
- Air distribution modifications
- Coal pipe orificing
- Mill Modifications
- Other
TOTAL SCORE (weighted average)
Score
Clarifications
B-11
Table B-6
Expected Performance Improvements with Optimization (Case 2)
Optimization Type
Stand-alone
5.0
0.250
On-Line/Advisory
7.5
0.375
Closed-loop
10.0
0.500
All the up front and annual costs, including in-house technical support
needed to install, calibrate, and maintain the software
Benefits such as fuel savings due to heat rate improvement, increase unit
output (regaining lost MWs), and reduction of NOx and SO2 emissions.
Based on this analysis, closed-loop is the most cost-effective option for Armstrong 1.
This option is also the most appropriate considering the relatively high number of
control variables (a total of 20-30 including the coal flow rate or primary air in 12 coal
pipes). With such high number of variables, stand-alone optimization would be timeconsuming and impractical.
Case 2
The cost-benefit analysis and the results for Case 2 (no coal flow distribution control)
are shown in Table B-8 indicating again that the most cost-effective option for
Armstrong 1 is the closed-loop optimization.
B-12
Optimization Types
Stand-alone
O n-line/advisory
Closed-loop
$70,000
$130,000
$170,000
$5,000
$0
$5,000
$0
$0
$5,000
$0
$5,000
$0
$0
$5,000
$0
$5,000
$0
$0
$57,000
$25,000
25
$800
$5,000
$32,000
40
$400
2.0
$0
$46,600
$14,600
12
$800
$5,000
$32,000
40
$400
2.0
$0
$53,000
$17,000
15
$800
$5,000
$36,000
45
$400
2.0
$0
Training ($)
$0
$0
$0
$0
$0
$0
$132,000
$181,600
$228,000
$5,000
$13,000
$17,000
$33,000
$17,000
15.0
$800
$5,000
$16,000
20
$400
2.0
$25,000
$13,000
10.0
$800
$5,000
$12,000
15
$400
2.0
$21,000
$13,000
10.0
$800
$5,000
$8,000
10
$400
2.0
$38,000
$38,000
$38,000
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
$0
$0
$0
$0
$88,080
10,000
0.750
1.12
171
70.0
$117,440
10,000
1.000
1.12
171
70.0
$146,800
10,000
1.250
1.12
171
70.0
Up Front Costs
Up Front Software License Fees ($)
Additional Computer Software and Hardware ($)
Computer software
Computer hardw are
Instrumentation and controls (I&C)
Data acquisition system (DAS)
BENEFITS (Losses)
Up Front (one time) Benefits
NPV of Deferred Costs ($)
Deferred costs
How long does optimization delay investment? (in months)
Annual Benefits
Annual Avoided Costs ($/yr)
Fuel Cost Savings ($/yr)
Baseline heat rate (Btu/kWh)
Unit heat rate improvement (% Pts)
Fuel cost ($/MBtu)
Unit output (MW)
Capacity factor (%)
B-13
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
0.0
0
$0
0.0
0
$0
0.0
0
$183,960
5
20.00
70
30
$367,920
10
20.00
70
30
$367,920
10
20.00
70
30
$0
$141,557
400
353.9
0.450
15.00
171
10,000
70
$0
$176,947
400
442.4
0.450
18.75
171
10,000
70
$0
$212,336
400
530.8
0.450
22.50
171
10,000
70
$0
$4,325
100
43.25
1.100
10,000
0.75
171
70
$0
$5,767
100
57.67
1.100
10,000
1.00
171
70
$0
$7,209
100
72.09
1.100
10,000
1.25
171
70
CO Emissions
Avoided fines/penalties ($/yr)
$0
$0
$0
Opacity
Avoided fines/penalties ($/yr)
$0
$0
$0
$0
0
8,199
208.52
10,000
0.75
171
70
$0
0
10,932
208.52
10,000
1.00
171
70
$0
0
13,666
208.52
10,000
1.25
171
70
$417,923
0.50
$668,074
0.75
$734,265
0.90
$208,961
$501,055
$660,838
5.0
5.0
5.0
10.6
10.6
10.6
$170,961
$638,265
$26,400
$463,055
$1,728,766
$36,320
$622,838
$2,325,298
$45,600
$506,265
110
9.3
