Remaining Life Assessment of Austenitic

Stainless Steel Superheater and Reheater

Tubes

Technical Report

13622634
13622634
Remaining Life Assessment of
Austenitic Stainless Steel
Superheater and Reheater Tubes

1004517

Final Report, December 2002

EPRI Project Managers

D. Gandy
B. Dooley

EPRI • 3412 Hillview Avenue, Palo Alto, California 94304 • PO Box 10412, Palo Alto, California 94303 • USA
800.313.3774 • 650.855.2121 • [email protected] • www.epri.com

13622634
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

EPRI

Aptech Engineering Services, Inc.

ORDERING INFORMATION
Requests for copies of this report should be directed to EPRI Orders and Conferences, 1355 Willow
Way, Suite 278, Concord, CA 94520, (800) 313-3774, press 2 or internally x5379, (925) 609-9169,
(925) 609-1310 (fax).

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 © 2002 Electric Power Research Institute, Inc. All rights reserved.

13622634
CITATIONS

This report was prepared by

Aptech Engineering Services, Inc.

1253 Reamwood Avenue
Sunnyvale, CA 94089

Principal Investigators
S. Paterson
R. Moser
T. Kuntz

This report describes research sponsored by EPRI.

The report is a corporate document that should be cited in the literature in the following manner:

Remaining Life Assessment of Austenitic Stainless Steel Superheater and Reheater Tubes, EPRI,
Palo Alto, CA: 2002. 1004517.

iii
13622634
13622634
REPORT SUMMARY

Austenitic stainless steel superheater and reheater tubing can fail in power plant operating
conditions by a variety of mechanisms. A comprehensive guide to assess these different
mechanisms and to provide a methodology for predicting future failures can assist in reducing
tubing failures.

Background
In recent years, in-service failures of austenitic stainless steel superheater and reheater tubes have
started to become a significant cause of lost plant availability. Steam-side oxide scale based
remaining life methods developed for ferritic tubing are commonly being used by utilities to
estimate austenitic tube life. Unfortunately, this approach does not provide a direct method for
estimating the life of austenitic stainless steel tubing.

Objectives
• To provide a compilation of current information and case histories associated with the
prediction of remaining life in austenitic stainless steel superheater and reheater tubes
• To document degradation and failure mechanisms experienced by austenitic stainless steel
superheater and reheater tubes
• To provide roadmap approaches for assessing the remaining life of austenitic stainless steel
superheater and reheater tubes
• To provide metallurgical methodologies and documentation for use in remaining life
calculations

Approach
The report begins with an overview of the basic design aspects and metallurgical characteristics
of new austenitic stainless steel (SS) tubes. Next, a description of microstructural degradation
modes commonly associated with the austenitic stainless steels is provided. Individual sections
of the report address primary damage, including: 1) fireside corrosion in both coal-fired and oil-
fired units, 2) pitting and crevice corrosion, 3) intergranular attack and stress corrosion cracking,
4) fatigue, 5) steam-side scale exfoliation, steam flow blockage, and short term overheating, 6)
dissimilar metal weld failures, and 7) long-term overheat creep. Remaining life models and
metallurgical tools to assess tube life were developed and tested.

v
13622634
Results
A thorough review of all known failure modes affecting austenitic stainless steel tubing is
provided to familiarize the reader with specific steps to: 1) identify the damage, 2) assess the
failure mechanism, 3) determine possible root causes, 4) determine the extent of the damage, 5)
determine and implement repairs, intermediate solutions and actions, 6) perform life assessment,
and 7) determine and implement long-term actions to prevent repeat failures. Detailed
methodologies have also been provided to assess the condition of austenitic SS tubing via
metallurgical evaluation. Finally, two remaining life models are presented to address long-term
overheat-creep failures.

EPRI Perspective
Successful remaining life methods for ferritic tubing have been available for more than a decade.
The increase in the number of failures of austenitic stainless steel tubing has promoted the need
for similar remaining life approaches. This report provides two comprehensive models to address
long-term overheat-creep failures. The first model is a cumulative damage model that provides
sufficient flexibility to: 1) predict the remaining life of the first and subsequent failures, 2)
perform sensitivity studies to determine the benefit of operational changes, repairs, retrofits, or
design alterations, and 3) provide a direct assessment of tube temperature or wall thinning rate
variations. The second model, referred to as a rank statistical model, is relevant to the situation
where failures have already occurred and an estimate of the rate of future failures is desired.
Both techniques may be applied by utilities to predict remaining life of austenitic stainless steel
tubing under long-term overheat-creep conditions.

Keywords
Austenitic stainless steels
Tubing
Remaining life assessment

vi
13622634
ABSTRACT

One of the increasingly significant causes of lost plant availability is in-service failures of
austenitic stainless steel superheater and reheater tubes. To address these failures, comprehensive
assessment techniques are provided to evaluate tubing damage mechanisms, determine the root
cause, implement remedial repairs/actions, and implement long-term actions to prevent future
failures. Many utilities are presently using steam-side oxide-scale remaining life methods
developed specifically for ferritic tubing to address austenitic tubing. This often leads to gross
inaccuracies in predicted life. In this project, two remaining life methods have been developed to
assist in predicting long-term overheat-creep failures. A flexible cumulative life model assists in:
1) predicting the remaining life of first and subsequent failures, 2) performing sensitivity studies
to determine the benefit of operational changes, repairs, retrofits, or design alterations, and 3)
providing direct assessment of tube temperature or wall thinning variations. The second model, a
rank statistical model, is applicable to situations where one or more failures have already
occurred and an estimate of the rate of future failures is desired. Additionally, comprehensive
metallurgical assessment techniques are presented.

vii
13622634
13622634
ACKNOWLEDGMENTS

Much of the text and many of the figures presented in this report were extracted directly or
adapted from material previously presented in the EPRI Boiler Tube Failures: Theory and
2
Practice volumes . Dr. Barry Dooley and Mr. Warren McNaughton are acknowledged for
allowing the use of their material in this report. The authors also wish to thank Clayton Lee and
Michael Cronin for their contributions to this report.

ix
13622634
13622634
CONTENTS

1 INTRODUCTION ....................................................................................................................1-1
1.1 Background ....................................................................................................................1-1
1.2 Objectives ......................................................................................................................1-3
1.3 Organization of This Report and How To Use It.............................................................1-3
1.4 References .....................................................................................................................1-4

2 STAINLESS STEEL SUPERHEATER AND REHEATER TUBING – DESIGN BASIS .........2-1

2.1 Common Alloys and Tubing Sizes .................................................................................2-1
2.2 Design-Basis Service Pressures and Temperatures......................................................2-1
2.3 Anticipated Design-Basis Life.........................................................................................2-6
2.4 References .....................................................................................................................2-9

3 IN-SERVICE DEGRADATION................................................................................................3-1
3.1 Introduction ....................................................................................................................3-1
3.2 Use of This Report for Identifying Specific In-Service Degradation Modes....................3-1

4 FIRESIDE CORROSION IN COAL FIRED UNITS .................................................................4-1

4.1 Introduction ....................................................................................................................4-1
4.2 Features of Failure and Typical Locations .....................................................................4-1
4.2.1 Features of Failure .................................................................................................4-1
4.2.2 Locations of Failure ................................................................................................4-5
4.3 Mechanism of Failure .....................................................................................................4-6
4.3.1 Mechanism of Attack by Alkali Salts.......................................................................4-6
4.3.2 Sulfidation of Stainless Steels ................................................................................4-8
4.3.3 Coal Composition and Corrosivity Indices..............................................................4-8
4.3.4 Effect of Chlorine in Coal......................................................................................4-12
4.4 Possible Root Causes and Actions to Confirm.............................................................4-13
4.4.1 Introduction...........................................................................................................4-13

xi
13622634
 4.4.2 Influence of Overheating of Tubes .......................................................................4-17
4.4.2.1 Poor Initial Design - Choice of Material.........................................................4-17
4.4.2.2 Poor Initial Design - Extra Gas-Touched Length ..........................................4-17
4.4.2.3 Internal Oxide Growth Which Occurs During Operation ...............................4-17
4.4.2.4 High Temperature Laning .............................................................................4-18
4.4.2.5 Tube Misalignment (Out of the Bank) ...........................................................4-18
4.4.2.6 Operational Problems When Coal Type Is Changed ....................................4-18
4.4.2.7 Rapid Startups ..............................................................................................4-18
4.4.3 Root Causes Related to Fuel Factors ..................................................................4-18
4.4.3.1 Change of Fuel to a Coal with an Unusually Corrosive Ash, Particularly
Those with High Content of S, Na, K or Cl.................................................................4-18
4.4.4 Root Causes Related to Incomplete or Delayed Combustion ...............................4-19
4.4.4.1 Actions to Confirm..........................................................................................4-19
4.5 Determining the Extent of Damage ..............................................................................4-19
4.6 Repairs, Immediate Solutions and Actions...................................................................4-19
4.6.1 Need for Remaining Life Assessment ..................................................................4-19
4.6.2 Repairs .................................................................................................................4-20
4.7 Life Assessment ...........................................................................................................4-20
4.8 Long term Actions and Prevention of Repeat Failures.................................................4-24
4.8.1 Remaining Life Assessment (Option 1, Figure 4-12)............................................4-25
4.8.2 Shielding (Option 2)..............................................................................................4-26
4.8.3 Coatings (Option 3) ..............................................................................................4-26
4.8.4 Replacement with Same Material, Same Thickness (Option 4) ...........................4-27
4.8.5 Change to a More Resistant Material, "Monolithic" (Option 5) .............................4-27
4.8.6 Change to a More Resistant Material, Composite (Such as Co-Extruded)
Tubing (Option 6)............................................................................................................4-29
4.8.7 Fuel Change, Blending, Washing (Option 7) ........................................................4-30
4.8.8 Limit Tube Temperatures (Option 8) ....................................................................4-30
4.8.9 Limit Steam Temperature (Option 9) ....................................................................4-30
4.8.10 Redistribute Steam Flow (Option 10) .................................................................4-31
4.8.11 Chemical Cleaning to Remove Steamside Oxide Scale (Option 11)..................4-31
4.8.12 Minimize Laning of Gases (Option 12) ...............................................................4-32
4.8.13 Correct Misalignment of Tubing (Option 13).......................................................4-32
4.8.14 Change Frequency, Check Effectiveness of Sootblowing (Option 14)...............4-32
4.8.15 Lowering Excess Air (Option 15) ........................................................................4-32

xii
13622634
 4.9 References ...................................................................................................................4-32

5 FIRESIDE CORROSION IN OIL FIRED UNITS .....................................................................5-1

5.1 Introduction ....................................................................................................................5-1
5.2 Features of Failure and Typical Locations .....................................................................5-1
5.2.1 Features of Failure .................................................................................................5-1
5.2.2 Locations of Failure ................................................................................................5-7
5.3 Mechanism of Failure .....................................................................................................5-9
5.3.1 Introduction.............................................................................................................5-9
5.3.2 Effect of Fuel Composition on the Rate of Corrosion .............................................5-9
5.3.2.2 Low Vanadium, Low Sodium Residuals........................................................5-11
5.3.3 Effect of Tube Temperatures on the Rate of Corrosion........................................5-11
5.3.4 Effect of Tube Material Composition ....................................................................5-11
5.3.5 Effect of Unit Operation on Corrosion Rates ........................................................5-12
5.3.6 Effect of Additives on Propensity for Corrosion ....................................................5-12
5.4 Possible Root Causes and Actions to Confirm.............................................................5-14
5.4.1 Introduction...........................................................................................................5-14
5.4.2 Influence of Oil Composition.................................................................................5-17
5.4.3 Influence of Tube Metal Temperatures.................................................................5-17
5.4.3.1 Excessive Temperatures Caused by Steamside Oxide Buildup...................5-17
5.4.3.2 Excessive Temperatures Caused by Operating Conditions .........................5-17
5.4.4 Influence of Operating Factors .............................................................................5-18
5.4.4.1 High Levels of Excess Oxygen and/or Periods of Very Low Excess
Oxygen ......................................................................................................................5-18
5.4.4.2 Poor Sootblowing Operations .......................................................................5-18
5.5 Determining the Extent of Damage ..............................................................................5-18
5.6 Repairs, Immediate Solutions and Actions...................................................................5-19
5.6.1 Need for Remaining Life Assessment ..................................................................5-19
5.6.2 Repairs .................................................................................................................5-19
5.6.3 Use of Mg-Based Additives ..................................................................................5-19
5.6.4 Other Short-term Options .....................................................................................5-19
5.7 Life Assessment ...........................................................................................................5-20
5.8 Longterm Actions and Prevention of Repeat Failures..................................................5-20
5.8.1 Remaining Life Assessment (Option 1, Figure 5-7)..............................................5-21
5.8.2 Shielding (Option 2)..............................................................................................5-21

xiii
13622634
 5.8.3 Coatings (Option 3) ..............................................................................................5-21
5.8.4 Replacement with Same Material, Same Thickness (Option 4) ...........................5-22
5.8.5 Change to a More Resistant Material, "Monolithic" (Option 5) .............................5-22
5.8.6 Change to a More Resistant Material, Composite such as Co-Extruded
Tubing (Option 6)............................................................................................................5-22
5.8.7 Monitoring of Fuel Changes (Option 7) ................................................................5-23
5.8.8 Use of Mg-Based Additives (Option 8) .................................................................5-23
5.8.9 Limit Tube Temperatures (Option 9) ....................................................................5-25
5.8.10 Limit Steam Temperature (Option 10) ................................................................5-26
5.8.11 Redistribute Steam Flow (Option 11) .................................................................5-26
5.8.12 Chemical Cleaning to Remove Steamside Oxide Scale (Option 12)..................5-26
5.8.13 Correct Misalignment of Tubing (Option 13).......................................................5-26
5.8.14 Change Frequency and Check Effectiveness of Sootblowing (Option 14) .........5-26
5.8.15 Lowering Excess Air (Option 15) ........................................................................5-27
5.9 Case Study...................................................................................................................5-27
5.10 References .................................................................................................................5-28

6 PITTING AND CREVICE CORROSION .................................................................................6-1

6.1 Introduction ....................................................................................................................6-1
6.2 Features of Failure and Typical Locations .....................................................................6-1
6.2.1 Features of Failure .................................................................................................6-1
6.2.2 Locations of Failure ................................................................................................6-2
6.3 Mechanism of Failure .....................................................................................................6-2
6.3.1 Development of Pitting Corrosion...........................................................................6-2
6.3.2 Development of Pitting from Oxygen-Saturated, Stagnant Condensate ................6-4
6.4 Possible Root Causes and Actions to Confirm...............................................................6-4
6.5 Determining the Extent of Damage ................................................................................6-5
6.6 Repairs, Immediate Solutions and Actions.....................................................................6-5
6.7 Life Assessment .............................................................................................................6-5
6.8 Longterm Actions and Prevention of Repeat Failures..................................................6-10
6.8.1 Instituting Proper Unit Shutdown and Layup Procedures.....................................6-10
6.8.2 Steam Chemistry Monitoring ................................................................................6-10
6.9 References ...................................................................................................................6-10

7 INTERGRANULAR ATTACK AND STRESS CORROSION CRACKING .............................7-1

xiv
13622634
 7.1 Introduction ....................................................................................................................7-1
7.2 Features of Failure and Typical Locations .....................................................................7-1
7.2.1 Features of Failure .................................................................................................7-1
7.2.2 Locations of Failure ................................................................................................7-3
7.3 Mechanism of Failure .....................................................................................................7-3
7.3.1 Basic Overview of the Mechanism of IGA ..............................................................7-3
7.3.2 Basic Overview of the Mechanism of SCC.............................................................7-5
7.3.3 Boiler Tubing Materials and IGA/SCC ....................................................................7-8
7.3.4 Sources of Corrosive Environments in Boiler Tubes That Can Lead to
IGA/SCC...........................................................................................................................7-9
7.3.4.1 Steamside Sources.........................................................................................7-9
7.3.4.2 Fireside Sources .............................................................................................7-9
7.4 Possible Root Causes and Actions to Confirm.............................................................7-10
7.4.1 Introduction...........................................................................................................7-10
7.4.2 Influence of Environment......................................................................................7-11
7.4.3 Influence of Stresses ............................................................................................7-12
7.4.4 Influence of a Susceptible Material.......................................................................7-12
7.5 Determine the Extent of Damage .................................................................................7-13
7.6 Repairs, Immediate Solutions and Actions...................................................................7-13
7.7 Life Assessment ...........................................................................................................7-13
7.8 Longterm Actions and Prevention of Repeat Failures..................................................7-13
7.8.1 Control of Environment.........................................................................................7-14
7.8.2 Control of Stresses ...............................................................................................7-14
7.8.3 Control of Sensitized Material...............................................................................7-14
7.9 Case Studies ................................................................................................................7-15
7.9.1 Case Study I - IGSCC ..........................................................................................7-15
7.9.2 Case Study II - IGA...............................................................................................7-18
7.10 References .................................................................................................................7-20

8 FATIGUE ................................................................................................................................8-1
8.1 Introduction ....................................................................................................................8-1
8.2 Features of Failure and Typical Locations .....................................................................8-1
8.2.1 Features of Failure .................................................................................................8-1
8.2.2 Typical Locations....................................................................................................8-2
8.3 Mechanism of Failure .....................................................................................................8-5

xv
13622634
 8.4 Possible Root Causes and Actions to Confirm...............................................................8-5
8.4.1 Introduction.............................................................................................................8-5
8.4.2 Excessive Strains Caused by Constraint of Thermal Expansion............................8-7
8.4.3 Poor Design and/or Manufacture Giving Rise to Excessive Mechanical
Stresses............................................................................................................................8-7
8.4.4 Vibration (Flue Gas-Induced) by Direct Flow or Vortex Shedding ..........................8-7
8.4.5 Poor Welding, Particularly Poor Geometry of Final Joint .......................................8-7
8.5 Determining the Extent of Damage ................................................................................8-8
8.6 Repairs, Immediate Solutions and Actions.....................................................................8-8
8.7 Life Assessment .............................................................................................................8-8
8.8 Longterm Actions and Prevention of Repeat Failures....................................................8-9
8.9 Case Study...................................................................................................................8-16
8.10 References .................................................................................................................8-17

9 STEAMSIDE SCALE EXFOLIATION, STEAM FLOW BLOCKAGE, AND SHORT

TERM OVERHEATING .............................................................................................................9-1
9.1 Introduction ....................................................................................................................9-1
9.2 Features of Failure and Typical Locations .....................................................................9-1
9.2.1 Features of Failure .................................................................................................9-1
9.2.2 Locations of Failure ................................................................................................9-2
9.3 Mechanism of Failure .....................................................................................................9-2
9.4 Possible Root Causes and Actions to Confirm...............................................................9-6
9.5 Determining the Extent of Damage ................................................................................9-8
9.6 Repairs, Immediate Solutions and Actions.....................................................................9-8
9.7 Life Assessment .............................................................................................................9-8
9.8 Longterm Actions and Prevention of Repeat Failures....................................................9-9
9.9 Case Study...................................................................................................................9-10
9.10 References .................................................................................................................9-10

10 DISSIMILAR METAL WELD FAILURES ...........................................................................10-1

10.1 Introduction.................................................................................................................10-1
10.2 Features of Failure and Typical Locations .................................................................10-1
10.2.1 Features of Failure .............................................................................................10-1
10.2.2 Locations of Failure ............................................................................................10-5
10.3 Mechanism of Failure .................................................................................................10-6
10.3.1 Introduction.........................................................................................................10-6

xvi
13622634
 10.3.2 Microstructural Changes in DMW During Service ..............................................10-7
10.3.3 Influence of Welding Variables ...........................................................................10-8
10.3.4 Influence of Stress and Temperature ...............................................................10-11
10.4 Possible Root Causes and Actions to Confirm.........................................................10-12
10.4.1 Introduction.......................................................................................................10-12
10.4.2 Excessive Tube Stresses Such as Caused by Improper Initial Design of
Tube Supports, or Supports That Have Stopped Functioning in the Designed
Manner .........................................................................................................................10-14
10.4.3 Excessive Local Tube Temperatures ...............................................................10-15
10.4.4 Changes in Unit Operation ...............................................................................10-15
10.4.5 Initial Fabrication Defects .................................................................................10-15
10.5 Determining the Extent of Damage ..........................................................................10-16
10.6 Repairs, Immediate Solutions and Actions...............................................................10-18
10.7 Life Assessment .......................................................................................................10-19
10.8 Longterm Actions and Prevention of Repeat Failures..............................................10-21
10.9 Case Study...............................................................................................................10-22
10.10 References .............................................................................................................10-23

11 LONGTERM OVERHEAT CREEP .....................................................................................11-1

11.1 Introduction.................................................................................................................11-1
11.2 Features of Failure and Typical Locations .................................................................11-1
11.2.1 Features of Failure .............................................................................................11-1
11.2.2 Locations of Failure ............................................................................................11-5
11.3 Mechanism of Failure .................................................................................................11-7
11.3.1 The Accumulation of Creep Damage in Superheater/Reheater Tubes ..............11-8
11.3.2 Steamside Processes.........................................................................................11-8
11.3.3 Fireside Effects...................................................................................................11-9
11.4 Possible Root Causes and Actions to Confirm.........................................................11-10
11.4.1 Influences of Initial Design and/or Material Choice ..........................................11-14
11.4.2 Buildup of Internal Oxide Scale ........................................................................11-15
11.4.3 Overheating Because of Restricted Steam Flow ..............................................11-15
11.4.4 Operating Conditions or Changes - Previous Similar Problems in Adjacent
SH/RH. .........................................................................................................................11-15
11.4.5 Combustion Conditions ....................................................................................11-15
11.4.6 Blockage or Laning of Boiler Gas Passages ....................................................11-16
11.4.7 Increases in Stress Due to Wall Thinning.........................................................11-16

xvii
13622634
 11.5 Determining the Extent of Damage ..........................................................................11-16
11.6 Repairs, Immediate Solutions and Actions...............................................................11-17
11.6.1 Need for Remaining Life Assessment ..............................................................11-17
11.6.2 Repairs .............................................................................................................11-17
11.6.3 Correction of "Off-Design" Conditions ..............................................................11-17
11.7 Life Assessment .......................................................................................................11-17
11.8 Longterm Actions and Prevention of Repeat Failures..............................................11-17
11.8.1 Remaining Life Assessment (Option 1, Figure 11-8)........................................11-18
11.8.2 Major SH/RH Redesign and Replacement (Option 2) ......................................11-18
11.8.3 Redistribute Steam Flow (Option 3) .................................................................11-19
11.8.4 Clean Tubes and Remove Sources of Blockage (Option 4) .............................11-23
11.8.5 Adjust Fireside Combustion Conditions and/or Burners (Option 5) ..................11-23
11.8.6 Minimize Laning of Gases (Option 6) ...............................................................11-23
11.9 Case Study...............................................................................................................11-23
11.10 References .............................................................................................................11-25

A METALLURGICAL CHARACTERISTICS OF SA213 TYPE 304H, 321H, AND 347H

STAINLESS STEEL SUPERHEATER AND REHEATER TUBING......................................... A-1
A.1 Tube Fabrication ........................................................................................................... A-1
A.1.1 Initial Melt .............................................................................................................. A-1
A.1.2 Manufacture .......................................................................................................... A-1
A.1.3 Heat Treatment ..................................................................................................... A-1
A.1.4 Pickling.................................................................................................................. A-6
A.1.5 Bending ................................................................................................................. A-6
A.1.6 Welding ................................................................................................................. A-6
A.2 Chemical Composition .................................................................................................. A-7
A.3 Physical Properties ..................................................................................................... A-11
A.3.1 Modulus of Elasticity ........................................................................................... A-11
A.3.2 Thermal Conductivity .......................................................................................... A-12
A.3.3 Thermal Expansion Coefficient ........................................................................... A-13
A.4 Mechanical Properties ................................................................................................ A-14
A.4.1 Tensile Strength .................................................................................................. A-14
A.4.2 Stress Rupture .................................................................................................... A-16
A.5 References.................................................................................................................. A-21

xviii
13622634
B METALLURGICAL EVALUATION TO ASSESS THE CONDITION OF AUSTENITIC
STAINLESS STEEL TUBING .................................................................................................. B-1
B.1 Common Degradation Modes ....................................................................................... B-1
B.1.1 Sensitization.......................................................................................................... B-3
B.1.2 Sigma Phase Formation........................................................................................ B-3
B.1.3 Grain Growth......................................................................................................... B-5
B.2 Metallurgical Characteristics of Tubing ......................................................................... B-7
B.2.1 Etch to Identify Second Phase Precipitates .......................................................... B-8
B.2.1.1 Example of Second Phase Etch Method ........................................................ B-9
B.2.1.2 Electrolytic Oxalic Acid Etch ........................................................................ B-14
B.2.2 Second Phase Compositional Analysis............................................................... B-17
B.2.3 Etch to Reveal Creep Microvoids........................................................................ B-22
B.2.3.1 Example of Microvoid Etch Method ............................................................. B-22
B.3 Remaining Life v. Metallurgical State.......................................................................... B-24
B.3.1 Sigma Phase Characterization............................................................................ B-24
B.3.1.1 Example of Sigma Phase Based Temperature Estimates........................... B-28
B.3.2 Creep Microvoid Characterization....................................................................... B-30
B.3.2.1 Microvoid Measurements ............................................................................ B-30
B.3.2.2 Grain Size Influence on Creep Microvoids .................................................. B-30
B.3.2.3 Example of Creep TVL and EVP Measurements ........................................ B-32
B.3.3 EVP Based Creep Damage Estimates................................................................ B-32
B.3.4 Recommended Reinspection .............................................................................. B-34
B.4 Case Study ................................................................................................................. B-35
B.5 List of Referenced Etchants........................................................................................ B-37
B.6 References.................................................................................................................. B-37

C RUL METHODS AND MODELS ........................................................................................... C-1

C.1 Roadmap ...................................................................................................................... C-1
C.2 Step I: Review design and historical information .......................................................... C-3
C.3 Remaining Life Model #1 - Rank Statistical Model ...................................................... C-3
C.3.1 Step II: Calibrated, rank statistical failure forecast (when 1 to 3 failures have
occurred) ......................................................................................................................... C-3
C.3.2 Step III: Rank statistical failure forecast (when more than a few failures have
occurred) ......................................................................................................................... C-5
C.3.3 Case Study: To predict the failure rate when one or more longterm
overheat-creep failures have already occurred ............................................................... C-6

xix
13622634
 C.4 Remaining Life Model #2 – Cumulative Damage Model............................................... C-7
C4.1 Step IV(a): Perform ultrasonic wall /steamside oxide thickness surveys and
selected tube sample metallurgical evaluations .............................................................. C-7
C.4.2 Step IV(b): Estimate the average and peak heat absorption rate .................... C-10
C.4.3 Step IV(c): Estimate the bulk steam temperature v. gas-touched length ............ C-11
C.4.4 Step IV(d): Estimate the steam velocity v. gas-touched length........................... C-12
C.4.5 Step IV(e): Estimate the tube metal temperature at selected, critical
locations ........................................................................................................................ C-16
C.4.6 Step IV(f): Estimate the tube stress at selected, critical locations.................... C-18
C.4.7 Step IV(g): Estimate the remaining longterm overheat-creep life at selected,
critical locations ............................................................................................................. C-19
C.4.8 Case Study: To predict the life when no failures have occurred or when it is
desired to include temperature, wall thinning, or other tube-to-tube variations into
the life assessment........................................................................................................ C-19
C.5.6 Step V: Refined Life Assessments...................................................................... C-30
C.5.7 Step VI: Reporting results, conclusions and recommended actions ................... C-31
C.6 References ................................................................................................................. C-31

xx
13622634
LIST OF FIGURES

Figure 2-1 Example of boiler superheater tubing materials and sizes per section.
1
Source: G. G. Stephenson and J.W. Prince ......................................................................2-3
Figure 2-2 Example of a temperature versus tube length diagram (adapted from
Reference 2). These design analyses provide an indication of which tube rows and
tube length locations might have the largest and smallest design temperature safety
margins. In this unit, longterm overheat failures of the hottest section of T11 tubing
(i.e., approximately 35 m from the inlet header) had occurred. No failures had
occurred in the austenitic stainless steel tubing. The T11 tube metal temperature
inferred from oxide scale measurements was as high as 593°C (1100°F). .......................2-3
Figure 2-3 Comparison of the longterm NRIM stress rupture test data and Larson Miller
and Manson Haferd regression models with the ASME Section I allowable stress-
temperature Values. The NRIM test data and the Larson Miller regression model
o
were in good agreement with the allowable values for temperatures up to 650 C
o
(1202 F). The Manson Haferd model was more conservative than the ASME values
o o
for temperatures exceeding approximately 575 C (1067 F). ..............................................2-8
Figure 3-1 Common in-service degradation modes for austenitic stainless steel tubes,
illustrating some of their macro- and microscopic characteristics and the report
section where additional information can be found. The most common location of
damage is indicated. The damage mechanisms shown with rectangles are most
commonly circumferentially oriented..................................................................................3-2
Figure 4-1 Schematic representation of fireside corrosion development for superheaters
and reheaters involving a molten intermediate layer (alkalis, sulfates). This case
shows maximum wastage at the 10 and 2 o'clock positions. .............................................4-2
Figure 4-2 Tube sample exhibiting fireside corrosion. Note the presence of multilayered
fireside scale along with wastage flats at the 10 and 2 o'clock positions of the tubes
4
circumference. Source: Paterson, et al. ...........................................................................4-3
Figure 4-3 Two tube sample segments showing fireside corrosion. The left shows the
ash pattern as removed from the furnace; the right shows the tube with the ash
removed. On this segment, the 12 o'clock position shows a smooth contour typical
of a fluxing fireside corrosion reaction, and the 10 and 2 o’clock positions show
alligator hide typical of longterm overheating/creep. ..........................................................4-3
Figure 4-4 Grooving of the tube’s external surface, known as “alligator hide”, associated
with coal ash corrosion. The fireside oxide scale and ash deposit have been
4
removed by glass bead blasting. Source: S. R. Paterson, et al. ......................................4-4
Figure 4-5 Typical locations where fireside corrosion can occur. .............................................4-5
Figure 4-6 Effect of metal temperature on fireside corrosion rate. Source: A.J.
1
Blazewicz and M. Gold ......................................................................................................4-7

xxi
13622634
Figure 4-7 Comparison of fireside corrosion resistance of various alloys as a function of
test temperature. Testing was performed for 50 hrs using a synthetic ash (37.5
mol% Na2SO4 + 37.3 mol% K2SO4 + 25 mol%Fe2O3) and a synthetic gas (80% N2 +
15% CO2 + 4% O2 + 1% SO2, including saturated H2O) and a vanadium catalyst
7
(V2O5). Source: S. Kihara, et al. ........................................................................................4-7
Figure 4-8 Nomograph for determination of corrosive index of coal. Source: R. Borio, et
11
al. .....................................................................................................................................4-9
Figure 4-9 Derivation of coefficients and effects of weight loss as a function of Na2O/K2O
13
ratio, content of CaO, MgO and SO2. Source: J. Shigeta, et al. ....................................4-11
Figure 4-10 The linear dependence of corrosion rate on coal chlorine content for
austenitic steels. 50 nm/hr is approximately 18 mils/yr. Source: W.H. Gibb and
15
J.G. Angus .....................................................................................................................4-13
Figure 4-11 Results of creep damage fraction, Dc, and plastic instability modeling of a
Type 304H reheater tube with an assumed fireside corrosion rate of (a) 7.5 nm/hr,
and (b) 25 nm/hr...............................................................................................................4-23
Figure 4-12 Strategies for preventing repeat failures by fireside corrosion in
superheater/reheater tubes for coal-fire units. .................................................................4-25
Figure 4-13 Effect of test temperature on corrosion loss for various alloys in 0.25 vol%
10
SO and 5 wt% alkali sulfate. Source: W. Wolowodiuk, et al. .........................................4-28
Figure 4-14 Effect of chromium content on resistance to attack after 16,000 hours
6
exposure at three plants. Source: J.L. Blough, et al. ......................................................4-29
Figure 5-1 General appearance of a 9% Cr final superheater tube containing fireside
corrosion deposits (top of figure) in an oil-fired boiler after 50,000 hours service.
The bottom of the figure shows a section of tubing having been acid cleaned to
removed the deposits and the ring section taken through the cleaned section shows
the general appearance of tube wastage on the outside surface. Source: J. Hickey,
Irish Electricity Supply Board .............................................................................................5-2
Figure 5-2 Equilibrium phase diagram for the V2O5-Na2O system showing PO2
2
dependence. (N) represents Na2O and (V) is V2O5. Source: J.R. Wilson .........................5-4
Figure 5-3 (a) Shows a secondary electron image of an oil ash scale and deposit on
2.25Cr-1Mo steel reheater tube that exhibited oil ash corrosion. (b, c, and d) Show
energy dispersive x-ray spectra from three layers in the ash. The spectra indicate
that the inner layer (b) may be an alkali vanadyl vanadate. Absence of significant
iron in the inner layer suggests that the vanadium compound is either directly
attacking the tube material or is fluxing away the indigenous iron oxide as rapidly as
it forms. The middle layer (c) is probably iron oxide which has precipitated at the
external surface of the molten layer. The outer layer (d) is comprised of ash deposit
constituents. .......................................................................................................................5-5
Figure 5-4 Typical boiler locations where oil ash fireside corrosion can occur. ........................5-8
Figure 5-5 Possible reactions that take place in liquid vanadate (oil-fired) corrosion.
2
Source: J.R. Wilson .........................................................................................................5-10
Figure 5-6 Equilibrium phase diagram for the V2O5 – MgO system showing that there is
2
no PO2 dependence. Source: J.R. Wilson .......................................................................5-13
Figure 5-7 Strategies for preventing repeat failures by fireside corrosion in
superheater/reheater tubes of oil-fired units. ...................................................................5-21

xxii
13622634
Figure 5-8 An example of fireside corrosion in a secondary superheater tube that
resulted in a longterm overheat – creep failure (rupture indicated by arrow).
Multicolored deposits are present on the fireside surface of the tube. Severe and
irregular tube wall loss (up to 47%) is apparent. ..............................................................5-28
Figure 6-1 Autocatalytic processes occurring in a corrosion pit. From: M.G. Fontana and
N.D. Greene, Corrosion Engineering, 1967, McGraw-Hill, New York, NY.
Reproduced with permission of the McGraw-Hill Companies. ...........................................6-3
Figure 6-2 Results of creep damage fraction, Dc, plastic instability, leak modeling of a
Type 304H reheater tube with axially aligned corrosion pits with a pit length to depth
ratio of one and a range of assumed fireside corrosion rates. ...........................................6-8
Figure 6-3 Results of creep damage fraction, Dc, plastic instability, leak modeling of a
Type 304H reheater tube with axially aligned corrosion pits with a pit length to depth
ratio of 10 and a range of assumed fireside corrosion rates. .............................................6-9
Figure 7-1 Cracking of SA-213 Type 304H base material near the edge of a weld
backing bar in a dissimilar metal weld joint. This tube contained through-wall
cracks after a few weeks service. The corrodant causing the cracking was
suspected to be a petroleum-based preservative which was not effectively flushed
1
form the weld backing rings prior to service. Source: S.R. Paterson, et al. .....................7-2
Figure 7-2 Intergranular stress corrosion cracking (IGSCC) of a SA-213 Type 304H
reheater tube. Away from the rupture, the IGSCC was limited to the tube’s internal
surface. Carryover of chlorides was believed to have been the corrodant
1
responsible for the cracking. Source: S.R. Paterson, et al. ..............................................7-2
Figure 7-3 Times of complete dissolution of various carbides and nitrides in austenite as
4 o
a function of temperature. (Note: 1600 F = 1144 K).........................................................7-5
5
Figure 7-4 General relationship for the penetration of SCC. .....................................................7-7
Figure 7-5 (a) Time-to-failure vs. stress for 40 sets of data measured for stainless steel
5
in 42% MgCl2 at open circuit conditions. , (b) Time-to-failure vs. 1/T for 23 sets of
data for stainless steel in boiling 35% to 45% MgCl2 solutions at open circuit
5
conditions. .........................................................................................................................7-8
Figure 7-6 Increasing the yield strength as a consequence of cold work can increase the
6
SCC growth rate and extent of TGSCC. ...........................................................................7-8
Figure 7-7 An example of an intergranular stress corrosion cracking failure of a
superheater wrapper tube. No appreciable wall loss or swelling was found. ...................7-16
Figure 7-8 Stress corrosion cracks initiated at the outside surface of the tube and
propagated as intergranular cracks without coalescence of microvoids. (Mag:
(upper) 20X, (lower) 500X, Electrolytic oxalic acid etch)..................................................7-17
Figure 7-9 Window blowout-type tube failure on the top of a horizontal reheater tube............7-18
Figure 7-10 Photomicrograph of OD-originating intergranular attack observed on area
near window fracture shown in Figure 7-9 (100x, unetched). ..........................................7-19
Figure 7-11 SEM photomicrograph intergranular morphology of the fracture surface. ............7-19
Figure 8-1 Thermo-mechanical failure of an SA-213 Type 304H superheater tube. The
portion of the rupture which is missing was believed to have contained a welded
1
attachment clip. Source: S.R. Paterson, et al. .................................................................8-2
Figure 8-2 Schematic showing typical locations of fatigue failures in steam-touched
2
tubing. Source: R.B. Dooley and H.J. Westwood. .............................................................8-3

xxiii
13622634
Figure 8-3 Three possible locations for tubing-related fatigue failures in tight 180° bends........8-3
Figure 8-4 Fatigue tube failures caused by differential thermal expansion of element
3
transfer tubes on a header. Source: W.R. Sylvester, et al. ...............................................8-4
Figure 8-5 Schematic illustrating failures due to inflexibility of movement between header
and waterwall. ....................................................................................................................8-4
Figure 8-6 Flowchart of the creep-fatigue modeling steps based on the ASME Section III,
Appendix NH guidelines...................................................................................................8-10
Figure 8-7 Flowchart of the creep-fatigue modeling steps based on the ASME Section III,
Appendix NH guidelines...................................................................................................8-11
Figure 8-8 Flowchart of the creep-fatigue modeling steps based on the ASME Section III,
Appendix NH guidelines...................................................................................................8-12
Figure 8-9 Flowchart of the creep-fatigue modeling steps based on the ASME Section III,
Appendix NH guidelines...................................................................................................8-13
Figure 8-10 Flowchart of the creep-fatigue modeling steps based on the ASME Section
III, Appendix NH guidelines..............................................................................................8-14
Figure 8-11 Flowchart of the creep-fatigue modeling steps based on the ASME Section
III, Appendix NH guidelines..............................................................................................8-15
Figure 8-12 An example of a failure that initiated at the toe of a circumferential butt weld
in a TP304H finishing superheater tube after six years of service. Flat crescent
shaped regions were identified as the crack initiation locations.......................................8-16
Figure 8-13 The fatigue crack initiated at the weld toe and propagated in a
nonbranching, transgranular cracking mode. (Mag: 20X, Electrolytic Oxalic Acid
Etch).................................................................................................................................8-17
Figure 9-1 Typical locations of failures by short term overheating in the
superheater/reheater. Exfoliating oxide in the austenitic leg AB drops to bend B
during shutdown and causes blockage. Failure typically occurs at “X” in the ferritic
material. .............................................................................................................................9-2
Figure 9-2 Growth and exfoliation of steamside oxide scale on austenitic stainless steel
4
tubing . For austenitic tubes the thermal contraction of the tubing (during cooling) is
greater than the steamside oxide. This results in a compressive strain in the oxide
of approximately 0.002mm/mm during large cooling transients. Under these
conditions exfoliation will begin to occur when the steamside scale thickness is
approximately 75µm (0.003-inch). The shaded box shown on the tension side of the
diagram is characteristic of the growth, strain, and exfoliation of ferritic tubing. ................9-4
Figure 9-3 Steam flow blockage caused by severe steamside scale exfoliation. ......................9-5
Figure 9-4 Short-term overheating caused by severe steamside oxide scale exfoliation. .........9-5
Figure 9-5 Irregular Steamside Oxide Growth on Type 304, 321, and 347 Tubing. (3) .............9-6
Figure 10-1 Typical appearance of a cracked dissimilar metal weld.......................................10-3
Figure 10-2 Typical cross-sectional appearance of a dissimilar metal weld failure after
long-time boiler service. This example is a DMW with stainless steel filler metal.
Note the oxide notch on the OD and the intergranular cracking adjacent to the weld
7
line. Source: S.R. Paterson, et al. . Inset shows further detail of the intergranular
creep cracking adjacent to the pressure weld. Note the cracking is oriented normal
to the hoop stress. Source: D. French. ............................................................................10-4

xxiv
13622634
Figure 10-3 Detailed metallographic appearance of cracking along the weld fusion line
associated with a line of carbides. This is typically observed in dissimilar metal
welds made with nickel-base filler metals. .......................................................................10-5
Figure 10-4 Microhardness profiles of dissimilar metal welds after 24 years of service
showing an increase in hardness near the weld interface. Source: K.H. Holko, et
2d
al. ...................................................................................................................................10-6
Figure 10-5 Relative performance of dissimilar metal welds with different commercial
filler metals and geometries under accelerated discriminatory testing. Source: D.I.
2h
Roberts, et al. ...............................................................................................................10-10
Figure 10-6 Possible weld geometries for dissimilar metal welds. Source: D.I. Roberts,
2h
et al. .............................................................................................................................10-11
Figure 10-7 Estimated weld performance factors for dissimilar metal weld joints made
7
using austenitic and nickel-based filler metals. Source: S.R. Paterson, et al. ...............10-12
Figure 10-8 Union Electric special radiographic technique for dissimilar metal welds.
5
Source: H.J. Grunloh, et al. ...........................................................................................10-16
Figure 10-9 Union Electric technique and multiple radiographic defect images. Source:
5
H.J. Grunloh, et al. ........................................................................................................10-17
Figure 10-10 Permissible time/temperature conditions for post-weld heat treatment of
dissimilar metal welds joining 2.25Cr-1Mo to 300-series stainless steel filler metal.
2h
Source: D.I. Roberts, et al. ...........................................................................................10-19
Figure 11-1 Types of high temperature creep failure in tubing. Source: G.A Lamping and
1
R.M. Arrowood, Jr. ..........................................................................................................11-2
Figure 11-2 Longterm creep overheat failure in a TP321 superheater tube. ...........................11-2
Figure 11-3 Typical boiler locations where longterm overheating/creep failures can
occur in vertical platen elements (a) and in horizontal platens (b). Solid circles
represent typical failure locations; letters in circles indicate locations of interest
described in the main text. ...............................................................................................11-6
Figure 11-4 Time-to-rupture curves for SA-213 Type 304 stainless steel...............................11-8
Figure 11-5 Schematic of the development of longterm overheating failures. (a)
Represents the normal situation. (b) Shows the development of “flats” at 10 and 12
o’clock. The flats can be present at other tube locations, such as the tube crown,
depending on tube alignment in the bundle. ....................................................................11-9
Figure 11-6 A reheater tube that contains steam at 510°C (950°F) will have varying
metal temperatures around its circumference and through its wall according to: gas
flow direction, ash deposit patterns, and internal scale formations. Typically, the
highest metal temperature will be achieved on the surface facing the gas flow.
1
Source: G.A. Lamping and R.M. Arrowood, Jr. .............................................................11-13
Figure 11-7 Typical thermocouple measurements across a superheater showing that
the actual tube metal temperatures can vary significantly with tube location and
1
boiler power level. Source: G.A. Lamping and R.M. Arrowood, Jr. ...............................11-14
Figure 11-8 Strategies for preventing repeat failures by longterm overheating/creep. .........11-18
Figure 11-9 Creep damage accumulation showing life extension obtained by reducing
8
the metal temperature. Source: K. Hara, et al. ..............................................................11-20
Figure 11-10 Schematic of a steam flow controller (SFC) and typical location in a
8
superheater tube bank. Source: K. Hara, et al. .............................................................11-21

xxv
13622634
Figure 11-11 Road map for achieving improved superheater longevity by steam flow
8
redistribution. Source: K. Hara, et al. ............................................................................11-22
Figure 11-12 An example of a longterm overheat- creep failure that initiated at a bend
extrados in a front pendant reheater tube. No appreciable deformation is apparent. ....11-24
Figure 11-13 The creep crack initiated at the tube external surface and propagated as
grain boundary creep cavities coalesced and linked together. (Mag: 16X,
Electrolytic oxalic acid etch) ...........................................................................................11-24
Figure 11-14 Final failure occurred by creep-rupture as evidenced by linking of creep
cavities to form unconnected intergranular cracks. (Mag: 100X, Villella's etch) ............11-25

Figure A-1 The rupture strength or life of Type 304 material is only very moderately
1
influenced by initial heat treatment or grain size. (Notes: 1) The numbers shown on
o o o o
the curves are the rupture ductility, 2) 1200 F = 649 C, 1300 F = 704 C) ......................... A-3
Figure A-2 The fupture strength or life of Type 321 material Is moderately influenced by
1
initial heat treatment or grain size. (Notes: 1) The numbers shown on the curves
o o
are the rupture ductility, 2) 1200 F = 649 C) ..................................................................... A-3
Figure A-3 The rupture strength or life of Type 347 material is strongly influenced by
initial heat treatment or grain size. The material with a coarser grain size and a
higher solution annealing temperature exhibited better performance. This was
believed to be a result of a greater amount of sigma phase formation during the
stress rupture test exposure for the material with the lower solution annealing
1
temperature. (Notes: 1) The numbers shown on the curve are the rupture ductility,
o o o o
2) 1200 F = 649 C, 1300 F = 704 C) ................................................................................. A-4
Figure A-4 The rupture strength or life of Type 347H material is strongly influenced by
initial solution annealing heat treatment. The material with a higher solution
annealing temperatures generally exhibited better performance. This effect is more
o o 2
significant at test temperatures above 650 C (1202 F) .................................................... A-5
Figure A-5 A Portion of an iron – chromium equilibrium phase diagram showing the
influence of dissolved carbon content on the stability of austenite (i.e., gamma iron,
γ-Fe plus carbon). ............................................................................................................. A-9
3

Figure A-6 Dissolved carbon under equilibrium conditions for Types 304, 321 and 347
stainless steels. Since less dissolved carbon will be present in the Type 321 and
347 stainless steels the amount of nickel is increased to 10% to 12% to ensure a
fully austenitic material. Lower levels of dissolved carbon are desirable because
they impart enhanced sensitization resistance and intergranular corrosion
2
resistance. ...................................................................................................................... A-10
Figure A-7 A portion of an iron-chromium- nickel- carbon equilibrium phase diagram
3
showing the influence on nickel on the stability of austenite. ......................................... A-10
4
Figure A-8 Yield and ultimate tensile strength of 304H material. (Ref: NRIM CDS 4B )........ A-14
5
Figure A-9 Yield and ultimate tensile strength of 321H material. (Ref: NRIM CDS 5B )........ A-15
6
Figure A-10 Yield and ultimate tensile strength of 347H material. (Ref: NRIM CDS 28B ).... A-15
Figure A-11 Normalized Larson-Miller Plot of 216 stress rupture tests of Type 304H
material. .......................................................................................................................... A-17
Figure A-12 Distribution of failure times for Type 304H laboratory stress rupture tests......... A-18

xxvi
13622634
Figure A-13 Normalized Larson-Miller Plot of 233 stress rupture tests of Type 321H
material. .......................................................................................................................... A-18
Figure A-14 Distribution of failure times for Type 321H laboratory stress rupture tests......... A-19
Figure A-15 Normalized Larson-Miller Plot of 289 stress rupture tests of Type 347H
material. .......................................................................................................................... A-20
Figure A-16 Distribution of failure times for Type 347H laboratory stress rupture tests......... A-21
Figure B-1 Appendix B roadmap for metallurgical evaluation of austenitic tubing. .................. B-2
Figure B-2 When austenitic stainless steels are exposed to prolonged exposure to
o o o o
temperatures in the range of 430 F (805 F) to 900 C (1650 F) chromium carbides
will form on the grain boundaries. This condition will result in a chromium-depleted
region along the grain boundaries. In this “sensitized” condition the material will
have increased susceptibility to intergranular corrosion, intergranular stress
corrosion cracking, and creep cavitation. Note that with increased levels of
dissolved carbon the rate and temperature range over which sensitization occurs
1
increases .......................................................................................................................... B-3
Figure B-3 Iron – chromium - nickel equilibrium phase diagram (section at 8% nickel).
The two phases that are relevant to austenitic stainless steel SH/RH tubes are
Austenite (Gamma Iron, γ + Carbon,) and Sigma Phase, σ (a grain boundary phase
comprised of approximately 50% chromium and 50% iron). The addition of carbon
will expand out the region of stability of Gamma Iron, γ-Fe. Note that even without
the benefit of carbon additions Sigma Phase is an equilibrium phase for chromium
2
levels above approximately 18% ...................................................................................... B-4
Figure B-4 Time-Temperature-Transformation curves for Types 304H, 321H, and 347H
3
materials . Note that even the stabilized grades of material will sensitize and form
sigma phase if they are exposed to prolonged temperatures approaching 600°C
(1112°F). At 650°C (1202°F) all three alloys will begin to form sigma phase after
approximately 10,000 hrs.................................................................................................. B-5
Figure B-5 Stress rupture curves for a "Weak" (A) and an "Average" Heat (B) of Type
304 stainless steel tubing material showing the four unique stress rupture cracking
mechanisms as a function of stress, temperature, and time-to-failure. The weak
heat contained a high (0.047wt%) aluminum content which lead to precipitation of
aluminum nitride associated with the grain boundary sigma phase. This caused a
4
reduction in life or strength due to early formation of cracks in the sigma phase .
Note: 1 kgf/mm2 = 9.8 MPa = 1422 psi............................................................................. B-6
Figure B-6 Etched surface of a service degraded Type 304H stainless steel tube sample
showing (A) and (B) outlined second phase particles with fully developed
microvoids (arrow). (A) and (B) were prepared with multiple 3, 3, and 2 minute
etches. (MAG: A 500X, B 1000X, Vilella’s Etch) ............................................................. B-10
Figure B-7 Etched surface of a service degraded Type 304H stainless steel tube sample
showing stained Sigma Phase particles with fully developed microvoids. Arrow in
(A) marks Sigma. Arrow in (B) marks a carbide. (MAG: 1000X, Vilella’s Etch plus
(A) NaOH and (B) KOH eectrolytic etch)......................................................................... B-11
Figure B-8 Etched surface of a “cold” side service exposed Type 304H stainless steel
tube sample showing second phase precipitation on grain boundary (arrow), twin
boundaries, and slip planes. These particles are likely carbides because they were
not stained by the concentrated caustic electrolytic etch. (MAG: 1000X, (A) Multiple
Vilella’s Etch and (B) same plus NaOH electrolytic etch)................................................ B-12

xxvii
13622634
Figure B-9 Etched surface of a service degraded Type 321H stainless steel tube sample
showing (A) outlined second phase particles (arrow) with aligned microvoids and
(B) stained particles (arrow) identified as Sigma. (MAG: 1000X, (A) Multiple Vilella’s
Etch and (B) same etch plus KOH electrolytic 1.5 volts for 8 seconds) .......................... B-13
Figure B-10 Etched surface of a service degraded Type 347H stainless steel tube
sample showing (A) outlined second phase particles (arrow) with aligned
microvoids (black cavities) and (B) stained particles (arrow) identified as sigma.
(MAG: 1000X, (A) Multiple Vilella’s Etch and (B) same etch plus KOH electrolytic
1.5 volts for 5 seconds) ................................................................................................... B-14
Figure B-11 Etched surface of a service degraded Type 304H stainless steel tube
sample showing (A) outlined second phase particles (arrow) and microvoids (black
cavities) and (B) sigma phase (arrow) dissolved away by the electrolytic oxalic acid
etch. (MAG: 500X, (A) Multiple Vilella’s Etch and (B) same etch plus electrolytic
oxalic acid 5 volts for 10 seconds) .................................................................................. B-15
Figure B-12 Etched surface of a service degraded Type 321H stainless steel tube
sample showing (A) outlined second phase particles (arrow) and microvoids (black
cavities) and (B) second phase particles (arrow) dissolved away by the electrolytic
oxalic acid etch. (MAG: 500X, (A) Multiple Vilella’s Etch and (B) same etch plus
electrolytic oxalic acid 5 volts for 14 seconds) ................................................................ B-16
Figure B-13 Etched surface of a service degraded Type 347H stainless steel tube
sample showing (A) outlined second phase particles (arrow) and microvoids (black
cavities) and (B) similar microstructure following the electrolytic oxalic acid etch.
(MAG: 500X, (A) Multiple Vilella’s Etch and (B) same etch plus electrolytic oxalic
acid 5 volts for 14 seconds) ............................................................................................ B-17
Figure B-14 Energy dispersive spectra from a Type 304H stainless steel tube showing
the composition of the (A) austenite matrix and (B) sigma phase particle. Note the
high chromium/iron (Cr/Fe) ratio of the sigma phase compared to the austenite
matrix. ............................................................................................................................. B-19
Figure B-15 Compositional dot maps showing (A) chromium and (B) iron contents of
the sigma phase and grain boundary carbide compared to the austenite matrix.
Sigma has a higher chromium content and lower iron content than the matrix. (C) is
a SEM image................................................................................................................... B-20
Figure B-16 Compositional dot map showing (A) chromium and (B) iron concentrations.
The titanium nitride particles (rectangular shapes) have higher titanium and lower
iron concentrations than the austenite matrix. ................................................................ B-21
Figure B-17 Service degraded Type 304H stainless steel tube sample showing (A) small
voids in the as-polished condition, (B) outlined second phase particles with some
microvoids, and (C) fully developed microvoids (black cavities). Arrows mark the
same location (A) as-polished. (B) 1 minute etch. (C) multiple 3, 3, and 2 minute
etches. (MAG: 500X, Vilella’s Etch) ................................................................................ B-23
Figure B-18 Etched surface of a service degraded Type 321H stainless steel tube
sample showing (A) outlined second phase particles with minor microvoids and (B)
fully developed microvoids (black aligned cavities). Arrows mark the same location.
(A) single 1 minute etch. (B) multiple 3, 3, and 3 minute etches. (MAG: 500X,
Vilella’s Etch)................................................................................................................... B-24
Figure B-19 Sigma phase thickness versus LMP for Types 304H stainless steels. .............. B-25
Figure B-20 Sigma phase thickness versus LMP for Types 321H stainless steels. .............. B-26

xxviii
13622634
Figure B-21 Sigma phase thickness versus LMP for Type 347H stainless steels. ................ B-26
Figure B-22 Microstructure from the hot side of a service degraded Type 321H furnace
tube showing creep cavitation that initiated at grain boundary sigma phase
particles. (MAG: 500X, Vilella’s Reagent Etch for 4+3+2 minutes) ................................ B-28
Figure B-23 Microstructure from the hot side of a service degraded Type 304H furnace
tube showing creep cavitation that initiated at grain boundary sigma phase
particles. (MAG: 500X, Vilella’s Reagent Etch for 4+3+2 minutes) ................................ B-29
Figure B-24 Preferential microvoid orientation normal to the principal stress. (MAG:
500X, Vilella’s Etch) ........................................................................................................ B-32
Figure B-25 Normalized strain versus exposure time data from austenitic stainless steel
creep-rupture tests. ......................................................................................................... B-33
Figure B-26 Microvoids of service degraded Type 304H stainless steel tube showing
the microvoid measurement direction and a typical grain size measurement. (MAG:
500X)(Vilella’s Etch)........................................................................................................ B-36
Figure C-1 Life assessment roadmap for austenitic stainless steel tubing subject to
longterm overheat – creep. ............................................................................................... C-2
Figure C-2 Predicted cumulative failure rate for a 624 circuit group of Type 304H tubes
with the first failure occurring in 200,000 hours................................................................. C-5
Figure C-3 Results of a nondestructive, ultrasonic steamside oxide survey, made on the
2-1/4Cr 1Mo outlet header stub tubes downstream of the austenitic stainless steel
tubing. ............................................................................................................................... C-8
Figure C-4 Estimates of the outlet steam temperatures based on the steamside scale
thickness survey shown in Figure C-2. These temperature estimations were derived
using the EPRI TUBELIFE III computer code. The “unit variable” was adjusted until
the average of the predicted outlet temperatures was equivalent to the bulk outlet
o
temperature of 1050 F....................................................................................................... C-9
Figure C-5 The relationship between the heat energy added to the outside surface of
the tubing balanced with the change in energy of the steam can be used to estimate
the average heat absorption rate of each individual tube circuit. .................................... C-10
Figure C-6 Steamside oxide scale thickness measurements made around the periphery
of a selected ferritic tube can be used to derive an estimate of the ratio of the peak
to average heat absorption rate at a given gas-touched tube location. As shown, the
peak to average heat absorption rate ratio is a function of the steamside heat
transfer coefficient. The ratio of peak to average heat absorption rate for the case
shown was 1.13, 1.24 and 1.35 for steamside heat transfer coefficients of 1700
2 2 2
W/m °C, 3400 W/m °C and 5100 W/m °C, respectively. ................................................. C-11
Figure C-7 The heat/steam energy balance relationships can be used to estimate the
bulk steam temperature as a function of gas-touched length.......................................... C-12
3
Figure C-8 The Darcy-Weisbach equations define the relationships between surface
roughness, flow rate and pressure drop with tube circuits. ............................................. C-14
Figure C-9 Local loss factors, KLlocal, are used to estimate the pressure drop within tube
circuits that are associated with steam contraction, steam expansion, directional
changes, etc. They are added to the frictional loss factors KLf = (ƒ)(L)/Di, associate
with the pressure drop in straight constant inside diameter segments of the tube
5
circuits. ........................................................................................................................... C-15

xxix
13622634
Figure C-10 For superheater or reheater tube designs with no tube bifurcations, the
pressure drops and steam velocities can be estimated using simple series and
parallel circuit analogies.................................................................................................. C-16
Figure C-11 Example of the estimation of the temperature rise between the bulk steam
temperature and the inside surface of the steamside oxide scale in a segment of
heat absorbing tubing...................................................................................................... C-17
Figure C-12 Configuration of the Ten Assembly, Three Row superheater (A). The tube
configuration at the inlet and outlet headers (B) is identical............................................ C-20
Figure C-13 The penthouse header stub tube steamside oxide thickness measurements
are used to estimate the distribution of outlet steam temperature for every tube
circuit............................................................................................................................... C-22
Figure C-14 Steam pressure and velocity along the length of tube Rows 1, 2, and 3 of
Assembly 4 of 10. These estimates were made using the Darcy-Weisbach
equation, and a series-parallel circuit analogy. The surface roughness, e, should be
adjusted to produce an estimated pressure drop through each circuit that is close to
measured values. In this case a surface roughness of 2.54E-5 m (0.001”) produced
good agreement. ............................................................................................................. C-27

xxx
13622634
LIST OF TABLES

Table 2-1 Example Austenitic Stainless Steel Tube Sizes .......................................................2-1

Table 2-2 Maximum Tube Metal Temperatures ........................................................................2-5
Table 2-3 ASME B&PV Code Allowable Stress Values For Section I Design (From
ASME B&PV Code , Section II-Part D, Table 1A, 1992) ....................................................2-5
Table 3-1 Screening Table for Austenitic Stainless Steel Tube Failures ...................................3-3
Table 4-1 Major Root Cause Influences, Confirmation and Corrective Actions ......................4-14
Table 4-2 Life Estimates of a 400 Circuit Reheater with 54 mm x 3.4 mm Type 304H
o
Tubes and a 5.69 MPa Steam Pressure and 600 C Midwall Tube Temperature.............4-24
Table 4-3 Life Estimates of a 500 Circuit Superheater with 54 mm x 6.0 mm Type 321H
Tubes and a 12.76 MPa Steam Pressure and 590°C midwall tube temperature.............4-24
Table 5-1 Melting Point of Some Commonly Observed Slag Deposits.....................................5-3
Table 5-2 Comparison of Characteristics of Longterm Overheating/Creep, Short-Term
Overheating, and Fireside Corrosion (Oil-Fired Units) In Superheater/Reheater
Tubing ................................................................................................................................5-6
Table 5-3 Impurity Levels of Residual Fuel Oils .......................................................................5-9
Table 5-4 Melting Points of MgO V2O5 Compounds.................................................................5-13
Table 5-5 Major Root Cause Influences, Confirmation and Corrective Actions ......................5-15
a
Table 5-6 Additives for High-Temperature Corrosion and Fouling from High-Sulfur Oil .......5-24
Table 5-7 Additives for High-Temperature Corrosion and Fouling for Intermediate-Sulfur
a
Oil ...................................................................................................................................5-25
Table 7-1 Major Root Cause Influences, Confirmation and Corrective Actions ......................7-10
Table 8-1 Major Root Cause Influences, Confirmation and Corrective Actions ........................8-6
Table 10-1 Distinguishing Features (Microscopic) of Failures in Fe Based Stainless Steel
and Ni Base Filler Metals in DMWs..................................................................................10-2
Table 10-2 Macro-Features Common to All Dissimilar Metal Weld Failures............................10-3
Table 10-3 Summary of Performance Characteristics of DMW Made With Various Filler
2h
Metals ............................................................................................................................10-9
Table 10-4 Major Root Cause Influences, Confirmation and Corrective Actions ..................10-13
Table 11-1 Comparison of Characteristics of Longterm Overheating/Creep, Short-Term
Overheating, And Fireside Corrosion (Coal-Fired Units) In Superheater/Reheater
Tubing ..............................................................................................................................11-4
Table 11-2 Maximum Tube Metal Temperatures ....................................................................11-7
Table 11-3 Major Root Cause Influences, Confirmation and Corrective Actions ..................11-10

xxxi
13622634
Table A-1 Chemical Requirements of SA-213/213M Type 304H, 321H, and 347H Steels ..... A-8
Table B-1 Summary of Failed Tube Data ............................................................................. B-31
Table B-2 EVP-Based Recommended Actions and Reinspection........................................ B-35
Table C-1 Local Loss Coefficients ......................................................................................... C-15
Table C-2 Dimensions of the tube segments for each of the 3 rows ..................................... C-21
Table C-3(a) Measured outlet stub tube steamside oxide scale thickness values and
estimated steam temperatures........................................................................................ C-23
Table C-4 Summary of the outside surface areas of the heat absorbing segments of
Assembly 4, Row 3 ......................................................................................................... C-24
Table C-5 Diameters, thermal conductivities, and estimated temperature distribution .......... C-29

xxxii
13622634
1
INTRODUCTION

1.1 Background

Austenitic stainless steel tubes are frequently used in superheaters and reheaters because of their
elevated temperature strength, corrosion, oxidation, and erosion resistance. The evolution of
austenitic stainless steels for boiling tubing is shown in Figure 1-1.

