ML19319D958
| ML19319D958 | |
| Person / Time | |
|---|---|
| Site: | Rancho Seco |
| Issue date: | 08/31/1972 |
| From: | Brandes W, Brown L, Lakner J BECHTEL GROUP, INC. |
| To: | |
| Shared Package | |
| ML19319D950 | List: |
| References | |
| NUDOCS 8003270639 | |
| Download: ML19319D958 (69) | |
Text
.-
~
.t O
f FINAL REPORT ON MINOR IMPERFECTIONS FOUND IN PIPE WELDS AT RANCHO SECO NUCLEAR GENERATING ' TATION o
.I i
BECHTE L CORPOR ATION AUGUST 31,1972 I
r 4
4 f
8003270[g r
=
- )
.I' l!
I FINAL REPORT ON MIN 0R D1 PERFECTIONS FOUND IN PIPEWELDS AT
- = -
RANCHO SECO NUCLEAR GENERATING STATION i
I I
lx d
4 l
-d
/
Prepared by:
J5 L. Lakner, Engineering Specialist 7
Reviewed by:
L. R. Brown, Engineering Supervisor Approved by:
WM e" M
W. A. Brandes, Project Engineer ha.
=.
N 4
7 --
TABLE OF CONTENTS TITLE PAgL INTRODUCTION 1
i
ONCLUSIONS 3
TELEDYNE STUDY 4
)-
WFTnING CONTROL' 7
ENGINEERING ANALYSIS 8
i APPENDIX EXPERIMENTAL EVALUATION OF THE EFFECT OF MINOR IMPERFECTIONS FOUND IN GIRTH BUTT WELDS AT RANCHO SECO NUCLEAR GENERATING STATION (Technical Report E-1526(b) by TELEDYNE MATERIALS RESEARCH
- [_.
c l
i I
f
- i e
0 4-4
.~
f*
i.
- a
INTRODUCTION This report is written as a final report covering the evaluation of minor imperfections found in girth butt welds in piping.
It describes the investigations made and the results developed in the analytical and physical testing program.
Conclusions have been drawn regarding the significance of the imperfection and certain practices have been put into effect at the jobsite to minimizc the formation of this condition.
^
Throughout this paper and the interim report, the imperfection under investigation is understood to be an extremely minute physical condition found on occasion at the root of a weld and observable only at high magni-fication on a polished weld cross section.
The basic background and planning for this document were covered in the interim' document entitled " Report on Minor Imperfections Found in Pipe
~^
Welds at Rancho Seco Nuclear Generating Station," dated May 30, 1972.
The following subjects were covered in the interim report:
1.
The discovery of the imperfections and a metallurgical evaluation of pipe butt welds.
This included an investigation of the correlation of " lines" found in radiographic film and the existence of the imperfection. An analysis of data taken from all imperfec-tions found in pipe welds showed them to vary from 1 to 15 mils in depth.
2.
A secondary examination, conducted by a third party, was made on welds which did not show linear indications on the radiographs.
No imperfections were found and it was concluded that minor a
" 19 1
-.-- - - - - -.a.:
-. ~
' ~
imperfections are not found in welds which do not have linear indications on the radiographs.
3.
A description of the analytical and physical testing program which had been authorized to evaluate the effect of the imperfection on the strength and fatigue life of the welds.
In addition certain controls had been implemented at the site to minimize the continued production of welds containing the imperfection.
4 b4 m
e te
$ 4 1
Eat M
7..
i u-2
CONCLUSIONS The conclusions developed in the erograms of investigation of the effect of the imperfection in piping butt welds are as follows:
a.
In stainless steel weldments, the presence of imperfections has been proven not to be significant by the physical testing program.
The fatigue life and the load carrying capacity o:i the welds were not reduced by the imperfection.
b.
In carbca steel weldments an increase of 15 percent in stress indices is indicated by the physical testing program for welds with the imperfection.
s c.
Piping for Rancho Seco is designed in accordance with the rules given in ANSI B31.7 Nuclear Power Piping and ANSI B31.1.0 Power L.
Piping.
These systems will be suitable considering the given
'~
loading condition even though imperfections may be present in the L'
butt welds with the single qualification that stress indices for carbon steel welds in safety related Seismic Class I piping be increased by the small amount indicated in the Teledyne report.
%M
+=
4 4 F'
%H
]
i
,w e *
+-se 3
L.
J
TELEDYNE STUDY On May 11, 1972 Teledyne Materials Research of Waltham, Mass. was commissioned by Bechtel to undertake a scudy to evaluate the effects of
" minor imperfections" in the structural integrity of butt welded pipe joints. The study was to include both analytical and physical investigations.
The analytical program was to include finite element analyses of pipe welds incorporating offset, rollover and " minor imperfections" to evaluate the effect of the " minor imperfection" on the strength and fatigue life of piping girth butt welds designed and executed in accordance with the applic-able codes for the specified loading conditions.
The offset and rollover (internal reinforcement) were to be to the maximum values allowed by ANSI B31.7.
The design of the joint was in agreement with this Code and the analysis was to be done in accordance with the Code.
Structucal and thermal loading conditions were selected as the most severe found in the piping stress analysis for Rancho Seco in systems where these types of welds would be found.
The physical testing program was to cover fatigue testing of weldments s
to determine the effects of offset, rollover and " minor imperfections."
(
The weldments and loadings were to effectively duplicate the configurations and stresses as used in the finite element analysis.
In addition, ultimate
'I' strength tests were to be conducted to determine if the " minor imperfection" affects the ultimate strength of the weldment.
The plan was based on the premise that the finite element analysis would be correlated with the physical testing program to establish a basis for stress analysis of welded piping systems which may contain imperfections.
be 4
Computations on the finite element analysis were started in advance of the testing program. However, the analysis was discontinued when the
~
physical test data began to accumulate and showed that the weld at the imperfection zone was superior to the adjacent pip'e in fatigue and ultimate strength.
The predominance of failures occurred at the edge of the weld and not through the imperfection.
Since the failures occurred at the edge of the weld joint, Teledyne did not complete the calculations of the imperfection zone; therefore, no conclusions can be drawn from the analytical part of their work.
A typical straight pipe joint was investigated by Bechtel in an inde-pendent finite element analysis. The model used gave information on the imperfection zone, however, as with the Teledyne analysis, it could not be l
related to the actual physical test failure modes.
No conclusion has been drawn from Bechtel's analytical work to evaluate the effect of these minor imperfections en the load carrying capacity or cyclic life of the weld joint.
As indicated in Teledyne's table on Page 34, minor imperfections and imperfection propagation in depths up to.049 inch deep were found in full
{
pipe section weld joints after failure.
Microscopic examination of the i.
cycled test joints was made but it was not possible to determine and measure the degree of propagation.
In any event, the cyclic failures of the full pipe test weld joints always started at the outside of the pipe and progressed inward.
It can be stated that the test results prove con-clusively that the minor imperfections do not play a role in joint failure.
The complete Teledyne report entitled " Experimental Evaluation of the Effect of Minor imperfections Found in Girth Butt Welds at Rancho Seco Nuclear Generating Station" is bound with this report as an Appendix.
i*
LJ 5
M
F Weld Specimens -
All wel.1 specimens were welded and assembled by Bechtel Corporation to dimensional specifications given by Teledyne.
Final machining of weld samples was handled by Teledyne. All welds containing " minor imperfections" in the specimens were made in the Bechtel Welding Laboratory in San Francisco.
All plate specimens were also welded at this facility.
All other welds and pipe machining operations were conducted under the direction of the Bechtel Welding Engineer at the Rancho Seco site.
All welds were radiographed to assure Code quality throughout.
Considerable difficulty was experienced in duplicating the imperfec-tion in the flat plate specimens. This was resolved by the manufacture of an artificial imperfection in these samples.