3.8
$1,547,166
235
4.7
8.5
$2,097,298
253
4.4
9.2
CO2 Emissions
Annual value of CO2 emission credits ($/yr)
-Value of CO2 credits ($/ton removed)
CO2 emissions reduction (tons/yr)
-Baseline CO2 emissions (lbs/MBtu)
-Baseline unit heat rate (Btu/kWh)
-Unit heat rate improvement (%)
-Unit output (MW )
-Capacity factor (%)
Annual benefits before diminishing effectiveness factor ($/yr)
Diminishing effectiveness factor
INTERMEDIARY FIGURES
Net annual benefits ($)
PV of future net annual benefits ($)
Annual depreciation (straightline)
B-14
Table B-8
Power Plant Optimization Cost-Benefit Analysis--Armstrong 1
Without Coal Distribution Control
COSTS
Optimization Types
Stand-alone
On-line/advisory
Closed-loop
$50,000
$130,000
$170,000
$5,000
$0
$5,000
$0
$0
$5,000
$0
$5,000
$0
$0
$5,000
$0
$5,000
$0
$0
$34,600
$14,600
12
$800
$5,000
$20,000
25
$400
2.0
$0
$46,600
$14,600
12
$800
$5,000
$32,000
40
$400
2.0
$0
$53,000
$17,000
15
$800
$5,000
$36,000
45
$400
2.0
$0
Training ($)
$0
$0
$0
$0
$0
$0
$89,600
$181,600
$228,000
$5,000
$13,000
$17,000
$33,000
$17,000
15.0
$800
$5,000
$16,000
20
$400
2.0
$25,000
$13,000
10.0
$800
$5,000
$12,000
15
$400
2.0
$21,000
$13,000
10.0
$800
$5,000
$8,000
10
$400
2.0
$38,000
$38,000
$38,000
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
$0
$0
$0
$0
$29,360
10,000
0.250
1.12
171
70.0
$44,040
10,000
0.375
1.12
171
70.0
$58,720
10,000
0.500
1.12
171
70.0
Up Front Costs
Computer software
Computer hardware
Instrumentation and controls (I&C)
Data acquisition system (DAS)
BENEFITS (Losses)
Up Front (one tim e) Benefits
NPV of Deferred Costs ($)
Deferred costs
How long does optimization delay investment? (in months)
Annual Benefits
Annual Avoided Costs ($/yr)
Fuel Cost Savings ($/yr)
Baseline heat rate (Btu/kWh)
Unit heat rate improvement (% Pts)
Fuel cost ($/MBtu)
Unit output (MW)
Capacity factor (%)
B-15
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
0.0
0
$0
0.0
0
$0
0.0
0
$183,960
5
20.00
70
30
$367,920
10
20.00
70
30
$367,920
10
20.00
70
30
NOx Emissions
Avoided fines/penalties ($/yr)
Annual value of NOx emission credits ($/yr)
-Value of NOx emission credits ($/ton removed)
NOx emission credits (tons/yr)
-NOx reference level for credits (lbs/MBtu)
-NOx emission reduction below reference level (%)
-Unit output (MW )
-Unit baseline heat rate (Btu/kWh)
-Capacity factor (%)
$0
$47,186
400
118.0
0.450
5.00
171
10,000
70
$0
$70,779
400
176.9
0.450
7.50
171
10,000
70
$0
$94,371
400
235.9
0.450
10.00
171
10,000
70
SO2 Emissions
Avoided fines/penalties ($/yr)
Annual value of SO2 emission credits ($/yr)
-Value of SO2 emission credits ($/ton removed)
SO2 emissions reduction (tons/yr)
-Baseline SO2 emissions (lbs/MBtu)
-Baseline unit heat rate (Btu/kW h)
-Unit heat rate improvement (%)
-Unit output (MW )
-Capacity factor (%)
$0
$1,442
100
14.42
1.100
10,000
0.25
171
70
$0
$2,163
100
21.63
1.100
10,000
0.38
171
70
$0
$2,884
100
28.84
1.100
10,000
0.50
171
70
CO Emissions
Avoided fines/penalties ($/yr)
$0
$0
$0
Opacity
Avoided fines/penalties ($/yr)
$0
$0
$0
$0
0
2,733
208.52
10,000
0.25
171
70
$0
0
4,100
208.52
10,000
0.38
171
70
$0
0
5,466
208.52
10,000
0.50
171
70
$261,948
0.50
$484,901
0.75
$523,895
0.90
$130,974
$363,676
$471,506
5.0
5.0
5.0
10.6
10.6
10.6
$92,974
$347,107
$17,920
$325,676
$1,215,875
$36,320
$433,506
