ML20134P090
| ML20134P090 | |
| Person / Time | |
|---|---|
| Site: | South Texas  |
| Issue date: | 01/31/1985 |
| From: | Capener E, Cipolla R, Egan G APTECH ENGINEERING SERVICES |
| To: | |
| Shared Package | |
| ML20134P044 | List: |
| References | |
| AES-8405461Q-3, NUDOCS 8509060132 | |
| Download: ML20134P090 (79) | |
Text
f p
IS APPLIED TECHNOLOGY AES 84054610-3 Final Report EVALUATION OF THE IllTEGRITY OF EMBED SYSTEMS AT THE SOUTH TEXAS ELECTRIC GEf1ERATIllG STATIO!!
Prepared by Erwin L. Capener Russell C. Cipolla Geoffrey R. Egan Thomas J. Feiereisen Steve R. Paterson Aptech Engineering Services, Inc.
795 San Antonio Road Palo Alto, California 94303 Prepared for Bechtel Energy Corporation 5400 Westheimer Court Houston, Texas 77252-2166 Attention:
Mr. flick Joonejo January 1985 8509060132 850118 PDR ADOCK 05000498 4
A PDR l
795 SAN ANTONIO ROAD 0 PALO ALTO O CALIFORNIA 94303 0 (415)858-2863
VERIFICATI0t; RECORD SHEET REPORT fiO: AES 8405461Q-3 TITLE:
Evaluation of the Integrity of Embed Systems at the South Texas Electric Generating Station DATE:
Janua ry 17, 1985
/
Originated by:
l- /7-b ~
Geoffrey R. Egan, P oject !!anager Date Checked and Verified by: _
/
iffre
. i over, Verifier Date MA I!il OI Quality Assurance Review by:
Je ffre:
, r aver, QA Engineer Date yy t,P l
Quality Assurance by:
g Je f frey 'D. Bykn, QA f4anager Date i
y-i ENGINEERING SERVICES
)
A TABLE OF CONTENTS Section Page ABSTRACT I
INTRODUCTION 1-1
Background
't-2 1-1 Integrity Evaluation Program Objectives 1-2 2
SCOPE 2-1 3
TEST MATRIX DESIGN 3-1 Introduction 3-1 Welding Parameters 3-1 Heat Af fected Zone Hardening 3-2' Susceptibility to Cracking 3-4 Test Matrix 3-4 Over Tenpering 3-6 Test Method and Parameters 3-6 4
TEST RESULTS 4-1 Introduction 4-1 Inspection Resul ts 4 Metallurgical Results 4-5 Hardness Results 4-5 Analysis of Results 4-14 Weld Strengths 4-14 Analysis of Olsplacement Measurements 4-19 Fractographic Analysis 4-22 Fracture Surf ace Appearance of Sample C8 4-22 Fracture Surf ace Appearance of Sample C3 4-22 Fracture-Analysis 4-30
)
Plastic Collapse 4-36 Fracture Mechanics Analysis 4-41 Root Cracks 4-41 Summary.
4-47 5
SIGNIFICANCE OF THE RESULTS 511 introduction 5-1 Design Basis.
5-1 Dynamic Loading Considerations.
.5-6 6
CONCLUSIONS AND RECOMMENDATIONS 6-1 REFERENCES R-1 APPENDIX A - Recommendations For Statistically-Based Lower A-1 Bounds'
i ABSTRACT l
This is a report on the structural integrity of embedded assemblies determined from an evaluation program perf ormed by Aptech Engineering Services, Inc.,
(APTECH) for Bechtel Energy Corporation.
The embedded assemblles in question were field-fabricated at the South Texas Project and consist of ASTM A36 plates with welded anchor rods specified to be A36 material. During febrication, some rods of ASTM A193-B7 material may have l
been inadvertently welded to the A36 plates using welding procedures intended i
for A36 materials.
The work described herein was perf ormed to determine the available load capacity of these potentially non-conforming weldments.
The loed capacity program was based on tests of A193-B7 welded rod specimens prepared to simulate and/or bound the production field welding.
Specific weld parameters were selected to conservatively represent metallurgical conditions that are consistent with the field welding characteristics.
Tensile tests were perf ormed on specimens of three rod dianeters with proportionate weld sizes.
All tests were perf ormed under. the supervision of APTECH and Bechtel Materials and Quality Services (M&QS) Department.
In addition to the tensile tests, detailed metallurgical and fractographic oxaminations were perf ormed to identify the material conditions resulting from these welds.
As expected, hard heat af fected zone regions were generated in the rod material and, in some cases, cracks attributed to the welding were observed.
The results of the tensile tests indicated 1 hat the available load capacity of
~
the A193 rod weldments is adequate to develop the required design load dictated by the originally specifled A36 rod material of one Inch, one and one-half Inch, and two inch disneters.
The tested ultimate loads of such weldments were found (1) to af ford adequate f actors of saf ety with respect to the specified allowable loads for A36 rod weldments, and (2) to be equal or higher than the specified minimum ultimate strength of the A36 rods.
i Examination of broken specimens indicated that the f ailure was controlled by the weld joint itsel f and not the properties of the heat treated A193 rod.
Furth ermore, fracture mechanics analysis confinned that there was no potential l
for low stress brittle fracture, and that limit load analysis was the appropriate basis for the f ail ure prediction model for the welded configurations under consideration.
In view of (1) the excellent load perf ormance of the weldments, (2) the fact that low stress brittle fracture is not developed, and (3) the original design margins are retained, it is recommended that the cmbedded plates with welded rods of potentially A193 material be accepted "use-as-is".
1 1-1 Section 1 INTRODUCTION BACKGROUND Based on a previous separate investigation, it was determined that past procedures used in the South Texas Project (STP) for handling round rod stock material for the field f abrication of embedded plates with welded anchor rods could have led to comingling of ASTM A36 and A193 Grade 87 rod materials.
The embedded plates are used in various buildings, of both Units 1 and 2 of the STP, including the reactor containment building (RCB), and the mechanical electrical auxiliary building (ME AB).
The welded rods range in diancter f rom three quarter inch to two inches.
The welding procedures used to fabricate these embecments were based on welding of A36 to A36 material, and accordingly, the inadvertent use of the procedures for welding comingled A193 rods to A36 plates was recognized to be a potential concern.
Separate studies perf ormed earlier Indicated that the actual extent and the projected probability of the material comingling were acceptably low.
Nevertheless, it was consicered desirable to determine the structural reliability of the onbedded plates by verifying the load capacity and fracture safety of plate assemblies postulated to have weldments of A193 rods to A36 plate.
The A193-87 material, because of its higher carbon and alloy contents, is more hardenable than A36 material, and consequently, hard, heat af fected zones (HAZ ) could be expected to develop in the A193-87-to-A36 weldments.
The higher hardness potentially developed in the HAZ may also f avor the f ormation of cracks as the wel dment cool s.
These two ef fects, hard HAZs and the likelihood of the presence of cracks, may contribute to a degraded strength condition in the A193-87 weldments.
G B
1-2 Because of the concern for the degraded weld condition, Bechtel Energy Corporation and Aptech Engineering Services, Inc., develeped an integrity eval uati on program, which is the subject of this report.
INTEGRITY EVALUATION PROGRAM Since hard HAZs will occur in the A193-87-to-A36 weldments, a testing program was developed to determine the available load capacity of these weldments.
Tne available load capacity can then be compared to the required design capacity to establish whether or not the f actors of saf ety prescribed by the cesign criteria are satisfied.
To develop such a program, it is necessary to consider the important parameters that will af fect the strength of the weldments made with the A193-B7 material.
It is also important to consider the field welding conditions and what influence they might have on the load capacity of such weldments, in addition, the tests have to be perf ormed in a controlled manner so that the results can be applied to the field condition.
Finally, if cracks may be present in these detail s, the potential for brittle fracture must be evaluated to estabilsh whether or not brittle f racture is likely in the field condition.
All of these aspects have been considered in the work described herein.
The mntallurgy of the af fected weldments has been established, and the test specimens used to determine load capacity have been designed to conservatively represent field welding conditions.
Strength and f racture mechanics analyses have also been perf ormed to conf irm the test resul ts.