It was made by welding I
two tightly fitted plates with incomplete penetration. This is considered to effectively duplicate the imperfection for test purposes, however its s
depth is not nearly so regular and wr.s found frequently to be excessive when compared to the 15 mil maximum depth found in the metallurgical investigations.
' I 4
't 4
5.
!f L
I 6
r-
4 WELDING CONTROL D' ring the period since March 1972 certain special welding control u
procedures have been in force at Rancho Seco to minimize the formation of
" minor imperfections." These include the following:
Purge gas quality - Liquid argon is being used with the purity of a.
each shipment certified by the supplier, b.
Allowable mismatch - Weld joint offset has been held to 1/16 in, wherever practical even though the Code allows up to 3/32 in.
Welding technique - Welders have been trained to minimize internal c.
weld reinforcement through a special training program.
d.
Root pass quality - Only selected welders have been permitted to perform root pass welding.
These measures have been found to be effective for the purposes intended.
This study indicates that they are not mandatory, however, the special weld-ing control procedures will continue to be in effect at the jobsite.
Approximately 90 percent of the pipe welds have been completed for this plant as of this date. To provide the highest quality welds possible the remaining 10 percent will be completed under the same upgraded controls.
All safety related butt welded pipe joints whose radiographs showed dark lines and could have masked some type of defect have been repaired, s.
Butt welds whose radiographs have evidenced linear indications and could be postulated to contain " minor imperfections" and meeting Code requirements in all other details have been accepted and will not be
'k repaired.
k
~.
a.
p U
ENGINEERING ANALYSIS Stress levels in carbon steel lines which could contain " minor imperfections" are being currently reviewed.
Since there are no nuclear Class I lines in this category the investigation is limited to safety related seismic Class I piping. Any welds will be repaired or resupported to reduce stress levels whose allowable limit is exceeded by the actual stress level corresponding to the increased stress index indicated in the Teledyne Report.
e
.L.
6 s o-eek 8
1 ".
L
e m
T IT1 2
Q
[
l
4
.e--
BECHTEL CORPORATION 12400 EAST IMPERIAL HIGHWAY
~~
NORWALK, CALIFORNIA 90650 TECHNICAL REPORT E-1526(b)
EXPERIMENTAL EVALUATION OF THE EFFECT OF MINOR IMPERFECTIONS FOUND IN GIRTH BUTT WELDS AT RANCHO SECO NUCLEAR GENERATING STATION
- Mu AUGUST 15, 1972 Prepared by:
R. F. Brodrick Senior Engineer l
J. E. Gallivan, Senior Engineer D. A. Van Duyne, Proj ect Manager Approved by:
, b, b - /rf b >>
D E. C. Rodabaugh, Baytelle Memorial Institute l
v p v-Dr'. W. E. Cooper, Vicf President & Technical Director TELEDYNE MATERIALS RESEARCH 303 BEAR HILL ROAD WALTHAM, MASSACHUSEITS 02154 1
Technical Report No. E-1526(b) i ABSTRACT Teledyne Materials Recearch was commissioned by Bechtel to initiate
~
an investigation to evaluate the effects of " minor imperfections" found in girth butt welds at the Rancho Seco Nuclear Generating Station. These
" minor imperfections" were associated with " lines" found on radiographs of pipe butt welds but these " minor imperfections" were not always present when " lines" were detected.
A fatigue test program was conducted to experimentally evaluate the significant loadings affecting the critical pipe welds. Tests were E
conducted on both stainless steel and carbon steel weldments.
In stain-less steel weldments the effect of the minor imperfections present in the test specimens was not significant as evidenced by the predominance of fatigue failures which did not pass thru the imperfections.
4, In carbon steel weldments, the limited test data (plate bending only) g I
indicate a slight increase in the present Code peak stress indices for
'I thermal and moment loading of 1.7 and 1.8 to a revised K-index of 2.0.
c l
l' i
L.
lI I
L.
- m i
t-,
r i.
Technical Report No. E-1526(b)
TABLE OF CONTENTS Page No.
INTRODUCTION 1
SELECTION OF TEST METHODS 2
V GENERAL DESCRIPTION OF TEST SPECIMENS 5
Materials 5
Welding and Fabrication Procedures 5
Evaluation of Imperfections 7
PLATE BENDING FATIGUE TESTS 22 Plate Bending Fatigue Specimens 22 t
Plate Bending Test Procedure 22 Plate Bending Test Results 23 Austenitic Stainless Steel 23 Carbon Steel 24 AXIAL FATIGUE TESTS 29 f
Axial Fatigue Specimens 29 Axial Fatigue Test Procedure 29 Axial Fatigue Test Results 30 I
PIPE BENDING FATIGUE TESTS 34 Pipe Bending Specimens 34 Pipe Bending Procedure 34 Pipe Bending Test Results 36 TENSILE TESTS 39 q
EVALUATION OF EXPERIMENTAL RESULTS 41 U-CONCLUSIONS 47
(!
REFERENCES 48 i.
ACKNOWLEDGEMENTS 49 a
- L 3
4 Technical Report No. E-1526(b)
LIST OF TABLES Page No.
(-.
I I.
SPECIMEN CONFIGURATION
SUMMARY
9 i"
II.
SUMMARY
OF PLATE BENDING TEST RESULTS 10 III.
SUMMARY
OF AXIAL TEST RESULTS 13 r
IV.
SUMMARY
- PLATE BENDING TESTS CARBON STEEL 14 t.
e r
i f '
L t
i 1
m o
g* %
e L
^
r, l
a.
Technical Report No. E-1526(b)
LIST OF FIGURES
~
Figure No.
Title Page No.
1 WELD JOINT DESIGNS 15 t..
2 TYPICAL WELD CROSS SECTIONS 16 c-
[,
3 IMPERFECTION - SPECIMEN 69C-2 17 r-4 IMPERFECTION - SPECIMEN D1-2 18 5
IMPERFECTION - SPECIMEN 69B-11 19 6
IMPERFECTION - SPECIMEN 4B-3 20 t
7 FRACTURE SURFACE - SPECIMEN 69B-2 21 8
PLATE BENDING FATIGUE SPECIMEN 25 9
PLATE BENDING FATIGUE MACHINE 26 10 TYPICAL BENDING FAILURE - EDGE OF WELD 27 11 TYPICAL BENDING FAILURE - THROUGH WELD 28 12 AXIAL FATIGUE TEST SPECIMEN 31 13 AXIAL FATIGUE TEST MACHINE 32
(
14 AXIAL FATIGUE TEST MACHINE 33 15 PIPE BENDING SPECIMEN 38
^
16 EXAMPLE OF ELONGATION OF TENSILE SPECIMENS 40 17 BENDING TEST RESULTS VS. INTENTIONAL VARIANT 45 18 BENDING TEST RESULTS VS. IMPERFECTION DEPTH 46
+
u
'sd -
r
4 1-Technical Report No. E-1526(b)
INTRODUCTION On May 11, 1972 Teledyne Materials Research was commissioned by Bech'tel to initiate an investigation to evaluate the effects of " minor
~
imperfections" on the structural integrity of butt welded pipe joints.
Dr. W. E. Cooper sponsored the Teledyne work and Mr. E. C. Rodabaugh of Battelle Memorial Institute served Teledyne as a consultant in areas of his expertise.
Bechtel Corporation Report, dated May 30, 1972, on this
,z subject (Reference 1) provided additional background information as stated in the Scope of that report as follows:
"This report containa a detailed account of the events associated with the ' lines' found in radiographs of pipe butt welds at the Rancho Seco Nuclear Power Station.
Included in the repor2 are discussions of the investiga-tions which were conducted and the test program that was performed to identify the cause of the ' lines' and to provide a means for the prevention of them."
~
The Teledyne work was to include a mechanical testing program covering fatigue aspects of all significant loadings. The tests covered config-
}
urations of unwelded material, offset, rollover and " minor imperfections" and combinations of these in a replication of samples of each. The program was to cover welds in both stainless steel and carbon steel materials.
