API 579: a comprehensive tness-for-service guide

Ted L. Anderson
a,
*
, David A. Osage
b
a
Structural Reliability Technology, 1898 S Flatiron Court, Suite 235, Boulder, CO 80301, USA
b
M & M Engineering, Shaker Heights, OH, USA
Received 4 August 2000; revised 11 December 2000; accepted 13 December 2000
Abstract
This article presents an overview of the recently published American Petroleum Institute (API) Recommended Practice 579, which covers
tness-for-service assessment of pressure equipment in petrochemical and other industries. Although API 579 covers a wide range of aws
and damage mechanisms, including local metal loss, pitting corrosion, blisters, weld misalignment, and re damage, the emphasis of the
present article is on the assessment of crack-like aws. The API 579 procedure for evaluating cracks incorporates a failure assessment
diagram (FAD) methodology very similar to that in other documents, such as the British Energy R6 approach and the BS 7910 method. The
API document contains an extensive compendium of K solutions, including a number of new cases generated specically for API 579. In the
initial release of the document, API has adopted existing reference stress solutions for the calculation of L
r
in the FAD procedure. In a future
release, however, API plans to replace these solutions with values based on a more rational denition of reference stress. These revised
reference stress solutions will incorporate the effect of weld mismatch. In addition to the Appendices of K and reference stress solutions, API
579 includes appendices that provide guidance on estimating fracture toughness and weld residual stress distributions. Over the next few
years these appendices will be enhanced with advances in technology. Recently, API has entered into discussions with the American Society
of Mechanical Engineers (ASME) to convert API 579 into a joint API/ASME tness-for-service guide. q 2001 Published by Elsevier
Science Ltd.
Keywords: American Petroleum Institute; Failure assessment diagram; Flaw assessment; Fitness for service; Fracture toughness; Reference stress; Residual
stress; Stress intensity factor
1. Background
Existing US design codes and standards for pressurized
equipment provide rules for the design, fabrication, inspec-
tion and testing of new pressure vessels, piping systems, and
storage tanks. These codes do not address the fact that
equipment degrades while in-service and deciencies due
to degradation or from original fabrication may be found
during subsequent inspections. Fitness-for-service (FFS)
assessments are quantitative engineering evaluations,
which are performed to demonstrate the structural integrity
of an in-service component containing a aw or damage.
The American Petroleum Institute (API) Recommended
Practice 579 [1] has been developed to provide guidance
for conducting FFS assessments of aws commonly encoun-
tered in the rening and petrochemical industry which occur
in pressure vessels, piping, and tankage. However, the
assessment procedures can also be applied to aws encoun-
tered in other industries such as the pulp and paper industry,
fossil fuel utility industry, and nuclear industry. The guide-
lines provided in API 579 can be used to make run-repair-
replace decisions to ensure that pressurized equipment
containing aws that has been identied during an inspec-
tion can continue to be operated safely.
API 579 is intended to supplement and augment the
requirements in API 510 [2], API 570 [3], and API 653
[4]: to ensure safety of plant personnel and the public
while older equipment continues to operate; to provide tech-
nically sound FFS assessment procedures: to ensure that
different service providers furnish consistent remaining
life predictions; and to help optimize maintenance and
operation of existing facilities to maintain availability of
older plants and enhance long-term economic viability. In
addition, API 579 will also be used in conjunction with API
580 Recommended Practice For Risk-Based Inspection [5]
that is being developed to provide guidelines for risk assess-
ment, and prioritization for inspection and maintenance
planning for pressure-containing equipment.
The initial impetus to develop an FFS standard that could
be referenced from the API inspection codes was provided
by a Joint Industry Project (JIP) administered by the
International Journal of Pressure Vessels and Piping 77 (2000) 953963
0308-0161/00/$ - see front matter q 2001 Published by Elsevier Science Ltd.
PII: S0308-0161(01)00018-7
www.elsevier.com/locate/ijpvp
* Corresponding author. Tel.: 11-303-415-1475; fax: 11-303-415-1847.
E-mail address: tanderson@srt-boulder.com (T.L. Anderson).
Material Properties Council (MPC). The driving force
behind this development was plant safety. The methodology
provided for in this document, together with the appropriate
API inspection code, had to ensure that equipment integrity
could be safely maintained when operating equipment with
aws or damage, and could also be used to demonstrate
compliance with US Occupational Safety and Health
Administration (OSHA) 1910 Process Safety Management
(PSM) Legislation.
A review of the existing international FFS standards by
the members of the MPC JIP was undertaken in 1991 as the
starting point for the development of a new FFS standard.
Based on the results of this review, it was determined that a
comprehensive FFS standard covering many of the typical
aw types and damage mechanisms found in the rening
and petrochemical industry did not exist. In addition, the
existence of many company-based FFS methods, the
complexity of the technology that no single company can
solve on its own, and the need to gain acceptance by local
jurisdictions in the US further indicated the need for a new
standard. Therefore, the JIP decided to start the develop-
ment of the required FFS technology that would be needed
to write a comprehensive FFS standard for the rening and
petrochemical industry. The results of this work were docu-
mented in a MPC FFS JIP Consultant's Report [6], and this
document was subsequently turned over to the API Commit-
tee on Renery Equipment (CRE) FFS Task Force charged
with development of the FFS standard.
In terms adopted by the API CRE FFS Task Group devel-
oping API 579, an FFS assessment is an engineering analy-
sis of equipment to determine whether it is t for continued
service. The equipment may contain aws, may not meet
current design standards, or may be subjected to more
severe operating conditions than the original or current
design. The product of a FFS assessment is a decision to
operate the equipment as is, alter, repair, monitor, or
replace; guidance on an inspection interval is also provided.
FFS assessments consist of analytical methods to assess
aws and damage and usually require an interdisciplinary
approach consisting of the following:
Knowledge of damage mechanisms/material behavior.
Knowledge of past and future operating conditions and
interaction with operations personnel.
NDE (aw location and sizing).
Material properties (environmental effects).
Stress analysis (often nite element analysis).
