12th Heat Rate Performance Improvement Conference

EPRI's Twelfth Heat Rate Improvement
Conference Proceedings
Proceedings
January 15-19, 2001 New Orleans, Louisiana The Hotel Inter-Continental New Orleans
EPRIs Twelfth Heat Rate

Improvement Conference
Proceedings
1001328
Proceedings, January 2001
EPRI Project Manager

J. Stallings
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
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN
ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH
INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE
ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:
(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)
WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR
SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR
INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL
PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER
(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR
SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
ORGANIZATION(S) THAT PREPARED THIS DOCUMENT
Electric Power Research Institute
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, (800) 313-3774.
Electric Power Research Institute and EPRI are registered service marks of the Electric Power
Research Institute, Inc. EPRI. ELECTRIFY THE WORLD is a service mark of the Electric Power
Research Institute, Inc.
Copyright 2001 Electric Power Research Institute, Inc. All rights reserved.
CITATIONS
These proceedings were compiled by
EPRI
3412 Hillview Avenue
Palo Alto, California 94303
This report describes research sponsored by EPRI.
The report is a corporate document that should be cited in the literature in the following manner:
EPRI's Twelfth Heat Rate Improvement Conference Proceedings, EPRI, Palo Alto, CA: 2001.
1001328.
iii
REPORT SUMMARY
The Twelfth Heat Rate Improvement Conference, sponsored by EPRIs Heat Rate and Cost
Optimization Value Package, is the latest in a series of meetings designed to assist utilities in
addressing problems with power plant performance and in identifying cost-effective solutions for
achieving and maintaining heat rate improvement. The previous conference was held in
Baltimore in September 1998.
Background
Deregulation in the utility industry has forced power plants to lower their costs of generating
electricity to become more competitive. Since the cost of fuel for coal-fired plants accounts for
60-80% of the overall cost of electricity, improvements in heat rate are at the forefront of these
cost-cutting efforts. In the long run the lowest-cost generators will be the ones that dominate the
power industry.
Objective
To summarize current efforts by EPRI and others to improve the heat rate of fossil-fired power
plants, including optimization, intelligent sootblowing, and heat rate performance and
monitoring.
Approach
EPRIs Heat Rate and Cost Optimization Value Package hosted a conference January 30 to
February 1, 2001 in Dallas, Texas. The conference was divided into six technical sessions, with
three additional panels designed to investigate individual topics in more depth. Panel topics
were:
How are utilities using heat rate information?
Why test?
What improvements in heat rate information are needed?
Results
Areas addressed in the individual sessions include:
The effectiveness and usefulness of On-Line Heat Rate Monitors
The trend for Optimization software tools to use heat rate as an input into total plant cost
minimization efforts
The potential for incorporating Intelligent Sootblowing applications into optimization

efforts
v
The possibilities for heat rate improvements from upgrades in Turbines and Auxiliaries
The latest trends in Heat Rate Testing
Actual Plant Experiences with heat rate improvement projects
EPRI Perspective
This conference was sponsored by EPRIs Heat Rate and Cost Optimization Value Package. As
such, the meeting reflects those topics considered most important by the members of the value
package in their continual efforts to improve heat rate and overall plant performance. The value
package is currently supporting demonstrations of on-line heat rate monitors, total plant cost
optimization, intelligent sootblowing, and steam quality assessment. The proceedings from the
previous conference in 1998 were published as Proceedings: 1998 Heat Rate Improvement
Conference (TR-111047).
Keywords
Heat Rate
Boiler Performance
Fossil-Fired Power Plants
Power Plant Optimization
Sootblowing
vi
AGENDA
EPRIs Twelfth Heat Rate Improvement Conference
January 30-February 1, 2001
Hotel Inter-Continental Dallas
Dallas, TX
Final Agenda
Tuesday, January 30, 2001
7:00 a.m.
Registration and Continental Breakfast
8:00 a.m.
Welcome
Conference Chair: Jeff Stallings, EPRI
Utility Host: Ron Seidel, Senior Vice President, Fossil Generation, TXU Energy
8:10 a.m.
Session 1: On-Line Heat Rate Monitors

Session Chairs: Tom Calle, TXU Energy and Charlie Rose, Consultant
8:10 a.m.
Real-Time Performance Monitoring of Coal-Fired Units

Duane Hill, Dairyland Power Cooperative
Sastry Munukutla, Tennessee Technological University
8:35 a.m.
F-Factor Method for Heat Rate Measurement and its Characteristics

Nenad Sarunac, Lehigh University
Carlos E. Romero, Lehigh University
Edward K. Levy, Lehigh University
9:00 a.m.
The Input/Loss Method

Fred D. Lang, Exergetic Systems, Inc.
9:25 a.m.
Recent Experiences with Performance Monitoring at TVA using EPRIs

Plant Monitoring Workstation
Robert Inklebarger, TVA
Eric Sikes, TVA
Cyrus Taft, EPRI I&C Center
vii
9:50 a.m.
Break
10:20 a.m.
Modular Heat Rate Calculations for Power Plants using OPC

Harry R. Winn, Westinghouse Process Control, Inc.
10:45 a.m.
Improving Results, Confidence and Decisions with Precise Data Validation

Marcus Caudill, Performance Consulting Services, Inc.
Ron Griebenow, Performance Consulting Services, Inc.
11:10 a.m.
Panel: How are utilities using heat rate information?

Moderator: Mark Ness, Great River Energy
Panelists:
Tom Calle, TXU Energy
Leeth DePriest, Southern Company Services
12:00 p.m.
Lunch
1:00 p.m.
Session 2: Optimization
Session Chairs: Darrell Howard, TVA and Stratos Tavoulareas, EnTEC
1:00 p.m.
Impacts of Combustion Optimization on Power Plant Heat Rate

1:25 p.m.
Obtaining Improved Boiler Efficiency and NOx using Advanced Empirical

Optimization and Individual Burner Instrumentation on a Boiler Operated
in Load-Following Mode
E. P. Payson, Allegheny Energy Supply
Dave Earley, Air Monitor Corporation
Rich Brown, EPRI
Carlos Moreno, Ultramax Corporation
1:50 p.m.
Application of GNOCIS
Neural Network Optimization Controller for
Boiler Efficiency Control
Darrell A. Howard, TVA
Lonnie Coffey, EPRI I&C Center/TVA
2:15 p.m.
Heat Rate Improvement at Dairylands Madgett Station using NeuSIGHT

Brad Radl, Pegasus Technologies
Glen Foster, Data Systems & Solutions
2:40 p.m.
Break
viii
3:10 p.m.
ProcessLink
at the Roanoke Valley Energy Facility
Don Keisling, LG&E
Peter Spinney, NeuCo
3:35 p.m.
Automatically Control NOx with Heat Rate Constraints, in a Coal-Fired

Power Plant
Kandi Forte, Reliant Energy
Tom Cowder, Reliant Energy
Russell F. Brown, Pavilion Technologies, Inc.
4:00 p.m.
Unit Optimization at Hammond Unit 4

John Sorge, Southern Company Services
5:30 p.m.
Reception and Exhibits
Wednesday, January 31, 2001

7:00 a.m.
Continental Breakfast
8:00 a.m.
Session 3: Intelligent Sootblowing

Session Chairs: William Yee, Reliant Energy and Rabon Johnson, EPRI I&C
Center
8:00 a.m.
Effects of Sootblowing in Coal-Fired Boilers on Unit Heat Rate and Nox

Edward K. Levy, Lehigh Univerisity
8:25 a.m.
Optimization of Boiler Sootblower Operation

Jeffery Williams, Westinghouse Process Control, Inc.
Xu Cheng, Westinghouse Process Control, Inc.
Bernie Begley, Southern California Edison, Inc.
Alex Smith, Southern California Edison, Inc.
Dale Hopkins, Southern California Edison, Inc.
8:50 a.m.
Intelligent Sootblowing Application Development

Neel J. Parikh, Pegasus Technologies, Inc.
Brad J. Radl, Pegasus Technologies, Inc.
9:15 a.m.
Intelligent Sootblowing Boiler Cleaning Management System

Randy Carter, Applied Synergistics
9:40 a.m.
Break
10:10 a.m.
Session 4: Turbines and Auxiliaries

Session Chairs: Jim Terrell, TVA and Tom McCloskey, EPRI
ix
10:10 a.m.
In-Situ Feedwater Flow Measurement

Izidro DiazTous, Encor-America
10:35 a.m.
In-Situ Enthalpy Measurements in Low Pressure Condensing Steam

Turbines
Steve Hesler, EPRI
Tom McCloskey, EPRI
11:00 a.m.
Steam Turbine Related Reseach at TVA

Jim Terrell, TVA
11:25 a.m.
On-Line Performance Monitoring and Condition Assessment of Steam

Turbines
Rolf F. Orsegh, Impact Technologies LLC
Michael Roemer, Impact Technologies LLC
Ben Atkinson, Impact Technologies LLC
Bill McGinnis, Reliant Energy
Scott McQueen, Reliant Energy
11:50 a.m.
Lunch
1:00 p.m.
Session 5: Heat Rate Testing

Session Chair: Sam J. Korellis, Dynegy Midwest Generation
1:00 p.m.
Performance Evaluation and Testing

Sam J. Korellis, Dynegy Midwest Generation
Philip Gerhart, University of Evansville
1:25 p.m.
A Procedure for Analyzing Power Plant Measurement Variances Associated

with Thermal Performance Testing
1:50 p.m.
Cycle Alignment Methods and Evaluation

2:15 p.m.
In Search of Unaccounted for BTUs via the Art of ASME PTC-6 Testing
Italo Liberatore, Constellation Power Source Generation
Allison Rossi, Constellation Power Source Generation
Donald Fyhr, Constellation Power Source Generation
2:40 p.m.
Break
3:10 p.m.
Panel: Why test?

Moderator: Sam Korellis Dynegy Midwest Generation
Panelists:
Matt Dooley, Alstom
Fred Lang, Exergetic Systems
Albert Lau, Reliant Energy
Dick Storm, Storm Engineering
4:30 p.m.
Adjourn
Thursday, February 1, 2001

7:00 a.m.
Continental Breakfast
8:00 a.m.
Panel: What improvements in heat rate information are needed?

Moderator: Duane Hill, Dairyland Power Cooperative
Panelists:
Ron Griebenow, Performance Consulting Services
Darrell Howard, TVA
Gary Walling, Alliant Power
William Yee, Reliant Energy
9:15 a.m.
Session 6: Plant Experiences

Session Chairs: Gary Walling, Alliant Energy and Jose Sanchez, EPRI
9:15 a.m.
Analysis of Variables for Predicting Power Output at the Columbia Power

Plant
Aravindan Rangarajan, Industrial Engineering MS, Iowa State University
9:40 a.m.
Experience of Moneypoint Power Station in Recovering Plant Heat Rate:

Focus on Control Valves
Michael Rocke, Electricity Supply Board, Ireland
Tom Canning, Electricity Supply Board, Ireland
Sanjay V. Sherikar, Control Components, Inc.
10:05 a.m.
Break
10:35 a.m.
Heat Rate Improvement in an Existing Multifuel Unit

Joaquin G. Blas, Hidroelectrica del Cantabrico, S.A.
Florentino Blanco, Hidroelectrica del Cantabrico, S.A.
11:00 a.m.
An Application of the Plant Performance Modelling Package PROATES to

Analyse the Cause of a Persistent Tube Failure Problem
K.R.J. Hartwell, Powergen
A.B. Ready, Powergen
xi
11:25a.m.
In-Situ O2 Probe Failure at Dairylands JP Madgett Station

11:50 a.m.
Closing Remarks
12:00 p.m.
Adjourn
xii
CONTENTS
1 SESSION 1: ON-LINE HEAT RATE MONITORS ................................................................ 1-1

Real-Time Performance Monitoring of Coal-Fired Units...................................................... 1-2
Sastry Munukutla, Tennessee Technological University
F-Factor Method for Heat Rate Measurement and its Characteristics................................1-23
The Input/Loss Method......................................................................................................1-61
Recent Experiences with Performance Monitoring at TVA using EPRIs Plant
Monitoring Workstation......................................................................................................1-73
Robert Inklebarger, TVA
Eric Sikes, TVA
Cyrus Taft, EPRI I&C Center
Modular Heat Rate Calculations for Power Plants using OPC ...........................................1-85
Harry R. Winn, Westinghouse Process Control, Inc.
Improving Results, Confidence and Decisions with Precise Data Validation....................1-109
Marcus Caudill, Performance Consulting Services, Inc.
Ron Griebenow, Performance Consulting Services, Inc.
2 SESSION 2: OPTIMIZATION .............................................................................................. 2-1
Impacts of Combustion Optimization on Power Plant Heat Rate ........................................ 2-2
xiii
Obtaining Improved Boiler Efficiency and NOx using Advanced Empirical Optimization
and Individual Burner Instrumentation on a Boiler Operated in Load-Following Mode .......2-30
E. P. Payson, Allegheny Energy Supply
Dave Earley, Air Monitor Corporation
Rich Brown, EPRI
Carlos Moreno, Ultramax Corporation
Application of GNOCIS Neural Network Optimization Controller for Boiler Efficiency
Control ..............................................................................................................................2-63
Darrell A. Howard, TVA
Lonnie Coffey, EPRI I&C Center/TVA
Heat Rate Improvement at Dairylands Madgett Station using NeuSIGHT......................2-68
ProcessLink at the Roanoke Valley Energy Facility ........................................................2-80
Don Keisling, LG&E
Automatically Control NOx with Heat Rate Constraints, in a Coal-Fired Power Plant ......2-107
Kandi Forte, Reliant Energy
Tom Cowder, Reliant Energy
Russell F. Brown, Pavilion Technologies, Inc.
Unit Optimization at Hammond Unit 4..............................................................................2-128
John Sorge, Southern Company Services
3 SESSION 3: INTELLIGENT SOOTBLOWING..................................................................... 3-1
Effects of Sootblowing in Coal-Fired Boilers on Unit Heat Rate and Nox............................ 3-2
Edward K. Levy, Lehigh Univerisity
Optimization of Boiler Sootblower Operation .....................................................................3-35
Jeffery Williams, Westinghouse Process Control, Inc.
Xu Cheng, Westinghouse Process Control, Inc.
Bernie Begley, Southern California Edison, Inc.
Alex Smith, Southern California Edison, Inc.
Dale Hopkins, Southern California Edison, Inc.
Intelligent Sootblowing Application Development...............................................................3-56
Neel J. Parikh, Pegasus Technologies, Inc.
Brad J. Radl, Pegasus Technologies, Inc.
Intelligent Sootblowing Boiler Cleaning Management System ........................................3-63
Randy Carter, Applied Synergistics
xiv
4 SESSION 4: TURBINES AND AUXILIARIES...................................................................... 4-1

In-Situ Feedwater Flow Measurement ................................................................................ 4-2
Izidro DiazTous, Encor-America
In-Situ Enthalpy Measurements in Low Pressure Condensing Steam Turbines.................. 4-3
Steve Hesler, EPRI
Tom McCloskey, EPRI
Steam Turbine Related Reseach at TVA ...........................................................................4-19
Jim Terrell, TVA
On-Line Performance Monitoring and Condition Assessment of Steam Turbines ..............4-27
5 SESSION 5: HEAT RATE TESTING ................................................................................... 5-1
Performance Evaluation and Testing.................................................................................. 5-2
Philip Gerhart, University of Evansville
A Procedure for Analyzing Power Plant Measurement Variances Associated with
Thermal Performance Testing ............................................................................................ 5-7
Cycle Alignment Methods and Evaluation..........................................................................5-14
In Search of Unaccounted for BTUs via the Art of ASME PTC-6 Testing ........................5-18
Italo Liberatore, Constellation Power Source Generation
Allison Rossi, Constellation Power Source Generation
Donald Fyhr, Constellation Power Source Generation
6 SESSION 6: PLANT EXPERIENCES .................................................................................. 6-1
Analysis of Variables for Predicting Power Output at the Columbia Power Plant ................ 6-2
Aravindan Rangarajan, Industrial Engineering MS, Iowa State University
Experience of Moneypoint Power Station in Recovering Plant Heat Rate: Focus on
Control Valves...................................................................................................................6-18
Michael Rocke, Electricity Supply Board, Ireland
Tom Canning, Electricity Supply Board, Ireland
Sanjay V. Sherikar, Control Components, Inc.
Heat Rate Improvement in an Existing Multifuel Unit .........................................................6-28
Joaquin G. Blas, Hidroelectrica del Cantabrico, S.A.
Florentino Blanco, Hidroelectrica del Cantabrico, S.A.
xv
An Application of the Plant Performance Modelling Package PROATES to Analyse

the Cause of a Persistent Tube Failure Problem ...............................................................6-63
K.R.J. Hartwell, Powergen
A.B. Ready, Powergen
In-Situ O2 Probe Failure at Dairylands JP Madgett Station................................................6-75
xvi
1
SESSION 1: ON-LINE HEAT RATE MONITORS
1-1
1-2
1-3
1-4
1-5
1-6
1-7
1-8
1-9
1-10
1-11
1-12
1-13
1-14
1-15
1-16
1-17
1-18
1-19
1-20
1-21
1-22
1-23
1-24
1-25
1-26
1-27
1-28
1-29
1-30
1-31
1-32
1-33
1-34
1-35
1-36
1-37
1-38
1-39
1-40
1-41
1-42
1-43
1-44
1-45
Test Instrumentation Error

D ifference in F lue G as F low R ate C alc ulated from S tandard and
Im proved Ins trum entation for Veloc ity Head M easurem ent
5
P e rc e n t F lo w D iffe re n c e
4 .5
3 .7
3 .5
S ite
S ite
S ite
S ite
1:
2:
3:
4:
S ho rt S tack,
S ho rt S tack,
T all S tac k,
T all S tac k,
RA =
RA =
RA =
RA =
4 Deg
9 Deg.
13 Deg.
6 D eg
1 .9
0
S ite 1
S ite 2
S ite 3
S ite 4
T est S ite
24
1-46
1-47
Achieved Reductions In Flow Bias Error

20
18.3
O ld E P A R eg u lation s
R ed u ctio n in C E M Bias E rro r [% ]
18
New E P A Reg u lation s
16
13.9
14
13.2
12
10.3
9.8
10
8
7.1
6
4
3.4
3.4
3.4
Un it A
Un it B
Un it C
4.2
4.1
Un it D
Un it E
3.8
2
0
T ested U n it
1-48
Un it F
26
1-49
1-50
1-51
Dilution Probe Errors

E rro r in C o n cen tratio n M easu rem en t [%]
6
Nominal P ara mete rs
T sta ck = 2 50 de g. F
4
P sta ck = 1 atm.
M W sta ck = 3 0 lb/lbmo l
P re g = 5 0 psig
T u mb .ca b le = 8 0 deg. F
6 0 D e g. F = 1% Erro r
-2
-4
-2 50
-2 00
-1 50
-1 00
-5 0
50
1 00
1 50
2 00
2 50
C h an g e in S tack T em p eratu re [F ]
30
1-52
1-53
1-54
1-55
1-56
1-57
F-Factor Method - Field Results

10 ,2 0 0
F lu e g a s flo w ra te m e a s u re d b y A u to p ro b e
TM
. G rid fin e n e s s = 4 8 p o in ts
10 ,1 0 0
Input/Output
Ne t Un it He a t R a te [B T U/k W h ]
B TC E
F F acto r, O2-B ase d
10 ,0 0 0
F F acto r, C O2-B ased
9,90 0
9,80 0
9,70 0
9,60 0
9,50 0
0
10
15
20
T e s t Nu m b e r
1-58
25
30
35
40
36
1-59
1-60
1-61
1-62
1-63
1-64
1-65
1-66
1-67
1-68
1-69
1-70
1-71
1-72
1-73
1-74
1-75
1-76
1-77
1-78
Bull Run DCS Graphics

Controllable Losses
1-79

Boiler Performance
1-80

Feedwater Heaters
1-81
1-82
1-83
1-84
Modular Heat Rate Calculations

for Power Plants using OPC
By:
Harry R. Winn
Westinghouse Process Control, Inc.
1-85
Modular Heat Rate Calculations

for Power Plants using OPC
Or how to use OPC in power plants
as a system wide performance
monitoring tool
1-86
Performance Monitoring
Traditional Methods...
Proprietary Code usually not supplied

Proprietary Datalink to DCS required
Difficult for users to make modifications
Costly to have the performance monitoring
vendor make modifications
Eventually does not match plant
configuration due to lack of maintenance
Difficult to get key performance data to
users on plant network
1-87
The challenge-- Develop performance monitoring system

which is user friendly, easy to maintain, and
provide needed data to the engineer in the
format required
Provide an standard communication link to
all DCS vendors that is easy to configure and
maintain and allows bi-directional
communication
Provide option to allow the performance
monitoring system to execute at the
enterprise level and send critical data
l
h l l l
1-88
The 1st Challenge-- Develop standard method of interface

to DCS systems
OPC was chosen as interface method
1-89
OPC--- What is it?

OPC (originally OLE for Process Control) is an industry
standard created with the collaboration of a number a leading
worldwide automation and hardware software suppliers working
in cooperation with Microsoft. The standard defines method
for exchanging realtime automation data among PC-based
clients using Microsoft operating systems.
The organization that manages this standard is the
OPC Foundation.
1-90
OPC--- Why use it?

Most DCS vendors have OPC server
Standard design of protocol based on OPC
specification
Browser capability
Easily configurable
No custom programming
Enterprise level of performance monitoring
www.opcfoundation.org
1-91
Typical OPC Performance

Monitoring Configuration:
DCS Vendor #1 OPC
Server
DCS PM Operator
Graphics
Bi-directional Data
Communication
Ethernet
Key Performance
Data to users on
plant network
1-92
DCS Vendor #3 OPC

Server
DCS Vendor #2 OPC

Server
DCS PM Operator
Graphics
Bi-directional Data
Communication
Ethernet
Performance
Monitoring System for
three units
DCS PM Operato
Graphics
Bi-directional Data
Communication
The 2nd Challenge-- Develop user friendly performance

monitoring system that is easily
configurable and maintainable
Complete point and click capabilty
Develop monitoring system for both
fossil fired utility type boilers and
combined-cycle plants
1-93
Performance Monitoring System

User Interface
Centralized performance monitoring
tag database
Complete point and click capability to
configure and maintain
Custom algorithms created for
performance monitoring modules
including condenser, boiler, turbine,
feedwater heater, etc.
1-94
Flowchart
workspace
Plant
attributes
Algorithm
selection
Drop and drag
1-95
Typical Condenser Calculation

Worksheet
1-96
Condenser Design Data Screen
1-97
OPC Tag Configuration
1-98
Event Logger to easily track

system
1-99
OPC tag selection output mapping
1-100
Run mode and What-If

manual constant entry
1-101
Run mode view of calculations
1-102
Any Plant
inputs
Any plant
equipment
type
Plant
Overview
Flow Chart
Live values
displayed
1-103
1st Installation ...

Implemented at AES Warrior Run
project in Cumberland, Maryland
210 gross megawatt rating

CE coal fired fluidized bed boiler
ABB steam turbine
Co-generation plant
Commissioned in May, 2000
1-104
Controllable Loss Operator

Graphic
1-105
Controllable Loss Operator

Graphic #2
1-106
Cooling Tower Operator

Graphic
1-107
Future Development...
Data Validation module with SoftSensor
replacement capability, Spring 2001
Data Replay module to permit retrieving
historical data, modifying, and re-executing
the calculations, Fall 2001
1-108
Improving Results, Confidence

and Decisions
with
Precise Data Validation
Duane Hill
Dairyland Power
Marcus Caudill
Ron Griebenow, P.E.
Performance Consulting Services, Inc.
1-109
Why Is Data Validation Important?

Instrument Errors:
Reduce Efficiency
z Increase Emissions
z Cause Unnecessary Power Purchases
z Create Monitoring System Errors
z Cost $$$$$$$$$$$
z
~$1,000,000 per year on a 440 MW fossil

plant
1-110
ALL Automated Systems Need

Accurate Data for Reliable Results
Control System
z Combustion Optimization System
z Expert / Advisory System
z Performance Monitoring System
z On-line Costing System
z
1-111
What is Advanced Data

Validation?
Recognizes Interdependence of All Data in
Man-Made Systems
z Utilizes Repeatable Patterns of Process
Data
z Applies Localized (Relevant) Models to
Current Operation
z Provides Precise Analytical Redundancy
z Provides Data Filtering for Automated
Systems
z
1-112
Data Filtering
Automated
System
ACM
Raw
Inputs
Filtered
Values
Outputs
1-113
Calibration Optimization /
Calibration Reduction
Early identification of instrument drift
z Identify instruments requiring calibration
during an outage
z Identify instruments that DO NOT require
calibration
z Utilize highly-skilled techs in higher
priority tasks
z
1-114
Duke Nuclear
Oconee Nuclear Station
z 900 MW PWR producing Superheated
Main Steam
z ACM Capability Demonstration Tracking
Calometric Data
z
44 points tracked over a 4 month period
1-115
With reference data primarily from the

first week of operation, this temperature
is predicted within 0.1% of actual value.
1-116
Drift in this main steam temperature

measurement is easily detected and
alarmed before the measurement had
drifted 4 degrees (less than 1%).
1-117
A loss of gross generation associated

with the error in calculated steam
temperature rise, feedwater venturi
fouling and changes in condenser
backpressure is also easily detected
1-118
Accurate replacement values

are provided during temporary
instrument faults.
1-119
Allegheny Harrison Unit 2

650 MW Rated Capacity
z Foster Wheeler Opposed-Wall, Coal-Fired,
Supercritical Boiler
z Westinghouse Single Reheat Turbine
z
1-120
Allegheny Project
Data Validation for Existing On-Line
Performance Monitoring System
z Initial Data Gathering and Model
Development in Early 1999
z Seven-Week Turbine Overhaul and Boiler
Outage Started April 1, 1999
z
System used to identify post outage process

changes
models updated for post outage operation
1-121
Range of reference data (~50100% load, consistent with

performance monitoring
system operation)
1-122
Very accurate modeling of main steam

temperature prior to unit outage (except shut
down periods where no reference data exists).
1-123
Increase in top heater drain temperature.

Subsequent analysis reveals corresponding
decrease in heater level and increase in DCA.
1-124
B-side heater shell pressure tracks

closely with prediction.
1-125
A-side heater shell pressure clearly

drops (<1%) after a unit shutdown
1-126
Allegheny Harrison Unit 2

System View
Early March
Late March
Post-Outage
Pre-Outage
1-127
Notable increase in gross generation

following turbine overhaul.
Turbine Overhaul
1-128
Turbine Overhaul
Increase in steam flow due to replacement

of many turbine seals.
increases in turbine train pressures
confirms increase in steam flow
supports generation increase
1-129
Turbine Overhaul
1-130
Expected improvements in heat rate due

to turbine overhaul (>1/2%) confirmed
through data validation. Validation of all
values used in heat rate calculation
increases confidence in calculated result.
No increase in HP turbine efficiency

after turbine overhaul.
Turbine Overhaul
1-131
Substantial (unrealistic) increase in

calculated IP turbine efficiency after turbine
overhaul. Performance monitoring system
includes assumed N2 leakage, which was
greatly reduced with seal replacement.
Calculations must be updated.
Turbine Overhaul
1-132
Turbine Overhaul
After transmitter replacement, hot

reheat pressure at the boiler shows
increase consistent with increased
steam flow (reduced turbine leakage).
1-133
Turbine Overhaul
1-134
Hot reheat pressure at the turbine

substantially lower than expected.
Screens placed in steam path after
work in boiler reheat section
produce large pressure drop.
Turbine Overhaul
Extensive air heater work during

outage, including replacement of
several baskets. Clear reduction
in air heater pressure drop on
both the air and gas sides.
1-135
Error in calculated BFPT steam flow.

Failure of BFPT discharge pressure
transmitter identified as root cause.
Turbine Overhaul
1-136
Turbine Overhaul
Several transmitters changed from

absolute to gauge during outage. Will
require update to performance
calculations and data validation system.
1-137
OPG Lambton Unit 3

z
Stand-alone Data Validation for Plant

Control System
Retrieves data from control system
Can send alarms back to control system
Initial Data Gathering and Model

Development in Spring 2000
z Installation October 2000
z Installation Being Expanded to 4 Units
z
1-138
ACM detects FWH drain temperature spikes

resulting from low heater level
1-139
During the daily automatic calibration cycle,

ACM can provide an accurate replacement
value.
1-140
Dairyland Power JPM

Installing NeuSIGHT Optimization System
Requires Extensive Data Set
Found O2 Calibration Errors After Data Set
Was Collected
Existing ACM Models Were Used to Assess
O2 Accuracy
ACM High Accuracy Predictions Were Used
to Validate and Replace Faulty O2 Data For
Training of the Neural Network Models
1-141
Measured value lower than prediction until

calibration on 2/3.
1-142
Drift begins on 2/1. No Change during 2/3

calibration.
1-143
Measured value lower than prediction for

entire time frame.
1-144
Matches well until 1/31, then measured value

drifts low.
1-145
Measured value higher than prediction until

calibration on 2/3. Begins to drift again on
2/8.
1-146
Measured value slightly higher than

prediction until calibration on 2/3, then a
good match.
1-147
APR Data Validation Has Been

Successfully Applied To:
z
Data filtering and validation for:

Performance Monitoring Systems
Control Systems
Neural Network-Based Optimization Systems
Stand-Alone Data Validation

z Calibration Optimization and Reduction
z
1-148
Conclusion
Precise Data Validation
Improves Performance Monitoring and
Optimization System Results
z Increases Operations and Engineering
Staff Confidence in On-Line Information
z Provides Reliable Information to Support
Operations and Maintenance Decisions
z
1-149
Improving Results, Confidence and Decisions

With Precise Data Validation
Duane Hill
Dairyland Power Cooperative, Inc.
Marcus Caudill
Ron Griebenow, P.E.
Performance Consulting Services, Inc.
Abstract
In the increasingly competitive electric power generation market, it is critical that all
generation resources be utilized in the most cost-effective manner. In particular, it is
essential that the operation and maintenance costs of steam power cycles be minimized
while maintaining peak availability, reliability, efficiency and environmental compliance.
Advanced control technologies and artificial intelligence are becoming more frequently
used to support these optimization efforts. However, these advanced technologies are
heavily reliant upon the validity of the input data.
Application of advanced data validation methods can improve the reliability of and the
confidence in intelligent control technologies. Using advanced data validation to preprocess the plant data that is used by performance monitoring, combustion optimization,
plant control, and artificial intelligence systems will provide these systems with accurate
and reliable information, increasing confidence in the calculated results and operational
recommendations.
In addition, advanced data validation can accurately identify instruments requiring
calibration. Calibration efforts can be then focused on only those instruments that need
attention, reducing total hours required for instrument maintenance.
Advanced data validation methods have been applied in various ways to a number of
generating units, including Dairyland Power Cooperatives J.P. Madgett and Genoa
Stations and Allegheny Powers Harrison Unit 2. This paper provides an overview of
various data validation methods and outlines some of the benefits of advanced data
validation in reducing operation and maintenance costs. It also presents some of the
specific findings from initial analyses at these sites. Oral presentation at the EPRI Heat
Rate Improvement Conference will include additional case studies from various utility
installations illustrating the integration of advanced data validation into plant automation,
monitoring, and intelligent control applications.
Page 1 of 1
1-150
What Is Data Validation and Why Is It Important?

