1 4945413584441049168
1 4945413584441049168
1 4945413584441049168
MPC – A Council of the WRC, the Materials Properties Council is dedicated to providing
industry with the best technology and the best data that can be obtained on the properties of
materials to help meet today’s most advanced concepts in design and service, life assessment,
fitness-for-service, and reliability and safety.
PVRC – A Council of the WRC, the goal of the Pressure Vessel Research Council is to
encourage, promote and conduct research in the field of pressure vessels and related pressure
equipment technologies, including evaluation of materials, design, fabrication, inspection and
testing.
WRC Bulletins contain final reports from projects sponsored by the Welding Research Council, important
papers presented before engineering societies and other reports of current interest.
No warranty of any kind expressed or implied, respecting of data, analyses, graphs or any other
information provided in this publication is made by the Welding Research Council, and the use of any
such information is at the user’s sole risk.
All rights are reserved and no part of this publication may be reproduced, downloaded, disseminated, or
otherwise transferred in any form or by any means, including photocopying, without the express written
consent of WRC.
ISSN 0043-2326
Library of Congress Catalog Number: 85-647116
r -
PROJECT FUNDS 1980 1982 1984 1986 1988 1990 TASK
1
Fig. 1-1—Dissimilar weld projects
a position where the materials are expected to be at phasis changed from one of validation of proposed
temperatures below that at which significant creep criteria to one of generation of new criteria.
takes place. This new emphasis was given further impetus by a
The program to evaluate the LMFBR DMW encom- program (Task 400 of Fig . 101), funded by EPRI, to
passed both testing and analysis and was funded by develop Design Criteria Strategies for Dissimilar
the Department of Energy (DOE) . A significant con- Welds in HTGRs . This project was established in 1984
clusion of this work was that the life achieved under as part of the Government-Industry HTGR program
test conditions simulating service transients and tem- and involved transfer of technology from the fossil
peratures was not predictable using state-of-the-art program to the HTGR program.
finite element methods . Although full elastic-plastic- The intention of the project was to define strategies
creep analyses were performed, magnitudes of stress for both near-term and long-term criteria develop-
and strain components calculated were found to de- ment ; and then, to implement the chosen strategies.
pend on the size of the mesh used in the model . This The major conclusion from the first phase of the pro-
conclusion is a direct result of the elastic stress singu- ject was that the best near term strategy for develop-
larity which theoretically exists at the interface for ing suitable criteria would involve the use of the con-
certain geometries of dissimilar weld . The validity of siderable data base available from the fossil program.
the predictions is questionable when the finite ele- A specific task to be carried out in the near term would
ment size approaches the grain size of the material and be to combine this service/laboratory data base with
significant material properties variations can occur the laboratory data base being generated (under DOE
over lengths of 1—100 µm. funding) for the HTGR tubing. The combining of
The HTGR work (Task 200) was initiated as a test these data would be done via PODIS and some refer-
program to validate the preliminary design criteria ence stress/strain parameter to account for the materi-
proposed in 1975 for the steam generator tubing al and geometric differences between the two weld
DMW . This program proceeded in parallel, to a cer- configurations . (See Fig . 1-2).
tain extent, with the fossil plant program and benefit- At this point, the PVRC Elevated Temperature De-
ed from it. As a result of the data generated, the em- sign Subcommittee also requested that a preliminary
Printed copies are not controlled. Printed on: 2018-05-22 (UTC)
2 WRC Bulletin 350
SERVICE FOSSIL
DATA
J
LAB TEST LIFE
PODIS
DATA PREDICTION
HTGR
ANALYTICAL
DAMAGE VERIFY
PARAMETER
and a preliminary mark up of the appropriate sections pressure-induced hoop stress but in the ab-
of N-47. sence of any other external loading. (Note:
pressure-induced axial stress is included in
2 . Description of PODIS Dp (below), not in D I .)