Data analysis (engineering reliability models).
T.L. Anderson, D.A. Osage / International Journal of Pressure Vessels and Piping 77 (2000) 953963 954
Table 1
Organization of each section in API 579
Section
subparagraph
number
Title Overview
1 General The scope and overall requirements for an FFS assessment are provided
2 Applicability and limitations of
the FFS assessment procedures
The applicability and limitations for each FFS assessment procedure are clearly indicated; these
limitations are stated in the front of each section for quick reference
3 Data requirements The data requirements required for the FFS assessment are clearly outlined; these data requirements
include:
Original equipment design data
Maintenance and operational history
Required data/measurements for a FFS assessment
Recommendations for inspection technique and sizing requirements
4 Assessment techniques and
acceptance criteria
Detailed assessment rules are provided for three levels of assessment: Level 1, Level 2,
and Level 3. A discussion of these assessment levels is covered in the body of this paper
5 Remaining life evaluation Guidelines for performing a remaining life estimate are provided for the purpose of establishing an
inspection interval in conjunction with the governing inspection code
6 Remediation Guidelines are presented on methods to mitigate and/or control future damage. In many cases,
changes can be made to the component or to the operating conditions to mitigate the progression of
damage
7 In-service monitoring Guidelines for monitoring damage while the component is in-service are provided, these guidelines
are useful if a future damage rate cannot be estimated easily or the estimated remaining life is short.
In-service monitoring is one method whereby future damage or conditions leading to future damage
can be assessed or condence in the remaining life estimate can be increased.
8 Documentation Guidelines for documentation for an assessment are provided; the general rule is A practitioner
should be able to repeat the analysis from the documentation without consulting an individual
originally involved in the FFS assessment
9 References A comprehensive list of technical references used in the development of the FFS assessment
procedures is provided; references to codes and standards are provided in this section
10 Tables and gures Tables and gures including logic diagrams are used extensively in each section to clarify assessment
rules and procedures
11 Example problems A number of example problems are provided, which demonstrate the application of the FFS
assessment procedures
Based on this denition, the API CRE FFS Task Group
modied and greatly enhanced the initial efforts of the
MPC JIP to produce the rst edition of API 579. The
MPC JIP continued to provide valuable technical contribu-
tions throughout this development effort and essentially
became the technical development arm of the API Task
Group. The MPC FFS JIP is still in existence and continues
to provide FFS technology development while working
closely with the needs of the API CRE FFS Task Group.
The overall organization and assessment procedures
in API 579 are reviewed below. This is followed by a
more detailed discussion of the API 579 assessment of
cracks.
2. Overview of API 579
2.1. Applicable codes
API 579 provides guidelines for performing FFS assess-
ments that can be used in conjunction with the API Inspec-
tion codes (API 510, API 570 and API 653) to determine the
suitability for continued operation. The assessment proce-
dures in this recommended practice could be used for FFS
assessments and/or rerating of components designed and
constructed to the following codes:
ASME B and PV code, Section VIII, Division 1
ASME B and PV code, Section VIII, Division 2
ASME B and PV code, Section I
ASME B31.3 Piping code
ASME B31.1 Piping code
API 650
API 620.
Guidelines are also provided for applying API 579 to pres-
sure-containing equipment constructed to other recognized
codes and standards, including international and internal
corporate standards.
2.2. Organization
API 579 is a highly structured document designed to
facilitate use by practitioners and to facilitate future
enhancements and modications by the API CRE FFS
Task Group. Section 1 of the document covers: introduction
and scope; responsibilities of the owner-user, inspector, and
engineer; qualication requirements for the inspector and
engineer; and references to other codes and standards. An
outline of the overall FFS assessment methodology that is
T.L. Anderson, D.A. Osage / International Journal of Pressure Vessels and Piping 77 (2000) 953963 955
Table 2
Overview of aw and damage assessment procedures
Section in
API 579
Flaw or damage mechanism Overview
3 Brittle fracture Assessment procedures are provided to evaluate the resistance to brittle fracture of in-service
carbon and low alloy steel pressure vessels, piping, and storage tanks. Criteria are provided to
evaluate normal operating, start-up, upset, and shutdown conditions
4 General metal loss Assessment procedures are provided to evaluate general corrosion. Thickness data used for the
assessment can be either point thickness readings or detailed thickness proles. A
methodology is provided to guide the practitioner to the local metal loss assessment procedures
based on the type and variability of thickness data recorded during an inspection
5 Local metal loss Assessment techniques are provided to evaluate single and networks of Local Thin Areas
(LTAs), and groove-like aws in pressurized components. Detailed thickness proles are
required for the assessment. The assessment procedures can also be utilized to evaluate blisters
6 Pitting corrosion Assessment procedures are provided to evaluate widely scattered pitting, localized pitting,
pitting which occurs within a region of local metal loss, and a region of localized metal loss
located within a region of widely scattered pitting. The assessment procedures can also be
utilized to evaluate a network of closely spaced blisters. The assessment procedures utilize the
methodology developed for local metal loss
7 Blisters and laminations Assessment procedures are provided to evaluate either isolated, or networks of blisters and
laminations. The assessment guidelines include provisions for blisters located at weld joints
and structural discontinuities such as shell transitions, stiffening rings, and nozzles
8 Weld misalignment and
shell distortions
Assessment procedures are provided to evaluate stresses resulting from geometric
discontinuities in shell type structures including weld misalignment and shell distortions
(e.g. out-of-roundness, bulges, and dents)
9 Crack-like aws Assessment procedures are provided to evaluate crack-like aws. Recommendations for
evaluating crack growth including environmental concerns are also covered
10 High temperature operation
and creep
Assessment procedures are provided to determine the remaining life of a component operating
in the creep regime. The remaining life procedures are limited to the initiation of a crack
11 Fire damage Assessment procedures are provided to evaluate equipment subject to re damage. A