OBJECTIVES The objectives of the proposed work are as follows:
o To establish the natallurgical characteristics of the welds that potentially have been used to join the A193 rods to the A36 embedded 1
l-3 plates.
Th is inf ormation is essential to be able to define designs f or test specimens that represent the f ield condition.
It is also necessary to determine the importance of the various welding para-meters within the constraints of the field welding procedures.
o To develop series of test specimens that bound the field conditions in terms of bolt material, size, and welding process.
These test series, representative of field conditions, can then be used to estab-lish a measure of the load capacity for the A193-87/A36 weldments, e To perf orm load tests to establish a statistical measure of the welded rod Icad capacity for the test series identified.
The ' data resul ting from these tests may be used to define an available load capacity.
This available load capacity can be established on a number of bases, including bounding statistics, o To compare the test results with the required load capacity for the design case.
This wil l give an assessment of the consequences of the A193-87 weltments on the structural integrity of the embedded plates.
i s
2-1 Section 2 SCOPE To achieve the above objectives, a program consisting of the following f our tasks was identified:
Task 1 - Design of Test Program to be a Conservative Representation of Field Conditions Task 2 - Monitoring of Tests Task 3 - Analysis of Results and Evaluation of the Significance of the Results Task 4 - Conclusions and Recommendations The following sections of the report detall the results of the tests and ana l y se s.
In Section 3, the basis f or the test specimen matrix is developed taking into account the field welding conditions and their influence on the test data.
The variables to be used in the production of the test specimens and the tests to determine load capacity values are established.
In Section 4, the results of the tests are described, including the metallurgical studies perf ormed by Bechtel (M&QS) and APTECH.
In Section 5, the significance of the results is outlined in a comparison of tested load capacity with required design capacities.
The concl usions and recommendations are outlined in Section 6.
3-1 Section 3 TEST MATRIX DES IGN INTRODUCTION To establish the available load capacity, a series of test specimens were developed to provide load capacity data for the weldment between A193 rod and A36 pl ate.
To meet the objectives of this program, the tests were designed to include all of the ef fects of field welding and to bound the metallurgical conditions that could result in these welds.
The objectives of the test plan were as follows:
o Determine the number of specimens to be tested to establish a measure of Icad capacity o Provide details of weld procedures ;o prepare specimens which conser-vatively represent the field welding conditions o Provide detail s of the test, including test specimen and fixture design and instrumentation o Provide a f ormat for perf orming the tests and recording the resul ts WELDING PARANETERS There are two conditions that may exist when low alloy quenched and tempered (LAQT) rods of A193-B7 material are welded to A36 plates using welding procedures intended f or A36 material.
First, because of the high hardenability of A193 material, a hard HAZ may be fcrmed.
The formation of hard HAZs is controlled by cooling rats (as af fected by arc energy or heat input from the welding arc, preheat, material size, Interpass termperature,
3-2 and material chemical composition) as it af fects the potential to form hard regions with heat treatment (harde na b i l i ty ).
Second, in cases where excess heat may be used during the welding process, it could be postulated that overtempering could lead to a sof tening of the subcritical HAZ and a resultant loss of strength. The weldability tests described herein were designed to simulate these two extreme welding conditions.
A detailed description of each of these conditions folicws.
HEAT AFFECTED ZONE HARDENING The propensity for a material to form a hard microstructure can be estimated from a hardenability index expressed as a carbon equivalent (CE ). This is an unpirically derived relationship that takes into account those chemical constituents of the metal that promote hardening when heat treated.
The CE is based on steel composition.
A f requently used f ormul a f or CE is:
CE C + Mn/6 + (Cr + Mo + Y)/5 +
(Ni + Cu)/15 (3-1)
=
If the chemistry tolerances of A36 are met, then the maximum CE level (assuming a weight % Cu = 0.2) obtainable for this material is 0.46.
In comparison, A193-87 material may have CE levels ranging f rom 0.66 to 0.96, indicating a significantly higher hardonability than A36 material (Tabl e 3-1).
This means that special precautions (such as preheat, high arc energy, post-weld heat treatment, etc. ) would normally be considered when welding these materials to avoid hard HAZs susceptible to cracking.
Since these precautions were not taken (the potentially comingled A193-87 rods were welded I
with a welding procedure intended f or A36 materials), the test specimen weld procedures should reflect the adverse welding conditions that will produce such HAZs.
For the case of the A193-87 rods, the adverse field welding will occur when:
o The rod is the largest size permitted by the design.
For a f ixed arc energy, the largest rod will give the greatest heat sink and,
3-3 Table 3-1 CHEMISTRY TOLERANCES OF A36 AND A193-B7 (WT f,)
A36 Element (3/4" 1/2" Thick)
A193-87 C
0.25 max 0.37 - 0.49 Mn 0.80 - 1.20 0.65 - 1.10 P
0.04 max 0.035 max S
0.05 max 0.040 max Si 0.15 - 0.35 Cu 0.20 min Cr 0.75 - 1.20 Mo 0.15 - 0.25 Carbon 0.46 0.66 - 0.96 Equivalent
3-4 hence, the highest cooling rate.
This, in turn, gives the greatest propensity to f orm martensite on cool ing, a transf ormation product with high hardness.
o The weld is the smallest.
This promotes low arc energy, which again implies high cooling rates and high hardnesses.
o The minimum preheat value is used.
The use of preheat provides slower cooling rates, and conversely, the minimum preheat will give the worst case cooling rate.
SUSCEPTIBlliTY TO CRACKING The adverse field welding conditions that may produce cracking were promoted by introducing conditions conducive to high hardness.
The presence of hydrogen in the weld pool, which may also lead to cracking, was not deliberately introduced since normal fleld electrode control procedures that ef fectively preclude that cracking were used in the prior field welding of plates as well as in the preparation of tne welded specimens.
The ways of controlling the volume of martensite (a hard susceptible microstructure) have been discussed previously. Again, the arc energy, preheat and postheat levels control the amount and type of transf ormation products f ormed.
TEST MATRIX in view of the above considerations, the test matrix shcwn in Table 3-2 was developed to simulate the adverse welding conditions expected to produce degradation of an AI93-87 rod to A36 plate weldment when it is perf ormed using the speci f ied wel ding procedures.
A nominal sample size of ten specimens for each of the three rod diameters was selected initially. The minimum rod size of one Inch for the tests was selected based on the f act that the one Inch rod with a 3/8 inch leg length fillet would harden mere than the 3/4 inch rod with a 5/16 Inch leg length fillet.
Table 3-2 TEST MATRIX AND WELDIfiG PARAMETERS Test f! umber Al A2 B
C D
A193 Rod Dia 1"
1" 1"
1 1/2" 2"
Preheat fi/A N/A N/A N/A 150 F Interpass N/A N/A N/A 50 F min.
150 F min.
Test Po si tio n IF 1G 1F IF IF (Fillet)
(Full penetration)
(Fillet)
(Fillet)
(Fillet)
CE 0.777 0.777 0.795 0.817
>0.8 Rod Heat No./ Code 8866608/ADV 8866608/ADV 93031/AVF 75973/AVM 8895931/AEA Weld Size 5/ 8" 5/8" 3/8" 1/2" 5/8" Y
w Bead Type Root Pass / Stringer Root Pass / Stringer All Passes / All Passes /
All Passes /
Capping Pass / Weave Fill Passes / Weave Stringer Stringer Stringer Capping Pass / Weave A36 Plate Size 4"x4"xl 1/2" 4"x4"xl 1/2" 4"x4"xl 1/2" 4"x4"x1 1/2" 5 1/2"x5 1/2x1 1/2" E7018 3/32" Dia 3/32" Dia 3/32" Dia 3/32" Dia 3/32" Dia Electrode Size Amperage 110-120 110-120 76-90 76-90 76-90 Arc Voltage 25-28 25-28 10-22 10-22 19-22 Travel Speed
- 2.1-2.6 in/ min 2.1-2.6 in/ min 3.1-4.3 in/ min 3.1-4.3 in/ min 3.1-4.3 in/ min Heat input (Max.)
70 KJ/in 70 KJ/in 25KJ/in 25 KJ/in 25 KJ/in fiumber/ Pa s s es
>2
>2 2
>2
>2 Number / Tests 2
2 22 12 12
- Travel speed should be monitored during welding.