$1,618,445
$45,600
$257,507
84
11.6
2.9
$1,034,275
159
6.7
5.7
$1,390,445
170
6.3
6.1
CO2 Emissions
Annual value of CO2 emission credits ($/yr)
-Value of CO2 credits ($/ton removed)
CO2 emissions reduction (tons/yr)
-Baseline CO2 emissions (lbs/MBtu)
-Baseline unit heat rate (Btu/kW h)
-Unit heat rate improvement (%)
-Unit output (MW)
-Capacity factor (%)
Annual benefits before diminishing effectiveness factor ($/yr)
Diminishing effectiveness factor
INTERMEDIARY FIGURES
Net annual benefits ($)
PV of future net annual benefits ($)
Annual depreciation (straightline)
B-16
Sherco 2 (a 725 MW T-fired boiler): retrofitted with low NOx burners and
overfire air
B-17
While these projects are expected to satisfy CAAA requirements, both at the unit and
power system level, NSP would like to have some additional means for controlling
NOx emissions as an "insurance policy" for unexpected events which can affect NOx
compliance. Such events include:
x
NOx reduction at King and Riverside 8 may be less than expected, especially
if the units experience waterwall corrosion problems due to reducing
atmosphere in the cyclone boilers; and
so far, CAAA Title I has not imposed any requirements which can not be met
with NSPs latest NOx compliance program, but it is possible that these
requirements may be more stringent in the future.
For this reason, power plant optimization is being considered for a number of units
including Riverside 7, Sherco 3, and Black Dog 3 and 4.
Description of Riverside 7
Riverside 7 is a 153 MW coal-fired unit consisting of two wall-fired boilers and one
steam turbine. The wall-fired boilers were built by Babcock & Wilcox in the 1950s and
burn subbituminous coal. The unit has two pulverizers feeding six burners arranged in
two rows (four burner in the bottom row and two in the top). Also, it is equipped with
a baghouse and a Westinghouse (WDPF) Digital Control System (DCS).
The unit is used for intermediate load, but recently has been operating more as
baseloaded. Full load NOx emissions average 675 ppm (0.90 lbs/MBtu). LOI averages
less than 1%, but in recent years has been highly variable. Occasionally, this variability
is not acceptable to buyers of the flyash and requires that it be trucked 53 miles (90 km
from Riverside) to Sherco for disposal.
B-18
included in this paper). The 5 steps outlined in the Guidelines are described in the
following paragraphs.
Step 1: Quick Tuning and Establishment of Baseline Performance
The first step was a quick walk-down/checking of the operating condition of key plant
components and control variable set-points. This review reveal the following:
x
excess O2 is set at 2.3%; most likely, the boiler can operate at a lower level, but
for safety purposes and because of the absence of a CO monitor, 2.3% has
been set as a minimum operating level
the air registers are set visually by operations. During this quick tuning
exercise, the burners were found to be set uniformly (in a scale of 1 to 9, the
registers were set at 6)
The first operating adjustment made was to change the settings of the air registers as
follows:
x
The result of these changes was an immediate NOx reduction of approximately 8%.
Following this quick tuning, baseline NOx emissions were measured at approximately
600 ppm (0.80 lbs/MBtu).