Three alloys are most commonly used: SA213 Type 304H, Type 321H, or Type 347H (UNS
Designations S30409, S33109, and S34709). Type 304H is an 18% chromium, 8% nickel alloy.
Alloy 321H is an 18% chromium, 10% nickel alloy with approximately 1/2% titanium. It
exhibits improved resistance to intergranular attack by corrosive condensates after being heated
in the sensitization range of 427°C (800°F) to 704°C (1300°F). Type 347H 321H is an 18%
chromium, 12% nickel alloy with approximately 3/4% of niobium (columbium) added to
stabilize the carbides and provide enhanced resistance to oxidation and intergranular attack. It
has the highest resistance to intergranular attack of the three alloys. Unlike the other two alloys
its stress rupture properties are somewhat sensitive to the initial solution annealing temperature.

Figure 1-1
Evolution of austenitic stainless steels for boiler tubing1.

1-1
13622634
Introduction

Austenitic stainless steel boiler tubes are supplied to the following specifications:

Country Specification

U.K. BS3059 Steel boiler and superheater tubes

Seamless and welded austenitic stainless steel pipes and tubes for
BS3605
pressure purpose

Seamless ferritic and austenitic alloy steel boiler, superheater and heat
U.S.A. ASTM A213
exchanger tubes

Germany DIN 17459 High temperature seamless circular stainless steel tubes

AD Merkblatt W2 Material for pressure vessels - austenitic steels

Steel tubes and seamless austenitic steel tubes for use at high
France NFA 49-214
temperature

TU 14-3-460-75E Steam boiler tubes and pipe-lines of high or subcritical steam parameter
Russia
TU 14-3-796-79 equipment

Japan JISG3463 Stainless steel and heat exchanger tubes

General British German French Japanese American

Grade BS W .-Nr. AFNOR JIS ASTM UNS

304H CFS304S51 1.4948 Z6CN 19-10 SUS304HTB Tp304H S30409

1.4878/ SUS321HTB
321H CFS321S51 Z6CNT 18-12B Tp321H S32109
(1.4941)

347H CFS354S51 1.4961 Z6CN Nb 18-12B SUS347HTB Tp347H S34709

These austenitic stainless steels have in general provided excellent service performance as
superheater or reheater tubes. However, as the population of fossil power plants ages, the number
of failures of these tubes, especially of the longterm overheat variety, has been increasing. As a
result, there is a need to be able to predict when these tubes will begin to fail and what the failure
rates will be. This type of life assessment technology is well developed for ferritic tubing and is
based on measurements of the tube wall thickness and the steamside oxide scale. This
assessment method is not directly applicable to the austenitic tubing because the austenitic
materials have excellent oxidation resistance and in general do not exhibit significantly thick
steamside oxide scales prior to failure. There is a need to develop a condition and life assessment
methodology that is unique from the methods in common use for the ferritic tubing. This report
has been prepared to fulfill this need.

1-2
13622634
 Introduction

1.2 Objectives

The primary objective of this report is to provide a compilation of currently known information
about how to predict the life of austenitic stainless steel superheater and reheater tubes that can
suffer a number of different degradation and/or failure mechanisms. Additional objectives are to:
• Document the degradation and failure mechanisms that are experienced by austenitic
stainless steel superheater and reheater tubes
• Provide direct, easy to follow, roadmap type approaches for assessing the remaining life of
austenitic stainless steel superheater and reheater tubes
• Provide sufficient background information and case studies so that the reader, if interested,
can understand the basis for the remaining life models.
• Provide supporting metallurgical documentation for use in remaining life assessments.

This compilation includes all known austenitic stainless steel tube failure mechanisms. It is never
possible to anticipate what future concerns might arise, but a final objective of the work reported
here is to provide enough information about how to perform remaining life assessments so that
rational remaining life approaches can be formulated for the analysis of as yet unknown damage
mechanisms or failure trends.

1.3 Organization of This Report and How To Use It

After this introduction, this report begins with an overview of basic design aspects and
metallurgical characteristics of new austenitic stainless steel tubes (Chapter 2). This chapter also
introduces and demonstrates how to perform bounding deterministic and probabilistic cumulative
damage life estimations for a stress rupture failure mechanism. Chapter 3 follows with a
description of in-service degradation modes for austenitic steels, providing insights into the
significance of this degradation in terms of its affect on remaining life estimations. These set the
stage for estimating the design-basis life of the tubing and what conditions may lead to tube
failures well before this life is achieved.

Chapters 4 through 11 provide descriptions of the common damage mechanisms that have been
experienced in austenitic tubes. The information presented in these chapters comes from case
studies and from the information compiled in EPRI’s three-volume Boiler Tube Theory and
2
Practice Handbooks . These chapters introduce the concept of damage rate influence factors, and
simple time to failure relationships derived from laboratory test data. Case studies are used to
demonstrate how these very simple methods can provide meaningful remaining life estimates
when one or more failures have already occurred.

There are also three appendices in the report. Appendix A provides compilations of physical,
mechanical, and metallurgical characteristics of the three commonly used austenitic stainless
steel tube materials. This information forms the basis for the remaining life estimations.
Appendix B provides a description of the factors which influence tube degradation and
metallurgical evaluation techniques that can be used to assess the condition of austenitic stainless
steel tubing.

1-3
13622634
Introduction

Appendix C provides detailed derivations of the more complex remaining life models. This
information was placed in an appendix with the hope that it would be useful to the reader that
wants to fully understand the basis for the life estimations and may have the desire to creatively
enhance the models. Appendix C brings together all the remaining life estimation concepts and
models that have been presented in the earlier chapters. It presents a roadmap approach to the
remaining life assessment of stainless steel tubing. It includes a comprehensive case study
demonstrating how the remaining life estimations rely on and provide input and decision making
capability to EPRI’s recommended seven-step method to the prevention or reduction of boiler
tube failures. The seven steps are:

1. Determine the Features of Failure and Typical Locations

2. Determine the Mechanisms of Failure

3. Determine the Possible Root Causes and Actions to Confirm

4. Determine The Extent of Damage

5. Determine and Implement Repairs, Immediate Solutions and Actions

6. Life Assessment

7. Determine and Implement Longterm Actions and the Prevention of Repeat Failures

It was the authors’ goal in preparing this report to provide a stand-alone document that could
guide the reader in how to predict the remaining life and make optimal, creative decisions that
involve the austenitic stainless steel superheater or reheater tubes. We encourage the reader to
use this reference as a starting point and to add their own knowledge and creativity to the
decision making process.

1.4 References
1
Viswanathan, R., and W. Baker, “Materials for Ultrasupercritical Coal Power Plants- Boiler
Materials:Part 1, Journal of Materials Engineering and Performance, Vol. 10(1), (Feb. 2001), pp.
81-93
2
Dooley, R.B., and W.P. McNaughton, Boiler Tube Theory and Practice: Theory and Practice,
EPRI TR-105261-V1, V2, and V3, 1996

1-4
13622634
2
STAINLESS STEEL SUPERHEATER AND REHEATER
TUBING – DESIGN BASIS

2.1 Common Alloys and Tubing Sizes

The most common outside diameter for superheater and reheater tubes are 50.8mm (2”) to 63.5
mm (2-1/2”). Some examples of tube thickness and inside diameters for these tubes are listed in
Table 2-1 below.

Table 2-1
Example Austenitic Stainless Steel Tube Sizes

Situation Design Pressure OD Minimum

Wall

Superheater/supercritical 26.72 MPa-g 50.8 mm 7.4 mm

coal-fired (304H tubing) (3875 psig) (2”) (0.293”)

Superheater/coal fired 16.90 MPa-g 54.0 mm 5.6 mm

(347H tubing) (2450 psig) (2-1/4”) (0.220”)

Superheater/oil fired 12.76 MPa-g 54.0 mm 6.0 mm

(321H tubing) (1850 psig) (2-1/4”) (0.238”)

Reheater/supercritical 5.69 MPa-g 54.0 mm 3.4 mm

coal-fired (304H tubing) (825psig) (2-1/4”) (0.135”)

Reheater/oil-fired 3.43 MPa-g 54.0 mm 3.76 mm

(304H tubing) (497psig) (2-1/4”) (0.148”)
Notes: Tubes purchased to ASTM A213 have the following tolerances:
OD (hot-finished) = specified plus 0.4 mm (1/64”) or minus 0.8mm (1/32”)
OD (cold-drawn) [for 50.8mm to 63.5 mm (2” to 2-1/2”) specified OD] = specified plus or minus 0.25 mm
(0.010”)
Maximum wall (hot-finished) = specified plus 28% (for specified wall > 4.6 mm (0.180”)
Maximum wall (cold-drawn) = specified plus 22% (for specified diameter > 38.1 mm (1-1/2”)

2.2 Design-Basis Service Pressures and Temperatures

Austenitic stainless steel superheater and reheater tubes are designed in accordance to applicable
codes such as the ASME Boiler and Pressure Vessel Code, Section I. The objective of the design
codes is typically to “afford reasonably certain protection of life and property and to provide a
margin for deterioration in service so as to give a reasonably long safe period of usefulness.”
2-1
13622634
Stainless Steel Superheater and Reheater Tubing – Design Basis

There is no specific design life/reliability expectation of boilers designed in accordance with this
Code. However, as will be presented later in this chapter, it is possible to back-out some
reasonable judgments regarding the design life of superheater/ reheater tubes that have been
designed to the Section I Code rules.

The design information that is useful to review (from the perspective of design-basis remaining
life assessments) includes:
• Sectional side views of the boiler and superheater or reheater tube banks of concern. These
provide the overall layout and arrangement of the tubing and will be used to make judgments
regarding where the tubing may experience locally high heat absorption, erosion, sootblower-
influenced erosion, etc.
• Material drawings of the tube bank (Figure 2-1). These provide the alloy specifications,
specified tube and fittings dimensions
•
2
Gas-touched length versus temperature diagram (Figure 2-2). These are typically not
provided to the owner with the original design documentation for the unit, but can sometimes
be obtained from the original equipment manufacturer/designer. These diagrams document
the designer’s estimation of the fluid and tube metal temperatures (including estimated
variability), and the design temperatures along the gas-touched length of each tube circuit
within a tube bank. These diagrams are extremely useful for identifying which tube rows
may have the lowest flow rates and highest tube metal temperatures and which locations have
the lowest design margins (safety factors on design temperature)
• The design pressure at the inlet and outlet of the tube bank

The predominant damage mechanism emphasized by ASME Section I design is longterm

overheating (i.e., stress rupture). The design basis life estimations presented in this chapter also
assume that longterm overheating combined with a moderate amount of tube wall loss due to
corrosion, oxidation, or erosion are the life limiting factors. Damage such as stress corrosion
cracking, out-of service corrosion, severe in-service wastage, short-term overheating, thermal-
shock, etc. are not considered in the design and are not considered in the design-basis life
estimations presented in this chapter. These mechanisms are addressed elsewhere in this
document.

2-2
13622634
 Stainless Steel Superheater and Reheater Tubing – Design Basis

Figure 2-1
Example of boiler superheater tubing materials and sizes per section. Source: G. G.
Stephenson and J.W. Prince1

Figure 2-2
Example of a temperature versus tube length diagram (adapted from Reference 2). These
design analyses provide an indication of which tube rows and tube length locations might
have the largest and smallest design temperature safety margins. In this unit, longterm
overheat failures of the hottest section of T11 tubing (i.e., approximately 35 m from the inlet
header) had occurred. No failures had occurred in the austenitic stainless steel tubing. The
T11 tube metal temperature inferred from oxide scale measurements was as high as 593°C
(1100°F).

2-3
13622634
Stainless Steel Superheater and Reheater Tubing – Design Basis

Many design and construction codes do not explicitly provide a design life. Once the general
design is specified (e.g., the tube surface area, steam pressure and temperature, fuel/fuel ash
characteristics, gas temperature. etc.), then a tubing material and its dimensions are chosen. This
selection process considers:
• Maximum allowable or manufacture set temperature limits (see Table 2-2)
• Allowable stress (see Table 2-3)
• Allowable wall thickness (which is a function of the allowable stress and the maximum
allowable working pressure, MAWP). The minimum allowable thickness is determined from
the following design formula:
PD
t= + 0.005D (2.1)
2S + P
where:
t = minimum required thickness, excluding wastage allowance
P = maximum allowable working pressure
D = outside diameter
S = maximum allowable stress (Table 2-3)

For example if the maximum allowable working pressure is 2000 psig (13.79 MPa-g), the
maximum allowable metal temperature is 1150oF (621oC), and the specified outside
diameter is 2.00” then the minimum allowable wall thickness for Types 304H, 321H, and
347H tubing would be:

t (304H, MAWP = 2000psi, Temperature =1150o F) =

(2000 psi )(2") + 0.005(2") = 0.240inch (6.1mm)
2(7700 psi ) + 2000 psi

t (321H, MAWP = 2000psi, Temperature =1150o F) =

(2000 psi )(2") + 0.005(2") = 0.263inch (6.7mm)
2(6900 psi ) + 2000 psi

t (347H, MAWP = 2000psi, Temperature =1150o F) =

(2000 psi )(2") + 0.005(2") = 0.184inch (4.7mm)
2(10500 psi ) + 2000 psi
• Corrosion/wastage allowance. This is a function of the allowable wall thickness and
perceived in-service wastage rate for the tube alloy selected. Reheater tubes typically have
thinner walls and much larger corrosion/wastage allowances than superheater tubes.
• Wall thickness manufacturing under-tolerance and commercially available tube dimensions.

2-4
13622634
 Stainless Steel Superheater and Reheater Tubing – Design Basis

Table 2-2
Maximum Tube Metal Temperatures

Stainless Steel Alloy ASME ASME Max.1 B&W Max.2 C-E Max.3 Riley Max.4
Specification

18Cr – 8Ni SA213 304H 816oC 760oC 704oC ~

(1500oF) (1400oF) (1300oF)

18Cr- 10Ni- Ti SA213 321H 816oC 760oC ~ 816oC

(1500oF) (1400oF) (1500oF)

18Cr – 12Ni-Nb SA213 347H 816oC ~ 704oC ~

(1500oF) (1300oF)

Notes:
1. ASME B&PV Code Section 1, Table PG-23.1. This is the highest temperature for which maximum allowable stress values are given
2. Stultz, S. C. and J. B. Kitto, Steam: Its Generation and Use, Babcock & Wilcox Company, Barberton, Ohio, Data from Table 32-2, 1978 Edition
3. Singer, J.G.,ed., Combustion Fossil Power, A Reference Book on Fuel Burning and Steam Generation, Combustion Engineering, Inc., Windsor,
Connecticut, Data from Table IV, 1981 Edition
4. French, D.N., Metallurgical Failures in Fossil-Fired Boilers, John Wiley & Sons, Wiley-Interscience Publications, New York, Data from Table VI, 1983
Edition

Table 2-3
ASME B&PV Code Allowable Stress Values For Section I Design
(From ASME B&PV Code , Section II-Part D, Table 1A, 1992)

Allowable Stress, MPa (ksi), For Alloy Type(1)

Metal Temperature, oC 304H 321H 347H

(oF)

371 (700) 109.6/ 76.5 (15.9/ 11.1) 108.9/ 75.8 (15.8/ 11.0) 101.4/ 88.3 (14.7/ 12.8)

399 (750) 107.5/ 74.5 (15.6/ 10.8) 108.2/ 75.2 (15.7/ 10.9) 101.4/ 88.3 (14.7/ 12.8)

427 (800) 104.8/ 73.1 (15.2/ 10.6) 106.9/ 74.5 (15.5/ 10.8) 101.4/ 87.6 (14.7/ 12.7)

454 (850) 102.7/ 71.7 (14.9/ 10.4) 106.2/ 73.8 (15.4/ 10.7) 101.4/ 87.6 (14.7/ 12.7)

482 (900) 101.4/ 70.4 (14.7/ 10.2) 105.5/ 73.1 (15.3/ 10.6) 101.4/ 86.9 (14.7/ 12.6)

510 (950) 99.3/ 68.9 (14.4/ 10.0) 104.8/ 73.1 (15.2/ 10.6) 100.7/ 86.9 (14.6/ 12.6)

538 (1000) 95.1/ 67.6 (13.8/ 9.8) 96.5/ 72.4 (14.0/ 10.5) 99.3/ 86.9 (14.4/ 12.6)

566 (1050) 84.1/ 65.5 (12.2/ 9.5) 80.7/ 69.6 (11.7/ 10.1) 97.2/ 86.2 (14.1/12.5)

593 (1100) 67.6/ 61.4 (9.8/ 8.9) 62.7/ 60.7 (9.1/ 8.8) 89.6/ 83.4 (13.0/ 12.1)

621 (1150) 53.1 (7.7) 47.6 (6.9) 72.4 (10.5)

649 (1200) 42.1 (6.1) 37.2 (5.4) 54.5 (7.9)

2-5
13622634
Stainless Steel Superheater and Reheater Tubing – Design Basis

Allowable Stress, MPa (ksi), For Alloy Type(1)

Metal Temperature, oC 304H 321H 347H

(oF)

677 (1250) 32.4 (4.7) 28.3 (4.1) 40.7 (5.9)

704 (1300) 25.5 (3.7) 22.1 (3.2) 30.3 (4.4)

732 (1350) 20.0 (2.9) 17.2 (2.5) 22.1 (3.2)

760 (1400) 15.9 (2.3) 13.1 (1.9) 17.2 (2.5)

788 (1450) 12.4 (1.8) 10.3 (1.5) 12.4 (1.8)

816 (1500) 9.7 (1.4) 7.6 (1.1) 9.8 (1.3)

>816 (> 1500) Not Permitted Not Permitted Not Permitted

Notes:
1) Where two values are listed the allowable stress is governed by the short time tensile properties. The
higher stress values exceed 62-1/2% but do not exceed 90% of the yield strength at temperature. Use of
these higher allowable values may result in dimensional changes due to permanent strain. These upper
values are not recommended where slight amounts of distortion can cause leakage or malfunction.

2.3 Anticipated Design-Basis Life

Although most boiler designs are not based on a specific design life, in reality, once the
maximum allowable stress and design temperature are selected then these can be used to estimate
the design life. For example, if the design temperature is 621oC (1150oF) then the allowable
stress for SA213 304H is 53.1 MPa (7.7 ksi). Recall that the allowable stress values (for
temperatures within the creep regime and for conditions which are not limited by strain limits)
are determined from the lowest of the following:
• 80% of the minimum 100,000-hour rupture stress (where the minimum is defined as the 5th
percentile based on a normal distribution of rupture stress values)
• 2/3 of the average 100,000-hour rupture stress

For the example being discussed either the minimum 100,000 hour rupture stress is 53.1 MPa/0.8
= 66.4 MPa (9.6 ksi) or the average 100,000 hour rupture stress is 53.1 MPa/0.67 = 79.7 (11.6
ksi). It should be clear from this discussion that the intent of the design allowable
stress/temperature values was to be sure that the stress rupture life of the tubing would exceed
100,000 hours. In reality the actual life will be dependent on the specific stress rupture strength
of the heats of tubing and a variety of service related factors that will be discussed in Chapter 3.

Prior to moving onto the service degradation effects it is useful to consider the effect of tube
property variations on the anticipated life of tubing. This is illustrated by comparing the
predicted life inferred from the extensive testing performed by the National Research Institute of

2-6
13622634
 Stainless Steel Superheater and Reheater Tubing – Design Basis

Metals/National Institute for Materials Science (NRIM/NIMS) in Japan for the allowable
stress/temperature combinations given in ASME Section I.

Time to rupture regression equations were developed using relevant NRIM test data. Least
square error fit Larson-Miller and the Manson-Haferd models were developed for each of the
austenitic stainless steel alloys. Test data with rupture times less than 1000 hrs and for
interrupted tests were excluded from the regression analysis. The model residuals (given as
actual time to rupture, trup, divided by the regression equation predicted time to rupture, truphat)
were fit to a lognormal distribution. The fit of the residuals provided an indication of the
accuracy of the regression model and provided an equation that can be used to estimate the
fraction of tubes that would be expected to fail for specific time/temperature/stress combinations.

The regression equations are documented in Appendix A. It was found that the Manson-Haferd
model had a slightly smaller spread in the residual values indicating that it provided a somewhat
better fit to the test data compared with the Larson-Miller model. The Larson-Miller model
seemed to provide better extrapolation characteristics (when compared to the ASME Section I
o o
allowable stress/temperature combinations) especially for temperatures above 650 C (1202 F).

The predicted stress and temperature to achieve a 100,000 hr life for the 50th (first tube in a group
of 2 tubes), 5th (first tube and a group of 20 tubes), were calculated for temperature levels from
o o o o
600 C to 700 C (1083 F to 1292 F). The results of these estimations (for 304H material) were
then compared with the ASME Section I allowable stress/temperature limits (Figure 2-3). It was
found that the NRIM stress rupture data and regression equation predictions were in general
agreement with the ASME allowable values with the exception that the Manson-Haferd model
indicated lower allowable stress levels for the higher temperatures. The NRIM test data seemed
to be in excellent agreement with the ASME allowable limits suggesting that the low Manson
Haferd estimation was an anomaly of the model, rather than a real material attribute. Based on
this behavior it was decided that the Larson Miller models would be used in the life assessment
examples used in the remainder of this report.

A close examination of the raw stress rupture data for each of the three austenitic alloys revealed
that the trends were not entirely consistent with either a Manson-Haferd or Larson-Miller model.
o
If life assessments are required for advanced plants (with metal temperatures exceeding 650 C,
then more advanced regression modeling and more longterm high temperature test data would be
required.

2-7
13622634
Stainless Steel Superheater and Reheater Tubing – Design Basis

Figure 2-3
Comparison of the longterm NRIM stress rupture test data and Larson Miller and Manson
Haferd regression models with the ASME Section I allowable stress-temperature Values.
The NRIM test data and the Larson Miller regression model were in good agreement with
the allowable values for temperatures up to 650oC (1202oF). The Manson Haferd model was
more conservative than the ASME values for temperatures exceeding approximately 575oC
(1067oF).

Using the Larson Miller regression model we can now illustrate the two steps in a design basis
life prediction. The first step estimates the average life of a group of tubes. The second step
estimates the anticipated life of the earliest failures within the group of tubes.

Consider a group of 400 finishing superheater outlet tubes fabricated from 50.8mm (2”) OD x
7.4mm (0.293”) MWT SA213 type 304H tubes with a maximum allowable working pressure of
26.72 MPa-g (3875 psig). Begin by rearranging equation 2.1 and solving for the allowable
stress, S:

 PD   26.72MPa (50.8mm ) 
 (t − 0.005 D ) − P   (7.4mm − 0.005 (50.8mm )) − 26.72MPa 
S= =  = 81.6 MPa (11.8ksi)
2 2

Interpolating within Table 2-3 (using the higher allowable stress limits) leads to an estimated
design temperature of 570oC (1058oF). The estimated average life (when 200 of the 400 tubes
o
would have failed) for a stress of 81.6 MPa and a metal temperature of 570 C can be estimated
using the Larson Miller regression model for Type 304H material (Figure A-11):

2-8
13622634
 Stainless Steel Superheater and Reheater Tubing – Design Basis

trup, 50th percentile = 10(-11.389+18746.602/(T+273.15)-268.416 Log(Stress)/(T+273.15)-969.675 Log(Stress) Log(Stress)/(T+273.15) = 1,075,719 hrs

The more useful life estimation is when the first of the 400 tubes are likely to fail. This can be
estimated using the distribution of the laboratory stress rupture tests (Figure A-12). The time to
rupture of the nth failure is estimated using the inverse lognormal distribution. Using the
nomenclature in an Excel spreadsheet:

trup, nth of X tubes = (trup, 50th percentile) LOGINV (number of failures/total tubes, 4.078E-8, 0.6580)

Using this relationship the estimated time to stress rupture failure for the first six failures is given
below:
Number of LOGINV (# of failures/total tubes, Service Time
4.078E-8, 0.6580)
failures

1 0.1577 169,643 hrs (19.4 yrs)

2 0.1836 197,521 hrs (22.5 yrs)

3 0.2018 217,073 hrs (24.8 yrs)

4 0.2164 232,761 hrs (26.6 yrs)

5 0.2288 246,141 hrs (28.1 yrs)

6 0.2398 257,965 hrs (29.5 yrs)

For the example presented the design life is approximately 20 years. The actual life of the tubing
might be longer if the service conditions are more benign than the design basis or may be shorter
if the tubing experiences unanticipated corrosion, thermal transients, or temperatures.

These simple design basis life estimation models will be used in more refined ways to make
remaining life estimations for more complex service conditions. These more complex models are
described in Appendix C.

2.4 References
1
Stephenson, G.G. and J.W. Prince, Guidelines on Fossil Boiler Field Welding, Final Report TR-
101699, EPRI, Palo Alto, CA, January, 1993.
2
Dooley, R.B., and W.P. McNaughton, Boiler Tube Theory and Practice: Theory and Practice,
EPRI TR-105261- V3, 1996, Figure 32-12.

2-9
13622634
13622634
3
IN-SERVICE DEGRADATION

3.1 Introduction

Austenitic stainless steel boiler tubing operating under high-temperature and high-pressure
conditions is subject to potential degradation by a variety of mechanical and thermal stresses,
and with the potential for both fluid- and fire-side attack. Specific potential damage mechanisms
include fireside corrosion, pitting and crevice corrosion, intergranular attack and stress corrosion
cracking, fatigue, short term overheating resulting from steamside oxide scale exfoliation and
blockage, dissimilar metal weld failure and longterm overheating/creep.

Historically, dissimilar metal weld failure has been the predominant damage mechanism
affecting austenitic stainless steel tubing. Due to improvements in the understanding of this
mechanism and how to prevent it, these failures have been reduced. Of the remaining
mechanisms affecting austenitic tubing, a compilation of recent tube failure statistics indicates
the following breakdown: longterm overheating/creep (29% of the total reported failures), stress
corrosion cracking (29%), exfoliation/short term overheating (18%), fatigue (12%) and fireside
corrosion (6%).

3.2 Use of This Report for Identifying Specific In-Service Degradation

Modes

Figure 3-1 illustrates the common in-service degradation modes for austenitic stainless steel
tubes and their characteristics. This figure and the specific features for each mechanism,
described below in Table 3-1, can be used as a rapid screening tool to identify the most likely
potential tube failure mechanisms. Reference can then be made to the appropriate report section
for details on how to confirm the specific failure mechanism, confirm its root cause(s), determine
the extent of damage, perform a life assessment and implement long term corrective actions.

For example, as shown in Figure 3-1, fireside corrosion (Sections 4 and 5) manifests itself as
general external wastage associated with a multilayered ash deposit. Pitting and crevice
corrosion (Section 6) is internal tube surface damage. Failures with an array of predominantly
straight cracks may be evidence of stress corrosion cracking (Section 7), while single,
circumferentially-oriented cracks, often at weld toes, is evidence of fatigue (Section 8). Severe
exfoliation of internal oxide scale (Section 9) is not externally visible but can be detected with
radiography or magnet testing. Accumulated creep damage on the ferritic sides of dissimilar
metal welds can lead to ruptures (Section 10). Long term overheating is manifested as
temperature- and stress-induced microstructural changes and creep damage; estimates of tube
temperatures and predictions of remaining life can be made (Section 11 and Appendix C).

3-1
13622634
In-Service Degradation

Figure 3-1
Common in-service degradation modes for austenitic stainless steel tubes, illustrating
some of their macro- and microscopic characteristics and the report section where
additional information can be found. The most common location of damage is indicated.
The damage mechanisms shown with rectangles are most commonly circumferentially
oriented.

3-2
13622634
 In-Service Degradation

Table 3-1
Screening Table for Austenitic Stainless Steel Tube Failures

Typical Fracture Other Likely Macroscopic and Typical Locations Possible Report Section
or Leak Metallographic Features Mechanism
Appearance

Thin-edged External thinning, longitudinal Highest temperature tubes Fireside Corrosion 4 (Coal-fired units)
fracture cracking; multilayered fireside deposit, (leading edge or out-of-alignment (coal- and oil-fired
possibly spalled, containing low tubes, outlet tubes, wrapper tubes, units) 5 (Oil-fired units)
melting point compounds; wastage tubes that surround a radiant
usually at 10 and 2 o’clock (flue gas at cavity).
12 o’clock); may have “alligator hide”

Pinhole leak Internal tube surface damage Locations where condensate can Pitting Corrosion 6
form and remain as a liquid during
offline periods; crevices formed by Crevice Corrosion
deposits, contact surfaces, etc.

Thick-edged Intergranular or transgranular cracking Bends and straight tubing with low Stress Corrosion 7
fracture or oriented perpendicular to the dominant spots where condensate can Cracking /
pinhole leak stress (circumferential or longitudinal); collect during offline periods; high Intergranular Attack
ID (most common) or OD initiation. stress locations

Thick-edged OD-initiated, straight transgranular Typically at tube attachments, Fatigue 8

fracture cracking (can be intergranular if creep bends and header connections.
also present).

Thin-edged Ductile “fishmouth” rupture; tube Near bottom bends in vertical Steamside Scale 9
fracture swelling; longitudinal rupture; tubing loops Exfoliation,
overheated microstructure. Blockage and Short
Term Overheating

Thick-edged Circumferential cracking in the ferritic Transitions between ferritic and Dissimilar Metal 10
fracture material; creep cavitation; external austenitic materials Weld Failure
oxide notching

3-3
13622634
In-Service Degradation

Typical Fracture Other Likely Macroscopic and Typical Locations Possible Report Section
or Leak Metallographic Features Mechanism
Appearance

Thick-edged Longitudinal rupture on the heated Highest temperature locations; Longterm Overheat 11
fracture side of the tube; internal and external near wall thickness transitions; Creep
oxidation; wastage flats; creep where there is a variation in gas-
cavitation and microstructural changes touched tube length; near radiant
(Appendix B). cavities

3-4
13622634
4
FIRESIDE CORROSION IN COAL FIRED UNITS

4.1 Introduction

Type 304H, 321H, and 347H tubing can be used in air up to 850°C (1560°F). In flue gases
containing sulfur, the corrosion resistance of Types 304H, Type 321H, and Type 347H tubing
are all reduced. In such environments these materials can be used at temperatures up to 600°C -
750°C (1110°F -1380°F) depending on service conditions. Factors to consider are whether the
atmosphere is oxidizing or reducing, (i.e. the oxygen and carbon monoxide/carbon dioxide
content), and whether impurities such as sodium and vanadium are present.

Superheater/reheater (SH/RH) fireside corrosion (also called "molten salt" attack or "coal ash"
corrosion or "liquid-phase" corrosion) is not currently a major problem for U.S. units using
538°C or 541°C (1000°F or 1005°F) steam. Historically, it has been a significant problem for
units operating in the U.S. at higher steam temperatures above 565°C (1050°F), and in those
burning coals with high fractions of chlorine such as is common in the United Kingdom (U.K.).

It is important to review the current understanding of this damage type as advanced coal-fired
units may operate at higher main steam temperatures as one means of achieving higher unit
efficiencies. Further, as the opportunity to burn coals of various compositions develops, it is
important to understand the effects of coal composition on fireside corrosion.

A third reason for reviewing the fireside corrosion mechanism is that it is often confused with
damage occurring by longterm overheating in superheater and reheater tubes (Chapter 11). Both
mechanisms eventually result in creep failure and share some common features of failure.
However, corrective actions differ as do the underlying causes.

Discussion of fireside corrosion in superheater/reheater tubing of oil-fired units can be found in

Chapter 5.

4.2 Features of Failure and Typical Locations

4.2.1 Features of Failure

Macroscopically, there will usually be a multilayered fireside scale and ash deposit. The deposits
found are generally tightly bound to tubes at room temperatures and will typically consist of
three layers1, 2:

4-1
13622634
Fireside Corrosion in Coal Fired Units

1. A hard, brittle and porous outer layer, which makes up the bulk of the deposit and has a
composition similar to that of boiler flyash.

2. A white intermediate layer consisting of compounds of complex alkali sulfates including

alkali iron trisulfates. When this layer has a chalky consistency, corrosion has been found to
be mild or non-existent; when fused and semi-glossy, corrosion has been found to be severe.

3. A black, glossy inner layer, composed primarily of oxides, sulfates and sulfides of iron.

Figure 4-1 shows the three deposit types schematically.

Deposits adjacent to the metal oxide are the ones that directly influence corrosion of the
substrate. Such deposits, if tightly bonded, should be suspected of covering corrosion sites,
whereas loosely bonded dust or thick, friable ash layers are less likely to be covering corrosion
sites.3

Fireside corrosion damage will be primarily distinguished from longterm overheating by the
presence of low melting point ash compounds. The fireside scale and ash should be examined
metallographically and chemically; the presence of low melting point (in the range 550°C to
620°C (~ 1025°F to 1150°F)) constituents in the ash is indicative of fireside corrosion. Similarly,
the presence of the alkali iron trisulphates (Na, K)3Fe(SO4)3 , particularly in the middle layer, is a
further indicator of an active fireside corrosion mechanism. The innermost layers will be
superficially similar for both mechanisms.

Figure 4-1
Schematic representation of fireside corrosion development for superheaters and
reheaters involving a molten intermediate layer (alkalis, sulfates). This case shows
maximum wastage at the 10 and 2 o'clock positions.

Tube wastage will often be evident and manifested as flat spots on the tube at the 10 o'clock and
2 o'clock positions (12 o'clock is the upstream position). A typical cross-section showing this
wastage and the presence of significant deposits can be seen in Figure 4-2. In this case tube
thinning will be most evident around the edges of the deposit. Alternatively, the maximum

4-2
13622634
 Fireside Corrosion in Coal Fired Units

wastage can occur at the 12 o'clock position. Figure 4-3 shows two tube sections with fireside
corrosion and longterm overheating/creep damage.

Figure 4-2
Tube sample exhibiting fireside corrosion. Note the presence of multilayered fireside scale
along with wastage flats at the 10 and 2 o'clock positions of the tubes circumference.
Source: Paterson, et al.4

Figure 4-3
Two tube sample segments showing fireside corrosion. The left shows the ash pattern as
removed from the furnace; the right shows the tube with the ash removed. On this
segment, the 12 o'clock position shows a smooth contour typical of a fluxing fireside
corrosion reaction, and the 10 and 2 o’clock positions show alligator hide typical of
longterm overheating/creep.

4-3
13622634
Fireside Corrosion in Coal Fired Units

Depending on the corrosion rate, removal of the fireside scale and ash deposit will reveal the
smooth contoured appearance typical of a fluxed corrosion reaction (the right hand tube
segment in Figure 4-3) or distinctive longitudinal grooving termed "alligator hide" as shown in
Figure 4-4.

Greatest wall loss will generally be seen in tubes that have been operated at the highest
temperatures over a period of time. The measurement of steamside oxide thickness by ultrasonic
or laboratory measurement will usually confirm this observation.

On austenitic steels, the external magnetite and spinel oxides become non-adherent when they
reach a thickness of about 0.2 mm (~ 0.008 in.) and tend to spall off along with any deposits that
have formed when the boiler cools. Bare metal can result and lead to rapid corrosion rates. Any
oxide scale that is thicker than about 1 mm (0.04 in.) should be considered as evidence of a fast
corrosion rate.3 There may also be localized, overlapping pits 0.5 - 2 mm (~ 0.02 - 0.08 in.) in
diameter.

The 300 series stainless steels can also be rapidly corroded if attacked by hydrogen sulfide gas,
H2S, which can be formed in flue gas that has high CO levels. Identification keys will include
carburization of the fireside surface of the stainless steel, the presence of iron sulfide scale or
discrete grain boundary sulfides.

Figure 4-4
Grooving of the tube’s external surface, known as “alligator hide”, associated with coal
ash corrosion. The fireside oxide scale and ash deposit have been removed by glass bead
blasting. Source: S. R. Paterson, et al.4

4-4
13622634
 Fireside Corrosion in Coal Fired Units

Fireside corrosion can be superficially similar to longterm overheating in superheater/reheater

tubing. The most distinguishing characteristic will be the presence of low melting ash
compounds in the deposits formed for fireside corrosion.

4.2.2 Locations of Failure

Figure 4-5 shows typical locations of concern. Corrosion will generally be the worst in the
highest temperature locations. Parts at highest risk therefore include:
• Leading sides of all tubes in pendant platens, especially hottest (leading) tubes, and steam
outlet tubes
• Tubes out of alignment that act as leading tubes
• Tubes in the outlet (final) sections towards the header, because these are at the highest
temperatures.
• Wrapper tubes
• Tubes that surround a radiant cavity (i.e., they may pick up more heat)
• At bottom bends of platens especially those facing the fireball.
• Tubes with a longer gas-touched length (GTL). GTL is the distance measured along the tube
circuit from the inlet header to the point of corrosion.
• Spacers and uncooled hangers, and the fins and studs on tubes.

Figure 4-5
Typical locations where fireside corrosion can occur.

4-5
13622634
Fireside Corrosion in Coal Fired Units

As noted above, on a particular tube, failure will often be at the 10 o'clock or 2 o'clock position
(flue gas impinging at the 12 o'clock position). Alternatively, the maximum corrosion damage
may be at the tube crown or on one side for a tube out of alignment.

4.3 Mechanism of Failure

4.3.1 Mechanism of Attack by Alkali Salts

Salts (alkali-iron trisulfates) of the type (Na, K)3 Fe (SO4)3 form by the reaction of alkali sulfates
with iron oxide in the presence of SO3. These salts can have melting temperatures as low as
552°C (1025°F) but are generally in the range 600-650°C (~ 1110-1200°F). This compares with
the much higher melting points of the simple sulfates, for example at 884°C (1623°F) for
Na2SO4 and 1069°C (1956°F) for K2SO4.

The formation of alkali-iron trisulfates has an upper temperature bound set by their stability,
which is a function of the partial pressure of SO3 at the ambient temperature.4 In the superheater,
combustion should be complete and the gas temperatures are lower than in the furnace, with
typical final superheater maximum tube temperatures around 600-650°C (~ 1110-1200°F). These
metal temperatures fall in the range in which alkali metal salts, condensed from the flue gas, can
react with the tube surface oxides in the presence of SO3, to form complex liquid sulfates, a
result that leads to rapid corrosion.

The precise mechanism by which corrosive attack occurs is still not fully understood. It is
presently held that the alkali iron trisulfates disrupt the protective tube oxide by a fluxing
mechanism and there is a subsequent transport of iron from the tube surface to form a non-
protective oxide in the salt. Subsequently, accelerated oxidation occurs because of (i) removal of
the protective oxide by mechanical loss, (ii) oxidation-sulfidation attack caused by sulfur in the
slag, and (iii) via fluxing of the oxide and sulfidation attack of the substrate metal

In a recent investigation of alloys for use in the SH/RH sections of advanced steam plant6,
laboratory and field probe tests indicated that the mechanism of coal-ash corrosion in those tests
was an accelerated oxidation/sulfidation process. In low-corrosion-resistant, low-chromium
alloys, oxidation/sulfidation occurred at a linear rate indicative of a fluxing action. In the higher
chromium-containing alloys, an initial resistance was broken down by a diffusion-controlled
step, the formation of a chromium oxide layer which caused depletion of the underlying zone and
rendered it more susceptible to oxidation/sulfidation.

Although the precise mechanism is still unknown, the results are well established. The wastage
rate is a function of metal temperature as shown in Figure 4-6 for 2¼ Cr ferritic steel and 18Cr-
8Ni stainless steel and in Figure 4-7 for many other alloys. This characteristic bell-shaped curve
has been confirmed by both laboratory work and plant probe trials for a wide range of
materials.2, 3, 8

4-6
13622634
 Fireside Corrosion in Coal Fired Units

Figure 4-6
Effect of metal temperature on fireside corrosion rate. Source: A.J. Blazewicz and M.
Gold1

Figure 4-7
Comparison of fireside corrosion resistance of various alloys as a function of test
temperature. Testing was performed for 50 hrs using a synthetic ash (37.5 mol% Na2SO4 +
37.3 mol% K2SO4 + 25 mol%Fe2O3) and a synthetic gas (80% N2 + 15% CO2 + 4% O2 + 1%
SO2, including saturated H2O) and a vanadium catalyst (V2O5). Source: S. Kihara, et al.7

These figures show several important factors. There is a rapidly increasing rate of corrosion with
temperature to a maximum around 700°C (~ 1300°F). If steam temperatures are controlled to
around 538°C (1000°F), tube metal temperatures should be limited to approximately 595°C (~
1100°F) which is outside the zone where the maximum attack occurs. Thus, the consideration of
fireside corrosion has led, in part, to placing limitations on superheater/reheater tube temperature
as an operating consideration. A second response to fireside corrosion in SH/RH tubes has been

4-7
13622634
Fireside Corrosion in Coal Fired Units

to use a more resistant material. The increased resistance of the high Cr alloys like IN 671 (50 Cr
- 50 Ni) is indicated in Figure 4-7.

Because the wastage rate is a strong function of tube metal temperature, the propensity for rapid
corrosion can be exacerbated by any of a number of problems that lead to even slightly elevated
tube temperatures, such as the presence of thick internal oxide scale, or a variety of operating
conditions such as excessive flue gas temperatures or local hot spots.

This sensitivity to temperature also explains why damage typically occurs at the 10 o'clock and 2
o'clock positions on the tubes. Deposits are usually thinner at these locations because of the flow
around the tube (Figure 4-1) and, as a result, these locations are generally slightly hotter than the
balance of the tube. Furthermore, a minimum level of SO3 (on the order of 250 ppm) is required
for the corrosion process to occur, and access of SO2 to the tube surface is also easier in those
locations of the tube because of the thinner deposits.

The accelerated wastage results in loss of wall thickness which causes an increase in hoop stress
and leads to eventual creep failure.

4.3.2 Sulfidation of Stainless Steels

Sulfidation is a second mechanism that may cause fireside corrosion in superheater/reheater

tubes. Stainless steels are rapidly corroded at high temperatures by hydrogen sulfide gas, H2S;
this can be formed in flue gas that has high CO levels, such as caused by incomplete combustion
or low excess air levels. Sulfidation attack and carburization of the stainless steel will result in
discrete iron sulfides at the grain boundaries of the corroded region.

It has also been proposed that a local carburizing atmosphere can be created during startup oil
firing when unburned oil coats the tubes. When this deposited oil subsequently burns, a
carburizing atmosphere is created which then makes the stainless steel more susceptible to
sulfidation.

4.3.3 Coal Composition and Corrosivity Indices

The most important species responsible for fireside corrosion in coal-fired boilers are alkali
metal compounds and sulfur.

As a general rule, for U.S. coals, eastern high-sulfur coals can be corrosive and prone to low
melting sulfate or alkali-iron trisulfate formation, whereas western low-sulfur coals usually do
not cause a coal-ash corrosion problem, but can cause slagging and fouling because of the high
levels of sodium and calcium in the coal.6

There is some evidence that the corrosiveness of these deposits increases with increasing coal
sulfur content over about 2 percent.9 However, in general, for coals with sulfur levels above
about 2 percent, changes in the content of volatile alkali metal compounds have a stronger
influence on corrosion than do changes in sulfur content.2

4-8
13622634
 Fireside Corrosion in Coal Fired Units

The mitigating properties of CaO and MgO have also been documented. Most recently in an
investigation of alloys for advanced steam plant, it was found that a 2½% addition of CaO was
beneficial in reducing the rate of fireside corrosion, particularly for the less-resistant materials.10
Calcium oxide lowers the potential for sulfur at the tube surface by forming a layer of calcium
sulfate in the deposit; this is expected to result in a lower corrosion rate. In the field experience,
there is often such a layer found. The effect could also be due to increasing melting point.

Several corrosion indices have been proposed to determine the likelihood that a given coal will
lead to a potential fireside corrosion problem, including methods proposed by Borio, et al.11,
Raask12, and Shigeta, et al.13

The index developed by Borio, et al.11 was based on an experimental study of U.S. coals from
six mines, specifically for Eastern U. S. coals. The index produced a value ranged from 0 -22,
with the potential for corrosion increasing with higher index values. It considered the effect of
alkali metals (Na + K), alkaline earth metals (Ca + Mg) and sulfur (as FeS) on the corrosion
potential of the coal. For example, the presence of CaO and MgO in coals elevates the melting
point of sulfates and lessens the possibility of molten salt attack. A nomograph (Figure 4-9) was
developed that incorporated four coal composition factors: (i) acid soluble K2O, (ii) acid soluble
Na2O, (iii) iron oxide Fe2O3, and (iv) CaO + MgO.

Figure 4-8
Nomograph for determination of corrosive index of coal. Source: R. Borio, et al.11

4-9
13622634
Fireside Corrosion in Coal Fired Units

This work lead to practical applications such as the blending of coal to adjust levels of CaO +
MgO as a function of the Na, K and S content of coal. The peak temperature and the corrosion-
temperature relationship at each temperature are likely to be different if coals are blended or
washed to meet emission standards, or if coals of different composition, such as Western coals
are considered.

The Borio index has had some success in ranking the potential for corrosion in units using
Eastern U.S. coals, as most recently confirmed in field probe tests of seven materials under
consideration for advanced plant conditions.6 Two units were part of the test program and burned
high sulfur subbituminous coal with Borio index ratings of 2.5 and 3.5. For five of the seven
alloys tested, the index ordered the corrosion behavior. However, an evaluation of the use of the
index to correlate with units which had severe corrosion problems and burned a wider range of
coal types met with only limited success.2

This limited applicability would indicate that other factors need to be considered when looking
for predictive methods associated with fireside corrosion, and that use of the Borio index should
probably be limited to bituminous coals of the type used to establish in the index.

Raask12 proposed a simple ranking system based on the sum of the percentages of water-soluble
sodium and potassium in coal as determined on the bomb residue from the measurement of coal
calorific value. The three categories are:

Water-soluble Na + K,
weight percent Corrosiveness

< 0.5 Low

0.5 - 1.0 Medium

> 1.0 High

A third index based on an extensive study of boiler deposits, test probe deposits, and laboratory
corrosion coupons was developed by Shigeta, et al.12 It relates weight loss to the amounts of
Na2O and K2O, and their ratio, the amount of SO2, and the inhibitive effects of CaO and MgO.
The expression developed was:

Weight loss (mg/cm2) = -0.08 + 0.13 (wt % acid soluble Na2O + K2O) (4-1)

x CNa2O/K2O x CCaO x CMgO x CSO2 x time (hrs.)

The four coefficients are read from Figure 4-9. The expression was validated for Type 347H
stainless steel and Alloy 17-14 CuMo in 6,000 hour probe tests. Expected weight loss for other
alloys was also determined and a ratio between the response of Type 347H and the alloy of
choice can be determined from Figure 4-7 as a function of expected metal temperature.

4-10
13622634
 Fireside Corrosion in Coal Fired Units

These indices can provide a qualitative guide to the effect of coal composition on the propensity
for fireside corrosion. This is one of the tools necessary to confirm the root cause of SH/RH tube
failures, as is outlined in Section 4.4 below.

Figure 4-9
Derivation of coefficients and effects of weight loss as a function of Na2O/K2O ratio,
content of CaO, MgO and SO2. Source: J. Shigeta, et al.13

4-11
13622634
Fireside Corrosion in Coal Fired Units

4.3.4 Effect of Chlorine in Coal

Chlorine appears to play a role in releasing the alkalis from the coal, but levels above about 0.2
weight percent appear to be required.3 Corrosion rate has been observed to be a strong function
of chlorine level for U.K. coals; increasing essentially linearly for superheater/reheater alloys
with increase in chlorine content of the coal over the range 0.1% to 0.5%.3 Past reasoning was
that the correlation with chlorine was indirect. The release of sodium contained in sodium
chloride in the coal triggered the release of potassium, so that the chlorine acted as a catalyst to
the formation of sulfate or trisulfate attack. There is also some evidence that the formation of
HCl in the flame can destroy the protective oxides such as Fe2O3.

A review of the effects of chlorine in U.K. coals was made in 1991.5 U.K. coals generally have
significantly higher levels of chlorine than typical U.S. coals of up to 0.7 weight percent
(averaging around 0.25%). The review includes details on fireside corrosion and points to the
linear dependence on chlorine for corrosion rates of austenitic superheater and reheater alloys.
Additional key aspects from that review are:
• Corrosion rates in plants burning high chlorine coals can reach well over 100 nm/hr (~ 35
mils/yr) at surface metal temperatures of 630°C (~ 1165°F).
• The source for chlorine in U.K. coals is mostly as weakly bound ions associated with organic
matter; with the balance being NaCl. Chlorine is released as HCl, about 80 ppm for each
0.1% Cl in the coal.
• HCl plays a key role in releasing both sodium and potassium from coal ash which leads to
aggressive sodium-potassium sulfate deposition.
• A correlation that relates corrosion rates to metal temperature, gas temperature, coal chlorine
content, leading/non-leading tube position and alloy composition has been developed based
on operating experience and plant corrosion measurements.14 It is of the form:

R = K L B (Cl-a) (Tm - b)c (Tg - d)e (4-2)

where R = corrosion rate
K, a, b, c, d, and e = empirically derived constants (For U.K. coals, a = 0.1,
b = 500, c = 4, d = 0, e = 2)
Cl = chlorine content (in wt %)
Tm = surface metal temperature (°C)

Tg = gas temperature (°C)

L = lead edge tube factor (1.0 for nonleading tube, 2.2 for leading tube)
B = alloy factor (unity for 300 series alloys)

The relationship has been found to hold for austenitic alloys 316, 321, 347 and Alloy E1250. The
B factor for more resistant alloys has also been evaluated and is: for Alloy 310 (B = 0.3) and
50% Cr-50% Ni (B= 0.1). The chlorine relationship for these more resistant alloys has not been
confirmed. Figure 4-10 shows a plot of corrosion rate versus chlorine content which has lead to

4-12
13622634
 Fireside Corrosion in Coal Fired Units

the observation that the increase in rate is approximately linear with chlorine content for U.K.
coals.

Figure 4-10
The linear dependence of corrosion rate on coal chlorine content for austenitic steels. 50
nm/hr is approximately 18 mils/yr. Source: W.H. Gibb and J.G. Angus15

It should be repeated that this correlation and the understanding of the effects of chlorine on high
temperature corrosion is only applicable to U.K. utility operations. A number of U.S. utilities
have been burning coals with elevated chlorine levels (up to 0.4 wt. %) from the Illinois Basin
and have not, to date, experienced the extensive corrosion that has been observed in U.K. plants.

Further field testing and laboratory work is currently being initiated to understand the basis for
the emphasis on chlorine in U.K. utility experience, and the extent, if any, of chlorine effects in
U.S. coals. There may be a difference in the way that the chlorine is bound in the coal and
released in the flame between U.S. and U.K. coals. Alternatively, the coal chlorine content may
simply be a useful measure of the availability in the coal of alkali metal salts that participate in
fireside corrosion of superheater/reheater tubes.

4.4 Possible Root Causes and Actions to Confirm

4.4.1 Introduction

Table 4-1 lists the primary root cause influences. In general, there are three groups of root
causes: (i) overheating of tubes, (ii) fuel factors, and (iii) combustion factors. Independent of the
root cause there are two actions that can be taken to confirm that the likely failure mechanism is
fireside corrosion, or to monitor its progress:

4-13
13622634
Fireside Corrosion in Coal Fired Units

(a). Wall thinning can be identified using non-destructive examination techniques, primarily
ultrasonic testing.

(b). Ash analysis to measure the melting point of ash layers, looking for low melting point
constituents. Thermogravimetry (ASTM E1131)16 or differential thermal analysis (ASTM
E794)17 can be performed to identify melting points of compounds in the deposit. Differential
scanning calorimetry can also be used to determine melting points of specific species.

The three groups of root causes are examined separately in the discussion that follows.
Table 4-1
Major Root Cause Influences, Confirmation and Corrective Actions

Major Root Cause Actions to Confirm Immediate Actions Longterm Actions and
Influences and Solutions Prevention of Repeat
Failure
4.4.1 Potential Actions for all Root Causes of Fireside Corrosion
(a) NDE measures • Choose repair • Perform remaining
(typically UT) to strategy based on life assessment.
identify wall severity of corrosion • Set up longterm
thinning. rate. monitoring and
(b) Ash and deposit • Implement longterm reevaluation
analysis to identify actions from choices program.
presence of low in Figure 4-12 in • Evaluate full range of
melting point conjunction with on- available options
constituents, going program of using roadmap in
particularly alkali remaining life Figure 4-12.
iron trisulfates. assessment and
monitoring.
4.4.2 Influence of Overheating of Tubes
(c) NDE of steamside • As above. • As above.
oxide thicknesses.
(d) Selective tube
sampling and
metallurgical
analysis to confirm
steamside oxide
and wall thickness.
(e) Monitoring of
thermocouples
installed across the
SH/RH outlet legs in
vestibule to identify
hottest platens
across the boiler.
4.4.2.1 Poor initial (f) Items (c) and (e) • As above, primary • As above, emphasis
design – choice of above. emphasis on will be on identifying
material upgrading to a more locations where
resistance material. material upgrading is
required.

4-14
13622634
 Fireside Corrosion in Coal Fired Units

Major Root Cause Actions to Confirm Immediate Actions Longterm Actions and
Influences and Solutions Prevention of Repeat
Failure
• May involve redesign
of circuit to extend
the use of the higher
grade material.
4.4.2.2 Poor initial (g) Evaluate • As above. • As above, emphasis
design extra gas- temperatures will be on identifying
touched length. across the element locations where
(via thermocouple material upgrading is
or steamside oxide required.
measurements) to
• May involve redesign
determine if
of circuit to extend
sections particularly
the use of the higher
near material
grade material.
changes are
running too hot .
4.4.2.3 Internal oxide (h) Items (c) and (d) • As in primary list • As in primary list
growth which occurs above. above (repairs, above, also see
during operation. followed by actions for the
longterm strategy) longterm
plus chemical overheating/ creep of
cleaning of tubes (Chapter 11).
steamside scale.
4.4.2.4 High (i) Monitor • As in primary list • Controlled with flow
temperature laning. temperatures as in above (repairs, distribution screens;
(e) above. followed by longterm in practice is difficult
(j) Laning can be strategy). to overcome
identified with cold because of high
air velocity temperatures in
technique. SH/RH.
• Review primary list
of alternatives in
Figure 4-12 for
options.
4.4.2.5 Tube (k) Visual examination. • Realign tubes, • Perform remaining
misalignment (out of implement on-going life assessment.
bank). program of • Set up longterm
remaining life monitoring and re-
assessment and evaluation program
monitoring.
4.4.2.6 Operational • Evaluate whether • Perform remaining
problems when coal operating procedures life assessment.
type is changed. such as sootblowing See discussion of
can be economically methods in Appendix
changed to protect C.
SH/RH tubes. • Set up longterm
monitoring and re-
evaluation program.
• Evaluate full range of

4-15
13622634
Fireside Corrosion in Coal Fired Units

Major Root Cause Actions to Confirm Immediate Actions Longterm Actions and
Influences and Solutions Prevention of Repeat
Failure
available options
using roadmap in
Figure 4-12
4.4.2.7 Rapid startups (l) Check startup probe • Modify startup • As above.
causing reheater to and that initial gas is procedures if
reach temperature limited to 1000°F feasible.
before full steam flow. (538°C) prior to RH
flow.