3-6 OVER TEMPERING Test Series A, as defined in Table 3-2 prescribes the highest expected arc energy and smallest rod dianeter used, th us, it provides a condition in which a maximum emount of subcritical HAZ tempering would occur.
These specimens were used for metallurgical evaluation only since the annealed or f ully tempered A193 would still be stronger than the A36 bolt material (.1). Test Series B, C and D prescribe conditions whereby minimum arc energy and preheat values were used to weld the A193-B7 rods of diameters consistent with the diameters used on the STP enbedded plates.
These tests were set up to maximize the cooling rates of the weldments and, thus, maximize the potential metallurgical degradation of the welded joint.
For each test series, two additional welded specimens were prepared for metallurgical evaluation so that the expected metallurgical conditions could be confirmed.
This test matrix and parameters permitted representation of the adverse welding conditions that could be associated with the STP welding of the A193-B7 rods to A36 plate.
Thus, using these parameters, a lower bound for
~
the available load capacity of the STP welded rods was obtained.
TEST FEBOD AND PARANETERS in the service situation, a particular embedded plate may be used for dif ferent f unctions, such as beam end connections, pipe whip restraint supports, pipe hanger attachments, etc.
Under all circumstances, the predominant loads on the unbedded plate will result in axial load and shear on the welds. Accordingly, as a conservative approach, all tests were perf ormed in axial tension.
This loading is considered to be the most discriminating case as it subjects the entire weld area to tension, and any expected cracks at the weld toe or root would be expected to perf orm worse under this loading condition.
The design of the specimen was such as to account for bearing stresses, plate thickness, and rod length for gripping.
The test specimen design shown in Figures 3-1 and 3-2 was used for alI tests.
3-7 V'[!
A N
l
,'i r
- 1 I
tit :
- l{,{
i rg.-
wr.- r,
77,,
- r.,r,
? -
M,
'?
I
~_ );, V i.'
W.
9, ';/.
,Q i,
- /,,,,.
y
,3
'h,'2
,'1
- v, i.
- ./
I,.
J
,,,, * /l f'jkkw(,')y,y
.g
\\
A36-ll" Plate
[
3 y
7 j',
Qi 1
)
~
.o 4
4-
,g o
//
. )
6
,.. ORZ flo tes :
l:
0 3 2 c Ro d should be at. least H
18" long to allow for grips ll g
p; i.
a o Rod shall be A193-B7 naterial 5 0 y
E a
E
'e
/
- <:3
~ [./
[
mm > l! ~
j-w Figure 3 Weld Test Apparatus and Specimen.
3-S ww..g; y.
s - - - -- - = -- --
- e v:m
[@n@. -S @
Il s-y.
c
.,W
[
,c
,,m.
..-m w
s s-g=_
y.,t
. g.
g,.
t, <
- o s.ms
. m w_ _ _ _ =
r: +3am,;
": ~,
.ar, M5W q..,
u,.
]
l v
,,c.,._
.t. ?. -j-y# s
'" 4 ' ; _
r
-2 l
8
!l i r_c ' -
i.77 V&.
i - ;,cao,e w4 j
' ;,;p!~e/
y, i
'.a r
.ae x _. ;a :.1 v._
- y v.--
v g.,j s;-
y,...,.
h O/
R..,!!
e 7
i:
O T:
i g
r 1
a E
\\
& Fillet l
L
'.f'
{t,[
\\
2 r.
G
[f p
p
\\
!i p,j a
i d
' "f'
// l
$ 'h
,1 4
s
\\
/'l, L I
. )d o
)
pA e
gn...
1
+
> c c-
,r.
f..; %,
u h
k
- ,./D.
p ;o.
(
t =
- a s
i 6t k4 s..,.. ; ;.w.?
~
- 1 whM TE '
..^
p 4+, w,i..t-!...
u,. p'c.n,ce,p:
D M i'"?" '"b.-'*"#*
Wifd6V D-
'O _-
.. %syyw--s=
c e.
n-2 w
, 2 i'I 3'
- 1. H
= ~U Cl 2.c A.a;fW
{-j s-- &
Clamp p
.nw;.
i;;
= --
ca a e!+
y
= y
_ y-f f
= -
r,p".4 q ;.
-2
- 3. Asi s
- s.t sr>,
y, m.,
, ~ : ),.
g.
- ..r
-- ; e r
s w
= Backing I Plate L
i e
i i
l 31" Long Llamp TYP 3
i f
Sh" Long_ Plate (Backing Plate) TYP
.i l
Plan View figure 3 Schematic o f Backing for Tension Tests.
4-1 Section 4 TEST RESULTS INTRODUCTION The speciaens f or the test matrix identif ied in Table 3-2 were welded at the STP site using f ield welding conditions modified as def ined in Table 3-2.
All welding was done by Ebasco under Bechtel's direction.
APTECH personnel also witnessed the production of some of these welds at the site.
Upon completion, the specimens were shipped to Bechtel M&QS in San Francisco f or visual and dye penetrant examinations.
In addition, metal lurgical examinations and fractographic studies were perf ormed by APTECH and Bechtel M&QS.
The tensil e tests were perf ormed at three testing laboratories based on the capacity of the testing equipment.
A typical test setup showing the dlal indicators f or displacement measurement is shown in Figure 4-1.
Bechtel M&QS started the B-series testing and broke several specimens.
The runalning B-series specinens exceeded their equipment capacity (60 KIPS) and were subsequently tested to f ailure at Anamot Laboratcries.
All C and D series specimens were broken on the 600 KIP machine at Pittsburg Testing Labora tor i es.
Fail ure strengths, weld leg lengths, and crack dimensions, if any, are shown in Table 4-1 (2).
{
L INSPECTION RESULTS The visual inspection revealed that the welds f or all test series (A, B, C, and D) were of acceptable quality, exhibiting some minor areas of undercut.
Typical weld appearance is shown in Figure 4-2.
Dye penetrant examination was perf ormed to detect the presence of surf ace cracking.
No surface cracking was observed.
E l
4-2 h
u.
..jy i
1.,
y e-
../, 3 r
.....e k
(.
.:6
._p
+
.......,.1, g., ". j-;
- 1. t
=
l 4
~
l ;...,.
b.
- ... ;c.. -
y
-~
..q..
., _ _ -.=
.c
~ %; _
.p g
)_'l
- [ l
.-., n.
'.}
c:
e 8
I k
.;h J [.,h$,[i
[
...' [,
y _ ': : :..;. 1 ;,., '
u
. 3
- g g
p '
<'..k f.'
4.y :: :,.[ ' '}. '. ~
4
.=
if'..
t
.. a.,..
a,
- m. y ',
.,~.a
..,t..
- 9. t, _..
.g.
Y.
L
. ; y,
- G)l.
l:
fr5 '..
3 q;;
- s..~.;
{
.c v.
- Q T;
+
- ~
'. :. pi. 9 p ; - v,... g..
.,.y
- - 9 c,.,
m
- ...'. e y.
u. '.'a
.s.
t.p.
... c :- : ' ;
.'>r.
is,
- .. ^:.;
. hx c :.
.,i.
_., y...
.y z
g...
. s g j e
..;h:
~ ~ ~ '
e s.
>V.1. [ 7k j...
. g
'# fy..s
....e, 9
y
=
(A)
(B)
Figure 4 Details Of Test Set-up.
TN c-y
~?
+ k p'. : x g :.
n...
.fi: *.. i '
g...
g 3... ',._.
L.j. 9' I
E' 1
i E (M i
Table 4-1 TEST RESULTS FOR A193 B-7 ANCHOR ROD WELDED TO A36 PLATE Average Specimen Rod Weld Crack Crack Weld Failure No.
Size (In.)
Length (In.)
Depth (In.)
Width (In.)