In an attempt to improve plant performance, reduce operating and maintenance costs,
and meet the requirements of new regulation, many utilities are implementing
automated systems and new technologies for plant monitoring and control. However,
the usefulness of these systems is dependent upon the reliability and accuracy of the
instrumentation. While the optimization algorithms are often mathematically perfect to
achieve best operating conditions, the data entering these systems is usually less than
perfect. Erroneous input data can result in misleading performance assessments,
inappropriate operator actions, inefficient plant operation, excessive plant emissions,
and a host of other undesirable effects. This is why accurate, reliable data is critical to
cost-efficient operation of any generating facility, and why implementing a cost efficient
method of plant data validation is essential.
Note that even though questionable data supplied to a performance monitoring system
will produce questionable results and reduce operator confidence in the system, the
data does not truly impact the integrity of the calculations. When good data is restored,
the calculated performance results will be correct. This is not necessarily true with
neural-network based optimization systems. Because these systems are developed
(trained) based on measured plant data, it is essential to provide accurate data for
initial training of the system and, possibly even more importantly, for subsequent on-line
re-training, if the system has this capability. Otherwise, errors will be built into the
network calculations causing incorrect optimization scenarios.
Measured data obtained from process sensors and associated electronic or pneumatic
equipment is the only path for the unit operator, automatic control system or computerbased operator support system to obtain process information. Therefore, it is very
important that the quality of process information be high. While instrument reliability
and quality has improved significantly in recent years, wholesale instrument
replacement is difficult to justify and usually not warranted. Erroneous data from
sensors must be quickly and effectively identified so that the operator and any
automated analysis or control functions can maintain reliable, efficient operation. Invalid
sensor signals may result in misleading or incorrect conclusions and inappropriate
responses (e.g., an invalid measurement could be mistaken for a process fault) by an
operator or computer-based system, possibly producing unnecessary process
shutdowns or equipment damage.
Historically, instrument calibrations have been performed on a time basis, with those
instruments that were deemed as critical to safety and efficient operations calibrated
more frequently than those viewed as less important. This method of calibration has
several side effects. First, the calibration intervals for the critical instruments may be
close enough together that the instruments would not change significantly, so while the
integrity of the instrument signals was maintained, the instrument maintenance staff
spent precious resources calibrating instruments that did not need calibration. Second,
due to the time required to maintain these critical instruments, the less important
instrument calibration intervals tend to get longer and longer until a problem with one of
these less important systems becomes critical.
Page 2 of 2
1-151
In short, quality data is essential to get effective results from and use of automated
controls, computerized optimization systems and operator response. Data validation
provides information to help distinguish measurement failures from process faults and
select which instrument signals to use in control and analysis functions. High quality
data validation can also reduce the time spent on calibration of instruments that are
within specifications, and can identify those instruments that are beginning to change,
so that a calibration check can be scheduled.
Data Validation Methods

Data validation has been an issue for power producers as long as there have been
power plants. In many cases, the traditional form of data validation was to take a test
instrument out into the plant and install it adjacent to a questionable instrument and
compare the two. If the test and plant instrument were close enough, then the plant data
was considered valid. As fuel cost became more of an issue in the industry and
additional automation required additional instrumentation, more complex and more
costly data validation schemes emerged. Many of these methods of data validation
resulted in significantly higher instrument maintenance costs and provided only
marginally better instrument accuracy and reliability. The more common methods of
maintaining instrument accuracy are presented below, along with a discussion of
advanced data validation.
Blanket Calibration
The data validation method that has been most commonly used throughout the industry
is blanket calibration. Historically, power generators have scheduled unit maintenance
on a time basis (12 month intervals, 18 month intervals, etc.) and during these
maintenance outages, many of the process instruments were scheduled for calibration.
When preparing for these outages, it was not unusual for the plant staff to budget 500 to
1500 hours or more for instrument calibrations. In addition to the staff time, high-quality
calibration equipment is required for each technician performing the calibrations. In
many cases, technicians found that they were calibrating instruments that did not need
calibration and, because of the length of time between calibrations, they were not
developing useful calibration history on the individual instruments. For example, they
might record that a particular pressure transmitter was out of calibration three years in a
row, but they had no idea when during that year the calibration drift occurred. Between
the intensive manpower requirements and the calibration history developed, there are
definite drawbacks to blanket calibrations.
Data Comparison
Comparison methods are based on the availability of at least two measurements (direct
or derived) for a desired process state. These redundant indications may then be used
to make some judgment about the validity of the measurement signals. The simplest
comparison methods involve the installation of two sensors at the same location for the
same process state. These redundant sensors are compared with each other, and
disagreement between them, larger than a threshold (which is typically related to the
Page 3 of 3
1-152
amount of anticipated measurement noise), is considered indicative of failure of one of

the measurements. However, if only two measurements are available, no decision can
be made on which to accept. An unambiguous measurement quality determination
requires more than two measurements. When at least three measurements are
available for comparison, it is possible to make some logical choice of which to accept
or reject, and to form a "best estimate" of the true value of the process state.
When there are no physically redundant devices, it is possible to use analytically
redundant measurements. An analytically redundant measurement is most often
derived using a process model to calculate a representative value of a directly
measured state from measurements of other states. Once three or more
representations of a particular measurement are available, there are methods which
may be used to discriminate failed measurements and select the most representative
"true" value of the measurement.
Advanced Data Validation
The usefulness of the methods described above has been severely limited by the cost
and complexity of their implementation, as well as the accuracy and reliability of their
results. Blanket calibrations require a significant investment in highly trained staff and a
continuous block of dedicated time for maintaining calibrations. Installing and
maintaining redundant sensors is costly and tends to be limited to a small subset of the
total number of sensors that impact process performance and reliability. Comparison
methods also generally compare only two or three measurements to each other or a
measurement to a fixed limit or a simple single-variant curve. Consequently, this
approach is not very precise or robust, and is therefore unlikely to cover a large subset
of process measurements. These limitations have deterred many utilities from initiating
a comprehensive data validation program. Fortunately, advanced data validation
technology resolves many of the shortcomings of these traditional methods.
Advanced data validation works in much the same way as the human brain. The human
control operator with an analog control system observes the positions of dials, gauges
and other indicators on the control panel and compares the current picture of unit
operation with previous pictures that are stored in memory to decide if the unit is
operating in an acceptable manner. If the picture matches something that the operator
remembers, either from training or from past operation, then the operator can identify
the current operation. If something in the picture deviates from past experience, the
operator might use part of the picture to make a prediction (that pressure should be
200 psi).
The advanced data validation technology works in a similar way. Snapshots of historical
data are collected, evaluated to assure that all of the information within the snapshot is
good, then included in a reference data file. Then as new data is evaluated, a snapshot
of current data is compared to the reference snapshots, and the system determines if
the current operation is consistent with past operation. A detailed explanation of the
methodology can be obtained in Reference 1.
Page 4 of 4
1-153
The advanced data validation approach brings several advantages over other numerical
methods. For example, since this technology recognizes that all monitored parameters
are interrelated, the underlying algorithm is highly fault tolerant; the effects of incorrect
or missing plant measurements are minimized.
It is important to realize that advanced data validation is not a replacement for
conventional performance monitoring, combustion optimization or other plant process
improvement system. It is designed to work in conjunction with these systems, providing
high quality data so that these conventional systems can achieve optimum results.
Advanced Data Validation Applications

Advanced data validation systems have been or are being installed at a number of
generating stations, both in the US and overseas. Personnel at these sites espouse a
wide range of performance monitoring philosophies and reasons for installing data
validation systems. However, all realize that the common problems with data reliability
that exist in all power plants will prevent them from achieving their lowest cost of
generation. Each of the examples below present the utilitys motivation for installing
data validation, their goals for the installation, and examples of the results achieved to
date.
Dairyland Application
Dairyland Power Cooperative is a long-time user of on-line performance monitoring and
has a history of being on the cutting edge of performance improvement technology.
After deciding to upgrade from a VAX-based performance monitoring system to a
Windows NT-based system, Dairyland staff recognized the need to verify data quality.
The data quality issue was to be even more important since Dairyland had plans to
install a neural network-based optimization system. Dairyland knew that both the
performance monitoring system and the optimization system were good programs but,
to work effectively, they need good data.
Another primary motivation for installing a data validation system at Dairyland was to
reduce the load on instrument technicians. The three technicians at JP Madgett station
were already overloaded, and each time that blanket calibrations were performed,
Dairyland technicians were finding a number of instruments were scheduled for
calibration when they didnt need to be. What was really needed was a tool to assist the
instrument technicians in determining when an instrument needed calibration or
replacement. To this end, advanced data validation was implemented to monitor
approximately 260 instruments.
Dairylands performance improvement program relies heavily on the existing
performance monitoring system, which is used not only for performance calculations,
but also as a data archival and retrieval system. In addition, the optimization system
being installed will acquire its data through the performance monitoring system. Based
on these considerations, Dairyland decided to integrate the data validation system with
Page 5 of 5
1-154
their existing performance monitoring system. Figure 1 depicts the data flow of the
Dairyland system.
In one of many examples of the systems benefits, it identified an anomaly in the highpressure feedwater heater drain temperature. Since this was happening at night at low
loads, the problem most likely would not have been identified through existing
monitoring activities. The controller was set improperly causing a drain valve to open
when it was supposed to be closed. Correcting the problem allowed the heater to
perform much more effectively, improving unit overall efficiency.
PC Windows NT 4.0
Unit Control
System
Perf. Monitor Database

DAS Value Storage
ACM Value Storage
Calc Value Storage
DAS
Value
DAS or
Replacement
Value
Performance
Monitoring
Calculations
Calculated
Results
Validated
DAS Value
DAS or
Replacement
Value
Advanced
Data
Validation
Process
Neural Network-based
Optimization System
(Planned)
Optimization
Instructions
Figure 1 -- Integration of Advanced Data Validation at Dairlyands JP Madgett
In another example, the performance staff noticed that one of the two Madgett
feedwater flow transmitters had failed. The operations staff had noticed the same thing
and were trying to determine the time of failure. Using the performance monitoring
system, a trend of both flow transmitters and the advanced data validation system
predicted flow value was displayed. The trend, depicted in Figure 2, shows that one
transmitter began to straight line, indicating the time of failure. Further, the predicted
value provided an accurate replacement for this failed transmitter that could be used for
continued operation until the failed instrument in replaced.
Page 6 of 6
1-155
3000
2500
F
W
2000
F
L
O
w
1500
1000
500
Time
Actual FW Flow 1
Actual FW Flow 2
ACM Predicted FW Flow 1
Figure 2 -- Dairyland Feedwater Flow Transmitter Failure

In yet another example, the plant had been experiencing unusual operating problems
for several days. Operators reported that the unit was operating strangely at lower
loads. A review of the advanced data validation system results revealed that the West
Primary AirFlow transmitter was fluctuating between 200 KPH and its maximum value of
531 KPH. The advanced data validation system prediction did not show this fluctuation,
indicating that there was a problem with the transmitter. Shortly after this problem was
discovered, the transmitter failed completely. This illustrates the value of data validation
in pinpointing the root cause of operating problems. Identifying the deficient instrument
saved Dairyland staff a significant amount of diagnostic time.
Finally, following weeks of parametric testing for the implementation of a neural
network-based combustion optimization system, performance engineering staff
determined that the boiler oxygen measurement probes had been experiencing
plugging resulting in an erroneously low indication of excess air. Gradual drift in the
boiler oxygen measurements was confirmed by reviewing the advanced data validation
system results. The data validation system predicted values were subsequently used
for successful training of the neural network saving weeks of repeated testing.
Based on the success of their current installation, Dairyland is looking to use the system
to provide early identification of performance degradation, in addition to expanding the
installation to additional units.
Allegheny Energy Harrison Station Application
Allegheny Energy realized that in order to remain competitive in a deregulated
environment that new approaches to traditional business practices were in order. While
Allegheny historically has been proactive in the performance improvement area,
personnel reduction and turnover has resulted in rethinking performance improvement
methods and turning to computerized performance tools. Two of the tools that were
implemented early in the process were a standardized information database and a
standardized performance monitoring system.
Page 7 of 7
1-156
Implementation of a information database provides Allegheny staff with instant access

to a wealth of plant process data, both at the plant and at the corporate level. With the
incorporation of performance calculations, operations, maintenance and engineering
staff not only have access to data, but to calculated performance results such as turbine
efficiency, boiler efficiency, and unit heat rate. While all of this information is
tremendously useful, it also points out a weakness that is common to all automated
systems: the results are only as good as the incoming plant data. Over time,
transmitters and other process indicators can drift and there is simply not enough
manpower available to continuously check the calibration of critical instruments. With
increased emphasis on heat rate and performance, Allegheny staff realized that
properly calibrated instrumentation is a key to being able to control the generation
process and maintain peak efficiency. Advanced data validation provides an economical
and efficient way to monitor the calibration of important sensors.
At Harrison Station, all historical data is stored in the OSI-PI data base on a DEC Alpha
with an Open VMS operating system. The existing performance monitoring system,
which also resides on the DEC Alpha, retrieves data from the PI database, runs the
performance calculations, and sends results back to the PI system. Since the initial
application of data validation was intended to improve confidence in the performance
calculations, there were a number of options for integration with this system. First, the
data validation system could directly filter data from the plant control system. Second, it
could communicate with the PI system and provide validation information back to the
database. Third, the data validation system could be interfaced directly to the
performance monitoring system, validating only the information used for the
performance calculations.
After considering all of the possibilities, Allegheny opted to configure the system so that
it communicated with the PI database. Since OSI PI is the standard database
throughout the Allegheny system, this interface would allow use of the same interface
software on all of the Allegheny units. In addition, all data from the plant computers are
stored in the PI system, allowing validation of all plant data, whereas only a limited
number of process points are transferred to the performance monitoring system. Figure
3 depicts the integration of the advanced data validation system with the plant database
and the performance monitoring system.
Data validation models were developed for Harrison Station Unit 2 in the spring and
summer of 1999. Installation of the system was scheduled to follow a seven-week
boiler outage and turbine overhaul that began April 1, 1999. In addition to identifying
several data problems prior to the outage, the system was used to identify changes in
the system following the outage.
Page 8 of 8
1-157
Alpha OpenVMS
Unit Control
System
DAS or
Replacement
Value
OSI PI System
DAS Value Storage
ACM Value Storage
Calc Value Storage
DAS
Value
Performance
Monitoring
Process
Calculated
Results
Validated
DAS Value
Intel NT PC
Advanced
Data
Validation
Process
Figure 3 -- Integration of Advanced Data Validation at Alleghenys Harrison

Station
System level displays available from the data validation system easily identified
substantial process change resulting from the turbine overhaul. Figure 4 shows
bullseye plots from both before and after the outage. Each hole in these plots
represents one of the 249 measurements or performance monitoring system results that
were analyzed with the data validation tool. Each ring of the bullseye represents one
standard deviation between the measured value and the data validation system
predicted value.
Figure 4 -- Allegheny Harrison Unit 2 System Level Displays
Page 9 of 9
1-158
In addition to several post-outage findings, the system clearly confirmed an increase in

steam flow (Figure 5) and turbine train pressures resulting from replacement of worn
turbine seals along with a corresponding increase in generation (Figure 6) and decrease
in unit heat rate (Figure 7).
Measured Value
Predicted Value
Dynamic Alarm Limits
Turbine Overhaul
Figure 5 Harrison Unit 2 Steam Flow
Measured Value
Predicted Value
Turbine Overhaul
Figure 6 Harrison Unit 2 Gross Generation

Page 10 of 10
1-159
Measured Value
Predicted Value
Turbine Overhaul
Figure 7 Harrison Unit 2 Net Unit Heat Rate

The system was installed in late-September 1999 and continues to provide real-time
validation for the on-line performance monitoring system.
Other Advanced Data Validation Applications
In the past, the unit operator could easily make adjustments for questionable plant data.
However, with the large quantity of data that is frequently used in advanced control
algorithms, it is impossible for an individual operator evaluate all of the input data
without assistance.
Many facilities are implementing neural network-based optimization systems, which are
designed to optimize many different aspects of plant operation, including NOx, heat rate,
SO2, and overall operating costs. Since these systems are configured based on actual
plant data from parametric testing, it is critical that high-quality data is used for
configuration of the systems and continued operation. Advanced data validation can be
integrated into an optimization program in a number of ways. It can be configured to
work directly with the optimization software and validate only the information that is
being assessed for optimization, or it can be interfaced with the plant data acquisition to
pre-process the data supplied to the optimization system. At a minimum, advanced data
validation technology should be employed to filter the data used in configuring a neural
network. This will reduce the chance of incorporating faulty data into the neural network
model and reduce the need for costly and time-consuming retraining. Further,
Page 11 of 11
1-160
implementing such a system in a closed-loop control mode requires high-quality, realtime data, making on-line advanced data validation essential.
Economics of Advanced Data Validation

In the current power generation market, most business decisions are driven by
economics. If there is not a definitive cost benefit to improving data quality, then the
prudent businessperson would not install a data validation system. In addition, many
utilities are implementing requirements for a one or two-year payback period. Advanced
data validation methodology can often provide full return in less than one year through
the following benefits.
Increases operator confidence in data

Aids management in making accurate O&M decisions
Reduces calibration resources by optimizing calibration scheduling
Reduces engineering analysis resources by identifying the root cause of a
problem (instrument or equipment)
Provides early warning of instrument drift or failure
Provides accurate replacement values for drifted or failed instruments
Provides early warning of process drift or change
Precisely quantifies the amount of instrument/process drift or change
While all of these are important benefits to power generators, only those which can be
assigned a dollar value can be used for cost justification. Examples of the cost benefits
available from advanced data validation are presented below.
Advanced data validation will help to reduce costs by streamlining the calibration
process. Utilities estimate that as many as 1500 man-hours are expended on instrument
calibrations during annual outages. If advanced data validation can reduce this by a
conservative 25%, an additional 375 man-hours would be available for more productive
tasks, such as controls tuning and optimization. Assuming that an instrument technician
with a loaded cost of $35 per hour is performing the calibrations, a reduction of 375
hours per unit results is a direct labor savings of more than $13,000 annually.
For the past 25 years the power generation industry has been on a quest to improve
fuel efficiency and reduce heat rate. One of the cornerstones of this quest has been the
reduction of controllable losses. A tremendous amount of research, time, training,
money and effort has gone into the effort to reduce controllable losses. While this effort
has provided exceptional payback, the weak link is again the primary instrumentation.
The pressure and temperature sensors must provide an accurate indication of the true
process value in order to minimize controllable losses. In a study performed on a single
450 Mw coal-fired generating unit, the impact of historical deviations in just three of the
instruments that impact controllable losses -- throttle pressure, throttle temperature, and
hot reheat temperature - was calculated to be more than $900,000 annually in additional
fuel consumption and replacement power. Since these instruments are part of the highprofile set of controllable loss instruments, it is assumed that higher priority was placed
Page 12 of 12
1-161
on maintenance of these instruments. With an advanced data validation system, these

calibration deviations would be readily detected. Add these tangible savings to the less
tangible, such as increased operator confidence in the input data, and it is clear that
data validation provides high returns.
Conclusions
Advanced data validation has been successfully applied to pre-processing of the plant
data that is used by performance monitoring and control systems and provides these
systems with accurate and reliable input data. This provides increased confidence in
monitoring and optimization system recommendations and reductions in plant operating
costs. In addition, advanced data validation has been used to accurately identify
instruments requiring calibration, refocusing calibration efforts on those instruments that
need attention and reducing instrument maintenance time.
As more utilities apply optimization programs, advanced controls, and other emerging
power plant technologies, the need for the improved accuracy, reliability, and
confidence in data becomes more imperative. Application of advanced data validation
will accelerate the practical and effective use of these advanced methods of plant
automation, helping to minimize the cost of production. Where quality data is critical for
efficient operations and accurate maintenance decisions, implementation of advanced
data validation is essential.
References
1. Application of Advanced Pattern Recognition to Power Plant Condition
Assessment; M.B. Caudill, R.D. Griebenow, E.J. Hansen; 1996 EPRI Heat Rate
Improvement Conference Proceedings, Dallas, Texas; 1996.
2. Advanced Calibration Monitor Users Manual; Performance Consulting Services,
Inc.; Montrose, Colorado; 1998.
3. "Applied Pattern Recognition for Plant Monitoring and Data Validation"; R.D.
Griebenow, E.J. Hansen, A.L. Sudduth; 6th Annual EPRI-ISA POWID; La Jolla, CA;
1995.
4. Similarity Based Regression: Applied Advanced Pattern Recognition for Power
Plant Analysis; E.J. Hansen, M.B. Caudill; 1994 EPRI-ASME Heat Rate
Improvement Conference; Baltimore, Maryland; 1994.
Page 13 of 13
1-162
2
SESSION 2: OPTIMIZATION
2-1
IMPACTS
IMPACTS OF
OF COMBUSTION
COMBUSTION OPTIMIZATION
OPTIMIZATION
ON
ON POWER
POWER PLANT
PLANT HEAT
HEAT RATE
RATE
EPRIs
EPRIsHeat
HeatRate
RateImprovement
ImprovementConference
Conference
Dallas,
TX,
January
30
February
Dallas, TX, January 30 - February1,
1,2001
2001
Carlos E. Romero
Edward K. Levy
Nenad Sarunac
Energy
Energy Research
Research
Center
Center
2-2
COMBUSTION OPTIMIZATION
Basic
Energy
Energy Research
Research
Center
Center
Steps:
Inspect and Maintain Related Equipment.
Define Relationships Between Boiler

Operating Parameters (Boiler Testing).
Identify Optimal Boiler Control Settings.
Maintain Levels of Improvement for NOx,
Combustion
Optimization
Heat Rate and Combustibles (CO and UBC).

2
2-3
Impacts of Comb. Optimization

Common
z
Energy
Energy Research
Research
Center
Center
Performance Impacts:
Boiler Efficiency (Excess O2, UBC, CO, Gas

Temperatures).
Turbine Cycle Heat Rate (Steam
Temperatures, Desuperheating Flow Rates).
Auxiliary Power Requirements (Fan Loading,
Pulverizers Settings).
3
2-4
Impacts of Comb. Optimization
Energy
Energy Research
Research
Center
Center
Combustion Optimization Parametric Impact on Unit Heat Rate
Boiler Component
Heat Rate Impact
Parameters Affected
Boiler
Excess Air
O2, UBC, CO, Tsteam, Tgas, Attemp.,
Pfan,
Turb.
PAux.
Pulverizer
Classification
UBC, CO,
Pmill
Primary Air
UBC, CO,
Pfan
Biasing
O2, CO, Tsteam, Tgas, Attemp.,
Pmill
Burner System
Secondary Air
O2, UBC, CO
Swirl
O2, UBC, CO
OFA
O2, UBC, CO, Tsteam, Tgas, Attemp.
Tilt
O2, UBC, CO, Tsteam, Tgas, Attemp.
Sootblowing
Tsteam, Tgas, Attemp., Pfan
2-5
Effect of O2 on CO and
Fly Ash LOI
Energy
Energy Research
Research
Center
Center
280 M W F W B o iler (T w in F u rn ace)

17
120
Average S ec ondary Air D am per at 65 %
110
16
C O [p p m ]
15
90
80
14
B a s e lin e
70
13
B a s e lin e
60
F ly A s h L O I [%]
100
12
50
40
11
2.0
2.1
2.2
2.3
2.4
2.5
2.6
2.7
A v e ra g e E c o n o m iz e r O 2 [%]
2-6
Effect of O2 on Unit Heat Rate
Energy
Energy Research
Research
Center
Center
300 M W B & W B o iler; F G R C ap ab ility

D e lta Un it He a t R a te [B tu /k W h ]
240
210
180
150
y = 103.06x 2 - 777.35x + 1490.6
120
90
60
H ig h S ta c k L o s s
H ig h L O I
30
0
2 .0
2 .3
2 .6
2 .9
3 .2
3 .5
3 .8
4 .1
4 .4
4 .7
5 .0
A v e ra g e E c o n o m iz e r O 2 [%]
2-7
Effect of FEGT on Unit Heat Rate
Energy
Energy Research
Research
Center
Center
585 M W C E B o iler (S p lit F u rn ace)
20 0
18 0
16 0
14 0
12 0
10 0
80
E x ce s siv e
60
D e su p erh e atin g S p ra ys
40
L o w S tea m T em p e ratu res
20
0
2,1 0 0
2,2 0 0
2,3 0 0
2,4 0 0
2,5 0 0
2,6 0 0
2,7 0 0
2,8 0 0
2,9 0 0
C a lc u la te d F E G T [d e g . F ]
2-8
Effect of Mill Configuration on

Unit Heat Rate
Energy
Energy Research
Research
Center
Center
3 0 0 M W B & W B o ile r; 6 E L -7 6 M ills

Ne u ra l Ne two rk M o d e l R e s u lts
1 0 ,10 0
O 2= 3% , F G R= 45
1 0 ,05 0
1 0 ,00 0
B o tto m M ills
9 ,9 50
U n lo a d in g
T o p M ills
U n lo a d in g
B & C -M ill O /S
B -M ill O /S
F -M ill O /S
D & F -M ill O /S
B & E-M ill O /S
E & F -M ill O /S
E-M ill O /S
9 ,8 50
C & E-M ill O /S
9 ,9 00
D & E-M ill O /S
Ne t Un it He a t R a te [B tu /k W h ]
O 2= 3.6% , F G R= 45
9 ,8 00
-1 .0
-0 .8
-0 .6
-0 .4
-0 .2
0 .0
0 .2
M ill B ia s P a ra m e te r
0 .4
0 .6
0 .8
1 .0
2-9
NOx vs. UNIT HEAT RATE

TRADE-OFF
In General, the Higher the Level of NOx

Reduction, the Greater the Potential for Unit
Performance Impacts.
NOx vs. Unit Heat Rate Trade-Off Affected by:

z
Optimization Objective.
Operational and Environmental Constraints.
Boiler Specifics (Design, Size, Fuel Blend, etc.).
Boiler Operability and Dispatching Demands.
2-10
Energy
Energy Research
Research
Center
Center
Other Impacts (Ash Disposal, Maintenance

Conditions, etc.).
NOX and Unit Heat Rate

Trending in Comb. Opt.
Baseline
Burner tilt / Sec. Air
Sec. Air
NOx, lb/MBtu
0.65
O2
Burner tilt
O2 / burner tilt
9200
Fuel Air
9150
Mills
0.60
9100
0.55
9050
0.50
9000
0.45
8950
"Final"
0.40
8900
NOx
0.35
Heat Rate, Btu/kWh
0.70
Energy
Energy Research
Research
Center
Center
8850
Heat Rate
0.30
8800
0
10
20
30
40
50
60
10
2-11
Typical NOx vs. Unit Heat Rate

Trade-Off
Energy
Energy Research
Research
Center
Center
Net Unit Heat Rate (Btu/kWh)
9200
9150
9100
9050
N o t e : H e a t R a t e s C a l c u la t e d U s i n g t h e H E A T R T C o d e
9000
0 .3 5
0 .4 5
NO
2-12
0 .5 5
x
0 .6 5
E m is s io n s (lb /M B tu )
11
BOILER OP STRUCTURE
Expert System - Guides

Engineer Through a
Series of Boiler Tests,
Builds the Database
Plant Data
Neural Networks Correlate Test Data
Optimization Algorithm Determines Best Control
Settings Satisfying
Optimization Goal and
Operational Constraints
Energy
Energy Research
Research
Center
Center
Boiler
Controls
Expert System
Recommended
Test Conditions
Plant
Engineer
Neural Networks
Optimization
Algorithm
Advice to
Plant Engineer
Personal Computer
12
2-13
UNIT A
Energy
Energy Research
Research
Center
Center
85 MW Tangentially-Fired CE Boiler.
Fires Eastern Bituminous Coal.
LNCFS-III Low NOx Firing System.
Four Burner Elevations Can Operate at Full
Load with Only 3 Mills in Service.
Optimization Objective: Improve Heat Rate and
Find Optimal Control Settings at Different
Target NOx Levels.
13
2-14
NOx vs. Unit Heat Rate TradeOff and Operator Variability
Energy
Energy Research
Research
Center
Center
8 5 M W C E B o ile r; L N C F S -III B u rn e r S ys te m
1 0,2 40
P redic t ed O ptim al S ettings
P aram etric Tes ts
1 0,2 20
1 0,2 00
O p e ra to r V a ria b il ity
1 0,1 80
1 0,1 60
1 0,1 40
1 0,1 20
0 .30
0 .35
0 .40
0 .45
0 .50
NO x E m is s io n R a te [lb /M B tu ]
0 .55
0 .60
14
2-15
Unit A Results
Energy
Energy Research
Research
Center
Center
Optimal Net Heat Rate is a Strong Function of

NOx Below 0.40 lb/MBtu.
Control Settings Traditionally Used by Individual
Operators Caused Heat Rate Penalties From 0
to 65 Btu/kWh.
OEM Control Settings Result in Emission OverCompliance and a 50 Btu/kWh Heat Rate
Penalty.
15
2-16
UNIT B
Energy
Energy Research
Research
Center
Center
Opposed Wall-Fired Supercritical 650 MW Boiler.

Phoenix Combustion Low-NOx Dual Register
Burners With OFA.
Periodic Opacity or CO Excursions Which,
Unchecked, Could Require Unit Derating.
Optimization Objective: Obtain NOx Reductions
with Minimum Impact in Unit Heat Rate, Subject
to Prescribed Constraints.
16
2-17
Air Flow Range Impact on CO

and Opacity
Energy
Energy Research
Research
Center
Center
650 M W F W B o iler; P h o en ix L o w -N O x B u rn ers

120
20
110
CEM CO
100
S tac k O p ac ity
Air Flow Range
18
16
14
80
Baseline Opacity
70
12
60
10
50
40
Baseline Air Flow
30
4
Baseline CO
20
10
0
67
68
69
70
71
72
73
74
S econdary Air Flow [%]
17
2-18
S tack O pacity [%]
C E M C O [ppm ]
90
Minimum Achievable NOx

and Optimal Heat Rate
Energy
Energy Research
Research
Center
Center
650 M W F W B o iler; P h o en ix L o w -N O x B u rn ers

8,520
N et U n it H eat R ate [B tu /kW h ]
8,500
8,480
T est Data Points
8,460
8,440
8,420
B a s e lin e
8,400
8,380
M in im u m
He a t Ra te
C urve
8,360
8,340
8,320
0.50
0.55
0.60
0.65
0.70
0.75
N O x Em issio n R ate [lb /M B tu ]
0.80
0.85
0.90
18
2-19
Unit B Results
Energy
Energy Research
Research
Center
Center
NOx Reduced to 0.60 lb/MBtu with Minimal

Impact on Heat Rate.
Minimum NOx Level of 0.55 lb/MBtu Achieved
Within CO and Opacity Constraints.
Implemented On-Line Closed-Loop
Combustion Control for CO and Opacity
Control.
19
2-20
UNIT C
Energy
Energy Research
Research
Center
Center
600 MW B&W Opposed Wall-Fired Boiler With a

Conventional Dual Register Firing System.
Unit Can Achieve Full Load With 6 of 8 Mills InService While Burning Western Coals.
Top Mill Row Secondary Air Can Be Used as
Simulated OFA.
Optimization Objective: Determine Best Heat
Rate and Minimum Achievable NOx Using Top
Mill Row Secondary Air as Simulated OFA.
20
2-21
Best Heat Rate and Minimum

Achievable NOx
Energy
Energy Research
Research
Center
Center
6 0 0 M W B & W B o ile r; S im u la te d O F A
9,750
28
6
9,720
11
4
10
9,690
8
26
21
24
27
9,660
19
14
1
9
12
13
9,630
P arametric Test
B aseline
S plit F ire
Optima l (S e ssion 65)
9,600
0.15
0.18
0.21
0.24
0.27
2-22
0.30
0.33
21
Unit C Results
Energy
Energy Research
Research
Center
Center
Simulated OFA Achieved a Minimum NOx Level

of 0.161 lb/MBtu (34% Reduction From the
Baseline Conditions).
Feasible to Achieve Simultaneous NOx
Reduction and Heat Rate Improvement (30
Btu/kWh).
Stack Opacity and Fly Ash LOI Were Not
Affected by the Change in Boiler Control
Settings.
CO Found to be a Strong Function of O2 and
Top Row Secondary Air Damper Position.
22
2-23
UNIT D
Energy
Energy Research
Research
Center
Center
300 MW Front-Fired B&W Boilers With a

Conventional Firing System.
Flue Gas Recirculation Fan Capability.
Pneumatic Combustion Control System and
Bailey INFI-90 Digital Control System.
Optimization Objective: Determine Minimum
Possible NOx Between the 75 MW to 300 MW
Load Range and Best Mill Combinations Over
the Load Range.
23
2-24
NOX vs. Unit Heat Rate

Trade-Off Over the Load Range
Energy
Energy Research
Research
Center
Center
300 M W B & W B o iler; F G R C ap ab ility

1 50
3 00
NO x Re d.
HR
2 50
[p p m , c o rr. @ 3 % O 2 ]
NO x R e d u c tio n
1 20
NOx Re d.
2 00
22.5 %
90
1 50
NOx Re d.
15.5 %
60
1 00
30
50
He a t R a te P e n a lty [k J /k W h ]
NO x
35.6 %
200
250
300
Un it L o a d [M W ]
24
2-25
Unit D Results
Energy
Energy Research
Research
Center
Center
NOx Reduced From Over 700 ppm to 550 ppm

(corr. to 3% O2 dry basis) Over the Load Range
(20% Reduction).
FGR Flow Important to Maintaining Steam
Temperatures When Reducing O2 and NOx.
Implemented On-Line Low-NOx Advisory System
for the Operators Which Provides an On-Line
Operator Interface Between Boiler OP and the
INFI-90 DCS for Continuous Low-NOx Advise.
25
2-26
Penalty Box Screen
Energy
Energy Research
Research
Center
Center
26
2-27
CONCLUSIONS
Energy
Energy Research
Research
Center
Center
Changes in Boiler Settings Result in Changes in

Operating Parameters Which Affect Boiler
Efficiency, Turbine Cycle Heat Rate and Auxiliary
Power.
Impacts on Unit Heat Rate Due to Changes in

Control Settings are Typically Limited to Within
100 Btu/kWh.
Additional Up-Front Reduction in Unit Heat Rate

Can Be Achieved Through Boiler Maintenance.
27
2-28
CONCLUSIONS
Energy
Energy Research
Research
Center
Center
Usually, the Higher the Level of NOx Reduction,

the Greater the Potential for Unit Performance
Impacts.
An Evaluation of the Trade-Offs Between Heat

Rate and NOx Emissions Should Consider the
Cost of Fuel and NOx Credits, as Well as Boiler
Constraints and Other Maintenance Costs.
28
2-29
Allegheny Energy Supply

Armstrong P. S.
Obtaining Improved Boiler Efficiency and NOx Using
Advanced Empirical Optimization and Individual
Burner Instrumentation on a Boiler Operated in Load
-Following Mode
E. P. Payson: Allegheny Energy Supply
Rich Brown: EPRI
Dave Earley: Air Monitor Corporation
Carlos Moreno: Ultramax Corp
2-30
Project Description
z
z
An EPRI Tailored Collaboration Project

Merge four technologies into plant operations:
Fuel Flow Measurement- all burners
Secondary Air Flow Measurement-- all burners
Variable orifices to balance primary air and coal flow - 1/3 of the
burners
Optimization Software- UltraMax Corp.s transient optimizer
Host unit is Allegheny Energy Supplys Armstrong #1

Unit
2-31
Armstrong #1 Unit
z
z
z
z
2-32
Foster Wheeler sub-critical boiler rated at 180 MW

Equipped with 12 Foster Wheeler IFS Low NOx Burners
Fires eastern bituminous coal: 124,500 lb/hr at MCR
Two D7-Class ball mills feed the 12 burners
A r m s tro n g
U n it s 1
&
2 B o ile r A r r a n g e m e n t
2-33
Burners
1A1
1A2
5
12
11
10
1B1
7
1B2
11
Coal
Conduits
Coal
Conduits
3 way distributor
Ball mill
pulverizer A
2-34
Ball mill
pulverizer B
10 12
T C Project Goals
z
Merge the new technologies into process control to

improve overall boiler performance and NOx reduction
Determine the impact on performance of each technology as it is
phased in and the overall impact
Determine the extent of fuel and air flow imbalance

Using existing controls and variable orifices on the coal pipes
attempt to balance flows.
Determine to what extent NOx; LOI, and CO can be

reduced by attempting to balance secondary air-to-fuel
ratios at each burner
2-35
Project Instrumentation & Control

z Air
Flow Measurement & Control
Air Monitors Individual Burner Air

Monitor (IBAM)
Sleeve damper control on FW/IFS
burners
z Coal
Flow Measurement & Control
Air Monitor/ Promecon Pf-Flow microwave flow measurement on each

burner
Power & Industrial variable orifices
2-36
IBAM Principal
Pitot-Fechheimer Probe
2-37
Individual Burner Airflow

Measurement (IBAM) Probe
2-38
Burner Register Flow Stratification

Multi-Point
Averaging IBAM Probe
Swirler
Secondary Air
P1
P2
P3
P4
Primary Air/ Coal

Secondary Air
P5
P6
P7
P8
2-39
2-40
Air Monitor Corporation Test

Chamber
2-41
Coal Flow Measurement - Pf Flow

Technology by Promecon
z
Absolute Measurement
Microwaved based
Output from each pipe not dependent on other pipes,
primary air, or other variables
Output not dependent on coal feeder information
No coal sampling required to yield relative flow distribution
Output is directly proportionate to coal flow in each pipe
Scaled Mass Flow

Apply scaling factor to obtain absolute output
Scaling factor may be obtained from gravimetric feed rate input,or
manual sampling of a single coal pipe.
2-42
PF Density Measurement
PF density determined by means of measuring microwave signal attenuation due to the

presence of coal particles.
Ty
Tx
Ry
Rx
=
=
=
=
Transmitter Y axis
Transmitter X axis
Receiver Y axis
Receiver X axis
The transmitter and receiver pair (Ty and Ry) are aligned linearly on the pipe. If installed
alone, the polarization is only in one plane, resulting in a dead spot at 90 to the transducer
pair. Therefore, a second pair of transmitter and receiver transducers (Tx and Rx) are
placed at 90, so that there is measurement in both y and x polarizations. This is essential
in that it allows measurement of the whole pipe cross sectional area.
Technology by Promecon
2-43
Velocity Measurement
2-44
2-45
Minimum Installation Requirements
2-46
ULTRAMAX
z
Sequentially adjusts the control inputs given the

values of the uncontrolled inputs to maximize a
scorecard metric
Learns about process characteristics from data
generated by the process
Does not require prior process models nor prior data
Converges to the optimum even if there has not been
experience with the process there
z
z
Can optimize in a dynamic-transient mode