methodology is provided to rank and screen components for evaluation based on the heat
exposure experienced during the re. The assessment procedures of the other sections of this
publication are utilized to evaluate component damage
common to all assessment procedures included in API 579 is
provided in Section 2 of the document. The organization of
Section 2 is shown in Table 1. This same organization is
utilized in all subsequent sections that contain FFS assess-
ment procedures.
Starting with Section 3, a catalogue of FFS assessment
procedures organized by damage mechanism is provided in
API 579. A complete listing of the aw and damage assess-
ment procedures currently covered is shown in Table 2.
These damage mechanisms can be grouped at a higher
level to form a degradation class (see Fig. 1). This higher
level of organization is useful in that it provides insight into
how the assessment procedures of different sections may be
combined to address complex aws in a component. As
shown in Fig. 1, several aw types and damage mechanisms
may need to be evaluated to determine the FFS of a compo-
nent. Each section in API 579 referenced within a degrada-
tion class includes guidance on how to perform an
assessment when multiple damage mechanisms are present.
When assessment procedures are developed for a new
damage mechanism, they will be added as a self-contained
section to maintain the structure of API 579. Currently, new
sections are being developed to address hydrogen induced
cracking (HIC) and stress-oriented hydrogen induced crack-
ing (SOHIC) damage, local hot spots, assessment proce-
dures for riveted components, and creep crack growth.
A series of appendices are provided which contain tech-
nical information that can be use with all sections of API
T.L. Anderson, D.A. Osage / International Journal of Pressure Vessels and Piping 77 (2000) 953963 956
Fig. 1. Schematic overview of the FAD procedure.
579, which cover FFS assessment procedures. The majority
of the information in the appendices covers stress analysis
techniques, material property data, and other pertinent infor-
mation that is required when performing a FFS assessment.
An overview of the appendices is provided in Table 3.
2.3. Assessment methodology
The API 579 FFS assessment methodology used for all
damage types is provided in Table 4. The organization of
each section of API 579 that covers an assessment procedure
is consistent with this methodology. This consistent
approach to the treatment of damage and the associated
FFS assessment procedures facilitates use of the document
in that, if a practitioner is familiar with one section of the
document, it is not difcult to utilize another section
because of the common structure. This assessment metho-
dology has proven to be robust for all aw and damage types
that have been incorporated into API 579. Because of this
success, when new sections are added to API 579, the
template used for the development will be based on this
assessment methodology.
2.4. Assessment levels
Three levels of assessment are provided in API 579 for
each aw and damage type. A logic diagram is included in
each section to illustrate how these assessment levels are
interrelated. As an example, the logic diagram for evaluat-
ing crack-like aws is shown in Fig. 2. In general, each
assessment level provides a balance between conservatism,
the amount of information required for the evaluation, the
skill of the practitioner performing the assessment, and the
complexity of analysis being performed. Level 1 is the most
conservative, but is easiest to use. Practitioners usually
proceed sequentially from a Level 1 to a Level 3 assessment
(unless otherwise directed by the assessment techniques) if
the current assessment level does not provide an acceptable
result or a clear course of action cannot be determined.
A general overview of each assessment level and its
intended use are described below.
Level 1 The assessment procedures included in this
level are intended to provide conservative screening
criteria that can be utilized with a minimum amount of
inspection or component information. The Level 1
assessment procedures may be used by either plant
inspection or engineering personnel.
Level 2 The assessment procedures included in this
level are intended to provide a more detailed evaluation
that produces results that are less conservative than those
T.L. Anderson, D.A. Osage / International Journal of Pressure Vessels and Piping 77 (2000) 953963 957
Table 3
API 579 appendices
Appendix Title Overview
A Thickness, MAWP and membrane stress
equations for a FFS assessment
Equations for the thickness, MAWP, and membrane stress are given for most of the common
pressurized components. These equations are provided to assist international practitioners who
may not have access to the ASME code and who need to determine if the local design code is
similar to the ASME code for which the FFS assessment procedures were primarily designed for
B Stress analysis overview for a FFS
assessment
Recommendations for stress analysis techniques that can be used to perform an FFS assessment
are provided including guidelines for nite element analysis
C Compendium of stress intensity factor
solutions
A compendium of stress intensity factor solutions for common pressurized components (i.e.
cylinders, spheres, nozzle, etc.) are given. These solutions are used for the assessment of crack
like aws. The solutions presented represent the latest technology and have been re-derived using
the nite element method in conjunction with weight functions
D Compendium of reference stress solutions A compendium of reference stress solutions for common pressurized components (i.e. cylinders,
spheres, nozzle, etc.) are given. These solutions are used for the assessment of crack-like aws
E Residual stresses in a FFS evaluation Procedures to estimate the through-wall residual stress elds for different weld geometries are
provided; this information is required for the assessment of crack like aws
F Material properties for a FFS assessment Material properties required for all FFS assessments are provided including:
Strength parameters (yield and tensile stress)
Physical properties (i.e. Young's Modulus, etc.)
Fracture toughness
Data for fatigue crack growth calculations
Fatigue curves (Initiation)
Material data for creep analysis including remaining life and creep crack growth
G Deterioration and failure modes An overview of the types of aws and damage mechanisms that can occur is provided,
concentrating on service-induced degradation mechanisms. This appendix only provides an
abridged overview on damage mechanisms; API 571 is currently being developed to provide a
denitive reference for damage mechanisms that can be used with API 579 and API 580
H Validation An overview of the studies used to validate the general and local metal loss, and the crack-like
aw assessment procedures are provided
I Glossary of terms and denitions Denitions for common terms used throughout the sections and appendices of API 579 are given