Strength (KIPS)
B1 1
0.44 55.7 B3 1
0.45 0.12 0.45 63.2 B4 1
0.49 68.8 85 1
0.46 61.8 B6 1
0.44 55.4 B7 1
0.39 47.3 B9 1
0.44 0.2, 0.1 0.4, 0.35 65.0 B10 1
0.44 0.25 0.4 62.2 B13 1
0.47 59.8 B14 1
0.44 67.2 C1 lb 0.54 0.25, 0.2 1.9, 0.3 125.4 C2 1h 0.64 0.1, 0.2 0.2, 0.9 151.0 1
C3 1h 0.58*
0.2, 0.2, 0.2, 0.1 0.5,1.1, 0.2, 0.2 135.0 f
C4 1h 0.62*
0.3, 0.2 0.8, 1.6 142.5 C5 1h 0.59 132.5 C6 1h 0.61 143.0 C8 lb 0.59 147.0 C9 1h 0.56 0.25 2.25 125.0 C10 1h 0.59 0.25 1.25 141.5 C12 1h 0.58 0.15, 0.25 0.8, 0.6 137.5 D1 2
0.73 240.5 D2 2
0.73 241.25 D3 2
0.72 0.2 0.4 222.5 D4 2
0.75 0.25, 0.25, 0.1 3.5, 0.3, 0.25 222.75 DS 2
0.74 247.25 D7 2
0.70 237.5 08 2
0.72 0.1 0.6 226.25 D10 2
0.69 226.5 D11 2
0.75 0.1 0.3 255.0 012 2
0.73 237.75
- APTECH measurements Interoffice memorandum, G. R. Schmidt to R. W. Straiton, Anchor Bolt Evaltation, South Texas Project, Job 146926-001, ORS-094-03 (September 10, 1984).
Interoffice nemorandum, G. R. Schmidt to R. W. Straiton, Anchor Bolt Evaluation, South Texas Project, Job 14926-001, GRS-074-13 (June 27, 1984).
t
4-4 4
f f
^-
a h,
e
? ~
)
e n
e e
Figure 4 Typical Appearance of Welded Test Specimens.
4-5 ETALLURGICAL RESULTS in the viginal test matrix, Series Al and A2 specimens were welded for metallurgical examinatien only.
These specimens were welded using high arc energy with a f il let (Series A1) and f ull penetration (Series A2) weld.
The metallurgical examination was done to conf irm that with these welding conditions (which produce slower cooling rates), the HAZ would not be as hard as for Series B, C, and D.
The specimens were sectioned on a diameter and polished and etched.
Figure 4-3 is typical of the microstructure from these welds.
For Series A2 (f ul l penetration wel ds ) the last pass of the weld was still exceptionally hard (%750 HV10) indicating that even for the high arc energy / slow cooling rate, hard HAZ regions are stilI formed (Figure 4-4).
Martensite typically is formed in such high hardness regions, and an example is shown in Figure 4-5.
Macro sections were also made f rom the two additional specirrens f rom Series B, C, and D, and these are shown in Figures 4-6, 4-7, and 4-8, respectively.
Regions of high hardness (martensite) are present, and in scce cases, root cracks were present.
In conclusion, the of fects of arc energy input and cooling rate do not significantly affect the hardness of the HAZ, and it is not essential to l
impose strict bounds on these parameters in order to obtain weld specimens
{
adversely affected by hard HAZ.
HARONESS RESULTS Microhardness measurunents were made for all test series welded, and the results are shown in Table 4-2 (2).
The peak values of HAZ hardness are summarized in Table 4-3.
There is no apparent HAZ hardness bias with respect to design cooling rate, which simply reflects the very high hardenability of the A193-B7 material.
This means that the variations in welding parameters
4-6
,%a,..-
..w.
~,
pg
- l:3.$;f '..
'~ -
- 4 13,
. t. - C g :- '
, :i.' :,,.
't..; ',. r g.j s.:
p,
V,, '
{x$ ','.'.. -
- .';t. -
v
.... 9. -
- ,, /** '-
e,,; e..:s C)
..g:.
p
- k...fI.
,{
s.
f..
Figure 4 Photomicrograph Typical of the Sections Made in Specimen Al-1.
4-7
-- m,e.
e 4
i.d.s.,....,":.. ;
..qd:.s.=-.
,Sk.E '
Gh >
1
, ;;4
. f.
_Q.
n',
,, a -
,1
. - ~
.y4 f.-
'eA 4'
i?
A -
+ :.
W,.;v "',.
~ >:, - ;.,
,, a b'
J.
5X, fli tal s.
Ah..nelm.eu A:I_..
Fi gure 4 Weld Profile of Specimen A2-1 (Full Penetration Weld).
4-8
.lh : '? g ;
_"q.
f 4
g,... '
,e
- a 7
qq s,
k l
[
4
- y
+
% i 1
e 6
'~ ';;-
500X, fli tal
- _ _ b' 2A w. A A 3
,w~,
Figure 4 Photomicrograph of Detail of Figure 4-4 White Area to Right is ifartensite.
4-9 s s..
,2:.
6
4.,t,y,
y, J h '-
2
- j
,K7'~r~.
~
~. i.
sx, nital Figure 4 Weld Profile for Specimen 82.
4-10 4
L 1
w f
? /*
.Y
,?
r'
,]h' v.
-7
) l'
'I f
.{.
,7 g" u%hf ',g*n;) ' f, ka;;
e. 3, 4., >..
_ s,..s.-
r:
- 4 4
< % w':,
.y i',
gJ; -', i, ap
_ t\\
.'S
[
g
'u_.yf5 Lp Qg.a; w
9..
e
^t ' u 1
.+ 2, _ c. r _
$.)5
'}'
SX, tiital
,J gd.j).SDu,.$ 4 0.-
2 Figure 4 Photomacrograph of Specirren Cll,
4-11 l
l
.,..w, 7,,
- e
,.s ;.N
's
.f kN '~
,,(,,
1 n.A
'd
",*,,4 ~
V y ' : 4-
'c.
g
,p l+_
l to.
.,< ~.W
- ?:
s t
3 g,-
5X, feital
... ~.
l figure 4 Photonacrograph of Specimen 09.
4-12 Table 4-2 MICR0 HARDNESS TEST RESULTS Bolt Plate Specimen we ld Me t a l Number Base llAZ llAZ Base Metal (Note 2)
(Note 3)
Metal Al-1A 333 493, 461, 471 178, 193, 171 191, 189, 176 175 Al-1B 362 498, 566, 560 180, 165, 178 211, 202, 198 174 Al-2A 306 327, 409, 525 193, 177, 183 188, 181, -
157 Al-2B 362 528, 665, 631 198, 170, 178 189, 189, 183 170 t
A2-1A 289 473, 560, 579 185, 180, 175 181, 192, -
155 A2-1B 360 724, 752, 6S1 213, 189, 205 210, 211, 220 177 A2-2A 355 357, 370, 368 185, 177, 170 173, 179, 180 162 A2-2B 342 325, 336, 365 188, 167, 170 181, 178, 182 165 B2A 343 686, 646, 631 207, 200, 204 219, 238, 238 172 B2B 336 669, 743, 752 216, 225, 209 266, 252, 232 167 B8A 281 536, 478, 413 225, 210, 193 213, 218, -
171 BBB 351 642, 711, 703 220, 197, 213 250, 249, 210 162 C7A 319 408, 642, 639 190, 179, 221 210, 193, -
159 C7B 353 563, 649, 698, 703 189, 193, 211 227, 207, 181 172 C11A 332 694, 686, 653 200, 193, 197 203, 206, 192 172 C11B 350 686, 703, 677 192, 198, 198 217, 219, 195 160 D6A 297 566, 642, 599 266, 230, 216 279, 213, 168 D6B 350 453, 466, 301 228, 216, 220 220, 219, 187 172 D9A 336 698, 653, 638 209, 197, 203 192, 194, 202 166 D9B 341 681, 711, 681 201, 199, 202 228, 222, -
167 NOTES:
- 1) See Table 4-4 for typical locations of hardness indentations.
l
- 2) Traverse from edge of base metal to near fusion line, i
- 3) Traverse from near fusion line to edge of bane metal !!AZ.
I 4-13 Table 4-3 PEAK VALUES OF HEAT AFFECTED ZONE HARDNESS Rod Maximum Diameter Hardness Series (In.)