4.4.3 Root causes related to fuel factors

4.4.3.1 Change to fuel (m) Evaluate coal • As in primary list • As above, plus
with unusually corrosive composition using above (repairs, • Develop a fireside
ash, particularly those corrosivity index. followed by longterm testing program
with high S, Na, K or Cl. (n) Analysis for low actions). using guidance
melting point of ash provided in fireside
components using testing guidelines.18
probes.
• Investigate coal
(o) Analysis of changes with Coal
metallurgical cross- Quality Impact Model
sections, particularly (CQIM) 24-26 or
for CI, S, C, Na, and equivalent, including
K. economics
(p) Install continuous evaluation.
readout corrosion
sensors if unit
switches coal or
uses spot market
coal.
4.4.4 Root causes related to incomplete or delayed combustion.
(q) Monitor for levels of • As in primary list • Perform remaining
CO and O2. above (repairs, life assessment.
(r) Check for unburnt followed by See discussion of
startup oil. longterm actions). methods in
Appendix C.
• Evaluate full range of
available options
using roadmap in
Figure 4-12.
• Develop a fireside
testing program
using guidance
provided in fireside
testing guidelines.18

4-16
13622634
 Fireside Corrosion in Coal Fired Units

4.4.2 Influence of Overheating of Tubes

There are a number of underlying causes of overheated tubes any one of which can elevate the
tube metal temperature into the highly corrosive regime. Independent of the underlying cause,
the following are actions that can be used to confirm whether tube overheating is contributing to
a problem with fireside corrosion in superheater/reheater tubes:

(c). Oxide thickness measurements (ultrasonically) on upstream or downstream ferritic tubing

provides an indication of tube temperatures within the element which can be extrapolated to the
stainless steel tubing locations (see Appendix C).

(d). Selective tube sampling and metallurgical analysis for evidence of microstructural
degradation (see Appendix B).

(e). Monitoring of thermocouples permanently installed across the SH/RH outlet legs in the
vestibule can provide an indication of the hottest platens across the boiler. This then pinpoints
where steps (a) and (c) should be performed.

4.4.2.1 Poor Initial Design - Choice of Material

A particular aspect of poor design may be evident by accelerated corrosion at or near changes to
higher grade material, where the original design should have changed to the higher grade
material sooner. Figure 4-7 shows the relative corrosion rates of various alloys that could be
considered for upgraded materials. Actions to confirm are:

(f). See items (c) and (e) above.

4.4.2.2 Poor Initial Design - Extra Gas-Touched Length

Poor initial design can also result in extra gas-touched length for similar (maybe adjacent) tubes.
This means that one tube will run much hotter. Possibilities include accelerated corrosion of lead
or wrap-around tubes (leading tubes are most affected), or a tube that surrounded a radiant cavity
and therefore picks up excessive heat. A discussion of gas-touched length of tubes can be found
in Appendix C. Actions to confirm:

(g). Evaluate temperatures across the element via thermocouple or steamside oxide
measurements to determine if sections are running too hot, particularly near material changes.

4.4.2.3 Internal Oxide Growth Which Occurs During Operation

This buildup of internal oxide gradually increases the tube temperature until it rises into the
regime where the corrosion rate rapidly increases. Typically the steamside oxide scales on
austenitic stainless steel tubing have thickness of 0.08 mm (0.003 inch) or less even after
longterm service. These scales will not result in a significant increase in tube metal temperature.
In less frequent situations, the austenitic tubes will exhibit much thicker steamside scales and

4-17
13622634
Fireside Corrosion in Coal Fired Units

corresponding significant temperature increases. In these cases it is useful to estimate the current
tube metal temperature and relate this to the corrosion rate (e.g., see Equation 4-2 above).

(h). See items (c) and (d) above.

4.4.2.4 High Temperature Laning

Actions to confirm consist of:

(i). Monitor temperatures as in (e) above.

(j). Laning of gases can also be detected using a cold air velocity test (CAVT).

4.4.2.5 Tube Misalignment (Out of the Bank)

In addition to above indicated actions:

(k). Visual inspection can detect if there is a misalignment problem.

4.4.2.6 Operational Problems When Coal Type Is Changed

This may change the relative absorption patterns between furnace and convective sections.

4.4.2.7 Rapid Startups

Rapid startups can cause reheater to reach temperature or even overheat before full steam flow
and thus cause overheating.

An action to confirm is:

(l). Check startup probe and check that the initial gas temperature is limited to 1000°F (538°C)
prior to RH flow.

4.4.3 Root Causes Related to Fuel Factors

4.4.3.1 Change of Fuel to a Coal with an Unusually Corrosive Ash, Particularly Those
with High Content of S, Na, K or Cl

Actions to confirm include:

(m). Evaluate coal composition using corrosivity index for Eastern coals.11 As noted above, there
is no universally applicable, quantitative index to the corrosivity of coals, however for Eastern
U.S. coals, the Borio index can give at least a qualitative indication of whether a fuel change has
lead to a potentially more corrosive condition. Western coals, generally do not cause corrosion

4-18
13622634
 Fireside Corrosion in Coal Fired Units

problems of this type, because of low sulfur and high alkaline earth oxide levels, and are not
amenable to analysis by the Borio index.

(n). If liquid ash corrosion is suspected, thermogravimetry (ASTM E1131)16 or differential

thermal analysis (ASTM E794)17 can be performed to identify melting points of compounds in
the deposit. Deposits can be collected using high temperature probes; the methods and analysis
are described in reference 18.

(o). Energy dispersive x-ray and/or x-ray dot mapping of metallographic cross sections through
damaged tubes can be used to detect the presence and distribution of S, C, Na, P, and Cl. This
type of analysis is needed to confirm whether the corrosive wastage is simple molten salt attack
or whether a local reducing or carburizing condition is also leading to attack via sulfidation.

(p). Install corrosion sensors (continuous readout), especially if the unit often switches coal or
uses spot market coal.

4.4.4 Root Causes Related to Incomplete or Delayed Combustion

4.4.4.1 Actions to Confirm

(q). Monitor for levels of O2, H2S, and CO. High levels of CO (> 1%) and low levels of oxygen
(< 0.1%) near tube walls are of particular concern.3 The level of CO can also be measured in the
flue gas at the economizer outlet or after the ID fans. Oxygen levels can be sampled at the
economizer exit to provide an overall indication of the combustion process in the boiler.

(r). Check for unburned startup oil deposits on tubes.

4.5 Determining the Extent of Damage

Ultrasonic testing (UT) can be used to measure both wall thinning and steamside oxide thickness
in upstream or downstream ferritic tubing to infer tube metal temperatures in stainless steel
sections. Selective sampling is recommended to assess the condition of the tubing.

4.6 Repairs, Immediate Solutions and Actions

4.6.1 Need for Remaining Life Assessment

Any immediate solution or repair strategy should consider that a longer term remaining life
analysis methodology should also be implemented. Such programs are discussed in detail in
Appendix C.

4-19
13622634
Fireside Corrosion in Coal Fired Units

4.6.2 Repairs

If the corrosion rate has been relatively modest (< 25 nm/hr (~ 9 mils/yr)) and is likely to
continue to be so, an acceptable strategy is to retube with the same alloy and monitor closely the
wastage rate. Another alternative, but not a preferred approach, is the use of a palliative coating
or tube shields. An overview of options is provided under longterm actions.

Pad welds should definitely not be used as a repair measure because of the uncertainty of the
tube condition, such as the presence of creep cracks and their depth, and the conditions of the
internal tube surface.

For higher corrosion rates that are resulting in rapid wastage of the existing alloy, the
replacement should be with a more resistant material.

In addition to the upgrading of materials, there are some steps that can be taken over a short term
to minimize the severity of fireside corrosion. These include: taking interim steps to limit tube
metal temperatures such as by limiting main steam temperature or chemical cleaning to remove
steamside oxide, aligning tubes, increasing sootblowing operation, and others. However, because
of the range of options and the interconnectedness of them, a complete overview of all options is
included in the next section.

4.7 Life Assessment

A remaining life assessment is required to relate the rate of corrosion wastage to the desired life
or to determine the time available to implement the desired option. Therefore such an assessment
should be undertaken in parallel with any of the other options. A systematic program will
include: baseline measurement, monitoring rates of wastage, application and monitoring of
control measures, and assessment of the effects on remaining life.

Monitoring of the flue gas, metal and steam temperatures, combustion conditions, and fuel
composition should also be considered, as these can determine corrosion rates while the unit is
still on-line.3 Particularly important are step changes in key parameters.

A critical step in determining the remaining life of a tube and in choosing the optimal solution is
knowing the rate at which damage is accumulating. A key goal of the remaining life assessment
will be to establish the acceptable rate of wastage, given the desired remaining life of the unit
and an analysis of damage accumulation. As a rule of thumb, rates of wastage above 25 nm/hr (~
9 mils/yr., 0.22mm/yr.) indicate the need for some periodic activity such as annual inspection
and for the explicit determination of re-evaluation periods. Rates above 50 nm/hr (~ 18 mils/yr.)
indicate a serious decrease in life of the tubing; an estimate of remaining life should be made and
the appropriate actions taken if that life is unacceptable.

The life of austenitic superheater and reheater tubes is relatively simple to predict using
incremental damage models. There are two potential failure modes that should be considered: (i)
plastic instability (i.e., hoop stress > yield strength), and (ii) creep/stress rupture failure.

4-20
13622634
 Fireside Corrosion in Coal Fired Units

The plastic instability limit can be determined by incrementally evaluating the wall thickness,
applied hoop stress for measured or assumed fireside corrosion rates. It is useful to use the
following mean diameter hoop stress formula for these analyses:

σMDF = P (Di +t)/2t

Where:

P = steam gage pressure

Di = inside tube diameter

t = time dependent wall thickness = to – R (HRS)

to = initial wall thickness

R = fireside corrosion rate

HRS = hours of service exposure

The plastic instability life is consumed when the estimated mean diameter stress exceed the
materials yield strength at the service temperature. The calculations can be easily performed
using a spreadsheet and a 2000 hr incremental time interval. The elevated temperature yield
strength values for Type 304H, 321H, and 347H are shown in Figures A-8 through A-10
respectively. For tube metal temperatures in the range of 500oC to 650oC, an assumed yield
strength value of 120 MPa will provide adequate plastic instability life estimations for all three
of the austenitic tube alloys.

The creep/stress rupture life can be estimated with the same incremental (2000 hour per time
step) spreadsheet model. In the case of creep damage, it is useful to evaluate the minimum and
median life estimations. In either case, the creep life is estimated using a cumulative damage, life
fraction rule:

Dc = ∆HRS / trup(σMDF, Tmidwall)

Where:

Dc = Creep damage fraction = 1 at failure

∆HRS = time interval with assumed constant σMDF, Tmidwall; a time interval value of 2000
hrs generally provides adequate accuracy

trup(σMDF, Tmidwall) = estimated time to rupture for a constant mean diameter stress, σMDF,
and midwall metal temperature, Tmidwall. Section A.4.2 provides median and lower bound
time to rupture equations for the three common austenitic tube alloys. Appendix C
provides a detailed description of methods that can be used to estimate the midwall tube
temperature at any particular gas-touched location.

4-21
13622634
Fireside Corrosion in Coal Fired Units

Figures 4-11(a) and 4-11(b) present the results of plastic instability and creep/stress rupture life
estimations of a Type 304H reheater tube operating at a midwall temperature somewhat below
it’s design limit. The analyses were performed for assumed fireside corrosion rates of 7.5 nm/hr
and 25 nm/hr. One of the important features that these calculations show is that are the fireside
corrosion rate increases, the importance of the scatter in stress rupture properties significantly
diminishes. Clearly there would be very little value in performing stress or creep rupture tests to
improve the accuracy in the life estimations for the reheater tube with 25 nm/hr fireside
corrosion rate. Improved accuracy in the life predictions for situations with severe fireside
corrosion of superheater tubes or moderate to severe fireside corrosion of reheater tubes will be
obtained by learning as much as possible about the rate of fireside corrosion. It would be useful
to know:
• The temporal variation in wall loss rate (i.e., is the wastage rate linear with time?)
• The spatial variation in wall loss rate (i.e., from ultrasonic wall thickness surveys, evaluation
of selected tube samples, and modeling to remove all possible measurement errors or
uncertainties)
• The relationship between the wastage rate and the estimated midwall tube temperature.
Appendices B and C describe various methods to estimate the midwall tube temperature of
any desired gas-touched tube location.

4-22
13622634
 Fireside Corrosion in Coal Fired Units

(a)

(b)
Figure 4-11
Results of creep damage fraction, Dc, and plastic instability modeling of a Type 304H
reheater tube with an assumed fireside corrosion rate of (a) 7.5 nm/hr, and (b) 25 nm/hr.

Tables 4-2, and 4-3 provide a comparison of the estimated minimum, median, and plastic
instability lives for the reheater tube considered in Figures 4-11(a) and 4-11(b) and a Type 321H
superheater. As expected, the life of the superheater tube is somewhat less sensitive to fireside
corrosion because of its greater wall thickness.

4-23
13622634
Fireside Corrosion in Coal Fired Units

Table 4-2
Life Estimates of a 400 Circuit Reheater with 54 mm x 3.4 mm Type 304H Tubes and a 5.69
MPa Steam Pressure and 600oC Midwall Tube Temperature

Fireside corrosion Minimum (1st in 400) Median Stress Plastic Instability Life,
rate, R, nm/hr Stress Rupture Life, Rupture Life, khrs khrs
khrs
0 691 4,387 Infinite
5 258 426 458
10 164 236 232
15 122 164 154
20 98 126 116
25 82 104 92
50 46 54 46

Table 4-3
Life Estimates of a 500 Circuit Superheater with 54 mm x 6.0 mm Type 321H Tubes and a
12.76 MPa Steam Pressure and 590°C midwall tube temperature

Fireside corrosion Minimum (1st in 500) Median Stress Plastic Instability Life,
rate, R, nm/hr Stress Rupture Life, Rupture Life, khrs khrs
khrs
0 295 1,958 Infinite
5 214 590 726
10 168 352 364
15 140 254 242
20 120 200 182
25 104 166 146
50 66 90 74

4.8 Longterm Actions and Prevention of Repeat Failures

As noted above, it may not be possible to remove the root cause for many fireside corrosion
problems. Knowing how to minimize the wastage rate and the application of a predictive
remaining life assessment process, including periodic inspection and monitoring, are the keys to
economic handling of fireside corrosion problems.

Figure 4-12 outlines most of the available strategies for superheater/reheater fireside corrosion.
As shown in that figure, there are two primary and not mutually exclusive routes that can be
followed: (i) materials options and (ii) operating options. If an initial design condition is at the

4-24
13622634
 Fireside Corrosion in Coal Fired Units

root cause of a fireside corrosion problem there are three approaches that can be taken: correct
the problem; accept the fault and the correspondent corrosion rates and expect an increased
outage frequency; or accept the fault and seek a material solution.

The circled numbers in Figure 4-12 are used to identify options for the discussion that follows
and no ranking is implied; however, boxes with bold outline indicate those options which have
been the most successful.

Figure 4-12
Strategies for preventing repeat failures by fireside corrosion in superheater/reheater
tubes for coal-fire units.

4.8.1 Remaining Life Assessment (Option 1, Figure 4-12)

Remaining life assessment is discussed above in Section 4.7.

4-25
13622634
Fireside Corrosion in Coal Fired Units

4.8.2 Shielding (Option 2)

The most common temporary measure is the installation of shields to the leading edges of
affected tubes. This option is generally used for ferritic tubing but could, in unusual situations, be
used to combat localized ash corrosion of austenitic tubing. High chromium alloys such as Type
304 or 347 (austenitic stainless steels) or 309 (25 percent Cr) are often used. The shield is curved
to fit the tube surface and tack-welded in place. As they are not cooled, the metal temperatures of
the shields are above the range for extensive attack indicated in Figure 4-7.

4.8.3 Coatings (Option 3)

A variety of processes have been suggested for applying metallized coatings to tubes in situ as a
means of increasing corrosion resistance. The advantage of coatings is that very corrosion-
resistant materials can be applied at specific susceptible sites, so it is not necessary to replace
entire sections of tubing in order to overcome localized problems.

Among the coating methods that have been tried for fireside corrosion resistance are chromizing
and aluminizing. Flame- or plasma spraying, with and without subsequent heat treatment, have
also seen significant development work. The various coatings and application techniques have
had variable and generally poor results in U.S. utilities. The primary problems have been
adherence of the coating, the reproducibility of coating techniques, and the need to have very
well prepared surfaces. In general, coatings should currently be regarded as a quick fix which
will require continued maintenance.

A number of coatings have been tried for use in corrosion and erosion resistance.19, 20 The
primary use of coatings has been for the prevention of fireside corrosion in waterwalls of coal-
fired units. There is less field experience for coatings on superheater/reheater tubing in either
coal- or oil-fired units.

If used, important features of coatings include19:

• Importance of clear specification of the areas to be coated, materials to be used and thickness
required. Time available and other restrictions to site applications should be clearly specified.
• Importance of surface preparation to provide a contaminant-free, roughened surface.
• Importance of producing samples throughout the application as a check that process
parameters were producing consistent and acceptable quality.

Currently, either replacement in-kind or with a material of high corrosion resistance, depending
on the wastage rate which has been experienced, are preferred options to the application of
coatings. Weld overlays can also be considered as an option for providing significant life
extension.

4-26
13622634
 Fireside Corrosion in Coal Fired Units

4.8.4 Replacement with Same Material, Same Thickness (Option 4)

If the corrosion rate is only slightly higher than that required to reach the desired life as
calculated from the remaining life assessment, tube replacement can be made in-kind.9

4.8.5 Change to a More Resistant Material, "Monolithic" (Option 5)

An upgraded material can be used where unit operation is at high temperatures and fireside
corrosion remains a problem despite the attempts at other fixes. A number of alternatives have
been used successfully. The material chosen will depend on what is currently being used, and
what the desired resistance is to be.

If serious fireside corrosion of austenitic materials is taking place, then consideration may be
given to more resistant materials. There have been a number of laboratory and field
investigations of superheater/reheater corrosion of materials, particularly for advanced steam
conditions. The following brief review of some of the laboratory and field investigations
highlights some of the key issues as they might apply to materials choices in currently operating
plant.

A laboratory investigation of the response of eleven high-temperature alloys (7 cladding alloys

for co-extruded tubes and 4 chromized steels) to a simulated field environment provided the
following general conclusions6:
• Corrosion rates increased with increasing levels of alkali (Na2SO4 + K2SO4). The effect of
alkali level was more significant than SO2 for the more highly resistant materials.
• The relation between tube material, temperature and corrosion rate is shown for these
laboratory tests in Figure 4-13. The peak temperature for corrosion loss is different for
different alloys, in part because the temperature range in which the molten salt is stable
depends on the surface character of the metal. Corrosion products can also affect the melting
temperature of salt. Generally, temperatures tend to peak at a lower level for more highly
resistant alloys. For example, the peak temperature is about 700°C (~1300°F) for low-
resistance alloys like 17-14 CuMo and 600°C (~ 1100°F) for highly resistant alloys like Type
310 stainless steel.

4-27
13622634
Fireside Corrosion in Coal Fired Units

Figure 4-13
Effect of test temperature on corrosion loss for various alloys in 0.25 vol% SO and 5 wt%
alkali sulfate. Source: W. Wolowodiuk, et al.10

An extensive field evaluation of the resistance of various materials was also conducted using air-
cooled, retractable corrosion probes constructed of various materials. Corrosion probes were
exposed for 4,000 hours, 12,000 hours, and 16,000 hours at three facilities. Two of the units
burned eastern high-sulfur subbituminous coals with Borio Index values of 2.5 and 3.5
respectively; the third burned western low-sulfur coal for which the Borio index was not
applicable.

Maximum wastage was observed on the high sulfur subbituminous probes at the 10 and 2 o'clock
positions (flue gas impinged at 12 o'clock) with an appearance that confirmed the occurrence of
liquid-phase (coal-ash) corrosion. The following additional observations were noted:
• Resistance to coal-ash corrosion increases with an increase in chromium level in the alloy. A
plot of alloy chromium content versus average metal loss is shown in Figure 4-14. Results
show that a Cr content over 25% Cr was required for stable corrosion resistance.21 This has
important implications for the choice of replacement materials.
• The rate of corrosion experienced by Type 347 stainless steel could be reduced by a factor of
two by using Type 310Nb stainless (SA213-T310 CbN).
• Further, some of the materials which show the most promise have been tested only in short-
term tests, their longer-term performance is as yet relatively unknown.

4-28
13622634
 Fireside Corrosion in Coal Fired Units

Figure 4-14
Effect of chromium content on resistance to attack after 16,000 hours exposure at three
plants. Source: J.L. Blough, et al.6

As a general rule, the more resistant the material, the more expensive it will be.

For this reason, it may be preferable to use duplex tubing such as discussed next. Such co-
extruded tubes can supply more reliable resistance, but they are rather expensive due to the
complexity in manufacturing. There is an economic advantage in using high strength and
resistant monolithic tubing, such as SA213-T310 CbN. A trial calculation has suggested that the
increased cost of such tube materials is around 10% compared with the use of Type 347H for
superheaters of advanced boiler designs, i.e., steam condition of 566°C (1050°F) and 31 MPa.10

4.8.6 Change to a More Resistant Material, Composite (Such as Co-Extruded)

Tubing (Option 6)

Highly resistant materials such as alloy 310 and 50% Cr- 50% Ni can provide excellent
resistance, but since they have insufficient creep strength their use is in the form of co-extruded
tubing. Metallurgically-bonded processing provides superior tubes for use in utility boiler
applications when compared to mechanically bonded or bi-metallic tubes.22 Specifications for
such tubes have existed for several years.23

Co-extruded tubes have been used on a large-scale since 1976. Over 230 km of tubing was been
installed in twelve units. Such a materials solution has been found to be the most economical
approach to dealing with fireside corrosion problems which are caused by the aggressive nature
of the coals burned in U.K. boilers.22

4-29
13622634
Fireside Corrosion in Coal Fired Units

The primary material for the outer layer of superheater/reheater tubing has been Type T310
stainless (25 Cr 20 Ni) or a modified Type 310 made by increasing the silicon content to 0.75 -
1.5%, the chromium to a minimum of 25%, and niobium to a minimum of 8X the carbon level).
Where additional resistance has been needed a 50Cr-50Ni alloy has been used as the outer layer.
The inner layer has typically been Eshete 1250 (a high-temperature, high-strength stainless
steel). A typical application used 54 mm (2.13 in.) outside diameter 310 with 2.9 mm (0.11 in.)
thickness on 2.0 mm (0.08 in.) Eshete 1250 inner. Performance improvement factors have ranged
from 2-5 times better than the original installed materials, depending upon the particular
application.22

Tube bending characteristics of co-extruded material have been found to be identical to

monolithic materials.22 Weld procedures have generally matched weld metals to the base metal
to maintain property levels. Conventional welding techniques and normal quality control have
been found sufficient to ensure good weld quality. No preheat or post-weld heat treatments have
been required.

4.8.7 Fuel Change, Blending, Washing (Option 7)

Fuel changing expressly for the short-term control of fireside corrosion is unlikely; most often
the motivating factor is to reduce emission levels, particularly of sulfur. However, as discussed
above, changes that adjust the presence of corrosive species, including fuel changes, blending or
washing, or that increase the amount of mitigating species will have positive effects on the
corrosion rate. Changing coals to some of the Western varieties or lignite will offer these benefits
inherently. Analysis of the likelihood that a change in fuel or in fuel handling will be beneficial
can be made by evaluating the fuel composition with a corrosivity index. As discussed at length
above, however, there is still no absolute agreement on a particular index for all U.S. coal types.

A method of investigating coal changes, blending and washing, such as the application of the
Coal Quality Impact Model (CQIM)24-26 can be used. Such a method will provide information on
the total economic impact as well as on the potential for fireside corrosion.

4.8.8 Limit Tube Temperatures (Option 8)

Given the strong correlation between tube metal temperature and the potential for corrosion
damage, several strategies can be implemented that directly address root causes of
overtemperature in the tubes; they are discussed separately below.

4.8.9 Limit Steam Temperature (Option 9)

One means to minimize tube metal temperatures is to limit the main steam temperature; this has
historically been a primary strategy for the control of superheater/reheater fireside corrosion.
With tube surface temperatures approximately 30 to 80°C (~ 50 to 150°F) higher than steam
temperatures, limiting the steam temperature to around 538°C (1000°F), or in some cases to
566°C (1050°F) should result in an upper limit on tube temperature that falls below the region of
maximum corrosion as shown in Figure 4-7. This approach has governed operating temperatures
in the U.S. since the late 1950s when extensive problems with corrosion were experienced in

4-30
13622634
 Fireside Corrosion in Coal Fired Units

boilers operating at 621°C (1150°F) and 650°C (1202°F).27 It has been estimated that as a result,
only about 8 percent of U.S. coal-fired boilers are still operating at 1050°F (566°C) steam.

There are several problems with this approach. The primary drawback is that it is not efficient;
there is a severe heat rate penalty. Predicting exactly what temperature is required is difficult,
given the considerable scatter in the corrosion rate as a function of temperature and its peak.

Finally, in a field survey of 42 operating units2 the relationship between operating steam
temperature and fireside corrosion was found not to be direct, but to involve the consideration of
the numerous other factors such as coal ash composition and flue gas temperature.

For these reasons, it was recommended that if limits on main steam temperature were to be used
as a control strategy, an empirical approach be used to set steam temperature by monitoring tube
wastage rates and ensuring that the rate was less than 25 nm/hr (~ 9 mils/yr.).2

4.8.10 Redistribute Steam Flow (Option 10)

The technique of steam flow redistribution in superheaters has seen recent significant
development.28 Significant nonuniformities in tube temperature can exist across a superheater
from a variety of causes. This results in a wide range of tube metal temperatures and damage
accumulation by creep, as well as making these tubes more susceptible to fireside corrosion
because of the characteristic "bell-shaped" curve of wastage.

Redistribution of steam flow can serve to equalize the temperatures across the superheater. With
steam flow controllers (orifices) of different sizes and lengths in the tubes coming out of the inlet
header, the flow can be decreased in "cold" tubes, which increases their temperatures. This
change will also increase steam flow to "hot" tubes with a concurrent decrease of the metal
temperatures. In conjunction with selected tube plugging and replacement, such a method offers
significant opportunity for extending the life of superheaters and reheaters.

This method was designed primarily as a means of extending the creep life of tubes subject to
overheating. If overheating of selected tubes is at the root cause of fireside corrosion, the
technique could also be considered to selectively lower tube metal temperatures and thus
decrease fireside corrosion. The method is discussed in more detail in Chapter 11 on longterm
overheating/creep in superheater/reheater tubes.

4.8.11 Chemical Cleaning to Remove Steamside Oxide Scale (Option 11)

If excessive temperature in the superheater/reheater tubes is a contributing factor to the corrosion

process, and if that condition has been exacerbated by the presence of increasing steamside oxide
scale thickness, a solution involving chemical cleaning may be in order.

4-31
13622634
Fireside Corrosion in Coal Fired Units

4.8.12 Minimize Laning of Gases (Option 12)

Laning or channeling of gases through certain tube sections can lead to overheated tubes. This is
very difficult to overcome; it can be recognized and monitored using the cold air velocity testing
technique. However, the application of flow distribution screens, a means to control locally high
velocities, is made more difficult in the SH/RH by the high temperature environment.

4.8.13 Correct Misalignment of Tubing (Option 13)

If misalignment of tubes has resulted in localized fireside corrosion, this problem should be
corrected by realigning the affected tubes. This will decrease the number of tubes that are
directly exposed to the gas flow.

4.8.14 Change Frequency, Check Effectiveness of Sootblowing (Option 14)

This can be an important action because it can stop the formation of excessive deposits which
result in laning or channeling in adjacent areas.

4.8.15 Lowering Excess Air (Option 15)

Lowering excess air could reduce the amount of SO3 and possibly the stability of the alkali-iron
trisulfates; however, such as change may also increase waterwall fireside corrosion.

4.9 References
1
Blazewicz, A.J. and M. Gold, "High-Temperature Gas-Side Corrosion in Coal-Fired Boilers",
presented at the ASME Winter Annual Meeting, New York, December, 1979.
2
Williams, D.N., H.R. Hazard, H.H. Krause, L.J. Flanigan, R.E. Barrett, and I.G. Wright,
Fireside Corrosion and Fly Ash Erosion in Boilers, Research Project 2711-1, Final Report CS-
5071, Electric Power Research Institute, Palo Alto, CA, February, 1987.
3
Laxton, J.W., D.B. Meadowcroft, F. Clarke, T. Flatley, C.W. King, and C.W. Morris, The
Control of Fireside Corrosion in Power Station Boilers, Third edition, Central Electricity
Generating Board, 1987.
4
Paterson, S.R., T.A. Kuntz, R.S. Moser, and H. Vaillancourt, Boiler Tube Failure Metallurgical
Guide, Volume 1: Technical Report, Volume 2: Appendices, Research Project 1890-09, Final
Report TR-102433, Electric Power Research Institute, Palo Alto, CA, October, 1993.
5
Latham, E., D.B. Meadowcroft, and L. Pinder, "The Effects of Coal Chlorine on Fireside
Corrosion", Chlorine in Coal, J. Stringer and D.D. Banerjee, eds., Elsevier Science Publishers,
Amsterdam, 1991, pp. 225-246.

4-32
13622634
 Fireside Corrosion in Coal Fired Units

6
Blough, J.L., M.J. Krawchuk, G.J. Stanko, and W. Wolowodiuk, Superheater Corrosion: Field
Test Results, Research Project 1403-19, Final Report TR-103438, Electric Power Research
Institute, Palo Alto, CA, November, 1993
7
Kihara, S., A. Ohtomo, I. Kajigaya, and F. Kishimoto, "Recent Plant Experiences and Research
Into Fireside Corrosion in Japan", Werkstoffe and Corrosion, 39, 1988, pp. 69-83.
8
Koopman, J.G., E.M Marselli, J. Jonakin, and R.C. Ulmer, "Development and Use of a Probe
for Studying Corrosion in Superheaters and Reheaters", Proceedings of the American Power
Conference, Volume 1, 1959, pp. 236-245.
9
Holmes, D.R. and D.B. Meadowcroft, "Fireside Corrosion and Problems of Tube Life
Prediction", Symposium on Thermal Utilities Boiler Reliability, McMaster University, Hamilton,
Ontario, May, 1983.
10
Wolowodiuk, W., S. Kihara, and K. Nakagawa, Laboratory Coal Ash Corrosion Tests, Topical
Report GS-6449, Research Project 1403-19, Electric Power Research Institute, Palo Alto, CA,
July, 1989.
11
Borio, R., et al., "The Control of High-Temperature Fireside Corrosion in Utility Coal-Fired
Boilers", U.S. Office of Coal Research, Research and Development Department, No. 41,
April,1969.
12
Raask, E., Mineral Impurities in Coal Combustion - Behavior Problems and Remedial
Measures, Hemisphere Publishing Company, Washington, D.C., 1985.
13
Shigeta, J., Y. Hamao, H. Aoki, and I. Kajigaya, "Development of a Coal Ash Corrosivity
Index for High Temperature Corrosion", ASME Joint Power Conference, Portland, Oregon,
October, 1986.
14
Meadowcroft, D.B, in T.N. Rhys-Jones, ed., Surface Stability, The Institute of Metals, London.
15
Gibb, W.H. and J.G. Angus, J. Inst. Energy, Volume 56, 1983, p. 149.
16
American Society for Testing and Materials, Standard E1131-86, "Standard Test Method for
Compositional Analysis by Thermogravimetry, 1992 Annual Book of ASTM Standards: General
Methods and Instrumentation, Volume 14.02, American Society for Testing and Materials,
Philadelphia, PA, 1992.
17
American Society for Testing and Materials, Standard E794-85 (1989), "Standard Test Method
for Melting Temperatures and Crystallization Temperatures by Thermal Analysis, 1992 Annual
Book of ASTM Standards: General Methods and Instrumentation, Volume 14.02, American
Society for Testing and Materials, Philadelphia, PA, 1992.
18
Sotter, J.G., J.A. Arnot, and T.M. Brown, Guidelines for Fireside Testing in Coal-Fired Power
Plants, Research Project 1891-3, Final Report CS-5552, Electric Power Research Institute, Palo
Alto, CA, March, 1988.

4-33
13622634
Fireside Corrosion in Coal Fired Units

19
Morgan-Warren, E.J., "Thermal Spraying for Boiler Tube Protection", Welding and Metal
Fabrication, Jan/Feb, 1992, pp. 25-31.
20
Bennett, A.P. and M.B.C. Quigley, "The Spraying of Boiler Tubing in Power Plants", Welding
and Metal Fabrication, November, 1990, pp. 485-489.
21
Anon., "Approaches for Predicting Corrosion Rates of Superheater Tubes in Coal-Fired
Boilers", Materials & Components in Fossil Energy Applications, U.S. Department of Energy
and the Electric Power Research Institute, No. 106, October, 1993.
22
Flatley, T. and T. Thursfield, "Review of Corrosion Resistant Co-Extruded Tube Development
for Power Boilers", 1984 ASM Conference on Coatings and Bimetallics for Energy Systems and
Chemical Process Environments, held at Hilton Head, South Carolina, November 12-14, 1984.
23
CEGB Standard 680 710, Engineering Documents Unit, London.
24
Davidson, P.G, et al., Development and Application of the Coal Quality Impact Model:
CQIMTM, Research Project 2256-2, Final Report GS-6393, Electric Power Research Institute,
Palo Alto, CA, April, 1990.
25
CQIM Computer Code and Manuals: Volume 1: CQIM: User's Manual and Theory Manual,
Volume 2: Data Collection Manual and Theory Manual, Electric Power Research Institute, Palo
Alto, CA, December, 1990.
26
Pavlish, J.H., P.R. Miller, N.C. Craig, and A.K. Mehta, "CQIM - An Analytical Tool Used to
Evaluate Performance and Economic Issues", Proceedings: Ninth Annual International
Pittsburgh Coal Conference, October, 1992.
27
Reid, W.T., External Corrosion and Deposits - Boilers and Gas Turbines, Elsevier, New York,
1971.
28
Hara, K., C. Lee, R. Moser, T. Rettig, and K. Clark, Improved Superheater Component
Longevity by Steam Flow Redistribution, Research Project 1893-13, Final Report TR-101697,
Electric Power Research Institute, Palo Alto, CA, December, 1992.

4-34
13622634
5
FIRESIDE CORROSION IN OIL FIRED UNITS

5.1 Introduction

Fireside corrosion in oil-fired units (also termed "oil ash corrosion" or "liquid phase corrosion")
is generally confined to the higher temperature sections of the superheater (SH) and reheater
(RH) where metal temperatures can exceed 600°C (~ 1100°F).

Damage produced by fireside corrosion is often confused with that from longterm
overheating/creep (see Chapter 11). As several of the key features of the two are similar, some
care is needed to correctly identify the primary mechanism, and thus take the proper corrective
action. Fireside corrosion of SH/RH tubes in coal-fired units is discussed in Chapter 4.

5.2 Features of Failure and Typical Locations

5.2.1 Features of Failure

Macroscopically, tube wastage will often be evident and manifested as undulations and
unevenness of the tube surface. A typical cross-section showing this wastage and the presence of
significant deposits can be seen in Figure 5-1. Tube thinning will be most evident around the
edges of the deposit.

Greatest wall loss will generally be seen in tubes that have been operated at the highest
temperatures over a period of time. The measurement of steamside oxide thickness on the
sections of ferritic tubing (by ultrasonic or laboratory measurement) upstream or downstream of
the austenitic tubing will usually confirm this observation.

On austenitic steels, the magnetite and spinel oxides become non-adherent when they reach a
thickness of about 0.2 mm (0.008 in.) and tend to spall off along with any deposits that have
formed when the boiler cools. Bare metal can result and lead to rapid corrosion rates. Any oxide
scale that is thicker than about 1 mm (0.04 in.) should be considered as evidence of a fast
corrosion rate.1 There may also be localized, overlapping pits 0.5 - 2 mm (~ 0.02 - 0.08 in.) in
diameter.

5-1
13622634
Fireside Corrosion in Oil Fired Units

Figure 5-1
General appearance of a 9% Cr final superheater tube containing fireside corrosion
deposits (top of figure) in an oil-fired boiler after 50,000 hours service. The bottom of the
figure shows a section of tubing having been acid cleaned to removed the deposits and
the ring section taken through the cleaned section shows the general appearance of tube
wastage on the outside surface. Source: J. Hickey, Irish Electricity Supply Board

Unburned oil on the upstream surfaces of tubes, such as caused by excess oil during startup, is a
source of carbon contamination and subsequent carburization of the coated surface. This is
particularly critical for austenitic tube materials which have a higher solubility for carbon than
ferritic materials.1

The tube will usually be coated with a multilayered fireside scale and ash deposits typically
consisting of two layers. Deposits adjacent to the metal oxide are the ones that directly influence
corrosion of the substrate. A black, glossy inner layer is typical and may be superficially similar
to that which occurs as a result of longterm overheating. If this layer is glassy or shows signs of
having been molten against the tube metal, then a very fast corrosion rate has probably occurred.
The absence of a layer of protective oxide adjacent to the tube surface is indicative of the fastest
corrosion rates.

A hard, brittle, and porous outer layer may have alternating dark/black/blue and light bands.

Fireside corrosion will be primarily distinguished by the presence of low melting point ash
compounds in these deposit layers. The fireside scale and ash should be examined
metallographically and chemically for the presence of low melting point constituents. Note that
oil ash deposits tend to be soluble in water and are more easily removed than coal ash deposits.

For high vanadium residuals these constituents will be vanadium oxides and/or vanadium sodium
complexes such as those listed in Table 5-1 and the eutectics shown on the V2O5- Na2O phase
diagram in Figure 5-2. For low vanadium residuals, these compounds will be essentially sulfatic
(sodium sulfates Na2SO4 and derivatives) with small amounts of NV6 and nickel vanadates. An

5-2
13622634
 Fireside Corrosion in Oil Fired Units

example of the distinctive layering and the presence of an alkali vanadyl vanadate in the inner
layer as detected by energy dispersive x-ray is shown in Figure 5-3.

Removal of the fireside scale and ash deposit will reveal an undulating surface and distinctive
longitudinal grooving and pitting appearance ("alligator hide") as was shown in Figure 4-5,
Chapter 4. Note that the appearance of alligator hide is indicative but not definitive as it can also
occur if the metal has been thermally shocked such as by sootblower condensate.3 Final failure
will typically be as a longitudinal crack associated with the alligator hide. If failure has been
longterm the fracture surface will appear brittle, rapid corrosion rates will result in fracture
surfaces that are ductile.

Table 5-1
Melting Point of Some Commonly Observed Slag Deposits

Compound Melting Point, °C

V205 670

V203 1,970

V204 1,970

Na20 • V205 605

2Na20 • V205 650

3Na20 • V205 850

Na20 • V204 • 5V205 625

Na20 • V204 • 11V205 575

Na2S04 884

Na3Fe(S04)2 624

Note: See also the V205-Na20 phase diagram (figure 34-2) which shows eutectics
down to 525°C (977°F).

5-3
13622634
Fireside Corrosion in Oil Fired Units

Figure 5-2
Equilibrium phase diagram for the V2O5-Na2O system showing PO2 dependence. (N)
represents Na2O and (V) is V2O5. Source: J.R. Wilson2

5-4
13622634
 Fireside Corrosion in Oil Fired Units

Figure 5-3
(a) Shows a secondary electron image of an oil ash scale and deposit on 2.25Cr-1Mo steel
reheater tube that exhibited oil ash corrosion. (b, c, and d) Show energy dispersive x-ray
spectra from three layers in the ash. The spectra indicate that the inner layer (b) may be
an alkali vanadyl vanadate. Absence of significant iron in the inner layer suggests that the
vanadium compound is either directly attacking the tube material or is fluxing away the
indigenous iron oxide as rapidly as it forms. The middle layer (c) is probably iron oxide
which has precipitated at the external surface of the molten layer. The outer layer (d) is
comprised of ash deposit constituents.

5-5
13622634
Fireside Corrosion in Oil Fired Units

The key differences between fireside corrosion and longterm overheating/creep are highlighted
in Table 5-2. The most distinguishing characteristic will be the presence of low melting ash
compounds in the deposits formed as a result of fireside corrosion, and the extent of wall
thickness loss.

Table 5-2
Comparison of Characteristics of Longterm Overheating/Creep, Short-Term Overheating,
and Fireside Corrosion (Oil-Fired Units) In Superheater/Reheater Tubing

Characteristic Longterm Short-Term Fireside Corrosion

Overheating Overheating
Fracture Surface and • Generally thick- • Usually thin- • Tube wastage in
Appearance of Failure edged, brittle final edged, ductile high
failure. final failures. temperature/heat
• Generally • Swelling of tubes flux portions of the
accompanied by without tube.
external tube ovalization. • Longitudinal
wastage at the 10 • Fish-mouth” cracking, final
o’clock and 2 appearance of failure can be
o’clock positions. tube rupture. (bun not
necessarily) by
overheating
Internal Scale? Yes, generally Not necessarily thick. Yes, particularly if
extensive, multi- Depends on age of tube metal over
laminated and tube at failure heating was a root
exfoliating cause
External Scaling? • Yes, thick, Not necessarily thick. Yes, with multi-layers:
laminated and (i) a hard brittle and
often porous outer layers
longitudinally which may have
cracked. alternating dark
blue/black bands and
• Usually two layers
(ii) a black, glossy
– (i) a hard porous
inner layer strongly
outer layer with
bonded to tube; has
composition
an appearance of
typically that of
molten deposits and is
flyash, and (ii) a
shiny.
black glossy inner
layer, mostly
oxide but may
contain some
sulfates and
sulfides of iron.
Outside surface Characteristic Swelling, stretch Characteristic
appearance after longitudinal grooving marks on tube metal. undulations or
removal of and pitting (“alligator unevenness of
scale/deposits hide”). surface. In worst
areas there might be
some “alligator hide”
and longitudinal
cracking

Composition of Does not contain low Not relevant. Does contain low

5-6
13622634
 Fireside Corrosion in Oil Fired Units

Characteristic Longterm Short-Term Fireside Corrosion

Overheating Overheating
External melting point ash melting point
Scales/Deposits compounds such as compounds. In high
alkali iron sulfates. vanadium oil will be
vanadium sodium
complexes. In low
vanadium oil will be
sulfuric (sodium
sulfates and
derivatives).
Wall Thinning? Typically wastage Only because of Primary feature of
flats at 10 o’clock and bulging of tube failure, worse on
2 o’clock positions material. portion of tube subject
caused by to high temperatures
accelerated oxidation and/or high heat flux.
There is always a
layer of oxide
adjacent to the tube
Ratio of wall loss to Typically less than 3:1 Not relevant. Typically greater than
steamside oxide 3:1; for ratios greater
thickness than 5:1 fireside
corrosion or erosion is
the dominant
mechanism.
Tube Material Yes, generally Depends on the If overheating has
Degradation extensive signs of material and the been a problem, yes;
overheating and/or of maximum otherwise, no.
creep damage, temperature reached Fireside corrosion can
particularly near to the For example, for the occur in a tube at
crack tip. Creep voids most rapid design temperatures.
will not be found overheating failures, Can be a carburized
removed from crack there will be relatively band adjacent to the
tip. little microstructural fireside deposits.14
change.
Change in material Localized softening Localized hardening Hardness change is
hardness near the rupture is near the rupture is not necessary; if there
typical. likely. has been no
overheating, there will
be no change in
hardness. corrosion,
then an increase in
hardness may be
observed.14

5.2.2 Locations of Failure

Typical regions of the boiler subject to tube failures by fireside corrosion are shown in Figure 5-
4. These locations are basically the same as those where longterm overheating/creep and fireside
corrosion in coal-fired units can occur. However, in oil-fired units, fireside corrosion tends to

5-7
13622634
Fireside Corrosion in Oil Fired Units

occur in higher temperature components, particularly those tubes with metal temperatures in
excess of about 600°C (~ 1110°F). As a result, failures may be more prevalent in outlet sections
and can occur in either austenitic materials or in ferritic materials exposed to the higher
temperatures.

Figure 5-4
Typical boiler locations where oil ash fireside corrosion can occur.

Specific locations at the highest risk therefore include:

• Leading sides of all tubes in pendant platens, especially hottest (leading) tubes, and steam
outlet tubes
• Tubes out of alignment that act as leading tubes
• In the outlet (final) sections towards the header because these are at the highest temperatures.
• Just prior to a change of material, e.g., in T22 just prior to the austenitic material, as the
lower Cr content material may be operating above its design point.
• Wrapper tubes
• Tubes that surround a radiant cavity (i.e., they may pick up more heat)
• At bottom bends of platens especially those facing the fireball.
• Tubes with a longer gas touched length (GTL). GTL is the distance measured along the tube
circuit from the inlet header to the point of corrosion.
• Spacers and uncooled hangers, and the fins and studs on tubes.

5-8
13622634
 Fireside Corrosion in Oil Fired Units

5.3 Mechanism of Failure

5.3.1 Introduction

Low melting point compounds produced by the firing of residual oils are aggressively corrosive
to the protective oxides which form on SH/RH tubes. As liquid phases begin to deposit and
solidify on tube surfaces, the temperature of the outside surface increases, leading to additional
deposition and melting of the deposits.

The primary compositional determinants from the fuel are the levels of vanadium and sodium in
the residual oil. The effects of two general classifications of residual oils are described here: high
vanadium, low sodium fuel typical of Venezuelan sources, and low vanadium, low sodium fuels
typical of some of those from the Persian Gulf. Typical compositions of each are shown in Table
5-3 along with a typical Mexican fuel oil.14

Table 5-3
Impurity Levels of Residual Fuel Oils

Fuel Source Vanadium, ppm Sodium, ppm Sulfur, Wt %

Venezuelan 280-300 10-20 2.5

Persian Gulf 45-50 35-70 2.5

Mexican14 150-340 1-100 3.2-4.5

For either fuel, corrosive slag occurs when two preconditions have been met: (i) operation occurs
with a residual oil that forms low melting point compounds, particularly those containing
vanadium and sodium, as well as sulfur, and (ii) tube metal temperatures that are above 600°C (~
1110°F). Final failure occurs as a stress rupture when thinned tube walls are no longer able to
contain the requisite pressure.

One of the primary effects of oil-firing is that there is relatively little ash produced (typically
around 0.2%) in comparison with coal-firing (around 20%) and therefore the beneficial effects of
ash components are not present. It has been reported that when both oil and coal are fired, that
the coal ash can dilute the vanadium content of deposits produced.1

Greater tube wastage has been observed if the tubes become carburized during service.15

5.3.2 Effect of Fuel Composition on the Rate of Corrosion

5.3.2.1 High Vanadium Residual Oils
In residual oils containing a high level of vanadium, it is generally agreed that the molten
vanadates increase the corrosion of boiler tubes by fluxing the normally protective metal oxide
layers, or by increasing the diffusion of oxidant to the protective layer.4 Although the exact

5-9
13622634
Fireside Corrosion in Oil Fired Units

mechanism of attack is still being refined, Figure 5-5 shows a ten-step process of oxygen
transport that illustrates the relevant factors.

Figure 5-5
Possible reactions that take place in liquid vanadate (oil-fired) corrosion. Source: J.R.
Wilson2

Table 5-1 summarizes the most commonly observed tube deposits that result from high
vanadium oils. Many have melting points that are below the expected temperatures in the hotter
sections of the SH/RH. Vanadium pentoxide (V2O5) is the highest oxidation product of vanadium
and is listed for completeness although it is rarely observed in operating boilers. Sodium vanadyl
vanadate (Na2O • V2O4• 5V2O5) is one of the many complex compounds usually present in the
ash and is the most commonly reported because there is X-ray data for it. In fact, it is not present
in the corroding deposit because it absorbs oxygen to become Na2O • 6V2O5 (referred to as NV6)
depending on the Na:V ratio.

Figure 5-2 is the equilibrium diagram for the Na2O - V2O5 system showing how the melting
points vary for these families of compounds. The dotted line on the figure illustrates the liquidus
for 0.2 atmosphere oxygen partial pressure; the solid line for 1.0 atmosphere. The substantial
differences lead to a corresponding significant difference in corrosion rates in pressurized fired
equipment.2 It also indicates how operation at low levels of excess oxygen can help mitigate the
effects of fireside corrosion by raising the melting point of the vanadates.

5-10
13622634
 Fireside Corrosion in Oil Fired Units

The greatest corrosion is not related to the absolute levels of V and Na but the ratio of their
oxides and temperature.13 The fastest corrosion rate (for temperatures up to about 900°C (~
1650°F)) is at a ratio of V2O5: Na2O of 6:1 (NV6), which corresponds to a Na:V ratio of about
0.1:1. Thus for a typical Venezuelan residual containing 300 ppm vanadium, only about 30 ppm
sodium is required to form a very corrosive medium. Other oxides at low concentrations exert
little influence on corrosion rates. The corrosivity can also depend on the sulfur content of the
oil-ash deposit in cases where the sulfur content in the oil is greater than 3%.14 The presence of
refractory oxides (MgO, CaO, Al2O3, and SiO2) inhibit corrosion when present at high
concentrations in the ash, e.g., a ratio of 2:1, additive:ash by weight.2

5.3.2.2 Low Vanadium, Low Sodium Residuals

With a typical Persian gulf residual oil, for example, containing 50 ppm V, 70 ppm Na and 2.5
weight percent S, the deposits or slags are essentially sulfatic (sodium sulfates Na2SO4 and
derivatives) with small amounts of NV6 and nickel vanadates. The fireside corrosion rate is
controlled by the dissolution of otherwise protective scales in the liquid alkali sulfates.

Laboratory and corrosion probe trials have shown that the deposition rate of sodium is
proportional to its concentration, independent of the presence of vanadium. The deposition of
vanadium increases with vanadium content in the flue gas, but large amounts of sodium can
depress the rate. No realistic reductions in the fuel oil impurities will markedly improve the rate
of corrosion which is of the same order as those for vanadate corrosion. The same sharp increase
in rate for austenitics above 600°C (~ 1110°F) is evident.

5.3.3 Effect of Tube Temperatures on the Rate of Corrosion

Because of the much lower operating temperatures, fireside corrosion in the waterwalls of oil-
fired units is unlikely, however, superheater/reheater temperatures are in exactly the right range
during operation to experience a problem. The higher the temperature, the faster the corrosion
rate in these temperature ranges. As a result, if steamside oxide buildup leads to higher tube
temperatures, or if unit firing rates are increased to compensate for insulated tubes, the resultant
increase in tube temperature can lead to an outbreak of tube failures by this mechanism.

5.3.4 Effect of Tube Material Composition

The corrosion rate increases sharply for austenitic materials (less for ferritic) between 600 and
650°C (~ 1110 to 1200°F). With a Venezuelan residual, for temperatures less than 600°C
(~1110°F), the rate is less than 50 nm/hr (~ 18 mils/yr), whereas at 650°C (~ 1200°F), it can be
up to 200 nm/hr (~ 70 mils/yr). Clearly, these higher rates can cause a considerable problem if
left unchecked.

High chromium materials generally have better corrosion resistance than ferritic materials. Type
310 has about a factor of three better resistance than ferritic steels; an alloy of 50Cr 50 Ni, which
has seen significant development for applications where corrosion has proven to be particularly
troublesome, has extremely low corrosion rates particularly below 700°C (~ 1300°F). Its use is

5-11
13622634
Fireside Corrosion in Oil Fired Units

generally in duplex or coextruded tubing so that the superior corrosion resistance is combined
with a creep resistant core.

5.3.5 Effect of Unit Operation on Corrosion Rates

The amount of excess air is a contributing factor to the rate of corrosion. In high vanadium
residual oil, a level of excess air that is just above the minimum stoichiometric requirement will
reduce the corrosion potential by preventing the oxidation of vanadium to its highest oxidation
state (V2O5). The higher melting point oxides V2O3 and V2O4 will be more likely to form, and
will not deposit as corrosive compounds. However, the addition of Mg additives will probably
still be required to reach acceptable corrosion rates. The effect of such additives is described in
the next section. Note that ashes containing vanadium complexes are toxic and are slightly water
soluble which may lead to groundwater contamination if used for land fill.

In the case of low vanadium residuals, reasonable success has been obtained by the correct
control of oxygen to just above stoichiometric combustion requirements. Neither further excess
oxygen reductions, nor the use of Mg additives has had significant additional benefit. These
effects tend to confirm that the active process is corrosion due to sulfatic compounds; thus
control of SO3 production is an important criterion.

5.3.6 Effect of Additives on Propensity for Corrosion

A variety of additives, most notably magnesium, but also including manganese, aluminum,
silicon, and calcium, alone or in combination with each other, have been used to control
corrosion. It is important to understand how they affect the mechanism so that proper additive
choice can be made.

The three objectives of an additive are (i) to raise the fusion point of the ash, (ii) to combine
chemically with all the corrosive agents present, and (iii) to not cause plugging or wear in fuel
systems. The use of magnesium, as MgO, is based on increasing the melting point of the
vanadium complexes and to eliminate the formation of low melting point eutectics. Table 5-4
shows the success of that strategy as the melting point of typical complexes plus MgO increases.
Figure 5-6 shows the phase diagram for the addition of MgO. The aim is to form 3MgO•V2O5
which has a melting point of 1071°C (1960°F). The situation is complicated by the presence of
Na2O which reduces the melting point, or requires more MgO to control.

5-12
13622634
 Fireside Corrosion in Oil Fired Units

Table 5-4
Melting Points of MgO V2O5 Compounds

Compound Melting Point, °C

2MgO • 3V2O5 640

MgO • V2O5 742

2MgO • V2O5 980

3MgO • V2O5 1071

Figure 5-6
Equilibrium phase diagram for the V2O5 – MgO system showing that there is no PO2
dependence. Source: J.R. Wilson2

To prevent corrosion, it is necessary that the deposit on the tubes has a melting point above the
metal temperature. It is difficult to control vanadate corrosion by the use of additives without
accumulating large volumes of modified fly ash; thus, it is important to add only enough MgO to
counteract the corrosive vanadate. This entails fuel and slag analysis, reference to the phase
diagrams for the expected behavior of the proposed combinations, and analysis of oil and

5-13
13622634
Fireside Corrosion in Oil Fired Units

additive flow rates. Excessive magnesium slag buildup will cause sintering and fusing of the
outer layers, which may be more difficult to remove by sootblowing or water washing.

As noted above, the benefits of Mg additives for use with low vanadium residuals has not been
as marked as that for the high vanadium oils.

5.4 Possible Root Causes and Actions to Confirm

5.4.1 Introduction

Table 5-5 lists the primary root cause influences. In general there are three groups of root cause
influences: (i) fuel factors, (ii) excessive tube metal temperatures, and (iii) the effects of
operating conditions. Independent of root cause, the two most important actions are:

(a). Analyze the composition and melting point of deposits. A key confirmation that fireside
corrosion is the active mechanism is the presence of low-melting point ash constituents.
Metallographic sampling and analysis will be key identification tools. Chemistry and
constituents of the deposits can be determined using EDX mapping, spectrochemical analysis, x-
ray diffraction or atomic absorption spectroscopy. Melting temperatures of ash are measured by
the use of thermogravimetric analysis and/or differential thermal analysis.

(b). Monitor progression of wall thinning on a periodic basis. Significant wall loss predominantly
from the tube OD is an indication that fireside corrosion rather than longterm overheating/creep
is dominant.

5-14
13622634
 Fireside Corrosion in Oil Fired Units

Table 5-5
Major Root Cause Influences, Confirmation and Corrective Actions

Major Root Cause Actions to Confirm Immediate Actions Longterm Actions and
Influences and Solutions Prevention of Repeat
Failure
5.4.1 Potential Actions for all Root Causes of Fireside Corrosion
(a) Ash and deposit Choose repair strategy Perform remaining life
analysis to identify based on severity of assessment.
presence of low corrosion rate. Continual check on the
melting point Implement longterm use of Mg-based
constituents, strategy from additives.
particularly choices in Figure 5-7 Optimization of excess
vanadium/vanadium in conjunction with oxygen levels.
-sodium and sodium on-going program of Set up longterm
sulfate complexes. remaining life monitoring and
(b) NDE measures assessment and reevaluation
(typically UT) to monitoring. program.
identify wall thinning
Evaluate full range of
and steamside
available options
oxide scale buildup.
using roadmap in
Figure 5-7.
5.4.2 Influence of Oil Composition
(c) Monitor oil As above, plus As above with emphasis
corrosiveness using On-going consideration on additives.
corrosion or of the use of Mg-
deposition probes. based additives.
(d) Analyze ash
deposits as in
Item (a) above.
5.4.3 Influence of Overheating of Tubes
5.4.3.1 Excessive (e) NDE of steamside Choose repair strategy Perform remaining life
temperatures caused by oxide thicknesses. based on severity of assessment.
steamside oxide (f) Selective tube corrosion rate Consider program of
buildup. sampling and Implement longterm periodic chemical
metallurgical strategy from cleaning. See also
analysis to confirm choices in Figure 5-7 options for longterm
steamside oxide in conjunction with over-heating of tubes
and wall thickness. on-going program of (Chapter 11).
(g) Monitoring of remaining life Set up longterm
thermocouples assessment and monitoring and
installed across the monitoring reevaluation
SH/RH outlet legs in Institute periodic program.
vestibule to identify chemical cleaning. Evaluate full range of
hottest platens available options
across the boiler. using Figure 5-7.

5.4.3.2 Excessive (h) For high Modify operation to Evaluate full range of
temperatures as caused temperature laning: correct the specific available options
by operating conditions. monitor problem. using Figure 5-7.
High temperature laning temperatures as in Implement longterm

5-15
13622634
Fireside Corrosion in Oil Fired Units

Major Root Cause Actions to Confirm Immediate Actions Longterm Actions and
Influences and Solutions Prevention of Repeat
Failure
of gases (g) above and strategy from
Changes in absorption consider the use of choices in Figure 5-7
patterns between the cold air velocity in conjunction with
furnace and test (CAVT). on-going program of
connective sections. (i) For RH remaining life
RH overheating overtemperature assessment and
because of rapid during start monitoring.
startups sequences: check
Tube misalignments the startup probe
and limit
temperatures to
538°C (1000°F)
prior to RH flow.
(j) Visual inspection
can be used to
detect tube
misalignments.
5.4.4 Influence of Operating Factors
5.4.4.1 Operation with (k) Check operating Modify operating
high levels of excess logs for typical procedures, if
oxygen and/or periods excess oxygen economically
of very low excess levels. feasible to reduce
oxygen. levels of excess
oxygen.
Implement longterm
strategy from
choices in Figure 5-7
in conjunction with
on-going program of
remaining life
assessment and
monitoring.
5.4.4.2 Poor (l) Check sootblowing Evaluate whether Perform remaining life
sootblowing operations frequency, operating procedures assessment. See
effectiveness, and such as sootblowing discussion of
superheat level of can be economically methods in Appendix
blowing medium. changed to protect C
SH/RH tubes. Set up longterm
Implement longterm monitoring and
strategy from reevaluation
choices in Figure 5-7 program.
in conjunction with Evaluate full range of
on-going program of available options
remaining life using Figure 5-7.
assessment and
monitoring.

5-16
13622634
 Fireside Corrosion in Oil Fired Units

5.4.2 Influence of Oil Composition

There is a significant effect of oil composition on corrosion rate as discussed at length in the
mechanism section above. Changing to a more aggressive fuel can significantly increase
corrosion damage. Actions to confirm that this is a primary factor in the appearance of fireside
corrosion are:

(c). Monitor oil corrosiveness by placing controlled-temperature corrosion probes or deposition

probes into susceptible locations in the SH/RH.

(d). Analyze ash deposits using methods outlined in (a) above to periodically determine whether
significant deposition of low-melting point constituents has begun or has accelerated.