Can operate in advisory, closed-loop mode or
stand alone mode
2-47
Armstrong
Armstrong P.O.V.
P.O.V. Game
Game Plan
Plan
Decision Diagram
Control Inputs
O2 Setpoint
Top Sleeve Damper Pos.
Mid. Sleeve Damper Pos.
Bot. Sleeve Damper Pos.
A
A
Aux Air A1
Aux Air A2
Aux Air B1
Aux Air B2
M
M
Allegheny
Power
M
M
M
M
M
Armstrong
Power
Station
Unit #1
Boiler Efficiency
NOx
LOI
North A Economizer O2
M
M
M
M
M
M
M
M
M
Uncontrolled Inputs
Time Since Last Soot Blow
FW
Front wall
180 GMW
w/LNB
M
M
M
M
M
M
Advice
North B Economizer O2
South A Economizer O2
South B Economizer O2
AH Gas Out Temp. A
AH Gas Out Temp. B
CO
Opacity
AH Gas Inlet Temp. A
AH Gas Inlet Temp. B
LOI A
LOI B
Precip. Amps 1A1, 1A2, 1B1, 1B2
Precip. Amps 11A, 12A, 13A
Precip. Amps 11B, 12B, 13B
Measurements
(can be automated)
ULTRAMAX
2-48
TC Project Status
z
z
z
z
z
Characterize Project Status as Work In Progress

Initial UltraMax stand-alone optimization performed Nov.
1998
Promecon system installed March 1999. System chosen for
TC because this particular technology appeared to be most
robust, incorporating particulate velocity and absolute
mass flow in all coal conduits
IBAMs installed during Apr-June 1999 outage
UltraMax closed-loop feature installed and demonstrated
March 2000
2-49
TC Project Status Cont.

z
2-50
Currently finalizing validation of Pf-Flow fuel flow

measurement system. Recently installed Reflector Rods
on all coal conduits to mitigate microwave reflections in
conduits.
Use of Pf-Flow velocity measurement enabled a greater
understanding of the combustion process. Were able to
reduce particulate velocities from 95-110 fps in most cases
down to 75-85 fps, and demonstrated enhanced control
over CO formation in the furnace
Two to four of the twelve IBAMS appear to be
malfunctioning. Troubleshooting for problem resolution
TC Project results to date

z
UltraMax stand-alone optimization showed nominal 2530% reduction in LOI; .25-.5% improvement in efficiency
without violating NOx constraint
UltraMax optimization in closed-loop demonstration
yielded sustainable reductions in absolute NOx levels.
About 20% NOx reduction above 165 MW; Nominal 12%
NOx reduction below 150 MW
2-51
Ultramax Results
After 84 readjustments (40 operating hours),
Efficiency improved 0.27% while all other requirements were satisfied.
Variable Name
2-52
Units
Baseline
Optimization
conditions
conditions
Management
Objectives /
Requirements
O 2 Setpoint
3.21
3.24
2.8 < * < 5.5
Top Sleeve Dampers
inches
0.07
0.90
-0.82 < * < 3.
Middle Sleeve Dampers
inches
-0.01
0.92
-0.61 < * < 3.
Bottom Sleeve Dampers
inches
-0.12
0.98
-0.64 < * < 3.
Top Outer Register
% open
50
55
40 < * < 70
Middle Outer Register
% open
50
55
40 < * < 70
Bottom Outer Register
% open
50
50
40 < * < 70
Aux. Air A1
42.4
4.4
10 < * < 60
Aux. Air A2
42.6
46.6
10 < * < 60
Aux. Air B1
42.5
35.4
10 < * < 60
Aux. Air B2
42.6
48.6
10 < * < 60
Boiler Efficiency
88.01
88.28
maximize
NO x
lb./MBtu
0.360
0.352
< 0.45
LOI A
6.2
5.5
< 20
LOI B
7.5
10.9
< 20
CO
ppm
127
429
< 430
A rm stro n g #1 Un it B o ile r
1 1/1 8/9 9 B RO S S T ES T
Au x air im p ac t o n C O
S ta ck G a s C O
1A O 2
1A O 2
1A 1 A ux A ir
1 A 2 A ux A ir
1B O 2
1 B O2
1 B 1 A ux A ir
5 28
511
494
477
460
4 43
426
409
3 92
375
3 58
0.00
3 41
0.0 0
324
0.50
3 07
50.0 0
290
1.00
273
1 00.0 0
2 56
1.50
239
1 50.0 0
2 22
2.00
205
2 00.0 0
188
2.50
1 71
2 50.0 0
154
3.00
137
3 00.0 0
120
3.50
86
3 50.0 0
103
4.00
69
4 00.0 0
52
4.50
35
4 50.0 0
5.00
18
5 00.0 0
1 B 2 A ux A ir
2-53
Armstrong #1Unit Boiler

March 9-10, 2000
Open Loop Operation
9AM-9AM
55
0.5
50
0.45
45
0.4
35
30
0.35
25
20
0.3
15
10
0.25
60
70
80
90
100
110
120
130
140
150
160
170
180
Load MW
Excess Air
2-54
Nox
Poly. (Excess Air)
Poly. (Nox)
190
NOx lb/Mbtu
Excess Air %
40
Armstrong # 1 Unit Boiler

March 15, 2000
Closed Loop Operation (part time)
2 AM-10PM
55
0.5
50
0.45
45
0.4
35
30
0.35
25
20
NOx lb/Mbtu
Excess Air %
40
0.3
15
10
0.25
90
100
110
120
130
140
150
160
170
180
190
Load MW
Excess Air
Nox
Poly. (Nox)
Poly. (Excess Air)
2-55
2-56
Armstrong #1
Oct 08, 7 AM - Oct 09, 7 AM: 2000
1B2 IBAMS
20.000
190
18.000
170
16.000
150
14.000
130
12.000
110
10.000
90
8.000
70
6.000
50
4.000
30
2.000
10
0.000
08-Oct-00 04:48:00 08-Oct-00 09:36:00 08-Oct-00 14:24:00 08-Oct-00 19:12:00 09-Oct-00 00:00:00 09-Oct-00 04:48:00 09-Oct-00 09:36:00
Tim e
GROSS LOAD (MW)
8 IBAM
12 IBAM
10 IBAM
1A WB Pres
1B WB Pres
2-57
A rm strong #1
O ct 0 8 , 7 A M - Oc t 0 9 , 7 A M : 2 0 0 0
IB A M P lo t
120.00 0
140.000
100.00 0
120.000
80.000
100.000
60.000
80.000
40.000
60.000
20.000
40.000
0.000
20.000
130
140
150
160
170
180
190
L oa d M W
2-58
1 IB A M
2 IB A M
3 IB A M
4 IB A M
5 IB A M
6 IB A M
7 IB A M
8 IB A M
10 IB A M
11 IB A M
1 A W B P re s
1B W B P r es
12 IB A M
9 IB A M
Arm strong #1
Oct 08, 7 A M- Oc t 09 , 7 A M: 2 000
A ll 12 C oal Pipes
195
190
185
180
175
170
165
160
155
150
145
140
135
130
125
120
115
110
105
100
95
90
85
80
75
70
65
60
55
50
45
40
35
30
4:48:00 AM
9:36:00 AM
2:24:00 PM
7:12:00 PM
12:00:00 AM
4:48:00 AM
9:36:00 A M
T im e
GROSS LO AD (MW)
Tot Co al Flow
CEM Fuel F low
2-59
Arm stron g # 1
O c t 08, 7 A M - O c t 09, 7 A M : 20 00
1A 1 E XH .
R eflec tor R o ds ins talled S ept 18, 2 000
4 0.000
10.000
9.000
3 5.000
8.000
3 0.000
7.000
2 5.000
6.000
2 0.000
5.000
4.000
1 5.000
3.000
1 0.000
2.000
5.000
1.000
0.000
1 30
0.000
140
150
160
170
18 0
L oa d M W
1 C o a l Flo w
2-60
3 C oa l F lo w
5 Co a l Flo w
1 A1 to ta l
1 A1 cla ss d iff
1 A1 A u x A ir
19 0
A rm st ro ng # 1
Co a l F low v s L o a d
O ct 11 , 7 AM - Oc t 1 2, 7 A M : 20 00
All 1 2 C o al Pip es
170 .0 0
P ro m e co n F u e l F lo w
150 .0 0
130 .0 0
110 .0 0
90 .0 0
70 .0 0
C E M F u e l F lo w
50 .0 0
D es ig n Fu e l F low
30 .0 0
10 .0 0
85
90
95
1 00
10 5
11 0
115
1 20
12 5
13 0
135
1 40
1 45
15 0
155
160
1 65
17 0
175
180
1 85
L o ad ( M W )
T o t Co al F lo w
L in ear ( T o t C o al F lo w )
D es ig n F lo w
L in e ar (C E M F u e l F lo w )
C E M F u el F lo w
2-61
Future Directions
z
z
z
z
z
2-62
Extractive sampling for further validation of the

Pf-Flow instruments
Trouble shoot problematic IBAMs
Evaluate effectiveness of variable orifices on
primary air and coal
Balance secondary air flows using sleeve
dampers
Import measurement values into UltraMax for
ultimate control on mill output dampers, aux air
dampers and burner sleeve dampers to optimize
NOx, CO and LOI
Evaluate transient mode of UltraMax with all
signals.
Tennessee Valley Authority

Energy Research / Technology Advancements
Application of GNOCIS Neural Network
Optimization Controller for
Boiler Efficiency Control
By
Darrell A. Howard
And
Lonnie A. Coffey
January 4, 2001
2-63
Test Results-NOx and Boiler Efficiency

NOx Optimization Test Results
Operation of the GNOCIS system with its interaction with the DCS bias
settings is very stable. Closed loop control was implemented from the beginning
of testing and was the sole mode of operation. The impact on NOx reduction
was achieved without ideal plant equipment conditions. For example, many
sootblowers were not operational. Figure 1 shown below is actual data taken
from a 200 MW T fired unit normal burners and no overfire air operating on
automatic load control. The average load remaining approximately at the same
level throughout the duration of the test. This unit normally burns a high quality
eastern bituminous coal. The operator constraints were relatively conservative.
The optimization system was not allowed to drive the unit operating conditions
into producing CO. The unit CO monitors were not included in the control
scheme on this unit. The CO and NOx indications were used extensively as a
safety indicator to prevent creating overfueling conditions in the twin furnace
balanced draft system, since the excess oxygen instrumentation is affected by
boiler air in-leakage. The performance indicated in Figure 1 has proven to be
typical after two years of operation. The final model is still accurate after two
years and an outage cycle on the unit. In this example, GNOCIS becomes
active at 8:15AM.
Optimizing for NOx
SH and RH NOx Unit 9 GNOCIS Test
SH = Pink, RH = Blue: Units= Lb/MMBTU
1
0.9
0.8
0.7
9CBAI1:B29.PNT
9CBAI1:B25.PNT
0.6
0.5
0.4
0.3
9/17/98 9/17/98
8:00
8:18
9/17/98 9/17/98 9/17/98

8:36
8:54
9:12
9/17/98 9/17/98 9/17/98

9:30
9:48
10:06
FIGURE 1
2-64
9/17/98 9/17/98
10:24
10:42
9/17/98 9/17/98 9/17/98 9/17/98

11:00
11:18
11:36
11:54
NOx and Boiler Efficiency Test Results

Figure 2 shown below was generated from the DCS system historian data, and
documents the test results of NOx and boiler efficiency relationships at a steady
state load situation of approximately 175MW. This particular test was conducted
on July 4, 2000 with a NOx reduction by GNOCIS of approximately 10%, and
boiler efficiency gains of .27% (approximately 27Btu.) GNOCIS constraints
were set conservatively. The actual improvements recorded at the beginning of
the test indicate a NOx level of .495 LB/MMBtu and a boiler efficiency of 87.8%.
These values were obtained by the unit operator allowing some additional boiler
excess air bias of approximately .2% below the constrained value that is
determined by GNOCIS. The NOx and boiler efficiency improvements with the
additional bias allowed by the unit operator, resulted in a NOx reduction of 12%
and a boiler efficiency improvement of .58% or approximately 58 Btu/kWh
F ig u re 2 . G N O C IS T E S T , J U L Y 4 , 2 0 0 0 - 2 0 0 M W " T " F IR E D U N IT
0 .7 5
0 .7 1
0 .6 9
G N O C IS R E T U R N E D T O
C LOSED LOOP AND
D R IV IN G F O R B O T H B .E .
AND NOx
N O T E : A F T E R T E S T IN G O R IG IN A L
O P E R AT O R B IAS E S W E R E N O T R E S E T
R E S U L T IN G IN A L O W E R O V E R AL L
B O IL E R E F F E C IE N C Y A F T E R G N O C IS
W AS R E T U R N E D T O AU T O M AT IC
O P E R AT IO N
8 7 .7
B O IL E R
E F F IC IE N C Y
%
0 .6 7
0 .6 5
NOx #/MM BTU
8 7 .9
8 7 .5
0 .6 3
0 .6 1
0 .5 9
0 .5 7
NOx AND
B .E . V A L U E S
W IT H O U T
G N O C IS IN
O P E R A T IO N
N O x M O V IN G
A V E R A G E (2 0 M IN )
AVER AG ED FOR BOT H
FUR NAC ES
8 7 .3
8 7 .1
S H N O x (2 0 M IN M A )
0 .5 5
0 .5 3
BOILER EFFICIENCY %
A L L G N O C IS A N D
OPER AT OR
B IA S E S R E M O V E D A T
7 :1 5 C D T
0 .7 3
8 6 .9
0 .5 1
0 .4 9
8 6 .7
R H N O x (2 0 M IN M A )
0 .4 7
0 .4 5
7 :0 0 :0 0
0 7 /0 4 /1 0 0
8 :0 0 :0 0
0 7 /0 4 /1 0 0
9 :0 0 :0 0
0 7 /0 4 /1 0 0
1 0 :0 0 :0 0
0 7 /0 4 /1 0 0
1 1 :0 0 :0 0
0 7 /0 4 /1 0 0
1 2 :0 0 :0 0
0 7 /0 4 /1 0 0
8 6 .5
1 3 :0 0 :0 0
0 7 /0 4 /1 0 0
D a t e T im e ( 1 5 m in g rid s )
2-65
Other Optimization Requirements and Benefits
Process sensors are required for real time performance and emissions
Sensor validation provided an important predictive maintenance tool
Process controllers must be available
Neural Network Application
-System must handle unlimited inputs
-Update time critical for non-baseload or large units
-Correlation relationships understood
-Sensitivity analysis understood
-Combining models feature important
Process outside the boiler optimization control system may be affected
Unit net heat rate

Maximum generation capability
Unit AGC response capability (operating under variable pressure control for
NOx, steam temperature, and unit heat rate)
Conflicting Optimization Goals

Normally boiler optimization techniques consist of supplying the correct average
amount of air and fuel to each burner to obtain ideal stiochiometry to produce the
desired process responses. Total fuel and air control demand signals and
dampers for all burners are driven from the same process. This means that any
fuel and air imbalance between burners may generate problem zones that
require individual bias adjustments. Optimization of the combustion process can
create a set of conflicting goals. Normally boiler efficiency and NOx improve
together as dry gas losses are reduced until CO is produced. At this point, NOx
will continue to improve and boiler efficiency decreases.
In some cases, as total air is reduced to minimize NOx and improve boiler
efficiency, steam temperatures may suffer due to boiler design, slagging
conditions, or a number of other factors. Since changes in steam temperature
affect turbine cycle efficiency more than boiler efficiency, it becomes an
administrative decision to determine the acceptable targets for these processes.
A real example of conflicting goals follows. This is a unit where the MW output is
imposed by the turbine steam flow conditions. At rated steam flow, pressure,
and temperature, additional load may be obtained by overfiring the boiler and
using attemperation to keep the RH temperature down within normal limits. The
additional RH flow may generate 2-2.5% more power. The increased steam
from attemperation flow will provide additional generation capacity. NOx and
boiler efficiency cannot be optimized under these conditions. The value of
additional generation may be greater than the boiler efficiency losses many times
over. If the emissions average for the unit is good enough, limits may still be
attained because of the average emission credits in the bank.
2-66
Conclusion
The GNOCIS technology can consistently be applied to reduce NOx and improve
boiler efficiency at the same time while operating on load control. In most cases,
GNOCIS can do a better job than the unit operator can because it calculates and
implements a new set of recommendations every minute. The amount of
improvement is conservatively 10% to 15% NOx reduction and 30- to 40 Btu
improvement on boiler efficiency with very conservative constraints. Continued
research is needed to apply the technology to areas beyond boiler efficiency and
NOx control.
2-67
Heat Rate Improvement at

Dairylands Madgett Station
using NeuSIGHT
TM
Duane Hill, Dairyland Power Coop.

2-68
Madgett Station
Riley Turbo Boiler

375 MW Gross
4 Ball Mill Pulverizers
6 Burners per Mill
Analog Controls
PMAX Performance Monitoring System
2-69
Dairyland HR
2-70
Dairyland HR
2-71
NeuSIGHT Project
Main Goal - Heat Rate Improvement
Occurred in 3 Phases
Phase III implemented a NeuSIGHT in an
Advisory mode
Interfaced through custom displays built by
DPC personnel on PMAX system
Put into service July 2000
2-72
Load Profile
Must Optimize Through Continually
Varying Loads
Typical Profile
Full Load Only When Cows Come Home (early
am, early pm)
170-250 MW during day
100-200 MW during night
2-73
NeuSIGHT Project
Control Points
RH Damper Positions (East/ West)
Excess O2
Mill Coal Flow Bias
Interacts with
Firing Rate
Attemperation Flows
Boiler Cleanliness
2-74
Mill Flow Bias Testing Results

1 /2 7/0 0 HR a t 3 2 5 M W
40
10900
10800
35
10700
10500
20
10400
15
HRn e t BT U/KW
10600
25
mill 61 %
mill 62 %
mill 63 %
mill 64%
Nuhr
10300
Roc elle c oal
Blac k Thunder c oal
10
10200
5
10100
Mill bais ing tes t
58
55
52
49
46
43
40
37
34
31
28
25
22
19
16
13
10
10000
4
0
1
M ill c o a l % o f T o t a l
30
2-75
Multi-Parameter Results
3 5 0 M W 3 .5 Hr c o n t in u o u s t e s t
10700
100
% Total A ir Flow
90
10600
Ne t He a t Ra t e BT U/KW
10500
70
RH Spr ay
kpph
60
10400
70 %
RH Damper
50
60 %
10300
50 %
40
Mills 61&63 bias ed up

[ % total c oal f low ]
30
10200
20
Mills 62&64 c ontr olled low er by Total Fuel c ontr ol
10100
10
5 m in in t e r v a ls
2-76
41
39
37
35
33
31
29
27
25
23
21
19
17
15
13
11
0
1
10000
RH S p ,T o t A ir ,M ills , RH Dm p
80
Net Heat Rate
Nuhr
RH_SPRA Y _FLOW
RH_E_DA MP_62
MILL 61 %
MILL 62 %
MILL 63 %
MILL 64 %
Total A ir Flow
Custom Advisory Display
2-77
Reported Heat Rate Comparison
2-78
Conclusion
To-Date Comparison Shows 1.5% Imp.
Improvements came from Two Areas:
NeuSIGHT responsive to changing unit
conditions
Operator knowledge and willingness to share
insights
2-79
ProcessLinkTM at the Roanoke Valley

Energy Facility
Don Keisling, LG&E
EPRI Heat Rate Conference

Dallas, Texas
January 30, 2001
2-80
Agenda
Introduction
Plant Description
How ProcessLink Works
ProcessLink at ROVA
Initial Results
Conclusions and
Future Directions
2-81
NeuCo, Inc. & Babcock Borsig Power

NeuCo was formed in 1996 and is owned by Charles River
Associates, Advanced Energy (NStar), Babcock Borsig, and
employees.
NeuCo: Statistical, Neural Network Modeling, Software
Engineering, Optimization Methodologies, Chemical &
Combustion Engineering.
Babcock Borsig: Boiler & Environmental technology expertise.
NeuCo & Babcock Borsig joined forces to combine expertise
and deliver integrated solutions to our customers.
2-82
LG&E-Westmoreland Partners
LG&E Energy is a Fortune 500 energy services company
headquartered in Louisville, Ky which owns and operates
businesses in power generation; project development; asset-based
energy marketing; and retail gas and electric distribution services.
LG&E was recently acquired by PowerGen.
Westmoreland Energy is a wholly owned subsidiary of
Westmoreland Coal Company, headquartered in Colorado
Springs.
LG&E-Westmoreland is a limited partnership that jointly owns
ROVA
2-83
ROVA Plant Description

Coal-fired PURPA QF and EWG built in 1994-95.
Sells power under long-term contract and steam to
local host.
Two wall-fired units of 180 MW and 50 MW gross.
Riley boilers with Atrita mills.
Bailey Infi-90 DCS
Overfire air and individual burner shrouds
Both units have flue gas desulfurization systems.
Unit 2 has an SNCR due to stricter permitting
requirements
2-84
On-Line Real Time Boiler Optimization

ProcessLink
identifies the
dozens of bias
and trim settings
needed to
improve the
mixing of the fuel
and air within the
furnace.
ProcessLink
NOx
CO
O2
Flue
Gas
Temp
Increased Efficiency, Improved Emissions Control
2-85
Where ProcessLink Fits

Current Performance
Relative to Design
Conditions
Best Performance Under

Actual Conditions and
How to Achieve it
Performance
Monitoring System
DCS
ProcessLink
Optimizer
PI Data Historian
Optimum Set Points
ProcessLink Information Available on Plant LAN
2-86
ProcessLink at ROVA
Objectives include Boiler Efficiency and Emissions
Control NOx and CO
ProcessLink selected after rigorous evaluation
Unit 2 SNCR is included in the optimization
PI and CO monitors installed as part of the project
Power contract necessitated closed-loop deployment
under full load
Manipulated variables include (but not limited to):
Gross air (boiler O2)

Mill biases (primary air flow and mill outlet temperatures)
OFA
Burner shrouds
2-87
Implementation Steps
Typical Tasks/Milestones of a ProcessLink Project
Work Plan
And Kickoff
Hardware
& Data
Acquisition
Systems
Installed
DCS
Integration
Direct
Search
Optimizer
On-line
Data
Collection
Neural
Optimizer
on-line
Project Duration of 12-16 weeks.
2-88
Training
&
Acceptance
2-89
2-90
2-91
2-92
2-93
2-94
Summary of Initial Results

Installed and commissioned with zero plant downtime
Initial 0.3% efficiency gain after just 2 months on both
units
Improved control of NOx on Unit 1 and CO on both
units
ProcessLink often able to control CO much more
effectively than operators
Reduction in average Urea flow rates on Unit 2 (at
times limited by need to control CO)
Further work underway to realize additional benefits
2-95
Unit 1 Times Series Analysis

Analysis Timeframe
NOx #/mmbtu
Pre DOE
DOE
Neural w/CO Constraint
CO #/mmbtu
Pre DOE
DOE
Average O2
Pre DOE
DOE
Minimum O2
Pre DOE
DOE
Boiler Effic (Abbr Losses)
Pre DOE
DOE
2-96
Period
Mean
Std Deviation
Comparison
Abs Delta Rel Delta
2/18 - 3/3
3/4 - 4/21
4/22 - 6/26
0.290
0.306
0.304
0.0209 Delta DOE - Pre DOE

0.0278 Delta Neural - DOE
0.0265 Delta Neural - Pre DOE
0.0161
-0.0014
0.0147
5.55%
-0.46%
4.80%
2/18 - 3/3
3/4 - 4/21
4/22 - 6/26
N/A
0.171
0.232
N/A Delta DOE - Pre DOE

N/A
0.0609
N/A
N/A
26.22%
N/A
2/18 - 3/3
3/4 - 4/21
4/22 - 6/26
2.929
2.899
2.736

-0.0303
-0.1632
-0.1935
-1.04%
-5.97%
-6.68%
2/18 - 3/3
3/4 - 4/21
4/22 - 6/26
1.989
2.222
2.274

0.2331
0.0524
0.2855
11.72%
2.30%
12.85%
2/18 - 3/3
3/4 - 4/21
4/22 - 6/26
83.399
83.437
83.796

0.0388
0.3587
0.3975
0.05%
0.43%
0.48%
Unit 1 Boiler Efficiency (Abbr. Losses)

Blr Eff (Abbr Losses) All Data
85
84.8
84.6
84.4
84.2
84
83.8
83.6
83.4
83.2
83
4-Feb-00
24-Feb-00
15-Mar-00
4-Apr-00
24-Apr-00
14-May-00
3-Jun-00
23-Jun-00
13-Jul-
DateAndTime
2-97
Unit 1 Net Plant Heat Rate

ROVA 1 C o r r e c t e d N e t He a t Ra t e [ P o s t Out a g e ]
11000
10750
10500
10250
10000
9750
9500
9250
9000
29-May-00
3-J un-00
8-J un-00
13-Jun-00
Da t e A ndT i me
2-98
18-Jun-00
23-Jun-00
28-Jun-00
Post-Outage NOx Trend for Unit 1

Unit 1 CHIM NOX lb/mmBTU Post Outage
0.45
0.4
#/mmbtu
0.35
0.3
0.25
0.2
0.15
29-May-00
3-Jun-00
8-Jun-00
13-Jun-00
18-Jun-00
23-Jun-00
28-Jun-00
DateAndTime
2-99
Unit 2 Times Series Analysis

Analysis Timeframe
Period
Total Urea Flow
Pre DOE
2/18 - 3/3
DOE
3/4 - 5/30
Neural w/CO Constraint 6/1 - 6/26
CO #/mmbtu
Pre DOE
2/18 - 3/3
DOE
3/4 - 5/30
Average O2
Pre DOE
2/18 - 3/3
DOE
3/4 - 5/30
Minimum O2
Pre DOE
2/18 - 3/3
DOE
3/4 - 5/30
Boiler Effic (Abbr Losses)
Pre DOE
2/18 - 3/3
DOE
3/4 - 5/30
2-100
Mean
Std Deviation
Comparison
Abs Delta Rel Delta
18.253
18.910
18.739

0.6569
-0.1716
0.4852
3.60%
-0.92%
2.57%
N/A
0.261
0.149
N/A Delta DOE - Pre DOE

N/A
-0.1124
N/A
N/A
-75.60%
N/A
2.534
2.677
2.484

-0.0303
-0.1632
-0.1935
-1.04%
-5.97%
-6.68%
0.951
2.177
2.116

0.1424
-0.1926
-0.0502
5.62%
-7.76%
-1.88%
84.614
84.880
85.408

0.2664
0.5275
0.7939
0.31%
0.62%
0.94%
Unit 2 Boiler Efficiency (Abbr. Losses)

Unit 2 Blr Eff (Abbr Losses) All Data
86
85.8
85.6
85.4
85.2
85
84.8
84.6
84.4
84.2
84
4-Feb-00
24-Feb-00
15-Mar-00
4-Apr-00
24-Apr-00
14-May-00
3-Jun-00
23-Jun-00
DateAndTime
2-101
Unit 2 CO During Period for Which

Monitoring Was Available
U2 BGHSE OUT CO #/MMBTU Post Outage
1
0.9
0.8
#/mmbtu
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
14-May-00
19-May-00
24-May-00
29-May-00
3-Jun-00
8-Jun-00
DateAndTime
2-102
13-Jun-00
18-Jun-00
23-Jun-00
Post-Outage Unit 2 Urea Usage

Unit 2 Total Urea Flow Post Outage
50
45
40
35
gal/hr
30
25
20
15
10
5
0
14-May-00
19-May-00
24-May-00
29-May-00
3-Jun-00
8-Jun-00
13-Jun-00
18-Jun-00
23-Jun-00
28-Jun-00
DateAndTime
2-103
Current SNCR Optimization

Focused on minimizing NOx-in as function of urea
flow performance objective
No current NOx measurement before reactor
Overall urea flow determined by NOx setpoint
No ability to modulate urea flows by injector or level
Demonstrated ability to reduce urea flow
Benefits somewhat limited by need to manage CO
O2 Trim is vital to obtaining benefits with current
regime
2-104
Future Enhancements to
SNCR Optimization
NOx monitoring before SNCR inlet
Control of flow rates through individual nozzles and/or
vertical levels
Flow rates need to be incorporated for control via DCS
Temperature sensitivity of chemical reaction and
tuning experience indicates substantial potential upside
Continued progress with controlling CO will allow
more room for minimizing urea usage and heat rate
2-105
ProcessLink is providing measurable ongoing boiler

efficiency gains while managing emissions to defined
constraints .
CO limits boiler efficiency gains but is effectively
controlled by ProcessLink.
Not all optimization technologies are the same.
Auto-retuning essential for sustained benefits.
Further benefits can be achieved by enabling
additional controllable parameter and controls to
SNCR.
Optimization more is a journey, not an event!
Conclusions
2-106
Automatically Control NOx,

With Heat Rate Constraints, in
a Coal-Fired Power Plant
Kandi Forte
Conemaugh Generating Station
Reliant Energy
Tom Cowder
Keystone Generating Station
Reliant Energy
Russell F. Brown
Pavilion Technologies, Inc.
2-107
Outline
2-108
Multivariable control
Dynamic control
Neural-Net APC
Reliant Keystone and Conemaugh
Control objectives
Results
Summary
Multivariable Control
Many handles" for NOx control
Lowering NOx impacts other things
Constraints
2-109
Dynamic Control
Fast disturbance rejection sootblowing, coal changes
Control load following units
Change setpoint quickly (but who
cares?)
2-110
Neural-Net APC
Steady-state neural network model

Steady-state optimizer
Dynamic model
Control model
Dynamic optimizer
2-111
Knowledge Guided Training

Correlated data
Some knowledge of process
Enforce known relationships
2-112
Reliant Conemaugh And

Keystone
CE 900 MW T-Fired supercritical

Concentric firing retrofit
Low NOx burners & OFA
ABB/CE P2 low NOx burner
Honeywell TDC 3000 DCS
PHD historian upgrade
Extensive use of multiple fuel supplies
2-113
Control Objectives
2-114
Control or minimize NOx emissions

CO below a maximum constraint
Control reheat temperature
Reheat temperature difference <25F
Minimize dry gas loss
Constraint on the auxiliary air damper
Conemaugh Results
Reduced NOx Rate by 15% in 1999
using Open-Loop Control
Reduced NOx Rate an additional 2 to
5% in 2000 using Dynamic Closed-Loop
Control
2-115
Conemaugh
C o ne ma ug h S ta tio n O zo ne S e a so n B o ile r
B o ile r Effic ie n c y D e lta
E ffic ie n cy
4%
3%
2%
1%
0%
Ap ri l
U ni t 1 199 8
2-116
May
Un it 2 1 9 9 8
Ju ne
U nit 1 1 99 9
Jul y
U ni t 2 199 9
Conemaugh
P rec ip ita to r In le t L O I
9.4 9
12
10
% L OI
7 .13
11 .45
8.1 0
U nit
1 99 9 J uly A vg
1 99 8 J uly A vg .
2-117
Keystone Unit 2, NOx Without

APC
Without APC control on NOx rate, 1998 Ozone Season.
2-118
NOx, Open Loop SS Optimization
APC open loop SS control on NOx, 1999 ozone season.
2-119
Closed Loop Dynamic Control, Set-Point Mode
APC closed loop control (setpoint mode) on NOx, 1999 ozone

season
2-120
NOx Minimization Mode
APC closed loop control (NOx minimization mode) on NOx, 1999 ozone
season.
2-121
No APC
Without APC control on NOx rate, 1998

Ozone Season.
2-122
Without APC control on reheat

temperature (top) and CO (bottom), 1998
Ozone Season.
Open Loop SS Control: NOx, CO, Reheat
APC open loop control on NOx, 1999

ozone season.
APC open loop control on CO (bottom),

1999 ozone season. Reheat (top) still
controlled by DCS.
2-123
NOx Dynamic Control, Setpoint Mode
APC closed loop control (setpoint mode)

on NOx, 1999 ozone season.
2-124
APC closed loop control (setpoint mode) on

reheat temp. (top) and CO (bottom), 1999
ozone season.
NOx Minimization Mode
APC closed loop control (NOx

minimization mode) on NOx, 1999 ozone
season.
APC closed loop control (NOx minimization

mode) on reheat temperature (top) and CO
(bottom), 1999 ozone season.
2-125
Summary
Dynamic control > steady state
optimization > no control
Allows trade-offs -- NOx vs. steam
temperatures, efficiency
2-126
Further Work
Reliant rolling out to total of 10 boilers
At Keystone, added soot-blowing and
opacity
More results appeared in Power
Engineering, and will appear in EPRI
TC report
2-127
Unit Optimization at
Hammond Unit 4
John Sorge
Southern Company
EPRI Heat Rate Improvement Conference

January 30 - February 1, 2001
Dallas, Texas
2-128
Project Participants
Funding
EPRI, PowerGen, Southern Company
U.K. Department of Trade and Industry
U.S. Department of Energy
Participants
EnTEC
PowerGen
Southern Company
Tennessee Tech
URS / Radian
2-129
Unit Wide Optimization

Maximizing unit return
Fuel, emissions, ash marketing ...
Using real-time controllable process parameters
Complications
Complex, non-linear, non-stationary processes
Many important parameters are difficult to measure in
real-time (LOI, heat rate, others)
Process interaction
2-130
Excess Oxygen
Min Airflow
Requirements
Optimization Envelope
Fan Limitations
ESP Limitations
Opacity Problems
Mass Emissions
$O
ptim
al O
NO
xO
Nom
ptim
inal
O
al O
2
2
Low Steam Temperatures
UBC Limits
0%
Load
100%
2-131
Project Overview
Desired
Targets
Boiler
Optimization
Unit
Optimization
ESP
Optimization
Possible Operating
Envelope
Intelligent
Sootblowing
Steam Cycle
Optimization
Online Heat
Rate
Process Data Server

LAN
DCS
DAS
PCs
PLCs
PLCs
Field I/O
Plant
2-132
Boiler Optimization Package

Issues
Current optimization system running closed-loop
Want improvements but at low risk of breaking
Approach
Upgrade boiler optimization system to latest version

Investigate adding additional manipulated variables
Implement online model error correction
Improve what-if capabilities
2-133
Steam Cycle Package

Issues
As compared to boiler:
Fewer control handles
More deterministic and more linear
Steam conditions have high impact when no cost

attributed to NOx; more when NOx is considered
Approach
Looked at several tools
Boiler optimization tools are directly applicable
2-134
ISB Package
Issues
Non-continuous and many variables

Difficulties in creating robust models
Probable significant interaction with boiler package
Hardware (control elements and sensors) problems
Approach
Fuzzy-Logic Rule Based Method
Advantages
Relative low cost, complexity, and risk
Models an expert instead of the process
Disadvantages
Models an expert instead of the process
2-135
ESP Package
Issues
Non-continuous control variables
Performance affected by upstream conditions
Full controls not integrated into the DCS
Approach
Install PCAMS / ESPert
ESPert
1st principles model of ESP
Knowledge of ESP operating state
PCAMs
ESP supervisory control system
Employs low level optimization
2-136
On-Line Heat Rate Package

Objective
Provide real time heat rate and boiler efficiency
calculation for use in developing models
Tennessee Tech developed package

Two methods
Conventional output/loss (direct)
Output/loss method with CEMs data (indirect)
Makes prediction of some fuel properties
2-137
Unit Optimization Package

Purpose
Produce minimum cost operation
Maintain operating constraints
Possible Approaches
Global
Hierarchical
Each satellite optimizer gives different control settings
Each satellite associated with different plant item
Some method of reconciling these different settings is required
2-138

Cost Function
Each M satellite processes must have a cost
associated with it
Unit cost expression function of these high level plant
variables
2
2
+C Boiler
COverall =C ESP
C ESP = f ( x1 , x2 ,..., y1 , y2 ,...)