'~
The objective of this report is to provide the data and analysis i
}
required to experimentally evaluate the effect of the minor imperfections found in the girth butt welds at Rancho Seco Nuclear Generating Station
[-
in accordance with the rules of Section III of the ASNE Code (Reference 2) or the similar rules of USAS B31.7 for Nu, lear Power Piping.
s
(_.
~~
Technical Report No. E-1526(b)..
'~
SELECTION OF TEST METl!ODS r.
A girth butt weld in pipe, made in accordance with Code require-ments, does not decrease the strength of the pipe insofar as a single load application is concerned. This is insured by the requirement (ASME 3
Section IX, Par. Q-6) that the ultimate strength of the weldment must be not less than the specified ultimate strength of the base material. Tests of this type were conducted and the results are included herein.
e However, imperfections in butt welds that are acceptable under Code requirements do significantly decrease the cyclic load capacity of the weldment, as compared to the fatigue strength of polished bars of the a
base material. These imperfections, acceptable under Code requirements, are such things as:
a) the geometric notch caused by the intersection of the t
weld reinforcement and the pipe b) root irregularities
~
c) weld porosity d) slag inclusions The welds considered herein have essentially no porosity or slag inclusions.
l The test specimens were fabricated so that they have various degrees of the root imperfection, the evaluation of which was the objective of the 1
I
- The term " Code" as used in the subsequent text, refers specifically to I~
Section III of the ASME Code, Reference (2) which replaces USAS B31.7 l
for Nuclear Power Piping.
j.
j..
I M
> F" 1.
~~
Technical Report No. E-1526(b).a tests.
Inasmuch as the tensile tests were not expected to indicate possible weakness due to the weld imperfections, the major part of the test program consists of fatigue (cyclic loading) tests.
u..
A girth butt weld in the piping systems of interest will be 4
subjected to three types of loadings each of which, during the anticipated life of the plant, will be applied a number of times and thus constitute cyclic loadings. These types of loadings are:
(a)
Internal pressure Lt (b) Moment (c) Thermal gradient.
"~
Internal pressure produces a hoop stress, Pr/t, which is parallel to the girth butt weld and hence the hoop stresses are not significantly increased by weld imperfections. The axial stress is Pr/2t. The affect of weld imperfections in the axial direction can be evaluated by an axial test of a section of the pipe. This type of specimen was therefore included in the test program.
The moment loading most significant to the weld imperfections produces 6._
an essentially constant axial stress thru the wall thickness at two points around the periphery; these high-stressed points for moment loading are therefore also represented by an axial test specimen.
However, since this loading produces a non-uniform stress around the periphery, it was deemed desirable to run confirming tests on full-section pipes.
[
Lf.
n c.
~~
Technical Report No. E-1526(b),,
Thermal gradients produce a thru-the-wall bending stress which can be represented by applying a deflection to a cantilever beam; thereby producing a thru-the-wall bending stress.
Accordingly, " plate bending" tests were included in the test program.
The magnitudes of loadings were selected to be consistent with the maximum loadings that might be applied in the actual piping system and in conformance with the intent of Article 11-1000 of the Code.
In this s-respect, it should be noted that the evaluation of weld imperfections covered herein is intended to apply to piping systems in general. There are many butt velds at various locations in a piping system and each butt
)
weld may be subjected to different combinations of moment loading, thermal gradients and pressure. Accordingly, it is not possible to directly apply the requirements of Article II-1000 because the " design service loads", used as parameters in Article II-1000, muct be establish-ed for each particular location in the piping system.
w i*
i e
c p
&d Technical Report No. E-1526(b),
GENERAL DESCRIPTION OF TEST SPECIMENS
~
Materials u
Plate bending fatigue specimens were fabricated from 304 stainless steel plate material of approximately 0.328 inch thickness. Similar bending fatigue specimens were fabricated from carbon steel plate equivalent to ASTM-A-333 Grade 6.
Axial fatigue specimens and pipe bending fatigue specimens were i
fabricated from stainless steel welded pipe, ASTM-A-358, Grade 304, 14-inch 0.D. x 0.438 inch nominal wall. All of these specimens were machined to a thickness of approximately 0.328 inch adjacent to the test weld.
Extra axial fatigue specimens were used for the stainless steel m
tensile tests.
Carbon steel tensile tests were conducted using plate bending specimen coupons.
Welding and Fabrication Procedures The welded specimens for testing in this program were fabricated f.
by Bechtel. The intent was to produce a range of weld quality from " base-
"~
line" or defect-free welds through various sizes of minor imperfections.
7.
To assist in interpretation of the data, it was decided to include imper-fections over 20 mils in depth, although this exceeds those detected in L_
field welds.
To produce these welds Bechtel employed special techniques developed
~
to reproduce imperfections similar to those detected in the Rancho Seco i
field welds. The welding procedures, other than special techniques, were II.
m
?
I" lu
.p i
Technical Report No. E-1526(b).-
the equivalent of those used on the field welds.
E'
}
Table I summarizes the specimen configurations used in the attempt to produce a range of imperfection sizes.
Visual examination indicated that the stainless steel plate bending specimens Nos. 66, 67, 68 end 69A had been welded with full penetration welds.
Specimens 69B, 69C and 69D had been welded with incomplete pene-tration.
The samples with complete penetration were consistent across the width of the specimen.
Those with incomplete penetration showed variations with the extent of partial penetration varying across the specimen width.
The stainless pipe specimens for axial fatigue showed full penetra-tion on Nos. 1, 2, 3A and 3B with partial penetration on 4A and 4B.
The partial penetration on these was more uniform than on the plate bending specimens.
In addition, the cover pass was a wide string bead in all cases, but 4A, where a weave was used.
The carbon steel plate specimens had full penetration welds on Nos. 82, 83, 84 and 85, and partial penetration on Nos. 86, 87 and 88.
The extent of partial penetration was more uniform and consistent than on the stainless steel plate specimens. The cover pass for Samples 82-85 and half of 88 was a single bead, for 86, 87 and the other half of 88 was multiple beads.
The samples provided measured 71/2" x 21/2" with t!.? weld centerad across the long dimension. A udnimum of nine samples were provided of L_
each test number.
im hl I
e-L
~
Technical Report 1.
No. E-1526 (b) -
Evaluation of Imperfections The presence and extent of imperfections in the test specimens were evaluated metallographically by TMR. Tables II, III and IV list the values as reported by Bechtel and as measured by THR.
t-As stated in the previous section, welding c5 the plate bending specimens with incomplete penetration showed variations across the sample.
Accordingly, more extensive investigation of imperfection size was required on these specimens, and correlation of Bechtel and TMR l'
f measurements was not always possible.
Figure 2 shows typical weld cross sections and the type of variation that occurred in root weld bead a
For the axial specimens, where the welding was automated, more uniform and consistent penetration was achieved, and extensive evaluation of imperfection size was not required.
In some cases imperfection size was measured on the machined speci-mens; where testing schedules did not permit this method, measurements were made on cut-outs adjacent to the reduced section of the test L.
specimens.
This type of cut-out was preserved for all specimens.
i Typical imperfections are shown in Figures 3, 4, 5, and 6.
It was noted that the machined bevel of joint configuration B was usua31y not evident in the plate bending specimens.
Figure 3 represents the 1
root area of specimen 69C-2-L and shows an imperfectiou (0.019") extending L
from the notch (offset) to the root pass of the weld.
Figure 4 ehows a different type of imperfection, shorter but with greater width.
C' I1.
1 f
J j
r-l e
Technical Report No. E-1526(b) ?'
(Sample Dl-2-L).
Figure 5 is typical of the B group with a comb'ned i~
notch and crack type imperfection (Sample 69B-ll-L).