J Technical inquiries Guidelines for submitting a technical inquiry to API are provided. Technical inquires will be
forwarded to the API CRE FFS task group for resolution
from a Level 1 assessment. In a Level 2 assessment,
inspection information similar to that required for a
Level 1 assessment are required; however, more detailed
calculations are used in the evaluation. Level 2 assess-
ments are typically conducted by plant engineers or engi-
neering specialists experienced and knowledgeable in
performing FFS assessments.
Level 3 The assessment procedures included in this
level are intended to provide the most detailed evaluation
that produces results that are less conservative than those
from a Level 2 assessment. In a Level 3 assessment the
most detailed inspection and component information is
typically required, and the recommended analysis is
based on numerical techniques such as the nite element
method. The Level 3 assessment procedures are primar-
ily intended to be used by engineering specialists experi-
enced and knowledgeable in performing FFS
evaluations.
T.L. Anderson, D.A. Osage / International Journal of Pressure Vessels and Piping 77 (2000) 953963 958
Table 4
API 579 FFS assessment methodology for all damage types
Step Description
1 Flaw and damage mechanism identication The rst step in a FFS assessment is to identify the aw type and cause of damage. FFS assessments
should not be performed unless the cause of the damage can be identied. The original design and fabrication practices, materials of construction,
service history, and environmental conditions can be used to ascertain the likely cause of the damage. Once the aw type is identied, the appropriate
section of this document can be selected for the assessment
2 Applicability and limitations of the FFS assessment procedures The applicability and limitations of the assessment procedure are described in each
section, and a decision on whether to proceed with an assessment can be made
3 Data requirements The data required for FFS assessments depend on the aw type or damage mechanism being evaluated. Data requirements may
include: original equipment design data; information pertaining to maintenance and operational history; expected future service; and data specic to the
FFS assessment such as aw size, state of stress in the component at the location of the aw, and material properties. Data requirements common to all
FFS assessment procedures are covered in Section 1. Data requirements specic to a damage mechanism or aw type are covered in the section
containing the corresponding assessment procedures
4 Assessment techniques and acceptance criteria Assessment techniques and acceptance criteria are provided in each section. If multiple damage
mechanisms are present, more than one section may have to be used for the evaluation
5 Remaining life evaluation An estimate of the remaining life or limiting aw size should be made. The remaining life is established using the FFS
assessment procedures with an estimate of future damage rate (i.e. corrosion allowance). The remaining life can be used in conjunction with an
inspection code to establish an inspection interval
6 Remediation Remediation methods are provided in each section based on the damage mechanism or aw type. In some cases, remediation
techniques may be used to control future damage associated with aw growth and/or material degradation
7 In-service monitoring Methods for in-service monitoring are provided in each section based on the damage mechanism or aw type. In-service
monitoring may be used for those cases where, a remaining life and inspection interval cannot be adequately established because of the complexities
associated damage mechanism and service environment
8 Documentation The documentation of an FFS assessment should include a record of all data and decisions made in each of the previous steps to
qualify the component for continued operation. Documentation requirements common to all FFS assessment procedures are given in Section 2 of API
579. Specic documentation requirements for a particular damage mechanism or aw type are covered in the section containing the corresponding
assessment procedures
Fig. 2. Level 2 FAD, which shows typical cut-off values.
2.5. Remaining life and rerating
The FFS assessment procedures in API 579 cover both
the present integrity of the component given a current state
of damage and the projected remaining life. If the results of
a FFS assessment indicate that the equipment is suitable for
the current operating conditions, the equipment can
continue to be operated at these conditions, if a suitable
inspection program is established. If the results of the FFS
assessment indicate that the equipment is not suitable for the
current operating conditions, calculation methods are
provided in API 579 to rerate the component. For pressur-
ized components (e.g. pressure vessels and piping) these
calculation methods can be used to nd a reduced maximum
allowable working pressure and/or coincident temperature.
For tank components (i.e. shell courses) the calculation
methods can be used to determine a reduced Maximum
Fill Height. The remaining life calculation in API 579 is
not intended to provide a precise estimate of the actual
time to failure. Alternatively, the remaining life calculation
is used to establish an appropriate inspection interval in
conjunction with the governing inspection code and/or in-
service monitoring plan, or the need for remediation.
2.6. Relationship to other FFS standards
As previously discussed, members of the MPC FFS JIP
reviewed existing international FFS standards to determine
the suitability for use in the rening and petrochemical
industry. Although a single comprehensive standard did
not exist, technology contained in these international stan-
dards was identied that could be utilized for certain aw
types. Where possible, parts of these methodologies were
incorporated into API 579, and in many cases they were
signicantly enhanced. In some cases, where the technology
was not directly incorporated, the API CRE FFS Task
Group members felt that alternate approaches may be desir-
able for use by more advanced practitioners. Therefore, the
Level 3 assessment in API 579 permits the use of alternative
FFS assessment methodologies. For example, the Level 3
assessment in Section 9 of API 579 covering crack-like
aws provides references to British Energy R6 [7], BS
7910 [8], SAQ/FoU-Report 96/08 [9], WES 2805 [10],
and EPRI J-Integral methodology [11].
3. Overview of API 579 crack-like aw assessment
Section 9 of API 579 covers the assessment of cracks and
other planar aws. As is the case with other prominent
procedures, such as R6 and BS 7910, the failure assessment
diagram (FAD) methodology forms the basis of the aw
evaluation.
Fig. 1 illustrates the FAD concept. The toughness ratio,
K
r
, and the load ratio, L
r
, for the structure of interest are
plotted on the diagram. The FAD curve represents the
predicted failure locus. If the assessment point falls within
the curve, it is considered acceptable.
The toughness ratio is computed from the following
expression:
K
r