Description (HV 10)
Al 1
High heat input, fillet weld 665 A2 1
High heat input, full penetration 752 weld B
1 Low heat input. fillet weld 752 C
lh Low heat input 703 0
2 Low heat input and preheat 711 l
4-14 4
. outilned In' Table 3-2 have led to insignificant or indistinguishable dif ferences in metallurgical condition. All of the resultant HAZs are very hard, irrespective of the variations introduced in the welding of the
- specimens.
In addition to the hardness values reported in Table 4-2, specific hardness
' traverses were perf ormed in Specimens B8 and C7.
These results are shown in I
Table 4-4' Q).
It can be seen from these results that the hardest material resides in areas heated by the last weld bead and that regions near the root of the weld appear to iiave been tenpered by successive weld passes.
1 i
i ANALYSIS OF RESULTS 1
i Weld Strenaths r-The tested f ail ure ' loads plotted on logarithmic probability scales indicate a log normal distribution (Figure 4-9).
The statistical properties of these I
welded specimens and estimates of the 90% conf idence level, 95% probability
{
population lower bounds are in Table 4-5.
I The statistical interpretation of these lower population bounds is as follows.
it is estimated at the 90% conf idence level that 95% of the population of A193 i
rod to A36 plate weldnents resulting wher the specified weld procedure is used with the given rod sizes will have load capacities of the stated value or j
greater. The data for the one Inch diameter rod indicate at a 90% confidence, l
95% probability level e icwer bound load capacity that is slightly below the minimum specified ultimate load for a one Inch diameter A36 rod (i.e., 90%/95%
lower bound of 45.3 KIPS versus specified ultimate tensile load of 45.6 KlPS).
f for the one and one-hal f Inch and two inch diameter rods, the 90%/95% lower j
bound load capacities of the A193 weldments exceed the specified ultimate loads for A36 rods of equivalent diancter.
Theref ore, ombedded plates
(
potentially containing A193 rods of these sizes will exceed the load capacity of the same plates containing A36 rods. APTE01 considers the 90% conf Idence, i
95% probability level as a conservative bound (Appendix A).
t a
j
.,m
--+--y a
y--er.
g----e y
-+-y q'we y
7 em
,-.9w.
sky 9 - -
ym 39
-a,-
e
.. - -c-am--eo
,-+ -- 9m y
g=yp.y.,hg
4-15 Table 4-4 HARDNESS TRAVERSES FOR SPECIMENS C7A AND 88B C7A Hardness location HV10 Comments 1
228 Weld Metal 2
645 HAZ 3
329 HAZ 4
318 Base Metal i
5 194 Weld Metal 6
360 HAZ 7
339 HAZ 8
?94 Base Metal 9
210 Weld Metal 10 342 HAZ 11 380 HAZ 12 313 Base Metal i
13 325 Base Petal i
l 4
Ha.'dne s s
/
"i Locations v v. ~*4 i.-
1-4 3:_.y.
3_
-y,p..
s
,,s e,
- +,
,1
,i 1
5-8
,,.,.. s,,, '.
- 4 Q :y, 7 ? l' QW' : ~ll
_z 9 - 13 o ;...
,g:7 l M'y ; c ; ;?,Y,, ' '
s m
,_,( l ;!...%r p;-
Colt
- ,c,,,
.h
,; a[ "
Plate C7A, 5X, Nital
4-16 (TABLE 4 Cantinued)
B83 Hardness Location HV10 Comments 1
251 Weld Metal 2
715 HAZ 3
698 HAZ 4
505 HAZ 5
322 Base Metal 6
234 Weld Metal 7
715 HAZ 8
704 HAZ 9
525 HAZ 10 334 Base Metal 11 343 Base Metal 12 384 Weld Metal 13 411 HAZ 14 380 HAZ 15 343 Base Metal 1
Bolt e
Hardness Locations l
e' l
e 9..
\\
l
' h.
1-5 l
ze
- 6 - 11
- llj,
- n E il '
4 'i; "
r 1
- b ;'
12 - 15
..)
ep %
Plate B8B, 5X, Hital
\\,
i
___.________J
l t
J
?O r
e te r
m r
e a
e t
i t
e D
e m
m a
h a
i c
i D
n D
I h
h 9
f c
2 c
9 n
/
n I
1 I
f 8
2 1
1 9
I 5
9 e
c d,
O ta 0
l P
9 6
3
?;
a A
0 o
f t
8
)
s k
d 0 n e
d f
7 a
l R
e 0 n W
i 6
a s
n-r i
d d
o 0 e R
i 5 M o
o
(
r 0 e h
I 4
c v
P n
cn er A
I 0 r O
3 3
u 9
c 1
c A
0 O f
2 f
O u
fo o
s y
h t
t f
0 i g
1 l
n i
n_
U b
e r
a t
b S
5 or P
d le W
9 f
1 4
erug i
F f
I 1
0 1
0 0
0 0
0 0
0 0
0 5
0 5
0 5
3 2
2 1
1 Ex EE"m y%*
m 4-10 Table 4-5 STATISTICAL PARAMETERS FOR SPECIMENS AND REPRESENTATIVE POPULATIONS Rod Size, In.
1.0 1.5 2.0 Ln Mean (KIPS), x 4.09959 (60.3) 4.92575 (137.8) 5.46170 (235.5)
Ln Standard Deviation, a 0.11108 0.063335 0.046345 Number of Specimens, n 10 10 10 K Factor, One-Sided, 90%/95% (1) 2.568 2.568 2.568 Lower Population Bound, Ln (KIPS) (2) 3.81434 (45.3) 4.76311 (117.1) 5.34268 (209.1)
Specified Ultimate Tensile Load for A36 Rod, KIPS, Per Original Design 45.6 102.5 182.2 (1) " Factors for One-Sided Tolerance Limits and for Variables and Sampling Plans,"
Sandia Corporation,tionograph SCR-607 (March 1963),
(2) Lower Bound Limit = x - Ko; Charles lipson and Narenda Sheth, " Statistical Design and Analysis of Engineering Experiments," McGraw Hill Book Company,
- p. 79 (1973).
l l
4
/
(
4-19 ANALYSIS OF DISPLACEhENT hEASUREFENTS The testing system shown in Figure 4-1 or a modification thereof was used to test the welded rod specimens.
The rods were gripped at a convenient distance from the weld (from 11. to 15 inches depending on rod size), and the displacement was measured with a dial Indicator over this gauge length, in addition, on some of the B series tests a near-weld displacement measurement was made by placing a dial indicator to measure the displacement over a gauge length of approximately 0.65 inch, which encompased the weld.
All of the displacement data from the dial Indicators measuring over the rod gauge length (2) were plotted against load, and a typical result from the 30 tests is shown in Figure 4-10.
These load displacement plots indicate that for all tests, the rod is expertencing elasti,c loading only and the controlling element for joint capacity is the weld.
In some cases, some nonlinearity was detected at the beginning of the test as the grips bedded into the bolt and as maximum load was approached. This latter non-linearity is attributed to the large displacements associated with weld pull out and f ail ure.
To estimate the app. oximate strain developed within the weld joint region, the elastic displacement of the rod length (derived from the failure stress and Young's modulus) was calculated and subtracted from the total measured displacement accounting f or the non-linearity. The resultant displacement remaining is that which takes place in the weld joint region.
A weld joint pseudo strain can then be estimated by assuming c gauge length. equal to the weld leg length. The displacements and resultant pseudo strains are-summarized in Table 4-6.
Typically, the tabulated valu'es indicate that the weld joints are capable of undergoins considerable def ormation, which is an Indication of favorable ductile behavior.
b,......
t I
I I
i 150.0 I
Specimen:
C1 120.0 --
,/'f
/
/
/
G
/
S 90.0 -
/
21
/
3
/
Ultimate Load = 125.4 KIPS 0 0.745 in.
~
O
/
60.0 -
/
?
/
/
/
/,/
30.0 -
f
/
/
/
/
/
/./
0.0 -
/
I
{
l l
1
[
t j
i
[
I
[
s
[
s 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 Deformation (In.)
Figure 4 Load Versus Dial Gage Extension.