C Boiler = f ( x1 , x2 ,..., z1 , z 2 ,...)
Two dimensions
Many dimensions (N)
2-139

Methodology
Minimize unit cost function treating high level plant
variables as independent with constraint on variable
change
Use recommended plant settings as extreme points of
convex set
Full unit cost optimisation using convex set as search
space - maintains plant constraints
Reduces problem from M dimensions to N dimensions
2-140
Summary
Benefits
Flexible, cost effective method for improving plant
operating margins
Keys to success
Obtaining accurate accounting of plant costs
Unit must not be in critical condition
User acceptance
2-141
3
SESSION 3: INTELLIGENT SOOTBLOWING
3-1
Effects of Sootblowing in Coal-Fired

Boilers on Unit Heat Rate and NOx
Presented at EPRIs Heat Rate Improvement Conference
January 30 - February 1, 2001
Carlos E. Romero
Nenad Sarunac
Edward K. Levy
Energy Research Center
Lehigh University
3-2
SOOTBLOWING AFFECTS:
Slag and Fouling Deposits

Gas Temperatures, Steam Temperatures and
Attemperating Sprays
NOx Emissions
Stack Opacity
Other Impacts
3-3
Effect of Sootblowers on
Steam Temperature (Boiler A)
Im p act o f IR s an d IK s o n H o t R eh eat, B o iler A
55 0
50
T e m p e ra tu re (d e g . C )
A v e ra g e Ho t R e h e a t S te a m
40
35
54 0
D e sign H T R T = 538 C
30
53 5
25
20
IK 63 (S HT - North & S outh S ides )
53 0
15
IK 58 (RHT - North & S outh S ides )
10
52 5
5
52 0
0
6:00
7:12
8:24
9:36
10:48
T im e (A p ril 2 7 , 2 0 0 0 )
3-4
12:00
13:12
14:24
Va lv e P o s itio n [%]
45
S ide IR's
54 5
R e h e a te r A tte m p e ra tin g S p ra y
Hot Corners IR's
Effect of Sootblowers on
Steam Temperature (Boiler B)
Hot Reheat Temperature (F)
IK 1-2, 5-6, 9-10, 11-14 Tests (2/96)

Boiler B
1,040
IK 1-2
RHTCF
1,020
IK 5-6
LTSHCF
HTSHCF
1,000
IK 9-10
IK 11-12
980
IK 13-14
960
60
65
70
75
80
85
90
HTSHCF, RHTCF, LTSHCF (%)
3-5
Effect of Wallblowers on
Performance (Boiler C)
Wall Blowers C and D Row, Boiler C
8,950
HEAT RATE
1,000
8,900
980
MAIN STEAM
960
8,800
940
HOT REHEAT
920
8,750
0
0.5
Time (hrs)
3-6
8,850
1.5
Heat Rate (Btu/kWh)
Steam Temperatures (F)
1,020
NOx Emissions (Boiler D)
E ffect o f S lag g in g o n E m issio n s, B o iler D
0 .9 0
0 .8 5
0 .8 0
0 .7 5
0 .7 0
0
E la p s e d T im e [h r]
3-7
NOx Emissions (Boiler E)
W A L L B L O W E R S E L E V . 101 ft., B O IL E R E
0.40
2,900
S o o tb lo w
0.39
2,890
(25 O /S )
S o o tb lo w
0.38
2,880
IR #18 ,2 1,2 3,26,29

(31 O /S )
0.37
2,870
0.36
2,860
0.35
2,850
0.34
2,840
0.33
2,830
0.32
2,820
0.31
2,810
0.30
6:57
2,800
7:33
8:09
8:45
9:21
9:57
10:33
11:09
T im e (No v . 0 2 , 1 9 9 9 )
3-8
11:45
12:21
12:57
F E G T [d e g . F ]
IR #17 ,1 8,1 9,26,27
SOOTBLOWING
OPTIMIZATION IS NEEDED TO:
Maintain Appropriate Level of Slagging and
Fouling in Various Heat Transfer Sections

Minimize Impact on Unit Heat Rate
Reduce NOx Emissions
Minimize Opacity Excursions
Meet Other Objectives
What portions of the boiler to clean and
on what schedule, considering the
trade-offs between NOx, steam temp.,
heat rate, etc.
3-9
SOOTBLOWING OPTIMIZATION
METHODOLOGY
Instrumentation/Calculations Setup
Sootblower Characterization
Data Analysis
Development of Sootblowing Strategy
Implementation and Evaluation of the
Strategy
3-10
UNIT B
CE Tangentially-Fired, 108 MW Boiler.
Subcritical, Single Reheat Unit With
Conventional Burners (Original Firing
System).
Unit Equipped With Cold and Hot ESP
VAX-Based PMW Available for Data
Archiving and Cleanliness Calculations.
3-11
Sootblower Locations at
Boiler B
RHTR
IK 1&2
HIGH
TEMP
SUPER
HEAT
IK 5&6
IK 9&10
IK 11&12
LOW TEMP
SUPERHEAT
24 25
IRs EL 96
23
26
22
27
21
28
IK 13&14
ECON
30 29
FURNACE
14 15
IRs EL 85
13
16
12
17
11
18
Hot
Corner
3-12
20 19
E
Cold
Corner
Odd numbered IK blowers

are located on the south
side of the boiler.
Sootblower Characterization,
Boiler B
B o iler B ; W all B lo w ers - H o t C o rn er
1,120
85
M ain S te am
Rehe at S te am
S te a m T e m pe ra ture [de g. F ]
1,080
80
1,060
1,040
75
1,020
1,000
0.00
S e c tio n C le a n lin e s s F a c to r [%]
WWCF
1,100
70
0.25
0.50
0.75
1.00
1.25
Tim e [hr]
3-13
Boiler B
S o o tb lo we r C h a ra c te riz a tio n T e s ts , B o ile r B
W a ll B lo w e r s
S IDE W A LL: IR 17
1 ,0 4 0
CO LD CORNE R: IR 18-19
[d eg . F ]
H o t R eh eat S team T em p eratu re
1 ,0 5 0
1 ,0 3 0
IK 1& 2
1 ,0 2 0
HO T CO RNE R: IR 11-20
1 ,0 1 0
1 ,0 0 0
0 .0 0
0 .2 5
0 .5 0
0 .7 5
T im e [h r]
3-14
1 .0 0
1 .2 5
Boiler B
S o o tb lo w er C h aracterizatio n T ests, B o iler B
1 .0 2
N o rm alized N O x [lb /M B tu ]
1 .0 0
0 .9 8
0 .9 6
Ho t C o rne r: IR 1 1 -2 0
0 .9 4
C o ld C o rne r: IR 1 8 -1 9
C o ld C o rne r: IR 2 8 -2 9
R e trac tab le s : IK 1 -2
0 .9 2
70
75
80
85
90
95
100
W ater W all C lean lin ess F acto r, C F -W W [%]
3-15
Data Analysis, Boiler B
NOx
Reduction
[%]
Change in
Main Steam
Temperature
[ooF]
Change in
Reheat
Steam
Temperature
[o F]
Change in
Unit Heat
Rate
[Btu/kWh]
Hot Corner
-45
-40
+110
Cold Corner
-35
-30
+85
Furnace Side
-25
-20
+60
IK 1-2
-15
-25
+50
IK 5-6
+2
-10
IK 9-10
-15
+20
+1
IK 11-20
+7
+7
-15
IK 13-14
+15
+7
-30
Sootblowers
3-16
Data Analysis,Boiler B
N O x E m is s io n R a te [lb /M B tu ]
N O x vs. F u rn ace E xit G as T em p eratu re, B o iler B

0 .7 0
0 .6 8
0 .6 6
0 .6 4
0 .6 2
0 .6 0
0 .5 8
1 ,8 4 0
1 ,8 8 0
1 ,9 2 0
1 ,9 6 0
2 ,0 0 0
2 ,0 4 0
C a lc u la te d F u r n a c e E x it G a s T e m p e r a tu r e [d e g . F ]
3-17
Data Analysis, Boiler B
Heat Rate Penalty (Btu/kWh)
NOx vs. Heat Rate Tradeoff, Boiler B

Low-NOx Mode
150
0.55
Heat Rate Penalty
0.52
90
60
0.49
30
0.46
NOx
-30
0.43
-60
Off-Line Heat Rate Penalty due to changes in Tms and Thr.
-90
0.40
60
65
70
75
WWCF (%)
3-18
80
85
NOx (lb/MBtu)
120
UNIT C
CE Tangentially-Fired, 585 MW Boiler.
Supercritical, Single Reheat Unit Retrofitted
With CE LNCFS-III Low-NOx System.
Unit Subject to Opacity Excursions.
VAX-Based PMW Available for Data Archiving
and Cleanliness Calculations.
3-19
Boiler C
Im p a c t o f E C O N C F o n S ta c k O p a c ity , B o ile r C
20
18
S ta c k O p a c ity [% ]
16
14
12
10
8
6
4
2
0
50
55
60
65
70
75
E c o n o m iz e r C le a n lin e s s F a c to r [% ]
3-20
80
85
Data Analysis,Boiler C
W W C F vs. H eat R ate T rad eo ff
1 ,0 2 0
H R S te a m T e m p . S e tp o in t
200
1 ,0 0 0
H o t R e h e a t S te a m T e m p .
980
150
960
940
100
920
900
50
880
D e lta U n it H R
860
0
[d e g . F ]
1 ,0 4 0
Ho t R e h e a t S te a m T e m p e ra tu re
250
840
70
75
80
85
90
95
100
C a lc u la te d W W C F [%]
3-21
UNIT E
B&W Opposed Wall-Fired, 640 MW Boiler.
Supercritical, Double Reheat Unit Retrofitted
With B&W DRB-XCL Low-NOx System.
Fires US and Foreign Coals.
Unit Constrained by NOx and CO Limits of
0.45 lb/MBtu and 160 ppm, Respectively.
PI-DAS and FEGT Instrumentation Available.
Reheat Attemperation Used for Extra Load
Generation.
3-22
Boiler E
W allb lo w er C h aracterizatio n fo r N O x , B o iler E
1 .0 4
N o rm a liz e d N O x
1 .0 2
1 .0 0
0 .9 8
0 .9 6
0 .9 4
8-IR20,22,23,24,28,29,30,32
S o o tb lo w in g
8-IR1,2,3,8,9,10,11,16
6-IR17,18,19,25,26,27
0 .9 2
5-IR17,18,19,26,27
4-IR20,24,28,32
0 .9 0
0
10
15
20
25
30
35
40
45
50
55
60
E la p s e d T im e [m in ]
3-23
Data Analysis,Boiler E
F u rn ace E xit G as T em p eratu re vs. N O x E m issio n s, B o iler E
0.440
June 26, 2000
0.430
A verage O 2 = 2.27 + /- 0.13%
0.420
0.410
y = 0.0006x - 1.1381
0.400
R 2 = 0.5718
0.390
June 22, 2000
0.380
A verage O 2 = 1.98 + /- 0.07%
0.370
y = 0 .0 0 0 7 x - 1 .6 6 7 6
2
R = 0 .6 7 1 3
0.360
2,740
2,750
2,760
2,770
2,780
2,790
2,800
F u rn a c e E x it T e m p e ra tu re [d e g . F ]
3-24
2,810
2,820
Data Analysis,Boiler E
F u rn ace E xit G as T em p eratu re vs. U n it L o ad , B o iler E
6 34
6 33
A ve ra ge E xce ss O 2 = 2.2 7 +/- 0.13 %
6 32
Un it L o a d [M W g ]
6 31
6 30
6 29
6 28
6 27
6 26
6 25
6 24
6 23
6 22
6 21
6 20
2 ,73 0
2 ,74 0
2 ,75 0
2 ,76 0
2 ,77 0
2 ,78 0
2 ,79 0
2 ,80 0
2 ,81 0
2 ,82 0
F u rn a c e E x it G a s T e m p e ra tu re [d e g . F ]
3-25
DEVELOPMENT OF
SOOTBLOWING STRATEGY
Use sootblower characterization test data to
create database concerning the effect of
individual sootblowers and sootblower groups on:
Boiler section cleanliness

Steam Temperatures
Attemperating Sprays
Gas Temperatures
NOx Emissions
Opacity
Develop sootblowing strategy that satisfies

optimization objective (goal) and unit constraints.
3-26
Example of a Sootblowing
Schedule, Boiler B
After the 4th mill is put on-line, alternate the activation of all
side blowers. (Avoid Steam Temp. Overshoot)
Once the unit is at full-load and settled-out, bring the WWCF to
about 85% by activating the sootblowers at the lower elevation
corners. (Thermal NOx Reduction)
About one hour into the steam temp. transient, alternate the
activation of IKs 11 to 14. (Steam Temp. Recuperation)
If the WWCF drops below 80%, activate the higher elevation
corner sootblowers. Otherwise activate them before unit cycles
back to minimum load. (Slagging and NOx Control)
At operator discretion, activate HT superheater and reheater
blowers on a one-a-day frequency. (Fouling Control)
Recommended daily frequency: 5 Wallblower (IRs), 1 to 2 Retractables (IKs).
Heat Rate Penalty Due to Steam Consumption = 3 Btu/kWh
3-27
Temperature Overshoot During

Load Ramp-Up
L o a d R am p -U p , B o iler B
120
1,120
Unit L o ad
Main S te am
1,100
R e he at S te am
1,060
U n it L o a d [M W ]
Tmst,de s ign
1,040
100
1,020
Trht,de s ign
1,000
90
980
960
80
0.00
940
0.25
0.50
0.75
1.00
T im e [h r]
3-28
1.25
1.50
1.75
2.00
S te a m T e m p e ra tu re [F ]
1,080
110
Sootblowing Impact on Steam

Temperature During Ramp-Up
S o o tb lo w in g E lim in ates T em p eratu re Oversh o o t D u rin g
L o ad R am p -U p , B o iler B
M a in S te a m T e m p e ra tu re [F ]
1 ,100
1 ,050
1 ,000
1 /3 0 /9 6
1 /3 1 /9 6
2 /4 /9 6
9 50
0 .0
0 .5
1 .0
1 .5
2 .0
T im e [h r]
3-29
IMPLEMENTATION OPTIONS
SCHEDULE DRIVEN:
MANUAL:
Provide written schedule to the operator.
SEMI-AUTOMATED/AUTOMATED:
Plant control system generates alarms to prompt the
operator or automatically activates sootblowers at
appropriate times.
INTELLIGENT SOFTWARE:
Artificial-based system provides expert advice or

activates sootblowers.
3-30
Example of a Sootblowing
Schedule Evaluation, Boiler C
O p tim ized S o o tb lo w in g E valu atio n , B o iler C
40
1 00
37
90
HE P enalty
34
80
31
70
28
60
25
50
22
40
19
30
16
20
13
10
10
He a t R a te P e n a lty [B tu /k W h ]
W a te rwa ll C le a n lin e s s F a c to r [%]
W W CF
0
0
10
15
20
25
30
35
40
45
50
55
60
65
70
Nu m b e r o f D a ys
3-31
INTELLIGENT SOOTBLOWING
SOFTWARE CONCEPT
R e t ra c t a b le
S o o t b lo w e r s
DA T A BA SE
Live
P lant Data
Fur nace
EXPERT SYST EM
W a l lb lo w e r s
Exp e r t A d vice
A dvic e and
W arnings
W hat-If
A naly s is
Op e r ato r
3-32
Schematic of the Intelligent

Sootblowing Approach
SOOTBLOW CHARACTERIZATION
(Field Tests)
DATABASE
(Emissions, Performance,
Opacity)
1. Select Optimization Goal

2. Define Goal Functions
3. Identify Operating Constraints
Expert
Knowledge
LIVE
PLANT
DATA
MEMBERSHIP FUNCTIONS
EXPERT SYSTEM
Expert Advice
or Action
3-33
CONCLUSIONS
The ERC has Developed and Successfully

Implemented Optimized Sootblowing Schedules at
Several Utility Boilers.
Sootblowing Optimization Helps to Determine the

Optimal Balance Between Slagging and Fouling
Levels and Steam Temp., Attemperation, NOx, etc.
Sootblowing Operation Can be Optimized for

Different Objectives (i.e., to Help Reduce the Cost
of Controlling NOx During the Ozone Season).
The ERC is Currently Developing Intelligent

Sootblowing Software.
3-34
Optimization of Boiler Sootblower Operation
Jeffrey Williams
Xu Cheng
Westinghouse Process Control, Inc.
Bernie Begley
Alex Smith
Dale Hopkins
Southern California Edison, Inc.
[W estinghouse ProcessControl,
Inc.
A Fisher-Rosemount Company
3-35
Overview - Intelligent Sootblower Scheduling

Project Outline
Model-Based Cleanliness Factor Calculation
Sootblower scheduling Optimization Scheme
Opacity Reduction Mode
Steam Temperature Control and Steam Saving Mode
Implementation Results
Summary
Inc.
3-36
Intelligent Sootblowing Goals

z
Maintain clean heat exchanger surfaces
Enhance heat transfer efficiency
Avoid excessive sootblowing
Improve steam temperature control
Avoid opacity spikes and generate steam

savings
Inc.
3-37
System Overview
Firing Rate
Heat Content
Air Flow
Feedwater
Furnace/Windbox DP
Flue Gas Temperature
Burner Tilt
Spray
Boiler
Boiler
Section Temperature &
Pressure Measurements
Step 1
Step 2
Neural Network
Neural Network
Model
Model
Neural Network
Neural Network
Model Training
Model Training
Actual
Actual
Enthalpy
Enthalpy
Calculation
Calculation
(Steam
(SteamTable)
Table)
Performance
Performance
Monitoring
Monitoring
Actual
ActualHeat
Heat
Absorption
Absorption
Expert Rule
Expert Rule
Formation
Formation
Steam Temp.
Opacity
Desired Cleanliness
Ideal Heat
Ideal Heat
Absorption
Absorption
Cleanliness
Cleanliness
Factor
Factor
Calculation
Calculation
Sequence Selection
Steam Saving
Mode
Opacity
Reduction Mode
Sootblower Sequence Logic
DCS Controller
3-38
Cleanliness Factor Calculation

Actual heat absorption rate
CF =
Ideal heat absorption rate
Implies section heat transfer effectiveness.

Values range from zero to one.
Better to have flue gas temp measurement.
Inc.
3-39
Motivation for Dynamic Model Based

Cleanliness Factor Calculation
z
First principles model very difficult to formulate
Direct flue gas temperature measurements not available at all

boiler sections.
Frequent load changes.
Employing empirical model makes sense.
Inc.
3-40
Formulation of Cleanliness Factor Model

z
Qact = Fs (Ho - Hi)
Qideal = f (Fs, Tsi, Tgi)
Neural Network is used to approximate

the non-linear function to calculate Qideal
A boiler section is assumed to be clean

immediately after soot-blowing.
Inc.
3-41
Intelligent Sootblower
z
Sootblower Optimizer can operate in two modes.
Steam Savings Mode - Avoids unnecessary sootblowing,

prevents excessive desuperheater spraying, improves steam
temperature control, and reduces heat rate.
Opacity Reduction Mode - Emphasizes uniformity of blowing

sequence. Adaptively adjust desired cleanliness factor in realtime.
The Result
Balanced sootblowing strategy ---heat rate is minimized
and plant performance are maximized.
Inc.
3-42
Boiler Heat Exchangers
Reheater
Final
Superheater
Primary
Superheater
Furnace Wall
Inc.
3-43
Opacity Reduction Basic Strategy
Maximum blow idle

time violation check
Load Change
Check
Cleanliness factor
check
Sootblowing
Opacity check
Violation ?
Modify desired
Cleanliness factor
Inc.
3-44
Steam Saving Mode Basic Strategy
In addition to section cleanliness consideration:
Superheat/reheat temperature too low

--->
Blow convective section
Superheat/reheat temperature too high

---> Blow furnace wall section
Inc.
3-45
Steam Temperature Change versus Sootblowing Location
R eh e at S tea m
T e m p e ra tu re
C h a n g e (F )
20
C lo se r t o fur n a ce w a ll
S oo tb lo w in g
L o ca tio n
-20
Inc.
3-46
Sootblower Location Fuzzy Membership Function Illustration
VL
F lu e G a s P a th
Fu r n ace
w a ll
VS
P rim a ry
S up erh ea t
S
R e he ate r
(I)
MS
F ina l
S u pe rhea t
ML
Re he at er
(II)
VL
10
VS
Inc.
3-47
Exemplary Rule Base

(Reheat steam temperature control)
Reheat
Temp.
Reheat
Spray
Reheat Super
Wall
Blower Heat Blower
Idle TM. Temp. Idle TM.
Blowing
Location
VL
ANY
ANY
NOT_VS
VL
VS
VL
ANY
ANY
MS
ANY
ML
MS
ML
ML
ML
ML
MS
MS
MS
ML
VS
VS
VL
ANY
VS
VL
Inc.
3-48
Result: Waterwall Cleanliness Factors

1 .1
S e r ie
S e r ie
1351
13 01
1251
12 01
1151
1101
1051
951
1 0 01
901
851
801
751
70 1
651
601
55 1
501
451
401
351
301
25 1
201
151
0 .9
101
0 .9 5
51
C le a n l in e s s F
1 .0 5
T im e
Inc.
3-49
Result:
Division Superheater Cleanliness Factor
1 .1
1 .0 5
C le a n lin e s
0 .9 5
0 .9
0 .8 5
0 .8
1
6 3
1 2 5
1 8 7
2 4 9
3 1 1
3 7 3
4 3 5
4 9 7
5 5 9
6 2 1
6 8 3
7 4 5
8 0 7
8 6 9
9 3 1
9 9 3
1 0 5 5
1 1 1 7
1 1 7 9
1 2 4 1
1 3 0 3
t im e
Inc.
3-50
1 3 6 5
Test Result : Stack opacity
O pacity-M e gawatt Ratio

0.035
(% Opacity / MW)
0. 03
0.025
0. 02
0.015
0. 01
0.005
0
July
Augus t
Sep1~13
Sep14~26
Inc.
3-51
Sootblowing Affects Steam Temperature
88 0
73 0
Temperature ( F)
Wall blowers on
8 70
86 0
72 0
8 50
84 0
710
83 0
82 0
70 0
0
50
10 0
150
Time (minute)
Superheat Temperature
Reheat Temperature
Inc.
3-52
Sootblowing Affects Spray Flow
300
Spray flow (KPPH)
Wall blowers on
250
200
150
100
50
0
0
50
100
150
Time (Minute)
Superheat Spray
Reheat Spray
Inc.
3-53
User Interface DCS
DCS Operator Graphics
3-54
SUMMARY
z
Neural network model based boiler cleanliness factor

calculation implemented
Optimized sootblowing strategy based on Steam

Saving or Opacity Reduction objective
Software tool and user interface integrated with

existing Westinghouse DCS
Initial test shows promising result
Inc.
3-55
Intelligent Sootblowing
Application Development
Neel J. Parikh and Brad J. Radl
Pegasus Technologies, Inc.
Confidential - Pegasus Technologies
3-56
Key Criteria (1/3)

Intelligent
Learn
Identify Current Conditions
Recognize Changes
Respond Accordingly
Asynchronous
Not Just Time Based
Does Not Rely On Pre-defined Sequence
3-57
Key Criteria (2/3)

Adaptive
Adjust to
Fuel Changes
Seasonal Changes
Equipment Wear and Tear
Operating Preference and Practices
3-58
Key Criteria (3/3)

Event Driven
Examine Current Operating Conditions
Evaluate Boiler Cleanliness
Identify Desired Objective
Determine Need to Sootblow
Signal Activation of Desired Sootblower
3-59
Desired Objectives
Multiple Goals
Avoid
Unplanned Outages
Temperature Variations
Slagging, Fouling, Tube Erosion
Auxiliary Power Consumption, etc.
Improve
Emissions, Heat Rate, etc.
3-60
Pegasus Approach
Combination of AI Techniques
Boiler Cleanliness Information
Utilize N-Net Models
Plant Specific Customization
Monitoring and Backup Processing
Intelligent, Asynchronous, Adaptive,

Event Driven, Sootblower Activation
3-61
Considerations
Utilize Modular Components of the Existing
Combustion Optimizer Application
Interface with Plant DCS/Sootblowing System
Facilitate Smooth Transition
Easily Switch Operating Modes
Incremental Benefits as Project Progresses
Current R&D Project
3-62
EPRI Heat Rate

Improvement Conference
Intelligent Sootblowing
Boiler Cleaning Management System
January 31, 2001
Presented By: Randy Carter
Applied Synergistics Proprietary Information 2000

Printed with permission of the copyright holder.
3-63
Presentation Topics
Sootblower Operation Past Practices

Intelligent Sootblowing ASI Approach
Furnace Cleanliness Module

Sootblowing Cleanliness Module
Boiler Efficiency Module
System Benefits

3-64
Boiler Cleaning Management

System
Basis for Sootblower Operation Past Practices
Boiler walk down (observation)

Time of day / elapsed time
Overall measurements
Calculations
Spray flows
Steam temperatures
Gas temperatures
Recoverable losses
Operating rules
Temperature constraints
Opacity
Others

3-65

System
ASI Approach
Provide direct measurements

for determining boiler
cleanliness in furnace
Assimilate plant process
information within sootblower
control system
Use existing instrumentation

Incorporate plant experience/
operating rules
Provide flexible platform to

implement real working system
Make it easy for operators!

3-66

System Optimization System
Purpose
Provide ability to clean boiler based on need

Provide real-time information on heat transfer
surface cleanliness
Provide real-time feedback on sootblowing
effectiveness
Provide measure of system performance
Provide full integration with sootblower controls

3-67

System Optimization System
System components
Measurements/analytical
modules to determine
surface cleanliness
Furnace Cleanliness
Module (direct
measurement based)
Sootblowing Cleaning
Expert (calculation based)
Performance assessment
modules
Average furnace heat flux

Furnace surface utilization

3-68
Sootblowing
Cleaning
Expert
Furnace
Cleanliness
Module

Control of furnace heat

transfer is critical
Rate of heat transfer
degradation varies with
furnace location
Sootblowing effectiveness
varies with furnace location
Therefore furnace
sootblowing optimization
requires both spatial and
time (demand) dimension

3-69

Furnace subdivided into

zones to provide ability
to clean where required,
spatial dimension
Direct measurement
used to determine
cleaning requirement in
each zone, time/demand
dimension

3-70

Heat Flux Sensor
Measures heat transfer per

unit area perpendicular to
the tube OD, KBTU/HR-FT2
Sensor located in each zone
Provides input for
determining clean/dirty
conditions and performance
measures
Sensor is an integral part of
furnace wall tube
3-71

Heat Flux (kBTU/hr-ft2)
Heat flux alone does

not indicate cleanliness
Furnace cleanliness
module software
interrogates heat flux
to find sintering point,
based on
Frequency
Rate of change
Minimum heat flux
100
Curve 1
90
Curve 2
80
Sintering Point
70
60
50
40
30
20
10
0

3-72
3
Time

When and where

to clean
Locations which
require cleaning
are color coded.
Same signal
drives sootblower
controls,
automatic control
module

3-73

When and where to

clean
Sootblower
effectiveness
Trend plot provides

direct feedback on
cleaning
effectiveness

3-74

System
System components
surface cleanliness

(direct measurement
based)
Expert (calculation
based)
Sootblowing
Cleaning
Expert
Furnace
Cleanliness
Module
Performance assessment
modules
Average furnace heat flux

Furnace surface utilization

3-75

Sootblower Cleaning Expert
Distribution of heat
transfer is just as
important as local
cleanliness
FEGT provides boundary
condition for furnace
(outlet) and hanging Furnace Cleanliness
Module
surface/convection
Direct measurement
region (inlet)
based
FEGT measurements
Optical, SpectraTemp
or
Acoustic, BoilerWatch

3-76
Sootblowing
Cleaning Expert
Calculation based
FEGT
EGOT

Heat transfer of
convection
surfaces strongly
interrelated
Convection
region heat
transfer affects
furnace
conditions

3-77

Calculation based
Calculations with
measurement to define
boundary conditions of
each section
Calculated cleanliness
factors determine
cleaning requirements
Cleanliness factor
Cf = Umonitored/Uideal

3-78

Engineering
information
Cleanliness factor
Cf =
Umonitored/Uideal
Heat transfer section

absorption
Gas path
temperatures
Provides input for
boiler efficiency
module

3-79

Operator interface
When and where to

clean
Sootblower
effectiveness

3-80

System Assessment Modules
System components
surface cleanliness
Furnace Cleanliness
Module (direct
measurement based)
Expert (calculation
based
Sootblowing
Cleaning
Expert
Furnace
Cleanliness
Module
Performance
assessment modules
Average furnace
heat flux
Furnace surface
utilization
Boiler Efficiency
Module

3-81

0.9
Average furnace heat

flux
Provides comparison to
boiler design basis
Furnace heat transfer

surface utilization
Indicates if areas over or

under cleaned
Indicates non-uniform
heat transfer

3-82
0.8
Furnace Surface Utilization
0.7
0.6
0.5
0.4
0.3
0.2

Boiler Efficiency
Module
Dry gas loss

Loss from water in
fuel
Loss from moisture
in air
Convention/radiation
loss
Unburned carbon
loss

3-83

System Under Automatic Control
Furnace/convection heat
transfer distribution,
permissives
FEGT
RH spray
Load
Steam temperatures
Adaptive cleaning device

control
Number of operations
Wait time
Jet progression velocity
Cleanliness bypass

3-84

System - Benefits
350 MW tangentially fired unit

Powder River Basin coal
Water and system cleaning devices
ASI Boiler Cleaning Management
System

3-85
90
Heat Flux kbtu/hr-ft2
45
80
70
40
60
35
50
40
30
30
20
25
10
20
0
Baseline
Auto
Avg Heat Flux

SB Ops Per Hour

3-86
0.85
900
850
0.8
800
0.75
750
0.7
700
0.65
650
0.6
600
Baseline
Auto
Boiler Effectiveness
EGOT
Economizer Gas Outlet Temperature, F
100
Boiler Effectiveness
50
Number of Sootblower Operations per hour

System - Benefits
Summary
Boiler Cleaning Management System (BCMS) is an integration of

sootblower controls, advanced heat transfer measurement
devices, existing plant sensors, and performance software
BCMS offers heat transfer optimizing for cleaning when and
where needed, that is demand based cleaning
BCMS benefits include
Increased boiler heat transfer

Reduced gas path temperatures
Reduced sootblowing frequency
Improved sootblowing maintenance practices
More stable steam temperatures and spray flows

3-87

3-88
4
SESSION 4: TURBINES AND AUXILIARIES
4-1
In-Situ Feedwater Flow Measurement

Presentation was not available at time of publication.
4-2
In-Situ Enthalpy Measurements

in Low Pressure Condensing
Steam Turbines
Steve Hesler
Tom McCloskey
Electric Power Research Institute
2001 EPRI Heat Rate Improvement Conference; Dallas Texas
4-3
History of Technology
Various optical transmission probes developed in U.K, Japan,
Russia, Germany, France, and Czechoslovakia
earliest applications in 1970s, included droplet sizing and
wetness fraction tests to support blade design
EPRI develops aerodynamic probe technology and sponsors
optical probe field evaluation in 1992
EPRI wetness probe development initiated in 1995, as part of LP
turbine efficiency improvement program
Initial plant tests of EPRI wetness probe in 1997-8; improved
model probes fabricated in 1999 and 2000
Benchmarking in subscale test turbine successfully completed
EPRI awarded U.S. patent in 2000
Full-scale verification benchmark system test in planning phase
4-4
Purpose of Enthalpy
Measurements
Turbine operators:
measurement of UEEP enthalpy
verify LP turbine performance
upgrades
measurement of hood losses
trend LP turbine performance
Turbine OEMs
measure flow conditions prior to
design of advanced replacement
blading
detailed feedback to designers
on characteristics of proposed
stage
4-5
Parameters Measured
Aerodynamic probe
total pressure
static pressure steam tables
pitch & yaw angle
Wetness probe
specific
volume
steam tables
wetness
fraction
droplet volume
concentration
steam wetness probe

subsonic aero probe
4-6
enthalpy
Flow-Weighted Averaging
= f(pstatic, W)
pstatic
Mach no. = f( ptotal
hstatic = f (pstatic, W)
htotal = hstatic +
.
V2
2Jg
m = 2
V
i
ax
ri ri
i =1
n
Htotal =
2 htotaliVax riri
i =1
2 i Vax riri
i =1
Ptotal =
2 ptotali Vax riri

i =1
2 iVax riri
i =1
4-7
Expansion Line Analysis
T T
Hinlet
Houtlet
T
T
Hinlet
Hideal
T
T
4-8
Aerodynamic Probe
null balancing
disk type
wide pitch range
large tap diameters
wet steam
calibration facility
4-9
Optical Probe Principal

probe head
100 %
spectral transmission plot

3 averaged cycles
0%
215 nm
930 nm
UV
infrared
4-10
Mie-Scattering Analysis
100 %
3 averaged cycles
0%
215 nm
930 nm
UV
infrared
4-11
Field Data from Full-Scale Turbine

Transmission Data
Droplet Size Distribution
1.0
0.9
position #4
peak droplet dia. => 0.17 micron

Sauter dia. => 0.19 micron
Transmission Ratio (I/Io)
Relative Distribution
4
3
2
1
0
0.8
0.7
0.6
0.5
0.4
0.3
measured
calculated
0.2
0.1
0.0
0.0
0.2
0.4
0.6
0.8
0.1
0.2
Diameter (micron)
0.4
0.6
0.7
0.8
0.9
1.0
Transmission Data
1.0
0.9
position #6