Figure 6 shows a wide and deep imperfection found in Sample 4B-3-R.
In those cases where the imperfection was uniform, the original machined bevel became apparent.in the fracture surface.
Figure 7 is n.
typical of failure appearance under these conditions.
>~
w w -
4 L
km Resp i
L.
T~
i.s F
~
L
_9 TABLE I.
SPECIMEN CONFIGURATION
SUMMARY
PLATE BENDING FATIGUE SPECIMENS (304 STAINLESS STEEL)
Intended Test No.
Joint Design (1) Offset Reinforcement Imperfection 66 A
No Remove None p
66 C
No None 67 A
No Intact None 68 A
3/32" Intact None 69A A
3/32" Intact None
[
69B B
3/32" Intact 0.005" 69C B
3/32" Intact 0.010" 69D B
3/32" Intact
>0.010" Note:
- 1) Joint designa - see Figures A,B,C AXIAL FATIGUE SPECIMENS (304 STAINLESS STEEL PIPE)
~
Intended Test No.
Joint Design (1) Offset Reinforcement Imperfection 1
C No Intact None 2
C 3/32" Intact None 3A A
3/32" Intact 0.010 to 0.015 3B A
3/32" Intact 0.010 to 0.015 4A B
3/32" Intact 0.010 to 0.015 4B B
3/32" Intact 0.030 to 0.060 PLATE BENDING FATIGUE SPECIMENS (CARBON STEEL)
Intended Joint Design (1) Offset Reinforcement Imperfection Test No.
82 C
No Intact None 83 A
No Intact None 84 C
3/32" Intact None a
85 A
3/32" Intact None 86 B
3/32" Intact 0.005" 87 B
1/8" Intact 0.010" 88 B
1/8" Intact
>0.010" PIPE BENDING FATIGUE SPECIMENS (304 STAINLESS STEEL PIPE)
Intended Joint Design (1) Offset Reinforcement Imperfection Test No.
14-C 1/8" Intact
>0.010 15
.C 1/8" Intact
>0.010 16 C
3/32" Intact
>0.010
~
For Joint Designa see Figure 1.
U$
h"
=-----
- v--_--
-(-
,1;,
,2,-
7: 2., -
-a i
[
TABLE II SUNIARY OF PLATE BENDING TEST RESULTS Actual Imperf'n.
Stress Joint (3)
Off-Intended Bechtel DfR Level Cycles to Fracture Test No.
Design Set Reinforcement Imperf'n L
R(2)
L R(3)
(kpal)
Failure Locationg 66-1 A
None Ground None None None
+ 40 63600 E0W 2
A None Ground None None None I 30 128600 E0W 3
A None Ground None None None I 40 52800 E0W 4
A None Ground None None None I 30 Invalid Test (5)
E0W 5
A None Ground None None None
[30 202300 E0W j
6 A
None Ground None None None
+ 40 56400 NS 9
C None Ground None None None
[30 118600 E0W 67-2 A
None As Welded None 0.008 0.005 0.007 0.007 1 30 326600 E0W 3
A None As Welded Nanc 0.009 0.004
+ 40 61800 NS 4
A None As Welded None 0.003 0.006 0
[30 340100 E0W
[
5 A
None As Welded None 0
0
+ 40 73300 NS 6
A None As Welded None 0.002 0.005 0.004
[30 362800 E0W 8
A None As Welded None None None
+ 40 105300 E0W H
9 A
None As Welded None None None 140 66300 E0W
?
681-2 A
3/32" As Welded None 0
0 0
0
+ 30 154400 E0W 1-3 A
3/32" As Welded None 0.004 0.004 0
0
[30 165300 E0W 682-1 A
3/32
As Welded None 0.015 0.007 0.006 0.013
+ 30 144900 E0W 2
A 3/32" As Welded None 0.010 0.012 0.009 0.014 I 40 46400 E0W 3
A 3/32"
'As Welded None 0.007 0.009 0.010 0.010
[' 40 45700 E0W 683-2 A
3/32" As Welded None 0
0
+ 40 51200 NS 3
A 3/32" As Welded None 0
0.005 0
0.004
[ 4(,
64000 E0W 69A-2 A
3/32" As Welded None 0.009 0.012 0.008 0.008
+ 30 216500 FOW 3
A 3/32" As Welded None 0.003 0.009
[40 38100 NS(6) 4 A
3/32" As Welded None 0.010 0.008 0.009 0.006
+ 30 168700 E0W 5
A 3/32" As Welded None 0.010 0.007 0.004 I 40 66000 WELD 6
A 3/32" As Welded None 0.008 0.008 0.011
[30 131500 WELD 8
A 3/32" As Welded None 0.004 0.006 1 30 269700 WELD
)
9 A
3/32" As Welded None 0.007 0.013 0.007 1 30 121500 WELD 10 A
3/32" As Welded None 0.007 0.016 0.013 0.012 1 40 171100 E0W 11 A
3/32" As Welded None 0.018 0.018 0.011 0.011 1 40 46800 E0W 12 A
3/32" As Welded None 0.019 0.016 0.018 1 40 35400 WELD i
]
a
- 7.,
.~
-( :7 i
?
i4 i,
O TABLE II (Continued) y)
Actual Imperf'n.
Stress Joint Off-Intended Bechtel DGt Level Cycles to Fracture 4}
Test No.
Design set Reinforcement Imperf'n L
Rl2J L
R(3)
(kpsi)
Failure Location
(
695-1 B
3/32" As Welded 0.005" 0.055 0.060 0.035 0.040 1 40 50400 WELD 2
B 3/32" As Welded 0.005" 0.050 0.060 0.045 0.040
+ 40 57300 WELD 3
B 3/32" As Welded 0.005" 0
0.015 0.003 0.016 I 40 49800 E0W 4
B 3/32" As Welded 0.005" 0
0 0.004 0.008 I 40 114000 E0W 5
B 3/32" As Welded 0.005" 0
0 0
0.002 1 40 205100 E0W 6
B 3/32" As Welded 0.005" 0.040 0
0.023 0.002 1 30 328500 E0W 7
B 3/32" As Welded 0.005" 0
0 0.002 1 30 177000 WELD
~
is B
3/32" As Welded 0.005" 0
0.030 0.005 1 30 137600 WELD 10 B
3/32" As Welded 0.005" 0.025 0.015 0.011 1 30 155100 WELD 11 B
3/32" As Welded 0.005" 0.025 0.045 0.018 0.014 1 40 53300 E0W s
69C-2 B
3/32" As Welded 0.010" 0.020 0.040 0.019 0.020 1 30 161600 E0W U
3 B
3/32" As Welded 0.010" 0.045 0.055 0.005 1 40 52500 E0W I
4 B
3/32" As Welded 0.010" 0.020 0.030 0.008 1 30 194900 E0W 5
B 3/32" As Welded 0.010" 0.035 0.020 1 40 47100 HS 6
B 3/32" As Welded 0.010" 0.025 0.010 0.011 0.007 1 30 170400 E0W 8
B 3/32" As Welded 0.010" 0.007 0.026 1 30 123700 E0W i.
9 B
3/32" As Welded 0.010" 0.020 0.030
+ 30 154100 E0W
^
10 B
3/32" As Welded 0.010" 0.018 I 40 39800 NS(7) 11 B
3/32" As Welded 0.010" 0.007 0.003 I 40 58600 E0W 12 B
3/32" As Welded 0.010" 0.004 0.003
[40 106600 E0W 69D1-2 B
3/32" As Welded 0.010" 0.009 0.004 0.006 1 30 126000 E0W 3
B 3/32" As Welded 0.010" 0.004 0.003 0
0 1 30 139600 E0W 69D2-2 B
3/32" As Welded 0.010" 0.006 0.013 0.013
+ 30 134200 E0W 3
B 3/32" As Welded 0.010" 0.011 0.015 1 40 58200 NS 69D-4 B
3/32" As Welded 0.010" 0.005 0
0.007 0
1 40
<175000(8)
E0W i
5 B
3/32" As Welded 0.010" 0
0.032 0.003 0
+ 40
<147400(8)
E0W 7
B 3/32" As Welded 0.010" 0.047 0
0 0.005 I 30 135000 E0W 8
B 3/32" As Welded 0.010" 0.003 0
0.013 I 40 67500 E0W 9
B 3/32" As Welded 0.030" 0.015 0
0.015 1 40 64200 E0W 1
(-
p-
[
j--
.j p [-]
e-a ass
- k ig TABLE II (Continued) g
$I 91 NOTES:
1.