5.4.3 Influence of Tube Metal Temperatures

Unfortunately, the higher temperature sections of the SH/RH are well within the temperature
regime of significant corrosion susceptibility. Furthermore, increases in tube metal temperatures,
such as caused by the buildup of steamside scale, can increase the rate at which corrosion occurs
and increase the number of tubes at risk. As deposits form on the external tube surface, they will
insulate the tube metal but result in higher ash temperatures. Combustion conditions that lead to
higher flue gas temperatures can cause localized tube overheating. The influences are reviewed
below.

5.4.3.1 Excessive Temperatures Caused by Steamside Oxide Buildup

Actions to confirm that steamside oxide buildup and the resultant overheating is a contributing
factor to the appearance of fireside corrosion include:

(e). Oxide thickness measurements (ultrasonically) in the ferritic tubing upstream and
downstream of the austenitic tubing provides an indication of tube temperatures within the
element (see Appendix C).

(f). Selective tube sampling and metallurgical analysis of ferritic tubing can be used to confirm
the results of the NDE measures of steamside oxides. Samples of selected austenitic tubing can
be used to confirm the characteristics of the external deposit as outlined in item (a) above and to
determine if creep cavitation has initiated (see Appendix B).

(g). Monitoring of thermocouples permanently installed across the SH/RH outlet legs in the
vestibule can provide an indication of the hottest platens. This action can help prioritize where
inspections and sampling should occur.

5.4.3.2 Excessive Temperatures Caused by Operating Conditions

A variety of operating conditions might cause excessive tube temperatures: high temperature
laning, changes in absorption patterns between the furnace and convective sections, rapid

5-17
13622634
Fireside Corrosion in Oil Fired Units

startups that cause the reheater to reach temperature before full flow starts, tube misalignments,
etc. The actions to confirm these conditions are:

(h). For high temperature laning: monitor temperatures as in (h) above. Laning of gases can also
be detected using a cold air velocity test (CAVT).

(i). For reheater overtemperature during start sequences: check the startup probe and limit initial
gas temperature to 538°C (1000°F) prior to RH flow.

(j). For tube misalignment (out of the bank): visual examination can be used to detect a problem.

5.4.4 Influence of Operating Factors

5.4.4.1 High Levels of Excess Oxygen and/or Periods of Very Low Excess Oxygen

The amount of excess air is a contributing factor to the rate of corrosion. In high vanadium
residual oil, a level of excess air that is just above the minimum stoichiometric requirement (<
1% excess oxygen) will encourage the formation of higher melting point constituents that will
not deposit as corrosive compounds. There is relatively less effect of excess air levels when
burning low vanadium residuals. Actions to confirm include:

(k). Check operating logs for typical excess oxygen levels.

5.4.4.2 Poor Sootblowing Operations

Typically, the first layer of deposit on clean tube surfaces is dry and loose because the tube metal
temperature is low enough that the ash deposits are in solid form. Proper sootblowing can easily
remove these early deposits. However, if sootblowing operations are executed improperly,
deposits can remain on tubes, raising the outside deposit temperature, resulting in the deposition
of liquid-phase ash which is difficult to remove, and as described above, begins the rapid
corrosion of tube surfaces. An action to confirm is:

(l). Check sootblowing frequency, efficacy and superheat level of blowing medium.

5.5 Determining the Extent of Damage

Ultrasonic testing (UT) can be used to measure wall thinning. Steamside oxide thickness in the
ferritic tubing upstream and downstream of the austenitic material can be used to estimate tube
metal temperatures. Locations should be chosen which are the most susceptible to fireside
corrosion as described in Section 5.2.2 above. Locations on a particular tube should be
investigated to find maximum wastage. Selective sampling is recommended to confirm the
results of the NDE examinations and to evaluate external deposits.

Monitoring thermocouples, either permanently placed or temporary installations, can be used to

detect excessive metal temperatures that might foster future accelerated wastage or tube failures.

5-18
13622634
 Fireside Corrosion in Oil Fired Units

5.6 Repairs, Immediate Solutions and Actions

5.6.1 Need for Remaining Life Assessment

Any immediate solution or repair strategy should consider that a longer term remaining life
analysis methodology should also be implemented. Such programs are discussed in detail in
Appendix C, and are summarized under the longterm actions below.

5.6.2 Repairs

If the corrosion rate has been relatively modest (< 25 nm/hr (~ 9 mils/yr)) and is likely to
continue to be so, an acceptable strategy is to retube with the same alloy and monitor closely the
wastage rate. Another approach, although not preferred, is the use of a palliative coating or tube
shields. An overview of options is provided under longterm actions.

Pad welds should definitely not be used as a repair measure because of the uncertainty of the
tube conditions such as the presence of creep cracks and their depth, and the conditions of the
internal tube surface.

For higher corrosion rates that are resulting in rapid wastage of the existing alloy, the
replacement should be with a more resistant material.

Note that considerable care must be taken in any repairs, inspections or movements within oil-
fired units because of the extreme toxicity of vanadium dust. Breathing equipment or masks
must be used for any repairs. Note that as with other vanadium complexes, ashes containing
these oxides are toxic. They are also slightly water soluble which may lead to groundwater
contamination if used for land fill.

5.6.3 Use of Mg-Based Additives

Utilities that are burning, or know that they will be burning, high vanadium residual oils will
generally use an additive as the most cost-effective means to prevent fireside corrosion.

5.6.4 Other Short-term Options

Several other short-term options that can be considered depending on the root cause might
include (i) interim steps to limit tube metal temperatures such as by cleaning steamside oxide, (ii)
aligning tubes, and (iii) improving sootblowing operation. A complete overview of all options is
included in the next section.

5-19
13622634
Fireside Corrosion in Oil Fired Units

5.7 Life Assessment

A remaining life assessment is required to relate the rate of corrosion wastage to the desired life
or to determine the time available to implement the desired option. Therefore such an assessment
should be undertaken in parallel with any of the other options. A systematic program will
include: baseline measurement, monitoring rates of wastage, application and monitoring of
control measures, and assessment of the effects on remaining life. Monitoring of the flue gas,
metal and steam temperatures, combustion conditions, and fuel composition should also be
considered as these can determine corrosion rates while the unit is still on-line.1 Particularly
important are step changes in key parameters.

A critical step in determining the remaining life of a tube and in choosing the optimal solution is
knowing the rate at which damage is accumulating. The lifetime of superheater/reheater tubes
can be calculated using the techniques discussed in Section 4.7 and in Appendix C.

A key goal of the remaining life assessment will be to establish the acceptable rate of wastage,
given the desired remaining life of the unit and an analysis of damage accumulation. As a rule of
thumb, rates of wastage above 25 nm/hr (~ 9 mils/yr) indicate the need for some periodic activity
such as annual inspections and for the explicit determination of re-evaluation periods. Rates
above 50 nm/hr (17.2 mils/year) indicate a serious decrease in life of the tubing; an estimate of
remaining life should be made and the appropriate actions taken that life is unacceptable.

5.8 Longterm Actions and Prevention of Repeat Failures

As noted above, it may not be possible to remove the root cause for many fireside corrosion
problems. Knowing how to minimize the wastage rate and the application of a predictive
remaining life assessment process including periodic inspection and monitoring, are the keys to
economic handling of fireside corrosion problems.

Figure 5-7 outlines most of the available corrective actions for superheater/reheater fireside
corrosion. As shown in that figure, three primary, and not mutually exclusive, routes that can be
followed are: (i) fuel options, primarily additives, (ii) materials options, and (iii) operating
options.

The circled numbers used in Figure 5-7 are used to identify options for the discussion that
follows and no ranking of is implied; however, boxes with bold outline indicate those options
which have been the most successful.

5-20
13622634
 Fireside Corrosion in Oil Fired Units

Figure 5-7
Strategies for preventing repeat failures by fireside corrosion in superheater/reheater
tubes of oil-fired units.

5.8.1 Remaining Life Assessment (Option 1, Figure 5-7)

Remaining life assessment is discussed above in Section 5.7.

5.8.2 Shielding (Option 2)

The most common temporary measure is the installation of shields to the leading edges of
affected tubes. High chromium alloys such as Type 310 or 50Cr-50Ni are typically
recommended. The shield is curved to fit the tube surface and tack-welded in place. As they are
not cooled, the metal temperatures are above the range for extensive attack.

5.8.3 Coatings (Option 3)

A variety of processes have been suggested for applying metallized coatings to tubes in situ as a
means of increasing corrosion resistance. The advantage of coatings is that very corrosion-
resistant materials can be applied at specific susceptible sites, so it is not necessary to replace
entire sections of tubing in order to overcome localized problems.

5-21
13622634
Fireside Corrosion in Oil Fired Units

Among the coating methods that have been tried for fireside corrosion resistance are chromizing
and aluminizing. Flame- or plasma spraying, with and without subsequent heat treatment, has
also seen significant development work. A number of coatings have been tried for use in
corrosion and erosion resistance.5, 6 The primary use of coatings has been for the prevention of
fireside corrosion in waterwalls of coal-fired units. There is less field experience for coatings on
superheater/reheater tubing in either coal- or oil-fired units. Currently, either replacement in-kind
or with a material of high corrosion resistance, depending on the wastage rate which has been
experienced, are preferred options to the application of coatings.

5.8.4 Replacement with Same Material, Same Thickness (Option 4)

If the corrosion rate is only slightly higher than that required to reach the desired life as
calculated from the remaining life assessment, tube replacement can be made in-kind.

5.8.5 Change to a More Resistant Material, "Monolithic" (Option 5)

An upgraded material can be used where unit operation is at high temperatures and fireside
corrosion remains a problem despite the attempts at other fixes. The material chosen will depend
on what is currently being used and what the desired resistance is to be. The corrosion rate of
300 series austenitic stainless steels, because of their nickel content, is greater than for ferritic
stainless steels. Upgrading to an alloy such as 12 Cr Mo V may therefore be necessary.1

It should be noted that there may not be significant improvement in the resistance to fireside
corrosion per se, but longer times to failure may result because the upgraded material has a lower
creep rate and higher strength (so that a thinner failure thickness is achieved).1 As a general rule,
the more resistant the material, the more expensive it will be.

5.8.6 Change to a More Resistant Material, Composite such as Co-Extruded

Tubing (Option 6)

Co-extruded tubing, originally developed for SH/RH applications in coal-fired units, could be
used for severe corrosion in oil-fired units. A 50Cr-50Ni alloy is used as the outer layer and is
metallurgically bonded to a creep-resistant, but not corrosion resistant core, typically Eshete
1250 (a high-temperature, high-strength stainless steel). Metallurgically-bonded processing
provides superior tubes for use in utility boiler applications when compared to mechanically
bonded or bi-metallic tubes.7

Tube bending characteristics of co-extruded material have been found to be identical to

monolithic materials.7 Weld procedures have generally matched weld metals to the base metal to
maintain property levels. Conventional welding techniques and normal quality control have been
found sufficient to ensure good weld quality. No preheat or post-weld heat treatments have been
required.

5-22
13622634
 Fireside Corrosion in Oil Fired Units

5.8.7 Monitoring of Fuel Changes (Option 7)

Monitoring for increases in vanadium and sodium content whenever fuel changes are made will
highlight a potential increase in corrosion rates.

5.8.8 Use of Mg-Based Additives (Option 8)

A number of forms of magnesium and combinations with other metals have been found to be
effective in reducing fireside corrosion in oil-fired units, particularly the use of magnesium
hydroxide Mg(OH)2 or magnesia, MgO. The use of magnesium has been found to form a
magnesium vanadate complex (3MgO • V2O5) which is desirable because of it's high melting
point of 1071°C (1960°F) as shown in Figure 5-6.

The type of additive to be used will depend on fuel chemistry and boiler design. The most
effective product forms have been either a slurry or a solution. The form will determine the point
of delivery and kind of delivery system. Treatment rates will depend on a number of factors
including: particle size and distribution, reactivity of chosen additive, design of boiler, mode of
application, and some trial and error.8

Choice and proper application of additives is a complicated subject.8, 9 The most successful or
optimized additives for a number of fuel oils are shown in Tables 5-6 and 5-7 for high vanadium,
high sulfur and for a medium vanadium, low sulfur oil respectively.8 These are meant to be
starting points as the optimum additive choice for each boiler and type of operation will differ. It
is important to adjust for conditions different from those presented in these tables.

5-23
13622634
Fireside Corrosion in Oil Fired Units

Table 5-6
Additives for High-Temperature Corrosion and Fouling from High-Sulfur Oil a

Feed Point Sootblower

Boiler Design Furnace Injection
with Fuel Feed Point

Mg0-oil Mg0 dispersion

Coal converted to oil dispersion (H20 or oil) Not recommended
2 to 7 microns 2 to 7 microns

Oil dispersion of
Mn0 or
Mg0/Mn0
2 to 7 microns

As above but
As above but less
Oil or oil with higher in furnace Mg0-H20
than 1 micron
coal potential to avoid water wall slurry
size reactive forms
coating

Oil-soluble MgCl2 solution-

Mg/Mn emulsion

MgCl2 solution-
emulsion

Notes:
High sulfur (2-3.5%) and vanadium (300-700) ppm, low sodium/vanadium ratio.
Additives listed in preferred order of recommendation.
Source: J.E. Radway and M.S. Hoffman8

5-24
13622634
 Fireside Corrosion in Oil Fired Units

Table 5-7
Additives for High-Temperature Corrosion and Fouling for Intermediate-Sulfur Oil a

Feed Point Sootblower

Boiler Design Furnace Injection
With Fuel Feed Point

Dispersion of Mg/Al; Dispersion of

Coal converted to oil Mg/Al Mn; Mg/Si Mg/Al Not recommended
2 to 7 microns 2 to 7 microns

As above but
As above but less
Oil, or oil with coal higher in furnace
than 1 micron and Mg0-H20 slurry
potential to avoid water wall
more reactive forms
coating of Mg0/Al203

Oil-soluble Mg/Mn

MgCl2 solution-emulsion

High reactivity
Gas-oil Combustion catalyst Not usually available
Mg0 dispersion

Soluble additive Soluble additive

(water or oil) (water or oil)

High-reactivity
Mg0 dispersion

Notes:
1 to 3% sulfur, 150 ppm vanadium, sodium/vanadium greater than 0.15.
Additives listed in preferred order of recommendation.
Source: J.E. Radway and M.S. Hoffman8

Deposition probes which can provide useful short-term information on the quantity and
corrosiveness of depositing ashes should be used to optimize the use of additives.10

The disadvantages of additives may include the relatively high continuous operating cost and a
substantial increase in ash volume leading to additional downtime for cleaning.2 Increased
abrasive wear of oil gun components can also occur. Thus, there is a need to continually check
the addition rate and its efficacy through the use of probes.

5.8.9 Limit Tube Temperatures (Option 9)

Given the strong correlation between tube metal temperature and the potential for corrosion
damage, several strategies can be implemented that directly address root causes of
overtemperature in the tubes; they are discussed separately below.

5-25
13622634
Fireside Corrosion in Oil Fired Units

5.8.10 Limit Steam Temperature (Option 10)

A specific strategy to minimize tube metal temperatures is by limiting main steam temperatures;
this has historically been a primary strategy for the control of superheater/reheater fireside
corrosion, particularly in coal-fired units when reducing the main steam temperature from 565 to
538°C (1050 to 1000°F).

There are several problems with this approach. The primary drawback is that it is not efficient;
there is a severe heat rate penalty. Predicting exactly what temperature is required may also be
difficult.

For these reasons, it has been recommended that if limits on main steam temperature were to be
used as a control strategy that an empirical approach be used to set steam temperature by
monitoring tube wastage rates and ensuring that the rate was less than 25 nm/hr (~ 9 mils/yr).11

5.8.11 Redistribute Steam Flow (Option 11)

A technique of steam flow redistribution in superheaters has seen significant development.12

Redistribution of steam flow can serve to equalize the temperatures across the superheater. The
method which was designed primarily as a means of extending the creep life of tubes subject to
overheating, however, if overheating of selected tubes is at the root cause of fireside corrosion,
the technique can selectively lowering tube metal temperatures and thus decrease fireside
corrosion. The method is discussed in more detail in the Chapter 11 on longterm
overheating/creep of SH/RH tubes.

5.8.12 Chemical Cleaning to Remove Steamside Oxide Scale (Option 12)

If excessive temperature in the superheater/reheater tubes is a contributing factor to the corrosion

process, and if that condition has been exacerbated by the presence of increasing oxide scale
thickness, a solution involving chemical cleaning may be in order.

5.8.13 Correct Misalignment of Tubing (Option 13)

If misalignment of tubes has created localized fireside corrosion, this problem should be
corrected by realigning the affected tubes. This will decrease the number of tubes that are
directly exposed to the gas flow.

5.8.14 Change Frequency and Check Effectiveness of Sootblowing (Option 14)

This can be an important action because it can stop the formation of excessive deposits which
result in laning or channeling in adjacent areas.

5-26
13622634
 Fireside Corrosion in Oil Fired Units

5.8.15 Lowering Excess Air (Option 15)

Operating with as low an excess air as possible can help keep constituent ash melting points high
and therefore limit deposition of the liquid phase and the resulting rapid corrosion. The optimal
level of excess oxygen has been found to range from around 0.8%10 to 1%2 depending on local
CO generation. There may be problems holding to these levels in older units because of leakage
or incomplete combustion leading to unburned fuel as soot to the stack. Care must also be taken
to avoid lowering the excess air too much in cases where carburization of the tube is present.14

5.9 Case Study

Three superheater tubes ruptured a few feet above the bottom bend following 37 years (270,000
hours) of operation in an oil fired furnace. The tubes were specified as ASME SA213-Grade
TP321 with dimensions of 2.125 inch OD and 0.238 inch MWT. The superheater outlet pressure
is 1850 psig.

The tube samples were covered by yellow, gray-green, gray, rust, and white colored deposits
(Figure 5-8). The ruptures were open (fish mouth) in two cases and narrow in one case. Wall
thickness measurements disclosed severe and irregular wall loss from the external tube surface of
up to 47% below the specified MWT. The internal tube surface appeared to be reasonably close
to its original condition.

Analysis of the external deposits by energy dispersive spectroscopy disclosed high

concentrations of vanadium in the gray-green and yellow layers, and sulfur and sodium in the
gray, yellow, and white layers. The presence of vanadium and sodium suggests that vanadium-
sodium-oxygen compounds (nNa2O-V2O5 and nNa2-V2O4-mV2O5) with low melting temperatures
were present that fluxed away from protective oxide layer, thereby accelerating wall loss.

The unconnected, intergranular cracks in the tube wall adjacent to the rupture indicate that the
gradually increasing tube stress and high metal temperature eventually resulted in failure by
creep-rupture. The lack of appreciable failure deformation is indicative of longterm creep-rupture
and suggests that the operating temperature is only moderately higher than design.

The internal surface may be close to its as-installed condition. Although final failure may be by a
secondary mechanism, such as creep, the primary cause of the failure should be recognized as
wall thinning due to fireside corrosion.

5-27
13622634
Fireside Corrosion in Oil Fired Units

Figure 5-8
An example of fireside corrosion in a secondary superheater tube that resulted in a
longterm overheat – creep failure (rupture indicated by arrow). Multicolored deposits are
present on the fireside surface of the tube. Severe and irregular tube wall loss (up to 47%)
is apparent.

5.10 References
1
Laxton, J.W., D.B. Meadowcroft, F. Clarke, T. Flatley, C.W. King, and C.W. Morris, The
Control of Fireside Corrosion in Power Station Boilers, Third edition, Central Electricity
Generating Board, 1987.
2
Wilson, J.R., "Understanding and Preventing Fuel Ash Corrosion," Corrosion 76, Paper No. 12,
held March 22-26, 1976, Houston Texas, 1976.
3
Paterson, S.R., T.A. Kuntz, R.S. Moser, and H. Vaillancourt, Boiler Tube Failure Metallurgical
Guide, Volume 1: Technical Report, Volume 2: Appendices, Research Project 1890-09, Final
Report TR-102433, Electric Power Research Institute, Palo Alto, CA, October, 1993.
4
Pantony, D.A. and K.I. Vasu, J. Inorg. Nucl. Chem., Vol. 30, 1986, p. 423.
5
Morgan-Warren, E.J., "Thermal Spraying for Boiler Tube Protection", Welding and Metal
Fabrication, Jan/Feb, 1992, pp. 25-31.

5-28
13622634
 Fireside Corrosion in Oil Fired Units

6
Bennett, A.P. and M.B.C. Quigley, "The Spraying of Boiler Tubing in Power Plants", Welding
and Metal Fabrication, November, 1990, pp. 485-489.
7
Flatley, T. and T. Thursfield, "Review of Corrosion Resistant Co-Extruded Tube Development
for Power Boilers", 1984 ASM Conference on Coatings and Bimetallics for Energy Systems and
Chemical Process Environments, held at Hilton Head, South Carolina, November 12-14, 1984.
8
Radway, J.E. and M.S. Hoffman, Operations Guide for the Use of Combustion Additives in
Utility Boilers, Research Project 1839-3, Final Report CS-5527, Electric Power Research
Institute, Palo Alto, CA, December, 1987.
9
Krause, H.H., Action of Fuel Oil Additives Containing Magnesium and Manganese on
Superheater and Reheater Surfaces, Final Report CS-3281, Research Project 1839-1, Electric
Power Research Institute, Palo Alto, Ca, December, 1983.
10
Dooley, R.B. and H.J. Westwood, Analysis and Prevention of Boiler Tube Failures, Ontario
Hydro, Report 83/237G-31, Canadian Electrical Association, Montreal, Quebec, November,
1983.
11
Williams, D.N., H.R. Hazard, H.H. Krause, L.J. Flanigan, R.E. Barrett, and I.G. Wright,
Fireside Corrosion and Fly Ash Erosion in Boilers, Research Project 2711-1, Final Report CS-
5071, Electric Power Research Institute, Palo Alto, CA, February, 1987.
12
Hara, K., C. Lee, R. Moser, T. Rettig, and K. Clark, Improved Superheater Component
Longevity by Steam Flow Redistribution, Research Project 1893-13, Final Report TR-101697,
Electric Power Research Institute, Palo Alto, CA, December, 1992.
13
Wong-Moreno, A., Y. Mujica Martinez, and L. Martinez, "High Temperature Corrosion
Enhanced Residual Fuel Oil Ash Deposits", Corrosion 94, Paper 185.
14
Private Communication from D. Lopez Lopez (IIE, Mexico) to R. B. Dooley, May, 1995.
15
Lopez Lopez, D., A. Wong Moreno, and L. Martinez, "Unusual Superheater Tube Wastage
Associated with Carburization", Materials Performance, Volume 33, No. 12, December, 1994,
pp. 45-48.

5-29
13622634
13622634
6
PITTING AND CREVICE CORROSION

6.1 Introduction

Pitting and crevice corrosion are localized forms of attack at a point where the passive film
breaks down due to interaction with ions such as chloride. Pitting of austenitic stainless steel
requires contact with moisture with high levels of dissolved oxygen, low pH or chlorides,
conditions which result primarily from poor shutdown/layup practices which allow stagnant,
oxygen-saturated condensate to form in the pendant loops of tubing. The rate of pitting increases
with increasing temperature up to temperatures approaching the dew point. High chlorides, low
pH, and high temperatures make the environment more aggressive and make pitting attack more
likely. Chemical cleaning damage in SH/RH sections, which can also cause pitting, is another
manifestation of pitting.

6.2 Features of Failure and Typical Locations

6.2.1 Features of Failure

Pitting is localized part- or through-wall dissolution of tube metal. Pitting that occurs in SH/RH
sections is a result of stagnant oxygen-saturated water formed during shutdown or layup periods.
These pits can be either numerous and closely spaced, or isolated, and often covered with caps of
corrosion products. Such pitting is generally found to be filled with corrosion products,
particularly red hematite iron oxide (Fe2O3) which is the thermodynamically favored reaction of
iron with oxygen under conditions of high oxygen.

Pits associated with low pH (acidic) conditions tend to be numerous and closely spaced. The
affected metal is usually clean and free of deposits (if found soon after the acid attack); the
surface is jagged and rough. Pits are generally sharp-walled and free of oxide or corrosion
products if fresh. Damage from excessively aggressive chemical cleaning in SH/RH sections is
of this type. Pits are typically narrow at the surface, are often much deeper than they are wide
and may open up under the surface to occupy significant volume. The environment trapped
inside the pit often becomes more aggressive than the bulk solution outside the pit and the pit
propagates, even if the bulk environment becomes more benign. The metal surface around a pit
may eventually become rust-stained.

6-1
13622634
Pitting and Crevice Corrosion

6.2.2 Locations of Failure

Typical locations for pitting are those where condensate can form and remain as liquid during
shutdown periods if improper venting or draining procedures are utilized. Typically affected
areas for both superheaters and reheaters are (i) the bottoms of pendant loops and (ii) low points
in sagging horizontal tubes.

Crevice corrosion may occur when something blocks the surface of the stainless steel and
prevents oxygen from renewing the passive film. The crevice can be almost anything that
prevents the air from reaching the surface (e.g., deposits, weld spatter, heat tint zones adjacent to
welds, contact surfaces such as with an adjacent tube, etc.). The likelihood of crevice attack
increases as the crevice becomes tighter and if the process environment has high chlorides, low
pH, or elevated temperatures (approaching the dew point).

Pitting and crevice corrosion may occur in Type 321H material even in solutions of relatively
low chloride content. However, the stabilization with titanium results in a somewhat better
resistance than that of AISI 304.Type 347H may be sensitive to pitting and crevice corrosion
even in solutions of relatively low chloride content. Stainless steels with increased contents of
Cr, Mo, and N have increased resistance to the initiation of pitting or crevice corrosion.

6.3 Mechanism of Failure

6.3.1 Development of Pitting Corrosion

Pitting is a form of corrosion that is distinguished by the aspect ratio of the damage: it tends to be
deep through-wall relative to the defect length. It is an insidious form of damage because (i) a
relatively small amount of metal loss can lead to through-wall failure with catastrophic results,
(ii) the small size coupled with the fact that pits are often filled with corrosion products makes
them hard to detect, (iii) pits often undercut the surface, which can complicate their detection,
(iv) laboratory simulation of field pitting is difficult, (v) and, perhaps most interesting, pitting is
autocatalytic; that is, conditions within the pit stimulate the continued activity of the corrodant.1

A breakdown in the passivity of a metal surface initiates the pitting process. An electrolytic cell
is formed, the anode is a small area of active metal, the cathode a large area of passive metal.2 A
large potential difference exists (about 0.5 V for 300-series stainless steels) which results in
considerable current flow and rapid corrosion. Figure 6-1 shows the process for the growth of a
pit in a metal M caused by a concentrated solution of aerated NaCl. Dissolution of metal in the
pit forms metal ions (M+) which results in the inward migration of Cl- from solution to maintain
charge neutrality. The metal chloride (M+Cl-) formed combines with water to form hydroxide
and free acid:

(M+Cl-) + H2O -> MOH + H+Cl- (41-1)

6-2
13622634
 Pitting and Crevice Corrosion

Figure 6-1
Autocatalytic processes occurring in a corrosion pit. From: M.G. Fontana and N.D.
Greene, Corrosion Engineering, 1967, McGraw-Hill, New York, NY. Reproduced with
permission of the McGraw-Hill Companies.

Although the pH of the bulk solution remains neutral, the concentration of acid in the pit lowers
the pH values (to the range 1.5 to 1.0).2 This accelerates the dissolution of the metal. At the
same time, a cathodic reaction such as oxygen reduction, is occurring on nearby surfaces which
suppresses corrosion on these adjacent areas. The suppression of corrosion around the edges of
the pit and the acceleration of corrosion within the pit result in increasing rates of damage
accumulation.

This model for pitting initiation and growth provides an explanation for other observations about
pitting. For example, pitting generally displays long initiation periods followed by what can be
quite rapid propagation. Why? Initiation is thought to require a momentarily high concentration
of metal dissolution to start the process.3 Such an initial concentration might be caused by a
surface scratch, dislodged defect or random solution variation.1 However, just as it forms, the
initial concentration could be swept away by the flow along the surface. Thus initiation might
nearly start and stop many times before a "permanent' concentration forms that is sufficient to
initiate a pit. In tubing, other possible initiators are deposits of feedwater corrosion products,
which might contain a number of different cations (such as Fe, Cu, Ni, Zn, etc.) and anions (Cl,
SO4, etc.). A number of theories have been proposed to explain pitting initiation. Two general
classifications are: (i) kinetic theories that explain breakdown of passivity based on competitive

6-3
13622634
Pitting and Crevice Corrosion

absorption between chloride ions in solution and oxygen on the metal surface, and (ii)
thermodynamic theories that consider that the critical pitting potential to be the potential where
chloride ions are in equilibrium with the oxide film.2, 4, 5, 6

Pit growth is influenced by gravity, a fact that is also explained by the mechanistic process
described above. The formation of a dense, concentrated solution and its retention favors the
propagation of the pit, a process that is facilitated if the pit is oriented to the pull of gravity.

Similarly, it is now clear why pitting is most prominent in stagnant conditions. The fact that the
fluid is static allows the initiation of pitting to occur unimpeded and stagnant conditions allow
for the concentrations needed to accelerate the attack. Once initiation has begun, such as during
shutdown of the unit, the continued growth of the pit under the more turbulent conditions of
operation can continue unabated.

Pitting remains a topic of active research interest. Additional information about pitting
mechanisms and activities can be found in references 4, 7 and 8.

6.3.2 Development of Pitting from Oxygen-Saturated, Stagnant Condensate

The most common cause of pitting in steam-touched tubing is the presence of oxygen-saturated,
stagnant condensate formed as a result of unit shutdown. Forced cooling and/or improper
draining and venting procedures can result in excess moisture throughout the boiler. Oxygen is
present if improper blanketing and/or protection procedures are used during the shutdown. The
problem is particularly acute in local areas where collected liquid is exposed to the air (the
interface line).

6.4 Possible Root Causes and Actions to Confirm

As discussed above, pitting can form if shutdown procedures allow accumulations of stagnant,
oxygenated water. Actions to confirm this root cause include:

(a). Analysis of corrosion products present in and around the pitting will help to pinpoint the
source of the problem. In the case of oxygenated attack, hematite (Fe2O3) is generally present.

(b). Selective tube sampling can be used to evaluate whether localized corrosion cells have been
formed.

(c). A critical evaluation of shutdown procedures and unit conditions during shutdown should be
performed to ensure that they are optimal for the unit. A review of the chemistry logs of analyses
performed during the shutdown will provide clues. A partial list of questions to be assessed
might include the following: (i) Was the boiler pH increased to around 10 with ammonia and
around 200 ppm of hydrazine and were these monitored during shutdown/layup periods? (ii)
Were nitrogen blanketing and caps properly applied and checked? (iii) Was sufficient N2H4
added? (iv) Have sources of air inleakage been detected?

6-4
13622634
 Pitting and Crevice Corrosion

6.5 Determining the Extent of Damage

Pitting can be difficult to detect unless it is extensive. If there has been significant loss of wall
thickness, the standard method of NDE is ultrasonic testing. Typical problems with access and
with surface preparation will be present.

Selective tube sampling and metallographic analysis can be used to detect localized corrosion
cells.

6.6 Repairs, Immediate Solutions and Actions

The primary immediate actions are to: (i) identify damaged locations, (ii) replace the affected
tubes, and (iii) repair obvious contributing conditions such as drum furniture damage, sources of
air in-leakage, etc.

The balance of actions can be conducted as a part of longterm actions.

6.7 Life Assessment

Pitting and crevice corrosion are damage mechanisms which occur during specific periods of
time (shutdowns, during chemical cleaning, etc.) and at localized areas where the environment is
conducive to attack. Because of this, the concept of life assessment is not applicable to this
damage mechanism; any actions should focus on preventing the conditions which promote the
damage.

The remaining life of austenitic superheater and reheater tubes experiencing pitting or crevice
corrosion is a function of the current distribution of pit/crevice depths, acurrent , the growth rate, arate,
and the maximum tolerable size, acrit . If these are each known then the remaining life, RL,
estimate is:
n n
RL = (acrit) – (acurrent) /arate

Where:

n = 1 (linear pitting rate) to 2 (parabolic pitting rate)

In practice it is extremely difficult to forecast the life of austenitic superheater and reheater tubes
that are experiencing pitting or crevice corrosion because:
• The growth mechanisms are only active during “times-of-wetness”, and the time-of-wetness
is generally not known
• It is difficult to obtain accurate, comprehensive measurements of the size (i.e., depth, length,
width) of the corrosion attack.

6-5
13622634
Pitting and Crevice Corrosion

• The growth rate, arate, of pits or crevices is difficult to predict from 1st principles. The
corrosiveness of the local aqueous environments responsible for the corrosion attack may
change both spatially and temporally. With a detailed evaluation of the past operating history
it may be possible to make a reasonable estimate of the cumulative time-of-wetness, but the
chemistry and corrosiveness of the moisture will for the most part be unknown. More often
the assumption that the future exposure will mimic the past is made and the growth rate is
inferred from the current pit depth.

In spite of these challenges some life prediction efforts may still be useful. This should start by
determining the extent, locations and severity of the damage. This is followed by an estimation
of the pitting rate and remaining life for based on creep life exhaustion, plastic collapse or
through-wall growth. A final check to determine the safety margin against unstable ductile
fracture is also useful.

The current pit depth, acurrent, can sometimes be obtained from nondestructive measurements (e.g.,
radiography, ultrasonic testing, eddy current testing). These measurements often require use of
custom nondestructive examination protocols for specific tube geometry and corrosion geometry.
If the results of nondestructive measurements are to be relied on for life estimations then it is
imperative that laboratory evaluations of tube samples with various severities of damage be
evaluated to verify that the NDE method had a meaningful detection and sizing capability and to
quantify the systematic and random errors in the measurements. Laboratory measurements of the
damage severity on a sample of tubes can sometimes also be used to remove some of the
systematic measurement error inherent in the NDE process.

There are three potential failure modes that should be considered: (i) plastic instability (i.e., hoop
stress > yield strength), (ii) creep/stress rupture failure, and (iii) through-wall leakage

The plastic instability limit can be determined by incrementally evaluating the wall thickness,
applied hoop stress for measured or assumed fireside corrosion rates. The critical pit/crevice
depth can be estimated using the net section stress relationships from the EPRI Ductile Fracture
Handbook in combination with the incremental damage models discussed in Section 4.7. For
aligned pitting that is oriented along the axis of the tubing the net section stress is:

σnet section = P(Do-t)/2t (1-x/M)/(1-x)

Where:
σnet section = net section stress
P = steam pressure
Do= outside diameter
t = wall thickness
x = a/t
a = pit depth
M = {1+1.61 (Lpit)2/[(4)(R)(t)]}1/2

6-6
13622634
 Pitting and Crevice Corrosion

Lpit = pit length

R = mean radius = (Do-t)/2
The plastic instability life is consumed when the estimated net section stress exceed the materials
yield strength at the service temperature. The calculations can be easily performed using a
spreadsheet and a 2000 hr incremental time interval. The elevated temperature yield strength
values for Type 304H, 321H, and 347H are shown in Figures A-8 through A-10 respectively. For
tube metal temperatures in the range of 500oC to 650oC an assumed yield strength value of 120
MPa will provide adequate plastic instability life estimations for all three of the austenitic tube
alloys.

The creep/stress rupture life can be estimated with the same incremental (2000 hour per time
step) spreadsheet model. In the case of creep damage it is useful to evaluate the minimum and
median life estimations. In either case the creep life is estimated using a cumulative damage, life
fraction rule:

Dc = ∆HRS / trup(σMDF, Tmidwall)

Where:

Dc = Creep damage fraction = 1 at failure

∆HRS = time interval with assumed constant σMDF, Tmidwall; a time interval value of 2000
hrs generally provides adequate accuracy

trup(σMDF, Tmidwall) = estimated time to rupture for a constant mean diameter stress, σMDF,
and midwall metal temperature, Tmidwall. Section A.4.2 provides median and lower bound
time to rupture equations for the three common austenitic tube alloys. Appendix C
provides a detailed description of methods that can be used to estimate the midwall tube
temperature at any particular gas-touched location.

Figures 4-11(a) and 4-11(b) present the results of plastic instability and creep/stress rupture life
estimations of a Type 304H reheater tube operating at a midwall temperature somewhat below its
design limit. The analyses were performed for assumed fireside corrosion rates of 7.5 nm/hr and
25 nm/hr. One of the important features that these calculations show is that are the fireside
corrosion rate increases, the importance of the scatter in stress rupture properties significantly
diminishes. Clearly there would be very little value in performing stress or creep rupture tests to
improve the accuracy in the life estimations for the reheater tube with 25 nm/hr fireside
corrosion rate. Improved accuracy in the life predictions for situations with severe fireside
corrosion of superheater tubes or moderate to severe fireside corrosion of reheater tubes will be
obtained by learning as much as possible about the rate of fireside corrosion. It would be useful
to know:
• The temporal variation in wall loss rate (i.e., Is the wastage rate is linear with time?)

6-7
13622634
Pitting and Crevice Corrosion

• The spatial variation in wall loss rate (i.e., from ultrasonic wall thickness surveys, evaluation
of selected tube samples, and modeling to remove all possible measurement errors or
uncertainties)
• The relationship between the wastage rate and the estimated midwall tube temperature.
Appendices B and C describe various methods to estimate the midwall tube temperature of
any desired gas-touched tube location.

Figure 6-2
Results of creep damage fraction, Dc, plastic instability, leak modeling of a Type 304H
reheater tube with axially aligned corrosion pits with a pit length to depth ratio of one and
a range of assumed fireside corrosion rates.

6-8
13622634
 Pitting and Crevice Corrosion

Figure 6-3
Results of creep damage fraction, Dc, plastic instability, leak modeling of a Type 304H
reheater tube with axially aligned corrosion pits with a pit length to depth ratio of 10 and a
range of assumed fireside corrosion rates.

An evaluation to determine the critical length, Lcrit, of a through-wall pit is useful so that the
potential for large unstable ductile fractures of tubes are prevented. The critical flaw length of an
axial, through-wall pit is a function of the tube geometry and operating pressure and midwall
temperature. The maximum tolerable pit or crevice depth will also depend on its orientation with
respect to the tube axis. Pits that do not extend appreciably in length will grow through wall and
cause a local leak. Aligned pits or crevices that extend for some length may fail by plastic
instability or local, net-section creep/stress rupture failure, leading to extensive secondary
damage. Using the limit load relationship for a through-wall axial flaw from the EPRI Ductile
9
Fracture Handbook the plastic collapse criteria are:

SF = σy/[(M)(σMDF)]
Where:
SF = Safety factor on strength or pressure
σy = yield strength for the service temperature
M = [1+1.2987 λ – 0.026905 λ + 5.3549E-4 λ ]
2 4 6 1/2

λ = (L/2)/(Rt)1/2
R = mean radius = (Do-t)/2
σMDF = mean diameter hoop stress = P(R)/t

6-9
13622634
Pitting and Crevice Corrosion

For the reheater tube example presented in Figures 6-2 and 6-3, the relationships above indicated
a safety factor of 1 for a through-wall flaw length that was 47.8 mm (1.88”) or thickness. If
aligned pits or crevices are detected that are approaching or may approach this length based on
the pit length to depth ratios (i.e., in this case ratios >14) should be regarding as a high failure
consequence condition.

6.8 Longterm Actions and Prevention of Repeat Failures

Longer term actions will consist of establishing procedures to prevent precursor conditions from
occurring and monitoring of steam chemistry to ensure that carryover is not excessive.

6.8.1 Instituting Proper Unit Shutdown and Layup Procedures

For SH/RH circuits, the main objectives are to ensure that good dry-out practices and nitrogen
blanketing are used during shutdown.

6.8.2 Steam Chemistry Monitoring

A means of monitoring steam composition, particularly for evidence of excessive levels of Na,
SO4, and Cl should be in place for the unit. All representative regimes of operation should be
included, with particular emphasis on those periods of transient operations.

6.9 References
1
Fontana, M.G. and N.D. Greene, Corrosion Engineering, McGraw-Hill, New York, 1967.
2
Asphahani, A.I. and W.L. Silence, "Pitting Corrosion" in Metals Handbook Volume 13:
Corrosion, ASM International, Metals Park, OH, 1987.
3
Evans, U.R., Corrosion, Volume 7, Number 238, 1951.
4
Shreir, L.L, R.A. Jarman, and G.T. Burstein, eds., Corrosion Volume 1: Metal/Environment
Reactions, 3rd Edition, Butterworth-Heinemann, Oxford, 1994.
5
Uhlig, H.H. and J. Gilman, Corrosion, Volume 19, 1963, p. 261t.
6
Vermilyea, D., Journal of the Electrochemistry Society, Volume 118, 1971, p. 529.
7
Isaacs, H., U. Bertocci, J. Kruger, S. Smialowska, Advances in Localized Corrosion, NACE-9,
National Association of Corrosion Engineers, p. 221.
8
Evans, U.R., The Corrosion and Oxidation of Metals, Arnold, London, 1961.
9
Ductile Fracture Handbook, EPRI, Palo Alto, CA: June 1989. NP-6301-D

6-10
13622634
7
INTERGRANULAR ATTACK AND STRESS CORROSION
CRACKING

7.1 Introduction

Intergranular attack (IGA) and stress corrosion cracking (SCC) are related cracking mechanisms
which affect superheater/reheater tubing constructed of austenitic stainless steel. Both
mechanisms involve a susceptible material in the presence of a wet corrosive medium.
Environments conducive to attack typically form during shutdowns; IGA/SCC will not occur in a
dry steam environment. In the case of SCC, a tensile stress (either applied or residual) at the
time of the corrosive exposure causes cracking in either an intergranular (IGSCC) or a
transgranular (TGSCC) morphology, depending on the specific contaminant and the stress level.
It is very common so see IGA even in new tubes, however the depth of attack will be limited to a
few grains or less. This pre-service IGA results from exposure to moisture during fabrication,
transport and storage.

7.2 Features of Failure and Typical Locations

7.2.1 Features of Failure

IGA and SCC of boiler tubes occur primarily in austenitic materials, but may also occur in
ferritics. Macroscopically, final failures are evidenced as thick-edged, brittle failures, and may
often involve the blow out of small "window-type" pieces. Damage may also be manifested as a
pin-hole leak. There is generally little plastic deformation associated with failure by IGA and
SCC, and there is also little or no loss of wall thickness because of the damage.

Cracks can be oriented circumferentially or longitudinally as they will form perpendicular to the
dominant stress, and may have significant branching. Failures can initiate at either inside or
outside surfaces of tubes, although ID initiation is the most prevalent.

Microscopically, crack propagation can be either transgranular (across the crystal grains) or
intergranular (along grain boundaries). An example of transgranular cracking, also sometimes
referred to as "branched SCC", is shown in Figure 7-1. This can be contrasted with intergranular
SCC, also termed IGA as shown in Figure 7-2. Either mode of propagation can be observed in
boiler tube materials, depending upon the conditions, and in some cases both can occur
simultaneously or in sequence.

7-1
13622634
Intergranular Attack and Stress Corrosion Cracking

Figure 7-1
Cracking of SA-213 Type 304H base material near the edge of a weld backing bar in a
dissimilar metal weld joint. This tube contained through-wall cracks after a few weeks
service. The corrodant causing the cracking was suspected to be a petroleum-based
preservative which was not effectively flushed form the weld backing rings prior to
service. Source: S.R. Paterson, et al.1

Figure 7-2
Intergranular stress corrosion cracking (IGSCC) of a SA-213 Type 304H reheater tube.
Away from the rupture, the IGSCC was limited to the tube’s internal surface. Carryover of
chlorides was believed to have been the corrodant responsible for the cracking. Source:
S.R. Paterson, et al.1

7-2
13622634
 Intergranular Attack and Stress Corrosion Cracking

Boiler tube failures caused by intergranular SCC/IGA may be confused with those caused by an
intergranular creep mechanism. Intergranular creep will show the presence of grain boundary
creep cavitation under high magnification whereas intergranular SCC/IGA will not. During
polishing of material for metallurgical examination, SCC/IGA damage will typically result in
grains that fall out (see Figure 7-2 for an example), whereas the cracking tends to be more
discontinuous for intergranular creep cracking.1 A final complication is that cracking may
proceed part way through-wall by intergranular SCC then propagate by intergranular creep.

Failures caused by intergranular corrosion can also be confused with intergranular SCC.
Intergranular corrosion has occurred extensively in the superheater and reheater tubing of several
boilers.2 The attack originated on the OD of the tubes due to the effects of sulfates and chlorides
in the deposited fly ash during wash down of the tubing while the boiler was out of service. The
propensity to the corrosion is enhanced when the tubing surface is exposed to the high
temperature effects of the ash which causes depassivation during the off-load periods when the
surface is wet.2 IGSCC and intergranular corrosion have the same morphology, that is, in
austenitic stainless steels they both require sensitization. The two will be distinguished primarily
by damage location. IGSCC is usually located at stress concentrations and can be either ID- or
OD-initiated. Intergranular corrosion occurs in such locations as (i) on the tube OD as caused by
wash down or (ii) on the tube ID at low spots where liquids will accumulate.

7.2.2 Locations of Failure

The typical failure locations are in bends and straight tubing with low spots; these locations are
exposed to the highest concentrations of contaminants when water solutions are left in, and
boiled out of, the superheater. Lower bends in vertical pendants which are entirely austenitic and
in which condensate can form during shutdown are a typical location for the occurrence of SCC.
Other potentially moist environments include the boiler windbox and areas where refractory in
contact with tubes can become wet. For example, water washing during outages to remove ash
deposits has been implicated in failures (see Case Study). Conditions when units are still under
construction are also of concern, when moisture with high concentrations of dissolved oxygen
and other aggressive gases or solids (chlorides, sulfur compounds) can condense onto the tubing
resulting in stress corrosion cracking.

High stress locations are also particularly susceptible. Such locations include bends, welds, tube
attachments, supports or spacers, and locations near welds, especially where a change in
thickness occurs. Higher stresses, either residual from the fabrication process or as a result of
operating conditions, are more typical of these types of locations rather than straight tubing.

7.3 Mechanism of Failure

7.3.1 Basic Overview of the Mechanism of IGA

For intergranular corrosion or attack (IGA) to occur, the steel first needs to be exposed to
temperatures in the range of 430-900°C (800 to 1650°F) for sufficiently long time for carbon to
combine with the chromium to form chromium carbides on grain boundaries within the stainless
steel structure. The steel is said to be “sensitized.” If that happens, regions adjacent to the

7-3
13622634
Intergranular Attack and Stress Corrosion Cracking

carbides in the grain boundaries are depleted in chromium content and become susceptible to
corrosion if there is subsequent exposure to a corrosive medium. This in-service sensitization
isn’t necessary if the tubing was slowly cooled or heated in its sensitization temperature range
during its manufacture, fabrication, or erection.

The degree and rate of sensitization can be controlled by either using a low carbon grade of
material or by using one of the “stabilized grades” of material where the low carbon effect can be
achieved by tying the carbon up in highly stable, fine, dispersed, intragranular compounds (e.g.,
titanium carbide in the case of Type 321H material and niobium (columbium)/tantalum carbides
in the case of Type 347H material). To achieve an optimal level of stabilization these materials
must be given a stabilization heat treatment after completion of the required solution annealing
heat treatment.

In the case of Type 321H the stabilization heat treatment would typically consist of heating to
1040-1100°C (1920 to 2010°F = 1313 to 1373K) for 2 to5 minutes, followed by rapid cooling in
air or water. In the case of Type 347H the stabilization heat treatment would typically consist of
heating to 1000-1100°C (1830 to 2010°F = 1273 to 1373K) for 5 to 20 minutes followed by
rapid cooling in air or water. The objective of these heat treatments is to fully dissolved the
chromium carbides and maximize the amount of titanium or niobium carbides. As shown in
Figure 7-3, the chromium carbides will fully dissolve in a couple minutes (120 seconds) at a
o o
temperature of approximately 871 C (1600 F=1144K).

7-4
13622634
 Intergranular Attack and Stress Corrosion Cracking

Figure 7-3
Times of complete dissolution of various carbides and nitrides in austenite as a function
of temperature.4 (Note: 1600oF = 1144 K)

Unfortunately the stabilized grades of superheater or reheater tubing are rarely given a
stabilization heat treatment prior during their manufacture and as a result tend to sensitize and
form sigma phase at a rate that isn’t markedly different than Type 304 material.

7.3.2 Basic Overview of the Mechanism of SCC

Stress corrosion cracking (SCC) occurs as a result of the combined action of a susceptible
material, a corrosive environment and a tensile stress. Stresses may be present in the form of
applied or residual stresses and, as noted above, locations where damage by SCC accumulates
are often associated with a stress riser or concentration. The mechanism(s) of stress corrosion
cracking is not fully determined and continues to be one of the most researched topics in
corrosion (see, for example, references 3-6).

7-5
13622634
Intergranular Attack and Stress Corrosion Cracking

The development of stress corrosion cracking requires that a susceptible material be subjected to
both stress and environmental effects. SCC is an electrochemical phenomenon, an aqueous
solution is required, however local, for its occurrence. Stress only will result in fatigue (cyclic
stresses) or overload failures; environmental attack only will result in pitting or generalized
corrosion. SCC often occurs in ductile materials that form a passivation layer on the surface and
therefore are resistant to general corrosion.4

For purposes of introducing the influences on SCC, the discussion of mechanism is here divided
into transgranular attack and crack growth, and intergranular growth, although it should be kept
in mind that they can occur together either sequentially or simultaneously, and there is evidence
that the two propagation modes can be represented as a continuum.7

For transgranular SCC, initiation occurs at the surface, from pits, discontinuities, scratches, or,
from local modification of the protective oxide. A rupture of the protective surface film allows
an initial corrosive attack. Film rupture is determined by a variety of local factors, including local
stress, strain rate, film thickness, film rupture strength, substructure of moving dislocations, and
microstructure.4 A further influence on the rate of damage accumulation is the repassivation rate
of the crack walls controlled by the material and such environmental parameters as potential, pH,
solution, and composition.4 Subsequently, the geometric effects of stress concentration at the
root of the initial surface defect lead to crack formation and advance. Successive rupture and
reformation of the passive film occurs at the crack tip as a result of tensile stresses.

In the case of intergranular SCC, preferential attack occurs at the grain boundaries of
"sensitized" material. Intergranular stress corrosion cracking typically occurs when the
contaminant responsible for the cracking is a sulfur compound or when the material is severely
sensitized (in which case chloride-induced stress corrosion cracking may become intergranular).
Transgranular stress corrosion cracking indicates that chloride is playing a role in the cracking
mechanism.

IGSCC and TGSCC both require:

• A susceptible material
• Contaminants (chlorides, sulfur compounds, oxygen)
• Tensile stress (applied or more often residual)
• Temperatures above approximately 60°C (140°F) but below the dew point
• Time of exposure

Nickel controls susceptibility to chloride stress corrosion cracking. The range of greatest
susceptibility is from about 8 to 10% nickel, which, unfortunately, is close to the nickel content
of all three of the commonly used stainless steel tubing materials. Materials with little or no
nickel (duplex stainless steels and ferritic stainless steels) and those with high nickel content
(super-austenitics and Ni-base alloys) have significantly better resistance to SCC. As described
above, all three common materials are prone to in-service sensitization and sigma phase
formation. The susceptibility of these materials to IGSCC increases markedly with the degree of
sensitization or sigma phase development.

7-6
13622634
 Intergranular Attack and Stress Corrosion Cracking

The concentration of chlorides required depends on the stainless steel grade but for Type 304
material it is on the order of a few parts per million. Any situation that results in wet-dry
situations or where the chlorides may evaporate and concentrate is sufficiently aggressive to
produce stress corrosion cracking.

Tensile residual stresses can result from forming operations such as cold finishing, bending or
straightening, welding, or from operational stresses, such as those from thermal cycling (a wet
sootblower spraying on hot tubing).

Stress corrosion cracking does not occur instantaneously, but instead takes some time to initiate
even when all of the other factors are present. Unfortunately, the time to initiation for a particular
exposure may be as short as a few hours. Consequently, it is sometimes possible to operate for
many years without SCC, but then rapidly observe stress corrosion cracking with no obvious
changes to the environment. In these cases the concentration of contaminant may have built up
over time, over the exposure time (outage time, time-of-wetness) may be somewhat longer and
sufficient to initiate cracking.

Figure 7-4 shows a general relationship that has been developed for the depth of penetration due
to SCC for austenitic stainless steels, showing the various influences and contributions from a
variety of sources. Specific data showing the effects of stress and temperature on time to failure
are provided in Figure 7-5. Figure 7-6 shows how increasing the yield strength of austenitic
stainless steels via cold work can increase the growth rate and extent of cracking.

Figure 7-4
General relationship for the penetration of SCC.5

7-7
13622634
Intergranular Attack and Stress Corrosion Cracking

Figure 7-5
(a) Time-to-failure vs. stress for 40 sets of data measured for stainless steel in 42% MgCl2
at open circuit conditions.5, (b) Time-to-failure vs. 1/T for 23 sets of data for stainless steel
in boiling 35% to 45% MgCl2 solutions at open circuit conditions.5

Figure 7-6
Increasing the yield strength as a consequence of cold work can increase the SCC growth
rate and extent of TGSCC.6

7.3.3 Boiler Tubing Materials and IGA/SCC

When austenitic stainless steels are exposed to temperatures in the range 430 to 900°C (~ 805 to
1650°F) chromium will diffuse to grain boundaries to form chromium carbides. The depletion of
chromium from the matrix nearby the grain boundary reduces the local corrosion resistance of
the material. Stainless steel in this condition is termed "sensitized". Sensitized stainless steel is
particularly susceptible to attack by chlorides.

Since stainless steel used in SH/RH sections is routinely exposed to these temperature levels, the
material will develop a susceptibility to IGA/SCC. Two approaches to limit the degree of
sensitization are (i) the use of low carbon grades and (ii) the use of stabilized grades of stainless
steels. In the low carbon grades, the objective is to have insufficient carbon to form carbides.

7-8
13622634
 Intergranular Attack and Stress Corrosion Cracking

Unfortunately, as they generally have low creep strengths, these materials are generally not
suitable for SH/RH tubing applications. The second approach is to utilize grades, such as Type
321H and 347H that contain elements (titanium and niobium respectively) which are stronger
carbide formers than chromium; thus the material maintains its corrosion resistance in these
"stabilized" grades. However, both Types 321H and 347H will sensitize in SH/RH applications.

7.3.4 Sources of Corrosive Environments in Boiler Tubes That Can Lead to

IGA/SCC

7.3.4.1 Steamside Sources

Steamside contributors to IGA/SCC are usually chlorides or hydroxides; ferritic materials are
susceptible to NaOH, stainless steels to either NaOH or chlorides. There are two primary sources
for these contributing species: (i) contamination from chemical cleaning, and (ii) carryover of
volatile chemicals from the boiler.

Briefly the major problems are:

• Poor back-fill procedures that fail to protect SH circuits from the carryover of solvents, such
as HCl, during cleaning of waterwalls.
• Problems that develop during cleaning of SH/RH circuits such as a breakdown of inhibitors
caused by excessive temperatures, or leaving acids in circuits because of improper flushing
of chemicals. Furthermore, the inhibitors used sometimes contain sulfur, which, if it deposits,
can lead to a problem with intergranular attack.

The second source of contaminants that can foster IGA/SCC is the carryover of volatile
chemicals form the boiler. For example, there can be a problem if there is carryover of NaOH in
those units under either caustic treatment or phosphate treatment and operating with excessive
levels of NaOH. There is also some emerging information that high levels of organics (hundreds
of ppm) can result in IGA/SCC especially when they are oxidized by high dissolved oxygen
levels.8

IGA/SCC caused by chloride or caustic ingress is invariably transgranular with extensive crack
branching.

7.3.4.2 Fireside Sources

Fireside corrosives are either polythionic acids, or less commonly, nitrates and sulfates.
Polythonic acids are H2SXO6 compounds where X is equal to 3, 4, or 5. They form in oil-fired
units from reactions between sulfur corrosion products, SO2, moisture and air.

7-9
13622634
Intergranular Attack and Stress Corrosion Cracking

7.4 Possible Root Causes and Actions to Confirm

7.4.1 Introduction

Failure by IGA/SCC requires all three aspects: environmental, stress and a susceptible material.
Part of the challenge in developing solutions will be the clear identification of the influence of
the various contributors. However, the reader should be forewarned that preventing BTF by
IGA/SCC, except for some of the effects of the environmental contribution, is difficult.

The root cause influences of environment, stress, and a susceptible material are discussed in the
next three sections. Table 7-1 summarizes the major root cause influences, their confirmation,
and corrective actions.
Table 7-1
Major Root Cause Influences, Confirmation and Corrective Actions

Major Root Cause Actions to Confirm Immediate Actions Longterm Actions and
Influences and Solutions Prevention of Repeat
Failure
7.4.2 Influence of Environment, Mainly Contamination From:
Carryover of chlorides (f) Analyze steamside Institute repair and Develop chemical
from the chemical fracture surfaces replacement as control and
cleaning of and oxide deposits required. monitoring
waterwalls. for presence of procedures to
Clean up sources of
contaminant maintain and verify
Boiler water carryover. contamination, if
species such as possible. For appropriate
Introduction of high chlorides. chemistry.
levels of caustic from example, reclean
desuperheating or (g) Analyze fireside SH/RH circuits if Develop optimized
deposits for improper flushing of chemical cleaning
attemperator spray.
indications of solvents is procedures and
Condenser cooling aggressive underlying cause. safeguards.
water constituents corrodants.
from a condenser
leak. (h) Review chemistry
records, monitoring
Fireside contaminants records, etc. for
such as polythionic indication of source
acid. of contamination.
Ingress of flue gas Perform carryover
environment into test. See main text
tube through primary for additional
failure, especially in discussion on this
RH when vacuum is point.
drawn. (i) Review recent
chemical cleaning
operations, either
waterwalls or
SH/RH circuits, for
potential sources of
contamination.

7-10
13622634
 Intergranular Attack and Stress Corrosion Cracking

7.4.3 Influence of Excessive Stresses

(j) Visual examination Repair any obvious Ensure that any
for signs of obvious contributions to redesign of supports,
distress such as excess stress levels. etc. will actually
broken or missing lower the imposed
See longterm actions.
attachments or stress.
supports, etc.
(k) Review field welding
procedures for
details of post-weld
heat treatment.

7.4.4 Influence of Sensitized Material

(l) Perform a See longterm actions. Consider replacement
metallurgical of material with a
examination to stabilized grade of
determine whether stainless steel.
sensitization
developed in
service or was
present as a result
of the fabrication
process used.
(m) Test the material for
baseline
susceptibility using
ASTM Standard
Practice A262.
(n) Analyze the
chemical
composition of the
failed material to
see if the proper
material was used.

7.4.2 Influence of Environment

Sources of contamination can occur on either the steamside or fireside and include: (i) carryover
of chlorides from the chemical cleaning of waterwalls, (ii) introduction of high levels of caustic
by carryover from the boiler, (iii) condenser cooling water constituents in attemperator spray
water as a result of a condenser leak, (iv) boiler water carryover, especially during drum level
surges produced by steam line blowing, (v) sulfur compounds that form polythionic acid during
shutdown and attack the fireside, (vi) ingress of flue gas environment into tubes through a
primary failure, especially in the RH when a vacuum is drawn, or (vii) desuperheating or
attemperator water.

