Design Method Combining API and

ASME Codes for Subsea Equipment

for HP/HT Conditions up to 25,000-psi
Pressure and 400F Temperature
Parth D. Pathak, Christopher G. Kocurek,* and Samuel L. Taylor, Cameron International Corporation**

Summary head specifications were developed in 1952. The first 20,000-psi

A current challenge in the offshore industry is the design of subsea wellhead systems were developed in 1972, quickly followed by
equipment for pressures exceeding 15,000 psi and temperatures ex- 30,000-psi wellhead systems in 1974. These sytems have been suc-
ceeding 250F. This combination of pressure and temperature has cessfully deployed both in land- and platform-based fields in the
been fairly accepted as the start of the high-pressure/high-temper- Gulf of Mexico and in the North Sea. In the years from 1960 to
ature (HP/HT) region. The current American Petroleum Institute 1990, subsea wells rated at a working pressure of up to 5,000 psi
(API) standard for designing subsea equipment, API Specification at a water depth of 2,000 ft were being explored. In the 1990s,
(SPEC) 17D (2011), is limited to a working pressure of 15,000 psi the working pressure increased to 10,000 psi at a water depth of
and provides little guidance on temperature conditions exceeding 3,000 ft. From 2000 to 2012, subsea wells rated up to 15,000 psi
250F. This paper demonstrates a design methodology that com- and 350F at a water depth of 8,000 ft were explored. The trend of
bines the API and American Society of Mechanical Engineers the industry is moving toward wells being explored at 20,000psi
(ASME) Boiler and Pressure Vessels Code (BPVC) for designing and 400F or higher at water depths approaching or exceeding
an example subsea pressure containing component for HP/HT con- 10,000ft. These wells require HP/HT-rated equipment for oil pro-
ditions greater than 15,000 psi and 250F. duction, generating the need for a design method for such HP/HT-
This paper shows the evaluation of a combined load-capacity rated subsea production equipment.
chart for an API SPEC 17D flange flow-loop [API SPEC 6A The existing design methods use API SPEC 17D and the ASME
(2010), 4 in., 20ksi] for a design pressure of 20,000 psi and a tem- BPVC as the primary design codes. The ASME BPVC was a result
perature of 350F with external tension and bending loads. Both of a committee set up by ASME in 1911 for the purpose of formu-
the linear elastic and elastic plastic methods for protection against lating standard rules for the construction of steam boilers and other
plastic collapse are used to determine the structural capacity of the pressure vessels. This code is formulated for pressure vessels in the
flange body. These methods combine the API material and design nuclear industry and for designing storage and transportation tanks.
allowables and ASME design methods. Stress classification and API SPEC 17D was formulated for standardization of subsea pro-
linearization are used for the evaluation of design capacities with duction systems. Because API SPEC 17D was formulated specifi-
linear methods. Modified load-resistance design factors are used cally for subsea equipment, it will be mandatory to meet the API
both to evaluate design capacities and to account for the difference material requirements, design allowables, and test requirements.
in ASME and API hydrostatic-test pressures with elastic plastic Subsea-wellhead and tree equipment rated up to 15,000 psi can
methods. The structural capacity is combined with thermal analysis be designed with methods documented in API SPEC 17D. How-
to determine the effects of high temperature on the flange capacity. ever, this standard does not provide guidance for equipment rated at
To assess the cyclic-loading capacity of the flange, stress-based fa- pressures greater than 15,000 psi and combined with temperatures
tigue analysis and fracture-mechanics analysis are also compared. greater than 250F. The design methods given in API SPEC 17D
The results obtained are comparable to existing API Technical refer to the design methods of API SPEC 6A (2010), which allows
Report (TR) 6AF1 (1998) charts. This work has been performed to for the application of ASME BPVC, Section VIII Division 2 (2010)
demonstrate both the acceptance of existing methods for HP/HT design guidelines (as well as other codes) for equipment rated for
conditions and to introduce the advanced ASME design methods working pressures up to 20,000 psi and for temperatures up to
for designing API SPEC 17D subsea equipment. The methods pre- 650F. ASME BPVC, Section VIII Division 2 and ASME BPVC,
sented are acceptable for designing equipment for working pres- Section VIII Division 3 (2010) provide design methods for high-
sures up to 25,000 psi and temperatures up to 400F. pressure vessels. These design methods can be used for designing
the equipment; however, the design must meet the API SPEC 17D
Introduction design allowable limits and material and test requirements.
The industry (i.e., API and the Association of American Wellhead The ASME and API codes differ in the material, hydrostatic-test,
Equipment Manufacturers) has a long and established history in and nondestructive-examination (NDE) requirements in particular.
the development of new standards and procedures for land- and The API codes provide temperature derating factors for elevated-
platform-based equipment rated at or above 15,000 psi and 250F. temperature design. The standard hydrostatic-test requirement per
As described by Payne (2010), the original 15,000-psi well- API SPEC 17D is 1.5 times the rated working pressure, whereas in
ASME BPVC, Section VIII Division 2, the hydrostatic-test require-
ment is 1.43 times the working pressure. Additionally, the hydro-
**Now with Conoco Phillips
**Now OneSubsea LLC
static-test pressure in ASME varies between divisions and has also
fluctutated over time because of the experiences of the ASME com-
Copyright 2014 Society of Petroleum Engineers
munity. The material requirements to satisfy various codes or stan-
This paper (SPE 169813) was revised for publication from paper OTC 23928, first presented dards are also not common or aligned. The different code-specific
at the Offshore Technology Conference, Houston, 69 May 2013. Original manuscript
received for review 21 January 2013. Revised manuscript received for review 15 October
material requirements are provided in API SPEC 6A, Section 5 and
2013. Paper peer approved 20 November 2013. ASME BPVC, Section VIII Division 2, Part 3 and ASME BPVC,