See A, B, C in Figure 1 2.
As reported by Bechtel, reading on specimen edges 4
3.
As measured by DiR, locations notnecessarily same as Bechtel Y
4.
Fracture Location: EOW, edge of weld NS, narrow section, failure in base metal at possible plastic hinge.
WELD, failure thru root of weld l'
5.
Loose grip 6.
Overheated, invalid test 7.
Surface imperfection 8.
Specimen did not fully separate, so automatic shut-off did not actuate
~.. - _. _.
lrb
{
Technical Report IL; No. E-1526(5) -
1 f
TABLE III SIDefARY OF AXIAL TESTS
.e-1 c-3 Joint Off-Imperfection Offset Nom.
Cycles to Test No.
Design see Incended Measured Meas'd Stress Failure f*
1-1 C
No None 0
0 None 116667 99100 1
1-3 C
No None None 116667 483800 1-4 C
No None 0
0 None 116667 367200 1-5 C
No None None 116667 322200 l
]-
2-1 C
Yes None 116667 40400 2-3 C
Yes None 0
0 0.093 116667 40600 2-4 C
Yes None 0.094 116667 40400 2-5 C
Yes None 0.097 116667 67500 2-6 C
Yes None 0
0 0.075 116667 36400
~~
2-7 C
Yes None 0.080 116667 70300 2-11 C
Yes None 0.090 116667 102000 I'
I 3A-2 A
Yes 0.010-0.015 0.020 0.009 0.106 116667 225900
'~
5 A
Yes 0.010-0.015 0
0.007 0.104 116667 114000 7
A Yes 0.010-0.015 0.013 0.004 0.102 116667 95400 9
A Yes 0.010-0.015 0.014 0.018 0.112 116667 118900 11 A
Yes 0.010-0.015 0.003 0
0.075 116667 145400 3B-3 A
Yes 0.010-0.015 0.011 0.008 0.108 iL6667 72500 7
A Yes 0.010 0.015 0.019 0,007 0.125 116667 38700 9
A Yes 0.010-0.015 0.013 0
0.130 iL6667 45300 t-11 A
Yes 0.010-0.015 0.010 0.021 0.110 116667 70400 4A-1 B
Yes 0.010-0.015 0
0.110 L6667 176300 y'
3 B
Yes 0.010-0.015 0.023 0.127 iL6667 170400 5
B Yes 0.010-0.015 0.016 0.111
$6667 55500 i
4B-1 B
Yes 0.030-0.060 0.015 0.109
$6667 78100 i
1 3
B Yes 0.030-0.060 0.023 0.097
$6667 142300 5
B Yes 0.030-0.060 0.008 0.100
$6667 131800 7
B Yes 0.030-0.060 0.068 0.098
$6667 98200 i
i6 NOTE:
1.
Material, pipe A358, Gr 304 14" 0.D. x 0.438 Nom. Wall, Machined to 0.328 for welding and testing. All specimena
,~
failed at edge of weld.
a t
L s
r-
- +
i; TABLE IV
SUMMARY
- PLATE BENDING TESTS CARBON STEEL
. E' Actual Imperf'n.
Y Joint Off-latended Bechtel TMR(2)
Cycles to Fracture Tes t No.
Design set Imperf'n L
R L
R Eailure Location
$E' M
82-2 C
None None None None 308,000 E0W
N 5 C None None None None 232,700 E0W 7 C None None None None 445,100 E0W 83-2 A None None N'one None 221,200 E0W 5 A None None None None 138,400 E0W 7 A None None None None 166,800 EOW 84-11 C 3/32 None 0.002 0.004 0.004 160,000 E0W 13 C 3/32 None 0.000 0.003 0.004 208,400 W i 16 C 3/32 None 0.002 0.003 0.000 226,100 W/E0W E-I 85-2 A 3/32 None 0.023 0.008 0.006 85,400 W 13 A 3/32 None 0.008 0.006 0.003 185,200 W 23 A 3/32 None 0.013 0.005 0.005 81,900 W 86 -5 B 3/32 0.005 0.002 0.012 0.013 62,700 W 11 B 3/32 0.005 0.012 0.016 0.018 64,500 W 16 B 3/32 0.005 0.012 0.009 0.000 96,900 W 20 B 3/32 0.005 0.054 0.010 0.025 0.026 <90,700 W 4 22 B 3/32 0.005 0.010 0.010 0.018 0.016 54,800 W { 25 B 3/32 0.005 0.020 0.020 0.020 0.030 58,000 W 87-2 B 1/8 0.010 0.006 0.003 0.009 134,000 W 13 B 1/8 0.010 0.002 0.000 0.000 0.000 205,700 E0W 23 B 1/8 0.010 0.030 0.020 0.013 82,500 W i ~ 25 B 1/8 0.010 0.0 30 0.020 0.018 72,000 W 88-25 B 1/8 >0.010 0.000 0.000 0.003 142,000 W 31 B 1/8 >0.010 0.018 0.012 0.003 153,600 W 34 B 1/8 >0.010 0.005 0.002 0.014 136,900 W 5 B 1/8 >0.010 0.035 0.038 0.018 171,500 W 'l
i~ i-~ ' " ' ' ~ ', ~~ } l ~~ il } f ~l f~ ~ ^) [. ~] f~ T [ i f E! ' T ^ @W TABLE IV (Continued) i eq $ I A E. vi NOTES: $7 9E 1) For all samples, minimum weld reinforcement, no grinding, all tested at
- 4 a stress level of + 30 kpsi.
2) For failures through weld, single figure is average value of imperfection as evident after fracture. 3) Failure started at weld offset and progressed to edge of weld. 'I
r- ~5 E Technical Report [ No. E-1526(b) 37 1/2* 37 1/2" r-y s A 5/16" NOE. o - 1/8" l 1 37 1/2* 37 1/2* s 1~ 1/4" R J, 45' BEVEL 5/16" NOM. B k I ON LAND o i d l o - 3/32"
- .080" FLAT 304 STAINLESS STEEL
.050" 304 PIPE AND FLAT CARBON STEEL 37 1/2* 37 1/2* 4 If C 5 5/16 NOM. o t 1/16" ,---- 1/8" + J' FIGURE 1. WELD JOINT DESIGNS A, B, & C wJ
i F-I 'l Technical Report l No. E-1526 (b) I l ( \\ k \\ Specimen 69B-3 5X Ferric Chloride Etchant FIGURE 2. TYPICAL RELD CROSS-SECTION l'
L i F i: Technical Report No. E-1526(b) : n.. t. y ?- ..,m.-
- ?
4.. i,; f 1 -h^. , ', v%...<3
- a w
Y,- s . ) fr ~,y ~, ih hf i_ 't f' {f. 'fffh. ( g pi v' 04) th. -(.
- L
-T 69C-2-L 100X Ferric Chloride Etchant t 1 FIGURE 3. IMPERFECTION, SPECIMEN 69C-2, 0.019" lu. l l h 1 __,,-_.___~._._,,,._.,__.,,___,..__,___-,,_.____,n,._._,._,,,,._.___.
<l Tecitnical Report No. E-1526(b) 4;r l{. l ,3 - ' ri (Ni, v. <f,jyp ,4 .. "6 l n ,i Q--Q #s_ '\\ l%.,.