4-21 V
V o
G) n a
G Q.
e m
m o
c) cn cn 3
a o
to e
o c
oE E 0, O 04.N M > 4 D O > 4 - N N D
.T O CG - D e M - O ? > > >
cn ' e-4 - N O e N N -* N N N D - > M 4 N.s 4 5 rD D - T 4 - O > M.*
{jM O N N C1 - *
- N v4 N > D D 4 D D D.N T T D CD D N D D D T N 4 k
000000000000000000000000000000 u
I o
3 V
-- c
- U "-
C D D T - > M D M 4 N.N 1 1 M t M y N M O 4 4 7.D > b N S P O
[ Q *g Q O - - - - O -4
-4.-* M N N N N N N N N N M e N M N N N N M M
.c OC0000000000000000000000000000 u
ouum m
> om om 000000000000000000000000000000
_J m:
V
.a.
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNN g
O.000O0000000000000000000000000, s
c 5-I I i 4 : : : a t i I e I i -1 i i i i ' i i ; e i i !
O3"22
^
m
'JJ W W L.! W L:.1 W W W W W ld W W W W lt.1 W W W W J.! 'J.I U '.t.1 W W W W W 4
e cas-e s > 4 e O M 4 D > e :o D s D.s m e s 4 4 4 e < < D e e s a g
M N,N N N N N N O4 N N N N N N N N N N N N N N N N N N N N N c E, Q
H m
ae v)
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO cn v H
i cm m
c> -
8 J CL w
m z
w
~
t c; -
o u
- 3 4 T. C0 0 4 D D N D CO > O t h W O N O~ 'B h o~ < tD O~ D O O -t h cn o 5 <'
v y "a7 4 3 0 - 3 0 4 O M M v M 4 D D T 4 0 4 M 4 3 M O O 4 4 > N N):
4 3 N M T M > T T 4 M O > 7 O' T CD 5 3) O 0 00 N N D D e - O
*V m '
--ee
-* 4 M M N N N N M N N e D M D M. (9 M M D T.!
.3 R S, H
m ce ~.-.
O - - ** O O -
m v3cv j
ue
-o OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO, moa i
i m *- c LJ-I c-- p c; O
LAJ U E j
%.o o
-- U CC O% m OC0000000000000000000000000000 o-g E
v.c O00000000000000000000000000000
- oa 3
T D > 4 e > v v N T T v O'i 'N > - > 4 ? a C0 CO N D v O N > D M
- cn.
av c
< < T T 4 :91 4 1 4 D 4 D 4 D 4 D D D D.N h N N.N.N N 0 % h ano m
3o~
mu I"
OCOOC00000O0000000000000000000 oo-x F-m a
m 13 O m
H c) Cw Lw V
umn E,*e O 4
- Ct 0 0 c CD O f9 O N * > 4 M N N O N 4 4 M D ** > - M N
- bc 3 c.
O0ND.N.Oe44SO.NTMN3NN--T340->M2) DOT nm
-4 4 - D M ?? :q M F3 M q M M e-0 M D v M 07 M D e m
Nmc O -.-. -
V * ~j ctL SO l 000000000000000000000000000000 ggy L3
+J w n x"m&
% N CD C # M O N D) N 4 C O D D O O O D D D M D O M D M D O CD.
me-
, Q.cs mo
%m b
D M O - D h D N }.N De4 DUN 7.ND**bO*8NN5N44Dh h,g Z
H D 4) 3 2, D C O 4 D 4 M D M T M < 4 N 1 M W < N N ? M M N D M
- en 4 -
.*-4
- - - 4
.-4 -NNNNNNNNNN Lmm O LL C
I QJ eo C,
.c > >
i E
s-~ o <
O e4 N O M. T O N. - N M e D.N *0.*
~
+o u.--
- M e D 4.N y 4
NMeD4O>4 e.
m Q Q G 3 3 T 2 7J Q U U U U U U U U U U Q Q Q C Q Q Q O O Q m
i.-. ~ m I
i
4-22 FRACT0 GRAPHIC ANALYSIS in addition to the metallurgical analysis to identify the metallurgical structure of the welds,- two specimens that had been tested were examined to clarify the nature of the f ail ure.
a The two specimens examined were C8, which did not exhibit a pre-existing crack, and C3, which contained a pre-existing (i.e., prior to testing) crack near the root of the weld.
J Fracture Surf ace Appearance of Specimen C8 Figure 4-11 shows the f ailed weld af ter testing from Specimen C8.
This view is looking at the cut end of the one and one-hal f Inch dinneter rod and the fracture surface.
A close view of the f racture surf ace in the weld is shown In Figure 4-12.
The origin is at the center of this picture.
This region is shown at a higher magnif ication (13.5X) in Figure 4-13.
Here again, the j
origin region is shown clearly and it appears to indicate Intergranular initiation below the weld surf ace with the f racture propagating towards the surf ace and the root of the weld.
To get a better Indication of the fracture morphology, the specimen was cut and mounted f or the scanning electron microscope (SEM)..A series of pictures depicting the region of the origin and the features each side of it are shown in Figures 4-14 through 4-17.
It is clear from Figures 4-14 and 4-15 that the origin region is intergranular f racture, whereas, each side of the dimple rupture morphology is apparent as tiie crack progresses into 1he weld and to the surf ace and to 1he root of the weld (Figures 4-16 and 4-17).
{
i i
. Fracture. Surf ace Appearance _of Specimen C3 A detailed examination of Specimen C3 was also perf ormed to clarify the
- nature of the pre-existing cracks. A view of the fracture surface with the crack shown as~ the-darker region in the lower part of the picture is shown in Figure 4-18.
=
4-23
~
t.
r
!!f f lll i
m jJI 4
8 Figure 4 Fracture Surface of Specimen C8 1
l 1
4-24 i
i l
)
l l
I l
i i
I l
x.
i s
MAG:3.7X Figure 4 Origin Region of Test Specimen C8.
4-25 s*.
N
,i
.t 4:,.y$
'%-)
g
., vi
, -f%.S.fi
. [."
' ' ?.,l g[' l' b '-
. 7. '.'.
,... lpf,
4 6
.-, c,;g}ir I'
' f?,d,>, t,,,
IAG:~13.5
' ~.',' jf.; j
~., e 4
1 i
I k
_j
4-26 m
,,,,,,.m
.(l ; /h g Ti,yy,s,.j.sni:4 EARS:lbssN b
c y., :w lv.4 e
- -,-. e...,
,-a.,.::,
- a..
.:j
- -. w.,
u
,a v
_. p. nw~m s.
u ~.m, a ;:>
s.. r '.,...
,' i - Q.+ ;; ~%.' l'..,
A,_., g%
6 f.'
a,#,
,' h gj;5GI
)
- i, -
'; '. r,
7 MAG.~12X
.t.,py;.7;,;
s (g,s{e.. a..
,[... :.y t-u : ;_.c.._m,.-l.{.. L i. d,g.
f
...a.n..s y ? o -
n
..r,,,.
es. ~
'.%'s.L..'.<',*)z1>2Ngi^
S$ 'e.
,.+.7,.,.*A
...%4;
-,,p. ~,
- ~% y, ?v y
L -;; %;. s
- e,4_..g;9 9
'g.5 % v:s.w s *%..
. s g,
%2j;gHg;".3'
$*mp;;j' Figure 4 SEM Photograph of Specimen C8 Fracture Surface.
i i
l l
4-27 i
,- ).
.,f,
P
.~
\\,/,.
- t. - i' c.
M
~
-,% Q pc
}
T, - '
,]
e d y
s.
d
'C-
\\
g V
l
) ^j %
hm
,.9 s
O S.
,'~
s 2.
's
-f
_ y4 5 W
N 'f g ~' /
~
MAG:100X
,NA,
'. f*' a, 5,,'s Q ','. " k
.J 1 r,
,' ' '& L+f,.[Q py
{-
r o
s
\\
lt n, e
]
Figure 4 Intergranular Fracture on Fracture Surface of Specimen C8 l
4-28 V.~.p}$.,., dt
- 7},1....
.d)
._ s s
~
'3
..y f-s_
's _ -
- 1. s
,f
'^
- mfe payg.
c.
- 44 ~ ' 5' v '
t
..., L; y'.
(.
.,, 1
.ww z-
} - S ' (- (
s%
/s l
A.'
