Proceedings of the ASME 2016 Pressure Vessels and Piping Conference

PVP2016
July 17-21, 2016, Vancouver, British Columbia, Canada

PVP2016-63759

Weld Joint Efficiency in Design By Analysis

Trevor G. Seipp
Becht Engineering Canada Ltd.
4720-106th Ave SE, Suite 210A
Calgary, AB CANADA
Tel: 403-668-7274
Email: tseipp@becht.com

ABSTRACT merely a numerical quantity used in the design of a

In the original ASME Section VIII, Division 2, no welded joint as a multiplier on the allowable stress. It is a
consideration was given to partial weld joint efficiencies function of material test group, NDE method and extent of
(values of the factor E less than 1.0) because that version examination, wall thickness, welding process and service
required full radiography and only permitted weld joint temperature. However, ASME PTB-1 provides no basis
efficiencies of unity. In the new (post-2007) Section VIII, for the numerical quantities selected by the Code
Division 2, partial weld joint efficiencies as small as 0.85 Committee.
are now permitted. Furthermore, much Design By In Section VIII, Division 1 [1] partial weld joint
Analysis work is performed on vessels fabricated to efficiencies as low as 0.45 are permitted, as presented in
ASME Section VIII, Division 1 and the ASME B31 Codes, Table UW-12. However, due to the variety of geometries
which all permit partial weld joint efficiencies. However, covered by Table UW-12, there may be some inherent
no guidance is provided on how to account for these geometric factors in addition to weld quality factors
values in Deign By Analysis to ASME Section VIII, associated with the extent of NDE. Therefore, to simplify
Division 2, Part 5. the discussion, this paper will focus solely on full
This paper provides the technical justification for the penetration butt welds/joints.
proposed changes to ASME Section VIII, Division 2, Part Frith and Stone [5] provide a very good history of weld
5 and API RP-579/ASME FFS-1 regarding weld joint joint efficiency factors. They also try to examine weld joint
efficiency. Guidance is also provided on how to efficiencies from the perspective of probability of
incorporate this change into ASME Section VIII, Division identification (POI) of defects in welds.
1 by way of U-2(g) and the B31 Codes. While there is ample usage of weld joint efficiencies
in Design-By-Rules, there exists a dearth of justification
INTRODUCTION for their values. This lack of justification leads to a
The original Criteria of the ASME Boiler and Pressure difficulty when trying to implement a Design-By-Analysis
Vessel Code for Design By Analysis in Sections III and approach that implements weld joint efficiency.
VIII, Division 2 [4], (hereafter the Criteria Document) was
silent on managing weld joint efficiencies because both STATEMENT ON WELD JOINT EFFICIENCY
Codes required full inspection to ensure a joint efficiency For the purposes of this paper and this approach to
factor of 1.0. However, since the 2007 Edition of ASME implementing weld joint efficiency in Design-By-Analysis,
section VIII, Division 2 [2], weld joint efficiencies less than we make the following foundational statement:
1.0 (hereafter called partial weld joint efficiencies) down
to 0.85 are now permitted. The weld joint efficiency less than 1.0, as
This partial weld joint efficiency is used to account for described in the various ASME pressure vessel
volumetric inspection (RT – radiographic examination or and piping Codes, is intended to indicate, for the
UT – ultrasonic examination) of the welds less than 100% purposes of mechanical strength, that a welded
of their total length, as described by Table 7.2. As structure is weaker than a non-welded structure.
described in ASME PTB-1 [3], the weld joint efficiency is

1 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/asmep/90481/ on 01/29/2017 Terms of Use: http://www.asme.org/abo