7-11
13622634
Intergranular Attack and Stress Corrosion Cracking

Actions to confirm:

(a). Analyze steamside fracture surfaces and oxide deposits for the presence of chemical species
that would indicate a source of contamination. For example, the presence of chlorides has been
detected on fracture surfaces by EDX spectroscopy.

(b). Analyze fireside deposits, in the case of fireside attack, for indications of aggressive
corrodants. For example, in the case of SCC of ferritic materials, analysis of fireside deposits
typically show significant amounts of chloride and sulfate, and sometimes nitrate.

(c). Review of chemistry records may indicate the source of contamination. For drum units a
mechanical carryover test should be performed to determine if any of the drum internals are
cracked or by-passing moisture. This involves measuring the ratio of the contaminant in the
saturated steam to that in the boiler drum (CBD) and comparing that ratio to prior measurements
or to the manufacturer's specified limit.

(d). Review of recent chemical cleaning operations either in the waterwalls or in SH/RH circuits
may indicate whether a potential source of contamination was introduced in the SH/RH circuits.

7.4.3 Influence of Stresses

Since residual stresses from fabrication or welding alone, and/or in conjunction with normal
service stresses, can be sufficient to contribute to a problem with IGA/SCC in boiler tubes, it
may be difficult to provide significant mitigation of this individual contributor. Generally, an
outbreak of IGA/SCC occurs as a result of a contaminant getting into the SH/RH with the
affected areas being those at the highest imposed stresses. As a result, although stresses may be a
contributor, control of IGA/SCC in boiler tubes will more often be found in identification and
correction of the occurrence of contamination and/or environmental effects. There may be,
however, obvious sources of excessive stresses such as improperly designed or constructed
attachments, missing supports, improper field welding or heat treating procedures etc.

Actions to confirm:

(e). Visual examination for obvious signs of stress distress such as missing or broken
attachments, supports, etc.

(f). Review field welding procedures for details of post-weld heat-treatment.

7.4.4 Influence of a Susceptible Material

The third contributor to IGA/SCC in boiler tubes is a susceptible material. As noted above, the
operating temperature in SH/RH tubes corresponds to the range of typical sensitization of Type
300 stainless steels. Therefore, without changing to a less susceptible material, sensitization is
essentially a given. There are three actions that can be taken to evaluate the extent of
sensitization and characterize the boiler tube material:

7-12
13622634
 Intergranular Attack and Stress Corrosion Cracking

(g). A metallurgical examination can help determine the degree of susceptibility. French9 points
out that if the degree of sensitization is different on hot and cold sides of the tube, then the
sensitization has occurred during service and not as a result of the fabrication process, which
would result in uniform sensitization.

(h). Test the material for baseline susceptibility using the ASTM Standard Practice (A262).10

(i). Analyze the chemical composition of the failed material to see if proper material was used.

7.5 Determine the Extent of Damage

If the cracking is surface-connected, fluorescent dye penetrant or eddy-current testing can be

used carefully to detect it. Fluorescent magnetic particle is used for ferritic materials. Because of
the tightness of the cracks, methods such as radiography, and ultrasonics for detecting subsurface
damage are of limited value.

7.6 Repairs, Immediate Solutions and Actions

Seldom is there an occurrence of IGA/SCC in SH/RH circuits where only a single tube location
is affected. Outbreaks of IGA/SCC tend to be widespread and there have been instances of where
whole sections of the SH have been replaced after corrosive chemicals were carried over during
a waterwall chemical cleaning.

The primary immediate actions are to (i) determine the extent of damage, especially critical may
be bends, weld, attachments, etc. and other locations of highest stresses although, as discussed
above, this is not a simple step, (ii) make appropriate repairs or tube replacements, (iii) clean up
any sources of contamination that may have triggered the outbreak. The last is particularly apt
when a single, obvious event is at the root of the problem, e.g., improper chemical cleaning,
breakdown of inhibitors, unit startup after an improper shutdown procedure has introduced
significant contamination, or similar event.

7.7 Life Assessment

IGA and SCC are damage mechanisms which occur during specific periods of time and at
localized areas where the environment is conducive to attack (due to contaminant ingress, at
bends, attachments, etc.). Because the time of wetness is not known and the localized
environmental conditions are not well characterized, the concept of life assessment is not
applicable to this damage mechanism; any actions should focus on preventing the conditions
which promote the damage.

7.8 Longterm Actions and Prevention of Repeat Failures

Longterm actions can be reviewed for each of the contributing influences on IGA/SCC:
environment, stress, and material susceptibility.

7-13
13622634
Intergranular Attack and Stress Corrosion Cracking

7.8.1 Control of Environment

Control of environmental influences, primarily keeping contaminants out of the SH/RH, is

perhaps the most direct means to preventing IGA/SCC. A variety of steps can be taken; all are
good practices that have considerable value beyond the prevention of IGA/SCC. They include
control of steam composition during normal operation, proper execution of chemical cleaning
operations, and proper shutdown/layup procedures. Some examples are:
• Monitor SH and RH steam for composition, including key parameters such as Na and cation
conductivity levels as described in the appropriate cycle chemistry guidelines.
• A check for mechanical carryover can provide assurance in a drum unit that impurities and
concentrated boiler chemicals do not enter the steam.
• Ensure proper backfilling of SH during chemical cleaning of waterwalls, and proper conduct
of chemical cleaning of SH/RH including complete rinsing following the clean.
• Control the SH/RH atmosphere during shutdown.

7.8.2 Control of Stresses

Executing effective strategies for the control of the influence of stress on IGA/SCC in boiler
tubes is difficult. Some actions to reduce obvious stresses, either applied (such as where
attachments are poorly designed, missing, or broken) or residual (such as by the heat treatment of
bends and control of installation welding) may be performed. Heat treatment of austenitic
stainless steels is not recommended because of the likelihood that the temperatures used will
result in fully sensitized microstructures.

7.8.3 Control of Sensitized Material

Because the normal operating temperatures for SH/RH tubes are sufficient to cause sensitization
of the austenitic stainless steels, it is difficult to prevent sensitization from occurring. It may be
possible to change materials to stabilized grades of austenitic steels that can reduce sensitization.
However, this is not a "magic bullet". There have been a number of cases of SCC in Types 321
and 347.

One such example cited by French9 was of SCC in Type 347H which is normally resistant to
SCC unless elevated temperatures of at least 590 to 705°C (~ 1100 to 1300°F) have been
experienced. It was concluded that the tube had been exposed to higher than expected
temperatures as had nearby tubing that had not failed, but contain extensive SCC damage. The
tubes in question had seen 20 years life and a replacement in-kind was made with a
recommendation to minimize overfiring at low flows during startups.9

7-14
13622634
 Intergranular Attack and Stress Corrosion Cracking

7.9 Case Studies

7.9.1 Case Study I - IGSCC

Intergranular stress corrosion cracking (IGSCC) is characterized by cracks that propagate along
grain boundaries in a sensitized microstructure in a direction approximately normal to the tensile
stress. This cracking mechanism requires a susceptible material and the simultaneous action of a
corrodant and sustained tensile stress. Failure by IGSCC generally involves no appreciable
deformation or wall loss.

Intergranular corrosion by chemical attack and metal loss at grain boundaries in the absence of
significant stress is a similar degradation mechanism. Individual grains can separate from the
matrix and fall out during metallographic preparation.

Both IGSCC and creep produce intergranular cracks. IGSCC can be differentiated from creep
cracks because the latter are formed from microvoids that coalesced and linked together. Failures
that initiate by IGSCC, however, may later involve creep as the stress increases at the crack tip.

A leak developed in a wrapper tube at the bottom of a vertical platen superheater after 16 year of
operation (Figure 7-7). This tube was specified as ASME SA213-Grade TP347H with
dimensions of 1.875 inch OD and 0.260 inch MWT. The superheater outlet pressure is 2420
psig.

The tube was 0.271 inch thick at the rupture, which indicates no appreciable wall loss. An
average OD measurement of 1.851 inch indicates minor external wall loss and no appreciable
tube swelling.

Metallographic examination disclosed multiple external initiated intergranular cracks that

propagated into the tube metal (Figure 7-8). The leak resulted when one of these cracks
propagated through-wall. Since these cracks propagated without coalescence of microvoids, the
failure mechanism is identified as intergranular stress-corrosion cracking.

IGSCC requires a corrosive environment, tensile stress, and a susceptible material. Stress-
corrosion cracks can be either intergranular, transgranular, or both. Intergranular cracking
occurred in this case because of the sensitized microstructure. Stainless steels, including
stabilized grades (e.g., Types 347 and 321) can sensitize during high temperature service
exposure as chromium carbides precipitate on gain boundaries, thus, depleting the adjacent metal
of chromium. The presence of sulfur on the fracture surface suggests that polythionic acid
(H2S3-4O6) may be the corrodant.

7-15
13622634
Intergranular Attack and Stress Corrosion Cracking

Figure 7-7
An example of an intergranular stress corrosion cracking failure of a superheater wrapper
tube. No appreciable wall loss or swelling was found.

7-16
13622634
 Intergranular Attack and Stress Corrosion Cracking

Figure 7-8
Stress corrosion cracks initiated at the outside surface of the tube and propagated as
intergranular cracks without coalescence of microvoids. (Mag: (upper) 20X, (lower) 500X,
Electrolytic oxalic acid etch)

7-17
13622634
Intergranular Attack and Stress Corrosion Cracking

7.9.2 Case Study II - IGA11

Several SA-213 TP 304H horizontal reheater tubes failed due to localized IGA after 28,000
operating hours. The IGA occurred on the tops of the tubes, directly underneath Type 304
stainless steel ash erosion / sootblower-enhanced ash erosion shields that had been attached to
the tubes with straps, resulting in a window blowout-type rupture (Figure 7-9). Intergranular
attack was found, originating from the OD of the tube (Figure 7-10). The intergranular nature of
the attack was confirmed by scanning electron microscopy of the fracture surface (Figure 7-11).

Two other types of IGA / IGSCC failures were experienced in this reheater. The first of these
occurred where the tubing penetrated the sidewall and moisture accumulated between the tube
and the insulation. Water washing during outages was believed to have provided the source of
moisture. After water washing was discontinued and the erosion shields were removed, no
further failures occurred from the OD surface. After additional service, ID-initiated IGA /
IGSCC failures occurred in regions near welded tube attachments. On the tube ID surface, areas
with red rust exhibited IGA with depths up to 1.5 mm (0.06 inch). Where these locations
corresponded with the tube attachments, additional crack growth by IGSCC occurred, leading to
circumferential through-wall failures.

Figure 7-9
Window blowout-type tube failure on the top of a horizontal reheater tube.

7-18
13622634
 Intergranular Attack and Stress Corrosion Cracking

Figure 7-10
Photomicrograph of OD-originating intergranular attack observed on area near window
fracture shown in Figure 7-9 (100x, unetched).

Figure 7-11
SEM photomicrograph intergranular morphology of the fracture surface.

7-19
13622634
Intergranular Attack and Stress Corrosion Cracking

7.10 References
1
Paterson, S.R., T.A. Kuntz, R.S. Moser, and H. Vaillancourt, Boiler Tube Failure Metallurgical
Guide, Volume 1: Technical Report, Volume 2: Appendices, Research Project 1890-09, Final
Report TR-102433, Electric Power Research Institute, Palo Alto, CA, October, 1993.
2
Personal Communications from K. Woolhouse (FCA) and D. French (David N. French, Inc.) to
R.B. Dooley, February, 1995.
3
Logan, H.L., The Stress Corrosion of Metals, Wiley, New York, 1966.
4
Macdonald, D.D. and G.A. Cragnolino, "Corrosion of Steam Cycle Materials", Chapter 9 in The
ASME Handbook on Water Technology For Thermal Power Systems, American Society of
Mechanical Engineers, New York, 1989.
5
Staehle, R.W., A.J. Forty, and D. Van Rooyen, eds., Fundamental Aspects of Stress Corrosion
Cracking, National Association of Corrosion Engineers, Houston, TX, 1969.
6
Staehle, R.W., J. Hockmann, R.D. McCright, J.E. Slater, eds. Stress Corrosion Cracking and
Hydrogen Embrittlement of Iron Base Alloys, National Association of Corrosion Engineers,
Houston, TX, 1977.
7
Capener, E.L. and R.H. Richman, "A Re-Analysis of Results from Constant Extension Rate
Tests on Sensitized AISI 304 Stainless Steel", NACE Corrosion, Vol. 42, No. 8, 1986, pp. 482-
489.
8
Martynova, O.I. and A.B. Vainman, "Some Problems of Oxygenated Treatment Use in
Supercritical Units", Thermal Engineering, Volume 41, No. 7, 1994, pp. 497-503.
9
French, S.M., "Intergranular Failure of Tube 347H Stainless Steel Superheater Tubes", in B.
Dooley, ed., Proceedings: International Conference on Boiler Tube Failures in Fossil Plants,
held in San Diego, California November 5-7, 1991, Proceedings TR-100493, Electric Power
Research Institute, Palo Alto, CA, April, 1992, pp. 4-17 through 4-27.
10
American Society for Testing and Materials, Standard A262-93a (1993), "Standard Practices
for Detecting Susceptibility to Intergranular Attack in Austenitic Stainless Steels", 1994 Annual
Book of ASTM Standards, American Society for Testing and Materials, Philadelphia, PA, 1994.

11Waddingham, A., presentation made at the International Conference on Boiler Tube Failures
and HRSG Tube Failures and Inspection, November 6-8, 2001, Phoenix, AZ.

7-20
13622634
8
FATIGUE

8.1 Introduction

This chapter discusses boiler tube failures caused by fatigue, without regard to the source of the
cyclic stresses (thermal, mechanical, vibration, etc.). The common elements of failures in this
category are that they are manifested by cracking that is OD-initiated, are predominantly
transgranular, and are not primarily attributable to the effect of either the fireside or fluid
environment. That is, for the failures described here, the effect of environment is small enough
that even if all environmental effects were removed, the failure would still occur. Cyclic stresses
are obviously also an important contributor to cracking mechanisms that do have a major
contribution from the environment, such as stress corrosion cracking (Chapter 7). This Chapter
provides some means to distinguish among these effects; however, it concentrates on fatigue as
the dominant failure mechanism.

The generic BTF caused by fatigue, reviewed in this Chapter, are divided into two general
classifications: tubing (attachment)-related and header-related.

8.2 Features of Failure and Typical Locations

8.2.1 Features of Failure

Fatigue failures, either tubing- or header-related, generally result in thick-edged failures (Figure
8-1). Cracking is predominantly straight and transgranular, although it may also be intergranular
particularly at higher temperatures in the SH/RH sections. Cracks may be filled with oxide,
depending upon the service conditions.

8-1
13622634
Fatigue

Figure 8-1
Thermo-mechanical failure of an SA-213 Type 304H superheater tube. The portion of the
rupture which is missing was believed to have contained a welded attachment clip.
Source: S.R. Paterson, et al.1

Fatigue failures originate on the outside surfaces of tubes. The appearance of beach marks and
ratchet marks is typical, although they may be totally obliterated by oxidation. Fatigue cracks can
also occur in material that has been damaged by creep that is primarily stress-driven in which the
cracking will be intergranular and will occur in conjunction with void formation.

8.2.2 Typical Locations

Tubing-related failures are generally found in locations associated with (i) attachments,
particularly solid attachments or jammed sliding attachments, or (ii) bends in tubing.

At attachments, failures by fatigue are often associated with weldments, particularly in cases
where the weld or condition of the attachment does not allow for thermal expansion, such as
shown in Figure 8-2.

8-2
13622634
 Fatigue

Figure 8-2
Schematic showing typical locations of fatigue failures in steam-touched tubing. Source:
R.B. Dooley and H.J. Westwood.2

At bends in the tubing, particularly "U"-bends, failures can be initiated at the intrados, extrados
or neutral axis as shown in Figure 8-3. If failures occur in U-bends, they are usually those in the
cooler (below creep temperature) regions of the primary SH or RH.

Figure 8-3
Three possible locations for tubing-related fatigue failures in tight 180° bends.

Header-related fatigue failures in tubes occur most frequently at the ends of the header and are
related to restricted flexibility of the header and/or thermal expansion stresses. The problem of
header-related fatigue failures is more prevalent in SH/RH than in economizers. Two common
locations have been experienced. The first is in nipple or stub welds, as illustrated in Figure 8-4.
As tube temperatures increase, the tubes try to lengthen. Temperature differentials lead to
constraint from expansion of the hotter tubes and stretching of the cooler tubes and a resulting
strain with each thermal cycle. The cyclic strain leads to failures at locations such as at the toe of
header-to-tube welds, shown in Figure 8-4.

8-3
13622634
Fatigue

Figure 8-4
Fatigue tube failures caused by differential thermal expansion of element transfer tubes on
a header. Source: W.R. Sylvester, et al.3

A second form of cracking is caused by the lateral expansion of the header as shown in Figure
8-5. Here connecting tubes that penetrate through the waterwall or roof tubes will be subjected
to bending moments when the header expands more than the tubes; the effect is worst at the ends
of the header. During normal operation the strains are modest since temperatures are steady;
however the problem is exacerbated by two-shifting or cycling operation. Cracking is OD-
initiated, oriented circumferentially, and generally found at the toe of welds.

Figure 8-5
Schematic illustrating failures due to inflexibility of movement between header and
waterwall.

8-4
13622634
 Fatigue

Vibration fatigue failures induced by gas-flow usually occur in vertical screen tubes and
horizontal rear pass tubes and rarely in pendant sections. They are also more common in SH/RH
sections than in economizers.

Poor weld geometry or execution (overfill, poor fillet weld profile, undercut) may also lie at the
root of a fatigue problem.

8.3 Mechanism of Failure

Fatigue is the phenomenon of damage accumulation caused by cyclic or fluctuating stresses. It is

manifested as the initiation and stable propagation of a crack. Final failure in a component
subjected to fatigue ensues when a critical crack size is reached and failure occurs by fracture or
overload. Fatigue is dependent upon the frequency and magnitude of the stress cycles and is
generally independent of stress duration. Obviously, at high temperatures where creep fatigue
becomes the dominant mechanism, there is dependence on stress duration. Stress cycling may be
induced mechanically or thermally.

Fatigue in boiler tubes is generally driven by excessive local stress levels, typically at welded
connections. Such connections may be tube attachments and supports, or at header nipples and
stub welds, as described above. Stresses may be induced by excessive (unanticipated)
mechanical loads, restrained thermal expansion, vibration such as produced by the flow of
combustion gases, or as a result of poor weld geometry in which the global stresses are within
the design, but strains are locally excessive because of the geometry of the joint. In many cases,
the excessive stresses, particularly thermally-induced, are produced by unit cycling or two-shift
operation leading to conditions that did not arise during baseload operation.

8.4 Possible Root Causes and Actions to Confirm

8.4.1 Introduction

Unanticipated or excessive cyclic stresses (strains) are the primary cause of fatigue failures in
boiler tubes. Three basic sources of excessive strains are those caused by (i) constraint of thermal
expansion, (ii) excess mechanical loads, or (iii) flow-induced vibration. The first of these is
responsible for the bulk of tube failures occurring by this mechanism. Excessive stresses can also
arise in welded connections if the expansion of the welded joint is inadequate. These four causes
are discussed below. A summary of the appropriate actions to confirm, immediate actions and
longer term measures to prevent future failures are summarized in Table 8-1.

8-5
13622634
Fatigue

Table 8-1
Major Root Cause Influences, Confirmation and Corrective Actions

Major Root Cause Actions to Confirm Immediate Actions Longterm Actions and
Influences and Solutions Prevention of Repeat
Failure
8.4.2 Excessive strains (a) Visual examination Identify similar damaged Evaluate modifications
caused by constraint of for distortion or locations. to attachment design
thermal expansion. bending in adjacent Repair/replace affected or to header/tube
tubes. tubes. connection to reduce
(b) Strain gauging of stress levels.
suspect locations to Institute periodic
evaluate strains inspection program,
during unit starts particularly of
and cycling susceptible header
operation. locations in units that
(c) LVDT are now, or will by
measurements to cycled.
monitor the relative Improve header/tube
movement of the flexibility and confirm
header/tube during with LVDT.
transients;
particularly whether
header support
allows for
expansion.
8.4.3 Poor design (d) Strain gauging to As above. Evaluate modifications
and/or manufacture measure actual to attachment design
giving rise to excessive strains experienced or to header/tube
mechanical stresses. at the local area connection to reduce
during operation. stress levels.
(e) As in (c) above. Institute periodic
(f) For tight, hair-pin inspection program,
bends, determine particularly of
whether residual susceptible header
stresses are high. locations in units that
are now, or will be
cycled.
8.4.4 Vibration (flue (g) Metallurgical As above. Evaluate and install
gas-induced) by direct examination to modifications such
flow or vortex shedding. determine high as snubbers or
cycle fatigue. vibration restraints to
(h) Estimate natural reduce stresses
and forcing induced by vibration.
frequencies and
confirm by test.
8.4.5 Poor welding, (i) Visual and As above. Institute program of
particularly poor microscopic weld quality control
geometry of final joint. examination of weld based on guidelines
quality. such as provided in
Reference 4.

8-6
13622634
 Fatigue

8.4.2 Excessive Strains Caused by Constraint of Thermal Expansion

The typical locations for the occurrence of this problem were discussed above. Actions to
confirm this root cause include:

(a). Visual examination to detect distortion or bending of adjacent tubes. Tubes at headers may
also be physically distorted or even pulled away.

(b). Strain gauging of susceptible locations can indicate the level of strains experienced during
thermal excursions of the unit such as cycling operation.

(c). LVDT measurements to monitor the relative movement of the header/ tube during transients,
particularly to ensure that the header support allows for expansion.

8.4.3 Poor Design and/or Manufacture Giving Rise to Excessive Mechanical

Stresses

Actions to confirm this root cause include:

(d). Strain gauging of susceptible locations for indication of unexpected mechanical loads in
excess of design assumptions.

(e). as in (c) above.

(f). Tight hair-pin bends may contain high residual stresses. This can be determined qualitatively
by observing the extent of "spring-apart" that occurs when the tube is cut. Quantitatively, rosette
strain gauges can be used to measure strains. Typically such gauges are placed on the location of
interest and the center drilled so that the strain thus relieved can be evaluated.

8.4.4 Vibration (Flue Gas-Induced) by Direct Flow or Vortex Shedding

Actions to confirm this root cause include:

(g). Metallurgical examination and location usually provide confirmation. Also if vibration is
occurring, it is often extremely noisy and thus obvious.

(h). Analytic estimate of natural frequency and forcing frequency followed by confirmation by
cold or hot flow testing with strain gages and accelerometers.

8.4.5 Poor Welding, Particularly Poor Geometry of Final Joint

Actions to confirm this root cause include:

(i). Analysis of weld profile and quality to determine whether poor welding was a likely
contributor to the failure.

8-7
13622634
Fatigue

8.5 Determining the Extent of Damage

Detection of fatigue damage once a crack has initiated is well established by a number of
methods including, liquid penetrant, eddy current, ultrasonic testing, and radiography. Since
fatigue cracks in boiler tubes are OD-initiated identification of cracked locations is possible
within the constraints of access. The analysis of fatigue is likewise well established and
sufficient materials data exist for all common constructional materials to allow complete analysis
of the expected life of components subject to fatigue.

8.6 Repairs, Immediate Solutions and Actions

Cracks can usually be ground out and rewelded as a temporary measure. Tight "U"-bends should
be replaced. Most of the rest of the actions can be taken over the longer term.

8.7 Life Assessment

The analysis of fatigue is well established and sufficient materials property data exist for all
common constructional materials to allow complete analysis of the expected life of components
subject to fatigue. In general components subject to high cycle fatigue are designed to prevent
crack initiation because once a fatigue crack initiates, it can propagate to failure with fewer load
cycles than were required for crack initiation. Components that have failed by fatigue should be
repaired or replaced and the source of the cyclic stress that resulted in fatigue failure should be
identified and eliminated.

The potential contribution from exposure to an elevated temperature should be included if a life
assessment program is performed. Components that are subject to thermal induced stresses may
operate at sufficiently high temperatures that the interaction of fatigue and creep may
significantly reduce the component life. Several fatigue damage rules have been established to
estimate life assessment predictions for creep-fatigue interactions. These include the summation
5
of fatigue and creep damage by life fractions, strain fractions, and other analytical methods .

Figures 8-6 through 8-11 show the analysis steps required for a fatigue analysis of a superheater
or reheater tube. The steps shown generally follow the guidelines in ASME Section III,
Appendix NH. There are numerous methods and models in the literature describing fatigue
analysis of materials that operate at elevated temperatures. The popular models all have specific
methods for addressing and combining the elastic, inelastic, and time-dependent portions of the
stress/strain cycles. For example, the method used in ASME Section III, Appendix NH uses a
creep fatigue interaction curve (Figure 8-11) but also requires that the time dependent portion of
the strain range be included in the estimation of the fatigue damage fraction, DF. The creep
damage fraction, DC, used in this approach is based on a stress magnitude that is integrated over
each loading cycle rather than using the nominal, sustained load stress. The reader is cautioned
that it is generally not appropriate to combine the fatigue damage fraction from one model with
the creep damage fraction from a different model, since the creep fatigue interaction factors are
unique to the overall features of the combined model. The same considerations are also relevant
to the specific way that the multiaxial stresses/strains are addressed in each specific model.

8-8
13622634
 Fatigue

8.8 Longterm Actions and Prevention of Repeat Failures

Longer-term actions will consist of modifications to prevent a recurrence of the problem. For
example, if the failure was at an attachment and strains were caused by differential thermal
expansion, the attachment design should be modified to reduce the stress level, e.g. replacement
of large C and T sliding spacers can be by a "ladder"-type support or one that allows free
expansion.2 It is important to establish clearly that the stresses in the new design are in fact
lower than those of the original design.

In the case of vibration-induced strains, solutions are typically found in the installation of
vibration baffles to break up the critical vibration modes for gas flow, or by using snubbers and
vibration restraints.

A periodic inspection program is indicated for susceptible locations, particularly in the header-
related locations, and if the unit is cycled.

If the unit has recently been or is to be changed to cycling or to two-shift operation, it may be
advisable to perform an analysis of the stresses induced by constraint at potential failure
locations identified in Figure 8-4. Monitoring the relative movement of header/tube during
transients with LVDTs is typical.

8-9
13622634
Fatigue

Figure 8-6
Flowchart of the creep-fatigue modeling steps based on the ASME Section III, Appendix
NH guidelines.

8-10
13622634
 Fatigue

Figure 8-7
Flowchart of the creep-fatigue modeling steps based on the ASME Section III, Appendix
NH guidelines.

8-11
13622634
Fatigue

Figure 8-8
Flowchart of the creep-fatigue modeling steps based on the ASME Section III, Appendix
NH guidelines.

8-12
13622634
 Fatigue

Figure 8-9
Flowchart of the creep-fatigue modeling steps based on the ASME Section III, Appendix
NH guidelines.

8-13
13622634
Fatigue

Figure 8-10
Flowchart of the creep-fatigue modeling steps based on the ASME Section III, Appendix
NH guidelines.

8-14
13622634
 Fatigue

Figure 8-11
Flowchart of the creep-fatigue modeling steps based on the ASME Section III, Appendix
NH guidelines.

8-15
13622634
Fatigue

8.9 Case Study

Fatigue failures are characterized by little or no visible deformation and typically exhibit
crescent shaped beach marks (crack arrest lines) on the fracture surface that are centered about
the crack initiation. Cross-sections cut through the cracks will reveal a relatively straight,
nonbranching, transgranular crack morphology. These cracks typically initiate at a stress raiser,
such as an attachment weld or at the toe of a circumferential tube weld.

A steam leak near the bottom of a vertical pendant in the finishing superheater caused a forced
outage. This section of tube, which had been in service for six years, was specified as ASME
SA213-Grade TP304H with dimensions of 2.50 inch OD and 0.371 inch MWT.

Two crescent shaped regions that emanated from the toe of a circumferential weld were visually
identified as probable crack initiation locations (Figure 8-12). Tube dimensions, including at the
rupture, complied with the specification requirements and showed no appreciable wall thinning
or swelling. Metallographic examination revealed equiaxed austenitic grains without significant
sigma phase formation, characteristic of a tube that had not been overheated in service.
Examination of the intact side of the circumferential weld disclosed a relatively straight,
nonbranching transgranular crack (Figure 8-13) that initiated at the weld toe, which confirms the
failure mechanism as fatigue.

Figure 8-12
An example of a failure that initiated at the toe of a circumferential butt weld in a TP304H
finishing superheater tube after six years of service. Flat crescent shaped regions were
identified as the crack initiation locations.

8-16
13622634
 Fatigue

Figure 8-13
The fatigue crack initiated at the weld toe and propagated in a nonbranching,
transgranular cracking mode. (Mag: 20X, Electrolytic Oxalic Acid Etch)

8.10 References
1
Paterson, S.R., T.A. Kuntz, R.S. Moser, and H. Vaillancourt, Boiler Tube Failure Metallurgical
Guide, Volume 1: Technical Report, Volume 2: Appendices, Research Project 1890-09, Final
Report TR-102433, Electric Power Research Institute, Palo Alto, CA, October, 1993.
2
Dooley, R.B. and H.J. Westwood, Analysis and Prevention of Boiler Tube Failures, Ontario
Hydro, Report 83/237G-31, Canadian Electrical Association, Montreal, Quebec, November,
1983.
3
Sylvester, W.R., et al., "Locate Problems in Radiant and Convective Sections Early to Improve
Availability", Power, March, 1978.

8-17
13622634
Fatigue

4
Stephenson, G.G. and J.W. Prince, Guidelines on Fossil Boiler Field Welding, Research Project
2504-02, Final Report TR-101699, Electric Power Research Institute, Palo Alto, CA, January,
1993.
5
Viswanathan, R., Damage Mechanisms and Life Assessment of High-Temperature Components,
ASM International

8-18
13622634
9
STEAMSIDE SCALE EXFOLIATION, STEAM FLOW
BLOCKAGE, AND SHORT TERM OVERHEATING

9.1 Introduction

In addition to microstructural changes, austenitic stainless steel tubes will experience the growth
of fireside and steamside oxide scale during service. The fireside scale will typically have an
insignificant effect on the life of the tubing unless the fireside ash disrupts its growth and
dissolves away or creates a less protective scale such as iron sulfide. In this case rapid fireside
corrosion may occur. Mechanical disruption of the fireside scale by ash erosion or sootblower
enhanced ash erosion is also possible but usually less of an issue than can occur on low alloy
ferritic tubing.

Steamside oxide growth on austenitic stainless steel tubes is normally very slow relative to
ferritic materials and doesn’t cause much change in the temperature of the tubing. The primary
concern with the steamside oxide scale is exfoliation leading to tube blockage, severe reduction
in steam flow and short term overheating of the blocked tube circuits. Excessive tube
temperatures and subsequent failures can occur in a very short period of time. This is distinct
from longterm overheating/creep, described in Chapter 11, which is gradual and longterm
damage that occurs as a result of moderate overheating from a number of causes, including steam
flow restrictions, unbalanced steam distribution through the steam circuits, the insulating
characteristics of steamside oxide scale, and poor initial design. Solid particle erosion of first HP
or IP turbine stages, valves, nozzles and other downstream components is also of concern.

9.2 Features of Failure and Typical Locations

9.2.1 Features of Failure

Damage by short-term overheating will display distinctive characteristics including: (i) a

considerable increase in the diameter of affected tubes (swelling) without ovalization, (ii) a
ductile final failure showing a thin-edged fracture surface, and (iii) "fish-mouth" appearance.
Failures are generally oriented longitudinally.

Microstructural examination will display evidence of severe overheating of the tube metal.
Depending on the composition of the material and the time over which the excessive
temperatures occurred there may be microstructural changes similar to those for longterm
overheating, such as the presence of sigma phase.

9-1
13622634
Steamside Scale Exfoliation, Steam Flow Blockage, and Short Term Overheating

The key factor that distinguishes longterm and short-term overheating is short-term overheating
is a ductile, thin-edged failure whereas longterm overheating is manifested as brittle, thick-edged
failures.

Examination of debris contained in removed sections of tubing may provide clues about the
cause of the problem. For example, bottom bends may be blocked by exfoliated oxide. Analysis
of the exfoliated oxide, for example, can indicate whether its primary origin is austenitic or
ferritic tubing materials.

9.2.2 Locations of Failure

The most common location of failure is near the bottom bends in vertical loops of the SH/RH
(Figure 9-1). Failures have also occurred in outlet tube legs.

Figure 9-1
Typical locations of failures by short term overheating in the superheater/reheater.
Exfoliating oxide in the austenitic leg AB drops to bend B during shutdown and causes
blockage. Failure typically occurs at “X” in the ferritic material.

9.3 Mechanism of Failure

During the initial stages of growth, the steamside oxide scale will form a coherent three layer
scale. The portion of the oxide scale that penetrates into original tubing material will be a spinel
(Fe,Cr,Mo)3O4 and will tend to grow in a very irregular profile (Figure 9-2) that is a function of
the orientation of the underlying grain structure. This portion of the oxide will have a chemical
composition that is similar (but with slight enrichment in chromium content) to the tubing
material. The outwardly growing scale will be comprised of a magnetite (Fe3O4) layer and a
much thinner hematite (Fe2O3) layer. As the oxide thickens during prolonged exposure to service

9-2
13622634
 Steamside Scale Exfoliation, Steam Flow Blockage, and Short Term Overheating

temperatures, porosity will tend to form at the metal/oxide and spinel/magnetite interfaces. This
will tend to limit the outward diffusion of iron ions and the hematite islands on the outside
surface will increase in thickness. Since the thermal expansion coefficient of hematite is lower
than magnetite (see Appendix A, Section A.3.3), an increase in hematite thickness will increase
the strain energy that occurs during cooling of the tubing during shutdown events. The strain
energy present in the oxide during cooling is also dependent on the total thickness of the oxide
scale. When the oxide scale thickness, hematite/magnetite thickness ratio and resultant cooling
strains reach a critical level then exfoliation of the oxide will occur during shutdown events. All
of the exfoliation will occur at the spinel/magnetite interface, leaving the spinel layer attached to
the tubing. The pieces of exfoliated scale can be quite large, up to 2.5 cm (1 inch) across. Note
that the growth and exfoliation of the outer oxide layer is not dependent on the composition of
the steam. Specifically, the growth and exfoliation characteristics of steamside oxides in
oxygenated units has been identical to that seen in sister units that are not on oxygenated
treatment.

Long periods of continuous operation can result in a major exfoliation event. With prolonged
operation with no major shutdown transients, there is increasing voidage at the Fe3O4/FeCr spinel
interface and increasing amounts of Fe2O3 at the oxide/steam interface. Eventually, when the
density of voids becomes large (or nearly continuous), the thermal transient experienced at a
shutdown can result in massive exfoliation of scale.

If the amount of steamside oxide exfoliation is severe, then lower tube bends may fill with oxide
(Figure 9-3). During restarting of the unit, the oxide may either blow clear or may remain in the
lower tube bends and restrict steam flow, causing overheating of the blocked tube circuits. In
severe cases, tube failures and severe tube deformation (Figure 9-4) may occur.

Since the growth of the steamside oxide scales is very irregular and the scale thickness is in
general very thin (Figure 9-5) it is not possible to use measurements of the scale thickness to
infer the temperature history of the tubing.

9-3
13622634
Steamside Scale Exfoliation, Steam Flow Blockage, and Short Term Overheating

Figure 9-2
Growth and exfoliation of steamside oxide scale on austenitic stainless steel tubing4. For
austenitic tubes the thermal contraction of the tubing (during cooling) is greater than the
steamside oxide. This results in a compressive strain in the oxide of approximately
0.002mm/mm during large cooling transients. Under these conditions exfoliation will begin
to occur when the steamside scale thickness is approximately 75µm (0.003-inch). The
shaded box shown on the tension side of the diagram is characteristic of the growth,
strain, and exfoliation of ferritic tubing.

9-4
13622634
 Steamside Scale Exfoliation, Steam Flow Blockage, and Short Term Overheating

Figure 9-3
Steam flow blockage caused by severe steamside scale exfoliation.

Figure 9-4
Short-term overheating caused by severe steamside oxide scale exfoliation.

9-5
13622634
Steamside Scale Exfoliation, Steam Flow Blockage, and Short Term Overheating

Figure 9-5
Irregular Steamside Oxide Growth on Type 304, 321, and 347 Tubing. (3)

9.4 Possible Root Causes and Actions to Confirm

Tube blockage is the most common cause of tube failures by short-term overheating in the
SH/RH. The growth and exfoliation of oxide in SH/RH circuits is a normal event of operating
the materials of construction at high temperatures. The rate of scale growth and rate of
exfoliation depends on the material of construction1,2. Under normal operating conditions the
scale that grows in the SH/RH is the greatest concern. The two layer scale that is produced in
austenitic SH/RH tube materials is very susceptible to exfoliation when subjected to thermal
transients such as occur during shutdown and rapid cooling events. This is because of the marked
difference in the coefficient of thermal expansion between the scale and the austenitic stainless
steel base metal.

Once exfoliation of oxide flakes has occurred, the propensity for tube blockages is a function of4:
flake size and shape, strength and stiffness of the flakes, tube internal radius, constrictions and
imperfections in the tube, flow rates in the tube, friction coefficients, and total quantity of spalled
oxide in the tube. The size of the exfoliated particles can be large in comparison to scale
exfoliated from ferritic materials, up to 2.5 cm (~ 1 in.) in diameter. As a result, the accumulation

9-6
13622634
 Steamside Scale Exfoliation, Steam Flow Blockage, and Short Term Overheating

of these large particles, particularly at bends in smaller diameter tubing, can more readily result
in tube blockage. On startup, steam flow is limited through the blocked tube(s).

The ability of a blockage to withstand boiler pressures depends on the orientation of the flakes.
At low flows, flakes tumble until they become oriented in a hydrodynamically stable position;
once these positions are reached, subsequent high flows do not always disturb them. As a result
tube blockages can occur when these stable positions are reached during the low steam flow
associated with boiler startup. Flake transport studies4 found that an overlapping wedge of flakes
formed at low speeds in an arch formation. The most important parameter in blockage formation
was the ratio of flake dimension to tube radius. The ability of the arches to withstand subsequent
fluid forces was a function of flake thickness; 100 µm flakes were likely to withstand breakage,
20 µm thick flakes were likely to break.

Failure will usually occur away from the blockage, typically in locations where there is a
transition to a lower grade (less creep resistant) material, or in changes of wall thickness in the
same grade material.

Actions to confirm this root cause:

(a). Removal of the damaged tube to confirm the source of blockage. It is important to
distinguish between sources of blockage in order to apply the proper corrective action. For
example, steamside oxide from austenitic or ferritic origin should be distinguished from debris
remaining after chemical cleaning.

Metallurgical examination may be required to confirm the source of the blockage, including
energy dispersive X-ray spectroscopy to characterize the elemental composition of the blockage
material. Generally exfoliated oxide from austenitic materials consists only of Fe3O4, whereas
exfoliated oxide from ferritic materials usually consists of at least two layers, one of Fe3O4 and
one of an iron chromium spinel. Very often the oxide from ferritics is multilaminated.

(b). Examination for evidence of exfoliation. If exfoliation is occurring, confirmation can be

obtained by direct sampling of candidate tubing or through the interpretation of other signals
from the unit. For example, has there been solid particle erosion in the HP or IP turbine or
valves?

(c). NDE examination, for example by radiography, to confirm the existence of similar blockages
that have not lead to failure. It may be possible to examine nondestructively similar locations for
evidence of tube blockage. Flow verification can be done by running warm water through
suspect sections and checking by hand or by thermal monitors to see if tubes are also warm,
indicating adequate flow.

(d). A hand held magnet will detect spalled oxide in bends of austenitic tube materials.

9-7
13622634
Steamside Scale Exfoliation, Steam Flow Blockage, and Short Term Overheating

9.5 Determining the Extent of Damage

Inspection for additional blockage locations generally requires 100% coverage, although it may
be possible to identify patterns of damage, or isolated affected tubes if, for example, maintenance
errors are at fault. Inspection for signs of blockage in austenitic loops can be performed using (i)
magnets on the tube, (ii) radiography, (iii) tube removal, or (iv) measurement of tube outside
diameters for signs of swelling. A recommended procedure used to determine whether tubes are
blocked after a chemical cleaning is to circulate water of changing temperature. A "hand check"
or measurement of tube temperatures using a thermal meter can detect tubes for which there is no
flow.

9.6 Repairs, Immediate Solutions and Actions

It is important for this mechanism that immediate identification of the problem be made. A
typical field experience is a repetitive cycle of (i) tube failure, (ii) repair/replace the affected tube
bend, (iii) unit online, and (iv) additional tube failures. Both the specific root cause and the
extent of the problem should be addressed before the unit goes back online in order to avoid the
predictable forced outage that will occur otherwise.

One means that has been used to evaluate whether all damaged material has been removed is to
measure the outside diameter of the tubing in which clear evidence of swelling has occurred and
replace those tubes where the original diameter has increased by more than 1%.5

Removal of all tube blockages and repair/replacement of all affected tube bends should be
performed prior to the unit being returned to service.

If the problem is exfoliation from austenitic stainless steel material, an interim procedure that
limits large temperature differentials such as occur during forced cool downs can be imposed,
but only until a longer term approach is defined. This can be helpful since such scales are more
sensitive to thermal transients than those for the ferritic tube materials.

9.7 Life Assessment

Damage to high temperature furnace tubes that have been adversely affected by steam flow
blockage from exfoliated scale may range from the accumulation of minor creep damage to
creep-rupture failures. Several methods are available to evaluate the extent of damage for tubes
that have not failed. These include tube swelling based on diameter measurements, as discussed
above, and creep damage estimates based on microstructural examinations. Although
microstructural evaluations can be performed using a nondestructive surface replica technique,
these evaluations are best performed on the cross-sections of short lengths of tube samples
removed from the furnace, because of surface effects that may alter the apparent creep damage in
the surface replicated microstructures. Once the creep damage is quantified, a life fraction creep
model (e.g., EPRI TubeLife III) can project the future accumulation of creep damage with the
appropriate and possibly changing stress and temperature to estimate a predicted failure time.

9-8
13622634
 Steamside Scale Exfoliation, Steam Flow Blockage, and Short Term Overheating

9.8 Longterm Actions and Prevention of Repeat Failures

If the failure was caused by the exfoliation of built-up steamside oxide, a longterm program to
monitor and prevent recurrence of the problem is required. See the program described in the case
study for a view of an integrated approach to solving the problem. For severe oxide buildup a
chemical clean of the SH/RH is indicated. A detailed manual of practice is available.6

Once the situation has been controlled, a monitoring scheme should be developed to determine
the severity of oxide buildup in affected tubing. Such monitoring may consist of tube sampling to
confirm the type and extent of scale, and temperature determination such as by direct
thermocouple measurement or analysis of oxide scale buildup.

9-9
13622634
Steamside Scale Exfoliation, Steam Flow Blockage, and Short Term Overheating

9.9 Case Study

Superheater/Reheater Short Term Overheating Case Study:
Analysis of Field Failures and Chemical Cleaning of the SH
Background. Three supercritical units, rated at 600 MW and with about 25,000
operating hours began to manifest tube failures in the secondary superheaters. A total
of fifteen tubes burst, failure always occurred during cold startup of the units. The
failures were located in SA213-T12 ferritic material. Blockages were eventually traced to
Type 304 austenitic stainless steel, which makes up about 45% of the secondary
superheater material. Minimum tube inside diameter in tubes of Type 304 was 22.8 mm
(0.90 in.); minimum bend radius is 28.5 mm (1.12 in.).
Appearance of Failures. Failures showed classic signs of short-term overheating
including fish-mouth openings with thin edges and considerable increases in tube
diameters (up to 40%) with no sign of ovalization. Microstructural examination of the
T12 showed transformation indicating that its critical temperature, Ac1, which is 740°C
(1364°F), had been exceeded. Hardness had increased from 160 to 260 Vickers. An
oxide scale had formed rapidly. It was estimated from morphology and thickness to
have occurred in less than 15 minutes.
Inspection Results. A 100% gamma-ray examination of the lower bends was used to
identify blocked tubes. Some 50% of tube bends indicated partial or total blockage with
27 bends totally blocked. Measurement of the outside diameter of T12 and T22 tubes
showed sixteen swelled tubes. Selective sampling was performed on both Type 304
and ferritic tubes.
Analysis of Internal Scale. The formation of a duplex oxide layer was found on Type 304
samples, with only the external layer (nearest to the steam) showing signs of exfoliation.
Tube blockage was caused by this exfoliated scale as determined by analysis of the
chemical composition of the blockage material which was essentially only Fe3O4. Some
of the scale was up to 10 mm (0.39 in.) in width.
Resolution: Air blowing was used to clear all tubes where inspection had indicated a
blockage of 40% or greater. An interim operating procedure was used to limit
temperature excursions until the corrective actions were finalized. Emphasis was on
NDE to determine the extent of damage and periodical chemical cleaning of the
secondary superheater. A chemical cleaning procedure was developed. Hydroxyacetic
formic acid (6%) was used without ammonium bifluoride. Extensive monitoring was
performed during and after the cleaning process including: temperature, pH, total acidity
measurements, analysis of Fe, Cr and Ni ions, flow-rate measurements, continuous
corrosion rate monitoring, thermal monitoring of tube, and post-test sampling. The
results indicated excellent cleaning of T12, T22, and Type 304 materials.

9.10 References
1
Dooley, R.B., F.M Somerset, A.C. Davidson, V.R.D. Miller and A.V. Manolescu, "Ontario
Hydro's Experience with Solid Particle Erosion of Steam Turbines", in 1980 EPRI/ASME

9-10
13622634
 Steamside Scale Exfoliation, Steam Flow Blockage, and Short Term Overheating

Workshop on Solid Particle Erosion of Steam Turbine Components, Report CS-4683, Electric
Power Research Institute, Palo Alto, CA, 1980, p. 3-36.
2
Miller, V.R.D and A.V. Manolescu, "Maintenance Strategies and Specific Design Changes to
Ontario Hydro's Lambton TGS Coal-Fired Generating Units to Mitigate the Effects of Solid
Particle Erosion", Solid Particle Erosion of Steam Turbine Components: 1985 Workshop, Report
CS-4683, Electric Power Research Institute, Palo Alto, CA, 1985, p. 2-21.
3
Eberle, F., and J.H. Kitterman, Scale Formation on Superheater Alloys Exposed to High
Temperature Steam, ASME Conference on the Behavior of Superheater Alloys in High
Temperature, High Pressure Steam, p. 67-113.
4
Armitt, J, R. Holmes, M.I. Manning, D.B. Meadowcroft, and E. Metcalfe, The Spalling of
Steam-Grown Oxide from Superheater and Reheater Tube Steels, Technical Planning Study 76-
655, Final Report FP-686, Electric Power Research Institute, Palo Alto, CA, February, 1978.
5
Personal Communication D. Barnett (Pacific Power) to R.B. Dooley, February, 1995.
6
Bartholomew, R.D, W.E. Chesney, R.D. Hopkins, J.S. Poole, J.W. Siegmund, J.P. Williams,
and S. Yorgiadis, Guidelines for Chemical Cleaning of Fossil-Fueled Steam-Generating
Equipment, Research Project 2712-06, Final Report TR-102401, Electric Power Research
Institute, Palo Alto, CA, June, 1993.

9-11
13622634
13622634
10
DISSIMILAR METAL WELD FAILURES

10.1 Introduction

Dissimilar metal welds (DMW) are used to join ferritic and austenitic steel tubing in the final
outlet sections of superheaters (SH) and reheaters (RH). Prior to the mid-1980s, numerous
outages occurred as a result of failures via low ductility cracking in the low alloy ferritic steel
immediately adjacent to the weld fusion line. The problem was widespread; it was estimated at
the time that 50% of all boilers in North America with austenitic tubing had DMW failures.1
There was wide variability in the times to first failure ranging from as few as 30,000 hours to
over 150,000 hours.

A comprehensive effort over the last ten years has characterized the problem of DMW failures,
provided a set of analytical tools for analysis, and provided appropriate solutions to the
problem.2-4

10.2 Features of Failure and Typical Locations

10.2.1 Features of Failure

Dissimilar metal welds (DMW) are found in the superheater/reheater where austenitic materials,
generally Types 304H, 321H or 347H, are joined with ferritic materials, generally T22, T5, or
T9. There are three types of DMW. The most common weld is made with austenitic stainless
steel filler (E309). A second type of weld is made with a nickel-based filler such as Inconel 132,
Inconel 182, Inconel 82, or Inco Weld A; Inconel and Inco are trademarks of the International
Nickel Company. The third type is a pressure induction weld; welds made by this process will
have features similar to fusion welds made with stainless steel filler metals.

Table 10-1 provides some key means to distinguish between the major types of DMW failures.

10-1
13622634
Dissimilar Metal Weld Failures

Table 10-1
Distinguishing Features (Microscopic) of Failures in Fe Based Stainless Steel and Ni Base
Filler Metals in DMWs

Iron-Base Stainless Steel

Characteristic Nickel-Base Filler Metal
Filler Metal

Immediately along weld interface

Along prior austenite grain
Location of Cracking within associated with carbide
boundaries approximately 1-2
HAZ (generally) precipitation and creep
grain diameters from fusion line.
cavitation.

Carbide Morphology Generally Type II. Predominantly Type I.

Nature of Carbide Diffuse array of smaller carbides. Planar array of globular carbides.

Do carbides encourage
No. Yes.
interfacial growth?

Creep voids associated with this

No. Yes.
carbide type?

Carbon activity gradient of filler

Higher than for Ni-base fillers. Lower than Fe-base fillers.
with ferritic material?

Thermal expansion with ferritic

Worse than Ni-base fillers. Better than Fe-base fillers.
materials

About 1/3 to 1/5 of times for Ni- Three to five times longer than
Time to final failure
base filler metal welds Fe-base filler metals.

Note: Induction welded DMW will have similar properties to those listed for Fe-based fusion welds above.

Table 10-2 summarizes macroscopic features that will point to dissimilar metal weld failures in
contrast to other SH/RH tubing failure mechanisms. Failure is generally seen macroscopically as
a fusion line crack occurring at or near the heat-affected zone on the low-alloy side of the joint
(Figure 10-1). Cracking is typically oriented circumferentially around the tube. For induction
welds, the fracture may be flat and featureless; for fusion welds, the fracture will tend to follow
the contours of the fusion line. Final failures occur with low-ductility and manifest thick-edged
fracture surfaces.

10-2
13622634
 Dissimilar Metal Weld Failures

Table 10-2
Macro-Features Common to All Dissimilar Metal Weld Failures

Macro-features
Thick-edged fractures with signs of low-ductility.
Circumferential cracking in the ferritic material.
Located near a dissimilar metal weld.
Formation of an “oxide notch” on the outside surface of the tube in the ferritic material.
Flat, featureless fracture surface (typical of fusion welds).
Cracking following fusion line (typical of induction weld).
Failures may be associated with bent tubes or other signs of overstressing.

Figure 10-1
Typical appearance of a cracked dissimilar metal weld.

Damage accumulates primarily by creep, possibly with some contribution by fatigue. Three
microscopically distinct manifestations of the basic creep mechanism in DMW have been
identified; they can occur singly or in combination, as follows:
• Development of cracks along prior-austenite grain boundaries in the low-alloy steel heat-
affected zone (HAZ) one to two grains away from the fusion line. This is commonly
observed in welds made with stainless steel filler metal and occasionally in nickel-based
filler metal DMW. Figure 10-2 shows a typical cross-section.
• Development of cracking immediately at the weld interface on the low-alloy side of the weld,
along a planar array of globular carbides. This is commonly observed for DMW made with
nickel-base filler metal. Figure 10-3 shows this type of cracking.
• Propagation of an oxide notch from the external surface. Oxide notches are almost
universally seen in DMW pulled from service; however, in many cases the notches do not

10-3
13622634
Dissimilar Metal Weld Failures

propagate. Those that do are most commonly seen in thin-walled tubing and can be in either
stainless steel or nickel-base filler metals.

Figure 10-2
Typical cross-sectional appearance of a dissimilar metal weld failure after long-time boiler
service. This example is a DMW with stainless steel filler metal. Note the oxide notch on
the OD and the intergranular cracking adjacent to the weld line. Source: S.R. Paterson, et
al.7. Inset shows further detail of the intergranular creep cracking adjacent to the pressure
weld. Note the cracking is oriented normal to the hoop stress. Source: D. French.

10-4
13622634
 Dissimilar Metal Weld Failures

Figure 10-3
Detailed metallographic appearance of cracking along the weld fusion line associated with
a line of carbides. This is typically observed in dissimilar metal welds made with nickel-
base filler metals.

Failures in DMW are generally accompanied by, but not caused by, carburization of the weld
metal as indicated by increased microhardness. The degree of decarburization can be seen in
Figure 10-4 where the dramatic increase in hardness at the weld interface is shown in DMW
removed from service after 24 years.

Early stages of creep damage, such as microvoid formation, particularly in nickel-base filler
metals, may not be detectable by optical microscopy, but can be detected by scanning electron
microscopy (SEM) techniques.

DMW failures may have microstructures that show signs of overheating. This can be detected by
an analysis of the oxide scale thickness on the ferritic side of the joint, somewhat removed from
the DMW itself.

10.2.2 Locations of Failure

DMW are located in the superheater, reheater, vestibule and penthouse regions of the boiler in
transitions between austenitic and ferritic materials.

10-5
13622634
Dissimilar Metal Weld Failures

Figure 10-4
Microhardness profiles of dissimilar metal welds after 24 years of service showing an
increase in hardness near the weld interface. Source: K.H. Holko, et al.2d

10.3 Mechanism of Failure

10.3.1 Introduction

Superheater/reheater design temperatures can range from 400 to 600°C (~ 750 to 1110°F)
depending upon location. Increasingly higher tube metal temperatures demand either increased
wall thicknesses and/or a material change. Carbon steel is used in the primary stages, whereas
low alloy steels are used for most of the SH/RH, except for the finishing stages where austenitic
stainless tubes are normally used.

The DMW of concern are those that join the ferritic materials to the austenitic stainless steel. As
noted above, either fusion or induction welding processes are used. Filler metals are either
nickel-based or iron-based austenitic stainless steels. Welds made by an induction process have
properties that are similar to those for fusion welding with austenitic filler metals thus the

10-6
13622634
 Dissimilar Metal Weld Failures

comments made throughout the balance of this section pertaining to austenitic filler metals will
also apply to induction welds.

Differences in thermal expansion and creep behavior of the joined materials, and local
metallurgical changes at the low-alloy steel to weld metal interface make the DMW more
susceptible to failure than like-material welds.

10.3.2 Microstructural Changes in DMW During Service

A brief review of microstructural changes in DMW is presented here; additional detail can be
found in references 3-6.

The degradation of DMW after longterm service includes a number of observable features,
including (i) oxidation of the ferritic steel, including oxide notching, (ii) softening of the ferritic
steel HAZ, (iii) migration of carbon from the HAZ into the weld metal, (iv) precipitation and
growth of carbides at the weld interface and the HAZ prior-austenite grain boundaries, and (v)
the formation and growth of creep voids. These processes are strongly influenced by stress,
temperature and time. The times to failure for field DMW are strongly influenced by service
conditions.

The two modes of damage accumulation in DMW were introduced above: austenite grain
boundary cracking (dominant in iron-based filler-metal welds), and weld interface cracking
(dominant in nickel-base filler-metal welds).

There is a strong correlation between creep cracking and carbide morphology. Two types of
carbide morphologies form in DMW in service. Type I is a planar array of globular carbides.
This type is prevalent in nickel-base filler metal welds, occurring along about 80% of the weld
interface in a typical weld, and encourages interfacial cracking. Type II, is a diffuse array of
generally smaller carbides found in a wider band. It is predominant in stainless-steel filler metal
welds and also occurs in about 20% of the nickel-base DMW. Creep voids are only associated
with Type I precipitates.3 Type II carbides inhibit interfacial growth, to the point where a crack
developing in Type I material will stop or transfer to grain boundaries when Type II material is
encountered.

Decarburization occurs as a result of the migration of carbon from low-alloy steel to the weld
metal. This gross decarburization however, is not thought to be a major factor in DMW failures.

The ductility of low-alloy steels in the presence of a triaxial stress field is low. There is a strong
indication that the development of a triaxial stress state occurs within the weld when an axial
stress, such as caused by bending, is imposed on the radial and circumferential stresses caused by
differential thermal expansion.

Oxide notches that form at the outside surface in both nickel-base and iron-base welds were
previously thought to initiate DMW failures. However, it is now established that damage can
initiate at the outside surface, mid-wall or inside surface and that oxide notching is not a root
cause of the problem.

10-7
13622634
Dissimilar Metal Weld Failures

10.3.3 Influence of Welding Variables

Various filler metals and weld configurations/geometries can have significant effects on the life
of a DMW. Table 10-3 provides the key results of a comparison of different filler types.

Other factors such as post weld heat treatments, use of backing rings, and presence of pre-
existing weld defects can in some cases influence performance, but are generally not factors in
the development of DMW failures per se. Sources for the design of, and procedures for,
improved dissimilar metal welds and filler metals have been compiled.2h, 2i A summary of those
findings is provided here.

Under service conditions, nickel based filler materials have longer lives than stainless steel filler
metals by up to five times because2h (i) their thermal expansion better matches ferritic materials
than stainless steel fillers, and (ii) the carbon activity gradients between ferritic steel and nickel
base weld metal are lower than for stainless steel fillers, which limits the carbon migration from
the ferritic material. Because the carbon activity is significantly lower, the time for a critical
density of carbides to form along the weld line is much longer than the time for creep damage to
form in stainless steel materials, leading to longer times to failure in the nickel-base filler metals.
Service experience also indicates that repairs made with nickel-base fillers for welds originally
made with stainless steel fillers show considerable life improvement, even if part of the original
weld is left in place.

One potential drawback to the nickel-base filler metals was noted above. Once Type I carbides
form at the fusion line, the rupture strength of the weldment is equal to or less than that for the
stainless steel filler metal. Under usual service conditions, the time required for such carbides to
form is long relative to the time for damage to accumulate in the stainless steel filler. However, if
the weld is exposed to high temperatures, for example, above 595°C (~ 1100°F), even for
relatively short times, then the carbides can form. The relative life of a number of standard
nickel-base fillers compared to a standard E309 stainless steel weld in accelerated
"discriminatory" tests is shown in Figure 10-5.