)'
,4 f +
E!)
l l
1
~ r l, T
.[:
4
.$c r \\
1,4
+
tf MAG:500X s
Figure 4 Region of Dimpled Rupture Towards OD From Intergranular Region.
-w s
--w.,. < -
,_s-
,wa-
---,m-
,-a. -. _ _y m-y
l 4-29 b
N v
p f
ja hp I
s
,, I dp.-py
.,N l7T
.z
.- ( I, s a x\\
r
>4 e
i 3.m $m 9.)
'[
h
's
.s y
- 'qQ.![
s4 w
,?
t fl
-5
- *R' p.
f,
^
5,
~
2,*
k s
o e b_
'.'f'i;fAP ' ". p}#
k MAG:500X fv'i j'.'
l
- %
- , )
. 's Figure 4 Region of Dimpled Rupture Towards ID From Intergranular Region.
l
(
4-30 F
f 1
i l
MAG:~7X e
7 e'
Figure 4 Fracture Surface of Specimen C3 Showing Pre-Existing Crack (Dark Region).
4-31 Again, an intergranular fracture morphology is present near the origin of this crack.
This specimen was cut and examined in the SEM to determine the fracture micromorphology.
Figures 4-19A and 4-198 show the Intergranular nature of the crack al though the surf ace is also covered with oxide. Outside this region, the fracture turf ace shows a dimple appearance typical of ductile f ailure (Figure 4-20).
Both f racture surf aces show that the origin region (for the fracture and the crack) are intergranular and the crack propagates by a ductile rupture mechanism outside this area. This indicates that the strength of this detail is controlled by the strength properties of the material through which the crack propagates.
In summary, the cracks appear to initiate in the hard HAZ of the A193-B7 rod and link back to the root of the weld.
FRACTURE ANALYSIS Although all of the test specimens f ailed at loads greater than the specified ultimate capacity for A36 rods, a fracture analysis was perf ormed to ensure j
that there was no potential for a low stress brittle fracture, it was also noted f r cm the f ractographic analysis of the broken test specimens that the f ailures appeared to initiate below the veld metal surf ace neer the throat of the weld and propagated through weld metal and HAZ in e ductile j
manner in the rod material.
No rod f ail ures were initiated at the weld toe on the rod even though undercut was apparent in some specimens.
To analyze the load capacity of A36 or A193-B7 rods welded to A36 plate using a weld procedure for A36 materials and E7018 welding electrodes, considering weldments containing def ects or pre-existing cracks, various potential failure i
4-32 i
l l
l l
l l
l l
c' y '3
~
.d$
, '(($'.,,_
g.g MAG:20X
- 4'.c,.?.. a i
a aus
]
Figure 4-19A - SEM Fractograph of Specimen C3 in Region of Discoloration.
l
,1 '..
1 I
1
' ~f g
- t., yp,
':g
. r-8
- .% $,6 :.-3. /
b
.,..,. j ** '
I
.4 4's .
-g y '* ;~
'A s,
~
4 f
,[
E.
1 M.
A
/
m.4 %,
N.y'il
. C _,, %
-5 j
a V
K,)
p_l ?'t.,
4'
- ~,
Q
'vL n
.p.
a s
1
-i ~~ _p 9
<T b
k*5J('
i f,-
',4 xc.%
A.
MAG:100X s
1 Figure 4-19B - Region o f Integranularity in Specimen C3.
4
l i
4-33 l
. f.(.
' j '-
i, g '
y r,
i s,
,y S
w f
(,,,
e*
a s'm e
. T7'.f
_9 l 1 [ j ), ' ? )f p ] Q y,..
,h,,
X
"'W, h
j g$,.
' ~ f,*~'l N:
p q, Q, '.'.
A
.hy*
T
~
<g *, ',gr
,t l
\\
I MAG:500X Figure 4 Oxidized Region of Dimple Rupture Towards OD From Region cf Intergranular Fracture in Figure 4-19D.
4-34 I
l mechanisms and flaw locations were postulated.
The following models were 4
proposed:
i Type l A
- Fracture Initiation from toe cracks located around the anchor rods Type IB
- Plastic collapse in the net section near toe cracks located around the anchor rods Type l l A - Fracture initiation from root cracks located parallel to the j
circumference of the anchor rods Type 11B - Shear f ailure of the weld metal or the rod subcritical HAZ Type Illa - Fracture initiation from root crack parallel to the A36 plate i
Type 1118 - Plastic collapse of the weld metal with flaws located in the 5'
plane of the minimum weld throat Figure 4-21 is e schematic showing the location of the postulated f
discontinuities.
Examinations of the metallurgical sections and the fracture j
surf aces of the iest specimens indicate that only Type il discontinuities were present in these wel dments (2). These discontinuities were cracks extending approximately one-quarter inch fran the weld root in a plane parallel to the axis of the ancher rod. They are loca ted in the A193-B7 coarse grained HAZ and have an intergranular fracture a,1d ductile norphology.
Although 1he presence of these cracks on 1he f racture plane suggests that they ecy have participated in the f ailure of the test specimens, there are also test P
specimens in which no pre-existing defects were observed on the fracture surface.
Since the load capacities of the test specimens with pre-existing cracks on the fracture surf ace were similar to the capacities of those with no apparent defects, the reduction in load capacity due to pre-existing def ects of this type with a depth of one-quarter inch or less is minimal (2).
r e y~
e--
e, y.
w w
4 w-r,--r
-e, n,-,
7 y
~m--
e-q
4-35 I
IIIB (jll (f
IIIA Figure 4-21 -Schematic of Possible Flaws Which May Pesult in a Reduction ir, load Capacity of the Embed.
1
4-36 Fail ure of the weldments in the presence of Type 11 cracks could occur by either plastic collapse or brittle fracture initiation.
Both of these f ail ure mechanisms have been addressed.
Since it is clear f rom the f ractography that a ductile f ail ure mode predominates in the tests, the analysis methods developed f irst ere based on limit load.
To conf inn that there is no potential f or low stress brittle fracture, fracture mechanics analyses have also been perf ormed.
PL ASTIC COLLAPSE Plastic collapse for the weld conf igurations tested will occur when the net section shear stress, 0 is greater than or equal to the shear strength, T
, of the uncracked l igament.
The critical load, P, can thus be crit c
obtained by multiplying the critical shear stress, T, by the net area in c
the presence of a crack.
P
= T x Area (4-1) c c
net The critical sheer strength of E7018 weld metal was calculated as the value of the tensile flow stress, c, divided by vi, as is considered appropriate based on the Von Mises yleid criteria.
For limit load analyses of steel structures, it is common to use the average of the yield and ultimate strengths f or the flow stress.
This results in a value of critical shear stress of 60 ksi + 70 ksi
{_
- =
33 ksi (4-2)
=
.3 where, 60 ksi and 70 ksi are the specified minimum yield and ultimate strength of the weld metal, respectively.
4-37 The net section area of any particular rod was calculated by the f ollowing eq uation:
Area n x Naninal Rod Dianeter x (Weld Leg Length
=
net
- Crack Depth )
(4-3)
Figure 4-22 shcws the results of this analysis corresponding to collapse of E7018 weld metal.
This analysis is based on a crack in the weld throat plane, al l the way around the wel d.
The cracks observed on the test specimens and the subsequent crack path were in part located in the A193-B7 material (2), thus, root cracks would result in an increase in the net section shear stresses in the A193-87 HAZ more than in the E7018 weld metal. The plastic collapse analysis was, thus, perf ormed a second time using the strength properties associated with a A193-87 material tempered at 1290*F.
This should provide the strength value f or the sof test HAZ region (i.e., the subcritical HAZ). Figure 4-23 shows the hardness of A193-B7 material as a f unction of tanpering temperature.
For a 1290*F (about 700*C) temper, the hardness is expected to be HRC 24.5.
Using a standard iable (see ASTM A370), this hardness value correlates to an ultimate tensile strength value of 118 ksi. Figure 4-24 shows ultimate tensile and yield strength values f or A193-B7 tanpered at slightly lower temperatures.
The ratio of the yleid strcngth to the ultimate tensile strength is approximately 0.9 for these data.
Using this relationship, the yield strength for a 1290*F temper is estimated at 106 ksi.