• Dp = Damage caused by the primary (load-con-
The metallurgical characteristics of the damage ob-
trolled) components of the "system" loads
served in both service and laboratory test samples4
on the weld . "System" loads in this context
indicate that creep rupture is the dominant DMW
are those arising at the weld location within
failure mode . However, it was also observed that tem-
the tube assembly (platen) including the ax-
perature cycling contributes significantly to damage
ial effect of pressure. Loads in this category
and can cause failure even when primary stress levels
are deadweight and axial pressure load.
are relatively low. Therefore, a creep-fatigue concept
• Ds = Damage caused by the secondary (strain-
was used as the basis for the development of PODIS . A
controlled) components of the "system"
simple linear summation philosophy was incorporated
loads on the weld . Loads in this category are
which recognizes damage caused by self loads, sus-
those due to restrained thermal expansion
tained primary stresses and cyclic secondary loads.
of the tube assembly.
Failure Modes
The total damage is then given by the summation of
The failure modes of concern were studied in the
the values of all three damage components.
fossil-fueled plant program 4 and all involved the de-
velopment of creep damage in the heat-affected zone DToT = DI+DP+Ds
(HAZ) on the ferritic steel side of the DMW . Damage
The value of D T o T is a direct prediction of the amount
in this region is expected to occur first because the
of interface material damage to be expected . Material
creep resistance of the material, at the temperatures of
damage is specifically defined as the portion of the
interest, is much lower than that of the austenitic
weld interface unable to transmit load by virtue of
parent metal or the commonly used filler metals . The
voiding or cracking resulting from time-dependent
general macroscopic appearance of many failures sug-
processes . For an unfailed weld, it is computed by
gests that failure occurs by a relatively low ductility
subtracting from the original area of the interface (in a
process and the final fracture surface may show evi-
longitudinal cross section of the weld) the area of the
dence of the weld bead contours. However, detailed
interface remaining intact and dividing the result by
examination has shown that failure can occur by one of
the original interface area . Intact material is any mate-
two submodes of damage and crack propagation. Spe-
rial through which the interface is connected to the
cifically, failures can occur by:
base tube material . DTOT at failure is obtained by
Mode I. Integranular cracking which occurs along
subtracting from the original area of the interface the
prior austenite grain boundaries within the ferritic
area deemed to have failed by tearing (rapid tensile
HAZ adjacent (1 to 2 grains) to the weld metal inter-
failure) and dividing the result by the original inter-
face. This mode occurs for DMWs made with austenit-
face area.
ic stainless weld metal and sometimes for DMWs
The damage from the self-generated loads, DI , is
made with nickel weld metal . The initial intergranular
due to temperature cycling as a result of power
voiding or cracking does not normally start out at
changes and due to self-stresses induced during hold
either the inside or outside surfaces but develops with-
periods . These self-stresses are strain-controlled in
in the wall of the tube . Failure occurs by crack link-up
nature . As a result, the stresses will relax with time in a
and propagation to the surfaces.
manner dictated by the creep properties of the materi-
Mode II. Interfacial voiding in which voiding and
al and the multiaxial stress state in the region . For the
cracking occur along a planar array of coarse globular
sake of simplicity, a factored elastically calculated ini-
carbides that form along the ferritic HAZ to weld met-
tial stress is assumed to regenerate as a result of the
al interface . This mode occurs only for nickel weld
startup-shutdown cycle, and this value is used as the
metals but may be accompanied by some Mode I
nominal stress to evaluate the creep damage . Hence, in
cracking . The initial interfacial voiding again starts
the evaluation of the creep damage component, only
from within the tube wall. Final failure occurs by void
the duration at nominal operating temperature is re-
link-up and crack propagation to the surfaces.
quired . In the fatigue damage evaluation, the strain
Mode I Damage Assessment range is based on the mean wall temperature change
For the purpose of estimating Mode I damage aris- AT (°F) occurring during operation.
ing in stainless-steel filler metal welds, the loading to The load damage relationship for DI is:
which the DMW is subjected was divided into three
categories. The corresponding three damage compo- DI = k l n(EI )'' + k 2 th 4l0f(T° ) ,
nents are defined as follows:
• DI = Intrinsic- (or self-) damage, caused when where
temperature changes are applied to the dis- m = number of loading conditions,
similar weld in the presence of the normal a l = factored self-stress based on To,
Printed copies are not controlled. Printed on: 2018-05-22 (UTC)
4 WRC Bulletin 350
equation for carbide growth is6 stores the weld strength. The redissolution behavior
d = kt 1i3 10gm , has been approximated in PODIS by use of a time-
temperature algorithm based on the Larson-Miller pa-
where rameter. The algorithm ensures that redissolution
d = carbide length in the direction parallel to the starts to occur at an LMP of 38,500 . Subsequent redis-
weld interface, solution behavior is assumed to be a mirror image of
t = time, the growth behavior.