10-8
13622634
 Dissimilar Metal Weld Failures

Table 10-3
Summary of Performance Characteristics of DMW Made With Various Filler Metals 2h

Expansion
Tendency to
Difference DMW Conclusions
Form Type 1 Needs Thermal
Filler with Performance Relative to DMW
Interfacial PWHT? Stability
2¼Cr–1 Mo Observations Use
Carbides
(Rt-1000°F)
Commercial alloys
Used only in least
Fairly Gives poorest
E309 27% greater None No arduous
stable. performance.
applications.
Shows little
Marked age tendency to Use limited by
Inco 92 5% greater Slight No
hardening. interfacial age hardening.
failure.
Significantly
better than
Better than #309
Fairly E309 (factor of
Inco 132 7% greater Marked No in most cases,
stable. 3-5X) but shows
widely used.
interfacial
failure.
Significantly
better than
Considerab Better than E309
E309 (factor of
Inco 182 10% greater Marked10-9 No le age in most cases,
3-5X) but shows
hardening. widely used.
interfacial
failure.
Significantly Better than E309
better than in most cases,
Fairly E309 (factor of widely used
Inco 82 3% greater Marked No
stable. 3-5X) but shows where TIG
interfacial welding
failure. employed.
Significantly
better than
E309, last Some indications
longer than that this is the
Fairly
Inco A 3% greater Marked No most nickel best of the
stable.
welds but commercial
shows nickel fillers.
interfacial failure
tendency.
Experimental Alloy
Microfissuring
No tendency to tendency needs
HFS6 (See ref
7% greater None No Excellent. interfacial: control before
2h)
shows best life. widespread use
possible.

10-9
13622634
Dissimilar Metal Weld Failures

Figure 10-5
Relative performance of dissimilar metal welds with different commercial filler metals and
geometries under accelerated discriminatory testing. Source: D.I. Roberts, et al.2h

Pressure welds have similar or shorter lifetimes than conventionally welded stainless steel
(E309).

Weld configuration was also found to be a significant variable. Evaluation of four

configurations, shown in Figure 10-6, indicated that cracking in service welds is more prevalent
in regions of the interface normal to the tube axis than those parallel to the tube axis. Under
laboratory conditions, the longest lives in accelerated tests have been obtained with DMW
having interfaces at the greatest weld angles (Figure 10-6). Weld interface angle is therefore
considered to be an important parameter in DMW performance and repair.

10-10
13622634
 Dissimilar Metal Weld Failures

Figure 10-6
Possible weld geometries for dissimilar metal welds. Source: D.I. Roberts, et al.2h

10.3.4 Influence of Stress and Temperature

The most important factors governing the life of DMW are service conditions. Stresses of three
general types contribute to premature failure in DMW: (i) primary system stresses such as caused
by pressure and dead weight loads, (ii) secondary system stresses such as caused by the
constraint of thermal expansion, and (iii) intrinsic or self-stress caused by the differential thermal
expansion of the joined materials.

Note that the last of these, differential thermal expansion of the joint materials is not sufficient,
by itself, to cause failures during normal unit operation. For failures to occur, additional, and
usually abnormal, loading or temperatures must generally exist at the DMW. Field experience
shows many examples of the effect of local secondary system stresses, such as constrained
thermal expansion, leading to DMW failures. Cases have been observed of two sets of DMW,
one with noticeable constraint of thermal expansion which led to early tube failures, and a nearby
tube without such constraint but identical in all other ways, which had not failed.

A means to demonstrate the effects of stress level, service temperature, and composition of weld
metal is shown in Figure 10-7. That figure shows the calculation of a weld performance factor

10-11
13622634
Dissimilar Metal Weld Failures

(WPF) which is the ratio of the rupture stress of T22 ferritic material remote from the weldment
to the rupture stress of the DMW. The larger the value of WPF, the lower the performance of that
weldment compared to the ferritic base material.

All of these influences have been combined into several computer codes, such as PODIS and
DMW LIFE, which are available to predict the life of DMW. Additional information about such
procedures are included in the section on longterm actions below.

Figure 10-7
Estimated weld performance factors for dissimilar metal weld joints made using austenitic
and nickel-based filler metals. Source: S.R. Paterson, et al.7

10.4 Possible Root Causes and Actions to Confirm

10.4.1 Introduction

Table 10-4 summarizes the potential root causes, actions to confirm, and corrective actions. One
of the primary questions to be answered by the boiler tube failure team is whether the DMW
BTF is driven primarily by excessive stresses or excessive temperatures.

10-12
13622634
 Dissimilar Metal Weld Failures

Table 10-4
Major Root Cause Influences, Confirmation and Corrective Actions

Major Root Cause Actions to Confirm Immediate Actions Longterm Actions and
Influences and Solutions Prevention of Repeat
Failure
10.4.2 Excessive Tube Stresses Such as Caused by Improper Initial Design or Improper Tube
Supports.
Locating the DMW near (a) Visual examination Repair damaged Implement a damage
the roof, furnace wall of the unit to locations using either assessment code,
or other fixed points determine whether a “dutchman” such as PODIS, to
or near to the there are suspect (preferred) or in-situ optimize a program
header. locations or weld repair with of control and
Weld placement in the evidence of a nickel-based filler prevention of DMW
middle of a long problem such as metal. failures. Actions
span. bent tubes, warpage Determine the extent of may include
of tubes, damage through (i) predicting remaining
Inadequate allowance misalignment, life, relocating welds,
for thermal visual examination to
missing or broken detect adjacent upgrading to higher
expansion. supports. grade materials.
locations with
Support failures or slag (b) Perform a stress obvious signs of Implement a periodic
accumulation leading analysis of suspect distress, (ii) inspection program
to constraint of locations. Piping specialized for hangers,
thermal expansion. stress codes can be radiography, (iii) supports, and
used to determine oxide scale spacers, and a
both primary and measurements and temperature
secondary stresses. analysis, (iv) monitoring program.
selective sampling, Redesign SH/RH to
as required for locate DMWs in
confirmation. areas of lower stress
or lower
temperature.
10.4.3 Excessive Local Tube Temperatures
Tube temperatures (c) Review of available As above. Institute a program to
above those thermocouple data measure and
anticipated in the for indications of interpret oxide scale
design. overheating. thickness periodically
Variation across the (d) Perform oxide scale as a means of
SH/RH. thickness evaluation understanding tube
including ultrasonic temperature trends.
measurement and Apply damage
analysis of results. assessment code to
determine whether
temperature is the
predominant factor.
Redesign SH/RH so
that the DMW is in a
lower temperature
regime.

10-13
13622634
Dissimilar Metal Weld Failures

10.4.4 Changes in Unit Operation

To increased unit (e) Review operating As above. Implement a damage
cycling change of records with an eye assessment code,
fuel causing toward conditions such as PODIS, to
increased tube that may have optimize a program
temperatures. increased either of control and
Redesign of adjacent tube stresses or prevention of DMW
SH/RH that results in temperatures. failures. Actions
higher tube service may include
temperatures. predicting remaining
life, relocating welds,
upgrading to higher
grade materials.
Determine what effects
changes in operation
will have on DMWs.
Implement a periodic
inspection program
for hangers, supports
and spacers, and a
temperature
monitoring program.
10.4.5 Initial Fabrication Defects
(f) Metallographic As above. Monitor for outbreaks of
samples should be similar problems in
used to evaluate other locations; a
whether initial weld damage assessment
defects such as method such as
incomplete fusion or PODIS may provide
lack of penetration guidance about
are a contributing general DMW life,
cause. although it cannot
specifically pinpoint
initial fabrication
defects.

10.4.2 Excessive Tube Stresses Such as Caused by Improper Initial Design of

Tube Supports, or Supports That Have Stopped Functioning in the Designed
Manner

One cause of DMW failures is if the initial design did not anticipate and properly accommodate
loads on tubes. For example, excessive restraint on tubes, such as caused by locating the DMW
near to the roof, furnace wall, or other fixed points, can be a source of high stresses. Improper
tube supports can also cause excessive weld stresses. Support design faults can include (i) weld
placement in the middle of a long span, and (ii) design that does not allow adequately for thermal
expansion of the tube. In addition to poor design of supports, restraint on tubes can result during
operation if there are support failures or if slag and other debris accumulate so as to constrain
tube thermal expansion.

10-14
13622634
 Dissimilar Metal Weld Failures

Actions to confirm will include:

(a). Visual examination of the unit to determine whether there are likely susceptible locations or
evidence of a problem such as bent tubes, warpage, misalignment of tubes, missing or broken
supports, or other visual signs of overstressing.
5
(b). Perform a stress analysis of suspect locations. One such analysis used a piping stress code to
calculate primary loads (pressure and dead weight) and secondary loads, such as constrained
thermal expansion. Although this can be expensive and time-consuming, the information can be
invaluable, and is a required input for all damage assessment models.

10.4.3 Excessive Local Tube Temperatures

Excessive local temperatures will increase thermally-induced stresses and will also accelerate
damage accumulation in DMW. These causes include (i) variation of temperatures across the
SH/RH, (ii) tube temperatures higher than design values, (iii) partial tube blockages resulting in
lower steam flows and subsequently to higher tube metal temperatures. Tube failures by
longterm overheating/creep can also occur in the ferritic alloy near to the transition between
material types. A knowledge of operative temperatures is required by damage assessment codes.

A number of methods can be used to estimate temperatures, and thus determine whether this is a
primary root cause. These include:

(c). Review available thermocouple data from locations such as outlet tubes in the penthouse,
etc.

(d). Perform an evaluation of the oxide scale on the ferritic tube. In addition to addressing the
temperature needs for the DMW analysis, this will also provide a life assessment estimate of the
ferritic tubing. Oxide thickness, the key parameter, can be measured non-destructively by
ultrasonic methods.

10.4.4 Changes in Unit Operation

Changes in operation can be at the root of DMW failures. These might typically include (i) a
change from base load to cycling of the unit, (ii) a change of fuel causing increased tube weld
temperatures, (iii) a redesign of an adjacent SH/RH that causes an increase in service
temperatures experienced by the tube. Unit cycling in particular, can be a major contributor to
damage accumulation in DMW. Actions to confirm this root cause include:

(e). Review operating records to determine if it is likely that changes have caused an increase in
primary or secondary stresses, or the temperatures of suspect tube locations.

10.4.5 Initial Fabrication Defects

Initial fabrication defects can be a direct cause of DMW failures, particularly improperly
fabricated pressure welds. To confirm:

10-15
13622634
Dissimilar Metal Weld Failures

(f). Metallographic samples of the failed tubes should indicate an incomplete or improper weld,
such as a lack-of-fusion defect.

10.5 Determining the Extent of Damage

Determining the extent of damage is the first step in an assessment of remaining life analysis for
DMW; however, it can be tricky.

Visual inspection should note locations of obvious failures and general information such as
support locations and condition, platen distortion caused by support failure or slag/ash buildup,
repairs or replacements, alterations to the boiler.

The crux of the challenge is determining the degree of damage within the DMW itself. Surface
techniques such as dye penetrant that detect only the formation of an oxide notch will be
inadequate, particularly in thick-walled tubes. Often such notches may stop propagating and final
failure will be induced mid-wall. A technique that can assess the full extent of damage is
therefore needed. Another complicating factor is that the damage occurs at the interface between
austenitic and ferritic material, and therefore it can be difficult to separate damage from the
interface signal with either ultrasonic or eddy current testing.

A specialized radiographic technique, termed the Union Electric technique for its source of
origin, can detect well-developed damage in DMW made with stainless steel filler metal.5, 8 The
basics of the method are shown in Figures 10-8 and 10-9. Although the method can detect
damage levels down to about 5% of the interface, film interpretation was found to be difficult
below about 15% damage. Good correspondence between the damage predictions by the
technique and actual damage was confirmed for over 50 DMWs that had damage levels ranging
from 5 to 90% damage.5

Figure 10-8
Union Electric special radiographic technique for dissimilar metal welds.
Source: H.J. Grunloh, et al.5

10-16
13622634
 Dissimilar Metal Weld Failures

Figure 10-9
Union Electric technique and multiple radiographic defect images.
Source: H.J. Grunloh, et al.5

For detecting creep damage with nickel-base filler metal manifested as small voids along the
interface, and for small amounts of damage in all materials, sampling and destructive
examination will be required. This may be performed on complete tube samples or boat samples
of selected areas. The use of SEM and optical techniques to characterize the extent of creep
damage is required. Appendix B reviews some of these methods.

Oxide scale thickness has been measured by ultrasonic testing to find high-temperature ferritic
tubes that are then subject to the more detailed damage assessment available from the Union
Electric technique described above.

Etching of DMW is difficult because of the variety of materials and the range of their chemical
reactivities. One successful reagent is dilute aqua regia. Its composition by volume is 42% HCl,
17% HNO3 with the balance H2O. Success has also been achieved with a double etch, once with
Nital, followed by electrolytic oxalic acid.9 Electrolytic chromium trioxide (10% in water) is
sometimes used in place of oxalic acid because it etches the stainless steel without attacking the
ferritic side of the DMW.

10-17
13622634
Dissimilar Metal Weld Failures

10.6 Repairs, Immediate Solutions and Actions

Repair methods that are relatively simple to implement and have been found to provide extended
lives of DMW are briefly summarized here.

The optimal approach has been found to be replacement of each damaged DMW with a
"dutchman", a short section containing a dissimilar metal weld made in the shop by an automatic
TIG process. Then two field welds to insert the dutchman can be made in like material: ferritic-
to-ferritic and austenitic-to-austenitic.

If it is not possible to obtain a shop-welded dutchman, then the next best approach is to insert a
short piece of new tube by making two new field welds. This is a temporary measure; the field-
welded insertion should be replaced with a proper dutchman at the first opportunity.

In-situ weld repairs of the damaged location are not recommended, however, if necessary for an
emergency, the following steps are used: (i) remove all of the damaged weld metal, leaving as
little original weld metal as possible; (ii) grind out base metal to effect a large included angle so
that a large capping pass can be used, see the bottom geometry in Figure 10-6, for example; (iii)
make the repair weld with nickel-base filler metals, independent of the composition of the
original weld metal. Weld geometry should follow in the recommendations contained in Section
10.3.3 above and illustrated in Figure 10-6. Further, the in situ repair should be replaced with a
dutchman at the next outage.

Nickel-base welds are more difficult to make than those with stainless steel filler, however the
upgrade to such welds that have more margin against failure is recommended, particularly (i) if
stresses and temperatures are higher than prescribed limits, (ii) if significant boiler cycling is
anticipated, or (iii) if there is significant uncertainty about operating conditions.

Post-weld heat treatment (PWHT) has been found to have only a secondary effect on the lifetime
of DMW, but as that effect is detrimental, PWHT should be avoided. If PWHT is used with
stainless steel filler metal, limits similar to those in Figure 10-10 should be imposed. If used with
nickel-base filler metals, care is required to avoid the formation of Type I carbides.

10-18
13622634
 Dissimilar Metal Weld Failures

Figure 10-10
Permissible time/temperature conditions for post-weld heat treatment of dissimilar metal
welds joining 2.25Cr-1Mo to 300-series stainless steel filler metal. Source: D.I. Roberts, et al.2h

10.7 Life Assessment

There are several codes available to (i) predict the level of damage that may be anticipated in a
particular DMW, (ii) to highlight potential failure locations, and (iii) to evaluate options for
corrective actions. The features of one such code, developed by EPRI and designated Prediction
of Damage in Service (PODIS)2g are discussed next as a means to introduce the nature of the
calculations that are typically performed for such assessments.

The PODIS code is an empirical method based on the field experience and accelerated test
results. Life estimates are calculated from knowledge of, or an estimate of, the loading history to
which the tubing was and will be subjected. Loading history is defined in terms of time, weld
metal temperature, weld metal temperature change, number of cycles of temperature change,
axial stress at the weld due to pressure, deadweight, and restrained thermal expansion loads
within the tube assembly.

Creep damage is imposed by three types of stresses:

• Intrinsic damage (self damage).
• Damage caused by primary (load-controlled) components of the system loading on the weld,
including axial pressure loads and dead weight.
• Damage caused by secondary (strain-controlled) components of the system loading on the
weld such as those from restricted thermal expansion.

10-19
13622634
Dissimilar Metal Weld Failures

Any method to evaluate the accumulation of damage must provide for estimates of these stresses
so that total damage can be estimated. A number of steps are typically required to accumulate
and analyze the needed information including2g:
• Plant inspection to determine tube condition and contributing factors to DMW failures, such
as support conditions, slag buildup, misalignment, and tube warpage.
• Review operating history to obtain background information necessary to estimate tube
temperature levels and cycling.
• Analyze system load histories at DMW.
• Assessment of actual DMW damage. See section on determining the extent of damage above.
• Predict damage from a code such as PODIS.
• Correlate actual damage measurement with that predicted from PODIS to determine whether
it is likely that all significant effects have been incorporated into the analysis.
• Predict residual life based on projection of future unit operations.

Some knowledge is required of how much conservative the results of a specific code are. For
example, the availability of field failure experience to calibrate the empirical constants in the
PODIS model have led to a conservative treatment, particularly for nickel-based welds. The
degree of conservatism has been explored in research funded by Empire State Electric Energy
Research Corporation (ESEERCO).10, 11 As a result, a program called DMW LIFE was
developed which removes some of the conservatism in the PODIS analysis of nickel-based welds
via an analysis of industry experience.

Further, the ability to analyze fireside corrosion and/or longterm overheating may be necessary
to analyze properly the remaining tube life. For example, PODIS does not take account the
effects of fireside corrosion that will reduce tube wall thickness and thus increase stress levels.
Modifications to PODIS have used data from the literature and limited field corrosion test results
to include corrosion factors in the DMW LIFE code. Finally, neither the effects of longterm
overheating/creep in tubing nor oxide notch propagation are considered in PODIS.

10-20
13622634
 Dissimilar Metal Weld Failures

10.8 Longterm Actions and Prevention of Repeat Failures

A longterm program to ensure the integrity of DMW should be a part of every BTF Team
agenda. The use of a damage assessment code such as PODIS can be used to predict DMW
lifetimes, to identify where new or replacement welds are required, and to anticipate where
future actions may be needed.

Longterm actions may include steps such as relocating DMW, e.g., placing them away from
fixed supports or in locations subjected to lower temperatures. Where there is uncertainty about
the temperatures that are being experienced, prudent practice is to substitute with higher grade
material. If a total replacement of the SH or RH is required, the optimum DMW locations, weld
metal composition, and weld configurations should be "designed-in".

A program of periodic inspection of hangers, supports and spacers to ensure that secondary loads
do not develop is also indicated. Periodic checking of unit temperatures with either direct
measurement by thermocouples or by measuring oxide scales in ferritic tubes is also
recommended.

10-21
13622634
Dissimilar Metal Weld Failures

10.9 Case Study

Dissimilar Metal Weld Failures

Case Study: Field Experience and Integrated Analysis
Unit Description. The unit is a base-loaded, 350 MW, coal-fired unit put on line in 1970
with 165,000 operating hours. Superheater outlet design conditions are 17.2 MPa
(2,500 psig) at 540°C (1005°F); reheater outlet design conditions are 4.1 MPa (594
psig) at 540°C (1005°F).
Weld Locations. The unit contains approximately 1,675 DMW located in the division
panels, superheater platen, reheater front pendant, and superheater front pendant.
Original welds were made with stainless steel filler metal. Although no failures had yet
occurred, a similar unit had experienced DMW failures so that a program of analysis
and preventive measures was considered prudent. Four regions were evaluated: SH
penthouse, SH furnace, RH penthouse, and RH furnace.
NDE and Metallography. Ultrasonic testing was used to determine oxide scale thickness
in three of the four regions, the exception was the SH furnace tubes owing to access
constraints. Every third tube containing a DMW in the ferritic material was measured
close to the DMW so as to provide an indication of joint temperatures, but removed from
the weld geometry and associated discontinuities. Radiographic testing by the Union
Electric Technique8 was performed on 65% (1,081) of the unit's DMW. The method had
previously been validated by comparison with removed samples. Twenty-five percent of
the DMW radiographed indicated interfacial damage ranging from 5 to 90%. The
majority of the damage was found in the RH furnace.
Metallographic samples were taken of selected DMWs with extensive damage to
confirm the mechanism and further confirm the accuracy of the RT.
Temperature Estimates. Oxide thickness measurements and subsequent analysis of the
results were used in conjunction with available thermocouple data to estimate tube
temperatures. Boiler-wide temperature estimates were made.
Stress Analysis. A piping stress analysis code was used to estimate primary (dead
weight) and secondary (primarily restrained thermal expansion) stresses. Pressure
stresses were calculated and added to the results. Results for SH tubes found a range
in secondary stresses from 0.10 – 127.6 MPa (14 to 18,501 psi); for primary stresses
the range was 10.5 – 20.8 MPa (1,522 to 3,016 psi).
Damage Analysis. Current temperatures were estimated from the inspection results and
stress levels. A damage assessment code was used to formulate a strategy for DMW
actions. DMW with more than 50% interfacial damage were repaired immediately. A
total of 13 DMW in the RH and 5 DMW in the SH were repaired. Repairs were
performed by grinding out the original weld metal as completely as possible, grinding a
60° angle on the ferritic side, and welding with either Inconel 182 or Inco-Weld A filler
metal. Plans for future replacements with a "dutchman" were formulated.
Source: This case study is a summary of an evaluation first reported by H.J. Grunloh,
R.H. Ryder, and R. Hellner.5

10-22
13622634
 Dissimilar Metal Weld Failures

10.10 References
1
Dooley, R.B. and H.J. Westwood, Analysis and Prevention of Boiler Tube Failures, Ontario
Hydro, Report 83/237G-31, Canadian Electrical Association, Montreal, Quebec, November,
1983.
2
Prager, M., et al., Dissimilar-Weld Failure Analysis and Development Program, Research
Project 1874-1, Final Report CS-4252, 8 Volumes, Electric Power Research Institute, Palo Alto,
CA.
a. Prager, M., Volume 1: Executive Summary, November, 1985.
b. Roberts, D.I., C.C. Li, and R.D. Nicholson, Volume 2: Metallurgical Characteristics,
November, 1985.
c. Ryder, R.H., H.J. Grunloh, R.F. Stetson, K.J. Tong, K.H. Holko, D.I. Roberts, F.V. Ellis,
M.P. Borden, S.N. Cato, and B.W. Roberts, Volume 3: Accelerated Discriminatory Tests,
November, 1986.
d. Holko, K.H., C.C. Li, R.H. Ryder, D.I. Roberts, and C.F. Dahms, Volume 4: Utility Plant
Results; November, 1985.
e. Roberts, D.I., H.J. Grunloh, and K.H. Holko, Volume 5: Evaluation of Acoustic Emission
and Enhanced Radiography, November, 1985.
f. Prager, M., H.J. Grunlo, J.R. Foulds, R.H. Ryder, C.F. Dahms, and M. Krishnan, Volume
6: Weld Condition and Remaining Life Assessment Manual, August, 1988
g. Ryder, R.H., C.F. Dahms, M. Krishman, H.J. Grunloh, and J.R. Foulds, Volume 7:
Prediction of Damage in Service (PODIS) Code - Background Document, May, 1988.
h. Roberts, D.I., R.H. Ryder, H.J. Grunloh, and B.E. Thurgood, Volume 8: Design and
Procedure Guide for Improved Welds, November, 1989.
3
Roberts, D.I., R.H. Ryder, and R. Viswanathan, "Performance of Dissimilar Welds in Service",
Journal of Pressure Vessel Technology, Volume 107, August, 1985.
4
Li, C.C., R. Viswanathan, and R.H. Ryder, "The Microstructure and Remaining Life of
Dissimilar Metal Weldments After Service in Fossil-Fired Boilers", Proceedings of the ASME
International Conference on Advances in Life Prediction Methods, April 18-20, 1983, Albany,
New York.
5
Grunloh, H.J., R.H. Ryder, and R. Hellner, "Damage Assessment and Predictive Maintenance of
Dissimilar Metal Welds in Superheater and Reheater Tubes", in B. Dooley, ed., Proceedings:
International Conference on Boiler Tube Failures in Fossil Plants, held in San Diego, California
November 5-7, 1991, Proceedings TR-100493, Electric Power Research Institute, Palo Alto, CA,
April, 1992, pp. 7-51 through 7-76.
6
Ryder, R.H., D.I. Roberts, and R. Viswanathan, "Dissimilar Metal Weld Failure Reduction" in
B. Dooley and D. Broske, eds., Boiler Tube Failures in Fossil Power Plants: Conference

10-23
13622634
Dissimilar Metal Weld Failures

Proceedings, Conference held in Atlanta, Georgia, November 10-12, 1987, CS-5500-SR,

Electric Power Research Institute, Palo Alto, CA, 1988, pp. 2-63 through 2-72.
7
Paterson, S.R., T.A. Kuntz, R.S. Moser, and H. Vaillancourt, Boiler Tube Failure Metallurgical
Guide, Volume 1: Technical Report, Volume 2: Appendices, Research Project 1890-09, Final
Report TR-102433, Electric Power Research Institute, Palo Alto, CA, October, 1993.
8
Gurnea, R.F, "Radiographic Technique for Detecting Cracks in Dissimilar Weld Joints", in R.
Viswanathan and D.A. Roberts, eds., Proceedings: Seminar on Dissimilar Welds in Fossil-Fired
Boilers, held in New Orleans, LA., February 23-24, 1984, Research Project 1874-1, Proceedings
CS-3623, Electric Power Research Institute, Palo Alto, CA, July, 1985, pp. 4-47 through 4-60.
9
Personal Communication from S. Paterson (Aptech Engineering Services, Inc.) to R.B. Dooley,
February, 1995.
10
Sherlock, T.P. and D.N. French, Dissimilar Metal Weld Assessment Program, Final Report,
ESEERCO Project 91-20, Empire State Electric Energy Research Corporation, New York, NY,
December, 1991.
11
Sherlock, T.P. and D.M. Conklin, Dissimilar Metal Weld Assessment Program, Final Report,
ESEERCO Project 91-20, Phase II, Empire State Electric Energy Research Corporation, New
York, NY, May, 1994.

10-24
13622634
11
LONGTERM OVERHEAT CREEP

11.1 Introduction

Superheater (SH) and reheater (RH) tubes operate in a regime where creep is significant and
oxidation resistance is important. As discussed in Chapter 2, SH/RH tube design involves a
choice of wall thickness and alloy type to withstand the expected pressures and temperatures and
still provide at least a minimum specified level of life.

Creep damage is strongly dependent on stress level and on tube metal temperature. Therefore
conditions which are only a slight departure from design levels leading to increased stress (such
as wall thinning by corrosion) or increased temperature will result in operating tube lives that are
significantly shorter than expected.

Higher than expected temperatures in SH/RH tubes can be a contributor to a number of damage
types. The reader is specifically directed to discussions of other mechanisms which might be
confused with longterm overheating such as fireside corrosion and short-term overheating.

Estimating the life of austenitic stainless steel tubes experiencing long term overheat creep
damage is generally more difficult than in the case of ferritic tubing. Appendix C provides a
roadmap approach and illustrative examples of remaining life methods for this mechanism.

11.2 Features of Failure and Typical Locations

11.2.1 Features of Failure

Final failures by longterm overheating/creep are generally longitudinal (axial to the tube), and
located on the heated side of the tube. Failures are generally thick-edged, corresponding to low
ductility. The range of final failure appearance in austenitic materials is shown schematically in
Figure 11-1. A typical appearance is shown in the failed tube pictured in Figure 11-2. In reheater
tubes, because of the thinner-walled materials, final failures tend to appear more ductile than in
superheater tubes.

11-1
13622634
Longterm Overheat Creep

Figure 11-1
Types of high temperature creep failure in tubing. Source: G.A Lamping and R.M.
Arrowood, Jr.1

Figure 11-2
Longterm creep overheat failure in a TP321 superheater tube.

11-2
13622634
 Longterm Overheat Creep

Evidence for the overheating of SH/RH tubes can include thickened external scales, which may
be cracked longitudinally. There may also be a relatively thick internal oxide scale, also cracked
longitudinally. When ash erosion or corrosion is active, the final failure may include wastage
flats on the tube at the 10 o'clock and 2 o'clock positions (flue gas at 12 o'clock). Note that the
flats can be present at other tube locations depending on tube alignment in the bundle.

Microscopically, there will be evidence of damage to the material microstructure by overheating

and intergranular or transgranular creep; distinguishing features will be material and stress level
dependent. Austenitic stainless steels will manifest microstructural damage with sigma phase and
grain boundary cavities. The nucleation, growth and interlinking of creep voids or cavities,
termed creep cavitation, will be associated with the cracking. Note that improper etching may
etch out sigma phase and give the appearance of creep cavities.2 Refer to Appendix B for
recommended metallographic preparation methods.

It is important that the root cause of the failure be identified accurately. In particular, the fireside
wastage mechanisms, such as fireside corrosion or erosion, can lead to wall thinning which will
then result in the accumulation of creep deformation. Before settling on longterm overheating as
the primary damage mechanism, it is important to evaluate whether these other mechanisms are
also contributors.

Table 11-1 lists some key distinguishing differences among longterm overheating, short-term
overheating, and fireside corrosion in SH/RH tubing.

11-3
13622634
Longterm Overheat Creep

Table 11-1
Comparison of Characteristics of Longterm Overheating/Creep, Short-Term Overheating,
And Fireside Corrosion (Coal-Fired Units) In Superheater/Reheater Tubing

Characteristic Longterm Overheating Short-Term Fireside Corrosion

Overheating
Fracture Surface and Generally thick-edged, Usually thin-edged, Tube wastage,
Appearance of Failure brittle final failure. ductile final failures. particularly at the 10
Generally accompanied Swelling of tubes and 2 o’clock
by external tube without ovalization positions.
wastage at the 10 “Fish-mouth” Longitudinal cracking,
o’clock. and 2 o’clock appearance of tube final failure can be
positions. rupture. (bun not necessarily)
by overheating.
Internal Scale? Yes, generally Not necessarily thick. Yes, particularly if tube
extensive, multi- Depends on age of tube metal over- heating was
laminated and at failure. an influencing factor.
exfoliating.
External Scaling? Yes, thick, laminated Not necessarily thick. Yes, with multi-layers: (i)
and often a hard, porous layer–
longitudinally composition typically of
cracked. flyash, (ii) an
Usually two layers – (i) intermediate layer
a hard, porous outer containing complex
layer with alkali sulfates, and (iii) a
composition typically black, glossy inner layer
that of flyash, and (ii) mostly of oxides,
a black glossy inner sulfates, and sulfides of
layer mostly oxide iron.
but may contain
some sulfates and
sulfides of iron.
Outside surface Characteristic Swelling, stretch marks Characteristic
appearance after longitudinal grooving on tube metal. longitudinal grooving
removal of and pitting (“alligator and pitting (“alligator
scale/deposits hide”). hide”). Sometimes the
corroded area is smooth
and featureless.
Sometimes “orange
peel” appearance at
extremities of severe
corrosion.
Composition of External Does not contain low Not relevant. Does contain low
Scales/Deposits melting point ash melting point
compounds such as compounds such as
alkali iron sulfates alkali-iron sulfates (coal-
fired units).
Wall Thinning? Typically wastage flats Only because of bulging Primary feature of
at 10 o’clock and 2 of tube material. failure, may be worse at
o’clock positions caused the 10 and 2 o’clock
by accelerated positions. Can be at
oxidation. Can be at other locations
other locations depending on tube
depending on tube position. Depending

11-4
13622634
 Longterm Overheat Creep

Characteristic Longterm Overheating Short-Term Fireside Corrosion

Overheating
position. There is upon the rate of
always a layer of oxide corrosion, a protective
adjacent to the tube. oxide layer may remain
on the tube or may have
been fluxed off.
Ratio of wall loss to Typically less than 3:1. Not relevant Typically greater than
steamside oxide 3:1; for ratios greater
thickness than 5:1 fireside
corrosion or erosion is
the dominant
mechanism.
Tube Material Yes, generally extensive Depends on the If overheating has been
Degradation signs of overheating material and the a problem, yes;
and/or of creep maximum temperature otherwise, no. Fireside
damage, particularly reached. corrosion can occur in a
near crack tips. Creep tube at design
voids will not be found temperatures.
removed from crack tip.
Change in material Localized softening near Localized hardening Hardness change is not
hardness the rupture is typical. near the rupture is necessary; if there has
likely. been no overheating,
there will be no change
in hardness.

In evaluating the contributing mechanisms, distinguishing between fireside corrosion and

longterm overheating is perhaps the most common cause of confusion. A key difference is that if
corrosive attack has occurred, external deposits will contain low melting-point ash components.

Note that longterm overheating, where the damage accumulates because of temperatures that are
somewhat over the design level, should also not be confused with short-term overheating, where
a pressurized tube is heated to well above its design temperature and failure occurs in a much
shorter time. SH/RH tube failures by short-term overheating are the subject of Chapter 9.

11.2.2 Locations of Failure

Figure 11-3 shows the typical locations of damage by longterm overheating/creep. The following
locations are susceptible3:
• Near changes in wall thickness in the same alloy.
• Where there is a variation in the gas-touched length among tubes of the same material,
leading to the earlier failure of those subjected to higher temperatures.
• On the lowest tube in a horizontal platen or the leading tube in a pendant section.
• Either in a cavity or just beyond a cavity, where radiant effects can lead to higher tube
temperatures.

11-5
13622634
Longterm Overheat Creep

• In the final leg of tubing just before the outlet header, where steam temperatures are the
highest.

Figure 11-3
Typical boiler locations where longterm overheating/creep failures can occur in vertical
platen elements (a) and in horizontal platens (b). Solid circles represent typical failure
locations; letters in circles indicate locations of interest described in the main text.

11-6
13622634
 Longterm Overheat Creep

11.3 Mechanism of Failure

In the absence of other boiler tube failure mechanisms, it is not unexpected that creep will be the
life-limiting damage type for SH/RH tubing. In fact, such tubing is designed for a finite life
based on a conservative creep criterion. Design temperatures can be up to 600°C (1112°F) in the
finishing stages where austenitic stainless tubes are normally used. On the basis of laboratory
oxidation experiments, each manufacturer specifies a maximum operating temperature for each
material. Representative values are shown in Table 11-2.
Table 11-2
Maximum Tube Metal Temperatures

Stainless Steel Alloy ASME ASME Max.4 B&W Max.5 C-E Max.6 Riley Max.7
Specification (see note a)

18Cr – 8Ni SA213 304H 816oC 760oC 704oC ~

(1500oF) (1400oF) (1300oF)

18Cr- 10Ni- Ti SA213 321H 816oC 760oC ~ 816oC

(1500oF) (1400oF) (1500oF)

18Cr – 12Ni-Nb SA213 347H 816oC ~ 704oC ~

(1500oF) (1300oF)

Notes:
a. This is the highest temperature for which maximum allowable stress values are given

There can be major differences (100-150°F) between the operating temperature and the
temperature expected for the design. This can lead to significant shortening of actual tube life.
Figure 11-4 shows the effects of stress and temperature on the length of time to failure for a
typical boiler tube material.

A number of other factors, discussed in the root cause section below, can result in higher than
expected temperatures. In addition to poor initial design (such as tubes with gas-touched length
longer than the design estimate), the main factors in the case of austenitic stainless steels are: (i)
nonuniform hydraulic resistance factors across or through the tube bank leading to differences in
steam flow rates and steam cooling of the tubing, (ii) partial blockages can sometimes occur
from chemical deposits, scale, debris or deposits, (iii) excessive operating conditions such as
high flue gas temperature, displaced fireball, periodic overfiring, or uneven firing of fuel burners,
and (iv) increased gas temperatures resulting from low NOx burner and air port modifications,
etc.

11-7
13622634
Longterm Overheat Creep

Figure 11-4
Time-to-rupture curves for SA-213 Type 304 stainless steel.

11.3.1 The Accumulation of Creep Damage in Superheater/Reheater Tubes

The effects on material microstructure that occur as a result of exposure to temperatures

consistent with creep damage are discussed in Appendix B. Here the manifestation of creep in
SH/RH tubes is reviewed.

The accumulation of creep damage can occur through an elevation of tube metal temperature or
stress as noted above. Tube stresses increase primarily as a result of localized wall thinning and
correspondingly higher hoop stresses. Processes that occur on both the steamside and fireside of
the tube can contribute to premature tube failures by creep.

11.3.2 Steamside Processes

A normally thin protective oxide forms on austenitic stainless steels when they are exposed to
steam. The steamside oxide growth is usually very slow (with total oxide scale thicknesses less
than 100µm (0.004-inch)) and does not lead to significant increases in tube metal temperatures.
The primary concern with regard to the steamside oxide scale is exfoliation and steam flow
blockage. There have been some cases of austenitic tubing that have experienced buildup of
steamside scale that is greater than 100µm (0.004-inch). In these cases the increased tube metal
temperatures due to the insulating effect of the scale may be significant.

11-8
13622634
 Longterm Overheat Creep

11.3.3 Fireside Effects

In the normal (design) condition, a protective oxide scale forms on the exterior of SH/RH tubing.
Ash accumulates on the side of the tube facing into the flow of flue gas as shown in Figure 11-5.
On a tube that operates above the design condition, there can be a reduced thickness of ash at the
10 and 2 o'clock positions (flue gas flow at 12 o'clock) due to aerodynamic factors. Several
consequences result: (i) the thinner layer of ash at the 10 and 2 o'clock positions allows for
higher heat transfer rates and more radiant effects which raise the temperature of the tube, (ii) an
increased corrosion rate occurs (oxidation in a sulfur-containing environment) which also leads
to material removal and thus higher stress levels. The latter consequence also changes the
kinetics of oxide growth from parabolic to linear and, as a result, it becomes non-protective
(Figure 11-5).

Figure 11-5
Schematic of the development of longterm overheating failures. (a) Represents the normal
situation. (b) Shows the development of “flats” at 10 and 12 o’clock. The flats can be
present at other tube locations, such as the tube crown, depending on tube alignment in
the bundle.

The increasing stress level and overtemperature in the tube accelerate the creep process leading
to tube failure.

The process can be further accelerated if tubes are also subject to attack by other corrosive
processes, such as "molten salt" attack, the topic of Chapters 4 (for coal-fired units) and 5 (for
oil-fired units).

In summary, creep damage will accumulate at a rate higher than the design estimates if there is
(i) a relatively continuous period of slight overheating such as caused by shortcomings in the
initial design, (ii) a slowly increasing level of temperature such as by increasing internal oxide
thickness, or stress caused by decreasing wall thickness, or (iii) the accumulation of periods of
excessive overheating due to operating factors.

11-9
13622634
Longterm Overheat Creep

11.4 Possible Root Causes and Actions to Confirm

Table 11-3 summarizes the major root cause influences, actions to confirm each, and corrective
actions.
Table 11-3
Major Root Cause Influences, Confirmation and Corrective Actions

Major Root Cause Actions to Confirm Immediate Actions Longterm Actions and
Influences and Solutions Prevention of Repeat
Failure
11.4 All Causes of Overheating
(a)Direct measurement Local repairs as Determine remaining life
of temperatures by appropriate. of affected tubes based
thermocouples. Perform selective on actual temperatures,
(b)Metallurgical analysis sampling and/or stress levels and
of tube structure and ultrasonic measurement material properties.
oxide thickness and to determine extent of See Appendix C for
morphology of selected problem. additional detail.
tube samples. Perform remaining life Make the change to a
(c)Ultrasonic testing for estimate of affected higher grade of steel
direct measurement of tubes. (next higher chromium
steamside oxide level).
See longterm actions.
thickness. Tube/circuit
(d)Visual examination realignment; consider
for evidence of slag steam flow
buildup, laning, bowed redistribution. See main
or misaligned tubes text for summary of
acting as leading tubes. methodology.
Major SH/RH redesign
and replacement.
Redesign tube bank.
Retube with same
material depending on
expected and desired
life of SH/RH.
11.4.1 Influences of Initial Design and/or Material Choice
Original alloy (e)Review temperature As above. As above: steam flow
inadequate for actual data from redistribution will be
operating temperatures. thermocouples installed particularly applicable
Inadequate heat in vestibule or across for correcting side-to-
treatment of original the header. side and local variations
alloy. (f)Review SH/RH circuit in temperature,
material diagrams, upgrading will be
Tubes at failure location particularly applicable in
have gas-touched calculate and plot GTL
as a function of steam cases where the original
lengths longer than alloy was insufficient or
design estimate and/or and metal temperatures,
plot positions of failures. where the tube has a
row-to-row variation in gas-touched length that
gas-touched length. is longer than the
Side-to-side or local gas design estimate.

11-10
13622634
 Longterm Overheat Creep

Major Root Cause Actions to Confirm Immediate Actions Longterm Actions and
Influences and Solutions Prevention of Repeat
Failure
temperature differences.
Radiant cavity heating
effects.
Lead tube/wrapper tube
material not resistant
enough to temperature.
11.4.2 Build-up of Internal Oxide Scale
(g)See Items (a) Chemical cleaning to Address underlying
through (c). remove deposits. cause of overheating or
Periodic chemical
cleaning to mitigate
effects.
Determine remaining life
of affected tubes based
on actual temperatures,
stress levels and
materials properties.
See additional options
on primary list from
above.

11.4.3 Overheating because of Restricted Steam Flow due to Chemical or Other Deposits, Scale,
Debris, Etc.
(h)Selective sampling of Clean out tubes and Introduce measures to
suspect locations to remove source of prevent future
verify whether local blockages. blockages.
blockage is leading to
excessive temperatures.

11.4.4 Operating Conditions or Changes in Operation

Local repairs as Optimization of
appropriate. operation and fireside
Perform selective conditions must be the
sampling and/or emphasis. See
ultrasonic compilations of the
measurements to applicable methods
determine extent of (Reference 9).
problem.

Perform remaining life

estimate of affected
tubes.
See longterm actions.

11-11
13622634
Longterm Overheat Creep

11.4.5 Previous similar problems in adjacent SH/RH

(i) Check As above.
temperature distribution
through the circuit by
performing analysis of
GTL and measured
temperatures; see (e)
and (f) above.
Combustion conditions (j) Monitor gas As above, plus Optimization of fireside
can lead to tube temperatures with Restore boiler design conditions. See
overheating. pyrometers or infrared (or optimized) Reference 9.
Excessive flue gas instruments. conditions.
temperature
Displaced fireball
Delayed combustion
Periodic overfiring or
uneven firing of fuel
burners.
11.4.6 Blockage or Laning of Boiler Gas Passages
(k)Can be recognized Controlled with flow
using cold air velocity distribution screens; in
technique. See flyash practice is difficult to
erosion mechanism for implement because of
a discussion of the high temperatures in
technique. SH/RH.
(l)Visual examination to
identify local flow
blockages.
11.4.7 Increases in Stress due to Wall Thinning
(m)NDE evaluation to Initiate procedures to Check longterm actions
determine the extent of identify source of tube in wastage mechanism
wall thinning. wastage – of particular chapters particularly
(n)If another mechanism concern are fireside fireside corrosion
(corrosion, erosion) is corrosion or (Chapters 4 and 5).
suspended, initiate flyash/sootblower
actions to confirm their erosion processes that
involvement. may be contributing to
an increased oxidation
rate.

Verification of longterm overheating/creep, independent of cause, can be direct or indirect.

Direct verification and an indicator of the current temperatures can be obtained by
thermocouples. Indirect verification and an indication of the historical creep damage
accumulation can be obtained from analysis of the tube microstructure. Refer to Appendix B for
microstructural evaluation recommendations.

11-12
13622634
 Longterm Overheat Creep

Independent of the source of overheating, the following three actions are used to confirm that the
condition exists. Additional actions pertinent to specific root causes are provided below.

(a). Direct measurement by thermocouples.

Thermocouples can be used to measure tube temperatures directly and also to determine furnace
gas temperatures; in the latter case, they verify abnormal flow patterns. The placement of
thermocouples and interpretation of results is complicated by the variations in temperature
around an individual tube (Figure 11-6) and across a superheater or reheater (Figure 11-7).
General guidance can be obtained from the temperature profiles across the unit that are recorded
in the vestibule or header area on the permanently installed operating thermocouples.

(b). Perform metallographic analysis of tube samples.

Metallographic analysis is used to estimate the creep degradation of the tube material. In contrast
to ferritic steels that form a reasonable thick oxide scale in high temperature steam, the oxide
scale in stainless steel tubes is normally too thin and irregular to provide useful thickness
measurements.

(c). Visual examination to look for evidence of slag buildup, laning, bowed or misaligned tubes
acting as leading tubes.

Figure 11-6
A reheater tube that contains steam at 510°C (950°F) will have varying metal temperatures
around its circumference and through its wall according to: gas flow direction, ash deposit
patterns, and internal scale formations. Typically, the highest metal temperature will be
achieved on the surface facing the gas flow. Source: G.A. Lamping and R.M. Arrowood, Jr.1

11-13
13622634
Longterm Overheat Creep

Figure 11-7
Typical thermocouple measurements across a superheater showing that the actual tube
metal temperatures can vary significantly with tube location and boiler power level.
Source: G.A. Lamping and R.M. Arrowood, Jr.1

11.4.1 Influences of Initial Design and/or Material Choice

The aim of the design process is to provide a structure that functions within maximum allowable
stresses and oxidation rates. As the temperature capability of each grade of tube material is
reached, the designer specifies a transition to a higher grade material. If the actual operating
temperatures exceed those foreseen in the design, then overheating will result. Figure 11-3
illustrates several examples of how this may occur:
• Differences in gas-touched length of tube that can sometimes be as large as 20-30 feet.
• Radiant cavity heating effects. Tubes in a cavity or that surround a cavity, such as Tube "D"
at location "P" in Figure 11-3, or just past a cavity may experience excessive temperatures
because of radiant absorption effects.
• Other effects of location. Tube "A" in Figure 11-3 is the lead or wrapper tube and thus will
pick up more heat than tubes "B" - "H". There can also be side-to-side or local gas
differences; tubes in the center run hotter.

Many creep failures have been observed due to additional stresses which were not known during
the design process or were underestimated.2

Actions to confirm this root cause will include (a) through (c) above, plus:

(e). Review temperature data taken from across the tubes in the vestibule or header space. The
array of thermocouples attached to tubing coming out of the header (used by operators for steam
temperature control) should be used. See Figure 11-7 for an example of the process.

11-14
13622634
 Longterm Overheat Creep

(f). Perform an analysis of gas-touched length (GTL) and location of the tube failures. Review
SH/RH circuit material diagrams and calculate the GTL to the position(s) of failure. Plot the
manufacturers recommended temperature limits (based on acceptable oxidation rates as
discussed in Section 11.3 above) for each material in the circuit. Superimpose any information
on measured steam and tube metal temperatures to determine whether the actual temperatures are
over the design limits. It is recommended that this method be used to confirm this root cause
when it is suspected. The process will also provide an alert of the potential for other boiler tube
failure mechanisms that are exacerbated by excessive temperatures such as fireside corrosion.

11.4.2 Buildup of Internal Oxide Scale

Typically the steamside oxide scales on austenitic stainless steel tubing have thickness of 0.08
mm (0.003 inch) or less even after longterm service. These scales will not result in a significant
increase in tube metal temperature. In less frequent situations, the austenitic tubes will exhibit
much thicker steamside scales and corresponding significant temperature increases. In these
cases it is useful to estimate the current tube metal temperature and relate this to the creep
damage rate. To confirm that the buildup of steamside oxide is a primary influence on the
damage observed:

(g). Institute any of the actions (a) through (c) outlined above.

11.4.3 Overheating Because of Restricted Steam Flow

Partial blockages can sometimes occur from chemical deposits, scale, debris or deposits. Usually
they will result in short-term overheating failures (Chapter 9).

(h). Selective sampling of suspect locations can verify whether a local blockage condition is
responsible for excessive tube temperatures. However, the usual sequence of identification is
tube failure and then assessment of adjacent circuits.

11.4.4 Operating Conditions or Changes - Previous Similar Problems in Adjacent

SH/RH.

Overheating can occur in a SH/RH circuit if absorption patterns are changed as a result of a
redesign in another circuit. Such redesign may have been necessitated by prior failures.

(i). Check the temperature distribution through the circuit by performing an analysis of gas-
touched length and measured temperatures, as described in (e) above.

11.4.5 Combustion Conditions

Combustion conditions can lead to tube overheating. Such conditions include: excessive flue gas
temperature, displaced fireball, delayed combustion, periodic overfiring or uneven firing of fuel
burners.

11-15
13622634
Longterm Overheat Creep

(j). Gas temperature monitoring with various suction pyrometers or infrared measurements.

11.4.6 Blockage or Laning of Boiler Gas Passages

Actions to confirm:

(k). The cold air velocity technique discussed at length in the chapter on flyash erosion can be
used to identify locations of locally high gas velocities.

(l). Visual examination can often identify local flow blockages.

11.4.7 Increases in Stress Due to Wall Thinning

The previous root cause influences have dealt primarily with increases in tube metal temperature
as a contributor to excessive damage accumulation by creep. Any of the processes by which tube
wall thinning can occur, such as fireside corrosion or erosion, can lead to increased tube hoop
stresses which will also increased creep damage. Specific confirmation of wall thinning can be
conducted by:

(m). NDE methods, such as ultrasonic testing, can determine the extent of wall thinning.

(n). If another mechanism is suspected, e.g., fireside corrosion, erosion, etc., actions to confirm
those processes should be implemented. Reference to the chapter of this book dealing with each
mechanism will provide a list of recommended actions.

11.5 Determining the Extent of Damage

Determining the extent of damage will involve identifying locations with the thinnest walls.
Ultrasonic testing is the most applicable and inexpensive NDE method for measuring wall
thinning.

Note that among the hundreds of tubes in each superheater/reheater bank in a boiler there will be
a distribution of gas and metal temperatures. As a result there will be an equivalent distribution
of creep and corrosion rates. Therefore the key to a successful prevention program will involve a
search for the highest risk locations, and assessment of their conditions, not of the mean
condition. Wall thickness measurements around the entire tube will be useful, particularly to
distinguish the development of tube "flats".

Sacrificial tube samples are removed from select locations and subjected to laboratory
metallurgical analysis and/or isostress rupture testing to confirm NDE results and to refine the
estimates of accumulated damage.

Potentially affected areas can also be identified by reviewing data from existing thermocouples
or by installing thermocouples in at-risk locations.

11-16
13622634
 Longterm Overheat Creep

11.6 Repairs, Immediate Solutions and Actions

11.6.1 Need for Remaining Life Assessment

Any immediate solution or repair strategy should consider that a remaining life methodology be
implemented. Such programs are discussed in detail in Appendix C, and are summarized below
under longterm options.

11.6.2 Repairs

If the problem is highly localized, it may be possible to replace the affected tubes. Care should
be taken with this quick replacement method. It is important to remove the full extent of the
damaged area, as indicated by tube swelling, wall loss, tube "flats", or "alligator hide".
Removing an insufficient amount of damaged area will result in a repeat boiler tube failure.

Pad welding should not be used for repair because of the uncertainty of tube conditions,
particularly the presence and depth of creep cracks and the condition of the internal surface of
the tube.

11.6.3 Correction of "Off-Design" Conditions

In this category are those actions that can correct obvious conditions including tube blockages
and fireside combustion problems such as displaced fireball, uneven firing or fuel burners,
correctable excess flue gas temperatures, etc.

11.7 Life Assessment

The importance of a comprehensive remaining life assessment cannot be overstated. It is a

prerequisite to the optimal choice of longterm solutions. During the past twenty years a number
of techniques have been developed to assess remaining life of superheater/reheater tubing subject
to high temperature creep and corrosion. Appendix B contains a recommended microstructural
based method for estimating the creep damage and to predict the remaining creep life of
superheater/reheater tubing. This evaluation method should be performed at the failure and at
both ends of the removed tube sample to assist in evaluating the extent of damage relative to the
length of the tube.

Appendix C provides a detailed description along with illustrative examples of two life
assessment methods/models for long term overheat creep.

11.8 Longterm Actions and Prevention of Repeat Failures

Figure 11-8 provides an overview of the available options. The numbers in the figure are for
reference to the text only and do not imply a ranking among options; however, items outlined in
bold indicate the strategies that have been shown to be the most successful.

11-17
13622634
Longterm Overheat Creep

Figure 11-8
Strategies for preventing repeat failures by longterm overheating/creep.

Remaining life assessment is almost mandatory in order to determine the most economic strategy
among the available options. This includes understanding (i) the extent of damage, (ii) the
expected remaining life (RL) of affected areas, and (iii) the desired life (DL) of the unit,

The two primary approaches to prevent repeat failures are to (i) accept the temperatures being
experienced at the location of interest and upgrade the material used, or (ii) institute through re-
design or (less commonly) operating changes to lower the temperature at the susceptible
location.

11.8.1 Remaining Life Assessment (Option 1, Figure 11-8)

Remaining life assessment was discussed above in Section 11.7

11.8.2 Major SH/RH Redesign and Replacement (Option 2)

If the overheat problem is directly related to poor initial design of the superheater or reheater,
such as poor circuit design or different gas-touched lengths of the tubes between headers, then

11-18
13622634
 Longterm Overheat Creep

re-design might be the optimal economic solution. The main criteria should be to equalize the
gas-touched length of the tubes and to avoid large radiant cavities.

11.8.3 Redistribute Steam Flow (Option 3)

The technique of steam flow redistribution in superheaters has seen recent significant
development.8 Superheaters and reheaters are candidates for this method if the tubes are mostly
not at the end of expected life and there are significant temperature differences among the tubes.
Redistribution of steam flow can then serve to equalize the tube metal temperatures across the
superheater.

With steam flow controllers (orifices) of different sizes and lengths in the tubes coming out of
the inlet header, the flow can be decreased in "cold" tubes, which increases their temperatures.
This change will also increase the steam flow to "hot" tubes with a concurrent decrease of the
metal temperatures. Figure 11-9 shows that the decrease in tube metal temperature can lead to
significant increase in the remaining life of tubes. This tube which has operated for
approximately 200,000 hours would have an expected life of about 27,000 more hours.
Decreasing the tube metal temperature by ~ 17°C (30°F) would increase the expected life to over
75,000 hours.

11-19
13622634
Longterm Overheat Creep

Figure 11-9
Creep damage accumulation showing life extension obtained by reducing the metal
temperature. Source: K. Hara, et al.8

Steam flow redistribution occurs by modifying the tube-to-tube steam flow resistance through
the use of steam flow controllers (SFCs), Figure 11-10. The tube inner diameter is decreased for
a specific length, not with a sharp-edged orifice but with a tapered section. Typically the SFC is
installed just downstream from the inlet header in the penthouse, which allows easy access to
tubes and places the SFC out of the heat flux region.

11-20
13622634
 Longterm Overheat Creep

Figure 11-10
Schematic of a steam flow controller (SFC) and typical location in a superheater tube
bank. Source: K. Hara, et al.8

A roadmap of the overall process, which includes both technical and economic considerations, is
shown in Figure 11-11. Steps in the process include8:
• Analyzing existing information, sampling, and field measurement to determine the current
distribution of tube metal temperatures.
• Analysis of the steamside oxide scale on the ferritic sections of tubing to determine tube
temperatures.
• Determining the locations and dimensions of SFCs required to achieve the desired tube metal
and steam temperatures.
• Estimating the expected creep life improvement.
• Evaluating the economic and technical tradeoffs of identified options.

11-21
13622634
Longterm Overheat Creep

• Performing the optimal modifications.

• Validation testing and longterm monitoring of the results.

Figure 11-11
Road map for achieving improved superheater longevity by steam flow redistribution.
Source: K. Hara, et al.8

Technical factors to be considered include8: (i) analyzing the increase in pressure drop across the
superheater because too much plugging or flow restriction can lead to a heat rate penalty,
although this has been found to be very minor, (ii) the optimal inside diameter of each SFC
including some minimum value to avoid plugging, typically 12 mm (0.5 in.) or so, (iii) the
number of SFCs, (iv) the number of plugged and replaced tubes, and (v) the temperatures of
tubes after redistribution.

An overall strategy for the superheater will include consideration of SFCs along with selective
tube plugging and tube replacement. Plugging of tubes that are near the end of their expected
lives will not only extend the time to first failure but will have the additional advantage of
increasing flow that can be directed to decrease the temperature of "hot" tubes. Selective tube
replacement will also extend the time to failure.

11-22
13622634
 Longterm Overheat Creep

For a typical 400 MW unit, the cost to perform a steam flow redistribution modification is
around $210,000 ($525/MW). A condition assessment based on the oxide scale technique in a
similar size unit is around $25,000 (~ $60/MW).

11.8.4 Clean Tubes and Remove Sources of Blockage (Option 4)

Similarly, if the source of overheating is a blocked or partially blocked tube, high pressure fluid
flushing or chemical cleaning may be necessary, depending on the problem. This can be a time
consuming and repetitive option, which involves radiography of the bends, removal and
replacement of tubes, or cleaning of the accumulated scale, debris or deposits. Precautions
should be taken to avoid stress corrosion cracking of stainless steel tubes during cleaning. This
option has the added advantage of reducing exfoliation that can lead to solid particle erosion of
HP/IP turbines.

11.8.5 Adjust Fireside Combustion Conditions and/or Burners (Option 5)

This involves quantification of the problem through use of fireside testing and optimization.
Reference should be made to compilations of the appropriate procedures.9

11.8.6 Minimize Laning of Gases (Option 6)

Laning or channeling of gases through certain tube sections can lead to overheated tubes. This is
very difficult to overcome; it can be recognized and monitored by way of the cold air velocity
technique, which is described at length in the chapter on the flyash erosion damage mechanism.
However, the application of flow distribution screens, a means to control locally high velocities,
is made more difficult in the SH/RH by the high temperature environment.

11.9 Case Study

Creep cracking is characterized by the presence of grain boundary microvoids that coalesce and
link together to form a field of intergranular cracks that are approximately normal to the tensile
stress. These cracks can initiate midwall or from either the internal or external tube surfaces.
Time-to-failures may range from a few minutes to several years depending on the exposure
temperature and stress. Temperatures sufficient to cause creep damage often result in the
precipitation of massive grain boundary sigma phase particles.

A front pendant reheater tube developed a leak in the bend extrados after 43 years of service
(Figure 11-12). The tube was specified as ASME SA213-Grade TP304.

The rupture occurred with no appreciable tube swelling. Wall thicknesses in the bend had been
reduced to 80% in the extrados and increased to 115% in the intrados from the average thickness
away from the bend. Metallographic examination disclosed that the rupture occurred within a
field of unconnected, intergranular cracks that were most severe near the external tube surface
(Figure 11-13). These cracks were formed as grain boundary microvoids coalesced and linked
together (Figure 11-14), which identifies the failure mechanism as creep-rupture.

11-23
13622634
Longterm Overheat Creep

Figure 11-12
An example of a longterm overheat- creep failure that initiated at a bend extrados in a front
pendant reheater tube. No appreciable deformation is apparent.

Figure 11-13
The creep crack initiated at the tube external surface and propagated as grain boundary
creep cavities coalesced and linked together. (Mag: 16X, Electrolytic oxalic acid etch)

11-24
13622634
 Longterm Overheat Creep

Figure 11-14
Final failure occurred by creep-rupture as evidenced by linking of creep cavities to form
unconnected intergranular cracks. (Mag: 100X, Villella's etch)

11.10 References
1
Lamping, G.A. and R. M Arrowood, Jr., Manual for Investigation and Correction of Boiler
Tube Failures, Research Project 1890-1, Final Report CS-3945, Electric Power Research
Institute, Palo Alto, CA, April, 1985.
2
Personal Communication from E. Tolksdorf (VGB) to R.B. Dooley, February 16,1995.
3
Dooley, R.B. and H.J. Westwood, Analysis and Prevention of Boiler Tube Failures, Ontario
Hydro, Report 83/237G-31, Canadian Electrical Association, Montreal, Quebec, November,
1983.
4
American Society of Mechanical Engineers, Boiler and Pressure Vessel Code: Section I: Rules
for Construction of Power Boilers, ASME, New York. Data for Table 32-2 from 1983 edition.
5
Stultz, S.C. and J.B. Kitto, Steam: Its Generation and Use, Babcock & Wilcox Company,
Barberton, Ohio. Data for Table 32-2 from 1978 edition.
6
Singer, J.G., ed., Combustion Fossil Power: A Reference Book on Fuel Burning and Steam
Generation, Combustion Engineering, Inc., Windsor, Connecticut. Data for Table 32-2 from
1981 edition.
7
French, D.N., Metallurgical Features in Fossil-Fired Boilers, John Wiley & Sons, Wiley-
Interscience Publications, New York. Data for Table 32-2 from 1983 edition.