April 2014 Oil and Gas Facilities 47

Section VIII Division 3, Part KM. The primary choice of materials Hydrostatic-test-design allowable-stress values will be based on
for these design recommendations will be based on the limitation API SPEC 6A, Section 4.3.3.2 and/or Section 4.3.3.3. Hydrostatic-
of API SPEC 6A, Section 5. Specific guidance is provided in this test-design allowable-stress values can also be based on ASME
paper to generate a conservative use of ASME methods with API BPVC, Section VIII Division 2, Section 4.1.6.2. Working-load-
materials. These approaches are based on the normalization of the design allowable-stress values will be based on API SPEC 6A, Sec-
hydrostatic-test pressures across the various codes. The NDE re- tion 4.3.3.2. Working-load-design allowable-stress values can also
quirements in ASME can allow for larger flaw sizes as compared be based on ASME BPVC, Section VIII Division 2, Section 4.1.6.1.
with the API allowable limits. Thermal stresses, when considered, can use secondary stress al-
lowable values on the basis of ASME BPVC, Section VIII Division
Design Method 2, Sections 5.15 and 5.5.6.1.d and API TR 6AF1, Section 5.2. The
The design method refers to the procedures outlined in Kocurek material yield strength will need to be derated on the basis of the
et al. (2012). In addition to these results, it covers the methods to working temperatures per API SPEC 6A, Annex G.
handle the high-temperature and combination of high-temperature If appropriate, in vessels that are R/t4, the classification of
and high-pressure design of subsea equipment. Components or as- stress can be performed on the basis of ASME BPVC, Section
semblies classified as pressure-containing parts per API SPEC 17D VIII Division 2, Section 5.2.2.2 and Figure 5.1, the Hopper Dia-
will be evaluated by the following techniques. In addition, equip- gram. In vessels that are R/t<4, the classification of stress should
ment not defined as pressure containing, but which experience a not be used. Stresses may still be linearized, but all results should
pressure greater than their working pressure (e.g., tubing hangers) be considered as primary membrane stresses. If stress lineariza-
during testing before deployment in the field, will be designed by tion is used, the peak-stress intensity derived through calculations
these methods. These design methods include linear and elastic shall not be allowed to exceed yield strength for more than 5%
plastic protection against plastic collapse, linear and elastic plastic of the section thickness (path distance). This recommendation is
protection against local collapse, protection against cyclic loading, based on ASME BPVC, Section VIII Division 2, Section 5.2.1.3
and closure and critical-bolting design. For protection against col- (2010). This ensures that the stress distribution through the sec-
lapse from buckling, the same design methodology as suggested in tion of the material is not so extreme as to cause local failure in
ASME BPVC, Section VIII Division 2, 5.4 is recommended. a portion of the section before full section failure might occur.
To ensure that all modes of failure are addressed, three load The stress-linearization procedures should be aligned with ASME
cases are considered. A hydrostatic load case with a test-pressure BPVC, Section VIII Division 2, Annex 5.A. Thermal loads, when
value of at least 1.5 times the rated working pressure is consid- considered in the analysis, can be directly classified as sec-
ered per API SPEC 17D requirements. Additionally, a working ondary stresses regardless of the R/t ratio. In many scenarios, this
load case with appropriate load scenarios and temperature (radial will lead to a nonconservative load case, and, hence, a load case
wall thermal gradients) distributions, as given by the design code with no thermal loads will also need to be evaluated to ensure de-
used for c alculations, is addressed. A reference table is supplied in sign acceptability.
ASME BPVC Section VIII, Division 2, Table 5.1. Last, a maximum Generally, the analyst must follow the requirements of ASME
load case for each component or assembly is evaluated for stress BPVC, Section VIII Division 2, Part 5, with the exception that
and structural stability. Along with each of the three load cases stress intensity shall be used instead of equivalent stress. Although
listed, each part must take into consideration any fatigue effects ASME BPVC, Section VIII Division 2, Part 5 calculates stresses
on the design. If a part falls under the designation of heavy section through equivalent-stress equations (e.g., von Mises stress), it is
bodies (ratio of inner radius to thickness R/t<4), special require- recommended to continue to use stress intensity to satisfy API de-
ments are outlined under the respective design methods. Worst sign requirements. FEA should use a small-displacement theory.
material-tolerance conditions shall be used when evaluating struc- Thermal loading or thermal results can be calculated separately
tural capacities. To check for proper functionality of the equipment, or together with the structural model, depending on the particu-
the deflection of systems under load must be accounted for during lars of the problem. ASME BPVC, Section VIII Division 2, Annex
the calculation of capacity. As consistent with API, there shall be 5.A established the preferred linearization procedure and path
no derating of material strengths for equipment in use up to 250F. methodology. After evaluating the component or assembly by use
For anything greater than 250F, the material strength should be de- of linear elastic criteria for protection against general plastic col-
rated per the factors in API SPEC 6A, Annex G. Fracture-toughness lapse, further evaluations may be addressed to protect against other
and fatigue-related data need to be evaluated for elevated tempera- failuremodes.
tures because these are not readily available in the codes. Elastic-Plastic Method. The elastic plastic method can be used
to evaluate components and interactions for protection against
Protection Against Plastic Collapse. This design method address- plastic collapse. Generally, the analyst must follow the require-
es general plastic collapse in equipment because of gross distortion ments of ASME BPVC, Section VIII Division 2, Part 5.
across its membrane as a result of applied loads. This failure mode The von Mises yielding criterion and associated flow rules will
can be analyzed by use of both linear methods and elastic-plastic be used. Material data will be reported as true stress-true strain. It
analysis methods. is preferred to use actual-material test data for the elastic plastic
Linear Methods. The linear methods include stress-analysis methods, but a standard (and conservative) estimation of material
methods that use handbook solutions or other industry-accepted behavior is given in ASME BPVC, Section VIII Division 2, Part 3.
methods, as long as these solutions represent the component geo When temperatures greater than 250F are considered, derated ma-
metry and loading conditions appropriately. For temperatures terial yield strength can be used per API SPEC 6A, Annex G and
greater than 250F, thermal analysis shall be conducted to take into curves can be estimated per ASME BPVC, Section VIII Division 2,
effect the temperature distributions and thermal strains. Numerical- Part 3. A load and resistance factor design (LRFD), as documented
analysis techniques, such as finite-element analysis (FEA), shall be in ASME BPVC, Section VIII, Table 5.5, shall be used to determine
used to determine the stresses in equipment for which handbook so- if a structure is suitable for operation at a specific load case. How-
lutions cannot be applied appropriately. The calculation of stresses ever, given that the typical hydrostatic-test pressure established by
will be based on the stress-intensity equation on the basis of the ASME BPVC, Section VIII, Part 8 is a minimum of 1.43 times the
Tresca yield surface and maximum-shear-stress theory. Thermal design pressure (maximum allowable working pressure), this is in
stresses will be classified as secondary stresses per ASME BPVC, conflict with the standard API hydrostatic-test case. Therefore, it is
Section VIII Division 2, Section 5.12, paragraph 19 and per API TR recommended that the LRFD factors be increased by the ratio es-
6AF1, Section 5.2. tablished here:

48 Oil and Gas Facilities April 2014

 (1.50/1.43)=1.049, ..................................................................(1)
fracture-mechanics approach to design shall be used. ASME BPVC,
where 1.50 is the API hydrostatic-test multiplier, 1.43 is the ASME Section VIII Division 3, KD-4 provides the step-by-step process for
hydrostatic-test multiplier, and the factor 1.049 is the multiplier ap- the fracture-mechanics approach. It involves estimation of design
plied to increase the LRFD factors established by ASME. If the life cycles on the basis of crack propagation to critical crack depth,
capacity of a component under engineering evaluation is to be es- which is all defined in ASME BPVC, Section VIII Division 3, KD-4.
tablished by use of the elastic plastic methods, the load-case inputs ASME BPVC, Section VIII Division 3, KD-340 is a traditional
shall be increased uniformly until the solution is nonconvergent be- fatigue-analysis approach that is based on strain-life theory. It shall
cause of unbounded deformation. The loads at which nonconver- be used when welds are present as structural supports. If the as-
gence is established are then decreased by the appropriate LRFD sembly has nonintegral components in the structural load path that
factors and the value set in Eq. 1. could progressively distort through sequential load cycles, it must
As an example, if an internally pressurized vessel reached non- also be evaluated for protection against ratcheting failure. Exam-
convergence at 60,000-psi internal pressure, then the design pres- ples of this are screwed-on caps, screwed-in plugs, shear-ring clo-
sure would be established by dividing 60,000 psi by 2.40 per sures, and breech-lock closures. Other examples in the industry
ASME BPVC, Section VIII Division 2 Table 5.5, Global Criteria might include collet- or clamp-style connectors. Evaluation by use
1, and by 1.049 (per Eq. 1). This would establish the maximum al- of the procedure outlined in ASME BPVC, Section VIII Division 3,
lowable design pressure for the vessel at 23,832 psi, and an API KD-234 shall be followed.
hydrostatic-test pressure of 35,748 psi (calculated as 1.50 times the In most cyclic-load cases for pressure-controlling or -containing
design pressure). equipment with high internal-fluid temperatures, it is observed to
For temperatures greater than 250F, when considering thermal have more allowable design cycles than the same component eval-
loads because of temperature differentials, ASME BPVC, Section uated for conditions without temperatures. This is typically ob-
VIII Division 2, Table 5.5 gives LRFD of 2.1 for thermal loads and served because of compressive stresses developed at the internal
the pressure and design loads; however, it is more conservative to surfaces as a result of the temperature gradient. Hence, it is impor-
use a factor of 2.4 for the pressure and design loads and to use the tant to evaluate the component at a load case with no thermal con-
maximum possible differential temperature to which the equipment ditions to ensure design acceptabiltity.
will be subjected. If there are appertunance live loads and other
loads than just the pressure and design loads, then it would be ap- Closure-Bolting Design. Linear elastic methods outlined in API
propriate to use the equation, with LRFD of 2.1 for the design loads SPEC 17D and paralleled in ASME BPVC, Section VIII Division
and 2.7 for the appurtenance live loads. 2, Part 5.7 are recommended for closure-bolting design. In ana-
After evaluating the component or assembly for protection lyzing the stress capacity of closure bolting, the maximum allow-
against plastic collapse, further evaluations must be made to pro- able tensile stress should be no more than 83% of the materials
tect against local collapse. yield on the basis of the root area of the thread (API SPEC 6A,
4.3.4). According to API SPEC 6A, bolting stresses are to be de-
Protection Against Local Collapse. This design method addresses termined with all loads acting on the closure area, including the
local collapse in equipment in addition to protection against plas- pressure acting over the seal area, gasket loads, and any other
tic collapse, as described in the previous subsections. This failure mechanical and thermalloads. The maximum allowable stress is
mode can be analyzed with both linear methods and advanced elas- also determined in all conditions, such as working-pressure and
tic-plastic analysis methods. In general, the analyst must follow hydrostatic conditions. When thermal loads are considered for
the requirements of ASME BPVC, Section VIII Division 2, Section bolting, thermalstresses shall use secondary-stress allowable val-
5.3 to address protection againt local collapse. When temperatures ues on the basis of ASME BPVC, Section VIII Division 2, Sections
greater than 250F are considered, thermal loads also need to be 5.15 and 5.5.6.1.d. Material yield strength will need to be derated
considered when evaluating for protection against local collapse. on the basis of thefactors provided in API SPEC 6A, Annex G.
Material yield strength will need to be derated per API SPEC 6A, The change in bolt preload as a result ofthermal expansions shall
Annex G to determine the allowables. be captured duringanalysis by ensuring that the load step to ap-
Linear Method. The linear elastic method is used in addition ply the boltprloadisbeforethe load stepat which thermal loads
to the protection against plastic collapse, as described previously, are applied.
to guard against local failure. ASME BPVC, Section VIII Division Both closure and critical bolting require a preload to a high per-
2, 5.3.2 outlines this linear method. centage of the materials yield strength. Closure bolting of all API
Elastic-Plastic Method. The elastic plastic method is used 6BX and 17SS flanges is to be made up to approximately 67% of
with protection against plastic collapse to prevent local failure. the bolts material yield stress. Other studs, nuts, and bolts used on
ASME BPVC, Section VIII Division 2, 5.3.3 describes the elastic end connections of subsea equipment shall be made up to at least
plastic method used for protection against local failure. This 50% of the bolts material yield stress (API SPEC 6A, Annex D).
method of analysis is for a sequence of applied loads on the basis Structural bolting is to be made up to the manufacturers written
of the load cases discussed previously. specification. Studs, nuts, and other closure bolting for use in
subsea service are often manufactured with corrosion-inhibiting
Protection Against Cyclic Loading. After evaluation for general coatings or platings that can affect the stud-to-nut friction factor
and local collapse, further evaluation may be made to determine dramatically. The manufacturer shall document the recommended
that a failure mode generated by load fluctuations is not encoun- makeup torque for their fasteners.
tered. Examples of these failure modes are fracture formation or High-strength-alloy steel bolts, studs, and nuts shall be eval-
ratcheting of nonintegral connections. uated for cyclic operation by use of elastic stress analysis and
Components or assemblies having undergone evaluation to pre- equivalent-stress-amplitude loading established in ASME BPVC,
vent general plastic collapse shall adhere to the following criteria Section VIII Division 2, 5.5.3 and Annex 3.F. Stress concentrations
to protect against failure caused by cyclic loading. ASME BPVC, at the root of standard ASME B1.1 threads are not to be included in
Section VIII Division 2, 5.5 provides the fatigue-analysis approach. the fatigue evaluation. In cases in which loading is shared between
This is also the recommendation of API SPEC 17D, Section 5.1.3.1. components that have deformed, such as a metal gasket, the con-
ASME BPVC, Section VIII Division 3, KD-3 is a traditional fatigue- trolling stress for the fatigue evaluation will be the effective total-
analysis approach that is based on stress-life theory. It shall be used equivalent-stress amplitude. This value is defined as one-half of the
when a leak-before-burst mode of failure is established. If the effective total-equivalent-stress range calculated for each cycle in
leak-before-burst mode of failure cannot be shown to exist, then a the loading histogram.