Transmission Ratio (I/Io)
4
3
2
1
0
0.8
0.7
0.6
0.5
0.4
0.3
measured
calculated
0.2
0.1
0.0
0.0
0.2
0.4
0.6
0.8
0.1
0.2
0.3
Diameter (micron)
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Wavelength (micron)
Transmission Data
1.0
0.9
position #8

Transmission Ratio (I/I o)
0.5
Wavelength (micron)
0.3
4
3
2
1
0
0.8
0.7
0.6
0.5
0.4
0.3
measured
calculated
0.2
0.1
0.0
0.0
0.2
0.4
Diameter (micron)
0.6
0.8
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Wavelength (micron)
i i
4-12
Optical Probe System
4-13
Steam Droplet
Concentration Analysis
EPRI MIST program
fits droplet size
distribution to
measured light
extinction curve
integrates distribution
to yield liquid volume
4-14
Probe Access Ports

turbine exhaust ports can be installed during boiler outage
b a ll v a lve
w e ld ed f ittin g o n
e x h a u st flo w g u id e
fla n g e
g u id e tu b e
w e ld ed a d a p te r
L P b la d e s
e x h a u st flo w g u id e
m a n h o le
4-15
Probe Deployment
Traverse mechanism controls probe position
4-16
Model Turbine Benchmark Test

p robe s
W outle t , P t
outle t
, Ps
,
outle t
facility
.
inle t
,Tt
inle t
, Pt
facility
kw
inle t
four-stage
shaft torque dynamometer
condensate flow measurement
e x tra c tion
, T e xtra ct ion
facility
kW
kW
W outle t , P t
outle t
, Ps
&
m
&
m
,
o utl et
inlet
inlet
H
H
inlet
&
m
inlet
&
m
extraction
extraction
1- 2
extraction
extraction
?
probe
1- 2
&
m
&
m
exit
exit
exit
H 2 exit
generator
4-17
Future Activities
Model turbine benchmark test report
Comparative benchmark test against PTC-6
measurements
Droplet sizing experiments in two-phase flow
Extend steam wetness measurement technology
to higher pressure and temperature regimes
4-18
Steam Turbine Related Research

at TVA
Jim Terrell
Tennessee Valley Authority
4-19
Introduction
TVA - Largest wholesaler of electricity in
the United States
156 Billion kw-hrs sold in 1999
$ 6.6 Billion in 1999 revenues
Supply power to 158 distributors serving
~ 8 M people in seven states
62 direct-serve industrial and federal
customers
4-20
TVA Generation Mix

59 Coal-fired Units (61%)
5 Nuclear Units (28%)
113 Hydro Units (9%)
56 Combustion Turbines (2%)
4-21
Current Steam Turbine R&D

Low Pressure Turbine Efficiency Upgrade
EPRI Steam Wetness Probe Validation
Improved Mechanical Seal Designs
Member of EPRI Target 58: Steam
Turbines, Generators, BOP
4-22
LP Turbine Efficiency Upgrade

Developed as EPRI TC Project
New Exhaust Flow Guide Design (+0.33%)
New Last Stage Blade Designs (+1.0%)
70 MW of Additional Capacity
GE 200 MW
Westinghouse 250 MW
4-23
Validate EPRI Steam Wetness

Probe
TVA/EPRI Cooperation
Test Cell Constructed at Texas A&M
University
EPRI Providing Probe System
Testing to Begin Spring, 2001
Final Report Due Fall, 2001
4-24
Improved Mechanical Seals

GE 200 MW Steam Turbines
N-2 Packing Leakage (Leaking 5% of MSF)
NASA Seal Designs
TVA/University of Tennessee Space
Institute
Future Research: Generators, Boiler
Feedpumps
4-25
Member of EPRI Target 58

Information Clearinghouse for T-G
NDE Inspection Techniques for T-G
NDE of Stator Winding Insulation
Turbo-X Upgrades
Guidelines to Reduce Outage Time
SAFER Rotor Evaluation Code
FMAC
4-26
On-Line Performance Monitoring and Condition Assessment of Steam

Turbines
Rolf F. Orsagh, Michael J. Roemer, and Ben Atkinson
Impact Technologies, LLC
125 Tech Park Drive
Rochester, NY 14623
Bill McGinnis and Scott McQueen
Reliant Energy
Houston, TX
Abstract: An Internet-based performance monitor and condition assessment system has

been developed that allows off-site maintenance personnel to access steam turbine
diagnostic information in real-time. The Internet provides an inexpensive vehicle for the
collection and assessment of condition data from machines in distant plants that leads to
improved maintenance scheduling. The On-line health monitoring system is designed to
efficiently and reliably transfer acquired data to a remote server for processing and
analysis then publish the results on a secure Web-site on a minute by minute basis.
Sophisticated machine condition assessment techniques including sensor validation,
statistical trending, and fuzzy logic algorithms are implemented to detect incipient forms
of machine performance degradation. An Internet-based condition monitor for a
supercritical steam turbine is presented to illustrate the features, operation, and
architecture of the system.
Key Words: Condition monitoring, Internet-based, health monitoring, diagnostics,
statistical trending, real-time analysis
Introduction: Web-based performance monitoring and condition assessment of steam
turbines implemented over the Internet is an inexpensive vehicle to deliver real-time
performance and diagnostic information to key off-site personnel. Industries that could
benefit from this technology include electric power generation, petroleum suppliers,
chemical producers, or any other owner/operators of complex equipment in remote areas.
On-line machinery condition monitoring merges the latest Internet communication
technology with advanced diagnostic and prognostic algorithms to deliver sophisticated
health monitoring information in a highly accessible format. Intelligent diagnostic
algorithms continuously running on a web-server identify anomalies, and diagnose
incipient performance faults, which are then published on a secure web-site. From the
web-site, authorized personnel can easily view graphs showing performance trends, and
tables of performance parameters, anomalies, and diagnosed faults
A web-based health monitoring system for a high pressure (HP) section of a supercritical
700 MW steam turbine is used to illustrate the details of this information technology
4-27
application. The unit was chosen because of its criticality to plant operation and its
constantly varying load posed more of a condition monitoring challenge to the operators
and maintenance personnel.
System Architecture: The web-based performance monitoring system consists of three
components as shown in Figure 1. First, the turbine performance data from the plants
data acquisition system must be stored in a database on the utilitys WAN (Wide Area
Network) then transmitted using file transfer protocol (FTP) across the internet to a
remote file server. Second, intelligent algorithms on the server validate the sensor data,
detect performance anomalies, and diagnose the most likely cause of the performance
degradation. A complete diagnostic record is written to an output database on the server.
Third, active server pages (ASP) use VB Script programs and active data objects (ADO)
to update the web-site with real-time data from the output database. Authorized users can
access the information on the web-site from remote locations via a user name and
password.
Browser
Server
Web-Site
Output.db
Analysis
Algorithms
Plant
NewData.mdb
FTP
Input.db
Figure 1 Architecture of Web-Based Monitor

Monitoring and Diagnostic Algorithms: Key monitoring and diagnostic algorithms on
the web-site server consist of five components; performance parameters, data validation,
data correction, statistical trending analysis, and fuzzy logic based diagnostics. First, a
data correction routine calculates the equivalent sensor readings under standard operating
conditions. Second, data validation tests are then applied to detect abnormally low or
high readings. Third, statistical trending algorithms are used to detect incipient
performance faults by identifying shifts in the mean values of recent corrected and
calculated parameters. Finally, in the event of a statically significant and severe deviation
from the baseline values, fuzzy logic algorithms diagnose the most likely cause of the
performance degradation.
4-28
Performance Parameters
An optimal set of performance parameters must be selected during the design of an online monitoring system. The appropriate set of parameters includes a sufficient level of
redundancy to support sensor validation algorithms and ensure reliable results without
becoming overly complex.
For the steam turbine demonstration, a set of performance parameters were chosen to
detect the onset of the most common causes of performance degradation in HP turbines;
solid particle erosion, leakage, deposits, and blockage. Diagnosis of these faults is based
on techniques developed by Cottoni and Beebeii and relies on five commonly used
parameters (1st stage pressure, cold reheat pressure, mass flow, 1st stage efficiency, and
HP efficiency) for HP turbine diagnostics. The web-based monitoring system also
displays other parameters, as shown in Figure 2, which are not used by the diagnostic
system but may provide additional insights into the operational status of the unit.
Figure 2 Monitored Parameters

Data Validation
To provide reliable information, a health monitoring system must begin with a rigorous
examination of the measured parameters to identify sensor malfunctions before they
contaminate the diagnostic information. Sensor problems such as ground loop faults or
sensor drift are often misinterpreted as the onset of performance or vibration faults. In the
4-29
event of a sensor failure, sensor recovery is possible through the use of artificial
intelligence algorithms that can provide proxy data until the malfunctioning sensor can be
repaired.
The sensor diagnostic process should be performed using multiple and collaborative
techniques that offer advantages for isolating and detecting specific sensor failure modes.
Some available techniques that have been implemented with success include; trained
neural networks, and fuzzy logic analysis. The neural network operates by comparing the
physical relationships between signals as determined from either a baseline model or
equivalent computer model of the machines performance parameters. The fuzzy logic
based sensor analysis continuously assesses the normal bands associated with each
sensor signal at the current operating condition. When a signal goes outside these bands,
while others remain within, an anomaly is detected associated with those specific sensors.
These parallel algorithms are combined in a data fusion process that determines the final
confidence levels that a particular sensor has either failed or has suspect operation.
In the steam turbine monitoring example, data validation algorithms test each diagnostic
parameter to identify gross deviations from the expected operating range. Upper and
lower bounds for the each parameter were established from normative (baseline) data
supplied by the utility. The sensor validation status is shown for each monitored
parameter on the web page shown in Figure 2. Interruptions of the data transfer from the
plant are also detected to avoid corruption of the statistical database.
Data Correction
Variable speed or variable load machinery presents a significant health monitoring
challenge due to the difficulty involved in trending performance parameters. To
accurately trend the performance of variable speed or variable load equipment while it is
in use, the performance parameters must be corrected to their equivalent values at a
standard speed or load condition. Corrections are based on a-priori knowledge of the
relationship between the independent and dependent performance parameters in a
baseline model.
Corrections are performed using polynomials that represent baseline mean values of
performance parameters over the operational range while the equipment is in a healthy
condition. Correction curves were developed for the steam turbine monitor by fitting
polynomials to the normative data supplied by the utility. Care must be used when
selecting normative data to ensure that it is unbiased by periodic or seasonal effects. The
variation of each diagnostic parameter from its baseline (expected) value is calculated in
real-time, and translated to an equivalent value at the design conditions. Figure 3 shows
examples of uncorrected and corrected data. The discontinuity in the slope of the
uncorrected data (at mass flow = 1700 units) is due to the fact that the unit operates in
two distinct modes.
4-30
Figure 3 Uncorrected and Corrected Data

The steam turbine on-line monitor required corrections to the five diagnostic parameters
and load. Each record of data that is transferred from the plant may be collected under a
different load condition. Before the data is analyzed, it must first be corrected to
eliminate the effects of the varying operating conditions. First stage pressure, cold reheat
pressure, and the efficiencies are all corrected to their equivalent values at the 5% over
pressure equivalent design mass flow from the turbine manufacturers heat balance
diagram for the unit. The load on the turbine is calculated by correcting the electric power
generated for the non-linear electrical and mechanical power loss in the generator.
Finally, mass flow is corrected to its equivalent value at the 5% over pressure condition
using the corrected load. Performance trends may then be identified and evaluated under
any operating conditions.
Statistical Trending Analysis
Performance anomaly detection algorithms are designed to statistically detect the manner
in which machinery performance parameters are shifting over time. The mean values of a
recent sample of performance parameters are compared against the mean values in the
baseline model. Significant deviations of the corrected performance parameters from
their baseline values indicate the presence on an anomaly and possibly the onset of
performance degradation.
A statistical T-test is used to detect subtle shifts in the corrected performance parameters.
Mean and standard deviation values are calculated for the corrected performance
parameters that comprise the baseline model. Small deviations during operation may be
due to normal random fluctuations in the selected sample or due to real performance
changes. The confidence value reported by the T-test represents the probability that the
observed shift is not attributable to normal random fluctuations as shown in Figure 4.
4-31
Baseline mean 1
Current mean 2
t =
| 1 2 |
s1
s
2
n1
n2
Figure 4 T-test is Used to Detect Incipient Faults

An anomaly was detected shortly after the implementation of the steam turbine on-line
monitor. Figure 5 shows the corrected data and corresponding distributions for the
baseline and anomaly conditions. The T-test indicated a nearly 100% confidence that a
13% shift had occurred in the parameter shown. However, sensor validation alarms
indicated that the problem was not due to performance degradation, but to corrupt data.
Consultations with the utility revealed a data processing error in which data from a
different turbine was transmitted to the on-line monitor.
Figure 5 Anomaly Detection

Diagnostic Classification
The total combination of performance parameter trends is used to provide a diagnosis of the most
probable performance fault. Once an anomaly has been detected, the performance error patterns,
i.e., parametric trendline deviations from baseline performance are fed to a diagnostic module.
Artificial intelligence-based pattern recognition algorithms within the diagnostic module
compare the observed error pattern with those of known faults to determine the most likely cause
of the anomaly.
4-32
The error patterns associated with known faults must be determined through modeling or
analysis of the failure modes. For the web-based steam turbine monitor, computer simulations of
a similar unit were run using a through-flow, HP-section, performance model. The model has
the ability to simulate SPE or Deposit damage in the HP control stage, seal leakage, or flow
blockage anywhere in the unit. This is performed through altering a combination of chord length
and surface roughness for various levels of SPE or Deposit damage and their respective
efficiency effects, and gaging on the HP stator and rotating rows (Figure 6). Figure 7 shows a
simulated normalized performance error pattern for the case of SPE (top left) versus ideal fault
error patterns.
Figure 6 Modeling Degradation of Turbine Blades
Figure 7 SPE Diagnostic Results

Within the diagnostic module, fuzzy logic is used to examine error patterns and determine the
most likely cause of anomalous conditions. Fuzzy logic is an artificial intelligence technique that
has the ability to recognize error patterns that are not necessarily identical to that of a known
fault. It is deemed an intelligent tool because it can operate in gray areas just as a human
evaluator would, rather than in black or white. For instance, in the case of solid particle erosion
(SPE), the 1st stage pressure is a little high (Figure 8). If, in addition, we find the CRH pressure
4-33
is ok, the mass flow is high, and the 1st stage efficiency is a little low we can say with confidence
that the unit is probably experiencing the onset of (SPE) in the HP control stage based on its
similarity to the known error pattern for SPE.
Figure 8 Fuzzy Logic Classification

Results form the anomaly detection and diagnostic algorithms are immediately presented on the
web site along with supporting information. Users can dig down to learn more about the
condition of their equipment by displaying logs of past diagnostic results, and trend plots of key
raw and corrected parameters.
Conclusion: This information rich condition monitoring and assessment system offers access to
current plant data, plots of performance trends, and automated fault diagnosis that are useful to
both technical and non-technical plant personnel. Access to current performance and machinery
health information allows risk managers, and plant operation planners to improve maintenance
scheduling and reduce the machinery life cycle costs. Posting the information on a web-site
makes it readily available to all authorized personnel.
Acknowledgements: The on-line monitoring demonstration was developed under a grant from
the Electric Power Research Institute (EPRI). The authors wish to specifically acknowledge
Peter Millet, and Bill Reuland of EPRI for their support and technical assistance.
References:
i
ii
4-34
Cotton, K. C., Evaluating and Improving Steam Turbine Performance, Cotton Fact Inc., Rexford, NY, 1993
Beebe, R, Machine Condition Monitoring, Engineering Publications, Victoria, Australia, 1988.
5
SESSION 5: HEAT RATE TESTING
5-1
PERFORMANCE EVALUATION AND TESTING

SAM J. KORELLIS, P.E.,
Performance Group Leader
Dynegy Midwest Generation
Vice-Chair ASME Board on Performance Test Codes
PHILIP GERHART, Ph.D., P.E.
Dean of Engineering, University of Evansville
ASME VP-Performance Test Codes
NEW EQUIPMENT? MAJOR MODIFICATION?
ASME PERFORMANCE TEST CODES
HELP EVALUATE PERFORMANCE
When faced with the question of How well is this equipment
performing? ASME Performance Test Codes (PTCs) can help
guide us to an answer. ASME PTCs provide uniform rules
and procedures for conducting and reporting tests of
equipment and systems to determine if performance criteria
have been met.
What are ASME PTCs?
ASME Performance Test Codes (PTCs) provide uniform
rules and procedures for planning, preparing, and
executing performance tests and reporting the results. A
performance test is an engineering evaluation based on
measurements and calculations, whose results indicate
how well the functions of equipment are accomplished.
For over a century, the philosophy of ASME has dictated
that a PTC test can provide results with a high level of
accuracy, based on current engineering knowledge and
practices; taking into account the costs of the test and the
value of the information obtained.
Codes are, or are intended to be, legal documents. ASME
PTCs are written in a format suitable to be cited in
contracts as the standard methodology to determine if the
equipments guaranteed performance was attained.
Who uses ASME PTCs?
In simple terms, PTCs can be used by equipment owners,
equipment suppliers, and engineering personnel who
conduct and analyze tests.
When acquiring new equipment, purchasing agents or
sales staff can specify that the equipment guarantee will
be based on the results of a test conducted in accordance
with a specific ASME PTC. Next, the design engineers
consult the PTC document to ensure the proper instrument
connections will be available. Later, the test engineers
install the required instrumentation, and utilize the test
procedures and calculation methods to conduct a code test
Throughout the process, the focus is on building
5-2
on the new equipment. Representatives of the now

numerous parties to the test ensure the test methods are in
compliance with the Code. The test results are analyzed
by both the supplier and the equipment owner, and
compared to the performance criteria.
Also, manufacturers and suppliers may be interested in the
exact performance of their equipment, to better
understand design margins and impacts of manufacturing
tolerances on performance. Therefore, Code tests are
conducted outside of any performance guarantees.
In the commercial arena, ASME PTCs permit suppliers to
compete fairly while protecting users from complacency
and poorly-performing products. Purchase specifications
are greatly strengthened by citing the results of tests
dictated by PTCs.
Who issues ASME PTCs?
Performance Test Codes fall under the jurisdiction of the
ASME Council on Codes & Standards. Anyone
identifying the need for a test code or a test code revision
communicates that need to ASME staff. Based on the
substantiated need, the Board on Performance Test Codes
organizes and empowers a technical standards committee
to develop or revise a test code.
The technical standards committee is a team of experts;
highly qualified and technically competent engineers, with
expertise in some or all fields covered by the specific
PTC.
The committee membership is balanced between
equipment users, equipment manufacturers, and general
interest personnel. While their employment and, in most
cases, their support is based in one of the three
membership categories, standards committee members
represent no one but themselves. This group prepares the
code draft, while meeting periodically, and corresponding
continually. It should be noted that all meetings are open
to anyone interested in attending and participating.
A preliminary draft is sent to knowledgeable persons in
the industry for review. This review and the incorporation
of comments and recommendations strengthen the
document. The completed draft is then approved by the
technical committee membership and the Board on
Performance Test Codes. The new Code is sent to ANSI
for their approval, and ASME publishes the document and
sells it for barely more than the cost of the material,
printing, paper, and handling!
consensus. All ASME Codes have been developed by
seeking balanced input from all types of parties who may

be interested in the Code and/or the equipment or process
the Code deals with.
ASME Vice-President, elected by the entire Society. One

of the five Councils of ASME, the Council on Codes and
Standards governs all ASME Codes and Standards
activities.
Who, besides ASME, issues Codes and Standards?

Several other professional organizations develop codes
and standards by the consensus process; among them
IEEE and ASTM.
Codes developed by these
organizations often compliment ASME Codes, and ASME
PTCs often refer to IEEE or ASTM Codes for test
elements that fall within the areas covered by the
IEEE/ASTM.
Examples include electric power
measurements for motor drives, and the determination of
properties of materials such as coal or limestone.
Codes and standards are sometimes issued by trade
organizations. These documents are usually developed by
a team of manufacturers, and often do not include input
from equipment users or general interest parties such as
engineering consultants or representatives of government.
The federal government and, to a lesser extent, state
governments sometimes develop and issue codes and
standards. At the current time, the U.S. Federal
Government is attempting to move away from
governmentally-developed standards, and towards
utilizing voluntary, consensus-based standards.
Other codes and standards are also issued by groups
sponsored by some countries outside the U.S. These may
represent a single country, such as the German DIN
standards; or a multi-national effort, such as the ISO
standards, which include U.S. participation.
Who is ASME?
ASME, International (the American Society of
Mechanical Engineers) is a not-for-profit organization
dedicated to promoting mechanical engineering. ASMEs
mission is:
To promote and enhance the technical competency and
professional well-being of our members, and through
quality programs and activities in mechanical
engineering, better enable its practitioners to contribute
to the well-being of humankind...
ASME consists of more than 125,000 members, and has a
staff of 400. Performance Test Code committees are
staffed by volunteers and coordinated by an ASME staff
engineer. Committee membership is not limited to ASME
members. PTC committees report to the Board on
Performance Test Codes. The Chair of the Board is an
What equipment is covered by PTCs?

Performance Test Codes originated as Power Test Codes
decades ago; so the main emphasis has been on power
plants and their associated equipment. The first Code (of
any type) issued by ASME was Rules for Conducting
Boiler Tests, published in 1884. Today, nearly 50 PTCs
are available, covering individual components, entire
systems, and complete plants. In addition to these Codes,
Supplements on Instruments and Apparatus are available,
covering measurements and techniques common to
several Codes.
Equipment covered by PTCs encompass:
PTC-4
Fired Steam Generators
PTC-6
Steam Turbines
PTC-10
Compressors
PTC-11
Fans
PTC-12.2
Steam Surface Condensers
PTC-18
Hydraulic Turbines
PTC-23
Cooling Towers
PTC-25
Pressure Relief Devices
PTC-39
Steam Traps
PTCs applying to plants and systems include:
PTC-46
Overall Plant Performance
PTC-PM
Performance Monitoring Guidelines
Some Supplements on Instrumentation & Apparatus are:
PTC-19.1
Test Uncertainty
PTC-19.2
Pressure Measurement
PTC-19.3
Temperature Measurement
PTC-19.5
Flow Measurement
PTC-19.11
Steam and Water Purity
What is included in an ASME PTC?

Performance Test Codes have a standard format. Each
begins with an Object and Scope, describing the
equipment to be evaluated, the goals of the test, and the
expected uncertainty of the results. Definitions and
Description of Terms, which also includes mathematical
symbols and abbreviations, follows the Object and Scope
section.
Guiding Principles are covered in detail, including:
- items to be agreed upon before the test
- preliminary tests
- test preparations
- operating conditions
5-3
equipment set-up for the test

permissible deviations
constancy of operations
acceptability of test runs
frequency of observations and duration of test runs
The section on Instruments and Methods of Measurement

includes guidance on the choice of instruments, including:
sensitivity, precision, number, alternatives, and
duplication. Additional information is provided on the
use and placement of instrumentation, its limitations,
sources of error, and precautions. Test methods are also
described.
Another section outlines the Computation of Results. This
section contains the formulas and directions for
determining the equipment performance from the recorded
data. It also contains the formulas and directions to
determine the uncertainty of the results.
Each PTC contains a list of information that should be
included in the Report of Results section. PTCs may
provide the format or outline for the report.
Many PTCs have appendices. Topics covered in the
appendices may include:
- sample calculations
- derivation of formulas
- alternate test methods
With these state-of-the-art methods, why worry about
test uncertainty?
Uncertainty is associated with all measurements. PTCs
provide guidance on determining test uncertainty and the
typical or expected values for specific tests. Some PTCs
require that a certain level of test uncertainty be achieved
in order for the test to be a valid Code Test.
It is important to note that test uncertainty is not a
tolerance; but rather, provides insight to the quality of the
test and its results. Delving into the individual
contributions to overall test uncertainty, the test engineer
can determine where best to spend limited funds to
improve the quality of the test results. If alternate
methods are provided in the code or are used in the test,
the impact on uncertainty to the final result is determined.
Why include PTCs in purchase contracts?
Acceptance tests have a significant cost; however, the cost
is much smaller than the capital cost of the equipment, or
the net worth of the proper operation and performance of
the equipment.
5-4
From all parties point of view, a PTC test firms up the

expected and guaranteed performance. Everyone now
understands what is expected and how it is measured.
The value added by conducting the tests includes:
- the purchaser and manufacturer can agree on the
current level of equipment performance
- adjustments can be made based on the results of the
test, including partial refund of purchase price,
restoration of inadequate performance, or bonus/
penalty payments.
- the manufacturer gains baseline knowledge of the
equipments design, often learning the impact of
manufacturing tolerances, impacts of off-design
operation, or evaluation of the latest improvements.
- the equipment owner now has baseline information
against which to compare future performance and
predict future capability (including projecting
product cost).
What do PTC tests cost, and what are their worth?
The PTC documents can be purchased directly from
ASME, at prices ranging from $20 to $250, each. The
exact price depends on the specific test code.
Figure 1 lists typical test costs for selected components
and respective test codes. The table also contains
component capital and operating costs, and the cost of
potential performance losses for comparing to the value of
a PTC test.
Test costs also depend upon the complexity of the system
being tested. Typically, the tests can cost a few thousand
dollars for a small component, up to a quarter of a million
dollars for a complex plant. These costs are put in
perspective by understanding that it doesnt require many
$1,000 on-peak MWhours to recoup the cost of
identifying and correcting an efficiency loss.
For fuel-consuming equipment like gas turbines and
boilers, the cost of a PTC test is less than one (1) percent
of one year of operating, fuel, and investment costs. For
components not directly consuming fuel, like cooling
towers and steam turbines, the test costs are typically well
under one year of operating and investment costs for that
component.
FIGURE 1
PTC COSTS
Component
PTC
Typical
Test
Cost
($000)
Boiler (Fired Steam Generator)

Condenser
Cooling Tower
Feedwater Heater
Gas Turbine
Heat Recovery Steam Generator
Steam Turbine
Total Plant
4
12.2
23
12.1
22
4.4
6
46
100
35
20
17
50
30
250
225
Component
Capital
Cost
($000)
Annual
O&M
Costs
($000)
200,000
25,000
12,000
200
60,000
20,000
150,000
800,000
30,000
50
60
20
13,000
14,000
400
38,000
Fuel
Costs
Included
Included
Included
Included
Test Cost
as a % of
Annualized
Costs
Potential
Annual
Performance
Loss ($000)
Payback
of Test
Costs
(months)
0.3%
4.0%
4.3%
63.8%
0.3%
0.2%
4.6%
0.3%
540
60
60
8
50
260
270
810
2
7
4
26
12
1
11
3
*all costs in U.S. dollars
The cost savings arise when the test identifies a

correctable deficiency. For example, a test on a new
steam turbine can cost $250,000, and require 7-10 days to
complete. The purpose of the test is to verify the
performance of a $150 million machine, where a one
percent efficiency loss can be worth over $250,000 each
year.
Why does it take so long to publish Codes?
New test codes are typically issued about five years after
the formation of a standards technical committee. While
coordinated by ASME staff, the documents are developed
entirely by volunteers; therefore, work does not progress
continuously. The consensus process used by these
diverse teams of experts, which results in a fair and
technically excellent product, requires time and effort.
Recently, ASME recognized that five years is too long,
and has redesigned the process of developing all products
in the entire Codes and Standards Division. Time savings
will result from the increased use of new technologies,
including electronically routing draft documents; and
streamlining the administrative process; e.g., conducting
parallel reviews.
Where can I get more information on ASME PTCs?
Contact ASME:
- their web page at http://www.asme.org
- Director of PTCs, Mr. W. O. Hays at (212) 705-8550
- Mechanical Engineering Magazine
- attend (and/or participate in) technical committee
meetings. The times, dates, and locations are
available from the sources noted above.
What can you do?

Whenever you or your staff procure a major piece of
equipment, ensure the warranty calls for the acceptance
criteria to be based on the results of a PTC test. Support
your engineering personnel participating on committees,
reviewing drafts, and conducting official tests. Volunteer
to serve on a technical standards committee. Provide
feedback to ASME, the committees, and Board on the
usefulness of PTCs or other related services you desire.
What is the future of PTCs?

Performance Test Codes have been extensively used for
over one hundred years. While the test methods have
changed greatly with the advent of modern instruments
and portable computers, there is still a need for standards
and directions for conducting these near people-less tests.
As the power industry accelerates through deregulation,
additional emphasis will be placed on cost reductions;
therefore, knowledge of the component level of
performance will increase in value. Equipment guarantees
will become more important; the purchase specifications
will be tightened, extending the intensity of competition to
the equipment suppliers.
As previously mentioned, ASME Codes and Standards
have redesigned the processes used to develop and deliver
their products. This change resulted from the observed
increases in everything from number of Codes and
Standards, to their increased acceptance and use; pace and
volume of data interchange, and increased expectations.
With leaner corporate structures and fewer available
volunteers having less available time, ASME is committed
to:
- producing their C&S products on a time scale that
5-5
meets users needs.

providing timely responses to inquiries.
maintaining due process and consensus documents
with less impact on volunteers.
REFERENCES
1.
2.
New PTCs have been developed to accommodate the new and

complex needs of the industry. Several PTCs that consider
equipment that is not specifically focused on electric power
production are under development. PTCs for environmental
protection equipment and processes are being written.
Additional products and services are being evaluated to ensure
that ASME PTCs best serve global industries as the
preeminent provider of state-of-the-art standardized methods
for performance testing, monitoring, and analysis of energy
conversion and industrial processes, systems, and equipment.
5-6
3.
ANSI / ASME PTC 1, General Instructions, American

Society of Mechanical Engineers, New York, 1999.
Yost, J. G., The ABCs of Acceptance Testing, PWRVol. 30, Proceedings of the International Joint Power
Generation Conference, 1996.
Korellis, S. J., ASME Performance Test Codes, Electric
Power Research Institute Balance of Plant Performance
Group Meeting, 1992.
5-7
5-8
5-9
5-10
5-11
5-12
5-13
CYCLE ALIGNMENT METHODS AND EVALUATION

and plants today are much more in tune to optimizing
their heat rates, all should undertake a regular program to
ensure proper cycle alignment.
SAM J. KORELLIS, P.E.,

Performance Group Leader
Dynegy Midwest Generation
Abstract
The action of restoring cycle alignment has historically
yielded large heat rate improvements. The benefits, in
addition to the fuel savings realized by minimizing cycle
leakages, include: extended hardware life, improved
component operation, reduced condenser backpressure,
more reliable flow measurements, and enhanced
personnel safety.
Methods of improving cycle alignment are explained,
evaluated, and accompanied by economic analyses.
Potential heat rate effects of cycle isolation losses are
listed. Actual field results and examples of cost savings
from improved cycle alignment are provided.
Basis
Historically, restoring cycle alignment has provided the
largest heat rate improvements as compared to any other
action. It is usually the least costly to implement, making
it desirable in terms of financial return. Since generators
In addition to the heat rate improvement realized by

routing the heated fluids on their intended ways, several
other benefits exist. Those benefits include hardware life
extension, improved component operation, reduced
condenser backpressure, improved measurements, and
reduced personnel safety hazards.
Just as the turbine designer/manufacturer takes great care
to maximize the amount of steam flowing through the
blades and buckets in the turbine, the architect engineer
and plant operators try to ensure that the heated fluids in
the steam cycle are routed to where they provide the
greatest benefit. Todays steam turbines have variable
clearance interstage packings that open during startup,
causing a brief efficiency loss, but protect the packings to
provide improved performance during full load operation.
That is similar to the use of emergency drains on FW
heaters. They open to protect the heat exchanger, routing
heated water to the condenser, temporarily causing a large
heat rate loss. But just as damaged turbine interstage
packings result in poorer turbine efficiency, a leaking
emergency drain line causes poorer unit efficiency.
EffectsCC
To put the potential losses into perspective, the following table is based on a leakage equal to one tenth of a percent of
throttle steam flow.
Loss originates at
Heat Rate Effect
Main Steam
Cold Reheat
Hot Reheat
Highest Press FW Heater drain
FW Pump
Steam Drum
0.12 %
0.07 %
0.08 %
0.03 %
0.02 %
0.04 %
The turbine manufacturers have developed and published similar tables of

generic values for use in evaluating test results. Unit specific heat rate effects
can be determined via steam cycle evaluation using software and models,
testing, or rigorous hand calculations.
Cycle Alignment
5-14
Effects
To put the potential losses into perspective, the following
table is based on a leakage equal to one tenth of a percent
of throttle steam flow.
Loss originates at
Main Steam
Cold Reheat
Hot Reheat
Highest Press FW Heater drain
FW Pump
Steam Drum
Heat Rate Effect