K
P
I
1FK
SR
I
K
mat
; 1
where K
P
I
is the applied stress intensity factor due to primary
loads, K
SR
I
is the stress intensity factor due to secondary and
residual stress, K
mat
is the fracture toughness, and F is a
plasticity adjustment factor on K
SR
I
: Note that the above
formulation, which was recently suggested by Ainsworth
et al. [12], differs somewhat from that in the current versions
of R6 and BS 7910, which account for secondary and resi-
dual stress plasticity effects through the r factor, which is
added to K
r
. Eq. (1), which has a multiplying factor on K
SR
I
;
is a more rigorous formulation. Both the r and F formula-
tions were derived from the same analyses. However, the r
factor formulation implies a toughness dependence on plas-
tic zone formation, which has no theoretical basis. The more
correct form for K
r
in Eq. (1) will most likely appear in
future revisions of R6 and BS 7910.
The load ratio in API 579 is dened as
L
r

s
ref
s
ys
; 2
where s
ref
is the reference stress and s
ys
is the yield
strength. Eq. (2) is identical to the L
r
denition in R6 and
BS 7910. However, API 579 proposes an alternative deni-
tion of the reference stress, as discussed later in this article.
The main crack-like aw assessment in API 579 is Level
2, which uses the following FAD equation:
K
r
1 20:14L
r

2
{0:3 10:7 exp20:65L
r

6
}
for L
r
# L
rmax
;
3
which is the same as the R6 Option 1 FAD, as well as the
one of the available Level 2 FAD expressions in BS 7910.
This FAD has a cut-off at L
r(max)
, which is dened as
L
rmax