April 2014 Oil and Gas Facilities 49

 As a safeguard to prevent fatigue failure of threaded com-
ponents, a limit shall be placed on equipment so that no regular
load encountered during working conditions causes a separation
of the clamped joint of a threaded fastening system. An exception
to this would be extreme scenarios, such as a drive-off or impact
event. Clamped faces that separate during qualification of factory-
acceptance testing shall have their fasteners either replaced or re- Flange
tensioned, as per the manufacturers written specifications.
Bolts
Nondestructive Examination. For NDE of the component, rules (Four bolts in the
from API SPEC 17D, API SPEC 6A, and API Recommended half-symmetry model)
Practice (RP) 6HT (2005) should be followed. There are no spe-
cific rules for equipment that would be subjected to tempera-
tures greater than 250F; hence, the same rules apply for high- Rigid Surface
temperature equipment. (simulating
bottom flange)
Example Analysis of an API SPEC 6A, 4-in. 20-ksi, 350F
6BX-Type Flange
To demonstrate the application of the design methods outlined in
the previous subsections, an API SPEC 6A, 4-in., 20-ksi, 6BX-type
flange is analyzed when subjected to an internal-fluid temperature
of 350F and an external-fluid temperature of 35F. The flange was
assumed to be made from F22, low-carbon-alloy steel (2 Cr. 1
Mo.) with 75-ksi yield and 95-ksi ultimate tensile strength. The Fig. 13D 180 FEA model of the 4-in., 20-ksi, 350F API 6BX
bolts were assumed to be made of high-strength steel with 105- flange with 3D bolts and rigid surface simulating the bottom
ksi yield strength and preloaded at 67% of the bolt yield strength flange.
per the recommendation in API SPEC 17D, Section 5.1.3.5. The
bolts are assumed to be open to seawater at a temperature of 35F. with and without thermal loads. When thermal conditions of in-
Thermal conductivity of 24.85 Btu/hr ft2 F was assumed for both ternal-fluid temperature of 350F and external-fluid temperature of
the flange and the bolts at all temperatures. The flange body and 35F are considered, stresses developed were considered as sec-
bolts were given a thermal coefficient of expansion of 7.3106/F ondary stresses. Secondary stress allowable of 1.52/3Yield
at all temperatures. Both the bolt and the flange are assumed to Strength for membrane stress intensity and 32/3Yield Strength
have modulus of elasticity of 29.7106 psi at 70F (room tempera- for membrane-plus-bending stress intensity were used per API TR
ture) and 35F (lower temperature) and 28.2106 psi at 350F (el- 6AF1, Section 5.2. The yield strength of the material was derated
evated temperature). A pressure-tension-bending moment (P-T-B) on the basis of factors in API SPEC 6A, Annex G for temperature of
capacity chart, similar to that in API TR 6AF1 (1998) and API TR 350F for F22. When temperature was not considered, the working
6AF2 (2010), is generated by use of the design method of protec- allowables of 2/3Yield Strength for membrane stress intensity
tion against plastic collapse by linear elastic analysis and elastic and 1.52/3Yield Strength for membrane-plus-bending stress
plastic analysis. Protection against cylic loading is demonstrated intensity were used. Because the bolts are always at a temperature
by evaluating the design cycles for extreme load cases on the P-T-B lower than 250F, the allowable stress of 0.83Yield Strength was
chart by both fatigue and fracture-mechanics analyses. used for tensile stress and bolts were not derated for yield strength.
Elastic-plastic FEA was conducted with the same meshed model.
Protection Against Plastic Collapse. FEA was conducted by use The true-stress/true-strain curves with minimum yield strength for
of commercial software Abaqus 6.12. A 3D, 180 model of the 70 and 35F and derated material yield strengths for elevated tem-
flange with 3D bolts and a rigid bottom surface was used for both perature of 350F for the flange were calculated with ASME BPVC,
linear and elastic plastic analysis, as shown in Fig. 1. The gasket Section VIII Division 2, Annex 3.D and ASME BPVC, Section VIII
was not included in the 3D model to avoid numerical instability Division 3 KD-231.4. Linear elastic properties were used for the