g(T) = function of operating temperature, T, The Ds damage term due to secondary bending
k = a constant. loads arises from loading cycling due to power changes
and from creep-rupture damage induced during hold
It is observed that little or no additional carbide periods . This load is strain-controlled and relaxes at
nucleation occurs with time, and that carbide growth elevated temperature . However, the stress modifier
depends on time and temperature . Hence, the pitch "w" is again introduced (as for Dp) in the creep damage
(sum of ligament width and major carbide diameter) term to account for the effective loss of load-carrying
remains constant. The average pitch value was deter- area due to the growth of interfacial carbides.
mined6 to be 1.723 µm. The loss of the load-carrying The fatigue damage evaluation for Ds is the same as
section is then given by: that described for stainless-steel welds . The stress
k modifier "w" is not used in this fatigue damage evalua-
.723 X (diameter of carbide), tion. The load-damage relationship for Ds is:
w 1
m
_ kd Ds = k4n(Es) ' + k5t m10f(T°) [as/(1 — k710g(To) tn13 )] ,
1 .723
where
= k6t o 10g( To) ,
m = number of loading conditions,
where as = axial secondary stress,
Es = strain range based on AR,
w= effective fraction of interface lost due to car-
t,,, = incremental time for a constant as,
bide growth, t o = cumulative time from t = 0 to the end of time
k6 = a constant.
interval tm, adjusted for segments with differ-
The effective net section stress than takes the form, ent temperatures,
f(To) and g(To) are functions of operating tempera-
Q
ture, To,
y, ,B, k4, k5 and k6 are constants.
1 .2
each can be predicted from the knowledge of the mag-
nitude of the damage components D . The current algo-
rithms in PODIS 4,7 are set to calculate damage under a
given set of loading conditions. It is an average-life
prediction tool—there are no design margins included.
The algorithm for De and D f in N-47, in contrast, 0 .8- 0
Anticipated Uncertainty
1.0--
Figs . 3-1 and 3-24 show comparisons between PO-
0
DIS—calculated damage—and that observed for 0
from the figures, the scatter in the data is such that for
a predicted damage of 0 .1 actual damage is, for 95% of
0
data, no greater than 0.5 . Thus use of a design limit of
0.1 results in a high degree of confidence that at the
0 .2-
end of service, a design margin factor of 2 will still exist
in the component. ~-- PROPOSED DESIGN LIMIT
F f I
0 .0 0.1 0 .2 0 .3 0 .4 0 .6 0 .0 0 .7 0 .8 0 .9 1 .0 1 .1 1 .2
4 . Proposed Changes to Code Case N-47
OBSERVED DAMAGE
- 3100 Mode I Failure . The design life of the dis- DI - MODE I (FATIGUE)
STAINLESS-STEEL FILLER DEW
similar welds may be subdivided into segments each
representing a different level of either applied load or
operating temperature and the associated time dura-
tion . The total damage is calculated using the design
rules in -3110, -3120 and -3130 . The total damage of
the dissimilar welds during the design life is obtained
by cumulative sum of each damage contributor . The
value of the total damage shall not exceed unity.