1
2
1 1
s
ts
s
ys
_ _
; 4
where s
ts
is the tensile strength. Fig. 2 shows a plot of Eq. (3)
with typical cut-offs for various steels.
Level 2 utilizes partial safety factors (PSFs) on toughness,
aw size and stress, whereby the user can select a target
reliability and perform a deterministic analysis. If, after
adjusting the input values by the PSFs, the assessment
point lies inside the FAD, one can conclude that the actual
probability of failure is less than the target value. The PSFs
tabulated in Section 9 of API 579 were generated as part of
the MPC FFS project [13].
The API 579 Level 3 assessment is a more advance analy-
sis that gives the user a substantial amount of exibility. The
T.L. Anderson, D.A. Osage / International Journal of Pressure Vessels and Piping 77 (2000) 953963 959
available options for a Level 3 assessment include:
Method A Level 2 assessment with user-generated
partial safety factors or a probabilistic analysis.
Method B Material-specic FAD, similar to R6
Option 2.
Method C J-based FAD obtained from elasticplastic
nite element analysis, similar to R6 Option 3.
Method D Ductile tearing assessment.
Method E Use a recognized assessment procedure,
such as R6 or BS 7910.
The Level 1 assessment is very simple screening evaluation
that can be performed by a qualied inspector. Level 1
consists of a series of allowable aw size curves. These
curves were generated using the Level 2 assessment with
conservative input assumptions. Note that the API 579
Level 1 assessment of cracks is completely different than
the BS 7910 Level 1 assessment. The latter is a pseudo FAD
analysis that is intended to maintain backward compatibility
with the 1980 version of the BS PD 6493 procedure. Unlike
Level 1 of BS 7910, the API 579 Level 1 assessment
requires almost no calculations.
4. New K solutions in API 579
Appendix C contains an extensive library of stress inten-
sity solutions for cracked bodies. Many of these solutions
were obtained from the published literature as well as other
assessment procedures, including BS 7910. New K solutions
were also generated for inclusion in API 579. In particular, a
comprehensive set of solutions for cracks in cylindrical and
spherical shells was recently developed [14]. This study
involved over 2400 nite element runs. Of course, there
were a number of existing solutions for cylinders and
spheres, but these tended to cover a limited range of
radius/thickness and aw aspect ratios.
In a study commissioned by the MPC FFS project [14],
the following geometries and aw orientations were
considered:
Internal axial surface aws in a cylinder.
External axial surface aws in a cylinder.
Internal circumferential surface aws in a cylinder.
External circumferential surface aws in a cylinder.
Internal meridianal surface aws in a sphere.
External meridianal surface aws in a sphere.
Three load cases were analyzed:
Uniform crack face pressure.
Linearly varying crack face pressure.
Global bending moment (circumferential cracks in
cylinders).
The rst 2 load cases can be used to derive a weight func-
tion, which can be used to infer K for an arbitrary through-
wall stress eld. The procedure for generating weight func-
tions from the uniform and linear crack face pressures is
outlined in Appendix C of API 579.
The range of dimensional parameters for the cylinder and
sphere analyses is as follows:
R
i
=t 3; 5, 10, 20, 60, 100, 1.
a=t 0:2; 0.4, 0.6, 0.8.
c=a 0:5; 1, 2, 4, 8, 16, 32.
where R
i
is the inside shell radius, t is the wall thickness, a is
the depth of the surface aw, and 2c is the surface aw
length.
Fig. 3 is a plot of typical results from the recent analyses.
Uniform crack face pressure was applied, giving a stress
intensity solution of the following form:
K
I
pG
0

pa
Q
_
; 5
where p is the crack face pressure, G
0
is a dimensionless
geometry factor, and Q is the aw shape parameter:
Q 1 11:464
a
c
_ _
1:65
: 6
Note that there is a signicant R
i
/t effect on the nondimen-
sional stress intensity factor, G
0
. Consequently, using a K
solution for a surface crack in a at plate when assessing a
curved shell could lead to signicant errors.
The K solution library in API 579 will be expanded as
new cases become available. Currently, solutions for cylin-
ders with R
i
=t 1 are being computed. In the near future, K
solutions for cracks at structural discontinuities such as
nozzles and stiffening rings will be generated.
5. Fracture toughness estimation
Appendix F of API 579 contains information on material
properties, including toughness. This appendix does not
contain a database of toughness values, however. Rather,
it provides correlations and estimation methods. For ferritic
steels, there are lower-bound correlations of toughness to
Charpy transition temperature. These correlations were
adapted from Sections III and XI of the ASME boiler and
pressure vessel code. For static loading in the absence of
dissolved hydrogen, the lower-bound toughness correlation
is as follows:
K
IC
36:5 13:084 exp0:036T 2T
ref
156
MPa

m
p
; 8C;
7a
K
IC
33:2 12:806 exp0:02T 2T
ref
1100
ksi

in:
p
; 8F;
7b
T.L. Anderson, D.A. Osage / International Journal of Pressure Vessels and Piping 77 (2000) 953963 960
where T
ref
is the 20 J (15 ft-lb) transition temperature in the
case of carbon steels. For dynamic loading or for hydrogen-
charged steels, the following lower-bound correlation can
be used:
1
K
IR
29:5 11:344 exp0:0260T 2T
ref
189
MPa

m
p
; 8C;
8a
K
IR
26:8 11:223 exp0:0144T 2T
ref
1160
ksi

in:
p
; 8F:
8b
An upper-shelf cut-off must be imposed on the above
expressions. For older, high-sulfur steels, a cut-off of
110 MPa

m
p
(100 ksi

in:
p
) is recommended. For newer,
low-sulfur steels, a cut-off of 220 MPa

m
p
(200 ksi

in:
p
)
may be assumed.
For probabilistic fracture analyses of steel structures, API
579 endorses the use of the fracture toughness Master
Curve, as implemented in ASTM Standard E 1921-97
[15]. The Master Curve quanties the temperature depen-
dence of steels in the transition range, as well as the statis-
tical distribution of toughness at a given temperature. The
latter is characterized by a three-parameter Weibull distri-
bution with two of the three parameters specied:
F 1 2exp 2
B
25:4
K
Jc
220
K
0
220
_ _
4
_ _
mm; MPa

m
p
; 9a
F 1 2exp 2B
K
Jc
218:2
K
0
218:2
_ _
4
_ _
in:; ksi

in:
p
; 9b
where F is the cumulative probability, B the specimen thick-
ness (crack front length), and K
0
is the Weibull mean tough-
ness, which corresponds to the 63rd percentile value. The
temperature dependence of the median (50th percentile)
toughness is given by
K
Jcmedian
30 170 exp0:0190T 2T
0
MPa