resulting in premature failure of the model and because previously bolts. To respresent test conditions, the internal pressure was deter-
conducted linear elastic methods indicated that gasket groove was mined at the point at which the model failed to converge, which is
not an area of concern. The serviceability of the gasket is deter- called the plastic-collapse load. LRFD from ASME BPVC, Section
mined at the maximum working condition on the basis of contact VIII Division 2, Table 5.5 (hydrostatic-test conditions) was used
stresses, and can be validated by performance testing on the basis with the suggested hydrostatic-test factor of 1.049 (per Eq. 1) to de-
of API SPEC 17D test methods. A predetermined gasket load by termine the maximum allowable working pressure as 30 ksi. This is
use of axisymmetric FEA was applied in the model to simulate the observed to be the same as linear elastic analysis methods.
gasket for all analyses performed. Two flange-failure criteria were Elastic-plastic FEA was also conducted for working conditions
considered for the analyses: flange structural limit and bolt struc- to determine the P-T-B capacity chart. A thermal-gradient equiva-
tural limit. lent to the elastic analysis was applied to the model just after the
Linear elastic FEA for the hydrostatic-test pressure-load case bolt preload. Then, a known external pressure and tension multi-
was evaluated for the worst stress-classification line (linearized plied by the LRFD factor of 2.4 times the hydrotest factor were
path) for membrane stress intensity and membrane-plus-bending applied to the model. The bending moment that causes noncon-
stress intensity. The membrane stress intensity was limited to 83% vergence of the FEA because of structural instability was deter-
of the yield strength and the membrane-plus-bending stress inten- mined for that particular load case. Several cases were run, and
sity was limited to the yield strength. The maximum structural ca- results were compiled to determine the P-T-B chart for the flange.
pacity of the flange was determined to be 30 ksi, limited by the For elastic plastic analysis, LRFD of 2.4 times the hydrotest factor
hub-neck area. The overall capacity of the flange was limited by from ASME BPVC, Section VIII Division 2, Table 5.5, Global Cri-
the bolts and was 26 ksi. teria 1 was used for all loads applied (P-T-B).
Linear elastic FEA was conducted for the working conditions, The treatment of thermal stress was examined for ASME BPVC,
which involved the determination of the P-T-B charts for the flange Section VIII Division 2, Table 5.5 Global Criteria 1 and Global

50 Oil and Gas Facilities April 2014

 20
Bolt Bolt Bolt Bolt
T=0 kips

T=200 kips

T=400 kips
Bolt Bolt Bolt Bolt
15
T=600 kips

Bore Pressure, ksi T=External

tension in kips
Bolt Bolt Bolt Bolt
10

Bolt Bolt Bolt Bolt

5

Bolt Bolt Bolt Bolt

0
0 50 100 150 200

Bending Moment, ft-kips

Fig. 2Overall P-T-B capacity of the 4-in., 20-ksi flange at 350F by use of linear elastic- and elastic-plastic analysis methods, con-
sidering both bolt and flange structural limits.

Criteria 2. The maximum differential temperatures (maximum teria2). The external loads were not considered as live-appurte-
thermal stresses) experienced by the component at the input condi- nance loads, and, hence, were not used with the LRFD of 2.7. Two
tions were applied to the model. These thermal stresses were com- charts have been presented in the paper. Fig. 2 shows the P-T-B
bined with the 2.4 LRFD (Global Criteria 1) structural loads. This chart for the overall capacity of the flange, taking into account
criterion was found to provide a more-conservative result than both flange and bolt structural limits, with both elastic-plastic- and
the 2.1 LRFD for all loads, including thermal loads (Global Cri- linear elastic analysis methods. Fig. 3 shows the P-T-B chart con-

20

T=0, LE
T=200, LE
T=400, LE
15
T=600, LE
T=0, EP
Bore Pressure, ksi

T=200, EP
T=400, EP
10 T=600, EP

T = External tension in kips

LE = Linear elasc analysis
EP= Elasc-plasc analysis
0
0 50 100 150 200 250 300 350 400

Bending Moment, ft-kips

Fig. 3P-T-B capacity of the 4-in., 20-ksi flange at 350F to compare the linear elastic and elastic-plastic analysis methods, consid-
ering only flange structural capacity.