In addition, the following steps were identified to improve performance in order of
priority:
1. Improvement of coal fineness through mill adjustments
2. Air and coal flow balancing including biasing of air towards the upper
elevation of burners
3. Application of optimization software
B-19
Present Arrangement
Mill 2
Mill 1
Modification
Mill 2
Mill 2
Mill 1
Mill 1
Alternative 1
Alternative 2
Figure B-5
Alternative Coal Pipe Arrangements for Improved NOx Control
B-20
No systematic tuning of the boilers had been performed in the last two years
Excess air could be reduced from 2.3% to 2.0%; further reduction may be
possible, but a CO monitor needs to be added
Switching of two coal pipes to allow for fuel biasing is being included.
As shown in Table B-9, the Riverside optimization project scores 191 points (out of 300
points) which Table 3-2 on page 3-24 of the Guidelines manual classifies as a Group B
project. Given the fact that the previous tuning was done more than two years ago and
EPRI's "Retrofit NOx Controls for Coal-fired Utility Boilers-1996 Update Addendum" (TR-102906-A
dendum, May 1997), pp. I-11 and 6-2.
B-21
focused on heat rate improvement, Table 3-2b on page 3-25 provides an even more
refined estimate of potential performance improvements:
B-22
NOx reduction:
22.5-30.0%
Table B-9
Project Classification to Determine Potential Performance Improvements Due to OptimizationRiverside 7
Criteria
Information on the last tune up
- When was the unit last tuned?
- Tuning Objective/Tuned for:
Unit operating flexibility (at full load)
- Fuel flow biasing capability (e.g., mill capacity)
- Air flow biasing capability
- Excess O2 range (min-max)
- Burner tilts (for T-fired units)
- Air and Gas dampers
- PA/Fuel ratio
- Other control variables
Ability to change equipment settings
- Burner settings (e.g., registers, yaw, etc. etc.)
- Pulverizer settings (e.g., spring tension; classifier; exit temp
- Other equipment settings
Hardware modifications
- Air distribution modifications
- Coal pipe orificing
- Mill Modifications
- Other
TOTAL SCORE (weighted average)
Score
Clarifications
B-23
The projections for the various types of optimization are shown in Table B-10. (Note: it
is assumed that closed-loop will achieve the upper end of the projected range, the
stand-alone will achieve the lower end and the on-line/advisory the average).
Table B-10
Expected Performance Improvements with Optimization (Case 1)
Optimization Type
Stand-alone
22.50
0.500
On-Line/Advisory
26.25
0.625
Closed-loop
30.00
0.750
All the up front and annual costs, including in-house technical support
needed to install, calibrate and maintain the software,
Benefits such as fuel savings due to heat rate improvement, and reduction of
NOx and SO2 emissions.
Based on this analysis, closed-loop is the most cost-effective option for Riverside 7.
B-24
Table B-11
Power Plant Optimization Cost-Benefit AnalysisRiverside 7
COSTS
Optimization Types
Stand-alone
On-line/advisory
Closed-loop
$20,000
$130,000
$170,000
$5,000
$0
$5,000
$0
$0
$5,000
$0
$5,000
$0
$0
$5,000
$0
$5,000
$0
$0
$33,000
$14,600
12
$800
$5,000
$18,400
25
$320
2.3
$0
$44,040
$14,600
12
$800
$5,000
$29,440
40
$320
2.3
$0
$50,120
$17,000
15
$800
$5,000
$33,120
45
$320
2.3
$0
Training ($)
$0
$0
$0
$0
$0
$0
$58,000
$179,040
$225,120
$8,000
$13,000
$17,000
$31,720
$17,000
15.0
$800
$5,000
$14,720
20
$320
2.3
$19,040
$8,000
10.0
$800
$0
$11,040
15
$320
2.3
$15,360
$8,000
10.0
$800
$0
$7,360
10
$320
2.3
$39,720
$32,040
$32,360