0.12 %
0.07 %
0.08 %
0.03 %
0.02 %
0.04 %
The turbine manufacturers have developed and published

similar tables of generic values for use in evaluating test
results. Unit specific heat rate effects can be determined
via steam cycle evaluation using software and models,
testing, or rigorous hand calculations.
The information from this table can be used to determine
the cost of a leak. For a nominal 500 MWe coal fired
unit, each percent heat rate loss is worth over $300,000 in
annual fuel costs. So using the table, a 30,000 lb/hr leak
through a HP heater drain can cause a $90,000 annual
loss. This large leak is nearly a full percent of throttle
flow and 10% of the expected extraction steam flow, but
would not be readily noticed during plant operation.
valve and the sparger cost over $10,000, but by now

operating this system without the leak, the cost of sparger
repairs will not be repeated and a heat rate improvement
will be realized. Had the leak continued, condenser tube
damage would have occurred shortly causing a forced
outage. The repair was justified based on a heat rate
improvement alone, but that cost is small in comparison
to those of a forced outage.
There are additional positive effects of maintaining cycle
alignment. Other motivations include proper equipment
operation, reliable fluid measurement, and potential
personnel safety hazards. The valves on the vents on
feedwater heaters need to be open to ensure proper
venting of non-condensables to maintain optimal heat
transfer. If the orifice is either clogged or enlarged by
erosion, the incorrect flow rate will either reduce heat
transfer in the feedwater heater or waste heat by sending
excess steam to the condenser. Equipment bypass lines
should be completely closed when the equipment is in
service. Routing a portion of a units feedwater flow past
a feedwater heater instead of through it results in an
efficiency loss. Ensuring the entire flow passes through a
flow meter, with minimal pressure drop contribution from
isolation valves, is necessary to properly measure the flow
rate. Lines open to the atmosphere release the energy in
the fluid, lose the treated fluid itself, and may be a
personnel hazard.
Methods
Using the typical annual costs of the least expensive

evaluation methods and anticipated repairs, restoring
cycle alignment can pay for itself by finding and
correcting one leak of this magnitude every six years.
Conversely, repairing this one leak will pay for itself in a
few months.
The magnitude of plants heat rate improvement is much
greater than the tenths of percents listed in the table that
reflect a single leak. For example during initial plant
startups, heat rate losses caused by cycle alignment
problems have been in the 5-10% range. Typically, when
preparing for a turbine test or commencing on a cycle
alignment restoration program, the gains have been 0.52.0% improvement in heat rate.
A big portion of the savings comes from heat rate
improvement since fuel costs consume 60-75% of the
annual plant operating budgets. But, these leakages can
also affect maintenance costs. Identifying and correcting
a leak before it ruins the sealing surface of a valve seat or
erodes the downstream pipe or cuts a few tubes in the
condenser will save maintenance dollars. For an example,
a leaking recirc min-flow line on a feedwater pump was
identified during a condenser inspection. The associated
flow sparger that distributes the high-energy flow in the
condenser was severely eroded. Repairing that leaking
Most of the methods to evaluate cycle alignment are labor

intensive, requiring personnel to list hundreds of flow
paths and walk down plant systems to make simple
observations. The sensing devices span the range from no
tech to high tech. As with any rigorous action in a power
plant, a plan should be developed in advance.
The first step is to develop a list of valves / lines / traps to
be checked. This list can be developed from plant
drawings and verified by walking down the system. Since
the list can be quite large, it is recommended to prioritize
the potential flow paths by their potential heat rate effect,
the system, or by its location in the plant. Prioritizing it
by potential heat rate effect can be done simply by
ranking based on fluid enthalpy or more accurately based
on the results of steam cycle analysis or testing.
Once the priority is set, a unit specific frequency for
surveys can be established. It is recommended that the
entire system of flow paths be surveyed after each major
overhaul and no less frequently than once each year.
Those valves with either a high effect on heat rate or a
history of leakage should be checked more frequently.
Several generators reported checking several high impact
valves as often as once a week.
5-15
There are many methods to detect proper or improper

alignment. Methods include temperature measurements,
sonic measurements, and visual detection.
One of the simplest detection methods is measuring
temperature upstream and downstream of a valve. Since
most of the lines of interest are connected to the
condenser an elevated downstream temperature is strong
indication of a leak. Boring a small (1/4) hole in the
bottom (at 6:00) of the insulation surrounding a pipe
permits a surface temperature measurement to be made
with a small portable probe. For those pipes more
remotely located, infrared temperature measurements
(thermography) are recommended. Some boiling water
reactor plants have installed surface mounted
thermocouples on the lines with the highest potential of
leakage or heat rate loss and routed the lead wires outside
the shielding to permit periodic monitoring without
additional radiation dose.
Other methods include: observing, recording, and
trending the valve stem position, or using aerosols to
detect motion of the air near the pipe caused by either
heat convection, flow exhausted from an open line, or air
drawn into the condenser vacuum.
ASME equipment Performance Test Codes contain
additional recommendations for verifying proper cycle
alignment. Though many of these tests require strict
cycle isolation, which cannot and should not be
maintained during normal plant operation, they provide
excellent guidelines for the most efficient operation.
PTC-6 on steam turbine testing requires that the
unaccounted for leakage from a turbine steam cycle be
less than 0.1% of throttle flow. Using the values in the
table above, based on this limit the uncertainty on heat
rate caused by not knowing the source of the external
leaks can approach 10 btu/kwh.
Several empirical relations have been published to
estimate the leakage flow rate based on the temperature
differential, acoustical emissions, and associated plant
conditions. Turbine vendors, test consultants, and test
equipment suppliers have numerous methods to estimate
leakage flow rates. The primary reason to estimate the
flow rate is to determine the effect on heat rate and
operating costs. Computer models and steam cycle
analyses and heat rate testing also provide valid estimates
the effects, but are labor intensive, time consuming
methods.
Several systems have been developed to indicate a leak or
estimate a flow rate based on acoustic signals. The
acoustic monitor is placed on the exterior of the valve
body. Its output is compared to past readings, when there
was reasonable assurance the valve was not leaking, and
provides an estimated flow rate based on vendor propriety
correlations.
5-16
A portable ultrasonic flow meter can be strapped onto the

outside of an uninsulated pipe to measure flow. This is
useful where a measurable temperature differential is not
anticipated, like in the case of an open feedwater heater
bypass. It shouldnt be used downstream of a potentially
leaking isolation valve, since without a full pipe of water
the reading is most often erroneous or difficult to obtain.
On the lower tech end, a mechanics stethoscope can be
used to listen to steam traps and valves, though the
analysis is quite subjective.
Some lines are designed with a vent or drain line between
the valve and the condenser. Water will flow out of that
valve only if certain conditions prevail, otherwise opening
this valve might cause a huge increase in condenser air inleakage and not indicate leakage when some exists. If
this vent or drain is located between two valves upstream
of the condenser, then it can be used to determine if the
most upstream valve is secure. Use extreme caution in all
cases when opening a pressurized system to the
atmosphere.
Some leakages are non-continuous and occur on a cyclic
frequency. Automatic drain valves might cycle based on
incorrect high liquid level signals or controller problems.
Unless the monitoring frequency just happens to match its
cycle of opening, the leak indications will be more
difficult to measure. Noticing downstream temperature
variations in time is a signal that the pipe is cooling after a
periodic loss of high temperature fluid. Longer term
monitoring using an automated data acquisition system is
the best method to identify non-continuous leakages.
Not all of the leakages identified are caused by
mispositioned valves. Many are the result of control
system problems, mechanical problems with the valve, or
even misapplication of valve design. To improve cycle
alignment some emergency drain valves, used to control
feedwater heater level, have been replaced with globe
valves to better withstand throttling service without
damaging the seating surface. Again to improve cycle
alignment, gate valves, typically used for steam and
startup drain valves, have been replaced with quarter turn
ball valves.
The control systems of many valves
experiencing intermittent service have been modified to
improve sealing and permit less leakage. Maintaining
proper air pressure and correct controller gain settings and
to ensure the valve is not just cracked open help preserve
its seat and prevent long term leakages.
Returning to your specific actions, the list has been
developed and the potential loss flow paths have been
surveyed in the plant. So now you are hot and sweaty and
have a list of leaks causing potential losses. They are
reported to maintenance and you follow their restoration.
On the leaking lines containing a manual isolation valve,
a short test can be conducted by temporarily closing the
manual valve to determine how much heat rate
improvement is realized when the flow is stopped. The

process doesnt end here. To maintain improved plant
performance for the long term, this is a continuously
ongoing process for the life of the plant.
made on cycle alignment. The return on investment

improves on future surveys without the first time
preparation costs.
It is recommended to conduct this in-plant survey on a

periodic basis. Checking the temperatures or acoustics of
a valve immediately following repairs provides excellent
feedback to the maintenance crew and plant leadership in
addition to developing records for future comparisons.
References
Examples of flow paths that should not be flowing:
2. ANSI / ASME PTC 39.1, Steam Traps, American

Society of Mechanical Engineers, New York.
-Emergency drains to the condenser

-Pump recirculation lines
-High point vents
-Low point drains
-Equipment pressure relief valves
-Steam startup drains
-Steam traps
-Equipment bypasses
-Turbine hood sprays
-Condenser water box priming vents
-Any line routing heated fluid out of the steam cycle
Example and Evaluation
This is an example of a process to improve cycle isolation
on a modern 500 MW fossil station. The hours and costs
are typical of actual experience.
Review drawings and verify valve locations in the plant to
develop a prioritized list containing 120 high priority flow
paths and 250 lower priority flow paths.
3-5 days
Purchase a hand held pyrometer.
$300
Identify and prepare locations.
2 days
Conduct the survey.
1 day
Issue maintenance work requests.
2-4 hours
Subtotal of first time survey cost
<$4000
Repair leaking valves
$10,000
Recheck those repaired
2-4 hours
Subtotal resolution cost
<$10,500
Total cost
<$15,000
Repeat survey cost
<$600
1. ANSI / ASME PTC 6-1996, Performance Test Code

on Steam Turbines, American Society of Mechanical
Engineers, New York, 1996.
3. Heat Rate Improvement Reference Manual, TR109546, EPRI, July 1998.

4. Performance Monitoring Program Manual, Duke Power
Company, June 1985.
5. Condenser In-Leakage Guideline, TR-112819, EPRI,
January 2000.
6. Branco, M. M., Balasubramanian, K., Dimmick, J. G.,
Fitzgerald, W. V., Heat Rate Recovery by Cycle
Isolation and Control, EPRI BOPPMG, 1992.
7. Albert, P. G., Booth, J. A., Shafer, H. S., Effects of
Turbine-Generator Associated Equipment on Sustained
Thermal Performance, ASME paper 83-JPGC-PTC-4,
1983.
8. Moradian, M. A., Sandhu, S. S., Southall, L. R.,
Application of Diagnostic Performance Programs in
Power Plants, ASME paper 87-JPGC-PTC-5, 1987.
Identifying and repairing one major leak during an initial

survey provides a two-month return on the investment
5-17
In Search of Unaccounted for Btus via the Art

of ASME PTC-6 Testing
Italo Liberatore
Senior Engineer
Predictive Maintenance Engineering
Constellation Power Source Generation
Allison Rossi
Plant Performance Engineer
Brandon Shores Power Plant
Donald Fyhr
Performance Engineer
Predictive Maintenance Engineering
5-18
In Search of Unaccounted for Btus via the Art of ASME

PTC-6 Testing
Italo Liberatore
Allison Rossi
Donald Fyhr
Abstract:
Beginning in late 1997, the unit heat rate of Brandon Shores Unit 2 has steadily increased
with no clear explanation of the root cause. Of the increase of approximately 500
Btu/kwhr, our Performance Monitoring Program has not explained 400 Btu/kwhr. To
identify the cause of this increase, Constellation Power Source Generation plans to
perform a full scale ASME PTC 6-1996 turbine heat rate on the unit during early 2001.
General Introduction:
The Brandon Shores Unit 2 boiler is a Babcock and Wilcox balanced draft, natural
circulation, Carolina Type radiant boiler, firing pulverized coal with No. 2 oil igniters.
The boiler is rated for 4,425,000 lb/hr of steam at 2520 psig and 1005/1005F. The
turbine-generator, supplied by General Electric Company, is rated for 680,000 kw output
at 3600 rpm. The turbine is a tandem-compound, double flow reheat with two double
flow low pressure sections. Commercial operation began in 1991. An SCR was recently
installed in late 2000 to reduce NOx emissions. A baseline turbine heat rate test was
performed in 1993.
Existing Program:
The Performance Monitoring Program consists of three primary sections:
Heat Rate/Thermal Performance Tracking: Monitoring and trending heat rate

deviations and long term/gradual performance problems.
As part of the Heat Rate/Thermal Performance Tracking we conduct routine
testing at similar steady state full load conditions using plant instrumentation. We
set the main steam and reheat steam as close as possible to design/optimum rated
conditions. Pressure, temperature, and flow data is collected, as applicable, for a
period of one hour preceded by one hour of settling out. Simultaneously, we
collect coal and ash samples. This information is used off line to calculate the
5-19
monthly average unit performance using the Output/Loss Method, as well as the
Input/Output Method. These values are used for trending of unit performance, as
described below.
Performance Engineering Test: Major equipment testing, as part of the Predictive

Maintenance Program, using portable calibrated instrumentation.
When routine monitoring identifies the need or for plant equipment not
instrumented for test, the Predictive Maintenance Engineering Unit is called in to
conduct a performance test using portable calibrated instrumentation. Testing is
done following detailed test procedures, which are based on the ASME Test
Codes.
Monitoring & analysis: Evaluation of the operators Controllable Loss Monitor

parameters.
The unit is equipped with an online controllable loss monitor to aid the plant
operators in maintaining optimum conditions to improve heat rate. The data is
trended for any indications of problem areas that need to be addressed.
Historical NPHR:
Routine performance testing using plant instrumentation has being done on a consistent
basis on this unit since February 1995. The plant instrumentation used in this testing may
not always give accurate results, but the trend in heat rate should be reliable.
Parameters calculated include, but are not limited to net plant heat rate (NPHR), turbine
generator heat rate (TGHR), HP & IP turbine efficiencies, boiler efficiency, feedwater
heater terminal temperature differences (TTDs) and drain cooler approaches (DCAs),
etc. For each piece of major equipment, performance indicators are calculated. The
expected value at the test load and/or steam flow is also determined. Deviations from the
expected value are determined for most common causes of inefficiencies and the sum of
the total losses calculated. These are known as accounted for losses. The actual NPHR
is compared to the expected NPHR and the difference between the total NPHR deviation
and the accounted for losses is defined as unaccounted for losses. Attachment 1 is an
example of the monthly report generated from the above. Attachment 2 shows the annual
averages for accounted for and unaccounted for losses.
Since the beginning of this reporting/testing period, Unit 2 has had unaccounted for
losses that could not be explained or associated with any one cause. Since late 1997,
these losses have increased substantially, but the root cause(s) for these inefficiencies
have not been identified. Attachment 3 shows the trends of net plant heat rate deviation
and unaccounted for losses over time.
5-20
Trending and boiler testing has indicated that the increase in heat rate is in the turbine
cycle. Therefore, all efforts have been concentrated on the causes for higher turbine heat
rate.
Special Testing/Evaluation:
To ensure the increase in heat rate was not an instrument problem, the plant instruments
that affect heat rate the most were verified for accuracy by comparing their indications
with portable, calibrated test instrumentation. Any instruments exceeding acceptable
limits were calibrated/corrected. The results of this process found no significant
instrument problems.
Performance testing using calibrated test instrumentation, as part of the Performance
Monitoring Program, did not identify any major deviations that were not already
identified by the monitoring program.
In an attempt to narrow the search for the cause(s) of these inefficiencies, we compared
the performance of Unit 2 to Unit 1 while they were operating at approximately the same
conditions. The two units are very similar, allowing this comparison. This relative
comparison highlighted the differences between each other, as well as the differences
between them and their expected design (best achievable condition). The comparison of
the units heat rates showed Unit 2 NPHR calculated based on the Output/Loss method to
be higher than Unit 1. The NPHR calculated by the Input/Output method showed similar
results. Comparison of other parameters such as turbine efficiencies, pressure ratios,
feedwater heater performance, etc. only explained part of the higher heat rate.
Attachment 4 lists the comparison results.
Attachments 5, 6, 7 and 8 compare all of the major components in the turbine cycle. No
indication of the cause for the large deviation in performance can be attributed to any of
them. Although all indications show that the losses are not attributed to the boiler cycle,
it was also reviewed again. As suspected no major indication of efficiency changes over
time were found.
Therefore, we went back to the historical database and began to assess each of the input
data points used to calculate NPHR based on the output/loss method. The one variable
that stood out more than all of the others was the final feedwater (FFW) flow.
Comparison of the FFW flow measurement to the calculated FFW flow based on MBFP
suction flow measurements showed that the FFW flow measurement may be indicating
high (see Attachment 9). Comparison of the FFW flow measurement to the condensate
flow measurement supports this conclusion (see Attachment 10). If the indicated flow is
high, then the heat rate would indicate high as well. This would result in a higher than
actual increase in heat rate being shown. Although unlikely and uncommon, we
suspected that the FFW nozzle might be creating higher than actual readings. In order for
the nozzle to indicate higher than actual flow, it would need to have deposits at the throat
area of the nozzle. Since there were problems with copper deposits in the HP turbine in
the past, the historical trends were evaluated for possible explanation. The only way to
5-21
prove or disprove this conclusion is via a nozzle calibration test. The last time it was
done was in 1993, during a full scale turbine heat rate test. This test was run to establish
the baseline condition of the turbine.
To validate the historical trend for Unit 2, it was decided to compare the output/loss
method with another approach, based on continuous emission monitor (CEM) data. This
information is known to be indicating higher than actual, but since only the trends would
be compared, and therefore their relative changes, it was considered to be valuable. It is
noted that the heat rate based on output/loss method trended similar to the CEM heat rate
calculations. These two are completely independent, and therefore indicates that the FFW
flow may be indicating the correct/actual flow. Attachment 11 shows the trends of gross
plant heat rate based on the Output/Loss, Input/Output and CEM methods over time. The
MBFP speed increasing over time is also indicating that flow is probably increasing. If
the FFW flow indication is assumed correct, then the cause of the increase in heat rate is
still unknown.
When Industry Experts encounter this type of unit performance, and have exhausted all
other possibilities, they will normally conclude that the inefficiency is due to the LP
turbine performance, i.e., if it is not elsewhere, its in the LP turbine. A PTC 6 test will
thoroughly evaluate all other possibilities, and therefore could make/present such a
conclusion.
After rigorous and lengthy evaluation and comparison of the performance data collected,
the root cause(s) of BS2 unaccounted for losses could not be identified. The two areas
which are suspect, although not certainly proven, are the final feedwater flow and the LP
turbine efficiency.
To provide concrete evidence of these possible causes it was recommended to perform a
Turbine Heat Rate Test, in accordance with ASME PTC-6. After presentation of the
review findings to plant personnel, approval was received to perform a full scale turbine
heat rate test.
PTC 6 Test Plan
The test is being run using the ASME PTC 6-1996 full-scale test as a guideline. Turbine
heat rate, HP turbine efficiency, and IP turbine efficiency will be calculated. LP turbine
exhaust steam enthalpy and LP turbine efficiency will not be calculated due to unknown
split of flow between the two LP turbines (dual pressure unit).
In addition to the above, the following is also being performed:

Calibration of the plant final feedwater nozzle, main boiler feed pump
suction nozzle, and condensate flow orifice
Calculation of feedwater heater performance
Calculation of boiler feed booster pump and main boiler feed pump head
versus flow performance
5-22

Calculation of main boiler feed pump turbine efficiency (through 5th stage)
Plant heat rate based on input/output method (electrical output divided by
fuel input)
Six load points will be tested with two tests run at each load point for repeatability. Each
test will be two hours in duration following a one-hour settling-out period. The six load
points are:
Load Point
1
2
3
4
5
6
Valve Point
VWO
VWO
2nd
3rd
1st
VWO
Steam Conditions
2520 psig,1000F/1000F (5% overpressure)
2400 psig,1000F/1000F
2400 psig,1000F/1000F
2400 psig,1000F/1000F
2400 psig,1000F/1000F
2450 psig,1000F/1000F (normal full load)
The cycle will be isolated for testing load points 1 through 5. Load point 6 testing will be
run with normal operating valve positions. No soot blowing will be allowed during the
test periods. Turbine valve points will be set by measuring the feedback voltage signal
from the control valves to the EHC system.
Data points have been selected to meet the requirements of PTC 6-1996 except as
follows:

Design values will be used for boiler feed pump and boiler feed booster
pump seal flows (no flow elements installed in lines)
Feedwater heater pressure drops will be measured
Readings will be taken on plant final feedwater nozzle, main boiler feed
pump suction flow nozzle and condensate flow orifice (plant flow
elements are being calibrated for routine testing)
Main boiler feed pump turbine operating data (leak-off flow, 5th stage
pressure and temperature, etc.) will be read for efficiency calculation
Coal flows will be read and coal samples obtained for plant heat rate
calculation
Test instrumentation will be installed as shown in Attachment 13. There are 62 pressure
and 75 temperature readings required for the heat rate test. An additional 14 pressure and
1 temperature readings will be added for other testing. The total number of readings to
be taken each minute is 152.
All pressures (static and differential) will be read using calibrated pressure transmitters.
All temperatures, except some turbine seal flows, will be read using calibrated type E
thermocouples equipped with lead wires and reference junctions. The other turbine seal
flow temperatures will be read using uncalibrated type E strap-on thermocouples.
5-23
Data will be collected every one minute using a PC-based data acquisition system. Plant
operating data will be collected using the plant computer. Hotwell level, deaerator
storage tank level, and coal consumption will be read at the beginning and end of the test
and at thirty minute intervals between. Coal samples will be obtained from the coal
bunkers just above the feeders at the beginning of and at thirty-minute intervals during
the test. Pump speeds will be taken every ten minutes.
Condensate flow will be measured using a calibrated 16" Daniel throat tap nozzle
installed in a horizontal pipe run between feedwater heater No. 24 and the deaerator. The
nozzle was calibrated just prior to installation during an outage in December 2000.
The CPSG Electric Test & Generator Protection Unit will obtain the electrical
measurements using calibrated test instrumentation.
An Excel spreadsheet has been created to calculate test results.
The test series is expected to run be within a two-week period (schedule to be based on
dispatch requirements). Instrument set-up and data collection system checkout will be
done in a two-week period preceding the testing.
Post-Test Analysis
Upon completion of the test and calculation of results, several areas will be looked at
including the following.
The final feedwater nozzle calibration data will be loaded into the routine performance
test calculations. This will show if the indicated increase in net plant heat rate trend was
true.
A comparison will be made with the 1993 baseline test results. Some of the parameters
to be looked at are:

The change in turbine heat rate (calculated to determine if it is the cause of

the change in NPHR).
Performance of the HP and IP turbine sections
Feedwater heater performance
Main boiler feed pump and turbine performance
Seal flows
The effect of cycle isolation on unit output will be evaluated to determine its impact and
possible problem areas.
Plant instrument readings will be compared to test instrument readings to determine if
any corrections or calibrations are required.
5-24
Once all changes in turbine heat rate that can be attributed to specific sources (i.e.
HP/IP turbine, MBFPT, feedwater heaters, etc.) have been made, the remaining increase
in turbine heat rate (unaccounted for losses) will be attributed to the LP turbine.
Future Actions
In cooperation with EPRI, CPSG has agreed to use Brandon Shores Unit 2 as a
demonstration facility for testing two Real-Time Heat Rate Monitors.
The boiler performance will be tested with the new SCR in service to determine its effect
on unit performance.
The installation of data validation software on the plant computer system will be
evaluated.
5-25
Attachment 1
Brandon Shores Unit 2

Monthly Thermal Performance Evaluation Report
August 2000
MONTHLY AVERAGE PERFORMANCE
FULL LOAD PERFORMANCE
Monthly Average Operating Net Plant Heat Rate

1999 Monthly Average Operating Net Plant Heat Rate
10,099 Btu/kW-hr
10,329 Btu/kW-hr
Previous Rolling Average Operating Net Plant HR

Present Rolling Average Operating Net Plant HR
**** 2000 Year-End Rolling Avg. NPHR Goal ****
Change in Average Operating Heat Rate
10,137
10,125
10,228
-11
Fuel Cost Savings, Actual vs Goal
Btu/kW-hr
Btu/kW-hr
Btu/kW-hr
Btu/kW-hr
K$ saved
OTHER INFORMATION
Monthly Average Percent LOI
2000 YTD Percent LOI
1999 Average Percent LOI
8.00
7.14
7.29
Monthly Average Percent Auxiliary Load

2000 YTD Percent Auxiliary Load
1999 Average Percent Auxiliary Load
Percent Savings over 1999 Average Auxiliary Load
REMARKS (Affecting NPHR and/or efficiencies)

BOOS / 24FWH OOS
BOOS / 24FWH OOS
NONE
NONE
%
%
%
10.13 %
6.32 %
4.72 %
-33.92 %
Week
1
2
3
4
Date
08/09/00
08/23/00
#NUM!
#NUM!
673.2
Average Circ Water Temperature

Average Expected Net Plant Heat Rate
Average Measured Net Plant Heat Rate
Average Total Net Plant Heat Rate Deviation
94
9,753
10,506
753
Heat Rate Deviation Accountability:

Turbine Cycle
Average Gross Turbine Heat Rate, Btu/kW-hr
Average HP Turbine Efficiency, %
Average IP Turbine Efficiency, %
Average 27 Feed Water Heater TTD, F
Average 26 Feed Water Heater TTD, F
Average HP Condenser Pressure, in-Hg
Average LP Condenser Pressure, in-Hg
Average MS Temperature, F
Average MS Throttle Pressure, PSIG
Average Reheat Steam Temperature, F
Average Reheater Pressure Drop, %
Boiler & Auxiliaries
Average Boiler Efficiency, %
Average Excess Air, %
Average Exit Gas Temperature, F
Average Cegrit Percent LOI, %
Average Coal Moisture, %
Average Auxiliary Load Usage, MW-hr
Expected
Value
8,126
85.3
91.3
0.2
3.2
4.7
3.3
1,000
2,450
1,000
10.0
Actual
Value
8,681
83.3
89.7
12.5
0.2
4.7
3.4
998.7
2,453.6
998.4
9.7
88.1
19.0
323.5
7.0
6.0
36.8
86.8
21.0
331.9
9.9
7.2
35.4
Average of Total Accounted For Losses

Average of Total Unaccounted For Losses
12/20/00 12:44 PM
Attachment 2: Average Yearly Accounted For & Unaccounted For Losses
650
600
550
1998 Year Average
1999 Year Average
500
450
Btu/kwhr
400
350
300
250
200
150
100
50
0
-50
5-26
Mw-hr (gross)
2000 Year-to-Date Average
115
638
F
Btu/kW-hr
Btu/kW-hr
Btu/kW-hr
Plant Heat
Rate Deviation
(Btu/kW-hr)
37.3
19.2
23.7
3.6
-3.5
7.2
2.1
-0.9
2.2
-2.8
-12.2
17.8
35.0
10.6
-24.3
Btu/kW-hr
Btu/kW-hr
Attachment 3: Brandon Shores 2 NPHR Accounting of Losses

Heat Rate Deviation
Unaccounted For Losses
1000
800
BTU/KW-HR
600
400
200
8/21/00
5/21/00
2/21/00
8/21/99
11/21/99
5/21/99
2/21/99
11/21/98
8/21/98
5/21/98
2/21/98
8/21/97
11/21/97
5/21/97
2/21/97
8/21/96
11/21/96
5/21/96
2/21/96
8/21/95
11/21/95
5/21/95
2/21/95
-200
DATE
A ttachm ent 4: Heat R ate D eviation C om parison, B tu/kw -hr
G e n e ra to r g ro s s lo a d , m w -h r
U1
6 8 5 .4
U2
687
U2 - U1
1 .7
N e t p la n t h e a t ra te (b as e d o n b o ile r e ff)
C o rre c te d n e t p la n t h ea t ra te
N e t p la n t h e a t ra te , e xp e c te d
N e t p la n t h e a t ra te d ev ia tio n
9895
9913
9780
115
10205
10224
9609
596
310
331
-1 7 1
481
T u rb in e -g e n e ra to r h e at ra te
E x p e c te d tu rb -g e n h ea t ra te
T u rb -g e n h e a t ra te d ev ia tio n
8197
8103
94
8458
7998
460
261
-1 0 6
366
56
4
11
0
3
0
-1
-2
3
32
3
-1 8
7
9
108
43
8
17
4
-2
-2
0
-7
7
18
-1 1
29
43
0
147
-1 3
4
6
4
-5
-2
2
-6
4
-1 4
-1 5
47
36
-8
39
H e a t ra te d e v ia tio n - H P tu rb e ffic ie n c y
H e a t ra te d e v ia tio n - IP tu rb e ffic ie n c y
H e a t ra te d e v ia tio n - 17 /2 7 fe e d w a te r h e a te r
H e a t ra te d e v ia tio n - 16 /2 6 fe e d w a te r h e a te r
H e a t ra te d e v ia tio n - m a in s te a m te m p e ra tu re
H e a t ra te d e v ia tio n - m a in s te a m p re s s u re
H e a t ra te d e v ia tio n - re h e a t s te a m te m p
H e a t ra te d e v ia tio n - re h e a te r p re s s u re d ro p
H e a t ra te d e v ia tio n - H P c o n d e n s e r b p
H e a t ra te d e v ia tio n - LP c o n d e n s e r b p
H e a t ra te d e v ia tio n - flu e g a s o x y g e n
H e a t ra te d e v ia tio n - air h e a te r o u t te m p
H e a t ra te d e v ia tio n - LO I
H e a t ra te d e v ia tio n - m o is tu re in fu e l
Total deviation (accounted for)
NPHR Dev
481
121
39
Total Differential Losses

U n it 1 N P H R D e v ia tion
321
115
U2 "Unaccounted For" Losses
436
10
5-27
Attachment 5: BS NPHR Comparison (Test Date: 3/29/00)
10500
10400
Unit 1
Unit 2
10300
Btu/kw-hr
10200
10100
10000
9900
9800
9700
9600
9500
Net plant heat rate
(output/loss)
Net plant heat rate

(input/output)
Net plant heat rate,

expected
Corrected net plant

heat rate
(output/loss)
Corrected net plant

heat rate
(input/output)
Attachment 6: BS Turb Efficiency Comparison (Test Date: 3/29/00)
92
91
90
UNIT 1
UNIT 2
89
88
87
86
85
84
83
82
81
80
High pressure turbine
efficiency
Intermediate pressure
turbine efficiency
11
5-28
HP turbine efficiency,
expected
IP turbine efficiency,
expected
Attachment 7: BS Corrected Pressures Comparison (Test Date: 3/29/00)
2400
2200
2000
UNIT 1
1800
UNIT 2
1600
psia
1400
1200
1000
800
600
400
200
0
Corrected throttle Corrected 1st
pressure
stage pressure
Corrected hot
reheat pressure
Corrected cold Corrected 16/26

Corrected IP
reheat pressure fwhtr extract pres turbine exhaust
Attachment 8: BS LP Turbine Extraction Pressures Comparison (Test Date: 3/29/00)
80
75
70
UNIT 1
65
UNIT 2
60
55
50
psia
45
40
35
30
25
20
15
10
5
0
Corrected 14/24 Corrected 13/23
fwhtr extract pres fwhtr extract pres
Corrected
12B/22B fwhtr
extract pres
Corrected
12A/22A fwhtr
extract pres
Corrected
11B/21B fwhtr
extract pres
Corrected
11A/21A fwhtr
extract pres
12
5-29
5-30
13
DATE
400
200
0
8/21/98
5/21/98
2/21/98
11/21/97
8/21/97
5/21/97
2/21/97
11/21/96
8/21/96
5/21/96
2/21/96
11/21/95
8/21/95
5/21/95
2/21/95
8/21/99
11/21/99
2/21/00
5/21/00
8/21/00
8/21/99
11/21/99
2/21/00
5/21/00
8/21/00
5/21/99
600
2/21/99
800
5/21/99
1,000
2/21/99
Attachment 10: BS 2 Measured Final Feedwater Flow - Condensate Flow

11/21/98
DATE
11/21/98
8/21/98
5/21/98
2/21/98
11/21/97
8/21/97
5/21/97
2/21/97
11/21/96
8/21/96
5/21/96
2/21/96
11/21/95
8/21/95
5/21/95
2/21/95
KLBS/HR
KLB/HRS
Attachment 9: BS 2 Measured Final Feedwater Flow - (MBFP Suction Flow - Spray

Flows)
250
150
50
-50
-150
-250
Attachment 11: Brandon Shores 2 Gross Plant Heat Rate Comparison

CEM
Input/Output
Output/Loss
11250
Btu/kw-hr
10750
10250
9750
9250
Sep-00
Jul-00
May-00
Mar-00
Jan-00
Nov-99
Sep-99
Jul-99
May-99
Mar-99
Jan-99
Nov-98
Sep-98
Jul-98
May-98
Mar-98
Jan-98
8750
Month/Yr
Attachment 12: Brandon Shores Unit 2 Heat Rate Test Instrumentation

RHTR
SSH
F P
P T
T
P T
HP T urbine
P P
IP Turbine
P
T T
F
F
T T
P
P
P
T
P
B
C
D
LP B Turbin e
P
P
P
F
RHTR
LP A Turbin e
P
Generat or
A
E
F
T
T
T
C ond enser
F
P T
P
T
T
B FPT
SPE
F
P
P T
P T
P
T
P
T
L
T
TEST
NOZZLE
DEAER
PF
PF
MFBP
P
S T
BFBP
T - Tem pe rature
F - Fl ow
S
T
E - Ele ctric
T
P
PF
P - Pre ssure
P
T
T
E
SJAE
PF
P T
L - Le vel
S - Speed
- Nozz le
- Orifi ce
- GE F/R Tub e
DC F 12/18/00
14
5-31
6
SESSION 6: PLANT EXPERIENCES
6-1
Analysis of Variables for

Predicting Power output at the
Columbia Power Plant
Aravindan Rangarajan
6-2
Project Description
Columbia
Plant is 535 MW T-fired boiler

Project is intended to establish the
statistical relationship of critical
variables to the unit output in order to
optimize performance and identify
critical operator-controllable parameters
Project involved statistic processing of
large volume of operating data
6-3
Correlation of different variables

with Power Output
6-4
A list of 70 variables was examined for

predicting the power output at the Columbia
Power plant.
Data for all the variables was accumulated
between May 1st 2000 to August 15th 2000
A list of 13 variables have been identified as
critical variables which affect the Power
output the most
These critical variables either have a high
positive correlation or a high negative
correlation.
Examples of Critical Variables

Corr. Coefficient
0.955
Secondary Air temperature
Combined Air flow total
0.960
Coal Flow
0.976
Air Heater Cleanliness Factor
0.899
PSH Cleanliness Factor
-0.890
Waterwall Cleanliness Factor
-0.587
6-5
List of Critical Variables (contd.)

Economizer
Gas out Temp

Air Heater Gas Pressure Drop
Air Heater Gas in Temperature
Flue Gas Oxygen
Economizer Gas in Pressure
Economizer Gas out Pressure
6-6
0.943
0.926
0.894
-0.941
-0.759
-0.949
Examination of other variables

The
initial (raw) coal fuel flow correlation

was poor
A further examination was undertaken
to verify other variables that affect coal
flow.
These other variables were initially
categorized as non-critical or in some
cases are not recorded in the plant data
acquisition system.
6-7
Raw Fuel flow correlation with the

Power Output
600
500
GMW
400
300
200
100
0
500
Total
6-8
1000
List of Variables which affect the

Fuel vs Power Output Graph
Correlation Coefficient
Moisture
-0.49
Moisture vs Ht content - 0.8 (average)
Ash
vs Ht content -0.25 (average)
Sulphur vs Ht content -0.25
(average of all months)
Its apparent that moisture, sulphur and Ash
affect the Heat content. Corrected heat
content correlates directly to Power Output
6-9
Corrected Graph of Fuel

vs Gen MW
600
500
GMW
400
300
200
100
0
500
Total
6-10
1000
Regression Equation with Critical

Variables
A
regression equation was developed

for 14 critical variables as a Linear
equation
Analysis of the standard deviations
show that Air heater and Primary SH
cleanliness factors have high variability.
6-11
Standard Deviations of some

important predictor variables
Predictor
StDev
AH A Cleanliness
5.173
9.263
PSH Cleanliness
AH A Gas
0.2376
This shows that the AH A and PSH
cleanliness factors are not in Statistical
control.
There may be interactions that are not fully
developed in this analysis.
6-12
Variability in Power Output

X-bar Chart for GMW
580
6
540
Sample Mean
560
3.0SL=564.6
2
520
X=517.1
500
6
480
-3.0SL=469.5
460
440
420
400
0
10
20
30
Sample Number
6-13
Interaction among the Variables

Burner tilt position on water cleanliness factor
( 0.7 correlation)
Sec.SH cleanliness with Prim.SH cleanliness
(0.55 correlation)
Flue gas oxygen with economizer gas
pressure ( 0.9)
Interaction shows that the noncritical
variables could affect the correlation of the
critical variables to Power Output
6-14
BURNER TILT POSITION
Interaction among Variables
10
-10
0.8
0.9
1.0
1.1
1.2
1.3
Waterwall Clean Fact - Raw
6-15
Inferences and Conclusion
6-16
The graph of the power output vs coal input is

affected by variables such as moisture, sulfur
and ash present in the coal. This brings
variability in the raw data graph.
The non-critical variables affect the
correlation coefficients of the critical variables
and can be taken as secondary critical
variables.
Inferences and Conclusion
For a better prediction of power output it is

necessary to take into consideration other
variables which affect the critical variables.
Some of the critical variables have a high
standard deviation which explains why some
correlations appear to be out of statistical
control at times. Further examination of
additional variable relationships may be
necessary or better statistical control may be
considered.
6-17
12th EPRI Heat Rate Improvement Conference, Dallas, Texas, January 2001.
EXPERIENCE OF MONEYPOINT POWER STATION IN RECOVERING

PLANT HEAT RATE: FOCUS ON CONTROL VALVES
Michael Rocke, BE Chartered Engineer
Technical Services Engineer, Moneypoint Generating Station
Electricity Supply Board, Ireland
E-mail: mick.rocke@mail.esb.ie
Tom Canning, B.E.