April 2014 Oil and Gas Facilities 51

 20

560 cycles 214 cycles

T=0 kips

T=200 kips
15
T=400 kips

Bore Pressure, ksi

T=600 kips

10

226 cycles 247 cycles

0
100 150 200 250 300 350 400

Bending Moment, ft-kips

Fig. 4Design cycles by use of fatigue analysis for identified load cases of the P-T-B- capacity chart at 350F, generated with linear
elastic anlaysis for flange structural capacity.

sidering only the flange structural limit for both elastic plastic and slightly lower capacities. In contrasting these two methodologies,
linear elastic methods for comparison. API TR 6AF1 considers an axisymmetric FEA model to determine
It is observed from Fig. 2 that the capacity of the flange is lim- the P-T-B chart for this flange. The bolts are evaluated by use of an
ited by linear elastic analysis for all the load cases considered axisymmetric model, with the bolt stresses averaged for all eight
on the basis of bolt failure. The chart is comparable to the API bolts. The results reported in Fig. 2 use a 3D model to examine
TR 6AF1 charts for the 4-in., 20 ksi, 6BX-type flange and shows the stresses in each of the bolts. With external bending applied, the

20

1588 cycles 549 cycles

T=0 kips

T=200 kips
15
T=400 kips
Bore Pressure, ksi

T=600 kips

10

557 cycles 2013 cycles

0
100 150 200 250 300 350 400
Bending Moment, ft-kips

Fig. 5Design cycles by use of fracture-mechanics analysis for extreme load cases of the P-T-B capacity chart at 350F, generated
with linear elastic analysis.

52 Oil and Gas Facilities April 2014

outermost bolts reach the maximum allowable stress before other crack on the outer surface, with a radial depth, yielded the most-
bolts. This is the reason that the capacity observed is slightly less conservative result. For the other three load cases identified, the
than that in the API TR 6AF1 charts. flange was loaded in such a way that the stresses were predom-
As seen in Fig. 3, the elastic plastic method generates a more- inantly axial stresses, and, hence, a circumferentially oriented
conservative result than the linear elastic methods. This is because crack on the outer surface, with a radial depth, yielded the most-
for linear elastic anlaysis, the thermal loads are categorized as sec- conservative results. This places the largest initial crack perpen-
ondary stresses, whereas for elastic plastic analysis, they have dicular to the worst-loaded plane, and initializes crack propagation
been considered as design loads per ASME BPVC, Section VIII at the highest stress level in that plane. The crack was assumed to
Division2, Table 5.5, Global Criteria 1. Secondary stresses allow be semielliptical in shape, with a ratio of depth-to-surface length of
higher design allowables than the general-stress allowables, and 1:3 per ASME BPVC, Section VIII Division 3, paragraph KD-411.
the LRFD of 2.4 plus thermal stresses provide a conservative re- This initial flaw is placed along the failure path determined from
sult. However, as observed from Figs. 2 and 3, the overall capacity the structural linear elastic analyses. An initial flaw length of 3/16in.
of the flange is still limited by the linear elastic analysis on the basis was used for the analyses. Crack-face pressure was included in
of bolt-failure criteria. the evaluations when internal pressure was considered. Commer-
cial software SIGNAL Fitness-For-Service version 4.0 by Quest
Protection Against Cyclic Loading. To demonstrate the design Reliability was used for the analysis. The allowable final-crack-
methods for cyclic loading, the structural capacity of the flange was depth criteria per ASME BPVC, Section VIII Division 3, paragraph
evaluated for extreme load cases from the P-T-B chart generated KD-412 were used to determine the allowable cycles for a partic-
for flange stress capacity by use of linear elastic analysis. It was ular load case. Fig. 5 shows the allowable design cycles by use of
assumed that for one cycle, all the loads applied after bolt preload fracture-mechanics analysis for the load cases considered.
went from zero to the actual value and again went back to zero.
For example, for a load case of 20,000-psi internal pressure, 600- Validation of Design Methods
kips external tension, 120-kips-ft external bending moment, and Validation of the example analysis can be demonstrated by com-
temperature of 350F, one cycle of loading would be zero loads to paring the capacities with the published combined-load-capacity
20,000-psi internal pressure, 600-kips external tension, 120-kipsft charts for the 4-in., 20-ksi, 6BX-type flange in API TR 6AF1. Sim-
external bending moment, and temperature of 350F and back to ilar capacity ratings were obtained when analyzed with these design
zero loads. Design cycles were calculated for these cases by use of methods. API Report 86-21 (1987) demonstrates the validation of