m
p
; 8C;
10a
K
Jcmedian
27 164 exp0:0106T 2T
0
ksi

in:
p
; 8F;
10b
where T
0
is the index transition temperature material for the
material of interest. It corresponds to the temperature at
which the median toughness for a 25 mm (1 in.) thick speci-
men is 100 MPa

m
p
(91 ksi

in:
p
). The median and Weibull
mean are related as follows:
K
0

K
Jcmedian
220
ln2
0:25
120 MPa

m
p
; 11a
K
0

K
Jcmedian
218:2
ln2
0:25
118:2 ksi

in:
p
: 11b
By combining Eqs. (9a), (9b), (10a), (10b) and (11a), (11b),
we see that once T
0
is known, the toughness in the transition
T.L. Anderson, D.A. Osage / International Journal of Pressure Vessels and Piping 77 (2000) 953963 961
1
The rationale for using a dynamic crack arrest fracture toughness corre-
lation for hydrogen charged steels is as follows: If dissolved hydrogen is
present, it may degrade the material's ability to resist brittle fracture initia-
tion. Once rapid crack propagation begins, however, the hydrogen can no
longer inuence the material behavior. Therefore, the crack arrest tough-
ness should be a reasonable lower-bound estimate of the material's ability
to resist unstable crack propagation.
Fig. 3. Nondimensional stress intensity factor at the deepest point of a surface crack f p=2 as a function of thickness/radius ratio in cylinders and spheres.
region is completely described. ASTM E 1921-98 outlines
the procedure for determining T
0
from fracture toughness
testing in the transition region.
When fracture toughness testing is not feasible, T
0
can be
estimated from the 27 J (20 ft-lb) transition temperature:
T
0
T
27 J
2188C; 12a
T
0
T
20 ft-lb
232:48F: 12b
The above correlation has a standard deviation of approxi-
mately 158C (278F).
6. Reference stress and weld mismatch
Appendix D of API 579 contains reference stress solu-
tions for a variety of cracked bodies. For the most part, these
solutions were adopted directly from R6 and BS 7910 and
are based on limit load solutions.
The authors believe that the current denition of refer-
ence stress based on limit load is inappropriate and should
be replaced in the long run. When rigorous elasticplastic J
solutions for cracked bodies are plotted in terms of FADs,
the resulting curves exhibit a strong geometry dependence
when L
r
is computed based on the limit load solutions. This
apparent geometry dependence has led some to criticize the
FAD methodology as being inaccurate.
Appendix B of API 579 outlines a procedure to obtain a
self-consistent reference stress denition from the elastic
plastic J solution. This alternative denition removes
virtually all of the geometry dependence in the FAD. This
approach also provides an effective means to account for
weld metal mismatch through L
r
.
Of course, there is no ambiguity in the denition of the
vertical ordinate (y axis) of the FAD from an elasticplastic
J solution:
K
r