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
0
$0
$0
$0
$0
$0
$0
$48,036
10,240
0.500
1.00
153
70.0
$60,045
10,240
0.625
1.00
153
70.0
$72,053
10,240
0.750
1.00
153
70.0
Up Front Costs
Computer software
Computer hardware
Instrumentation and controls (I&C)
Data acquisition system (DAS)
BENEFITS (Losses)
Up Front (one time) Benefits
NPV of Deferred Costs ($)
Deferred costs
How long does optimization delay investment? (in months)
Annual Benefits
Annual Avoided Costs ($/yr)
Fuel Cost Savings ($/yr)
Baseline heat rate (Btu/kWh)
Unit heat rate improvement (% Pts)
Fuel cost ($/MBtu)
Unit output (MW)
Capacity factor (%)
B-25
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
0.0
0
$0
0.0
0
$0
0.0
0
$0
0
0.00
70
0
$0
0
0.00
70
0
$0
0
0.00
70
0
$0
$164,282
190
864.6
0.800
22.50
153
10,240
70
$0
$191,662
190
1008.7
0.800
26.25
153
10,240
70
$0
$219,042
190
1152.9
0.800
30.00
153
10,240
70
$0
$0
0
28.82
1.200
10,240
0.50
153
70
$0
$0
0
36.03
1.200
10,240
0.63
153
70
$0
$0
0
43.23
1.200
10,240
0.75
153
70
CO Emissions
Avoided fines/penalties ($/yr)
$0
$0
$0
Opacity
Avoided fines/penalties ($/yr)
$0
$0
$0
$0
0
5,008
208.52
10,240
0.50
153
70
$0
0
6,260
208.52
10,240
0.63
153
70
$0
0
7,512
208.52
10,240
0.75
153
70
$212,318
0.50
$251,707
0.75
$291,096
0.90
$106,159
$188,780
$261,986
5.0
5.0
5.0
4.9
4.9
4.9
$66,439
$288,442
$11,600
$156,740
$680,482
$35,808
$229,626
$996,916
$45,024
$230,442
95
10.5
4.0
$501,442
68
13.7
2.8
$771,796
82
11.8
3.4
CO2 Emissions
Annual value of CO2 emission credits ($/yr)
-Value of CO2 credits ($/ton removed)
CO2 emissions reduction (tons/yr)
-Baseline CO2 emissions (lbs/MBtu)
-Baseline unit heat rate (Btu/kW h)
-Unit heat rate im provement (%)
-Unit output (MW )
-Capacity factor (%)
Annual benefits before diminishing effectiveness factor ($/yr)
Diminishing effectiveness factor
INTERMEDIARY FIGURES
Net annual benefits ($)
PV of future net annual benefits ($)
Annual depreciation (straightline)
B-26
Quick tuning which should be carried out in Step 1 has the potential to reduce
NOx emissions by approximately 10-15%. Such tuning involves mainly:
Resetting the air registers in such a way that air biasing towards the top
elevation of burners is created,
Tuning of the mills to improve coal fineness, and
Addition of CO monitor to reduce the excess O2 level ,
Selection of specific power plant optimization software will be carried out by Northern
States Power following the guidance provided in Step 5 of the Guidelines manual.
Such evaluation is not described here because of the proprietary nature of the
information included.
The projections for the various types of optimization are shown in Table B-10. (Note: it
is assumed that closed-loop will achieve the upper end of the projected range, the
stand-alone will achieve the lower end and the on-line/advisory the average).
B-27
C
FORMULAS
C-1
SO2 Emissions
Annual SO2 reduction (tons/yr) = Baseline SO2 emissions (lbs/MBtu) x Baseline heat
rate (Btu/kWh) x Heat rate improvement (%) x Unit output (MW) x Unit capacity factor
(%) x (876 x 0.5/109)
Value of SO2 emission credits ($/yr) = Annual SO2 reduction (tons/yr) x Value of SO2
credits ($/ton of SO2 removed)
CO2 Emissions
Annual CO2 reduction (tons/yr) = Baseline CO2 emissions (lbs/MBtu) x Unit heat rate
(Btu/kWh) x heat rate improvement (%) x Unit output (MW) x Unit capacity factor (%)
x Pounds to million tons conversion factor (876 X 0.5/109)
Value of CO2 credits ($/yr) = Annual CO2 reduction (tons/yr) x Value of CO2 credits
($/ton of CO2 removed)
C-2