- .;9
fl ?.'pp a ~. . 'w..
- x-y %y A$
- .,e n,
.,. n s J{.s Q :; ).,. ~l:6l% (* ! ll o ,'6~ *:p R. f;T; 'i g,.~ ~r I7.", T,h w{,'y/w ., p- ~~$' 9 p j. 3 s, y ..;.1 1 g^ l( %;y . s. Y?A l 69Dl-2 100X Ferric Chloride Etchant l i< { lL. FIGURE 4. IMPERFECTION, SPECIMEN 69D1-2, 0.006" I L t
r. i l. I I i 4 i l l " Technical Report No. d-1526(b), l l.- i l
- ~w3.m.
- m. y v 3 y m,
~ y M2D ?. g- ' ~ hA .s \\ ~y6'*%';rD., m.w n y ~, is,,Rf:yg^+ QT : ?. ~ ~.
- w$wj
- .
y.,y-L.ag e*%.- 4dd[?q%l s y i.~, ^$?(**::35 +.
- 'l _~7
_ T15?$^ ?$ I' .] .. c c- - y_. ; = ,a,- "g g. r-*' .,. 9 .- w -
- x..
,,s-Y - m W s &. m.g y -Tg;.,.. ;e;y > -r" g ji {Q, 'y gh* 41.' ~5,', Is ., -<dy fiM' ' _...( '. ,,, e - 1 J#. j h . " ~
- N; i
'a m z w a w. - ; m; "'4n c.iJ 1 ,m.... ~,,, 47 A-gtp-g,y:.p ci r r,- p, ,e ~ [ 69B-11-L 100X Ferric Chloride Etchant s 4 FIGURE 5. IMPERFECTION, SPECIMEN 69B-11, 0.018" k j. A. t
!E ii Technic.a1 Report ; No. E-1526(b) r_ i k_ ~ ~ 2(~ ' l t -[ !{ a' ! ' t i[ .l. } ' i - I.. 4B-3 100X Unetched i I
- l.
i f FIGURE 6. IMPERFECTION, SPECIMEN 4B-3, 0.023" 6. ls. l l - - - - - - - - - - - - - - - - - - - - - - - - - - ^ -
Technical Report No. E-1526 (b)....,_.; + - ,, ~ - r' m + +- a cv kan
- m... p.4,
.. e: . e m y. ~ f c r. ~.., _ -,.. e 'e ~* ' ,?4 A ' y e. q, 'j' - . ' p; .+ p 4 s FIGURE 7. FRACTURE SURFACE, SPECIMEN 69B-2
s-4 Technical Report No. E-1526(b) ;_ PLATE BENDING FATIGUE TESTS Plate Bending Fatigue Specimens Specimens for the bending tests were cut from the welded flat plates. Specimen blanks were removed from each plate by sawcuts made perpendi-a. cular to the weld. The resulting blanks were then machined to the contour shown in Figure 8, with the weld at approximately midlength of the taper. The original surfaces were not machined except for the s. 4 rough removal of the reinforcement crown by griding on variant 66. r-i Plate Bending Test Procedure u. These tests were conducted on a Budd Model VS-150 Plate Bending r-L_ machine. The important features of this machine are shown in Figure 9. The tapered cantilever specimen is gripped by a stationary grip at the large end and by a moving grip at the small end. The small end is driven in a fluctuating motion by an adjustable rotating eccentric. The deflection amplitude of the moving end of the specimen is maintained at a constant value according to the setting of the eccentric. The point of load (i.e., deflection) application is at the imaginary apex of the specimen taper. Thus, the stress amplitude is constant along the test section of the specimen. Calibration of the apparatus was conducted by means of deadweight loading a specimen up to a point somewhat below the elastic limit. The deflection was measured by a dial indicator. The actual test conditions were then set up by extrapolating the deflection values to correspond to the nominal stresses of the test program. Nominal stress ranges of c_ m
.= s.. ~~ Technical Report No. E-1526(b). 60,000 pai and 80,000 psi were imposed. Essentially all of the tests at the lower range were completed prior to changing to the higher range. Thus, the eccentric drive setting was unchanged during each set of tests. A strain-gaged specimen was used to check the amplitudes when changing loading conditions. An operating speed of 2,000 cycles per minute was used for the tests at the lower stress. An operating speed of 700 cycles per minute was used at the higher stress in order that specimen heating not be excessive. This speed, combined with the use of fan cooling, resulted in a maximum specimen temperature of 140F at the higher stress u amplitude. "~ Plate Bending Test Results - Austenitic Stainless Steel Table II gives details of the specimen types, test conditions, and-tabulated results of the plate bending tests. The imperfection' depths were determined by the methods discussed earlier. It is seen that the majority of failures occurred at the edge of the weld (E0W) although several occurred through the weld. Examples of these types of failures are given in Figures 10 and 11. A few specimens at the higher stress level (80,000 psi range) failed away from the weld. These failures were all at the narrow end of the j tapered specimen. Although the tapered beam has theoretically constant "~ stress throughout its length, the plastic strain condition apparently causes a strain concentration at the narrow end. Data from specimens which failed in this manner have not been used in subsequent evaluations as the local values of stress may be in doubt. 6WW
~ ~ - - 7 L r-Technical Report L No. E-1526(b). Plate Bending Test Results - Carbon Steel Table IV gives details of the specimen types, test conditions, and tabulated results of the carbon steel plate bending tests conducted. The imperfection depths were determined by the methods discussed earlier. It may be observed that for the three tests of variant 82 (no offset, no intended imperfection) failures occurred at the edge of the weld (ECW) but the three tests of variant 86 (3/32" intended offset, intended imperfection) failures occurred at the imperfection (or at the root of the weld formed by the offset). On variant 86 specimens there was a relatively good weld reinforcement shape and a relatively severre notch effect due to the offset. r e
- O numeS
& 4 4 d 1 u. w n .~ IJ
_1 _ _ _ L_.,___--_____________.-.________p__--;---- I. f. I__, L ,_.,l [. _ _ ] t. t t i i H tD zo E mW i m HH u N m 7 1/2" L U* O a vg r, S/16" NOMINAL THICKNESS I WELD _1 / 6 w 7 1/2* J .t. ~~ _---a-4 W g \\ O i ,wu APEX OF ANGLE AND 7/8 R ^" l l FIGURE 8. PLATE BENDING FATICUE SPECIMEN i ) 1 i , \\ l i I
1. -__.,.m_- 9 l ~ Technical Report ' No. E-1526(b) u l1. P* 1 I ... + v - * * = *.e+ n -wee e <*=esumenee ho 9 k4 m 9 u 6e t t 1 6 P4 6 I t.4 lM d 't p e 3. FIGURE 9. PLATE BENDING FATIGUE MACHINE b4 .W l m !J m a-
--q m. r-Technical Report h" No. E-1526(b), AXIAL FATIGUE TESTS I Axial Fatigue Specimens i Specimens for the axial fatigue tests were made by machining the ends of pieces of 14-inch schedule 40 pipe to the configuration to give the desired weld preparation and offset for each case. The welding was per-i formed and the pipe then cut longitudinally to form 12 axial specimens, each having a section of the weld at midlength. E:ch piece was then machined further to the contour shown in Figure 12, with the weld being { at midlength. The details of welding and imperfection geometry are . discussed in the General Description of Test Specimens. I'. Axial Fatigue Test Procedure { The axial tests were conducted on an MTS servo-controlled electro-w. hydraulic test machine. (See Figures 13 and 14.) Special grips were T 1 made to accommodate the curved surfaces of these specimens which had been cut from pipes. In this machine, the load is applied through the grips, L one of which attaches to the hydraulic actuator and the other to a load f~ cell. Thus, the load is forced to follow a command signal which in these L tests was a sinusoidal wave. The machine incorporates accessory apparatus such as cycle counters, t. automatic shutoffs, recording cpparatus, etc. l The tests were run at cyclic load such as to give an alternating 6_. of 16,667 psi in the uniform section of the specimen. Cycling rate was 17 Hz. i L d
~ l P 6 7 Technical Report i No. E-1526(b) m Axial Fatigue Test Results Table III gives the details of the specimen types and the results of the axial tests. Imperfection depths were determined by the methods discussed in the General Description of Test Specimens. It is noted that all failures initiated at the edge of the weld. w. e e a t b & ?- amm P+4 bW W* .m 1 i P b. I s i enes d