fatigue analysis and fracture-mechanics analysis per ASME BPVC, the flange-capacity charts by testing a 71/16-in., 10-ksi, 6BX-type
Section VIII Division 3. flange for combined loadings. Cameron International Corporation
The flange under consideration is assumed to be made from (Cameron) has also successfully tested an 11-in., 25ksi flange for
F22, low-carbon-alloy steel with 75,000-psi yield strength and hydrostatic test pressure of 37.5 ksi with the flange and bolts strain
95,000psi ultimate tensile strength. Published material properties gauged, which provided accurate correlation to predicted values
have been used to determine the design cycles for both fatigue and from the finite-element-model analyses. Cameron has successfully
fracture-mechanics analysis. tested a 29/16-in., 25-ksi, Cameron Gate Valve for cyclic loading
Fatigue Analysis. Fatigue analysis was conducted per ASME from 0 to 30 ksi for 500 cycles, validating no change in preload
BPVC, Section VIII Division 3, KD-3. Because ASME BPVC, Sec- of the bonnet and end flanges. Additionally, post-test NDE inspec-
tion VIII Division 3, Table KD-320.1 provides the material-fatigue tions were performed to validate that no fatigue damage was expe-
curve for low-carbon-alloy steel for ultimate tensile strength of rienced by the valve body as a result of crack growth. This same
90,000 psi, this curve was used for fatigue analysis. Half of the valve was also tested successfully for a hydrostatic test pressure of
maximum stress intensity from the linear stress-analysis results 37.5 ksi. The bonnet, bolts, and flanges were strain gauged, which
was used for the equivalent stress amplitude, and the number of provided accurate correlation to predicted values from the finite-
design cycles was calculated. The allowable cycles were taken as element-model analyses. These test data demonstrate the validation
one-tenth of the number of cycles calculated by use of the formula- of the design methodology outlined in this paper.
tion from ASME BPVC, Section VIII Division 3. Fig. 4 shows the
design cycles for the extreme load cases considered. Application
Fracture-Mechanics Analysis. Fracture-mechanics analysis Two primary design methods have been presented and con-
was conducted per ASME BPVC, Section VIII Division 3, KD-4. For trasted in the preceding sections, with both of the results approxi-
material properties, the fracture-toughness value of 138 ksi in. mating each other for the overall capacity. During the writing of
was used from the published report by Young (2008), which cor- this paper and while analyzing the example flange, a senior ex-
responds to the seawater-with-cathodic-protection environment perienced analyst was used for approximately 2 months for ana-
and was determined per ASME BPVC, Section VIII Division 3 by lyzing, post-processing, verification, and documentation. During
the crack-tip-opening displacement method. Recommended crack- this period, approximately 8 days were spent on the linear elastic
growth-rate factors for steel in a marine environment with cathodic and fatigue portion of the work. Approximatly 60 days were spent
protection were used from BS 7910 (2005), as listed in Eq. 2 in the on the elastic plastic and fracture-mechanics portion of the work.
form of the Paris-law equation: With both methods providing approximately similar results in both
overall capacity and cycle life, the primary difference was a 650%
da in.
( ) increase in manpower needed to perform the elastic plastic and
1.4
= 5.22 107 K ksi in. . ..................................(2)
dN cycle fracture-mechanics analyses. Even with the analysts experience,
the elastic-plastic analysis results were often subjected to numerical
For the crack orientation, both circumferential and axial crack instability, boundary-condition influences, and interpretation of a
orientations were considered. Crack locations on both the inside point of nonconvergence as a result of unbounded d eformation.
surface and the outside surface were based on the stress distribu- These results must be accounted for during the planning of a
tion. This was needed because of the compressive thermal stresses work package in which elastic plastic and fracture-mechanics
on the inside surface, which prevent crack propagation. The analyses are required or anticipated. Staffing needs and experi-
thermal stresses were considered as secondary stresses during cal- ence levels must be planned, and the results/capacities must be re-
culations. For the load case with zero internal pressure and zero ex- viewed. With the proliferation of FEA software and easy-to-use
ternal tension, the flange was loaded in such a way that the stresses programs, the generation of results has become trivial. However,
were predominantly hoop stresses. Therefore, an axially oriented the establishment of the problem (boundary condition) and the in-