11-25
13622634
Longterm Overheat Creep

8
Hara, K., C. Lee, R. Moser, T. Rettig, and K. Clark, Improved Superheater Component
Longevity by Steam Flow Redistribution, Research Project 1893-13, Final Report TR-101697,
Electric Power Research Institute, Palo Alto, CA, December, 1992.
9
Sotter, J.G., J.A. Arnot, and T.M. Brown, Guidelines for Fireside Testing in Coal-Fired Power
Plants, Research Project 1891-3, Final Report CS-5552, Electric Power Research Institute, Palo
Alto, CA, March, 1988.

11-26
13622634
A
METALLURGICAL CHARACTERISTICS OF SA213
TYPE 304H, 321H, AND 347H STAINLESS STEEL
SUPERHEATER AND REHEATER TUBING

A.1 Tube Fabrication

A.1.1 Initial Melt

ASME Section I requires that boiler pressure parts exposed to fire or products of combustion
shall be made from open-hearth, electric furnace, or basis oxygen steel. The 300 series austenitic
stainless steels are electric furnace melted followed by argon-oxygen deoxidation (AOD) for
controlled oxidation of impurities such as sulfur and to obtain a desired level of carbon. In the
AOD process, the carbon monoxide partial pressure over the melt is lowered with argon or an
argon-nitrogen mixture thus enhancing the removal of sulfur and carbon by reactions with
oxygen.

A.1.2 Manufacture

Austenitic stainless steel superheater and reheater tubes are made in accordance to ASTM/ASME
SA-213/SA-213M. This specification requires that the tubing shall be seamless and shall be
either hot or cold finished, as specified.

A.1.3 Heat Treatment

The initial heat treatment of austenitic stainless steel tubing will have an important role in its
resistance to in-service degradation and resistance to creep or environmentally assisted cracking
or corrosion.

The required heat treating procedure for SA-213/213M Types 304H, 321H, and 347H depends
on the alloys and whether the tubes were hot –rolled or cold worked.

For Type 304H that has been either hot-rolled or cold worked, the required solution annealing
heat treatment consists of heating the material to a minimum of 1040oC (1900oF) and quenching
in water or rapidly cooling by other means.

For Type 321H and 347H that have been cold worked, the required solution annealing heat
o o
treatment consists of heating the material to a minimum of 1100 C (2010 F) and quenching in

A-1
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

water or rapidly cooling by other means. If these materials are hot-rolled then the minimum
solution annealing temperature is 1050oC (1925oF). Annealing above 1100°C (2010°F) may be
required in order to meet grain size requirements. This treatment should be followed by a
stabilization treatment.

In the case of Type 321H and Type 347H an optional stabilization heat treatment may be
performed after solution annealing to provide optimal corrosion and stress corrosion cracking
resistance. In the case of Type 321H the stabilization heat treatment would typically consist of
heating to 1040-1100°C (1920-2010°F) for 2 to5 minutes, followed by rapid cooling in air or
water. In the case of Type 347H the stabilization heat treatment would typically consist of
heating to 1000-1100°C (1830-2010°F) for 5 to 20 minutes followed by rapid cooling in air or
water.

The post-solution annealed grain size of cold worked Type 347H, as determined in accordance to
ASTM E112, shall be No. 7 or coarser. There is no grain size requirement for Types 304H, 321H
or hot-rolled Type 347H. Laboratory tests indicate that the stress rupture properties of Type 304
and 321 materials are somewhat insensitive to heat treatment and grain size (Figures A-1 and A-
2). The stress rupture properties of Type 347 material, on the other hand, are very sensitive to the
initial heat treatment and grain size (Figures A-3 and A-4).
o o
It is well known that solution annealing at temperatures above 1065 C (1950 F) may reduce these
alloys resistance to sensitization and intergranular cracking or corrosion because at these
temperatures the fine, dispersed, intragranular carbides begin to dissolve and the amount of
dissolved carbon increases. When the tubing is subsequently exposed to service temperatures this
dissolved carbon will tend to form chromium rich (Cr,Fe,M)23C6-type grain boundary carbides
which in turn deplete the grain boundaries of chromium and provide sites for intergranular
corrosion/attack (IGA) or intergranular stress corrosion cracking (IGSCC). The resistance of the
two stabilized alloys to sensitization can be increased to some extent by requiring that SA-
213/213M supplementary requirement S2 (stabilizing heat treatment) be performed. This
o o
requires that the material be re-solution annealed at a lower temperature (typically 900 C/1650 F)
than the initial solution annealing temperature and rapidly cooled. This lower temperature
solution annealing will provide the opportunity for the fine, dispersed, intragranular, somewhat
stable titanium carbides (in the case of Type 321H material), or columbium carbides (in the case
of 347H) to form thus lowering the level of dissolved carbon available for subsequent formation
of (Cr,Fe,M)23C6-type grain boundary carbides. This stabilization heat treatment may reduce the
stress rupture resistance of the alloy somewhat (Figures A-1, A-2, and A-3) and must be
considered in the design of the tubing.

A-2
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

Figure A-1
The rupture strength or life of Type 304 material is only very moderately influenced by
initial heat treatment or grain size.1 (Notes: 1) The numbers shown on the curves are the
rupture ductility, 2) 1200oF = 649oC, 1300oF = 704oC)

Figure A-2
The fupture strength or life of Type 321 material Is moderately influenced by initial heat
treatment or grain size.1 (Notes: 1) The numbers shown on the curves are the rupture
ductility, 2) 1200oF = 649oC)

A-3
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

Figure A-3
The rupture strength or life of Type 347 material is strongly influenced by initial heat
treatment or grain size. The material with a coarser grain size and a higher solution
annealing temperature exhibited better performance. This was believed to be a result of a
greater amount of sigma phase formation during the stress rupture test exposure for the
material with the lower solution annealing temperature.1 (Notes: 1) The numbers shown on
the curve are the rupture ductility, 2) 1200oF = 649oC, 1300oF = 704oC)

A-4
13622634
 Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

Figure A-4
The rupture strength or life of Type 347H material is strongly influenced by initial solution
annealing heat treatment. The material with a higher solution annealing temperatures
generally exhibited better performance. This effect is more significant at test temperatures
above 650oC (1202oF)2

In addition, supplementary requirement S4 (intergranular corrosion test) can also be required if

desired. This requires that the as-shipped material pass the intergranular corrosion tests
conducted by the manufacturer in accordance with ASTM A262, Practice E. This won’t assure
that the material will retain its resistance to intergranular corrosion after exposure to service
temperatures, but it will assure that its resistance to intergranular corrosion or stress corrosion
cracking during storage, shipping, erection, and early-service will be controlled to some extent.

Austenitic stainless steel tubing that has been cold worked during fabrication (e.g., cold
straightened, bent) after the required solution annealing or stabilizing heat treatments may have
locally high residual stresses (with magnitudes up to the materials yield strength or slightly
higher). The regions with high tensile residual stress will be likely sites for inter- or transgranular
stress corrosion cracking. Due to the slow creep relaxation rate of these alloys these fabrication-
induced residual stresses will persist throughout the service life of the tubing. Any out-of-service
periods, in which moisture can combine with caustic, sulfur-containing or chloride-containing
compounds on the tubing may result in stress corrosion cracking. The potential for stress
corrosion cracking can be minimized to a large extent by requiring that all cold worked tubing be
subjected to a resolution-annealing heat treatment. This requirement is not adequately addressed
in SA213/213M. The supplemental requirement that addresses this issue is S1. However
supplemental requirement S1 is not applicable to the high carbon, H-grades of austenitic stainless
steel. Specifically the range of stress relief heat treatment temperatures listed in S1.2 is not
optimal for the H-grades. The following post-cold-work stress relief temperatures are
recommended:

Alloy Post-cold work Re-Solution Annealing Treatment

o o
304H Heat to 1040 C (1900 F), followed by rapid cooling in air

A-5
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing
o o o o
321H Heat to 850 C to 950 C (1560 F to 1740 F for 10 to 15 minutes, followed by rapid
cooling in air
o o o o
347H Heat to 850 C to 950 C (1560 F to 1740 F for 10 to 15 minutes, followed by rapid
cooling in air

Keep in mind that tubing may distort during this stress relief heat treatment and require further
straightening. The goal is obviously to minimize the need for or extent of any final cold
straightening.

A.1.4 Pickling

Seamless tube is supplied in the solution-annealed and white-pickled condition or solution-

annealed in a bright-annealing process. During heat treatment of the tubing it will be exposed to
temperatures above the materials oxidation limits and unless bright annealing (i.e., heat treating
in a furnace with a protective atmosphere) is used the surfaces of the tubing will be coated with
an oxide scale that must be removed. This is removed by a pickling operation. This typically
o o o o
involves soaking the tubing in a hot (49 C/120 F to 60 C/140 F), agitated solution of 10% to 25%
nitric + 1% to 4% hydrofluoric acid, followed by dipping in a washtub and then application of a
preservative. This results in a scale free, somewhat roughened surface.

A.1.5 Bending

Annealing after cold bending is not normally necessary, but this point must be decided with
regard to the degree of bending and the operating conditions. Heat treatment, if any, should take
the form of stress relieving or solution annealing. The key issue with regard to cold bending is
the introduction of residual stresses that may lead to increased susceptibility to stress corrosion
cracking. It may be desirable to perform a stress relief, solution anneal, or solution anneal and
stabilization heat treatment following cold bending.

Hot bending is carried out at 1100-850°C (2010-1560°F) and should be followed by solution
annealing or solution anneal and stabilization heat treatment.

A.1.6 Welding

The weldability of Type 304H, 321H, and 347H is good. Suitable welding methods are manual
metal-arc welding with covered electrodes and gas shielded arc welding with the TIG and MIG
methods as first choice. Preheating and post-weld heat treatments are not normally necessary.

Since the material has low thermal conductivity and high thermal expansion, welding must be
carried out with a low heat input and with welding plans well thought out in advance so that the
deformation of the welded joint can be kept under control. If, despite these precautions, it is
foreseen that the residual stresses might impair the function of the weldment, it is recommend
that the entire structure be stress-relieved.

A-6
13622634
 Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

A.2 Chemical Composition

Type 304H is an austenitic, 18% chromium-8% nickel stainless chromium-nickel steel. It has a
nominal composition as shown below.

Example Chemical Composition of Type 304H Tube

C Si Mn P S Cr Ni

0.05 0.6 1.73 0.028 0.012 18.7 9.0

Type 321H is an austenitic, titanium-stabilized stainless chromium-nickel steel. It is suitable for

wet-corrosive service but also has good mechanical strength at high temperatures. It has a
nominal composition as shown below.

Example Chemical Composition of Type 321H Tube

C Si Mn P S Cr Ni Ti N

0.05 0.59 1.76 0.024 0.008 17.6 10.5 0.35 0.011

ASTM 347H is an austenitic, niobium-stabilized stainless chromium-nickel steel for use at

temperatures up to 816°C (1500°F). It has a nominal composition as shown below.

Example Chemical Composition of Type 347H Tube

C Si Mn P S Cr Ni Nb (Cb) N

0.05 0.59 1.64 0.019 0.014 17.6 10.4 0.87 0.025

Table A-1 provides a summary of the chemical requirements for SA213/213M Type 304H,
321H, and 347H.

A-7
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing
Table A-1
Chemical Requirements of SA-213/213M Type 304H, 321H, and 347H Steels

Grade TP 304H TP 321H TP 347H

UNS Designation S30409 S32109 S34709
Carbon 0.04 – 0.10 0.04 – 0.10 0.04 – 0.10
Manganese 2.00 max 2.00 m 2.00 m
Phosphorus 0.040 max 0.040 max 0.040 max
Sulfur 0.030 max 0.030 max 0.030 max
Silicon 0.75 max 0.75 max 0.75 max
Nickel 8.00 – 11.0 9.00 – 13.0 9.00 – 13.0
Chromium 18.0 – 20.0 17.0 – 20.0 17.0 – 20.0
Titanium NS 4 (%Carbon) – 0.60 NS
Niobium (Columbium) + NS NS 8(%Carbon) – 1.0
Tantalum

Due to their chemical composition, austenitic stainless steels combine the desirable properties of
elevated temperature strength and oxidation resistance, corrosion resistance and formability.
Carbon, an austenite stabilizer, plays a significant role in these properties by expanding the
region over which austenite is stable (Figure A-5) and by influencing properties through the
formation of a variety of metallic carbides in the microstructure, particularly chromium carbides.
During heat treatment and service exposure, however, the precipitation of chromium carbides on
grain boundaries can lead to sensitization and increased susceptibility to corrosion. For this
reason, the stabilized grades (Types 321 and 347) use the addition of strong carbide-forming
elements (primarily titanium and niobium, respectively) to lower the amount of dissolved carbon
present at a given heat treatment temperature (Figure A-6). This makes less carbon available for
chromium carbide formation, reducing the tendency for sensitization. As shown in Figure A-7,
nickel also plays a role in the formation of the desirable, stable austenitic structure; like carbon,
increasing nickel content also expands the austenite region.

A-8
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

Figure A-5
A Portion of an iron – chromium equilibrium phase diagram showing the influence of
dissolved carbon content on the stability of austenite (i.e., gamma iron, γ-Fe plus carbon).3

A-9
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

Figure A-6
Dissolved carbon under equilibrium conditions for Types 304, 321 and 347 stainless
steels. Since less dissolved carbon will be present in the Type 321 and 347 stainless steels
the amount of nickel is increased to 10% to 12% to ensure a fully austenitic material.
Lower levels of dissolved carbon are desirable because they impart enhanced
sensitization resistance and intergranular corrosion resistance.2

Figure A-7
A portion of an iron-chromium- nickel- carbon equilibrium phase diagram showing the
influence on nickel on the stability of austenite.3

A-10
13622634
 Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater Tubing

A.3 Physical Properties

The following tables list modulus of elasticity, thermal conductivity and thermal expansion coefficients for common boiler tubing
materials.

A.3.1 Modulus of Elasticity

Modulus of Elasticity; US Units

6
Modulus of Elasticity, E = Value Given (psi) for Temperature of:
70F X 10
200F 300F 400F 500F 600F 700F 800F 900F 1000F 1100F 1200F
Material Temperature
21C 93C 149C 205C 260C 316C 371C 427C 482C 538C 593C 649C
Carbon-moly steels 29.2 28.5 28.0 27.4 27.0 26.4 25.3 23.9 22.2 20.1 17.8 15.3
Intermediate alloy steels1 30.9 30.1 29.7 29.0 28.6 28.0 27.3 26.1 24.7 22.7 20.4 18.2
Austenitic stainless steels 28.3 27.6 27.0 26.5 25.8 25.3 24.8 24.1 23.5 22.8 22.1 21.2

Modulus of Elasticity; SI Units

6
Modulus of Elasticity, E = Value Given (kPa) for Temperature of:
21CX 10 93C 149C 205C 260C 316C 371C 427C 482C 538C 593C 649C
Material Temperature
70F 200F 300F 400F 500F 600F 700F 800F 900F 1000F 1100F 1200F
Carbon-moly steels 201.3 196.5 193.1 188.9 186.2 182.0 174.4 164.8 153.1 138.6 122.7 105.5
Intermediate alloy steels 1 213.0 207.5 204.8 199.9 197.2 193.1 188.2 180.0 170.3 156.5 140.7 125.5
Austenitic stainless steels 195.1 190.3 186.2 182.7 177.9 174.4 171.0 166.2 162.0 157.2 152.4 146.2
All data from ASME B31.1-2001.
Note 1: 5% Cr through 9% Cr.

A-11
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater Tubing

A.3.2 Thermal Conductivity

o
Coefficients of Thermal Conductivity; US Units (Btu/hr F)

70F 200F 300F 400F 500F 600F 700F 800F 900F 1000F 1100F 1200F
Material Temperature
21C 93C 149C 205C 260C 316C 371C 427C 482C 538C 593C 649C
1
Carbon steels 36.00 34.85 33.20 31.60 29.80 28.30 26.80 25.40 23.80 22.60 21.40 20.10
2
Austenitic stainless steels 8.35 8.90 9.35 9.80 10.23 10.70 11.10 11.55 11.95 12.42 12.90 13.30
21/ 4 Cr - 1 Mo (T22):
D. French (temp. range
16.70
unstated)
Bell ("high temperature") 22.90
Focke 18.80 17.90
Magnetite (Fe 3O4):
D. French 0.34
Armitt 1.81 1.76
Purdue (temp. range
0.26 - 0.46
unstated)
Bell ("high temperature") 0.35
Spinel Oxide (Fe,Cr,Mo)3O 4 3.33

o
Coefficients of Thermal Conductivity; SI Units (W/m K)

70F 200F 300F 400F 500F 600F 700F 800F 900F 1000F 1100F 1200F
Material Temperature
21C 93C 149C 205C 260C 316C 371C 427C 482C 538C 593C 649C
1
Carbon steels 62.31 60.32 57.46 54.69 51.58 48.98 46.39 43.96 41.19 39.12 37.04 34.79
Austenitic stainless steels 2 14.45 15.40 16.18 16.96 17.71 18.52 19.21 19.99 20.68 21.50 22.33 23.02
1
2 / 4 Cr - 1 Mo (T22):
D. French (temp. range
28.90
unstated)
Bell ("high temperature") 39.64
Focke 32.54 30.98
Magnetite (Fe 3O4):
D. French 0.59
Armitt 3.13 3.05
Purdue (temp. range
0.45 - 0.8
unstated)
Bell ("high temperature") 0.61
Spinel Oxide (Fe,Cr,Mo)3O 4 5.76

Note 1: This group includes carbon steel, carbon-moly steel, low-chrome steels (through 3% Cr).
Note 2: This group includes stainless steels having 18% Cr - 8% Ni through and including those having 25% Cr - 20% Ni.
All carbon and austenitic stainless steel data from ASME Boiler and Pressure Vessel Code, Sec.III; "Rules for Construction of Nuclear Power Plant Components"; 1971 edition.

A-12
13622634
 Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater Tubing

A.3.3 Thermal Expansion Coefficient

Thermal Expansion; US Units

o o 70F 200F 300F 400F 500F 600F 700F 800F 900F 1000F 1100F 1200F
Material 70 F (21 C) to
21C 93C 149C 205C 260C 316C 371C 427C 482C 538C 593C 649C
Carbon steel: carbon-moly steel low-
A 6.4 6.7 6.9 7.1 7.3 7.4 7.6 7.8 7.9 8.1 8.2 8.3
chrome steels (through 3% Cr)
B 0.0 1.0 1.9 2.8 3.7 4.7 5.7 6.8 7.9 9.0 10.1 11.3
5 Cr -1 Mo steels
A 6.4 6.7 6.9 7.0 7.1 7.2 7.3 7.3 7.4 7.5 7.6 7.6
B 0.0 1.1 1.9 2.8 3.6 4.6 5.5 6.4 7.4 8.4 9.4 10.4
9 Cr - 1 Mo steels A 5.8 6.0 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 7.1
B 0.0 0.9 1.7 2.5 3.3 4.1 5.0 5.9 6.8 7.7 8.7 9.6
Austenitic stainless steels A 8.5 8.9 9.2 9.5 9.7 9.8 10.0 10.1 10.2 10.3 10.5 10.6
B 0.0 1.4 2.5 3.7 5.0 6.3 7.5 8.8 10.2 11.5 12.9 14.3
1
TP 304; Oxide A 5.4 6.5 6.1 7.0 7.0 7.1 7.5 7.5 7.8 8.3 8.4
B 0.0 0.8 1.8 2.4 3.6 4.5 5.4 6.6 7.5 8.7 10.3 11.4
1 1
2 / 4 Cr - 1 Mo; Oxide A 6.2 4.3 6.1 5.8 6.6 6.5 6.8 7.1 7.0 7.3 7.3
B 0.0 1.0 1.2 2.4 3.0 4.2 4.9 6.0 7.1 7.8 9.0 9.9
1
9 Cr - 1 Mo; Oxide A 6.2 5.7 6.4 6.5 6.8 6.5 6.8 7.3 7.7 8.0 8.3
B 0.0 1.0 1.6 2.5 3.4 4.3 4.9 6.0 7.3 8.6 9.9 11.3
o -6 o o
A = mean coefficient of thermal expansion, in/in/ F x 10 in going from 70 F(21 C) to indicated temperature.
o o
B = linear thermal expansion, in/100 ft going from 70 F(21 C) to indicated temperature.
o
Multiply values shown in A by 1.8 to obtain coefficient of expansion in cm/cm/ C.
Multiply values shown in B by 8.33 to obtain linear expansion in cm per 100m.
All data, except as noted, from ASME B31.1-2001
Note 1: Date from Eberle, F. and J. H. Kitterman, "Scale Formations on Superheater Alloys Exposed to High Temperature Steam".

Thermal Expansion; SI Units

o o 21C 93C 149C 205C 260C 316C 371C 427C 482C 538C 593C 649C
Material 21 C (70 F) to
70F 200F 300F 400F 500F 600F 700F 800F 900F 1000F 1100F 1200F
Carbon steel: carbon-moly steel low- A 11.5 12.1 12.4 12.8 13.1 13.3 13.7 14.0 14.2 14.6 14.8 14.9
chrome steels (through 3% Cr)
B 0.0 8.3 15.8 23.3 30.8 39.2 47.5 56.6 65.8 75.0 84.1 94.1
5 Cr -1 Mo steels A 11.5 12.1 12.4 12.6 12.8 13.0 13.1 13.1 13.3 13.5 13.7 13.7
B 0.0 9.2 15.8 23.3 30.0 38.3 45.8 53.3 61.6 70.0 78.3 86.6
9 Cr - 1 Mo steels A 10.4 10.8 11.2 11.3 11.5 11.7 11.9 12.1 12.2 12.4 12.6 12.8
B 0.0 7.5 14.2 20.8 27.5 34.2 41.7 49.1 56.6 64.1 72.5 80.0
Austenitic stainless steels A 15.3 16.0 16.6 17.1 17.5 17.6 18.0 18.2 18.4 18.5 18.9 19.1
B 0.0 11.7 20.8 30.8 41.7 52.5 62.5 73.3 85.0 95.8 107.5 119.1
1
TP 304; Oxide A 9.7 11.7 10.9 12.6 12.6 12.8 13.5 13.5 14.0 14.9 15.1
B 0.0 7.0 14.9 20.0 30.0 37.1 44.7 54.7 62.2 72.5 85.5 94.9
2 1 / 4 Cr - 1 Mo; Oxide 1
A 11.2 7.7 10.9 10.4 11.9 11.7 12.2 12.8 12.6 13.1 13.1
B 0.0 8.1 9.9 20.0 24.9 35.0 40.9 49.6 58.9 65.1 75.1 82.5
1
9 Cr - 1 Mo; Oxide A 11.2 10.3 11.5 11.7 12.2 11.7 12.2 13.1 13.9 14.4 14.9
B 0.0 8.1 13.1 21.1 27.9 36.0 40.9 49.6 60.6 71.6 82.4 93.8
o -6 o o
A = mean coefficient of thermal expansion, cm/cm/ C x 10 in going from 21 F(70 C) to indicated temperature.
o o
B = linear thermal expansion, cm/100m going from 21 F(70 C) to indicated temperature.
o
Divide values shown in A by 1.8 to obtain coefficient of expansion in in/in/ F.
Divide values shown in B by 8.33 to obtain linear expansion in inches per 100ft.
All data, except as noted, from ASME B31.1-2001
Note 1: Date from Eberle, F. and J. H. Kitterman, "Scale Formations on Superheater Alloys Exposed to High Temperature Steam".

A-13
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

A.4 Mechanical Properties

A.4.1 Tensile Strength

The yield and ultimate tensile strength of Type 304H, Type 321H, and Type 347H materials is
shown in Figures A-8, A-9, and A-10

Figure A-8
Yield and ultimate tensile strength of 304H material. (Ref: NRIM CDS 4B4)

A-14
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

Figure A-9
Yield and ultimate tensile strength of 321H material. (Ref: NRIM CDS 5B5)

Figure A-10
Yield and ultimate tensile strength of 347H material. (Ref: NRIM CDS 28B6)

A-15
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

A.4.2 Stress Rupture

Regression analyses were performed on an extensive database of longterm stress rupture tests of
Types 304H, 321H, and 347H materials. Both Manson Haferd and Larson Miller models were
evaluated. Both models gave acceptable results however the model extrapolations to low stress
levels seemed to be more realistic for the Larson Miller models so these are presented as the best
fit models. It is anticipated that some additional improvement in these predictive models might
be attainable with the use of more complex models. The final model used had the following
form:
2
Log (truphat) = C1 + C2/T + C3 Log(S)/T + C4 [Log(S)] /T

Where:
truphat = median time to rupture, hours
S = stress, MPa
T = temperature, K = (oC +273.15)

The empirical coefficients for Types 304H, 321H, and 347H are summarized in the following
table:
Alloy C1 C2 C3 C4

304H -11.389 18,746.602 -268.416 -969.675

321H -10.108 16,527.621 315.797 -998.095

347H -6.8261 16,973.063 -3,610.762 137.473

The Larson Miller plots are presented in Figures A-11, A-13, and A-15.

The model residuals (variation of each test result from the median predicted life from the
regression model) were defined as the ratio of the test time to rupture to the median predicted
time to rupture. These ratios were then fit to Weibull, normal, and Log Normal distributions. The
Log Normal distributions provided the best fit of this data. Using the LOGNORMAL function in
an EXCEL spreadsheet provided the following function:

Fraction failed = LOGNORMAL (trup/truphat, µ, σ)

Where:
µ = Mean of all the ln(trup/truphat) values
σ = Standard deviation of all the ln(trup/truphat) values

A-16
13622634
 Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

The values of µ and σ that were determined for each alloy are summarized in the following table.
The graphical fit of the resulting equations is documented in Figures A-12, A-14, and A-16.
Alloy µ σ

304H 4.0781E-8 0.65797

321H 3.6410E-8 0.65110

347H -8.7119E-6 0.85120

Figure A-11
Normalized Larson-Miller Plot of 216 stress rupture tests of Type 304H material.

A-17
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

Figure A-12
Distribution of failure times for Type 304H laboratory stress rupture tests.

Figure A-13
Normalized Larson-Miller Plot of 233 stress rupture tests of Type 321H material.

A-18
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

Figure A-14
Distribution of failure times for Type 321H laboratory stress rupture tests.

A-19
13622634
Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

Figure A-15
Normalized Larson-Miller Plot of 289 stress rupture tests of Type 347H material.

A-20
13622634
 Metallurgical Characteristics of SA213 Type 304H, 321H, and 347H Stainless Steel Superheater and Reheater
Tubing

Figure A-16
Distribution of failure times for Type 347H laboratory stress rupture tests.

A.5 References
1
Digest of Steels For High Temperature Service, 6th Edition, The Timken Roller Bearing
Company, Steel & Tube Division, 1957
2
ASM Metals Handbook, 10th Edition, Elevated Properties of Stainless Steels, pg 945
3
Stainless Welding Engineering, Stoody Company, 1977
4
NRIM Creep Data Sheet Number 4B
5
NRIM Creep Data Sheet Number 5B
6
NRIM Creep Data Sheet Number 28

A-21
13622634
13622634
B
METALLURGICAL EVALUATION TO ASSESS THE
CONDITION OF AUSTENITIC STAINLESS STEEL
TUBING

This appendix describes metallography-based techniques for assessing the current condition of
austenitic stainless steel tubing, primarily by evaluating the presence and extent of i) creep
microvoids and ii) sigma phase in the microstructure. Tube sampling can be used to assess the
condition of specific tubing if severe operating conditions or a history of failures/degradation
indicate that significant damage is accumulating. This appendix shows how the presence of creep
microvoids and, to a lesser extent, sigma phase in the tube sample microstructures can both be
used to provide an estimate of the condition and remaining life of austenitic tubing, thereby
increasing the amount and usefulness of information gained by tube sampling. A roadmap for the
evaluations described in this appendix is provided in Figure B-1.

B.1 Common Degradation Modes

Austenitic stainless steel superheater and reheater tubing failures have occurred following
service exposures that are much less severe than the anticipated design lives discussed in Chapter
2. The reverse is also true, with tubes sometimes exposed to service conditions that are more
severe than anticipated, resulting in failures. To understand why this occurs requires an
understanding of the ways that these materials degrade when exposed to elevated temperatures
and the unanticipated environments and stresses that may occur during service.

The common microstructural degradation experienced by austenitic stainless steel superheater or

reheater tubes includes sensitization, sigma phase formation and grain growth. These in turn can
play a role in and/or lead to degradation by specific tube damage mechanisms. A description of
the common microstructural degradation modes is presented in the remainder of this chapter. The
common tube damage/failure mechanisms are each presented in Chapters 4 through 11.

B-1
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

Figure B-1
Appendix B roadmap for metallurgical evaluation of austenitic tubing.

B-2
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

B.1.1 Sensitization

When austenitic stainless steels are exposed to temperatures within the range of 430°C (805°F)
to 900°C (1650°F), chromium carbides will form on the grain boundaries (Figure B-2). This
condition will also result in a chromium-depleted region along the grain boundaries. In this
“sensitized” condition, the material will have increased susceptibility to intergranular corrosion,
intergranular stress corrosion cracking, and creep cavitation. All of the common austenitic
stainless steel tubing alloys will sensitize in service. The stabilized grades such as 321H and
347H will sensitize somewhat slower but after a few years of typical SH/RH service they will
also sensitize. The same is true of the lower carbon grades of these alloys.

Figure B-2
When austenitic stainless steels are exposed to prolonged exposure to temperatures in
the range of 430oC (805oF) to 900oC (1650oF) chromium carbides will form on the grain
boundaries. This condition will result in a chromium-depleted region along the grain
boundaries. In this “sensitized” condition the material will have increased susceptibility to
intergranular corrosion, intergranular stress corrosion cracking, and creep cavitation.
Note that with increased levels of dissolved carbon the rate and temperature range over
which sensitization occurs increases1.

B.1.2 Sigma Phase Formation

Sigma phase (σ -phase) is a hard (>HRC 60), brittle, non-magnetic phase that forms during
service in austenitic stainless steels superheater and reheater tubes. The approximate composition
of σ -phase is FeCr (Figure B-3), although it is commonly listed as Fe(Cr,Ni,Mo)because its
composition depends on the specific alloy system in which it forms.

Sigma phase usually forms when the material is exposed to temperatures in the range of
approximately 550°C to 900°C (1022°F to 1652°F). Sigma phase is unstable and redissolves if
heated to a temperature of about 871°C (1600°F), the exact dissolution temperature depends on

B-3
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

the composition. For Type 304H material σ-phase formation requires more than 1000 hours of
exposure in this critical temperature range (Figure B-4).

Sigma phase can be extremely deleterious to material performance. The formation of even a few
volume fraction percentage points of σ-phase can reduce the stress rupture (Figure B-5) and
corrosion resistance of the material. Other material properties, such as ductility and toughness
are significantly degraded by the presence of even small amounts of σ -phase.

Figure B-3
Iron – chromium - nickel equilibrium phase diagram (section at 8% nickel). The two phases
that are relevant to austenitic stainless steel SH/RH tubes are Austenite (Gamma Iron, γ +
Carbon,) and Sigma Phase, σ (a grain boundary phase comprised of approximately 50%
chromium and 50% iron). The addition of carbon will expand out the region of stability of
Gamma Iron, γ-Fe. Note that even without the benefit of carbon additions Sigma Phase is
an equilibrium phase for chromium levels above approximately 18%2

B-4
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

Figure B-4
Time-Temperature-Transformation curves for Types 304H, 321H, and 347H materials3. Note
that even the stabilized grades of material will sensitize and form sigma phase if they are
exposed to prolonged temperatures approaching 600°C (1112°F). At 650°C (1202°F) all
three alloys will begin to form sigma phase after approximately 10,000 hrs.

B.1.3 Grain Growth

Another change in the microstructure that will occur over long periods of time is grain growth.
Depending upon the time and temperature, grain growth can begin at temperatures as low as
1150°F-1200°F if the time is long enough. For conventional fossil-fuel fired plant the metal
temperature will typically be below the temperatures where grain growth will be significant.
While grain size can play a role in the number and distribution of creep microvoids, the effect is
relatively minor, as discussed in Section B.3.2.2.

B-5
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

Figure B-5
Stress rupture curves for a "Weak" (A) and an "Average" Heat (B) of Type 304 stainless
steel tubing material showing the four unique stress rupture cracking mechanisms as a
function of stress, temperature, and time-to-failure. The weak heat contained a high
(0.047wt%) aluminum content which lead to precipitation of aluminum nitride associated
with the grain boundary sigma phase. This caused a reduction in life or strength due to
early formation of cracks in the sigma phase4. Note: 1 kgf/mm2 = 9.8 MPa = 1422 psi

B-6
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

B.2 Metallurgical Characteristics of Tubing

Prior to embarking on a thorough life assessment, it is crucial that the active tube damage
mechanisms be verified and the current metallurgical state of selected tubes within the
population of concern be determined. This section provides a detailed description of the
assessment of metallurgical features, in particular sigma phase formation and creep cavitation. A
careful determination regarding the extent or severity of these metallurgical features will not
provide a direct estimate of remaining life but will provide a general idea if the tubes are
approaching their end-of-life or still have considerable life remaining.

Austenitic stainless steel tubes are generally supplied in the solution annealed condition with a
microstructure consisting of single phase equiaxed austenitic grains. Microstructural changes,
however, occur in these tubes during long-term exposure during normal elevated temperature
service. Microstructural examination can provide information useful for estimating the thermal
history, accumulation of creep damage, and remaining creep life. Metallographic based methods
may be used to supplement creep damage calculations based on measured or estimated
temperature, stress, and material properties; or may be used when adequate temperature and
stress information is not available. Creep damage and remaining life results obtained using this
method should be considered as semiquantitative estimates.

Carbides and sigma are second phase particles that precipitate during service preferentially at
grain boundaries, twin boundaries, and slip planes. In addition, if the tube has operated at
sufficiently high temperature and stress, in the creep regime, microvoids will form at the
interface to second phase particles, primarily at grain boundaries. As the tube accumulates
additional creep damage, these microvoids will increase in number and size and eventually align
normal to the principal stress to form microcracks and macrocracks. The microvoid morphology
is a function of the stress level at which the creep damage occurred, which can provide useful
information regarding the service stress.

Fracture mechanism maps based on long-term creep-rupture tests have been developed that show
a transition from intergranular to transgranular creep fracture with increasing stress.
Intergranular fractures are further categorized as: i) wedge-type cavitation, ii) grain boundary
cavitation, and iii) cavity initiation at the interface between grain boundary sigma phase and the
5
austenite matrix . A reduction in creep-rupture properties has been observed that coincides with
the formation of intergranular sigma phase.

Creep-rupture failures of stainless steel tubes that have operated at conditions moderately above
the normal design stress and temperature will initiate creep cavities at the interface between
sigma phase and austenite matrix. In these cases, the rupture life is controlled primarily by
growth of these creep cavities. Tube degradation and creep damage accumulates in the
5
following sequence :

• Sigma phase precipitates at grain boundaries

• Creep cavities nucleate at the interface between the sigma phase and austenitic matrix

B-7
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

• Creep cavities grow to crack-like voids along grain boundaries

• Fracture occurs due to interlinking of the crack-like voids

Although cavitation is considered the primary controlling and recognizable feature in elevated
temperature degraded tubing, metallographic evaluation of sigma phase may provide useful
information regarding the lifetime average tube metal temperature before cavitation becomes the
dominant metallographic feature. Combined with tube stress and creep-rupture material
properties, this information can provide an approximate indication of the tube service exposure
and degradation relative to the accumulation of creep damage. Some uncertainty, however, is
introduced into this analytical method because, in addition to time and temperature,
compositional differences also influence sigma formation.

B.2.1 Etch to Identify Second Phase Precipitates

Second phase particles that precipitate in stainless steel boiler tubes during long-term high-
temperature exposure are sigma (FeaCr, a = ¼ to 4) and carbides (M23Cr6, M=Cr,Fe). Sigma
phase forms at an elevated temperature in the range of approximately 550°C to 900°C (1020 -
1650°F), with a maximum rate of formation between 750°C and 850°C (1382 - 1560°F). Sigma
will dissolve into the austenitic matrix by solution annealing. Sigma nucleates first at triple
points and then at grain boundaries. The kinetics of sigma precipitation increases with delta
ferrite content, deformation, and recrystallization. Creep strain in furnace tubes, therefore, is
expected to enhance precipitation of sigma phase. Some elements (molybdenum, silicon,
vanadium, tungsten, titanium, and niobium) promote its formation and some elements (carbon,
6
nitrogen, and nickel) retard sigma formation . At ambient temperature, sigma phase is a hard,
brittle, nonmetallic intermetallic compound. It is generally undesirable because it imparts
ambient temperature brittleness and degrades creep strength.

Carbides precipitate within a temperature range of 430°C to 900°C (800 - 1650°F), with the
fastest reaction rate between 650°C and 700°C (1200 - 1290°F). Precipitation occurs at grain and
twin boundaries and at inclusion/matrix-phase boundaries. Alloy addition of titanium or niobium
cause the carbides to form as MC (M=Ti or Nb) rather than M23C6.

Various methods are available to differentiate between sigma phase and carbide particles. These
include special etching techniques to selectively stain sigma phase and elemental chemical
analysis (for example, energy dispersive spectroscopy using electron microscopy).

A two step etching technique is described below in which the sample is first etched using
Vilella’s reagent to outline the second phase particles. The sample is next electrolytically etched
using concentrated sodium or potassium hydroxide (NaOH or KOH) to stain the sigma phase.
The sample must not be wiped following this etch because it will remove the stain.

B-8
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

B.2.1.1 Example of Second Phase Etch Method

Identification of sigma phase in Type 304H stainless steel is illustrated below. The
microstructure as delineated by alternate polish-etch sequence using Vilella’s reagent contains
both large and small outlined second phase particles (Figure B-6). Electrolytic etching using
concentrated NaOH (1.5 volt for 20 seconds) and KOH (1.5 volt for 10 seconds) stained the large
particles but had no effect on some of the smaller particles (Figure B-7). The large particles,
therefore, are identified as sigma and the smaller particles that remain clear are identified as
carbides.

Small second phase particles precipitated in grain boundaries and slip planes in the “cold” side
(trailing side of the tube in the gas path) of a ruptured Type 304H tube (Figure B-8). Although
the small size of these particles makes it difficult to observe if they are stained, the larger of
these particles were not stained by the concentrated caustic electrolytic etch, which is consistent
with carbide particles (Figure B-8).

B-9
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)
Figure B-6
Etched surface of a service degraded Type 304H stainless steel tube sample showing (A)
and (B) outlined second phase particles with fully developed microvoids (arrow). (A) and
(B) were prepared with multiple 3, 3, and 2 minute etches. (MAG: A 500X, B 1000X, Vilella’s
Etch)

B-10
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)
Figure B-7
Etched surface of a service degraded Type 304H stainless steel tube sample showing
stained Sigma Phase particles with fully developed microvoids. Arrow in (A) marks Sigma.
Arrow in (B) marks a carbide. (MAG: 1000X, Vilella’s Etch plus (A) NaOH and (B) KOH
eectrolytic etch)

B-11
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)

Figure B-8
Etched surface of a “cold” side service exposed Type 304H stainless steel tube sample
showing second phase precipitation on grain boundary (arrow), twin boundaries, and slip
planes. These particles are likely carbides because they were not stained by the
concentrated caustic electrolytic etch. (MAG: 1000X, (A) Multiple Vilella’s Etch and (B)
same plus NaOH electrolytic etch)

Relatively large sigma phase and smaller carbides phases in Type 321 stainless steel are also
illustrated. A multiple polish-etch sequence using Vilella’s reagent outlines both second phase
particles and reveals creep microvoids, which are aligned approximately normal to the principal
stress (Figure B-9A). The large particles that are stained by the concentrated caustic electrolytic
etch are identified as sigma (Figure B-9B). Microvoids visible in this creep damaged tube are
associated with sigma phase. The smaller particles within the austenitic matrix that are not
stained are identified as carbides. Creep microvoids and second phase particles in Type 347H are
illustrated in Figure B-10.

B-12
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)
Figure B-9
Etched surface of a service degraded Type 321H stainless steel tube sample showing (A)
outlined second phase particles (arrow) with aligned microvoids and (B) stained particles
(arrow) identified as Sigma. (MAG: 1000X, (A) Multiple Vilella’s Etch and (B) same etch plus
KOH electrolytic 1.5 volts for 8 seconds)

B-13
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)

Figure B-10
Etched surface of a service degraded Type 347H stainless steel tube sample showing (A)
outlined second phase particles (arrow) with aligned microvoids (black cavities) and (B)
stained particles (arrow) identified as sigma. (MAG: 1000X, (A) Multiple Vilella’s Etch and
(B) same etch plus KOH electrolytic 1.5 volts for 5 seconds)

B.2.1.2 Electrolytic Oxalic Acid Etch

Oxalic acid is a common electrolytic etchant used for stainless steels. Its effect on service
exposed Types 304H, 321H, and 347H stainless steel tubes is illustrated below. Etching
parameters are commonly specified as volts and times. Etching parameters of 5 volts and 10
seconds were used in the following examples. Metallographic samples were first etched using the
three stage Vilella’s reagent procedure to reveal second phase particles and microvoids. The
samples were then electrolytic etched with oxalic acid. Large second phase grain boundary
particles (identified previously as sigma) outlined by the Vilella’s reagent are dissolved by the

B-14
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

oxalic acid etch in Type 304H (Figure B-11) and Type 321H (Figure B-12). Sigma phase,
however, is not apparent in Type 347H (Figure B-13).

(A)

(B)
Figure B-11
Etched surface of a service degraded Type 304H stainless steel tube sample showing (A)
outlined second phase particles (arrow) and microvoids (black cavities) and (B) sigma
phase (arrow) dissolved away by the electrolytic oxalic acid etch. (MAG: 500X, (A) Multiple
Vilella’s Etch and (B) same etch plus electrolytic oxalic acid 5 volts for 10 seconds)

B-15
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)
Figure B-12
Etched surface of a service degraded Type 321H stainless steel tube sample showing (A)
outlined second phase particles (arrow) and microvoids (black cavities) and (B) second
phase particles (arrow) dissolved away by the electrolytic oxalic acid etch. (MAG: 500X, (A)
Multiple Vilella’s Etch and (B) same etch plus electrolytic oxalic acid 5 volts for 14
seconds)

B-16
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)

Figure B-13
Etched surface of a service degraded Type 347H stainless steel tube sample showing (A)
outlined second phase particles (arrow) and microvoids (black cavities) and (B) similar
microstructure following the electrolytic oxalic acid etch. (MAG: 500X, (A) Multiple Vilella’s
Etch and (B) same etch plus electrolytic oxalic acid 5 volts for 14 seconds)

Since electrolytic oxalic acid can dissolve second phase particles (both sigma and carbides
depending on the etching parameters), its use is not recommended for revealing the true
microstructure of service degraded stainless steel boiler tubes.

B.2.2 Second Phase Compositional Analysis

Elemental analysis using an energy dispersive spectroscopy (EDS) attachment to a scanning

electron microscope (SEM) is useful to confirm the metallographic identification of sigma and
carbide phases. EDS spectra obtained from the Type 304H austenite matrix and a large second
phase particle (metallographically identified as sigma) are shown in Figure B-14.

B-17
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

Semiquantitative analysis of these spectra indicate the following approximate compositions for
the matrix and sigma phase:

Element Matrix Sigma

Iron 73% 61%
Chromium 18% 32%
Nickel 8% 4%
Silicon 0.6% 0.6%

The higher chromium/iron ratio in the second phase compared to the matrix is consistent with
sigma. The high chromium and low iron content of the sigma phase relative to the matrix is also
illustrated in a compositional dot map (Figure B-15).

B-18
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)
Figure B-14
Energy dispersive spectra from a Type 304H stainless steel tube showing the composition
of the (A) austenite matrix and (B) sigma phase particle. Note the high chromium/iron
(Cr/Fe) ratio of the sigma phase compared to the austenite matrix.

B-19
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)

(C)
Figure B-15
Compositional dot maps showing (A) chromium and (B) iron contents of the sigma phase
and grain boundary carbide compared to the austenite matrix. Sigma has a higher
chromium content and lower iron content than the matrix. (C) is a SEM image.

B-20
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

The dot map includes several narrow second phase particles that precipitated on grain boundaries
that were not stained by the concentrated caustic electrolytic etch (arrow in Figures B-7B and
B-15C), and are therefore metallographically identified as a carbides. This downward tail from
the sigma phase is shown to be low in iron (Figure B-15), as expected for chromium carbide.

Compositional dot maps of the Type 321H stainless steel reveal that the distinct yellow
rectangular precipitates common in this alloy are high in titanium and low in iron compared to
the matrix (Figure B-16). This composition is consistent with titanium nitrides found in titanium
stabilized stainless steel. Other related compounds found in this alloy are titanium carbides and
titanium carbonitrides. Titanium carbide is gray in color and has a less distinct rectangular shape.
Titanium carbonitrides have an intermediate appearance that varies with the C/N ratio.

(A)

(B)
Figure B-16
Compositional dot map showing (A) chromium and (B) iron concentrations. The titanium
nitride particles (rectangular shapes) have higher titanium and lower iron concentrations
than the austenite matrix.

B-21
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

B.2.3 Etch to Reveal Creep Microvoids

Microstructural characterization of creep microvoids requires a distinct preparation technique

compared to normal metallographic preparation that does not involve microvoids. The special
procedure must be capable of removing the disturbed layer of metal that will otherwise cover and
mask microvoids. It must also leave second phase particles intact since their dissolution would
produce false positive indications of creep damage.

The final and unique steps in this procedure consist of alternate polishing and etching. The
method described below has been found to give satisfactory results for service degraded Types
304H, 321H, and 347H stainless steel boiler tubes:
• Rough grind on water lubricated 80 and 120 grit abrasive belts.
• Hand grind on successively finer water lubricated 240, 320, 400, and 600 grit papers.
• Rough polish with a napless nylon cloth impregnated with 6 micron diamond paste.
• Etch with Vilella’s reagent for 4 minutes.
• Polish with a napless nylon cloth impregnated with 1 micron diamond paste.
• Etch with Vilella’s reagent for 3 to 4 minutes.
• Final polish with a low nap cloth impregnated with 0.05 micron alumina paste.
• Etch with Vilella’s reagent for 2 to 4 minutes.
Etching times are given as a range to accommodate the different etching characteristics of the
various stainless steel alloys. The minimum etching times are usually adequate for Type 304H.
Longer times may be required for Types 321H and 347H.

Since etching will be accelerated by heat generated during polishing that is retained in the
sample, the sample should be cooled under running water before etching. A convenient method
for applying Vilella’s reagent is to drip it onto the polished surface using a disposable pipette.
Swabbing is not recommended because it can cause stains.

B.2.3.1 Example of Microvoid Etch Method

The importance of proper metallographic preparation to reveal microvoids is illustrated in the

following example. The as-polished surface of a Type 304H service exposed superheater tube
reveals few small isolated voids in the microstructure (Figure B-17A). Microvoids are apparent
at this stage of preparation only if they are filled or lined with oxides. Second phase particles are
revealed by a one-minute etch with Vilella’s reagent (Figure B-17B). This single etch, however,
does not reveal the full extent of the microvoids. Upon completion of the recommended
procedure, the number and size of the microvoids reveals severe creep damage in this tube
sample (Figure B-17C). A similar example illustrating the necessity of proper metallographic
preparation for Type 321 stainless steel is shown in Figure B-18. In both of these examples,
creep damage is sufficiently severe to result in the alignment of microvoids normal to the
principal stress. These microvoids formed preferentially at the interface to second phase
particles.

B-22
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)

(C)
Figure B-17
Service degraded Type 304H stainless steel tube sample showing (A) small voids in the
as-polished condition, (B) outlined second phase particles with some microvoids, and (C)
fully developed microvoids (black cavities). Arrows mark the same location (A) as-
polished. (B) 1 minute etch. (C) multiple 3, 3, and 2 minute etches. (MAG: 500X, Vilella’s
Etch)

B-23
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

(A)

(B)
Figure B-18
Etched surface of a service degraded Type 321H stainless steel tube sample showing (A)
outlined second phase particles with minor microvoids and (B) fully developed microvoids
(black aligned cavities). Arrows mark the same location. (A) single 1 minute etch. (B)
multiple 3, 3, and 3 minute etches. (MAG: 500X, Vilella’s Etch)

B.3 Remaining Life v. Metallurgical State

B.3.1 Sigma Phase Characterization

Various methods are available to quantify sigma phase precipitation as a function of the time-
temperature exposure history. The Larson-Miller parameter (LMP) is one method that has been
shown to be useful in the temperature range of interest, between 600°C and 750°C (1202 -
1382°F). This method does not reflect the decreased sigma precipitation at higher temperatures.
The LMP is defined as:

B-24
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

LMP = T x [C + Log(t)] (B-1)

Where

LMP = Larson-Miller parameter

T = Temperature (°K = °C + 273.15)
C = Constant
t = Exposure time (hours)

Unstressed stainless steel coupons that had been exposed up to 50,000 hours at temperatures
between 600°C to 750°C (1112 - 1382°F) were metallographically examined to characterize the
sigma phase precipitates7. The average widths of the six largest sigma phase particles in
representative photomicrographs were selected for measurement. Sigma phase measurements
from these coupons are plotted as a function of the LMP in Figures B-19, B-20, and B-21. These
data indicate that titanium and niobium stabilized stainless steels (Types 321H and 347H) have
similar sigma responses, whereas, unstabilized Type 304H stainless steel has significantly slower
sigma precipitation kinetics.

5
Type 304H
4 2.0430E-03x
y = 4.2758E-03e
3
2
1
0
2000 2500 3000 3500
LMP = K [log(Hr) -1.4)]

Figure B-19
Sigma phase thickness versus LMP for Types 304H stainless steels.

B-25
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

8
Type 321H
6.7954E-04x
6 y = 2.4647E-05e

0
15000 16000 17000 18000 19000
LMP = K [ log(Hr) +13.1]

Figure B-20
Sigma phase thickness versus LMP for Types 321H stainless steels.

6
Type 347H
5
4.3304E-04x
4 y = 1.1168E-03e
3
2
1
0
15500 16500 17500 18500 19500
LMP = K [log(Hr) +13.9]

Figure B-21
Sigma phase thickness versus LMP for Type 347H stainless steels.

B-26
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

The trend curves provide a method of estimating the thermal history of tubes based on sigma
phase thickness measurements provided the service hours are known. The lifetime average
temperatures can be calculated from these curves as:

T = A1 x ln(A2 x WS) / [A3+log(t)] (B-2)

Where: T = Temperature (°K = °C + 273.15)
WS = Sigma width (microns)
t = Exposure time (hours)

Alloy A1 A2 A3

Type 304H 489.48 233.87 -1.4

Type 321H 1471.58 40573 13.1

Type 347H 2309.26 895.42 13.9

If the tube stress is known, the total creep life can be estimated by combining the temperature
inferred from sigma phase measurements with the creep-rupture relationships given in Appendix
A, which are repeated below:

Log (truphat) = C1 + C2/T + C3 Log(S)/T + C4 [Log(S)]2/T (B-3)

Where:
truphat = median time to rupture, hours
S = stress, MPa
T = temperature, K = (oC +273.15)

Alloy C1 C2 C3 C4

304H -11.389 18,746.602 -268.416 -969.675

321H -10.108 16,527.621 315.797 -998.095

347H -6.8261 16,073.063 -3,610.762 137.473

Creep damage calculations based on metallographic evaluations should be considered as

approximate because of scatter inherent in both sigma phase kinetics and creep-rupture
properties. The actual sigma phase reaction kinetics and creep-rupture properties of any given
heat of material may differ from the mean trend curves given above because of the influence of
minor compositional changes.

B-27
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

B.3.1.1 Example of Sigma Phase Based Temperature Estimates

Examples in which sigma phase measurements reveal the limited remaining life of service
degraded tubing are provided below. The pertinent features are:

Example 1 Example 2

Alloy Type 321H Type 304H

Pressure (MPa) 12.8 18.3

OD (mm) 54 64

Wall (mm) 6 7

Service time (hours) 270,000 72,000

Coupons removed from the “hot” sides of the tubes facing the gas flow were metallographically
prepared to reveal sigma phase. This preparation included three stages of etching with Vilella’s
reagent to outline the second phase particles and to reveal any creep cavitation that might be
present, as shown in Figures B-22 and B-23. Sigma phase was confirmed using an electrolytic
sodium hydroxide etch.

Figure B-22
Microstructure from the hot side of a service degraded Type 321H furnace tube showing
creep cavitation that initiated at grain boundary sigma phase particles. (MAG: 500X,
Vilella’s Reagent Etch for 4+3+2 minutes)

B-28
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

Figure B-23
Microstructure from the hot side of a service degraded Type 304H furnace tube showing
creep cavitation that initiated at grain boundary sigma phase particles. (MAG: 500X,
Vilella’s Reagent Etch for 4+3+2 minutes)

Tube metal temperatures were calculated from the average of six largest sigma phase precipitates
and service hours using Equation B-2 as:

Example 1 Example 2

Sigma width (microns) 2.3 2.1

Tube temperature (°C) 636 604

(°F) 1176 1120

The total tube creep lives were next calculated from temperatures and stresses using Equation
B-3. The mean diameter stress formula was used in these examples.

Example 1 Example 2

Total creep life (hours) 291,000 341,000

Since the tube in Example 1 has already accumulated 270,000 hours of operation, these
calculations indicate significant distress of this tube and that failure may be imminent, as
confirmed by the presence of creep cavitation. The tube in Example 2, however, has expended
only 72,000 hours of the estimated total 341,000 hours, which indicates only a moderate 21% life
exhaustion. The presence of extensive creep cavitation in this tube microstructure, however, is
an overriding indication that this tube has accumulated significantly more creep damage than
indicated by sigma phase measurements.

B-29
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

B.3.2 Creep Microvoid Characterization

Microvoid measurements provide a recognized microstructural-based method to characterize

creep damage in high temperature components. Surface replicas are sometimes used to obtain the
necessary microstructural information, however, this method may compromise accuracy because
of stress or compositional differences near the component surface that may have modified
microvoid formation. For this reason, microvoid measurements are best obtained from
metallographic samples removed from the component cross-section. Welds and heat-affected
zones should be included where applicable.

The as-installed microstructures of stainless steel components are in a metastable state that
usually evolve to a more stable state during elevated temperature plant operation. These
microstructural changes are expected to influence grain boundary cavitation. Grain boundary
sigma phase precipitates, for example, provide an initiation point for cavitation. Successful use
of grain boundary cavitation to characterize creep damage, therefore, is dependent on evaluation
of microstructures that have evolved to a similar state. Additionally, only service aged material
should be used to establish the correlation between grain boundary cavitation and the
accumulation of creep damage.

B.3.2.1 Microvoid Measurements

Microvoids have been characterized for life assessment purposes by their distribution, density,
size, and fraction of cavitated grain boundaries. The method described below characterizes creep
microvoids by their cumulative length, which provides a relatively simple and repeatable
measurement method.

Microvoids are measured in this method as the total void length (TVL) divided by the surface
area examined. Microvoid lengths are measured normal to the principal stress in microns per
square millimeter (µ/mm ) on a metallographically polished and etched surface. The principal
2

stress direction will be apparent in severely creep damaged materials by the preferential
orientation of the microvoids. If the microvoids do not exhibit a preferred orientation normal to
the principal stress and are reasonably represented as circular shapes, the TVL may be calculated
as the microvoid density (number per mm ) multiplied by their average diameter (µ). An
2

example of severely creep damaged material, shown in Figure B-24, illustrates microvoid
preferential orientation and the principal stress direction. Microvoid measurements are best
obtained at magnifications between 200X and 500X. The optimum magnification depends on the
number and size of the microvoids. Photomicrographs provided in this section are at a
magnification of 500X and include an area of 0.0277 mm2.

B.3.2.2 Grain Size Influence on Creep Microvoids

Since microvoids typically initiate at the interface between the austenitic matrix and grain
boundary second phase particles, the material’s grain size is expected to influence the number
and distribution of microvoids. A study involving three grades of austenitic stainless steel tubes
that failed by creep was performed to characterize the grain size influence. Both grain size and
TVL measurements were obtained adjacent to the rupture in each sample. Grain size was

B-30
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

determined as the average intercept distance (ASTM Method E112) and TVL was measured
normal to the principal stress.

Data acquired from the failed tubes are provided in Table B-1. Statistical analysis of these data
suggests that there is little or no material dependency and indicates a grain size dependency, in
which less cavitation (low TVL) is associated with large grain material.
Table B-1
Summary of Failed Tube Data

Material 3 2 3
TVL (10 µ/mm ) Aver. Intercept (µ) EVP / 10

304 10.3 41 26.1

304 12.4 29 28.8

304 6.9 50 18.3

304 13.7 36 33.6

321 10.0 81 30.0

321 10.0 27 22.8

347 8.2 62 23.0

The equivalent void parameter (EVP) incorporates the grain size influence. It is defined as the
th
TVL multiplied by the grain size (average intercept distance) raised to the n power as follows:
n
EVP = TVL x GS (B-4)

where:
EVP = Equivalent void parameter
= Total void length per area (µ/mm )
2
TVL
GS = Grain size, average intercept distance (µ)
n = Constant (0.25)

Analysis of TVL and EVP data indicate that the data scatter, as represented by the coefficient of
variance (COV), is reduced from 0.23 before the grain size adjustment to 0.20 after the
adjustment. Although the addition of grain size provides a moderate improvement, it suggests
that normal variations in grain size have only a minor effect relative to cavitation.

B-31
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

B.3.2.3 Example of Creep TVL and EVP Measurements

The sum of the microvoid lengths measured perpendicular to the principal stress (vertical height
of the 34 microvoids in Figure B-24 is 170 mm at a magnification of 500X. Applying the
appropriate factors (i.e., 1000 µ/mm / 500X / 0.0277mm2) gives:

TVL = 12.3 x 10 µ/mm

3 2

The average grain size intercept distance is measured from Figure B-24 as 29µ (i.e., 14.6 mm x
1000 µ/mm / 500X). From the TVL and grain side intercept distance, the EVP is calculated from
Equation B-4 as:
3 0.25 3
EVP = 12.3 x 10 x 29 = 28.5 x 10

Figure B-24
Preferential microvoid orientation normal to the principal stress. (MAG: 500X, Vilella’s
Etch)

B.3.3 EVP Based Creep Damage Estimates

Previous studies of ferritic tubular components have shown that the EVP in a component
progresses proportionally with creep strain. Microstructural based creep damage estimates for
austenitic stainless steel assume that this remains a valid relationship. Since EVP is proportional
to creep strain, then, normalized EVP is equal to normalized creep strain as:

EVP / EVPf = ε / εf (B-5)

The relationship between strain and time is obtained from uniaxial creep-rupture test data.
Normalized strain versus time curves developed from test data, from which the primary stage

B-32
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

creep strain was omitted, are shown in Figure B-25. For information, the average EVP from
Table B-1 is provided on the right vertical axis.