 FAILURE MODES LIMIT-LOAD ANALYSIS METHOD
From this foundational statement, it can stated that As a logical extension to the Elastic Stress Analysis
weld joint efficiency is only considered when determining Method, we propose that the weld volume, plus 1*t
Protection Against Plastic Collapse. The effects of welds adjacent to the weld, have a plastic limit set at 1.5SE. In
on Protection Against Failure From Cyclic Loading: this way, the plastic limit of the weld (effectively its
Fatigue is a completely separate topic, and one that is strength) is now multiplied by the weld joint efficiency
adequately handled already by ASME Section VIII, knockdown factor for that weld.
Division 2. Since weld joint efficiency affects mechanical
strength, it will have minimal-to-no effect on the structural ELASTIC-PLASTIC ANALYSIS METHOD
stability of pressure equipment. Accordingly, weld joint The next incremental expansion of this concept is for
efficiencies are not considered in determining Protection the elastic-plastic analysis method. The sole question is:
Against Collapse From Buckling (also note that the how to de-rate the strength of an elastic-plastic material?
Design-By-Rules in Division 1 and 2 also neglect weld We propose that, identical to the Limit-Load Analysis
joint efficiencies for buckling failure). Similarly, since Method, the weld volume plus 1*t adjacent to the weld be
Protection Against Local Failure is not governed by modeled as a weaker material. That weaker material is
mechanical strength, weld joint efficiency does not need calculated in the same manner as the non-weld material:
to be considered for this failure mode. using Appendix 3-D, however, both the yield strength and
the ultimate tensile strength shall be multiplied by the weld
GENERAL APPROACH joint efficiency factor for that weld.
The general approach to implementing partial weld
joint efficiencies is that the weld, itself, is what makes the IMPLEMENTATION
welded structure weaker than a non-welded structure. It One sample problem is presented to demonstrate the
is, in effect, as knock-down factor on the strength of the implementation and the consequences of the
weld. implementation.
In the next three sections, the implementation of this
general approach will be shown for the three different
Sample Problem E=0.85
analysis method for demonstrating Protection Against
The first sample problem is a simple cylinder with an
Plastic Collapse: Elastic Stress Analysis Method, Limit-
OD of 120 in (3048 mm), 1 in (25.4 mm) thick and 125 in
Load Analysis Method, and Elastic-Plastic Analysis
(3175 mm) long. A material of SA-516 Gr. 70 at 70°F
Method.
(21°C) was used, resulting in an allowable stress from
Table 5A of ASME Section II, Part D [3] of 25.3 ksi (175
ELASTIC STRESS ANALYSIS METHOD MPa). A design internal pressure of 485 psi (3.344 MPa)
Evaluating stresses from an elastic analysis using this was applied. The model is shown in Figure 1, with the
method involves creating Stress Classification Lines plate shown in green and the weld shown in grey.
(SCLs). Our proposal is that if the SCL is located at or
adjacent to (less than 1*t away from) a welded joint, then
the allowable stress shall be multiplied by the weld joint
efficiency factor for that weld. The intent of the 1*t
adjacency limit is that, for thin vessels, this distance is
typically the extent of the deposited weld metal and the
weld heat-affected zone. For thicker vessels, and/or
vessels fabricated using a narrow-gap welding technique,
this adjacency limit may be adjusted.
This implements the general approach by penalizing
only the weld (and the adjacent heat affected zone) where
flaws not detected may exist. This implementation is
consistent with the existing Design-By-Rules regarding
reinforcement of openings. In those rules, provided that
the opening does not go through a welded joint, credit can
be taken for the full wall thickness of the vessel shell,
despite the fact that the wall thickness may have been
calculated using a partial weld joint efficiency – see ASME Figure 1. Finite element model of sample problem
Section VIII, Division 1 [1], paragraph UG-37.
The finite element mesh is shown in Figure 2, while a
close-up of the mesh of the weld is shown in Figure 3.

2 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/asmep/90481/ on 01/29/2017 Terms of Use: http://www.asme.org/abo

 The model was meshed with quadratic 3D-solid elements sole material property input. Since there is no
of type C3D20R. Five elements were used in the through- discontinuity at the weld, the membrane stresses are
thickness direction resulting in an element size of 0.2 in classified as Pm. Accordingly, the allowable stress for the
(5.1 mm). In the weld, an element size of 0.2 in (5.1 mm) plate is 25.3 ksi (175 MPa). Using a weld joint efficiency
was also used in the circumferential/hoop direction. Away factor of 0.85, the allowable stress for the plate would be
from the weld, the size in the circumferential direction 21.505 ksi (148.75 MPa).
increases to 3 in (76.2 mm). Since very little variation in For a SCL whether in the weld or not in the weld, the
the longitudinal direction was expected, an element size calculated membrane equivalent stress is 24990 psi
of 3 in (76.2 mm) was used. Based on the results of a (172.3 MPa). Therefore, the plate material is acceptable
discretization study, the results for this mesh have an but the weld material is not, and the maximum internal
estimated discretization error of less than 2%. pressure would need to be reduced from 485 psi (3.344
MPa) to 417.4 psi (2.878 MPa).
For the limit load analysis, the yield stress for the plate
material was set at 38.0 ksi (262.0 MPa), while the yield
stress for the weld material was set at 32.3 ksi (222.7
MPa). The limit load analysis was performed, and the
maximum permissible internal pressure was calculated to
be 418.07 psi (2.882 MPa), or essentially the same result
as the elastic analysis.
For the elastic-plastic analysis, the true stress-true
strain curve for the plate material was calculated using the
procedure in ASME Section VIII, Division 2 [2], Annex 3-
D, using an engineering yield strength of 38 ksi (262 MPa)
and an engineering ultimate strength of 70 ksi (482.6
MPa). The true stress-true strain curve for the weld
material used these engineering yield and ultimate values
multiplied by 0.85. The resulting true stress-true strain
curves are shown in Figure 2.