7 . __ y
- 7.. )
~, I i i ! sm EE i
- DmX
.L E U:= Re wo 11" N I. l 3 1/2" y i + + + \\ + g 4. f 4 1 y w _A e / 7 / + + 1" + / 3/4" R / SPECIMEN CROSS SECTION 5/16" NOMINAL WALL i FIGURE 12. AXIAL FATIGUE SPECIMEN i W
2 P'" Technical Report No. E-1526(b) i I ] 4 1 'I 'e l
- pty t
l l' 1 d 'e A f.t'. L FIGURE 13. AXIAL FATIGUE TESTING MACHINE u w ~1
r- -1s. I I Technical Report _33_ No. E-1526(b) h O ~ 337,7?- * - v.w - r x 3.. 3 l' l .*M 4
- (
II .N i g I i I h i' i FIGURE 14. AXIAL FATIGUE TESTING MACHINE .M Iy I e i
r-. i '~ Technical Report f No. E-1526(b) PIPE BENDING FATIGUE TESTS Pipe Bending Specimens The test specimen is shown in Figure 15. The test weld was made between two 9" long pieces of 14" Sch. 40 pipe, ASTM A358 Grade 304, machined to a wall thickness of 0.328". The remainder of the pipe is ~i 14" Sch. 40; left at its original thickness of 0.438" so that the auxiliary welds would not fail. The two 9" lengths wer-a offset from each other as indicated in Figure 15. The weld preparation detail for the test weld is as shown in Figure 1-C. The offsets and imperfection lengths, as measured on L the failed specimens, were as tabulated below. Ieperfection and/or Crack length i' Spec. Offset measured af ter testing (l) No. 0* 180* O' 180* 14 3/32 >3/32 0.049 0.029 15 >3/32 1/16 0.009 0.037 16 3/32 3/32 0.002 0.005 Pipe Bending Test Procedure ( } The free end of the test specimen was loaded by means of a hydraulic L cylinder as indicated in Figure 15. Because of the long distance between the loading point and the test weld, the loading effectively produces a f ( Indicated lengths include original imperfection and resultant fatigue crack. ( The test procedure used is the same as that used by Mark 1(5) Mark 1's 1 data forms the major part of the basis for stress intensification factors for girth butt welds given in ANSI piping codes and for the stress indices for moment loadin B31.7(4) and ASME Section III(2)g for girth butt welds given in ANSI C u >1 I L
[" Technical Report No. E-1526(b).. pure moment at the test weld. The pipe was oriented so that the maximum offset was aligned with the maximum stress due to the moment. A load-displacement relationship was established by applying loads in increments; reading the dial gages (See Figure 15) at each load incre-ment during the 1st, 2nd and 10th cycles. A cycle consists of applying loads to maximum in one direction, reducing to zero, increasing to maximums in the opposite direction and reducing the load to zero. Strain gages were placed on the test specimens as shown in Figure 15; strain P readings were taken at the same loads as the dial gage readings. The test specimen was mounted as shown in Figure 15 and was filled i'~ with water to a head (above the test weld) of about six feet. This was done so that a fatigue crack penetrating the wall of the pipe would be indicated by leakage of water. The fatigue test was " displacement controlled" by means of travel limit switches which controlled the hydraulic cyclinders. Cycles were applied at a rate of 10 per minute. The displacements, corresponding moments, and nominal stresses at the test weld were: Test Displacement Load Nominal Ratio to Axial No. Range Amplitude Stress Amplitude Test Stress Level ( in. Ib. psi c 14 1.384 8880 30,000 1.8 i 15 0.977 6410 21,700 1.3 16 0.759 4930 16,700 1.0 r lu t. r-i l7
P-L. P Technical Report No. E-1526(b), The nominal stress is based on the test section and test weld nominal e, dimensions of 14" outside diameter, 0.328" wall thickness. The nominal stress, S, is equal to MC/I, where M = load x 13.25 x 12 in-lb C = 7.00 in I = (w/4) (7 - 6.672 ) = 329.4 in' The load and nominal stress ranges are twice the amplitudes tabulated above. Pipe Bending Test Results Cycles-to-failure and failure locations are tabulated below. Test Nominal Stress Cycles-to-Thru-Wall Failure No. Amplitude, psi Failure Locations 14 30,000 11,671 E0W at 0* 15 21,700 18,988 E0W at 0* 16 16,700 24,500* E0W at 180* It can be observed from the above tabulation that all failures occurred at the edge of the weld reinforcement. Examination of the failed No. 14 and 15 test specimens showed that both specimens had secon-dary cracks that apparently started at the root of the weld (imperfection location) but these carcks did not penetrate the wall thickness. s. Cycles-to-failure for Test No. 16 were estimated from strain and cycling rate records because the automatic shut-off device failed to actuate. L w e-
e. Technical Report No. E-1526(b). The results of test numbers 14 and 15 contain an anomaly in that as the load was decreased from its maximum value to about 90% of its maximum value, the displacement increased slightly instead of decreasing as would be expected. This oc. curred for both directions of load applica-tion. Strain gage readings exhibit a similar anomaly. The load-displace-I ment relationships are based on averages of all readings except for the suspect readings, hence the load-displacement relations are probably valid despite the annmaly. In any event, the relative effect of minor imperfections in the weld versus other stress raisers (e.g., the edge l of the weld reinforcement) would not be effected by the absolute value of the nominal bending stress. S b a d L 4 N. ) i ~ i 1 i
k F l ' / AY .328" STRAIN GAGES r DIAL GAGES & LIMIT TEST WELD SWITCHES - / I p r p _.m-14" SCH. 40 A / r \\ [' I PIPE ( g( 1 1 1 / 7 A+ D i /- liin h LOAD FROM 9 / 159" = DISPLACEMENT REFERENCE FRAME ( i O# 1 t 7 f 3/32" I s (SECTION A-A) s. h 1 FIGURE 15. PIPE BEN"'NG SPECIMEN w f' U 1
Technical Report No. E-1526 (b) TENSILE TESTS r-A few tensile tests were performed on welded samples of the austenitic steel and of the carbon steel. These tests were performed for the purpose of confirming that the materials met the Code ultimate strength, values of 75,000 psi and 60,000 psi, respectively. Test results are given in the following table. Specimen No. Ultimate Strength Condition Offset Imperfection Austenitic Steel 1-6 86,470 None None 4A-9 85,440 3/32 in. 0.018 in. 4B-9 73,940 3/32 in. 0.045 in. Carbon Steel 82-11 82,600 None None 85-5 81,600 3/32 in. 0.010 in. Specimen 4B-9 is discounted since the imperfection is significantly deeper than any imperfection found in otherwise acceptable welds in the ~ plant. Specimens 4B-9 and 85-5 failed through the weld. All other specimens failed well away from the welds. Figure 16 shows the degree of elongation typical of these tests. O }. M em i b
r t. e I t Technical Report l. No. E-1526(b) ! t l 't l' L Y,, ', Q*r.,,, [ { eg g 1 I c -ll'f s by. f 1 W IW g N e i 1 \\.. } m i 1 f. FIGURE 16. EXAMPLE OF ELONGATION OF TENSILE SPECIMEN L 9