- 3110 Intrinsic Damage. The damage (DI) created
by the differential thermal expansion between the
weld filler and the base materials is calculated using
the following equation:
M
,
DI = k 1 nJ( EJ) + k 2 tL (crL) R 10f1TL)
J=1 L=1
where
M = total number of different types of loading
cycle,
nJ = total number of fatigue cycles for the jth TEMPERATURE DEC F
loading cycle type,
Fig . 4-1—Failure curve for Mode I (fatigue) in stainless-steel filler
Ej = (Da) X (iTJ), differential thermal expansion D,
DMW
strains for the Jth loading cycle type,
Act= differential value of coefficient thermal ex-
pansion between the filler and base metal
(in./in./°K),
OTJ = mean wall cyclic temperature range in the N
Jth loading cycle type, Dp = N k 3 tL vL10f(TL) ,
TL = the operating temperature (°F) for the Lth L=1
loading condition,
N = total number of steady-state loadings, where
tL = time duration (hr) of the Lth steady-state
loading condition,
~L = 1/2E(&a)(TL — 70) ( p s i ),
E = Young's modulus (psi),
DI - MODE I (CREEP)
k 1 , k 2 , y, Q, f(T) are defined in Table 4-1. STAINLESS-STEEL FILLER DMW
ki 6.8 X 10-4 0
k2 2.5 X 10- 10 0
k3 2.4 X 10-' 1 .6 X 10- 7
k4 1 .0 X 10 -2 1.0 X 10 -2
k5 5.5 X 10- 10 5.5 X 10- 10
0.2 0.2 TEMPERATURE DEG F
2 .7 2.7
31032 31032
f(TL ) 13 .26 - 13.26 - Fig. 4-2—Failure curve for D, Mode I (creep) in stainless-steel filler
(T + 460) (T + 460) DMW
Printed copies are not controlled. Printed on: 2018-05-22 (UTC)
8 WRC Bulletin 350
DP - MODE I (CREEP)
QL = primary stresses (psi) created by pressure and NICKEL BASE FILLER DMW
where
D5 - MODE I (FATIGUE)
DP - MODE 1 (CREEP) STAINLESS-STEEL OR NICKEL BASE FILLER DMW
STAINLESS-STEEL FILLER DMW
188
Fig . 4-3—Failure curves for Dp Mode I (creep) in stainless-steel filler Fig . 4-5—Failure curves for Ds Mode I (fatigue) in stainless-steel or
DMW nickel-base filler DMW
Printed copies are not controlled. Printed on: 2018-05-22 (UTC)
Dissimilar Metal Welds 9
DS - MODE I (CREEP)
STAINLESS-STEEL OR NICKEL BASE FILLER DUB
TL = operating temperature (°F) for the Lth
steady-state loading,
tL = time duration (hr) for the Lth steady-state
loading. Smaller time steps may be required
for accuracy depending on the specific case.
where
(L-1)
tL _ 25665[1/(TL+460)-1/(T K +460)]
tk10
K=1
MTL = 10 [385001(T1+460)-20]
where
k 3 , k 6 , 13, AT') and g(TL ) are defined in Table 4-2,
oL = primary stresses (psi) for the Lth steady-state
loading,
N = total number of different types of steady-state
loading,
DS - MODE II (CREEP)
NICKEL BASE FILLER DMW
Ds = k 4nJ( t
J=1
N
aL ( TL )
k5
lof tL,
+ L= 1 — k 6 10g(TL) (TWL + 0.5 TL)1/3
DTOT = Dp + Ds
TIME MRS
5 . References
1. Ryder, R. H ., "Design Criteria Strategies for Dissimilar Welds in High
Temperature Gas-Cooled Reactors," EPRI-HTGR-85-02, Electric Power Re-
search Institute, Dec. 1985.
2. Sartory, W. K., "Revised Analysis of the Transition Joint Life Test,"
ORNL/TM-9211, July 1984.
3. Case N-47 of the ASME Boiler and Pressure Vessel Code, Version N-47-
21, Dec. 11, 1981.
4. "Dissimilar Metal Weld Failure Analysis and Development Program,"
EPRI CS-4252, Nov . 1985.
DS - MODE II (FATIGUE) 5. Johnson, J . J ., "MODSAP—A Modified Version of the Structural Analy-
NICKEL BASE FILLER DMW sis Program SAPIV for the Static and Dynamic Response of Linear and
Localized Nonlinear Structures," GA-A14004, GA Technologies Inc ., June
1978 .
6. Foulds, J. R ., "Kinetics of Interfacial Carbide Growth at Ni-Based
2 1/4 Cr-lMo Transition Joints," GA-C18533, GA Technologies Inc ., June 1986.
7. "Transition Joint Development Program—Final Report," GEFR-00645,
General Electric, Advanced Reactor Systems Department, Dec . 1982.
1BBBB
NUMBER OF CYCLES