J
elastic
J
:
_
13
This is plotted against the load ratio, as dened in Eq. (2).
The potential geometry dependence of the FAD curve arises
in the denition of reference stress. A self-consistent deni-
tion of s
ref
can be derived from the R6 Option 2 FAD
equation, which is material-specic but is assumed to be
geometry-independent. Setting L
r
1 in this expression
leads to
J
J
elastic
_
_
_
_
L
r
1
1 1
0:002E
s
ys
1
1
2
1 1
0:002E
s
ys
_ _
21
: 14
The above expression assumes that s
ys
is the 0.2% offset
yield strength. The reference stress is linearly related to the
nominally applied stress through a geometry factor, H:
s
ref
Hs
nominal
; 15
where H is inferred from the nominal stress at L
r
1:
H
s
ys
s
nominal
j
L
r
1
: 16
Thus, the reference stress is chosen in such a way that the
Option 2 FAD will nearly match a rigorous elasticplastic J
analysis. That is, given the above denition of reference
stress, Option 2 and Option 3 FADs will be virtually
identical.
The forgoing begs the question: if an elasticplastic J
analysis is required to determine L
r
, what is the point in
using the FAD methodology? Traditionally one of the
advantages of the FAD approach has been that it is consid-
erably simpler than a rigorous elasticplastic analysis. At
rst glance, the proposed denition of s
ref
would seem to
eliminate this advantage. Such is not the case, however, as
discussed below.
When L
r
is set to unity in the Option 2 FAD expression,
the strain hardening dependence disappears. Consequently,
the geometry factor H, dened above, should be insensitive
to the shape of the stressstrain curve. If an elasticplastic
analysis is performed once for a given cracked body, it
should not have to be repeated for other stressstrain curves.
Non-dimensional reference stress solutions can be
computed and tabulated for use in standard FAD analyses,
much like compendia of K solutions are currently published.
As part of the ongoing efforts to enhance the technology
in API 579, a project is planned in which reference stress
solutions (based on the above denition) will be generated
for a range of cracked bodies. This project will also address
weld mismatch effects.
7. Residual stress distributions for FFS assessment
One of the key assumptions in fracture assessments of
welded structures is the residual stress distribution. Earlier
assessment procedures, such as PD 6493 (both the 1980 and
1991 versions), made the very conservative assumption of
yield-magnitude membrane residual stresses in as welded
components. More recent assessment procedures, including
API 579, have removed much of this conservatism.
Appendix E of API 579 contains a compendium of resi-
dual stress distributions for various weld geometries. These
distributions are based on nite element analyses of weld
residual stresses in a series of pipe girth welds, seam welds,
and nozzle-to-head attachment welds performed under MPC
sponsorship. Based on these results, a series of parametric
residual stress distributions were developed and included in
API 579 Appendix E. However, an in-depth review of the
residual stress analyses performed thus far and a large body
of recent residual stress results from other sources over the
last few years suggest that additional work should be
performed to improve the current FFS procedures for pres-
sure vessel and piping components. An upcoming research
T.L. Anderson, D.A. Osage / International Journal of Pressure Vessels and Piping 77 (2000) 953963 962
project will address the following issues:
Conrmation of some of the parametric distributions in
Appendix E.
A clear criterion for selecting `bending' and `self-equili-
brating' types of residual stress distributions in pipe/
vessel welds.
Development of improved residual stress distributions
for llet welds at corner joints, nozzle welds, and repair
welds.
Incorporation of local post-weld heat treatment effects.
Appendix E will continually be expanded and revised as
new results become available.
8. API and ASME FFS activities
The American Society of Mechanical Engineers (ASME)
has formed a new main committee, the Post Construction
Main Committee, with a charter to develop codes and stan-
dards for in-service pressure containing equipment covering
all industries. Currently, standards development activity is
underway in the areas of Risk-Based Inspection (RBI) and
repair methods (e.g. leak sealing, boxes, patches, etc.).
In the area of FFS, API and ASME are working to create a
new standards committee that will jointly produce a single
FFS standard in the US that can be used for pressure
containing equipment. It is envisioned that once the nego-
tiations and operating procedures for the new committee
structure are complete, API 579 will form the basis of the
joint API/ASME standard that will be produced by this
committee. The initial release of the new standard will
include all topics currently contained in API 579 and will
also contain an FFS assessment procedure for the evaluation
of creep crack growth. This assessment procedure is
currently being developed jointly by the Pressure Vessel
Research Council (PVRC), Continued Operation of Equip-
ment (COE) Division and the ASME Post Construction
Committee Subgroup on Creep and Fatigue Growth, and
is being sponsored by Edison Electric Institute.
The agreement to produce a joint standard on FFS tech-
nology is a landmark decision that will focus resources in
the US to develop a single document that can be used in all
industries. This will help avoid jurisdictional conicts and
promote uniform acceptance of FFS technology. It also
provides an opportunity for pooling of resources of API,
ASME, PVRC, and MPC to develop new FFS technology
as required by the standards committee. Discussions are
already in progress, and suggestions have been made to
have the new standards committee meetings in conjunction
with PVRC. This would help to create a focal point for FFS
technology development in that the PVRC COE and MPC
FFS JIP have previously met at this time. In addition, the
members of the standards committee could directly
interface with members of these groups to dene technology
needs and help arrange for appropriate funding levels.
References
[1] API. Recommended practice for tness-for-service. API 579.
Washington, DC: American Petroleum Institute, 2000.
[2] API. Pressure vessel inspection code: maintenance inspection,
rerating, repair and alteration, API 510. Washington, DC: American
Petroleum Institute, 1999.
[3] API. Piping inspection code: inspection, repair, alteration, and
rerating of in-service piping systems. API 570. Washington, DC:
American Petroleum Institute, 1998.
[4] API. Tank inspection, repair, alteration, and reconstruction, API 653.
Washington, DC: American Petroleum Institute, 1998.
[5] API. Recommended practice for risk-based inspection, API 580 (in
development). Washington, DC: American Petroleum Institute.
[6] MPC. Fitness-for-service evaluation procedures for operating pres-
sure vessels, tanks, and piping in renery and chemical service,
FFS-26. New York, NY: The Materials Properties Council, October,
1995.
[7] British Energy. Assessment of the integrity of structures containing
defects. British Energy R-6, 1999.
[8] BSI. Guide on methods for assessing the acceptability of aws in
structures, BS 7910. British Standards Institute, 1999.
[9] SAQ/FoU. A procedure for safety assessment of components with
cracks Handbook. SAQ/FoU-Report 96/08, 1997.
[10] Method of assessment for aws in fusion welded joints with respect to
brittle fracture and fatigue crack growth, WES 2805, 1997.
[11] Kumar V, German MD, Shih CF. An engineering approach for elas-
ticplastic fracture analysis. EPRI Report NP-1931, Palo Alto, CA:
EPRI, 1981.
[12] Ainsworth RA, Sharples JK, Smith SD. Effects of residual stress on
fracture behavior experimental results and assessment methods.
J Strain Anal 2000:53.
[13] Osage DA, Shipley KS, Wirsching PH, Mansour AE. Application of
partial safety factors for pressure containing equipment. Presented at
the 2000 ASME Pressure Vessel and Piping Conference, Seattle, July,
2000.
[14] Anderson TL, Thorwald GV, Revelle DJ. Stress intensity solutions for
surface cracks and buried cracks in cylinders, spheres, and at plates.
Presented at the 2000 ASME Pressure Vessel and Piping Conference,
Seattle, July, 2000.
[15] ASTME 1921-97 Standard test method for determination of reference
temperature, T
0
, for ferritic steels in the transition range. Philadelphia:
American Society for Testing and Materials, 1997.
T.L. Anderson, D.A. Osage / International Journal of Pressure Vessels and Piping 77 (2000) 953963 963