Manager, Thermal Performance
Electricity Supply Board (E.S.B.), Ireland.
E-mail: tom.canning@mail.esb.ie
Sanjay V. Sherikar, Ph.D., P.E.

Manager, Plant Betterment
Control Components Inc.,
Rancho Santa Margarita, CA, U.S.A.
E-mail: svs@ccivalve.com
Abstract
This paper reports the results of a program undertaken to recover plant heat rate by
Moneypoint Power Station, a 3 x 315 MW coal-fired power station, owned and operated
by the Electricity Supply Board (ESB) of Ireland. The focus of this program was control
valves. The results of to date indicated that an overall performance improvement of 4%
for the station is possible by eliminating valve-related inefficiencies.
The priorities at this station in recent years have been, and still are, availability of the
Units and NOx control. Having recorded improvements in these areas over the years, a
systematic program focussing on valves was undertaken to recover efficiency loss
attributable to valves. An initial study was done as the first part of this program. This
study showed that performance improvement equivalent of 37 MW for the station is
achievable with the correct valve performance. The problems/losses in all the cases were
caused because the technology in those valves was not suited for the specific
applications. Leaking valves constituted a majority of the sources of loss.
Recommended changes in seven of the problem valves applications in Unit 2 were made
in May 2000. Measurements at these valves indicate an estimated gain of 7.4 MW.
6-18
Background at Moneypoint Power Station

Moneypoint station of Electricity Supply Board (ESB) has 3 identical coal-fired Units,
305 MW each, with Foster Wheeler drum boilers and Asea Brown Boveri Steam
Turbines. They were commissioned successively in 1985, 86 and 87. The nominal
rating for the main steam on these Units is 160 bar at 538C, with reheat steam also at
538C. The station operates at base load for most part. Over the years, major valve
problems were solved, or resolved, at least to the degree that they did not appreciably
affect the plant operation and its ability to produce at the rated capacity.
In recent years, efforts were focussed on improving availability of the station, which
brought in good results in that respect. Next, Moneypoint station took the initiative to
improve the heat rate of the Units. It was recognized that, given tight plant margins and
availability constraints, the elimination of valve-related losses held the most significant
potential for quick improvement.
Influence of Control Valves on Plant Operation & Heat Rate

Control valves are the final control elements in the operation of a power plant.
Therefore, under-performance of the valves directly affects plant operation - in terms of
output, heat rate and/or in terms of reliability and availability. Recent studies indicate
that eliminating control valve problems alone can improve the heat rate of power plants
in the range of 2% to 5%.
Figure 1 shows a simple process in which the control system generates a control signal, a
valve that operates according to the signal and then a feedback sensor that relays the
parameter being monitored to the control system. The weakest link in this control loop
will be the limiting factor in the control of such a process. Even sophisticated digital
control systems (DCS) or modern feedback sensors cannot make up for the limitations in
the performance of a control valve.
CONTROL
SYSTEM
Feedback
loop
Control
signal
Sensor
PROCESS
Figure 1. Simplified diagram of process control.
6-19
Severe Service Valves. Of the hundreds of control valves in any power plant, some
valves experience particularly tough operating conditions either at all times or under
some operating conditions. These are known as severe service valves. Although they are
few in number, they pose challenges to maintenance and operation. In most cases,
problems are caused by the misapplication of general service valves into severe service
duty. These severe service applications commonly also affect efficiency more than the
rest of the valve population. Conversely, eliminating problems in a stations severe
service valves offers one of the quickest and most effective means of improving the
efficiency of an existing power plant.
There are three basic modes through which valve-related penalties occur:
Loss of production (MW-hours) due to their unavailability/unreliability,
Penalty in heat rate or MW-capacity due to leakage,
Penalty in heat rate or MW-capacity due to poor control.
In addition, other components have to work harder to compensate for such losses, which
affects their life-cycle.
While contributions to plant efficiency loss from individual valve applications may be
small, together they can sum together to significant levels. When the invisible effects of
the valve problems are taken into account, the net impact is even greater.
Methodology
The program to recover the performance loss attributable to control valves can be divided
roughly into three broad phases:
1. A study to identify valves which affect plant performance and solutions for the
same,
2. Implementation of recommended solutions, and,
3. Follow-up, or monitoring, to ensure that the desired results are achieved.
The first phase of this effort, the study focussing on valves, had the following objectives:
Identify the loss in efficiency and the penalty in heat rate due to the poor performance
of control valves in the system,
Diagnose the root cause of these valve problems,
Recommend solutions, which will eliminate the root causes of these problems, and,
Recommend changes in valve-related systems to improve the reliability and
operation.
6-20
In gathering the necessary information, assistance was solicited from different groups at
the site. Discussions were held with plant staff from the Control and Instrumentation
Group and the Technical Services Group as well as Maintenance and Operations staff
generally. This staff is involved in day-to-day operation, either directly or indirectly, and
has good visibility of problems that are being experienced in this plant.
Based on discussions with the staff, the operation of the system and the design basis for
each of the valves under study was reviewed. Plant data was gathered for review, which
included the following:
Flow diagrams for the startup system and the Unit

Startup and shutdown procedures
Heat Balance Diagrams for the system
Operating parameters from the plants data acquisition system
Datasheets for control valves where available
Operating Reports for Moneypoint station
Most importantly, the historical performance of severe service control valves of interest
in this study was reviewed. In addition, their operation was observed from close
proximity. This allowed checking of the valves in their closed position for leakage, as
evidenced by abnormally high downstream temperature, and/or sounds generated by
leakage flow.
Calculations for quantification of losses were based for the most part on standard
procedures and had been established earlier [Ref. 1-3].
The second phase of this program was to implement reliable long-term measures to
recover the losses caused by individual valves. This required the station engineers and
the valve experts responsible for designing and manufacturing them to work closely
together. Good co-ordination between the two was necessary to make sure that all the
design/operating conditions were considered and that different options based on the
system requirements were evaluated. Unit 2 overhaul was due in when the results of the
study became available. In the limited time available, seven (7) valves were identified
for replacement.
The third phase in this program is the follow-up, or monitoring to ensure that the valves
performed as expected. As poor isolation was the main problem in most cases,
temperatures upstream and downstream of these valves were key indicators.
Measurements were made both before and after the overhaul.
6-21
Results
Results from the study, Phase 1 of the program, identifying the losses attributable to
individual valves are shown in Table 1 below. The loss in each of Units 2 & 3 was
equivalent of 14.2 MW. Unit 1, which already has a new Start-Up valve capable of tight
shutoff, was losing 8.4 MW. By far, the biggest component of loss was attributable to the
leakage past the Start-Up valve at 5.8 MW. The other major components of loss were
emergency heater drain valves (0.8 MW), steam drain valves (2.3 MW), drum blowdown
valve (1.7 MW), and spraywater valves (1.7 MW).
Application
TURBINE BYPASS SYTEMS:
Start-Up Valve, Units 2&3 only
HP Bypass Valve
Penalty in Heat
Rate
Estimated
Eq. MWe-Loss
Estimated
Leak rate
1.9 %
0.02 %
5.8 MWe
0.06 MWe
5 kg/s
0.43 kg/s
0.18 %
0.54 MWe
0.43 kg/s
LP Bypass Valve
FEEDWATER CONTROL SYSTEM
FW Control Valves
HEATER DRAIN SYSTEM:
#7 Heater Emergency Drain
0.12%
0.36 MWe
0.15 %
0.10 %
0.44 MWe
0.30 MWe
1.28 kg/s
1.06 kg/s
#6 Desuperheater Emergency Drain
0.04 %
0.12 MWe
0.42 kg/s

DRAIN & BLOWDOWN SYSTEMS
Drum blowdown, V35
Main Steam Drain Valves
(RT01/RT02/RT05 S001/S002)
Steam drain, V61 & V66
0.06 %
0.17 MWe
0.80 kg/s
0.54 %
1.74 MWe
2.70 kg/s
0.27 %
0.81 MWe
0.63 kg/s
0.33 %
1.00 MWe
0.90 kg/s
Other steam drain valves

SPRAYWATER SYSTEMS
Aux. Steam Spraywater valves
Superheat Spraywater valves
0.16%
0.47 MWe
0.30 kg/s
0.22 %
0.14 %
0.66 MWe
0.43 MWe
0.98 kg/s
HP Bypass Spraywater valves
0.21 %
0.64 MWe
2.5 kg/s
UNRELIABILITY {lost production}
0.21%
0.62 MWe
(Unit 1)
2.8%
8.4 MWe
(Units 2&3)
4.7%
14.2 MWe
TOTAL
Table 1. Estimate of loss due to individual valves
6-22
The quantification of losses attributable to individual valves helped in prioritizing actions

to eliminate these losses. Unit 2 was getting ready for the outage about the time the study
was concluded.
In the time-frame available, the following valves were replaced in May 00:
Startup valve, 12 x 12
Emergency Heater Drain valves (#7, #6 and #5), 6 x 6
Main Steam Drain valves (RT01, RT02 and RT05), 2 x 2
The temperatures in the vicinity of these valves were recorded before and after the
overhaul, i.e. before and after the old valves were replaced with new valves featuring
fluid velocity control. These are shown in Table 2. The low temperatures after the
modifications indicate that the new valves shutoff tight and have practically eliminated
losses that were occurring earlier in these services.
VALVE/
TAG #
Main Steam Drain
RT01S001
Main Steam Drain
RT02S001
Main Steam Drain
RT05S001
Startup Valve
SF012S001
HP Heater #7
Emergency Drain
02RP70S001
HP Heater #6
Emergency Drain
02RP40S001
HP Heater #5
Emergency Drain
02RP20S001
TOTAL
TEMPERATURE
UPSTREAM
Before
After
326 OC
27 OC
TEMPERATURE
DOWNSTREAM
Before
After
180 OC
26 OC
ESTIMATED
GAIN
323 OC
27 OC
223 OC
27 OC
0.27 MWe
319 OC
27 OC
137 OC
27 OC
0.27 MWe
125 OC
57 OC
5.8 MWe
236 OC
128 OC
50 OC
37 OC
0.44 MWe
175 OC
45 OC
30 OC
40 OC
0.30 MWe
34 OC
42 OC
29 OC
37 OC
0.27 MWe
7.35 MWe
Table 2. Unit 2 Temperature survey of critical valves before and after May
2000 overhaul.
No significant changes in temperatures were observed after six months in service, which
indicated continued tight shutoff.
6-23
Discussion
The estimate of the improvement potential in the Moneypoint study is consistent with
published literature and industry experience [Ref. 4]. A coal-fired supercritical station in
the U.S. had noted a 1.2% heat rate penalty within a year after a major overhaul due to
just the drain valves [Ref. 5].
It was observed that, among all the valves, the losses of significance occur mostly in the
severe service applications. Further, in most such cases, the problems occur as a result of
the existing valves being inadequate for that service from technology/design and/or
selection perspective. Maintenance and/or operational changes are not a solution when
this is the case because they does not address the root cause of those problems.
Controlling fluid velocity along the flowpath is perhaps the single-most important feature
that is necessary when p across the valve is high. Failure of conventional valve designs
in high p applications is attributable to high fluid velocities, which lead to problems
such as premature erosion leading to loss of shutoff capability, high vibrations,
cavitation, and so on [Ref. 6]. Figure 2 illustrates the origin of high velocities along the
flowpath inside such valves, in which the process of pressure reduction occurs in one
step. High velocities occur at the vena contracta, or minimum flow area, which may
cause the liquid to boil; subsequent pressure-recovery downstream causes cavitation.
VALVE
Inlet
Pressure
Inlet
Velocity
Pinlet
Pressure
Inlet
Outlet
Vvc
vc = Vena
Voutlet
Vinlet
Velocity
Poutlet
Va p o u r
Pressure
Cavitation occurs in this

pressure-recovery region
Pvc
Figure 2. Pressure letdown process and velocity along the flowpath inside
conventional valves for liquid flow.
6-24
All this points to the fact that having the correct technology in the severe service valves is
a must if such losses have to be eliminated. One tool for ensuring this is to have a good
specification it forces better selection of valve for a given application [Ref. 7, 8].
Example of a good generic specification, specifically for severe service control valves,
can be found in Reference 9.
DRAG technology for severe service valve applications. DRAG technology was
used in the new severe service valves at Moneypoint Power Station. It is a combination
of an unique hardware design and good engineering practices, all of which have evolved
from experience in solving severe service valve problems over thirty-five years. It
attends to application details for each service as part of the design procedure. Such an
approach eliminates the risks, and losses, that are inherent with trial-and-error or
solution-by-trials.
Fluid velocity is controlled in DRAG valves by providing many tortuous path stages of
pressure reduction. This eliminates high fluid velocities as shown in Figure 3.
VALVE
Inlet
Pressure
Inlet
Velocity
Pinlet
Inlet
Outlet
Pressure
Vinlet
Voutlet
Velocity
Va p o u r
Pressure
Poutlet
Figure 3. Pressure letdown process and velocity along the flowpath inside DRAG
valves for liquid flow. Fluid velocity is controlled within safe limits all along the
flow-path, which eliminates cavitation.
6-25
A section of a DRAG valve is shown in Figure 4 to illustrate the velocity control

principle. The multi-path, multi-stage flowpath effectively dissipates the fluid energy
associated with the valve pressure-drop within the trim itself. This type of design allows
20, 30, 40 stages whatever number that is required, to keep the fluid velocity at the
trim exit within the safe limit. As a result, the fluid exiting the trim has much lower
levels of kinetic energy, typically less than 5% of the total energy dissipated in the valve.
Consequently, destructive and unwanted effects such as cavitation, premature erosion,
vibration, noise, etc. are eliminated at their source, which is the high fluid kinetic energy
at the trim exit. In conventional technology valves, most of this energy dissipation occurs
through fluid turbulence outside the trim within the valve body cavity, or in the
downstream pipe, leading to the problems described earlier.
Figure 4. Sectional view of DRAG trim inside a valve body.
6-26
The velocity control principle was applied, beyond the valves, to

the main steam drain lines at Moneypoint. Occasionally blowing of
holes in the drain line elbows, typified by Figure 5, delayed startups
of the Units. This was attributed to magnetite and high velocities in
the drain line during startup when the drain valves are open. By
installing new correctly sized valves, steam velocity in each of the
critical lines was reduced by a factor of 7; Since erosion is
proportional to (velocity)n where the exponent n varies between 2
and 4, the corresponding erosion rates are expected to drop
drastically and thereby practically eliminate the problem.
Figure 5
Conclusions
Major findings from the initiative at Moneypoint, ESB, are:
1. Heat Rate can be improved by an estimated 4.1% by eliminating the inefficiencies
attributable to valves. This is consistent with previous industry experience.
2. Long-term reliable performance in severe service valve applications requires that
correct technology be used. Experience shows that fluid velocity control along the
flow path, as in DRAG technology, is necessary for long-term reliable performance.
3. A systematic approach, focussed on eliminating valve-related losses, can improve
plant heat rate significantly. Both the initial system-wide study and the follow-up in
monitoring performance continually are key to achieving the full potential benefits.
Expertise is available for reliable analysis of inefficiencies due to control valves in
existing power stations, and for the specification of long-term solutions.
4. Ease of implementation, quick payback and low risk all combine to make this
approach very attractive for heat rate improvement.
References
1. Byrne, L. and Sherikar, S.V., Improving Efficiency at Moneypoint Power Station: Focus On Control Valves,
January 2000.
2. Sherikar, S.V. and Puri, A., Importance of Startup System Isolation in Recovering MW-Loss and Heat Rate in
Power Plants, International Conference on Power Plant Operation, Efficiency and Environmental Protection,
New Delhi, India, 8-11 February, 2000.
3. Sherikar, S.V., Sterud, C.G., Bhate, B.H. and Strother, J., Modernization of the Startup System at Paradise
Power Plant, Report for TVA Contract #99-PYN-247276 (1999).
4. Cotton, K.C., Evaluating and Improving Steam Turbine Performance, 2nd edition, p. 296, pp. 305-313, (1998).
5. Weeks, Ed, Personal Communication, (May 2000)
6. Miller, H.L., and Stratton, L., Fluid Kinetic Energy As A Selection Criteria For Control Valves, ASME Paper
FEDSM97-3464.
7. Miller, H.L., Frequent Control Valve Problems, Seventh EPRI Valve technology Symposium, Incline Village,
NV, May 26-28, 1999.
8. Sherikar, S.V., Technology In Severe Service Control Valves, 15-th Annual Air-Operated Valve Users Group
(AUG) Meeting, Tucson, AZ, June 9-12, 1998.
9. Control Valve Technical Specification, ISA Guideline Compliant Specification for Control Valves (based on
Control Valves Practical Guides for Measurement and Control, ISA, 1998), CCI-LIT-350.
6-27
HEAT RATE IMPROVEMENT IN AN

EXISTING MULTIFUEL UNIT
Joaqun G. Blas
Florentino Blanco
Hidroelctrica del Cantbrico, S.A.

Heat Rate Improvement Conference January, 30 2001 February, 1 2001
6-28
The New Scenario

In the new competitive scenario, power
stations must face:
To Reduce the generating costs.
To Maintain high availability, efficiency and operational
flexibility.
To Meet strict environmental conditions.
To Manage and extend the equipment life, including
systems modernization.
TO REDUCE THE GENERATION COSTS

Heat Rate Improvement Conference
January, 30 2001 February, 1 2001
6-29
The Generation Cost Reduction

THE SITUATION
The kWh fuel cost (Europe) = 70% to 90% the variable overall cost
The Marginal cost The Variable Cost

The Marginal Cost, decides the competitiveness of the electric
units in a generating pool
The Fuel cost components:
The kilocalorie cost (pta/kcal)
The Net Heat Rate (kcal/kWh)
OBJETIVE
To reduce the marginal cost through the heat rate improvement
TO IMPROVE THE HEAT RATE

6-30
Heat Rate improvement in an existing unit
n (%) = 100 860

NHR
n = b
tg
n = Unit net efficiency

b = Boiler efficiency
tg= Turbo Generator efficiency
a = Transformation and Auxiliaries efficiencies
NHR = Net Heat Rate
OPPORTUNITIES TO INCREASE EACH OF THE ABOVE FACTORS

6-31
Boiler efficiency improvement

Can only actuate significantly on:
Heat loss due to unburned fuel (latent heat).
- Fuel characteristics.
- Combustion condition (air excess and its distribution).
- Other factors.
Heat loss due to heat in the flue gas (sensible heat).
- Adjust exit flue gas temperature and mass flow.
Heat loss due to surface radiation and convection.
- Boiler insulation.
- Preferential path ways and convection
cooling air flows.
6-32
Turbo Generator efficiency improvement

Can only normally act on:
Losses caused by internal and external leaks.
- Through the seals between stationary and rotating
parts.
Losses caused by pressure and/or temperature deviations
in the exhaust.
- Cleaning system of the condenser tubes
- Feed water heaters drainage system
Losses in the electric generator (without considering both,
the core and armature)
- Improving the efficiency of the excitation system
6-33
Transformation and Auxiliaries efficiencies

The actions are centered on:
Improving the efficiency of the Boiler Feed Pumps (BFP)
drive
Improving the efficiency of the IDF and FDF
Improving the efficiency of the other electric consumptions
and energy savings, where possible

6-34
Application
Aboo Multifuel Power Station
Two Units (I & II) rated respectively 360 MW and 543 MW
ABOO I
FUELS:
Pulverized Coal (domestic and imported)
Low BTU siderurgic gases (BFG and COG)
Heavy Fuel Oil
Schematic Diagram
UNIT I
Rated Output
Commisioning date
Accumulated operation hours
360 MW
1974
167,000

6-35
Actuations: Boiler Area (I)

New IDF and FDF.
- IDF: New induced draft fans with variable pitch blades.
- FDF: New high efficiency centrifugal fans.
Investment: 750 Mpta.
Operation and Diagnosis operation System (ODIS)

- Phase I: Continuous operational data collecting, its
analysis and comparison, for both, to improve the
process knowledge and to transmit them to other systems.
- Phase II: Use of the neural networks and ODIS connection
with the maintenance and alarm management systems.

6-36
Actuations: Boiler Area (II)

Economizer modifications.
- Surface extension by 15% increase

- Flow guide baffles
Actuation on the soot-blowing process
- New supervisory system for process optimization

Air heaters
- Replacement of the original Rothemles soot blowing
system
- New flow guide baffles installation in the tubular primary
air heaters

6-37
Actuations: Turbogenerator Area

HPT and MPT packings.
- Change of shaft packings and strips seals repairs.
Generator Excitation System.

- Change of the power equipment and the voltage regulator.
Condenser and Feed Water Heaters

- Condenser cleaning system improvement.
- Feed Waters Heaters drainage system optimization.

6-38
Actuations. Auxiliaries Area

Induced Draft Fans and Forced Draft Fans.
- As mentioned previously in Boiler Area.
Electrostatic Precipitators.
- New control system.
Boiler Feed Pumps.

- Variable speed gearings.
- Pump seals modifications.

6-39
Results: General
Relative improvement of the net efficiency = 3.1%
Net Heat Rate passed from 2,598 kcal/kWh to 2,520 kcal/kWh
Savings
1. Under operating conditions of 7,500 equivalent hours per year
and with a fuel cost of 1.20 pta/termia, the amount saved is
233 Mpta per year
2. Amount obtained by being able to burn a greater percentage
of BFG, which goes from 20% to 30% in terms of energy,
112 Mpta per year
Total savings:
345 Mpta per year
Total investment:
1,500 Mpta
Internal rate of return (IRR): 25%
6-40
Discussion: Impacts assessment

The separation of the impacts was made as follows:
Boiler :
ASME test
Specific test and Plant data
Turbogenerator : Complete Heat Balance of the Unit

Manufacturer estimations
Specific test and Plant data
Transformation
and auxiliaries:
Directly from Wh meters

6-41
Boiler Impacts Separation

Relative Efficiency Improvement= 1.88% = 47.5 kcal/kWh
(from 86.06% to 87.68%)
15C reduction in exit gas and

better control BFG temperature = 0.8 % = 20
kcal/kWh
Unburned reduction (excess air,

more BFG) and fuel mix (ODIS) = 0.6 % = 15
kcal/kWh
Soot blowing and gas flow

redistribution

6-42
= 0.5 % = 12.5 kcal/kWh
Turbogenerator Impacts Separation

Relative Gross Efficiency Improvement= 0.6 % = 15 kcal/kWh
(from 41.81% to 42.06%)
Shaft Packing changes and Strip

Seals repairs
= 0.35% = 9 kcal/kWh
Excitation system modernization
= 0.05% = 1 kcal/kWh
Condenser cleaning and FWH

optimization
= 0.2 % = 5 kcal/kWh

6-43
Transformation and Auxiliaries

Impacts Separation
Relative Net Energy Improvement = 0.6% = 15 kcal/kWh
(from 91.5% to 92.05%; 2 MW auxiliaries reduction)
FDF and IDF replacement
= 0.3% = 7.5 kcal/kWh
Electrostatics Precipitators
consumtion optimization
= 0.05% = 1 kcal/kWh
BFP speed gearing and

seals modifications
= 0.25% = 6.5 kcal/kWh

6-44
Impact of the Actions. Summary

Element
NHR improvements
UNIT
BOILER
Economizer and S IOD
IDF and FDF and Fuel
S oot-blowing and AH
%
3.1
1.88
0.8
0.6
0.5
Kcal/Kwh
78
47.5
20
15
12.5
T URBOGENERAT OR
MPT and IPT
0.6
0.35
15
9
Generator
Condens er and FWH
0.05
0.20
1
5
AUX ILIARIES
IDF and FDF
Electros tatic precipitator
0.6
0.3
0.05
15
7.5
1
BFP (drive and pump)
0.25
6.5
Initial Unit NHR

Final Unit NHR
--3.1
2,598
2,520
S teps performed
S ee below
S ee below
S urface extens ion and new ODIS
New fans and fuel mix control
S hoot blowing optimization baffles
ins tallation and others
S ee below
New s haft packings and s trip s eals
repairs
Modernis ation of the excitation
Cleaning s ys tem improvement and
controls optimization
S ee below
Fans replacement
Optimization of electric
cons umption
Variable s peed gearing and pump
s eals modifications
--S ee above

6-45
Conclussion
Improvements in the Heat Rate can
make competitive existing units.
A complete and systematic study most be performed, including:
- Potencial improvements in the Heat Rate

- Associated operating costs reduction
- Investments
- Profitability
Also, should include:
- Unit actual condition and heat rate
- Remanent and extended life
- Fuel costs
- Environmental restrictions, and others
6-46
Blast Furnace Heating Circuit

6-47
ABOO POWER PLANT
Unit I Boiler Sectional view

6-48
Neural Network for BFG heating control

Neural Network topology used for the prediction
of the heating water flow for the BFG (C512).
Error evolution during the network training and

histrogram error between real and estimated values

6-49
HEAT RATE IMPROVEMENT IN AN EXISTING MULTIFUEL UNIT
Joaqun G. Blas
Florentino Blanco

Abstract- With the electricity markets deregulation, and particulary the competition in
the electricity generation, power producers are forced to reduce their kwh costs. This
paper proposes a systematic actuation for tackling the potential oportunities of increase the efficency of the conventional thermal power stations in service and so reduce the kwh cost.
The application to a multifuel thermal unit shows, that with a reasonable investment, is
possible to improve the competitiveness of the existing plants in the spanish new scenario through a systematic study of the potential improvements in the heat rate and
the associate savings and the profitability of the required investment to get it.
Index terms: Deregulation, efficiency, net heat rate, generating costs.
1. INTRODUCTION
In the new competitive scenario, power stations must face:
-
To reduce the generating costs.
To maintain high availability, efficiency and operational flexibility.
To meet strict environmental conditions.
To manage and extend the equipment life, including systems modernization.
In Europe, the kWh fuel cost - depending on the specific circumstances of each plant
represents approximately between 70 and 90% of the variable overall cost of the generated kWh or, which is equivalent, the marginal cost, which decides the competitiveness of
the electrical units in a generating pool, where an abundant offer exists.
Therefore, the first step to be taken is to reduce the both components of fuel cost: the kilocalorie cost (pta/kcal), and the net heat rate (kcal/kWh).
Leaving aside the steps to be taken to ensure the adequate, economical and secure supply of fuel, we are going to deal with the net heat rate and the steps to reduce it and how to
do it.
In the following paper, we will examine the possibilities of improving the operating cost of a
existing unit, by increasing efficiency, refered to the net heat rate, NHR measured in
terms of kcal/kWh related to the HHV of the fuel, from which we obtain the net efficiency
860
n , whose value in percent is : n=100
NHR
Moreover, the net efficiency is the product of the efficiency of the three classic components
of the installation: i.e. the steam generator or boiler (B) the turbo generator (TG) and the
auxiliaries (A) so that n = b x tg x a where b , tg and a are respectively, the efficien-
6-50
2
cies of the boiler, the turbo generator (thermodynamic, mechanical, electrical) and transformation and auxiliaries.
The systematic work process leads us to deal with the opportunities, existing in real life, to
increase each of the above examined factors.
2. POSSIBLE ACTIONS
Bearing in mind that we always refer to a existing Unit, with up-to-date maintenance, reliability centered, with an adequate diagnosis system, the process leads us to the following
considerations:
2.1 BOILER EFFICIENCY
Among the various types of boiler losses for a given fuel (single or multifuel), can
only normally act significantly on the following:
-
Heat loss due to unburned fuel (latent heat).
Heat loss due to heat in the flue gas (sensible heat).
Heat loss due to surface radiation and convection.
The losses due to unburned fuel are related to the fuel characteristics, and the conditions and combustion characteristics (air excess and its distribution, pulverized
coal size, etc.) and other different factors. In our practical application, this aspect is
one of the most important elements that were involved: new axial Induced Draft
Fans (IDF) and centrifugal Forced Draft Fans (FDF) of higher efficiency, which
made possible to increase the excess of O2 when burning Blast Furnace Gas
(BFG), reducing simultaneously the auxiliaries consumption, and actions on the fuel
mix: Pulverized Coal, Blast Furnace Gas, Coke Oven Gas and Heavy Fuel Oil (PC
+ BFG + COG + FO), using a new Operation and Diagnosis Information System
(ODIS).
With regard to the loss due to heat in the flue gas, and maintaining always the boiler
exit gas temperatures according to the fuel sulphur (S2) content, the steps to take
will be focused on adjusting that temperature, working on the heat transference in
the recovery exchangers, basically by adecuate cleaning of the exchange surfaces
(air heaters, economizers, etc) using the soot-blowers, and in the case of wide temperature deviation, acting on the exchange surface itself and/or gas distribution
(pre-economizers installation or surface extension, distribution baffles installation,
etc). We will see that this area has also been included: we are refering to the Soot
Blowing Optimatization Progam, economizer modification and the work performed
on the Air Heaters (AH).
Finally, the steps taken to reduce the losses due to surface radiation and convention are included in the normal boiler maintenance, in order to keep the boiler insulation in a good condition, and avoiding preferential path ways and natural cooling
air flows.
2.2. TURBOGENERATOR EFFICIENCY
6-51
3
Among the different types of losses found in a particular turbogenerator, in the most
common cases we would normally act on the following:
-
Losses caused by internal and external leaks (through the seals between stationary and rotating parts).
Losses caused by pressure and/or temperature deviations in the exhaust.
Losses in the electrical generator.
The best way of reducing the losses caused by internal leaks, which could add up to
1% of the energy available in the turbine, is to keep in good condition the seals between the rotating blade tips and the shell, the seals between the stationary blades
and the shaft, and finally, the shaft end packings. This has also been a specific
case of our application where we have opted to change the shaft end packings in
the High and Intermediate Pressure Turbines (HPT and IPT) and to repair both stationary and rotating blade strip seals.
Normally, it is also strongly recommended to work on the cleaning system of the
Condenser tubes, on the condensate levels controls in the Feed Water Heaters
(FWH) and on the integrity of the preboiler system. In our case, improvements to
the Condenser cleaning system was given preferential attention and full advantage
was taken of the operating conditions of the feed water preheating system.
The losses reduction in the generator, without considering both the armature and
rotor winding, is focused on improving the efficiency of the excitation system, taking
advantage of modern electronic power in static systems, more efficient, which, in
our case, was an important area to work on.
2.2 TRANSFORMATION AND AUXILIARIES EFFICIENCIES
Leaving aside major modifications to the plant design, the possibilities of reducing
the auxiliaries consumption are centered on actuations such as:
-
Improving the efficiency of the Boiler Feed Pumps (BFP) drive.
Improving the efficiency of the IDF and FDF.
Improving the efficiency of other electric consumptions and energy savings,

where possible.
With regard to the first point, the steps that were taken includes the installation of
new internal in the BFP drive gearings, and the replacement of the pumps original
labyrinthic seals by mechanical seals, with cero leaks.
Regarding to IDF and FDF, the inicial fans were replaced by new axial IDF with
variable pitch blades, and by more efficient centrifugal fans in the case of FDF.
Finally, the actuation on the auxiliaries consumption was completed with a thorough
study, including each one of them. As a result of it, and among other measures, the
optimization of the energy consumption of the boilers electrostatic precipitators was
identified as one of the main potential source of savings.
6-52
4
3. APPLICATION
3.1.
GENERAL
In the following paper, we will examine the results, when applied to the Unit 1, 360
Mw, multifuel thermal power station in Aboo, owned by Hidroelectrica del Cantabrico, S.A. This plant was commisioned in 1974 and has a total of 167.000 operating
hours. A sectional view of the boiler is shown in Figure 1.
The plant uses pulverized coal, heavy fuel oil, and low Btu siderurgic gases (BFG
and COG). These gases are used in variable quantities depending on the operation
of the nearby integral iron and steel mill. The flowing characteristic of the siderurgic
gas supply drives the plant to operate more than 8,000 hours per year.
The total investment, corresponding to the measures taken, rises to 1500 Mpta,
splitted among the various items examined below.
3.2.
BOILER AREA
3.2.1 The New IDF and FDF
In order to increase the operational flexibility beyond the base load, while maintaining efficiency, new induced draft fans with variable pitch blades were installed and
the original forced draft fans were also replaced by high efficiency centrifugal fans.
As a result of it, was possible, in first place, to operate the boiler with higher excess
air, and in second place, to increase the quantities of BFG to be burnt in the boiler.
Consecuently, the losses due to unburned fuel were reduced by two ways: higher
excess air and more BFG used with cero unburned, as well as the energy consumed by the fans.
The investment reached 750 Mpta.
3.2.2 Operation and Diagnosis Information System (ODIS)
With a view to the optimization of the plant control proces, the project was carried
out in two phases:
-
Phase I.: The continuous operational data collecting, its analysis and comparison, for both to improve the process knowledge and to transmit them to other
systems.
Phase II: The use of neural networks and the ODIS connection with the maintanence and alarm management systems, in order to reduce the forced shutdown periods as well as to centralize the control and to improve the efficiency.
The completion of Phase I was a success, while the testing period for Phase II is
well underway, with the installation of the neuronal networks for the prediction of the
feed water flow used for BFG heating (Figures 2 and 3), which mass flow is subjected to frecuent and strong fluctuations because it is a process gas produced in
the nearby integral iron and steel mill plant, in spite of the use of a gas holder. The
investment of the project was approximately 100 Mpta, and as a result we have
6-53
5
achieved to optimise the heat exchange of hot water- BFG, keeping under control
both temperatures.
3.2.3 Economizer Modifications
With the target of reducing the temperature of the gases at the air heater outlet and
therefore improving boiler efficiency, it was decided to extend the economizer surface, by installing four new rows of tubes at the inlet side, taking advantage of the
available place, and so increasing by 15% the total economizer surface. Flow guide
baffles in the economizer inlet gas duct were also installed. As a result, the gas
temperature at the inlet of the air heaters, which are placed in the gas circuit behind
the economizer, was reduced by 25C, and consecuently, the temperature of the
boiler exit gases was reduced by 15C.
The investment was 300 Mpta.
3.2.4 Actuation on the soot-blowing process
To take full advantage of the Unit operation, a supervisory system was installed with
an investment of 30 Mpta. The most representative parameters of the plant operation are taken as a base, and boiler and Unit efficiencies are calculated on line
every five minutes. Based on it, the supervisory system detects and assesses economically the deviations and calculates the most reasonable and economical alternatives and, among other factors, it stablish the optimal soot blowing program.
The optimization of the soot blowing program not only reduces steam consumption
and the erosions caused by the soot blowing, but also improves the overall efficiency of the Unit.
3.2.5 Air Heaters
In the regenative air heaters, Rothemhle type, the original soot blowing system
with rotating type valves, was modified for an oscillatory type system, which ensures a more effective and uniform cleaning, improving the air heater efficiency.
In the tubular-type primary air heater, flow guide baffles were installed in the inlet
gas duct, and as a result tube pluggings were avoided, the primary air temperature
increased, and the exit gas temperatures reduced, improving therefore the boiler efficiency.
The required investment for all the above modifications and improvements reached
60 Mpta.
3.3 TURBOGENERATOR AREA
3.3.1 HPT and MPT Packings
Taking advantage of the overall inspection of High and Intermediate Pressure Turbines, the shaft packings were changed because the excesive leaks due to the
acumulated operating hours, as wells as rubbing, deformation caused by the shaft
and shell temperatures, operating incidents, and others. The rotating and stationary
blade strips seals were simultaneously repaired, with a total investment of 50 Mpta.
6-54
6
3.3.2 Generator Excitation
In the Generator excitation system, the power equipment was changed by installing
new thyristors which showed a significant reduction in its number and corresponding losses, modifing also the cooling air control system.
At the same time, the electronics of the original voltage regulator was replaced by
microprocessor technology, including additional functions (V/Hz limitation, phase
swingings stabiliser, etc).
The investment required was 100 Mpta.
3.3.3. Condenser and Feed Water Heaters
In the on line condenser cleaning system, several improvements ware made, with
include a continuous monitoring of the cleaness factor leading to take full advantage
of the automatic cleaning.
In the Feed Water Heaters (FWH), the existing partition plate leaks were eliminated
and the condensate level controls and drainage systems were tuned and improved,
ensuring its optimal performance. The required investment was 20 Mpta.
3.4
AUXILIARIES CONSUMPTION
3.4.1 Induced Draft Fans and Forced Draft Fans
See 3.2.1
3.4.2 Electrostatic Precipitators
In the electrostatic precipitators, a new control system based on microprogramed
remote units were installed, connected to two remote terminal units (RTU) and with
a central supervision station. The system has various control strategies according to
the fuel mix burned (domestic/imported coal and siderurgic gases) incorporating
several algorhythms: sparks and detection extinction strategy, automatic adjusting
of the electric consumption according to the gas opacity measurements, energy reduction to the precipitator during the collection plates hammering, and finally different secuences of the plates and electrodes hammering.
With this new control system, an important reduction of about 50% of the precipators electric consumption has been achieved, maintaining the efficiency.
The required investment was 55 Mpta.
3.4.3 Boiler Feed Pumps
In the boiler feed pumps-3 pumps of 55% of unitary capacity, hydraulic variable
speed gearing, -the speed torque convertor gears were replaced to get the maximum efficiency.
The original pump seals, floating ring type, wich require an important condensate
flow injection, were replaced by mechanical seals, with external cooling in a closed
circuit, and with negigeable leaks.
6-55
7
The total cost of the investment was 35 Mpta.
4. RESULTS AND DISCUSSION