1.0
Limit Curves 25
0.9 316 30 KSI 1100 F 2200 Hr
316 25 KSI 1100 F 16500 Hr
0.8 304 30 KSI 1100 F 1270 Hr 20
0.7 304 30 KSI 1100 F 1780 Hr
304 30 KSI 1100 F 1780 Hr
0.6 304 30 KSI 1100 F 2574 Hr
15
304 30 KSI 1100 F 2750 Hr
0.5
0.4 10
0.3
0.2 5
0.1
0.0 0
0.0 0.2 0.4 0.6 0.8 1.0
Normalized Time (t/tf)

Figure B-25
Normalized strain versus exposure time data from austenitic stainless steel creep-rupture
tests.

The relationship between normalized time and creep strain is approximated from these data as:

t/tf = (1 – Exp(-A x ε/εf)) / (1 – Exp(-A)) (B-6)

where:

t = Exposure time
tf = Time at failure
ε = Creep strain
εf = Strain at failure
A = 3.2 +/- 1.8

Creep damage (CD) in this section is defined as the exposure time divided by the time-to-failure
under constant load and temperature conditions as:

B-33
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

CD = t/tf (B-7)

From the above discussion, the relationship between EVP and creep damage (CD) is expressed
as:

CD = (1 – Exp(-A x EVP/EVPf)) / (1 – Exp(-A)) (B-8)

where:

EVP = Equivalent void parameter

CD = Time based creep damage (1 = failure)
A = 3.2 +/- 1.8
3
EVPf = 26 +/- 7.7 x 10

This relationship involves two parameters, A and EVPf. Parameter A is the shape factor and
EVPf is the EVP at failure. Ranges for A and EVPf accommodate varying materials and exposure
conditions to provide a measure of the calculation uncertainty. More limited ranges of A and/or
EVPf may be established for specific cases.

Calculating the remaining creep life from the estimated CD is straightforward for components
that operate at constant stress and temperature (e.g., pipes and headers). The following
relationship is useful for these cases:

Remaining Life = Operating Hours x (1 / CD – 1) (B-9)

The limited growth of steamside oxide scale and minor wall thinning typically observed in
stainless steel tubes generally permit successful use of this method.

B.3.4 Recommended Reinspection

Tubes that are metallographically determined to have a measurable amount of creep damage
should be re-inspected at a future date to monitor the additional accumulation of damage.
Reinspection of tubes by more detailed methods and possibly accelerated creep-rupture testing is
recommended for tubes that have sustained significant creep damage. Creep damage should be
3
calculated by the method described above using A = 5.0 and EVPf = 18.3 x 10 to provide
conservative predictions on the pessimistic side of the uncertainty range. Table B-2 summarizes
the reinspection and additional testing recommendations.

B-34
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

Table B-2
EVP-Based Recommended Actions and Reinspection

3 Creep Damage Recommended Action Recommended

EVP / 10
(1 = Failure) Reinspection Interval

0 to 2.5 0 to 0.50 Reinspect, same method 0.33 x Oper Time

>2.5 to 5.0 >0.50 to 0.75 Reinspect, higher level 0.11 x Oper Time

>5.0 >0.75 Reinspect, higher level and test 0.02 x Oper Time

B.4 Case Study

An example of a microstructural-based creep damage estimate is provided below. The

microstructure from a service degraded Type 304 stainless steel tube was developed using the
etching procedure previously described. It included the following polish/etch sequence using
Vilella’s regent:

Polish Stage Etch Times

6µ diamond 4 minutes
1µ diamond 3 minutes
0.05µ alumina 2 minutes

Microvoids revealed by this procedure appear to have a slight aspect ratio with the long axis in
the vertical direction (Figure B-26). The TVL is measured, therefore, in the vertical direction. A
total of 21 microvoids in this photomicrograph have a cumulative measured length of 58
millimeters at 500X magnification. The TVL, therefore, is 4.2 x 10 µ/mm (58mm x 1000 µ/mm
3 2

/ 500X / 0.0277mm ). The average grain intercept distance is measured as 25µ (12.4mm at 500X
2

magnification, ASTM grain size of 7.5). Because of the difficulty in discerning individual grains
in some service degraded tubes, the interference contrast feature available on some
metallographs may be used to enhance contrast between grains.

B-35
13622634
Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

Figure B-26
Microvoids of service degraded Type 304H stainless steel tube showing the microvoid
measurement direction and a typical grain size measurement. (MAG: 500X)(Vilella’s Etch)

The EVP is calculated from the TVL and grain size using Equation B-4 as:
3 0.25 3
EVP = 4.2 x 10 x 25 = 9.4 x 10

The average estimated CD is calculated from the EVP using Equation B-8 as:

CD = (1 – Exp(-3.2 x 9.4 / 26)) / (1 – Exp(-3.2))

= 0.686 / 0.959

= 0.71

Inserting the stated ranges for A and EVPf yields a CD range of 0.43 to 0.93. Since failure is
predicted when CD = 1, the tube represented by this microstructure may be, but not necessarily,
very near end-of-life. This tube is a candidate for higher level inspection and accelerated creep
rupture testing because the upper range of creep damage exceeds 0.75 (Table B-2). Increased
inspection and tests will more accurately identify the level of creep damage, estimate its
remaining life, and assist in making run/repair/replace decisions.

B-36
13622634
 Metallurgical Evaluation to Assess the Condition of Austenitic Stainless Steel Tubing

B.5 List of Referenced Etchants

The various microstructural etchants listed in this Appendix are described below:

Etchant Composition Etching technique

Vilella’s 5 ml HCl Immerse or flow onto surface

1 g picric acid 2 to 4 minutes
100 ml ethanol

NaOH 50 g NaOH Electrolytic

100 ml H2O 1.5 volts, 10 - 20 seconds

KOH 56 g KOH Electrolytic

100 ml H2O 1.5 volts, 5 - 10 seconds

Oxalic 10 g oxalic acid Electrolytic

100 ml H2O 5 volts, 10 – 20 seconds

B.6 References
1
Bruemmer, S. M., “Quantitative Modeling of Sensitization Development in Austenitic Stainless
Steel”, Corrosion, September 1990, p. 698 - 709
2
The Making, Shaping, and Treating of Steel, United States Steel, 1964, pg. 1115
3
Minami, Y., H. Kimura, and Y. Ihara, Microstructural Changes in Austenitic Stainless Steels
During Longterm Aging”, Materials Science and Technology, Vol. 2, (August 1986), pp. 795-
806
4
Shinya, N., J. Kyono, H. Tanaka, M. Murata, and S. Yokoi, “Creep Rupture Properties and
Creep Fracture Mechanism Maps for Type 304 Stainless Steel, National Research Institute of
Metals, 19861.
5
Creep rupture Properties and Creep Fracture Mechanism Maps for Type 304 Stainless Steel,
Shinya, Kyono, Tanaka, Murata, and Yokoi, National Research Institute for Metals, Japan, 1986
6
Handbook of Stainless Steels, Peckner and Bernstein, McGraw Hill
7
“Microstructural Changes in Austenitic Stainless Steels During Long-Term Aging”, Minami,
Kimura, and Ihara, Materials Science and Technology, Vol.2 (August 1986), pp. 795-806

B-37
13622634
13622634
C
RUL METHODS AND MODELS

This appendix describes and demonstrates the use of two remaining life models that are relevant
to longterm overheat-creep failures. To some extent they can be adapted to other damage
mechanisms such as fireside corrosion or stress corrosion cracking.

The first remaining life model is a cumulative damage model that has enough flexibility to allow
the following:
• Predict the remaining life of the first and subsequent lead-the-fleet failures
• Sensitivity studies to determine the benefit of operational changes, repairs, minor or major
retrofits, or design alterations such as steam flow redistribution, alloy or tube dimensional
changes, etc.
• Provide a direct assessment of tube temperature or wall thinning rate variations

The second remaining life model is a rank statistical failure rate model that is relevant to the
situation where failures have already occurred and it is desired to estimate the rate of future
failures. A key assumption with this model is that the future service conditions are the same as
the past. This rank statistical model can be adapted with moderate variations for any of the
austenitic stainless steel damage mechanisms but is presented in this appendix for longterm
overheat-creep failures.

C.1 Roadmap

The steps used to make remaining life assessments often must be adjusted to compensate and use
the best available information available. The reader is encouraged to be creative and to use the
methods described in this report in combinations that are not necessarily identical to the roadmap
approaches described below. As a starting point, the following roadmap (Figure C-1) can be used
to obtain life estimations for the situation where no LTOC failures have occurred, or where some
lead the fleet failures have already occurred. Detailed descriptions of each of the roadmap steps
are included in this appendix. Case histories are used to illustrate the analytical steps and use of
the key equations.

C-1
13622634
RUL Methods and Models

Figure C-1
Life assessment roadmap for austenitic stainless steel tubing subject to longterm
overheat – creep.

C-2
13622634
 RUL Methods and Models

C.2 Step I: Review design and historical information

The first step in any life assessment of superheater or reheater tubing is a thorough review of the
design and the operational, maintenance, and capital improvements history. ·Information that
may be useful to review includes:
• Design drawings for the unit showing the general arrangement of the tube bundles, division
walls, burners, headers, sootblowers.
• General attributes of the unit including maximum continuous rating (MCR), typical operating
mode (load following, peaking, base loaded, etc.), service hours, number of start/stop cycles,
and fuel characteristics
• Detailed dimensional and material drawings of the superheater/ reheater. This should include
the lengths, outside and inside diameters and alloys of all the tubing, the lengths and
diameters of all the segments of each circuit including any orifices. It should also include the
radii of all bends. Any field changes (such as replaced sections of tubing with diameters that
were different than the design dimensions) need to be thoroughly documented. Field
measurements of the tube dimensions, rather than design-basis dimensions are preferable.
• If detailed gas side flow modeling is to be performed then the tube layout/spacing must also
be obtained
• At least six months (preferably a few years or more) of hourly records of the following
measured values: generator load (MW), total fuel burned, furnace air flow rate, SH/RH steam
bulk mass or volumetric flow rate, SH/RH inlet and outlet pressure, SH/RH inlet and outlet
bulk steam and penthouse tube temperatures
• All available wall and steamside oxide scale thickness measurements made on ferritic tubes
within the furnace and on the outlet penthouse tubes
• Wall thickness of the stainless tubing at locations where wall loss has been most severe
• Results of all metallurgical and accelerated stress rupture tests performed on selected ferritic
and austenitic tubing
• Metallurgical and root cause investigation reports of all boiler tube failures.

C.3 Remaining Life Model #1 - Rank Statistical Model

C.3.1 Step II: Calibrated, rank statistical failure forecast (when 1 to 3 failures have
occurred)

To predict the failure rate when one or a few longterm overheat – creep failures have already
occurred can be done by using the calculations described in Step IV(a) through Step IV(g). This
is done by “forcing” or “calibrating” the model to adequately predict the lead-the-fleet failures,
then evaluating the life of the tube circuits/segments that are predicted to fail in the future. The
model can usually be calibrated sufficiently by adjusting the material properties (choosing a
failure percentile that leads to an accurate prediction of the lead-the-fleet failures).

C-3
13622634
RUL Methods and Models

It is also possible to forecast the failure rate using rank statistical models. If only one to three
failures have occurred then these forecast should be combined with information about expected
rates of failures for the active damage mechanism. In the case of longterm overheating of
austenitic tubing the failure rate of laboratory stress rupture tests will provide an upper bound
estimate of the rate of field failures. Keep in mind that this same approach will in general not
lead to conservative failure rate estimations for ferritic superheater and reheater tubes. In ferritic
tubes the steamside oxide scale growth is often sufficiently rapid to produce a significant and
ever-increasing tube metal temperature throughout the service life. This systematic, ever
increasing tube metal temperature will lead to more rapid tube failure rates than laboratory tests
performed with constant metal temperatures. For austenitic stainless steel superheater and
reheater tubes the steamside oxides are usually very thin even after prolonged service exposure;
thus the metal temperature variations are more random and load dependent and will not
increasing significantly throughout the lifetime of the tubing. In the case of a group of austenitic
alloy tubes the life influencing factors (i.e., temperature, and stress) will be more widely
distributed than in well controlled laboratory experiments. This being the case, the rate of
failures of laboratory stress rupture tests will be more rapid than a population of superheater or
reheater tubes.

The steps in determining the failure rate when just a few tubes have failed are listed below.

Step 1: Evaluate the failed tube(s) and sample a few of the adjacent austenitic stainless steel
tubes to confirm that the life limiting damage mechanism is longterm overheat - creep, and to
verify the dimensions, microstructural characteristics, and tube alloy

Step 2: Using the time to failure of the first failure and the tube population size as input, estimate
when 50% of the population of tubes that are similar to the failed tube would be expected to fail
(using the lognormal distribution function of time to rupture to median time to rupture ratio for
the tube alloy of concern)

Step 3: Estimate when the lead-the-fleet failures in this population are likely to occur (using the
lognormal distribution function of time to rupture to median time to rupture ratio for the tube
alloy of concern). These steps are demonstrated in the following example.

For a Type 304H superheater with 52 assemblies and 12 rows (624 tube circuits) that has the first
th
LTOC failure in 200,000 hours of service the estimated 50 percentile time to rupture (the
estimated service hours when 312 of the 624 circuits would have experienced LTOC failures) is
estimated using the distribution of failure times of laboratory stress rupture tests given in
Appendix A.

trup50th = Time to the first Failure / LOGINV(1/624,4.0781E-8, 0.657969) = 1,390,723 hrs

The estimated time for the second failure is then:

trup2nd of 624 = (1,390,723 hrs) LOGINV(2/624,4.0781E-8, 0.657969) = 231,354 hrs

C-4
13622634
 RUL Methods and Models

th
The estimated time for the n failure is:

trupnth of 624 = (1,390,723 hrs) LOGINV(n/624,4.0781E-8, 0.657969)

The estimated time for the first 10 failures for this example is shown in Figure C-2.

Figure C-2
Predicted cumulative failure rate for a 624 circuit group of Type 304H tubes with the first
failure occurring in 200,000 hours.

Note that this failure rate estimation requires no estimation of the stress or temperature of the
tubes. The basic premise of this approach is that all of the tubes in the population will have
similar stress and temperature histories. In reality the first tube failure probably was exposed to
significantly higher stress and or temperatures than its neighbors. If this is the case and the
overall operating environment doesn’t become more severe then the actual rate of failures will be
slower than predicted by this approach. As more failures occur this conservatism can be removed
from the model by making small adjustments to the standard deviation coefficient in the
lognormal distribution such that error in the predictions for all of the LTOC failures in this group
of tubes is minimized.

C.3.2 Step III: Rank statistical failure forecast (when more than a few failures
have occurred)

If more than a few LTOC failures have already occurred within a group of similar tubes then the
rate of additional failures can be estimated directly using a rank statistical model such as the
lognormal distribution presented in Step III. The rank statistical failure rate estimation can also
be made using commercially available computer programs or spreadsheet add-ins such as

C-5
13622634
RUL Methods and Models

WeibullsmithTM 1 or @Risk2. These will provide both best-estimate failure forecasts and tolerance
bounds to the estimates.

C.3.3 Case Study: To predict the failure rate when one or more longterm
overheat-creep failures have already occurred

Objective:

To forecast how many additional LTOC failures will occur in a finishing superheater during each
of the next 10 years (i.e. during the service periods from 202,000 hrs to 210,000 hrs, 210,000 hrs
to 218,000 hrs, etc.).

Background:

The finishing superheater has 10 rows and 50 elements (500 tubes) fabricated from SA213
TP321H tubing. Three LTO-C failures have occurred. The service hours to failure were 162,000
hrs, 184,000 hrs, and 202,000 hrs. The failures occurred in three different tube circuits but were
all located within 20 ft (gas-touched length) of each other. Metallurgical analysis confirmed that
the failure mechanism was LTOC.

Failure Rate Forecast:

The first of 500 tubes failed in 162,000 hours. The predicted median time to failure (when 250 of
500 tubes would have failed) is estimated using the lognormal distribution of stress rupture
failure times for Type 321H material.

From Appendix A, the relationship between fraction of stress rupture samples failed and the ratio
of the time to rupture and the median life is given as:
Fraction failed (Type 321H stress rupture tests) = LOGNORMAL(time to rupture/median time to rupture, 3.64E-08, 0.65107)

The inverse lognormal function in an EXCELTM spreadsheet can be used to estimate the median
time to rupture if the time to first failure and the tube population are known:
Median time to rupture = Time to the first failure/ LOGINV (1/Population size, 3.64E-08, 0.65107)

= 162,000 hrs/LOGINV(1/500, 3.64E-08, 0.65107) = 1,055,179 hrs

The predicted time for the nth failure is estimated by rearranging the same equation and solving
for time to nth failure:
Time to nth failure = Median time to rupture [LOGINV(n / Population size, 3.64E-08, 0.65107)]

The predicted times for the first ten failures are given in column (B) of the table below. A minor
improvement in the forecast could be made by altering the second coefficient in the failure rate
relationship to minimize the errors in the predictions of the three failures that have already
occurred. The adjusted model produced the following relationships:

C-6
13622634
 RUL Methods and Models

Median time to rupture = Time to the first failure/ LOGINV (1/Population size, 3.64E-08, 0.658308)

= 162,000 hrs/LOGINV(1/500, 3.64E-08, 0.658308) = 1,077,401 hrs

Time to nth failure = 1,077,401hrs [LOGINV(n / Population size, 3.64E-08, 0.658308)]

The predicted times for the first ten failures are given in column (C) of the table below. Column
(D) presents the results of both the direct and “calibrated” model in terms of operating years
(based on 8,000 operating hours per year).
(A) Number of (B) Predicted time to (C) Predicted time to (D) Actual time to
failures failure failure (calibrated failure
model)

1 162,000 162,000 162,000

2 187,688 184,118 184,000

3 205,593 201,887 202,000

4 219,883 216,080 Expected in 2 to 3 yrs

5 232,029 228,153 Expected in 4 yrs

6 242,723 238,789 Expected in 5 yrs

7 252,367 248,384 Expected in 6 to 7 yrs

8 261,207 257,182 Expected in 7 to 8 yrs

9 269,407 265,347 Expected in 8 to 9 yrs

10 277,089 272,998 Expected in 9 to 10 yrs

C.4 Remaining Life Model #2 – Cumulative Damage Model

C4.1 Step IV(a): Perform ultrasonic wall /steamside oxide thickness surveys and
selected tube sample metallurgical evaluations

The second step in the roadmap is the measurement of wall and steamside oxide scale thickness
of the ferritic (typically 2-1/4Cr-Mo) tubing in the outlet penthouse and located just upstream of
the austenitic tubing. The objective of these measurements is to make an estimate of the tube
bulk steam temperature up and downstream of the austenitic tubing for every tube circuit. These
steamside oxide scale inferred bulk steam temperature estimates are used to estimate the local
heat absorption rate (Step IV(b)), steam temperature (Step IV(c)), and radial temperature
gradient within the austenitic sections of tubing (Step IV(e)).

A penthouse oxide scale thickness survey including many hundreds of tubes can often be
completed in one or two days. Ideally the wall and steamside oxide thickness on every penthouse
tube should be made since this will provide the most accurate, comprehensive estimations of

C-7
13622634
RUL Methods and Models

remaining life of the austenitic tubing. Figure C-3 shows the results of a typical penthouse oxide
scale thickness survey performed on a four row by 55 assembly (220-circuit) secondary
superheater.

Figure C-3
Results of a nondestructive, ultrasonic steamside oxide survey, made on the 2-1/4Cr 1Mo
outlet header stub tubes downstream of the austenitic stainless steel tubing.

Figure C-4 shows the distribution of estimated tube bulk steam temperatures inferred from the
measured oxide scale thickness values. These estimates were made using the EPRI TUBELIFE
III computer code. The systematic error in these temperature estimates was corrected by
adjusting the “unit variable” in TUBELIFE III to a value that resulted in the average of all the
estimated tube temperatures being equivalent to the measured bulk outlet steam temperature. If
the unit load-follows or low load cycles, then a review the time history of the bulk steam
temperature should be made to determine the time averaged bulk outlet steam temperature.
Skipping this set may lead to a somewhat conservative estimation of temperatures and life.

C-8
13622634
 RUL Methods and Models

Figure C-4
Estimates of the outlet steam temperatures based on the steamside scale thickness
survey shown in Figure C-2. These temperature estimations were derived using the EPRI
TUBELIFE III computer code. The “unit variable” was adjusted until the average of the
predicted outlet temperatures was equivalent to the bulk outlet temperature of 1050oF.

It is important to verify the accuracy of the steamside oxide scale readings by sampling a few
tubes. The most useful approach for this measurement accuracy verification/calibration is to
sample a few sections of penthouse tubing that are easily repaired and had the “thick”
ultrasonically measured steamside oxide scales.

Sample a few of the austenitic stainless steel tubes in the hottest tube circuits to verify the
dimensions, microstructural characteristics, equivalent void parameter EVP (see Appendix B),
material properties, and to confirm that the life limiting damage mechanism is longterm
overheat-creep. These tube samples may also be used for accelerated creep or stress rupture
testing to calibrate or fine-tune the remaining life estimations.

C-9
13622634
RUL Methods and Models

C.4.2 Step IV(b): Estimate the average and peak heat absorption rate

Once the outlet steam temperature of each tube circuit is determined (Step IV(a)) then this
information can be used to estimate the effective average heat absorption rate, q’’OD, of the tube
bundle. The effective average heat absorption rate can be estimated by balancing the heat
absorption energy with the increase in steam energy as shown in Figure C-5.

Figure C-5
The relationship between the heat energy added to the outside surface of the tubing
balanced with the change in energy of the steam can be used to estimate the average heat
absorption rate of each individual tube circuit.

Note that this relationship assumes the heat absorption occurs evenly around the entire perimeter,
π Do of the gas-touched length, LGT, of tubing. In reality there will be a characteristic variation in
heat absorption rate around the periphery of the tubing and the peak value will be somewhat
higher than the value estimated with the global energy balance relationships described above.
The peak heat absorption rate will typically occur at the 2 o’clock and 10 o’clock positions of the
tube with respect to the primary direction of flue gas flow. Since the location with the peak heat
absorption rate will also be the location with the highest metal temperature and lowest life it is
useful to develop a meaningful adjustment factor between the average and peak heat absorption
rate. This can be derived from modeling or published empirical relationships but probably the
best method is to measure the steamside oxide scale thickness around the periphery of a few of
the ferritic tubes just upstream of the stainless steel tubing and then “back out” the heat
absorption rate distribution from the distribution in steamside oxide temperature inferred from
the oxide thickness readings and common heat transfer relationships. It is recommended that at
least 6 evenly spaced readings of the steamside oxide thickness be made around the perimeter of
a tube with one of the readings being the maximum detected scale thickness. Figure C-6 presents
an example of this approach for a reheater tube.

C-10
13622634
 RUL Methods and Models

Figure C-6
Steamside oxide scale thickness measurements made around the periphery of a selected
ferritic tube can be used to derive an estimate of the ratio of the peak to average heat
absorption rate at a given gas-touched tube location. As shown, the peak to average heat
absorption rate ratio is a function of the steamside heat transfer coefficient. The ratio of
peak to average heat absorption rate for the case shown was 1.13, 1.24 and 1.35 for
steamside heat transfer coefficients of 1700 W/m2°C, 3400 W/m2°C and 5100 W/m2°C,
respectively.

Further refinements in estimating the heat absorption rate distribution across and through the
tube bundle and as a function of gas touched length can be made from steamside oxide scale
measurements of the ferritic tubes within the heat absorbing sections of the tube circuits. Direct
in-situ measurements can also be made using heat absorption rate measuring instruments such as
“flux domes”.

C.4.3 Step IV(c): Estimate the bulk steam temperature v. gas-touched length

Since it is not possible with today’s technologies to make a nondestructive measurement of some
feature of the austenitic tubing that can be directly correlated to metal temperature or remaining
life we must use the measurements of steamside oxide made up and downstream of the austenitic
stainless steel in combination with analytical modeling to estimate the metal temperature of
specific tube segments. From Steps IV(a) and IV (b) we can obtain the outlet steam temperature
and the average and peak heat absorption rate for each tube circuit. The next step is to make
estimates of the bulk steam temperature along the entire length of each (or a selected group of
tube circuits with the greatest wall loss or highest estimated outlet temperatures. Estimates of the

C-11
13622634
RUL Methods and Models

bulk steam temperature versus gas touched length can be made in a variety of ways. For example
a linear interpolation between the inlet temperature (obtained from thermocouple readings) and
the outlet temperature (obtained from steamside oxide thickness measurements) can be made.
Slight engineering adjustments to these linear extrapolations can be made for each pass of tubing
to reflect the variation in heat absorption throughout the tube bundle. On the other end of the
spectrum, full steam and gas side computation fluid dynamic models of the tube bundle can be
made. Another approach is to use the energy balance equation presented in Step IV(b) in
combination with some engineering judgment regarding the distribution of heat absorption rate.
For example if the heat absorption rate near the outlet of the tube bundle is known then the heat
balance equations can be used to estimate the bulk steam temperature in regions just upstream of
the header. An example of this approach is shown in Figure C-7. In this case the steam
temperature 10.4 meters (34 feet) from the position where the tube exits the furnace is estimated
based on known tube mass flow rate, and outlet pressure, temperature, and enthalpy.

Figure C-7
The heat/steam energy balance relationships can be used to estimate the bulk steam
temperature as a function of gas-touched length

Some additional accuracy in the bulk steam temperature estimates is possible if the slight
pressure difference between the outlet and the upstream positions are estimated. In addition the
mass flow rate distribution between all the individual tube circuits should be estimated. These
are considered in Step IV(d).

C.4.4 Step IV(d): Estimate the steam velocity v. gas-touched length

Since the outlet steam temperatures are determined in a region with no heat absorption this is not
the correct value to use at the steam/steamside oxide interface in the heat absorbing (gas-
touched) sections of austenitic stainless steel. In the gas-touched sections of tubing there will be
a temperature gradient between the bulk steam temperature and the inside surface of the
steamside oxide that is controlled by radial convective heat transfer. This temperature gradient is

C-12
13622634
 RUL Methods and Models

a function of the local peak heat absorption rate, the tube inside and outside dimensions, the
surface roughness/friction factor, ƒ, of the steamside oxide, physical properties of the steam (i.e.,
thermal conductivity, specific heat, viscosity), and the local steam velocity.

The surface roughness for new or service degraded tubes carrying superheater steam is
1
anticipated to be in the range of 0.020mm to 0.10 mm (0.0008” to 0.004”) with a typical value
near 0.025 mm (0.001”). An indirect estimate of the surface roughness can be estimated from the
measured pressure drop from inlet to outlet header. The steam properties can be estimated using
the outlet steam pressure/temperature and steam tables. The local steam velocity, V, can be
estimated by dividing the bulk mass flow rate by the number of tube circuits and then
multiplying by the steam density and the inside surface area:

V = Q/Ai = mcircuit/{(ρ)(Ai)}

Where:

V = local bulk velocity, m/sec

Q = volumetric flow rate, m3/sec

mcircuit = average mass flow rate for a specific tube circuit, kg/sec

= total mass flow rate/number of tube circuits

ρ = steam density, kg/m3

Ai = cross-sectional area of the inside of the tube segment = π Di2/4

Di = tube inside diameter, m2

The actual mass flow rate and velocity in each tube circuit will be a function of its length, and
internal dimensions compared to the other circuits within the bundle. Tube circuits with the
shortest length or largest inside diameters will have the lowest resistance to flow and will thus
carry proportionally more steam than other circuits. An estimation of the mass flow rate in each
circuit can be estimated using handbook solutions or computational fluid dynamics modeling, or
from in-situ measurements. The distribution of mass flow rate per circuit is typically within ±
15% or less of the average mass flow rate per circuit, mcircuit. In unusual/unbalanced designs the
range of mass flow within individual tube circuits can be as much as ± 30% of the average mass
flow rate. Recall from Step IV (b) that for a constant heat absorption rate, the tube circuits with
the highest mass flow rate will have the lowest increase in bulk steam temperature. As will be
discussed in Step IV(e) the tube circuits with the highest mass flow rate will have the smallest
temperature drop (for a given heat absorption rate) between the bulk steam and the inside
diameter of the tube.

The local velocities or volumetric flow rates in each tube circuit and segment can be estimated
using handbook equations and an iterative solving routine to account for the changes in steam
properties along the length of the circuit. The basic approach requires field measurements of the

C-13
13622634
RUL Methods and Models

mass, kg/sec (or volumetric m3/sec) flow rate through the inlet or outlet header and the pressure
drop from the inlet to outlet header. Next the dimensions (inside diameter, length, bend radii) of
each segment of each circuit is determined. The basic equations relating flow rate with tube
geometry and pressure drop are shown in Figure C-8. For a given steam flow rate and surface
roughness the pressure drop is proportional to the tube length divided by the inside diameter to
th 5
the 5 power, L/Di . The pressure drop through a tube circuit comprised of multiple segments and
with various bends can be determined by determining the incremental pressure drop through
each tube segment and fitting. For tubes circuits in parallel the distribution of flow rates must
first be determined then the pressure drop through each circuit can be determined. Alternatively,
if the pressure drop is known (from in-situ measurements) then the flow rate through each tube
circuit can be calculated.

Figure C-8
The Darcy-Weisbach3 equations define the relationships between surface roughness, flow
rate and pressure drop with tube circuits.

The local loss coefficients, KLlocal, relevant to superheater or reheater tube bundles are shown
graphically in Figure C-9 and in Table C-1. The factors are added to the frictional loss
coefficients to account for pressure drops associated with the contraction of steam at the inlet
header, contractions or expansions of steam at tube inside diameter changes, directional changes
at bends, and expansion at the outlet header.

C-14
13622634
 RUL Methods and Models

Table C-1
Local Loss Coefficients

Situation Configuration Range of Loss Typical Value

Coefficient Used
Entrance Entrance with sharp corner 0.5 0.5
Entrance – well rounded 0.04 to 0.25 0.04
Contraction d/D = 0.75 0.18 to 0.20 0.20
d/D = 0.5 0.30 to 0.33 0.33
d/D = 0.25 0.40 to 0.43 0.43
o
Bend 180 Return bend 0.40 to 2.2 2.2
Expansion d/D = 0.75 0.18 to 0.20 0.20
d/D = 0.5 0.55 to 0.60 0.60
d/D = 0.25 0.88 to 0.92 0.90
Exit Sharp or well rounded 1.0 1.0
Adapted from Mays, L.W., Hydraulic Design Handbook, McGraw Hill, Table 22.3, 19994

Figure C-9
Local loss factors, KLlocal, are used to estimate the pressure drop within tube circuits that
are associated with steam contraction, steam expansion, directional changes, etc. They
are added to the frictional loss factors KLf = (ƒ)(L)/Di, associate with the pressure drop in
straight constant inside diameter segments of the tube circuits.5

C-15
13622634
RUL Methods and Models

The pressure drop and flow rate distributions throughout a tube bundle can be modeled using a
series-parallel circuit model. The series-parallel pressure drop/flow rate relationships are
presented in Figure C-10. Somewhat more complex modeling steps are required if the tube
circuits contain bifurcations. The overall objective of this step is to determine the local steam
velocity and pressure as a function of distance from the inlet or outlet headers. The steam
velocity will be used to estimate the temperature rise between the bulk steam temperature and the
inside surface of the tube. The local pressure will be used to estimate the hoop stress at selected
locations.

Figure C-10
For superheater or reheater tube designs with no tube bifurcations, the pressure drops
and steam velocities can be estimated using simple series and parallel circuit analogies.

C.4.5 Step IV(e): Estimate the tube metal temperature at selected, critical
locations

The mid- or mean diameter tube metal temperature will provide conservative estimations of the
creep rupture life of austenitic stainless steel tubing. The midwall temperature of a specific, gas-
touched section of superheater tube can be estimated from the local bulk steam temperature, the
local peak heat absorption rate, and local steam properties and velocity, in combination with a
convective and conductive radial heat transfer model. The first step is to estimate the convective
heat transfer between the inside surface of the tube and the bulk steam. An example calculation
of this convective, radial heat transfer at a specific gas-touched location within a tube circuit is
shown in Figure C-11.

C-16
13622634
 RUL Methods and Models

Figure C-11
Example of the estimation of the temperature rise between the bulk steam temperature and
the inside surface of the steamside oxide scale in a segment of heat absorbing tubing.

Where:
2
Dvsteam = dynamic viscosity of the steam, kg/m-sec (N-s/m )
Cpsteam = specific heat of the steam, kJ/kg-K (kW-sec/kg-K)
ksteam = thermal conductivity of the steam, kW/m-K
ρsteam = steam density, kg/m3
m = average mass flow rate for a specific tube circuit, kg/sec
f = friction factor from Moody chart
f ≅ 0.25/{log[e/[(3.7)(Di)]+5.74/Re0.9]}2 where e = surface roughness, m
Ai = cross-sectional area of the inside of the tube segment = π Di2/4
Di = inside diameter, m2
Pr = Prandtl number
Nu = Nusselt number
2 o
hi = convective heat transfer coefficient, W/m - C
TID = Temperature at the steam/steamside oxide interface, oC

C-17
13622634
RUL Methods and Models

Similar conductive heat transfer relationships can be used to estimate the temperature gradient
across each of the oxide and tube metal interfaces. The steamside oxide scale thickness should be
set to a value that is equivalent to its anticipated value at the end of the “evaluation period”. For
example if the current thickness of the oxide (determined from laboratory evaluation of the
hottest anticipated segment of stainless steel tubing) is 0.05 mm after 100,000 hours of service
and you want to determine if LTOC failures could occur prior to 200,000 hours of service then a
0.10 mm scale should be used in the heat transfer analysis. For ferritic tubing this very
conservative assumption would produce overly conservative life estimation, but for austenitic
tubing the steamside oxide thickness are generally very thin and this conservative approach to
the estimation of future scale thickness will not introduce a significant amount of conservatism in
the life estimation. If steamside oxide thickness values much greater than 0.10 mm are present
then a more realistic projection of the oxide scale thickness at the end of the evaluation period
should be made. These oxide scale thickness projections can be made using the EPRI TubeLife
III code algorithms.

C.4.6 Step IV(f): Estimate the tube stress at selected, critical locations

Using the estimated steam pressure and the current tube dimensions, the mean diameter hoop
stress of each tube segment can be determined.

σMDF = P (Do-t)/2t

Where:

σMDF = mean diameter formula hoop stress

P = local steam pressure
Do = outside tube diameter
t = minimum anticipated wall thickness at the end of the evaluation period

For a tube with a 0.045m (45mm) outside diameter and 0.0054m (5.4mm) wall thickness and a
14.98 MPa steam pressure the hoop stress is:

σMDF = 14.98 MPa (0.045m -0.0054m)/[2(0.0054m)] = 54.9 MPa

In austenitic stainless steel tubes residual stresses may also be present. If longterm overheat
failures occur well in advance of the estimated life determined based on the assumption that only
pressure-induced stresses are active then the presence of residual stress should be considered.
The residual stresses may be introduced during tube manufacturing or may be induced during
service as a result of severe thermal transients such as water spraying of hot tubes or forward
flow of cold condensate into hot tubing. Due to the excellent creep strength of austenitic stainless
steel residual stresses may be unable to relax away during service exposure. Residual stress
measurements can be made using a variety of techniques such as sectioning or hole drilling with
strain gages, x-ray diffraction, tube splitting with incremental measurement of the width of the

C-18
13622634
 RUL Methods and Models

cut as the cut progresses across the tube wall, etc. If the magnitude of the measured residual
stress is significant then it should be added to the applied, mean diameter formula stress.

C.4.7 Step IV(g): Estimate the remaining longterm overheat-creep life at selected,
critical locations

Using the results of Steps IV(a) through IV(g) as input the remaining life of each segment of
austenitic stainless steel of each tube circuit can be made using the stress rupture formulas in
Appendix A. For example if the tube is a type 304H alloy, the estimated hoop stress is 54.9 MPa,
and the estimated midwall temperature is 634oC then the median and 5th percentile life
estimations are:
o o
trup50th = 10^(-11.389+18746.602/(634 C+273.15)-268.416*LOG(54.9 MPa)/( 634 C
2 o
+273.15)-969.675*[LOG(54.9 MPa)] /( 634 C+273.15))) = 336 khrs

trup5th=(trup50th)[LOGINV(5/100, 4.0781E-8, 0.657969)]=336 khrs (0.339) = 114 khrs

If the analyses outlined in Steps IV(a) through IV(g) are linked together in a spreadsheet or
similar computer code then it is possible to estimate the life of all the austenitic stainless steel
segments of each tube circuit. These life estimations can then be combined to determine the
likely failure rate of the austenitic tubes during future operating years. This information can be
used as input into a cost model to evaluate the need and benefit of the various corrective actions
discussed in Section 11, Figure 11-8.

C.4.8 Case Study: To predict the life when no failures have occurred or when it is
desired to include temperature, wall thinning, or other tube-to-tube variations into
the life assessment

Objective:

To forecast the LTOC based remaining life of each segment of austenitic tubing in the ten
assembly by three row (30-circuit) superheater shown in Figure C-12. The superheater design
used in this case study is fictitious (to simplify the presentation of the analysis steps), but was
developed based on realistic distributions of steamside oxide scale thickness values and realistic
tube dimensions.

C-19
13622634
RUL Methods and Models

Figure C-12
Configuration of the Ten Assembly, Three Row superheater (A). The tube configuration at
the inlet and outlet headers (B) is identical.

Background:

The example superheater was fabricated from SA213 T22 (2-1/4Cr-1Mo) and TP304H tubing.
No LTOC failures have occurred. The unit has accumulated 92,000 hrs mostly under base load
o o o
operating conditions. The bulk inlet and outlet temperatures were 475 C (887 F) and 565 C
o
(1049 F) respectively. The inlet and outlet steam pressures were 15 MPa (2175 psia) and 14.38
MPa (2085 psia) respectively. The total mass flow rate of the superheater was 35 kg/sec.

The tube circuit segment dimensions are shown in Table C-2. All ten assemblies have identical
tube dimensions. The first and last segments in each row correspond to a small step increase in
inside diameter between the header stub tubes and the header boreholes.

C-20
13622634
 RUL Methods and Models

Table C-2
Dimensions of the tube segments for each of the 3 rows

Assembly Row Segment Tube Material Length, m Inside Diameter, m Outside Diameter, m
1 1 A T22 0.03 0.036 NA
B T22 0.3 0.034 0.060
C T22 0.2 0.034 0.060
D T22 30 0.034 0.060
E T22 0.25 0.034 0.060
F T22 6 0.034 0.060
G 304H 0.25 0.034 0.045
H 304H 30 0.034 0.045
I 304H 0.2 0.034 0.045
J T22 0.3 0.034 0.080
K T22 0.03 0.036 NA

1 2 A T22 0.03 0.036 NA

B T22 0.3 0.034 0.060
C T22 29 0.034 0.060
D T22 0.25 0.034 0.060
E T22 5 0.034 0.060
F 304H 0.25 0.034 0.045
G 304H 29 0.034 0.045
H T22 0.3 0.034 0.080
I T22 0.03 0.036 NA

1 3 A T22 0.03 0.036 NA

B T22 0.3 0.034 0.060
C T22 0.2 0.034 0.060
D T22 28 0.034 0.060
E T22 0.25 0.034 0.060
F T22 4 0.034 0.060
G 304H 0.25 0.034 0.045
H 304H 28 0.034 0.045
I 304H 0.2 0.034 0.045
J T22 0.3 0.034 0.080
K T22 0.03 0.036 NA

C-21
13622634
RUL Methods and Models

Evaluation and Life Assessment:

The evaluation will follow Step IV(a) through IV(g) of Figure C-1.

Step IV(a): The objective of this step is to estimate the outlet steam temperatures of every tube
circuit and to verify that LTOC is the only active damage mechanism on the austenitic segments
of tubing.

Steamside oxide scale measurements were made on each of the 30 outlet header T22 stub tubes.
Tube samples of a few sections of stub tubes confirmed that the oxide scale thickness values
were accurate and that no significant steamside oxide scale exfoliation had occurred. The outlet
steam temperatures for each tube circuit were estimated from the steamside oxide scale thickness
readings using the oxidation rate algorithms in EPRI TubelifeIII. The relationship between oxide
scale thickness and temperature was adjusted with the “unit variable” such that the average of all
the estimated steam temperatures entering the outlet header was equal to the 565oC bulk outlet
temperature. The steamside oxide thickness readings and temperature estimates are presented in
Tables C-3(a) and C-3(b). The distribution of estimated temperature is also presented in Figure
C-13. Row 3 of Assembly 4 exhibited the thickest measured steamside oxide scale. The
o o
estimated outlet steam temperature for this tube circuit was 595 C (1103 F).

Figure C-13
The penthouse header stub tube steamside oxide thickness measurements are used to
estimate the distribution of outlet steam temperature for every tube circuit.

C-22
13622634
 RUL Methods and Models

Table C-3(a)
Measured outlet stub tube steamside oxide scale thickness values and estimated steam
temperatures

Oxide Thickness, mm Temperatures, oC

Assembly Row 1 Row 2 Row 3 Row 1 Row 2 Row 3

1 0.28 0.31 0.31 551 559 559

2 0.34 0.30 0.51 567 555 592

3 0.34 0.48 0.52 567 589 593

4 0.28 0.37 0.54 551 573 595

5 0.28 0.24 0.28 551 537 551

6 0.27 0.27 0.30 547 547 555
7 0.30 0.27 0.30 555 547 555

8 0.51 0.51 0.49 592 592 590

9 0.52 0.51 0.37 593 592 573

10 0.25 0.28 0.27 541 551 547

Table C-3(b)
Measured outlet stub tube steamside oxide scale thickness values and estimated steam
temperatures

Oxide Thickness, mils Temperatures, oF

Assembly Row 1 Row 2 Row 3 Row 1 Row 2 Row 3

1 11.2 12.3 12.3 1024 1038 1038

2 13.5 11.7 20.0 1052 1031 1097

3 13.5 18.8 20.6 1052 1092 1100

4 11.2 14.7 21.2 1024 1064 1103

5 11.2 9.4 11.2 1024 998 1024

6 10.6 10.6 11.7 1016 1016 1031

7 11.7 10.6 11.7 1031 1016 1031

8 20.0 20.0 19.2 1097 1097 1094

9 20.6 20.0 14.6 1100 1097 1063

10 9.9 11.1 10.6 1006 1023 1016

C-23
13622634
RUL Methods and Models

Metallurgical analysis of selected sections (including a section of, Assembly 4, Row 3, Segment
I) of the austenitic tube of the hottest tubes confirmed that no failure mechanism other than
LTOC was active in the tubing and that the lower bends had been solution annealed and had no
measurable residual stress. The steamside oxide on the stainless steel tubing had a maximum
total thickness of approximately 0.0254mm (0.001”). For the heat transfer analyses performed in
Step II(e), the steamside oxide was assumed to be four times thicker than the current thickness.

Step IV(b): The objective of this step is to estimate the average heat absorption rate of each tube
circuit based on their outlet steam temperature, the bulk inlet temperature, and the bulk inlet and
outlet temperatures

This analysis step is completed for every tube circuit. The calculations below show an example
of the estimation of the average heat absorption rate for the hottest tube circuit (Assembly 4,
o o
Row 3). The outlet steam temperature of this tube circuit was estimated to be 595 C (1103 F).
The inlet and outlet steam pressures were 15 MPa (2175 psia) and 14.38 (2085 psia)
o o
respectively. The bulk inlet temperature was 475 C (887 F). It is generally assumed that the inlet
temperature is uniform for all tube circuits. For unusual inlet header/piping designs this
assumption should be verified with thermocouple measurements of selected inlet header stub
tubes. The total mass flow rate of the superheater was 35 kg/sec. The corresponding average
mass flow rate per tube circuit is 1.17 kg/sec. The outside surface areas, (π)(Do)(L), of the heat
absorbing segments of tubing are shown in Table C-4.
Table C-4
Summary of the outside surface areas of the heat absorbing segments of Assembly 4,
Row 3

Assembly Row Segment Mtrl. L, m Do, m Gas-touch Surface Area, m2

4 3 A T22 0.03 NA Outside furnace

B T22 0.3 0.060 Outside furnace

C T22 0.2 0.060 Outside furnace

D T22 28 0.060 5.278

E T22 0.25 0.060 0.047

F T22 4 0.060 0.754

G 304H 0.25 0.045 0.035

H 304H 28 0.045 3.941

I 304H 0.2 0.045 Outside furnace

J T22 0.3 0.080 Outside furnace

K T22 0.03 NA Outside furnace

C-24
13622634
 RUL Methods and Models

2
The total heat absorbing outside surface area of the Assembly 4, Row 3 tube circuit is 10.055m .

The average heat absorption rate for this tube circuit is estimated using the energy balance
equations. From steam tables, the steam enthalpy, h(p,T), at the inlet of this tube circuit is:
o
h(15,000 kPa-a, 475 C) = 3237 kW-sec/kg

The steam enthalpy at the outlet is:

h(14,380 kPa-a, 595oC) = 3572 kW-sec/kg

The increase in energy of the steam is:

∆Esteam = Mass flow rate (Change in Enthalpy)

= 1.17 kg/sec (3572 kW-sec/kg - 3237 kW-sec/kg) = 392 kW

The heat absorption rate is the ratio of the increase in energy of the steam divided by the gas-
touched outside surface area:
2 2 2
Heat absorption rate = 392 kW/10.055m = 38.98 kW/m (12,357 Btu/h-ft )

Further refinements in the estimation of the local, segment by segment, and circumferential
distribution of heat absorption rate can be made by more comprehensive modeling, by direct, in-
situ measurements or can be inferred from steamside oxide scale measurements made of the
ferritic tubing in the furnace, just upstream of the austenitic stainless steel tubing.

Step IV(c): The objective of this step is to estimate the distribution of bulk steam temperature
along the length of each tube circuit.

The energy balance equation used in Step II(b) can be used to estimate the change in steam
energy at any location along the length of a given tube circuit. The best approach is usually to
start at the outlet header stub tube and estimate the bulk steam temperature in the upstream heat
absorbing segments of austenitic stainless steel tubing. In the case of Assembly 4, Row 3 the
final heat absorbing segment of tubing before the outlet penthouse segments of tubing is tube
2
Segment H. This tube segment has an outside surface area of 3.941 m . Using the estimated
average heat absorption rate for illustrative purposes we can estimate the steam enthalpy value at
the upstream end of segment H:

h(p,T) = h(outlet) – (Heat absorption rate)(Outside Surface Area)/Mass Flow Rate

2 2
= 3572 kW-sec/kg – (38.98 kW/m )( 3.941 m )/ 1.17 kg/sec = 3440 kW-sec/kg

The bulk steam temperature at the upstream end of Segment H can be estimated by using the
steam tables and looking up the temperature that corresponds to an enthalpy value of 3440 kJ/kg
and a steam pressure of 14,380 kPa. The temperature obtained is 544oC (1011oF).

C-25
13622634
RUL Methods and Models

Step IV(d): The objective of this step is to estimate the distribution of steam pressure and
velocity along the length of each tube circuit.

Since the superheater configuration does not include any bifurcations the distribution of steam
flow and absolute steam pressure can be estimated using the series-parallel models presented in
the earlier discussion of Step IV(d) in combination with the Darcy-Weisbach equation:

∆P = (ρ) (g) (hL) = (½) (ρ) (KL) (Q/Ai)2 = (½) (ρ) (KL) (V)2

If these models are combined with the heat/steam energy equations used in Steps II(b) and II(c)
then it is possible to iteratively solve the equations together and include the varying steam
properties such as steam density, ρ, in the estimations of local pressure, and steam velocity.
Using this approach the mass flow rate, volumetric flow rate, Q, and velocity, V, in Assembly 4,
Row 3, Segment H was estimated to be 1.0229 kg/sec, 0.0258 m3/sec, and 28.4 m/sec
respectively. The estimated mass flow rate is therefore approximately 14% slower than the
average mass flow rate for all the tube circuits. The pressure drop through Segment H can
estimated as follows:

∆P = (½) (ρ) (KL) (V)2 = (½) (ρ) (ƒ) (L) /( Di) (V)2

For the high fluid velocities anticipated in superheaters or reheaters the friction coefficient can be
4
estimated using the following relationship :

ƒ = 0.25 / [log(e/3.7Di + 5.74/Re0.9)]2

Where:
Re = (V)(Di)( ρ)/Dv
Dv = dynamic viscosity of the steam, kg/m-sec
ρ = steam density, kg/m3
3
For an estimated dynamic viscosity of 3.31E-5 kg/m-sec, and a steam density of 39.68 kg/m the
Reynolds number is:

Re = (28.4 m/sec)(0.034m)(39.68 kg/m3)/ 3.31E-5 kg/m-sec = 1,160,000

For a surface roughness, e, of 2.54E-5 m (0.001”) the friction factor, ƒ, is:

ƒ = 0.25 / {log(2.54E-5 m /[(3.7)(0.034m)] + 5.74/1,160,0000.9)}2 = 0.0187

The pressure drop through this 28-meter long by 34 mm inside diameter segment of tubing is:

∆P = (½) (39.68 kg/m3) (0.0187) (28m) (28.4 m/sec)2 /( 0.034m)

= 246,783 kg/m-sec2 (= 246,783 Pa = 36 psi)

C-26
13622634
 RUL Methods and Models

The distribution of pressure throughout a particular tube circuit can be estimated using this
segment-by-segment approach. A small improvement in accuracy can be obtained by also
including the local pressure drop factors due to entrance, exit, contraction, expansion, and
directional changes (i.e., bends). For example with the same surface roughness assumed above
o
the local loss factor, KL for a 90 bend with a bend radius to tube outside diameter ratio of one is
approximately 0.2. The pressure drop resulting from this bend (in addition to the frictional loss
associated with the length of the bend) is:

∆P = (½) (ρ) (KL) (V)2 = (½) (39.68 kg/m3) (0.2) (28.4 m/sec)2

= 3,200 kg/m-sec2 (=3,200 Pa = 0.5 psi)

Combining the relationships in all the preceding steps and solving the relationships iteratively
produced the steam velocity and absolute pressure values shown in Figure C-14.

Figure C-14
Steam pressure and velocity along the length of tube Rows 1, 2, and 3 of Assembly 4 of 10.
These estimates were made using the Darcy-Weisbach equation, and a series-parallel
circuit analogy. The surface roughness, e, should be adjusted to produce an estimated
pressure drop through each circuit that is close to measured values. In this case a surface
roughness of 2.54E-5 m (0.001”) produced good agreement.

Step IV(e): The objective of this step is to estimate the peak midwall temperature along the
length of each tube circuit

To estimate the peak midwall temperature at any particular tube location a radial heat transfer
analysis is performed. The boundary conditions for the analysis are:

C-27
13622634
RUL Methods and Models

• The bulk steam temperature

• The local heat flux (adjusted if necessary to provide the peak value around the perimeter of
the tube)
• The steam velocity
• The local physical properties of the steam (density, dynamic viscosity, thermal conductivity,
specific heat)
• The thermal conductivity of the tube alloy and the steamside oxide layers (magnetite, and
spinel)
• The thickness of the steamside oxide layers (adjusted to correspond to the thickness at the
end of the service interval of interest)

The first step is to estimate the steamside heat transfer coefficient, hi.

hi = Nu (ksteam)/Di

Where:

Nu = Nusselt number6 = (ƒ/8)(Re)(Pr)/[1.07 +12.7(ƒ/8)1/2(Pr2/3-1)

Pr = [(Dv)(Cp)/k]steam

For Assembly 4, Row 3, Segment H:

o o
Pr = (3.31E-5 kg/m-sec)(2.590 kW-sec/kg- C)/(8.996E-5 kW/m- C) = 0.953

Nu = (0.0187/8)(1,156,807)(0.953)/[1.07 +12.7(0.0187/8)1/2(0.9532/3-1) = 2456

hi = 2456 (8.996E-2 W/m-oC)/(0.034 m) = 6499 W/m2-oC (1144 Btu/hr-ft2-oF)

Tsteam/oxide = Tsteam + (q”OD, peak)(Do/ Di)/ hi

For a location of Segment H where the estimated bulk steam temperature, Tsteam, is 592oC and the
peak heat absorption rate, q”OD, peak, is 43,000 W/m2 the steam/steamside oxide interface
temperature, Tsteam/oxide, is estimated to be:
o 2 2 o o
Tsteam/oxide = 592 C + (43,000 W/m )( 0.0448 m/ 0.034 m)/ 6499 W/m - C = 600.8 C

The increase in radial temperature from the inside surface of the steamside oxide scale out to the
midwall position of the tube can be estimated as a series of conductive resistances. In the case of
austenitic tubing the steamside oxide thickness assumed in the analysis should be set to a value
that is thicker than the current value to simulate the temperature that will be present after future
growth. In this case study the steamside oxide thickness was increased fourfold for the analysis.
The temperature increase for each resistive layer is estimated as follows:

C-28
13622634
 RUL Methods and Models

∆T = q”OD, peak (Do/2) ln(Douter/Dinner)/klayer

Where:

Douter, and Dinner = the outer and inner diameter of the layer that has a thermal conductivity, klayer

The diameters, thermal conductivities, and estimated temperature distribution for the location
evaluated on Assembly 4, Row 3, Segment H are shown in Table C-5.
Table C-5
Diameters, thermal conductivities, and estimated temperature distribution

Location Diameter, m Thermal conductivity, Temperature. oC

W/m-oC

Steam/steamside oxide 0.0340 600.7

interface

Magnetite/hematite oxide 0.590

Magnetite/spinel oxide 0.0341 605.5

interface

Spinel oxide 5.76

Spinel/metal interface 0.0342 606.0

Type 304H material @ 22

600oC

Midwall 0.0395 612.3

Outside diameter 0.0448

Step IV(f): The objective of this step is to estimate the hoop stress due to steam pressure at any
desired location

For a local gage pressure of 14.662 MPa the estimated hoop stress for Assembly 4, Row 3,
Segment H is:

σ MDF = P (Do-t)/2t = (14.662 MPa) (0.0448 m – 0.0054 m)/[ 2(0.0054 m)] = 53.6 MPa

Step IV(g): The objective of this step is to estimate the stress rupture life for given midwall
temperature and mean diameter hoop stress values

For 304H material the lower bound (5th percentile) stress rupture life is:

C-29
13622634
RUL Methods and Models

trup5th = (0.339) 10^(-11.389+18746.602/(Tmidwall+273.15)-268.416*LOG(σMDF)/(Tmidwall

+273.15)-969.675*[LOG((σMDF)]2/(Tmidwall +273.15)))

Where:

σMDF = mean diameter hoop stress (or the total stress if global bending or residual stresses
are present), MPa
o
Tmidwall = midwall temperature, C

For the tube location with a stress of 53.6 MPa and a midwall temperature of 612oC the estimated
life is 321,000 hours.

This ends the steps in the detailed life assessment. The reader should keep in mind that the steps
in this case study were often simplified for illustrative purposes. Meaningful analyses can be
made with the steps presented but the reader is again encouraged to use these analytical tools in
combination with creative thought and engineering judgment.

As a final thought, the detailed version of the fictitious case study used as an illustration revealed
that more than two dozen failures would be expected to occur in the T22 portions of this
superheater prior to the first failure of austenitic tubing. This knowledge could be used to provide
additional calibration to the overall life assessment model and would provide invaluable
knowledge with regard to when and where both ferritic and austenitic segments of tubing may be
approaching end of life. If premature LTOC failure of austenitic tubing occurred prior to
numerous T22 tube failures then this would signal the need to evaluate global and residual
stresses and possible perform accelerated stress rupture testing of both the austenitic and ferritic
sections of the hottest tube circuits.

C.5.6 Step V: Refined Life Assessments

Two approaches to life estimation of LTOC failures of austenitic stainless steel have been
presented above. Each approach is based on some simplifying assumptions. The analyst must
carefully review the particular circumstances and determine if the assumptions required in each
model are acceptable. The analyst is encouraged to perform sensitivity studies to determine if a
particular assumption has a significant effect on the life predictions and decisions that must be
made based on the predictions. For example, if the uncertainty in stress rupture properties is so
large that clear decisions cannot be made then it may be worthwhile to remove some of the
hottest tubes in the population and perform accelerated stress rupture tests. If on the other hand,
lead-the-fleet failures have already occurred then accelerated testing is probably not justified. In
this case the life assessment model can be readily calibrated based on the times to failures and
characteristic of the lead-the-fleet failures.

Issues that may need to be considered and evaluated include:

• More detailed heat absorption rate, steam velocity, temperature measurements
• More detailed thermal-hydraulic modeling

C-30
13622634
 RUL Methods and Models

• Detailed stress analysis (e.g., if global bending stresses are significant or if the geometry of
bends is resulting in premature failures)
• Residual stress measurement (e.g., if the magnitude of the applied stress doesn’t explain the
time to failure or if the failures are occurring in bends, near malfunctioning sootblowers, or
are occurring in tubes that are not the “hottest” tubes)
• Metallographic evaluation (e.g., Appendix B).
• Accelerated creep/stress rupture testing of selected tubes
• More detailed evaluation of fireside wastage and its effect on the tube stress history

C.5.7 Step VI: Reporting results, conclusions and recommended actions

The overall objective of life estimations is to provide input into short or longterm decision
planning. The life estimations in Step IV(a) through Step IV(g) offer the most flexibility in terms
of decision making because the benefit of operational or design changes can be evaluated with
the model. In other cases the simple models presented in Steps II and III will provide all the
information needed to make realistic cost forecasts and meaningful decisions.

C.6 References

1 Abernethy, R.B., WeibullSmith, Fulton Findings, North Palm Beach, Florida

2 @RISK: Risk Analysis and Simulation Add-In for Microsoft Excel or Lotus 1-2-3,
Palisade Corporation, Newfield, New York
3 Idelchick, I.E., Handbook of Hydraulic Resistance Factors, 3rd Edition, Begell House,
(1996), Table 2-5.
4 Mays, L.W., Hydraulic Design Handbook, McGraw-Hill, NY, (1999), pg. 2.17.
5 Munson, B.R., D.F. Young, and T.H. Okiishi, Fundamentals of Fluid Mechanics, 4th
Edition, John Wiley & Sons, NY, (2002), pg. 445
th
6 Incropera, F.P. and D.P. Dewitt, Fundamentals of Heat and Mass Transfer, 4 Edition,
John Wiley & Sons, NY, (1996), pp. 483-487.

C-31
13622634
13622634
13622634
Target: About EPRI
Fossil Materials and Repair 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 energy-
related organizations in 40 countries. EPRI’s
multidisciplinary team of scientists and engineers
draws on a worldwide network of technical and
business expertise to help solve today’s toughest
energy and environmental problems.
EPRI. Electrify the World

© 2002 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

1004517

EPRI • 3412 Hillview Avenue, Palo Alto, California 94304 • PO Box 10412, Palo Alto, California 94303 • USA
800.313.3774 • 650.855.2121 • [email protected] • www.epri.com
13622634