r-Technical Report ~ No. E-1526(b) P L. EVALUATION OF EXPERIMENTAL RESULTS m l Austenitic Stainless Steel The location of failure in a fatigue test is a highly significant 7 part of the test results because the failure will occur at the " weak link" t- ~ in the test specimen. A girth butt weld, fully meeting Code requirements, ordinarily contains geometric and/or metallurgical " notches" which reduce the cyclic load capacity as compared to a polished bar of the base material. If the imperfections deliberately introduced into the test specimens reduce the load capacity of the weld, this will be indicated by the fatigue failure initiating from, or at least passing thru, the imperfection. Accordingly, the most significant part of the test results with respect to evaluation of the effect of the imperfection is the failure location. The failure locations are summarized below. Type of Total Failure thru Failure thru Fatigue Test Tests Imperfection Edge of Weld Plate Bending 61 10 51 Axial 27 0 27 .I.f Pipe Bending 3 0 3 / The above tabulation indicates that, in most of the tests, the i geometric notch at the edge of the weld reinforcement produced a larger f[ fatigue-effective stress than the imperfection. 7 Figure 17 shows results of the plate bending tests grouped into inten-tional variants. Other than variant 67 (no offset, no intentional imper-fections) the variants give essentially the same cycles-to-failure. The w parallel relationship shown in Figure 17 between the results for 30,000 psi w I h
r Technical Report [ -. No. E-1526(b) p-I and 40,000 pai nominal bending stress suggests that there is some para-meter which is related to the variant type but that parameter is not the imperfection length. Figure 18 shows results of the plate bending tests grouped by the measured (average of Bechtel and THR) lengths of the imperfections. The L. results are random in nature, indicating that the imperfections are gen-( erally not significant in controlling the fatigue life of the test specimens. 1!_ The random nature of the results with respect to the imperfection is, [ of course, due in large part to the fact that most of the fatigue failures L. were not associated with the imperfections. A comparison between those t plate bending specimens which failed thru the imperfection and those which failed at the edge of the weld is given below. Nominal Failure thru Failure thru 1 2"
- Stress, Edge of Weld Imperfection (N /N ) ! '
y 2 psi N L, mils N L, mils y 2 30000 205000 9.7 165000 9.1 1.05 40000 85000 9.6 52000 36.4 1.12 {' In the above tabulation, N is the average cycles to failure and L is the i average length of imperfection. The comparison indicates that in these '~ cases where failure did occur thru the imperfection, the number of cycles-to-failure was, on the average, somewhat less than where the failure s.- occurred thru the edge of the weld. It should be noted that the differ-ence in cycles represents a small difference in fatigue-effective stress s_ 1 L +
- 9 b
c-Technical Report F No. E-1526(b) ! '~ 4 level. Using the relationship S a 1/N *3, then the relative stresses are as shown in the last column of the above tabulation. In the axial tests, all failures occurred through the edge of the weld; none thru the imperfection. Accordingly, no correlation with either the i intended imperfection length or measured imperfection length would be r {, expected. In the pipe bending tests, the failures which penetrated the wall c. thickness all initiated at the edge of the weld. Carbon Steel Results of the plate bending fatigue tests on carbon steel indicate a that the imperfection (or notch at the root of the weld formed by the offset) is a larger stress raiser than the edge of the weld reinforcement. I This appears to be due, at least in part, to a relatively better weld reinforcement shape on the carbon steel specimens as compared with the stainless steel specimens. f The fatigue lives of variant 82 (no offset, no intended imperfections) were substantially greater than implied by the K factor of 1.8 or K 2 3 I factor of 1.7 given in the Code. Comparison of the cycles-to-failure of L variant 82 with the ASME Criteria Document ( } S-N curve for carbon steel f i i indicates a peak stress index of about 1.4 rather than 1.7 or 1.8. This p i relatively " good" weld is believed to be due to the smooth reinforcement contour and root pass. See II-1520 of ASME Section III. [ L
Technical Report No. E-1526(b) i The fatigue lives of variant 86 (3/32" intended offset, intended imperfections) were significantly lower than variant 82 but, when the cyclic lines of variant 86 specimens are compared with the S-N curve for carbon steel, a peak stress index of about 2.0 is indicated instead of the Code K index of 1.8, K index of 1.7. Accordingly, the limited 2 3 j data on girth butt welds with minor imperfection indicates that the peak stress index should be increased to a slightly higher value; i.e., K =K = 2.0. 2 3 It should be noted that the offset itself, in the variant 86 specimens tested, produces a significant notch and the results obtained may be due to that notch rather than the imperfections. Again it should be emphasized that plate bending fatigue specimens were the only ones tested on carbon steel material. However, results of axial and bending fatigue tests conducted on stainless steel indicated that the plate bending was most sensitive to this notch or " minor imperfection". i a i 4 I 4 l t
p. m-- 350 aa sb or D + 30.000 PSI en a l l D ~ 300 G" $5 o - + 40.000 PSI ne$8 O e "o 250 N s
- 3 150 c
kn a = o 8 a 100 50 1 4 0 1 l 66 67 68 69A 69B 69C 69D ~ i SPECIMEN GROUP i i FIGURE 17. BENDING TEST RESULTS VS, INTENTIONAL VARIANT k
Technical Report No. E-1526(b) 6 I' l. 250 m e y 225 m I sh 2E W 200 u gi o m EE a m [ 8;. w< .g h 175 ~ L si< r / L d 150 NONE O TO 10 10 TO 20 20 AND LARGER v MEASURED IMPERFECTION DEPTH (INCHES X 10-3) [ r i a FIGURE 18. BENDING TEST RESULTS VS. IMPERFECTION DEPTH I l l l i 6
~~ Technical Report No. E-1526(b), CONCLUSIONS 1. In stainless steel weldments, the effect of impa.rfections present in the test specimens was not significant as evidenced 1 by the predominance of fatigue failures which did not pass through the imperfections. 2. In carbon steel veldments, the limited test data (plate bending only) indicate that either the imperfection or the notch due t-the offset is significant. A slight increase in peak stress (~ indices is indicated from the present Code indices of K2 " 1*0 c, for moment and K3 = 1.7 for thermal loading to K2"K3 = 2.0. 3. The difference in results for stainless steel as compared to "~ carbon steel is believed to be due, at least in part, to the L relatively smoother weld reinforcement contour on carbon steel t, test specimens. i w bd i (L afM
Technical Report No. E-1526(b) REFERENCES 1. Bechtel Corporation " Report on Minor Imperfections Found in Pipe Welds at Rancho Seco Nuclear Generating Station." dated May 30, 1972. 2. ASME Boiler and Pressure Vessel Code, Section III,1971 Edition with Addenda through Winter 1971. 3. ASME Boiler and Pressure Vessel Soce, Section IX,1971 Edition with Addenda through Winter 1971. 4. USAS B31.7 - 1969 for Nuclear Power Piping. 5. A.R.C. Markl, " Fatigue Testing of Piping Components." Trans, ASME Vol. 74, 1952, pp. 287-303. Ie. 6. Criteria of the ASME Boiler and Pressure Vessel Code for Design by Analysis in Sections III and VIII, Division 2. L. l I t I t. k L 1 L
Technical Report No. E-1526(b) ACKNOWLEDGEMIRTS On May ll, 1972 Teledyne Materials Research was commissioned by ~ Bechtel to iniciate an investigation to evaluate the effects of "ntinor aperfections" on the structural integrity of butt welded pipe joints. Dr. W. E. Cooper sponsored the Teledyne work and Mr. E. C. Rodabaugh of Battelle Memorial Institute served Teledyne as a consultant in areas of his expertise. L-The pipe bending fatigue tests were conducted by Southwest Research under the technical leadership of Mr. S. C. Grigory. All other fatigue tests and the tensile tests were conducted by Teledyne Materials Research. ee 4 1 i 1. ) f.] J}}