April 2014 Oil and Gas Facilities 53

terpretation of the results (nonconvergance) are still extremely dif- BPVC to achieve working pressures greater than 15,000 psi and
ficult. The burden of establishing if a component is safe for use temperatures greater than 250F because there is not a published
shifts from the design engineer to the analyst, resulting in a historic opinion. The intention of this work is to show that existing API
reversal of roles. methods are adequate for working pressures up to 25,000 psi and
Given the geometries associated with pressure-containing temperatures up to 400F. The 400F limit is currently based on the
equipment rated for more than 15,000 psi, the classification and material-data availability. The recommendations presented in this
treatment of stress needs to be evaluated carefully. In many cases, paper are not meant to act as a new standard; an in-depth analysis
the elastic plastic methods presented previously are the only ap- should be performed on each component or assembly to verify that
propriate tools to use for the definition of component capacity. the proper requirements are met.
This is because stress linearization is not appropriate for heavy- The methods published in this paper are limited to API Subcom-
walled vessels, especially around local or discontinuous regions. It mitee 17, Subsea Production Systems, but these methods can be ex-
is also seen that when large thermal gradients (stresses) are classi- tended to other API subcommitees by use of similar methodology
fied as secondary stresses, the capacity charts indicate the increase and changing the factors on the basis of respective applicable de-
in capacities because of higher design allowables for the secondary sign codes subject to sufficient verification and validation.
stresses. However, it should also be noted that for the evaluation
of more-standard load cases (e.g., P-T-B with no thermal consid- Acknowledgments
erations), the linear elastic methods provide equally conservative The authors would like to extend special thanks to Stuart Harbert,
results. Both methods can be used together to provide efficient an- Larry Earles, and Paul Bunch from Cameron International Corpo-
swers, with linear elastic methods being used initially and elastic ration for their valuable input regarding analysis by use of these
plastic methods used to justify the linear elastic assumptions. design methods.
Reviewing fatigue and fracture-mechanics analyses for cyclic
loading, both methods provide approximately equal results for al- References
lowable design cycles. The primary contrast between the methods API RP 6HT, Heat Treatment and Testing of Carbon and Low Alloy Steel
lies in the material-data requirements. Fracture mechanics requires Large Cross Section and Critical Section Components. 2005. Wash-
more material testing and characterization, even between the same ington, DC: API.
material batches, while fatigue analysis relies on more generally API SPEC 17D, Design and Operation of Subsea Production Systems
accepted material assumptions with larger safety factors. Both Subsea Wellhead and Tree Equipment, second edition. 2011. Wash-
methods are well posed to analyze ductile materials, which are re- ington, DC: API.
quired for pressure-containing subsea equipment. API SPEC 6A, Specification for Wellhead and Christmas Tree Equipment,
20th edition. 2010. Washington, DC: API.
Conclusions API TR 6AF1, Technical Report on Temperature Derating on API Flanges
Several design-analysis methods are given in this paper to verify Under Combination of Loading, second edition. 1998. Washington,
the capacity of components or assemblies under consideration, in- DC: API.
cluding protection against plastic collapse, protection against local API TR 6AF2, Technical Report on Capabilities of API Integral Flanges
collapse, and protection against cyclic loading. An approach to clo- Under Combination of LoadingPhase II, fourth edition. 2010.
sure bolting is also presented. Both linear elastic and elastic plastic Washington, DC: API.
methods are given for the protection against local and plastic col- API REPORT 86-21, API Research Report Project 86-21, Capabilities
lapse. Using these methodologies, an example design analysis was of API Flanges Under Combination of Loading. 1987. Washington,
presented for a 4-in., 20-ksi API 6BX flange at 350F. DC: API.
The linear elastic and elastic plastic methods of analyzing the ASME Boiler Pressure and Vessel Code, Section VIII Division 2Alter-
sample flange gave the same pressure rating when only internal native Rules for Construction of Pressure Vessels. 2010. New York:
pressure load was applied. The overall capacity of the flange for hy- ASME.
drostatic-test condition is limited by bolt, which is analyzed using ASME Boiler Pressure and Vessel Code, Section VIII Division 3Alterna-
linear elastic methods. When external loads were used to determine tive Rules of Construction of High Pressure Vessels. 2010. New York:
the combined structural P-T-B capacity of the flange at 350F, it ASME.
was observed that the linear elastic methods gave the most-conser- BS 7910:2005. Guide to methods for assessing the acceptability of flaws in
vative results. When examining only the capacity of the flange and metallic structures. 2005. London: British Standards Institution (BSI).
allowing the bolt stresses to exceed allowable values, the elastic Payne, M. 2010. HP/HT Challenges. J Pet Technol 62 (10): 70.
plastic method gave more-conservative results than the linear Kocurek, C.G., Pathak, P., Melancon, C. et al. 2012. Merging ASME
elastic methods. This was based on two assumpations: the use of and API Design Methods for Subsea Equipment Up To 25,000 PSI
stress categorization in linear methods of thermal stresses as sec- Working Pressure. Presented at the Offshore Technology Confer-
ondary stresses and the choice of ASME BPVC, Section VIII Di- ence, Houston, 30 April-3 May. OTC-23063-MS. http://dx.doi.
vision 2, Part 5, Global Criteria 1, which conservatively applied org/10.4043/23063-MS.
thermal loading in the elastic plastic methods. The overall capacity Young, K. 2008. Characterizing Material Performance for Design of High-
chart reported is comparable with the API TR 6AF1 chart and veri- Pressure High-Temperature (HPHT) Equipment in Accordance with
fies the results obtained. API RP 6HP Practices. Technical Report PB2011-105599, Contract
Fatigue and fracture-mechanics analyses for cyclic loading are No. MMS MPO7PC13016, Minerals Management Service (MMS),
reported for the identified load cases on the P-T-B chart at 350F. Mohr Engineering Division of Stress Engineering Services (SES),
The reported cycles show that the flange design has a cyclic- Houston, Texas (March 2008).
loading capacity that meets the standard cycle requirements (200
cycles) for subsea equipment, but should be carefully evaluated for Parth D. Pathak is a new-product-development engineer in the Subsea
higher-cycle applications. Trees Group at OneSubsea LLC and is also involved in representing
OneSubsea and Cameron on the API 17TR8 industry work group.
Limitations of Work. The work and results presented in this paper He worked for Cameron for 1 year on an engineering-graduate rota-
are not meant to provide a definitive solution to building equipment tional program, which involved working in multiple oilfield-equipment
for working pressures greater than 15,000 psi and temperatures groups. Previously, Pathak worked for Mercedes-Benz as an assembly-
greater than 250F. The purpose of this paper is to show that it is line engineer. He has coauthored two published OTC conference pa-
possible to combine the methods of both API SPEC 17D and ASME pers involving HP/HT design methods. Pathak holds an MS degree in

54 Oil and Gas Facilities April 2014

mechanical engineering, specializing in FEA, from Ohio State Univer- career with Cameron as an apprentice draftsman. His research focus
sity and a BE degree in mechanical engineering from the University of is on seals and sealing technologies for oilfield environments. Ko-
Pune, India. curek holdsaBSdegree inmechanical engineering from the Univer-
sity ofHouston.
Christopher G. Kocurek is currently with Conoco Phillips. Previously,
he was the Subsea Tree Product Manager for OneSubsea LLC, man- Samuel L. Taylor is a mechanical engineer with OneSubsea LLC. He
aging the new-product-development program and representing One- currently works on the design and implementation of subsea gate
Subsea LLC on various API industry committees. His other roles at valvesinprojects all over the world and has worked with all manner of
OneSubsea LLC have included engineering manager, research spe- subseaoilfield equipment during his tenure at OneSubsea. Taylorholdsa
cialist, and designer. Before moving to OneSubsea, K
ocurek began his BSdegree in mechanical engineering from Ohio State University.

April 2014 Oil and Gas Facilities 55