Figure 2. Finite element mesh

100000

80000
True stress (psi)

60000

40000

20000

0
0 0.05 0.1 0.15 0.2 0.25 0.3
True Total Strain
Figure 3. Close-up of the finite element mesh of the weld
Figure 4. True stress-true strain curves
The applied loads are the internal pressure, which is
applied to the interior surface, and a pressure thrust,
The elastic-plastic analysis was performed, and the
which is applied as a negative pressure to the “near” end maximum permissible internal pressure was calculated to
of the cylinder, as shown in Figure 2. For the boundary be 464.6 psi (3.203 MPa).
conditions, a cylindrical coordinate system was
The results of these three different analyses are
generated; the “far” end of the cylinder, as shown in
summarized in Table 1.
Figure 2, was restrained in the R and Z directions, while
the “near end was restrained only in the R direction.
For the elastic analysis, the material properties of the
plate and weld are identical, as Young’s Modulus is the

3 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/asmep/90481/ on 01/29/2017 Terms of Use: http://www.asme.org/abo

 Table 1. Summary of results design-by-analysis rules in ASME Section VIII, Division 2.
Analysis Type Design Reduction For ASME Section VIII, Division 1, paragraph U-2(g)
Pressure (psi) Factor should be modified to indicate that the weld joint
Elastic 417.4 0.850 efficiencies for full-penetration butt welds calculated in
Limit Load 418.07 0.851 UW-12 shall be used when using ASME Section VIII,
Division 2, Part 5. For ASME B31.3, paragraph 304.7.2
Elastic-Plastic 464.6 0.946
(d) should be modified to indicate that the quality factor
from Table A-1B shall be used when modeling welds in
DISCUSSION ASME Section VIII, Division 2, Part 5. When modeling
Both the elastic analysis and the limit load analysis castings, the quality factor from Table A-1A shall be
method produce a resulting design internal pressure that applied to the entire material.
is 0.85 of the value that would otherwise be used for non- For other Codes and Standards, similar paragraphs
welded construction. This result is consistent with a may need to be modified.
manual calculation approach whereby the design internal
pressure is equal to the non-welded design pressure
REFERENCES
multiplied by the weld joint efficiency. Therefore, this
1. ASME, 2015, ASME Boiler and Pressure Vessel
result is entirely consistent.
Code, Section VIII, Division 1, American Society of
The elastic-plastic analysis produces a result that is
Mechanical Engineers, New York, NY
0.946 of the value that would otherwise be used for non-
welded construction. This result is less conservative than 2. ASME, 2015, ASME Boiler and Pressure Vessel
the results from the elastic and limit load analysis Code, Section VIII, Division 2, American Society of
methods, but represents the reinforcing effect of the Mechanical Engineers, New York, NY
adjacent material. However, based on the original
assumptions inherent in the approach, it is not surprising 3. ASME, 2015, ASME Boiler and Pressure Vessel
that the result is less conservative; it is generally Code, Section II, Part D, American Society of
understood that the elastic-plastic analysis method is Mechanical Engineers, New York, NY
more accurate/less conservative. Therefore, it is deemed
that this approach is acceptable. 4. Osage, David A, Sowinski, James C., ASME PTB-1-
2014, ASME Section VIII - Division 2, Criteria and
CONCLUSION Commentary, American Society of Mechanical
Engineers, New York, NY
One of the approaches for managing partial weld joint
efficiency in design by analysis is presented, and a simple 5. Anon, 1972, Criteria of the ASME Coiler and Pressure
example problem is presented. Vessel Code For Design By Analysis in Sections III
and VIII, Division 2, American Society of Mechanical
FUTURE WORK Engineers, New York, NY
Additional sample problems will be presented in the
future to describe how these rules would be implemented 6. Frith, Robert and Stone, Mark, 2015, “Weld Efficiency
in more complex geometries and loading situations. Factors Revisited”, presented at the 14th International
Conference on Pressure Vessel Technology, Elsevier
INCORPORATING INTO THE CODE 7. API and ASME, 2007, API 579-1/ASME FFS-1
The methodology described above for implementing Fitness For Service, API Publishing Services,
the weld joint efficiency into a design-by-analysis can be Washington, DC
directly implemented into both ASME Section VIII,
Division 2 [2] and API 597-1/ASME FFS-1 [7]. In the 8. ASME, 2014, ASME Code for Pressure Piping, B31:
upcoming Edition of API 579-1/ASME FFS-1, the Annex Process Piping, American Society of Mechanical
Stress Analysis Overview for An FFS Assessment is Engineers, New York, NY
moved from Annex B1 to Annex 2D, and the paragraphs
implementing this methodology are 2D.2.5. For ASME
Section VIII, Division 2, the following paragraphs need to
be modified: 5.2.2.4, Step 5 (a), 5.2.2.4, Step 5 (b),
5.2.3.5, Step 3, and 5.2.4.4, Step 3.
For ASME Section VIII, Division 1 [1] and ASME
B31.3 [8] (and other design Codes that do not have
design-by-analysis rules), these rules can be
implemented by modifying the paragraph that refers to the

4 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/asmep/90481/ on 01/29/2017 Terms of Use: http://www.asme.org/abo