4.1
GENERAL
The results show that the net efficiency made a relative improvement of 3.1% as
the NHR passed from 2,598 kcal/kWh to 2,520 kcal/kWh, which means a reduction
of 78 kcal/kWh.
Therefore, if the Unit is operating 7,500 equivalent hours at full load per year, with a
fuel cost of 1.20 pta/termia, (1 termia = 1,000 kcal), then the amount saved is 233
Mpta per year (7,500 x 360 x 0.9236 x 78 x 1.20 = 233) without taking in account
the improvement to the operation conditions flexibility, safety and a small load increase which was achieved, since was possible to operate the Unit with more excess air.
Besides the amount saved, another 112 Mpta per year can be added. This last
amount was obtained by being able to burn a greater percentage of BFG, which as
mentioned previously, goes from 20 to 30% in terms of energy, as the cost of the
BFG termia is 15% less than that of coal.
Consecuently, the total savings can be estimated at 345 Mpta per year, and the required investment was, as before mentioned, 1,500 Mpta, so the internal rate of return (IRR) of the investment over15 years is approximately 25%.
The assessment of the impact of the different measures taken, and the separation
of each result, shown in Table 1, was carried out in the following way:
-
The ASME test was used to measure the efficiency of the Boiler (separate
losses).
The complete heat balance of the Unit, performed simultaneosly with the boiler
efficiency test, was used for the gross heat rate determination. The different
contributions were separated by using the manufacturers estimations and the
plant available tests results.
The transformation efficiency and auxiliaries consumption was obtained directly

from the Wh meters.
4.2
BOILER
After the different actuations carried out, the boiler test shows that efficiency has
improved from 86.06% to 87.68%, that means a relative increase of 1.88%.
This 1.88% (47.5 Kcal/Kwh) is distributed in this way:
-
6-56
0.8% (20 Kcal/Kwh) due to the 15C reduction in the boiler exit gas temperature
because the economiser surface extension and the better control of the hot BFG
temperature.
8
-
0.6% (15 kcal/kwh) as a result of the improvement in the unburned fuel losses,
since it was possible: to increase the excess air due to the fans changes, to
burn more BFG with cero unburned, and less losses due the formation of H2O
from the fuel- as well as to improve the control of the fuel mix (PC + BFG +
COG + FO) through the new of Operation and Diagnosis Information System.
0.5% (12.5 kcal/kwh) as a result of taking full advantage of the soot blowing
program, the flows redistribution, the installation of guide baffles and others.
4.3
TURBOGENERATOR
The gross efficiency of the turbogenerator, calculated from the gross heat rate and
boiler efficiency, passes from 41.81% to 42.06% which means a relative increase of
0.6% (15 kcal/kwh). This improvement is distributed as follows:
-
Improvement due to shaft packing changes and strip seals repairs in HPT and
IPT, according to manufacturers estimation: 0.35% (9 Kcal/Kwh).
Improvement due to the excitation modifications: 0.05% (1 Kcal/Kwh), according

to the losses calculation.
Improvement due to the condenser cleaning and the feed water preheaters optimization: 0.20% (5 kcal/kwh).
4.4
TRANSFORMATION AND AUXILIARIES
The net energy at rated load now represents 92.05% of the gross energy generated, compared to 91.5% before the implementation of the related measures. The
reduction in the auxiliaries was 2 MW at full load, showing a relative improvement of
0.6% (15 kcal/kwh) due to the following savings:
-
IDF and FDF = 0.3% (7.5 kcal/kwh)
Electrostatic Precipitators = 0.05% (1 kcal/kwh)
BFP = 0.25% (6.5 kcal/kwh)
5. CONCLUSIONS
The results show that with reasonable investments it is possible to make older power stations more competitive, in the new spanish scenario, after performing a systematic study
considering the potential improvements in the heat rate and associated operating costs,
and the profitability of the necessary investments to attain it.
A specific study must be carried out for each Unit and the external conditions of each area
competitiveness, as well as the Unit internal conditions must be taken into consideration:
in this case we are referring to the Unit general condition, remanent and extended life, actual heat rate, specific operating problems, fuel costs, environmental retrictions and others.
The cut cost study should include other areas apart from the fore-mentioned heat reate,
such us operation and maintenance cost, possible avalability improvements and others.
6-57
9
In any case, it is almost obligatory to consider at the same time, the necessary modifications for the efficiency and availability improvements with those relating to life extension
and modernization of the Unit and its regulatory environmental adaptation.
REFERENCES:
6-58
rd
Jornadas Hispano-
(1)
Gonzlez Blas, J.: Future Electricity Generating Cost with Fossil Fuels, 3
Lusas de Ingeniera Elctrica, Barcelona, 1-3 Julio 1993.
(2)
Gonzlez Blas, J.: Operation and Diagnostic Operation System in Power Installations, 5 Jornadas
Hispano-Lusas de Ingeniera Elctrica, Salamanca, 3-5 Julio 1997.
(3)
Grhn, Michael; Tielsch, Hans-Peter: On Line Expert Diagnostic Siemens Power Journal. Enero
2000.
th
10
T ABLE 1
IMPACT OF T HE ACT IONS
Element
NHR improvements
UNIT
BOILER
Economizer and S IOD
IDF and FDF and Fuel
S oot-blowing and AH
%
3.1
1.88
0.8
0.6
0.5
Kcal/Kwh
78
47.5
20
15
12.5
T URBOGENERAT OR
MPT and IPT
0.6
0.35
15
9
Generator
Condens er and FWH
0.05
0.20
1
5
AUX ILIARIES
IDF and FDF
Electros tatic precipitator
0.6
0.3
0.05
15
7.5
1
BFP (drive and pump)
0.25
6.5
Initial Unit NHR

Final Unit NHR
--3.1
2,598
2,520
S teps performed
S ee below
S ee below
S urface extens ion and new ODIS
New fans and fuel mix control
S hoot blowing optimization baffles
ins tallation and others
S ee below
New s haft packings and s trip s eals
repairs
Modernis ation of the excitation
Cleaning s ys tem improvement and
controls optimization
S ee below
Fans replacement
Optimization of electric
cons umption
Variable s peed gearing and pump
s eals modifications
--S ee above
6-59
11
ABOO POWER PLANT

UNIT I Boiler Sectional View
HEAT RATE IMPROVEMENT IN AN EXISTING MULTIFUEL UNIT Fig.1
6-60
12
6-61
13
NEURAL NETWORKS
Neural Network topology used for the prediction

of the heating water flow for the BFG (C512).
Error evolution during the network training and

histrogram error between real and estimated values
HEAT RATE IMPROVEMENT IN AN EXISTING MULTIFUEL UNIT- Fig.3
6-62
AN APPLICATION OF THE PLANT PERFORMANCE MODELLING PACKAGE

PROATES TO ANALYSE THE CAUSE OF A PERSISTENT TUBE FAILURE
PROBLEM
Mr K R J Hartwell
Mr A B Ready
Powergen, Power Technology Centre
Ratcliffe-on-Soar
Nottingham, NG11 0EE
Nottinghamshire
UK
Tel: +44 115 936 2219/2232 Fax +44 115 936 2711
Email kevin.hartwell@powertech.co.uk
Abstract
TM
Powergens proprietary whole plant modelling software PROATES has been developed and
applied to solve a variety of plant performance problems, examples of which have been
TM
discussed in previous papers. This paper describes in detail an analysis, using PROATES , of a
particularly persistent tube failure problem.
During the course of a month, six tube leaks occurred on the same tube in a Generating Stations
boiler furnace and were incurring considerable costs and loss of generating income. The Station
had tried promoting extra flow in the problem tube by enlarging the tube orifice. Cold flow tests
had shown that this action had resulted in the tube taking 30% more flow than neighbouring tubes
but the failures still occurred. To gain a better understanding of the causes of the high tube
TM
temperatures which had led to failures, a detailed PROATES model was built of the heat
transfer/flow circuit and was used to demonstrate the likely reasons for the tube failure and also
remedial actions which could be taken. As a result of the analysis advice was given and
implemented and the tube failure problem was successfully overcome.
Introduction
Powergens Power Technology Centre was contacted by a Utility following a series of failures in
the furnace of one of their boilers. One particular tube failed six times within a period of one
month resulting in an estimated loss of revenue of $2 million. This tube was repeatedly replaced
until very little of the original tube remained; however overheating failures still occurred in the
replaced tubing. The Station could not detect any flow blockages and even increased the tube
orifice diameter to promote more flow, but still the tube failed. Power Technology Centre were
asked to undertake a theoretical study to identify the mechanisms leading to the observed high
tube temperatures which were resulting in tube failure.
Study Procedure
To build a model of the thermo-hydraulic system where tube failure was occurring using the
PROATES modelling package
To examine the recorded plant data to provide an insight into the failure problem
To validate the model against cold flow test data
To use the model to determine the sensitivity of the failed tube outlet temperature to
imbalances in tube heat absorption
To determine the effects of orifice size in reducing the sensitivity of tube outlet temperature to
imbalances in heat flux
To suggest ways of preventing future tube failures
6-63
Model Construction
The furnace circuit is shown in Figure 1a. Furnace corner tubes flow down from the top to the
bottom of the furnace. At the bottom of the furnace these tubes turn and form the up flowing
centre serpentine section panels of each furnace wall.
TM
TM
The modelling was performed using PROATES . PROATES , a commercial product available
from Power Technology, is a steady state plant modelling software package enabling a range of
modelling to be undertaken from whole boiler/turbine/feedheater models to more detailed models
of particular components of the plant. PROATES has been described in detail, together with
examples of its application, in previous EPRI heat rate papers (Refs 1,2,3).
A model of the circuit was constructed using the tube pressure drop module within the PROATES
modelling system.
This module includes the facility to model two-phase frictional pressure
losses in tubes or pipes with a given roughness length. In addition, to allow the modelling of
geometrical features of the tube, such as bends and orifices, a further loss term has been
incorporated. This Other geometric loss term is specified in terms of so-called K-values which
2
give a lumped parameter representation of resistance corresponding to the number of v
losses through the tube. The frictional flow conductance parameter is represented by a version of
the well-validated Colebrook-White equation. This equation closely approximates the curves on a
Moody chart in the turbulent region of flow. The module also allows a heat absorption input to be
specified, in this case used to represent the furnace heat input.
The furnace tube circuit, shown schematically in Figure 1b, can be considered as two basic
parallel paths. A path representing the long path of the front and rear wall tubes and a path
representing the shorter path of the side wall tubes. Each of the major paths additionally has an
associated parallel single tube path represented to enable studies of single tube abnormalities to
be undertaken. In this analysis only abnormalities associated with the short side wall tube circuit,
which contained the failing tube, were considered.
The model contained formulations allowing the pressure drop across the tube orifices, given the
orifice size and profile, to be represented. Heat flux profiles to the tubes were assumed to be the
same as those determined in a previous study and the total heat gain of the circuit was set to
agree with plant data.
Initial Data Examination

Figure 2 shows a typical plot of the tube outlet steam temperatures from the wall section
containing the failing tube. A study of this data and subsequent similar plots indicated the
following:
6-64
The high temperature of the failing tube followed the natural temperature fluctuations of its
neighbouring tubes indicating that the overheating had occurred due to increased heat
absorption rather than due to a flow instability
The failing tube was more sensitive to general wall heat absorption fluctuations than
neighbouring wall tubes
At the beginning of the week the failing tube temperature tended to be higher than at the end
of the week. This is thought to be due to slag spalling off at weekend shutdowns. As
operation continued during the week slag would build up again and heat absorption would be
reduced.
Model Analysis
Cold Flow Tests
Cold flow tests data was compared to the model predictions. The model predictions were close to
the measured data. The percentage flow to the front and rear walls measured and predicted was
57% and 57.6% respectively. The increase in flow to the failing tube resulting from the orifice
enlargement in that tube measured and predicted was 28.6% and 29.1% respectively. This
added confidence that the orifice sizes, tube path lengths, tube geometries, tube roughness used
in the model and the construction of the model were correct.
Static Stability
Figure 3 shows the pressure drop flow relationships for a typical furnace side wall tube with a
0.28in (7.1mm) orifice operating at different pressure and temperature boundary conditions. In
constructing the curves the heat input to the tubes has been assumed to be proportional to the
flow to the whole of the furnace section. The plot shows that at inlet conditions of 572F (300C) /
flow<2143 lb/h (0.27 kg/s) and 653F (345C) / flow<1746 lb/hr (0.22 kg/s) the pressure drop in
the furnace section goes negative. Under negative pressure drop conditions it is possible for a
tube to operate alternatively in a reversed flow mode, which will satisfy the negative pressure
differential. If a reversed flow mode is established in a tube then this can lead to high tube
temperatures and failure. To reduce the tendency for this condition to arise, operation above
loads of 30% MCR should be aimed for. Additionally the tendency for static instability is reduced
by maintaining the inlet temperatures as high as possible (>345C) whilst ensuring that the outlet
temperature does not exceed its design value of 860F (460C). However operation at high inlet
temperature can exacerbate other problems as discussed in the next Section.
Effects of Increased Heat to a Single Tube

Figure 4 shows the effect of increased heat to a single tube at different inlet temperature and
pressure conditions. The heat absorption factor is the ratio of heat absorbed by a single tube
compared to the rest of the tubes. The general heat distribution to the four furnace walls is
assumed to be even. As can be seen the higher the heat absorption factor the higher the tube
temperature, and as would be expected the higher the inlet temperature the greater will be the
outlet steam temperature. The effect of different operating pressures is small. Figure 5 shows
the additional effect of +-5% uneven heat distribution to the furnace walls, on a tube which has
increased heat absorption compared with its neighbours but which has the same orifice size. It
can be seen that outlet temperatures of 1290 F (700C) can be obtained when heat absorption is
increased to a single tube by 33% if at the same time there is an uneven heat distribution to the
wall containing the tube of greater than 5%.
Increased heat absorption by a single tube can arise for the following reasons.
If a tube becomes proud of its neighbours (due to repairs)

If its emissivity (absorptivity) is higher than its neighbours. This can occur if the tube is new
since the metal surface of these are quoted to have high emissivity (>0.8) compared with
tubes which have a slag coating which are quoted to have an emissivity around 0.7
If the new tube is relatively clean compared with long term fouling of old tubes.
Uneven distribution to the furnace walls can arise due to the furnace fireball not being central, or
if the walls are not equally fouled.
6-65
Effects of Increasing Orifice Size on the Failing Tube

Figure 6 compared with Figure 5 shows how increasing the orifice size from 0.28in (7.1) to
0.354in (9.0mm), on the tube with enhanced heat absorption, affects the sensitivity to increased
heat absorption. It is seen that increasing the orifice size has counteracted the effects of
increased heat absorption by promoting extra flow through the tube and has effectively reduced
the steam temperatures by approximately 230F (130C). Figures 5 and 6 can also be used to
explain the observed data. The data showed that sometimes the failing tube temperature is
higher, sometimes similar and sometimes lower compared with other tubes in the group. It has
been observed that the failing tube temperature tends to be higher than its neighbours at the
beginning of the week and lower at the end of the week with step changes in its temperature
relative to its neighbours sometimes occurring. It is postulated that at the beginning of the week
the tubes are cleaner than at the end of the week, due to the action of slag spalling when the unit
is off at weekends. This explains why the tube with increased heat absorption has a relatively
higher absorption at the beginning of the week. It is believed that the step changes in the failing
tube temperature relative to its neighbours are due to operational changes such as sootblowing
or firing which cause imbalances in heat distribution to the furnace walls.
Causes of Tube Failure

From the above simulations and comparisons with the observed data it is concluded that the
cause of the tube failures is due to the replacement tube initially absorbing more heat than
neighbouring tubes. This results in a greater frictional pressure drop contribution compared with
other tubes and hence a lower flow which together with the extra heat absorption results in high
tube temperatures and failure.
Measures to Overcome the Problems of Tube Failures

Orifice Size Alterations
The orifice size in the front and rear wall tubes is larger than in the side wall tubes. This is aimed
at compensating for the longer path length and heat transfer area of the front and rear wall tube.
The model indicates that this compensation is too high resulting in the average side wall tubes
operating approximately 45F (25C) hotter. This result is supported by limited plant data
obtained from a small number of tubes. Using the model it is predicted that reducing the orifice
sizes in the front and rear wall tubes from 0.409in (10.4mm) to 0.339in (8.6mm) would enable a
better balance of outlet temperatures. If this was undertaken it is predicted that the increase in
pressure loss would be (15psi) 1 bar at a load of 350MWe.
Figure 7 shows the effect of reduction of orifice size of all tubes on the outlet temperature of a
single tube receiving increased heat absorption, and Figure 8 shows the effect of reduced orifice
size on pressure loss. Figure 7 shows that there is some benefit to be obtained from reducing
orifice size, however this is at the expense of a significant increase in pressure loss. A reduction
in orifice size equivalent to a 290 psi (20 bar) increase in pressure loss would reduce the tube
outlet temperature by 140F (60C) at a heat absorption factor of 1.4 and 86F (30C) at a heat
absorption factor of 1.25. It is considered that the increase in pressure loss is not acceptable as
the benefits are relatively small.
Capping off the Failing Tube

This option was modelled. Provided the capped off tube does not corrode away it is not thought
that this will cause any problematic temperature rise effects in other tubes. There will be an
increased heat transfer to tubes either side of the capped tube due to the high operational
temperature of the capped tube. It is estimated that this is not likely to result in more than a 36F
(20C) increase in steam temperature in these tubes. However if the capped tube corrodes away
6-66
with time then more and more surface area of the adjacent tubes will be exposed to furnace
radiation. This will result in an increased heat gain by these tubes and steam temperatures could
increase significantly. For example a 20% increase in heat absorption would result in a steam
temperature increase of 153F (85C).
Cap off the Up Flowing Section of the Failing Tube and Merge the Exit of the Down
Flowing Section into the Exit of an Adjacent Down Flowing Tube
This is definitely NOT recommended since as shown by Figure 9, at flows below 3300 lb/h
(0.42kg/s) (equivalent to loads below 44% MCR) there is both a downward flow and an upward
flow which can satisfy the pressure drop between top and bottom headers i.e. static instabilities
can occur. This could result in one of the joined downward tubes operating with a reversed
(upward) flow causing overheating and failure. In addition the possible problems resulting from
the hot dead tube as described above would also be present.
Reducing Imbalances in Tube to Tube Heat Absorption

The following were suggested as ways to reduce the imbalances in tube to tube heat transfer
Ensuring all the tubes are in line so that the area of each tube exposed to radiation is the
same. Note that the practice of replacing tubes by the use of small rods as a means to attach
the replacement tube to neighbouring tubes will effectively increase the area of the replaced
tubes.
Ensure that the levels of fouling on the four walls of the combustion chamber are similar. If a
wall has a tube with a high temperature then let it foul up and this should reduce the
temperature.
Try to keep the furnace firing pattern balanced to achieve a central fireball and an even
distribution of incident radiation to the tubes.
When a failed upflowing serpentine tube is replaced by a new clean tube then it would be
beneficial to try and condition this tube to ensure its heat absorption was not higher than the
other tubes. A thin coating designed to reduce the surface emissivity might be applied to the
tube. For instance the emissivity of oxidised aluminium is 0.2-0.3 compared with that of shiny
oxidised steel of 0.8, and the emissivity of aluminium paints range between 0.3 and 0.7.
Such a low emissivity coating, if durable, will reduce the tube heat absorption whilst the tube
becomes conditioned by slagging during the normal course of operation.
Concluding Remarks
The PROATES model was able to explain the reasons for tube failure. Subsequently newly
repaired tubes have been given a coat of aluminium paint which has been effective in avoiding
increased heat absorption in the initial period before the tube surface has become naturally
conditioned during operation. This measure has been successful in preventing failures due to
overheating.
References
1. Green C.H., Ready A.B., Rea J., The Powergen Experience in off-line and on-line modelling
of power plant for efficiency improvement , EPRI Heat Rate Improvement Conference,
Birmingham, 1992.
6-67
2. Green C.H., Ready A.B., Rea J., PROATES. A computer modelling system for power plant:
Its description and application to heat rate improvement within Powergen, EPRI Heat Rate
Improvement Conference, Baltimore, 1994.
3. Green C.H., Ready A.B., Chew P.E., Hartwell K.R.J., Application of the Whole plant
Modelling System package PROATES to improve power plant performance, EPRI Heat Rate
Improvement Conference, Baltimore, 1998.
6-68
Figure 1a Boiler tube geometry
Figure 1b - Schematic representation
6-69
Figure 2 Typical plot of Furnace tube outlet temperatures

Note Tube 13 is the failing tube with an orifice enlarged from 0.28in (7.1mm) to 0.354in (9 mm)
6-70
Figure 3 Expected flow characteristics of a typical side wall tube with a 0.28in (7.1mm)
orifice
25
326.5
20
226.5
15
10
126.5
26.5
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
-5
-73.5
Single tube flow - kg/s

220 bar 345C inlet T
220bar 300C inlet T
180 bars 345 C inlet T
180 bar 300 C inlet T
Figure 4 Effect of increased heat absorption to a single tube with even furnace heat
distribution to furnace walls
800
1460
750
700
1260
650
600
1060
550
500
860
450
400
350
660
1
1.1
1.2
1.3
1.4
1.5
1.6
Single tube heat absorption factor

6-71
Figure 5 Effect of increased heat to a single tube with an orifice size on 0.28in (7.1mm)
and the additional effects of uneven furnace wall heat distribution
900
850
1550
800
750
1350
700
650
1150
600
550
950
500
450
400
750
1
1.1
1.2
1.3
1.4
1.5
1.6
Heat absorption factor

even furnace heat distribution
5% less heat to wall with problem tube
5% more heat to wall with problem tube
Figure 6 Effect of increased heat to a single tube with an orifice size of 0.354in
(9.0mm) and the additional effects of uneven furnace wall heat distribution
750
700
1260
650
600
1060
550
500
860
450
400
350
660
1
1.1
1.2
1.3
1.4
1.5
1.6
Heat absorption factor

Even furnace distribution
6-72
5% less heat to wall with problem tube
5% more heat to wall with problem tube
Figure 7- Effect of side wall and front/rear wall orifice sizes on desensitising the
effects of increased heat transfer to a single tube
750
1350
700
650
1150
600
550
950
500
450
400
750
1
1.1
1.2
1.3
1.4
1.5
1.6
Single tube absorption factor

7.1 and 8.6 mm orifices
Figure 8 - Effect of side wall and front/rear wall orifice size on pressure loss
50
700
45
600
40
35
500
30
400
25
300
20
15
200
10
100
5
0
0
40
50
60
70
80
90
100
Boiler Load % MCR

orifice 7.1 and 8.6
orifice 6.0 and 6.7
orifice 5.0 and 5.4
orifice 7.1 and 10.43
6-73
Figure 9 Pressure drop flow stability curve of a down flowing tube which has been
joined at exit to an adjacent tube
4
45
2
-0.8
positive flow indicates downflow

negative flow indicates upflow
-0.6
-0.4
0
-0.2
0.2
0.4
0.6
0.8
-5
-2
-55
-4
-6
-105
-8
-10
-155
-12
-14
-205
Single tube flow kg/s

Downflow
6-74
Upflow
In-Situ O2 Probe Failure at

Dairylands JP Madgett Station
Duane Hill
Manager, Performance Administration
Dairyland Power Cooperative
6-75
Description of Problem
In November 1999, Operators report the
following:
having trouble making full load at JPM; reaching
maximum ID Fan amps approximately 25 MW
lower than full load
indicated air flow appears to be high
NOx levels rising
6-76
Initial Test Results

Instrument techs conduct calibrations of O2
probes; all probes calibrate with no
corrections
Performance department conducts ID fan
tests; both fans performing on curve; high
flow rate through fans
Performance department conducts air heater
tests; high DPs on both air heaters and
elevated leakage rates
6-77
A PH G a s DP (I n W C )
1 3.5
13
1 2.5
12
1 1.5
11
1 0.5
10
198 6
1988
1 990
19 92
199 4
1996
1 998
20 00
Ye a r
Av g AP H G as D P
6-78
Avg AP H G as O ut Pres s
200 2
W C)
-15
-15.5
-16
-16.5
-17
-17.5
-18
-18.5
-19
-19.5
-20
-20.5
-21
-21.5
-22
14
A PH G A s O u t P re s s (In
JPM APH G a s D P & P re ss At U R G E
JPM NOx Jan99-Apr00

0.5200
0.4800
0.4600
320-330
0.4400
0.4200
Apr
Mar
Feb
Jan
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
0.4000
Jan
lb NOx/MBTU
0.5000
6-79
Initial Plant Response

Plant schedules an outage to clean air heaters
and look for sources of air leakage
In preparation for outage, a complete leakage
test of boiler gas cycle is performed
6-80
JPM Air Heater Test Results 2/00

Econ.
Out
%CO2
15.28
61 61 AH 62 AH 62 AH
AH
Out
In
Out
In
13.62 12.88 13.55 13.14
%O2
1.55
3.65
4.6
3.75
4.26
5.8
%
Leakage
------
11.89
5.70
12.52
3.00
8.88
Stack
11.93
6-81
JPM Air Heater Test Results 4/00
6-82
Furnace
Exit
Econ.
Out
61
AH In
61
AH
Out
%CO2
17.55
18.96
16.93 15.42 17.11 15.84 12.65
%O2
3.07
1.63
3.70
5.24
3.52
4.81
8.05
%
Leakage
------
------
11.25
9.25
10.17
7.57
32.58
62 AH
In
62
AH
Out
Stack
Instrument Evaluation
Economizer outlet O2 was measured with insitu instrumentation
When using test instrumentation, economizer
outlet O2 and furnace O2 were the same
Compared all analyzers with calibration gas;
all responded correctly
6-83
Which O2 is Right?
Compared test instrument with new O2 probes on
another boiler, and test instrumentation read what
the new O2 probes read
Test instrumentation appeared correct, but what
caused in-situ probes to read low?
Further analysis of O2 probes showed 4 of 6
reading low, one reading correct and one giving
poor reading (out of average)
6-84
Why?
Contacted vendor to find out what would
cause O2 probes to read 1.5% low?
First guess from vendor was combustibles
burning on probe; assumed CO; CO measured
less than 100 ppm
Second guess from vendor was plugged
diffuser; will cause unit to read low but will
calibrate
6-85
Plugged Diffuser test

Vary calibration gas flow rate up and down.
Under normal conditions, resultant O2 will
vary no more than .2%. If variance more than
.2%, diffuser is plugged
JPMs O2 probes varied 1-2.5%
Visual inspection not good enough
6-86
Final Results
Brought unit up on load with test
instrumentation and had no problems making
full load; changed diffusers, in-situ probes
working properly
Diffuser pluggage test added to monthly
calibration checks
Still had problems with air heater DPs, but no
a load limiting problem
6-87
Target:
Heat Rate and Cost Optimization
About EPRI
EPRI creates science and technology solutions for
the global energy and energy services industry. U.S.
electric utilities established the Electric Power
Research Institute in 1973 as a nonprofit research
consortium for the benefit of utility members, their
customers, and society. Now known simply as EPRI,
the company provides a wide range of innovative
products and services to more than 1000 energyrelated organizations in 40 countries. EPRIs
multidisciplinary team of scientists and engineers
draws on a worldwide network of technical and
business expertise to help solve todays toughest
energy and environmental problems.
EPRI. Electrify the World
2001 Electric Power Research Institute (EPRI), Inc. All rights

reserved. Electric Power Research Institute and EPRI are registered
service marks of the Electric Power Research Institute, Inc.
EPRI. ELECTRIFY THE WORLD is a service mark of the Electric
Power Research Institute, Inc.
Printed on recycled paper in the United States of America
1001328
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

12th Heat Rate Performance Improvement Conference

Uploaded by

Copyright:

Available Formats

12th Heat Rate Performance Improvement Conference

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

12th Heat Rate Performance Improvement Conference

Uploaded by

Copyright:

Available Formats

EPRI's Twelfth Heat Rate Improvement

EPRIs Twelfth Heat Rate

Proceedings, January 2001

EPRI Project Manager

DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

How are utilities using heat rate information?

What improvements in heat rate information are needed?

The effectiveness and usefulness of On-Line Heat Rate Monitors

The potential for incorporating Intelligent Sootblowing applications into optimization

The latest trends in Heat Rate Testing

Actual Plant Experiences with heat rate improvement projects

Registration and Continental Breakfast

Session 1: On-Line Heat Rate Monitors

Real-Time Performance Monitoring of Coal-Fired Units

F-Factor Method for Heat Rate Measurement and its Characteristics

The Input/Loss Method

Recent Experiences with Performance Monitoring at TVA using EPRIs

Modular Heat Rate Calculations for Power Plants using OPC

Improving Results, Confidence and Decisions with Precise Data Validation

Panel: How are utilities using heat rate information?

Impacts of Combustion Optimization on Power Plant Heat Rate

Obtaining Improved Boiler Efficiency and NOx using Advanced Empirical

Heat Rate Improvement at Dairylands Madgett Station using NeuSIGHT

Duane Hill, Dairyland Power Cooperative

Automatically Control NOx with Heat Rate Constraints, in a Coal-Fired

Unit Optimization at Hammond Unit 4

Reception and Exhibits

Wednesday, January 31, 2001

Session 3: Intelligent Sootblowing

Effects of Sootblowing in Coal-Fired Boilers on Unit Heat Rate and Nox

Optimization of Boiler Sootblower Operation

Intelligent Sootblowing Application Development

Intelligent Sootblowing Boiler Cleaning Management System

Session 4: Turbines and Auxiliaries

In-Situ Feedwater Flow Measurement

In-Situ Enthalpy Measurements in Low Pressure Condensing Steam

Steam Turbine Related Reseach at TVA

On-Line Performance Monitoring and Condition Assessment of Steam

Session 5: Heat Rate Testing

Performance Evaluation and Testing

A Procedure for Analyzing Power Plant Measurement Variances Associated

Cycle Alignment Methods and Evaluation

Panel: Why test?

Thursday, February 1, 2001

Panel: What improvements in heat rate information are needed?

Session 6: Plant Experiences

Analysis of Variables for Predicting Power Output at the Columbia Power

Experience of Moneypoint Power Station in Recovering Plant Heat Rate:

Heat Rate Improvement in an Existing Multifuel Unit

An Application of the Plant Performance Modelling Package PROATES to

In-Situ O2 Probe Failure at Dairylands JP Madgett Station

1 SESSION 1: ON-LINE HEAT RATE MONITORS ................................................................ 1-1

4 SESSION 4: TURBINES AND AUXILIARIES...................................................................... 4-1

An Application of the Plant Performance Modelling Package PROATES to Analyse

Session 1: On-Line Heat Rate Monitors

Session 1: On-Line Heat Rate Monitors

Session 1: On-Line Heat Rate Monitors

Session 1: On-Line Heat Rate Monitors

Session 1: On-Line Heat Rate Monitors

Session 1: On-Line Heat Rate Monitors