Usin0 these values, an estimated shear strengtn is given as:
118 ksi + 106 ksi 2
64.7 ksi (4-4)
=
/T This value was used to generate another series of load capacity versos crack depth to weld leg length ratios and is included in Figure 4-25.
As is shown, the calculated load capacities of the embeds with axisymmetric weld root
150--
130-x 120 -
N 7a Dia.
0.625" U 110-x 100-90 -
E
.5" Dia, 0.V d G 80 -
D 70 -
la Dia =
" LL
'G E 60 -
N N
n N
5 50 '
C" Dia, 0. 8 u 40 -
~
30 -
20 -
N x
10 '
0 r
t i
I 3
\\
i 0
0.2 0.4 0.6 0.8 1.0 Crack Dept))
Weld L*9 L 9 Figure 4 Load Capacity Versus Crack Depth to Weld Leg Length Ratio For Plastic Collapse of E7013 Weld Metal (See Type II Cracks in Figure 4-21),
4-39 T. mw..e3. "....i.. e ' l 27 290 at:
67' sw icco i100 taca es r;
i ri rm r T '
m....
i i
j i
i i
i
.._4-_. j__j I
l
.i i
__._ 4. -._... [
,2.___._
I I
l i
t c t.__..__ L. _._. _
t
.3 c
ict 2c:
)ca ex sce sa tea sca t,. v... m.,
c Figure 4 Hardness Versus Tempering Temperature for A193-B7
!!aterial (1).
l IP
- C f.ed O' g
}
g$Q
~
~
- _ -
.s, e.., x 3,
gq i- -
7,g 7 p _.
10 0 b-
--+- -
l l
l l
d
,,..e,.,....,..,
C0 400 8bO L200 0 400 6GO 1200
- r. y..n,.a... r Figure 4 Yield and Ultimate Tensile Strength Values for A193-B7 in Various Conditions and at Various Temperatures (1).
i 260 -
i 240 -
2" Dia., 0.625" LL 220 --
x\\
200 -
180 -
160 -
2 N 1.5" Dia., 0.5" LL
.S 140 -
ti 120 -
- 0. 2 100 -
\\
b v
g 80 -
1" Dia 0.375" Li 60 -
\\
40 -
20 -
NN 0
i i
i i
i i
i 0
0.2 0.4 0.6 0.8 1.0 Crack Depth Weld Leg Length (LL)
Figure 4 Load Capacity Versus Crack Depth to Weld Leg Length For Plastic Collapse of A193-87 Subcritical HAZ (See Type II Cracks in Figure 4-21).
4-41 cracks located within the A193-87 HAZ are higher than the corresponding load capacity based on the weld metal properties for the same flew size.
With larger flaws present in this region than in the weld metal, this region could control.
The tested f ailure loads correspond to values analytically predicted f or very small flaws by the f oregoing model, which considers a crack all the way around the HAZ in the rod material.
Therefore, it appears that the assumption of an axisymmetric continuous crack in the HAZ is too severe and an analysis considering short aspect ratio cracks would be more realistic.
The model as postulated, however, does indicate that there is very little influence of the cracks in the HAZ on the measured f ail ure loads.
As long as the cracks located in the weld HAZ do not appreciably increase the stresses in the weld metal, then significant cracks can reside in the A193-87 HAZ without a loss in load capacity since even without any def ects in the weld metal, the weld metal will be stressed above its critical shear stress prior to the net section stress in the HAZ becoming critical.
This means that f ail ure through the weld metal is likely in the tests since the limiting strength condition is the weld metal. This is, in fact, what occurred in the tests.
Even with cracks present, ductile f ailure of the weld occurred.
FRACTURE EOiANICS ANALYSIS Root Cracks As a f urther check on the ef fect of the A193-B7 HAZ cracks, a linear clastic fracture mechanics analysis was perf ccmed to mode! these cracks.
The cracks were modeled as edge cracks loaded by Inplane shear stress.
For this case, the Mode li stress intensity f actor, K is given (4):
K T/Ea F (4-5)
=
4-42
- Where, 1
Shear stress in the crack plane
=
Crack length a
=
b Thickness
=
F Mode iI stress intensity correction f actor = f (fIaw and
=
specimen geonetry) 2 3~
1.122 - 0.561 (f) + 0.085 (f) + 0.180(f) _
=
/1 - a/b Fracture initiation wilI occur when:
K
<K (4-6)
IIC ll However, for this particular detail and materials, no K data are available.
In the f racture testing of steels, K
<K IC llc, and hence, K
values (the critical Mode I stress intensity factor) were used.
A crl+Ical stress intensity f actor for the A193-87 HAZ was estimated from Ref s. (5) and (f).
A value of 30 ksi/E was used, which represents a lower bound of the reported values in Ref s. (3) and (f).
Figures 4-26 through 4-29 show the results of this analysis ccrresponding to f racture initiation in A193-B7 coarse grained HAZ. Again, it is found that failure would be expected to occur by shear f ailure in the weld, even if pre-existing def ects are present in the A193-87 coarse grain HAZ.
Since the strength properties used in these analysis were not based on the actual shear strength and fracture toughness levels of the A193-87 HAZ and E7018 weld materials but rather were based on lower bound material properties, the calculated load capacities are less than the capacities actually measured
150 -
140 -
130 -
120 -
110 -
7 Fracture initiation in A193-B7
- ** 9# "
"^
100-
,y 90 -
Plastichollapse of A193-87 80 -
0 subcritichi HAZ 70 -
3
?
60 -
g 50 -
Plastic collapse of 40 -
30 d 20 -
10 -
0 i
i i
i q
0 0.2 0.4 0.6 0.8 1.0 Crack Depth held Leg Length (LL)
Figure 4 Load Capacity Versus Crack Depth to Weld Leg Length Ratio For Type II Cracks (1" Diameter Rods With 3/8" Fillet Weld (0.375" LL)).
- I l
/
.[
o
/
0 1
g e
,8 h
t 0
g neh L t i
gW e
L s d
d o l R eW r
)
e L
ot L
t e m 6
(
h a i.0 h h
ti t
t pD p
g e
e n D"
)
D e
l)
L k( L k
c L
c g
as Z
a e rk 5 r L C c2 fA C
a6 oH d
sr el l
uC0 e
s
(
s, a W
rI rc eI d ai
,4 V
l l
t ee l i 0
ypW or t y cc iT t b
c e
n cu arl i
i s pol t
aFi n
s3 C
F oZ a9 o
iA l 1 di" tH P A a
at 8 oa/
id f
LR5 t e o
in ni l
ia sa 2
7 r
pt i.
2
\\ae eg 0
r l m ue 4
l t s od e
cr cl r
aa e
u ro cw F c g
i i
t8 F
s1 a0 l 7 PE
,0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 5
4 3
2 1
0 9
8 7
6 5
4 3
2 1
1 1
1 1
1 1
^ $ E 3 g2 O
150 -
Fracture initiation in A193-B7 140 -
coarse grained HAZ 130 -
\\
120 -
110 -
Pla ic collapse of A193-87 subtriF cal HAZ 100 -
90 -
Plastic collapse'h 7
E7018 weld metal
.S 80 -
5 70 -
x
?
3
{
60 -
50 -
0 40 -
N 30 -
20 -
\\
10 -
0 i
i i
i i
i i
0 0.2 0.4 0.6 0.8 1.0 Crack Depth Weld Leg Length (LL)
Figure 4 Load Capacity Versus Crack Depth to Weld Leg Length Ratic For Type II Cracks. (1-1/2" Diameter Rods With 1/2" Fillet Weld (0.5" LL)).
.z.
~
.a t
vc
,,300 -
\\
280 -
k~Fractureinitiationin
~
- 260 -
A193-87 coarse grained HAZ 240 -
.,~
220 -
'x, 200 -
E 180 -
Plastic collapse of A193-87 L
subcritical HAZ 160 -
3 Plastic collap of
" 140 -
5
?
P 120 -
E 3 100 -
80 -
60 -
40 -
20 -
0 i i
i i
i i
i i
0 0.2 0.4 0.6 0.3 1.b Crack Depth Weld Leg Length (LL)
Figure 4 Load Capacity Versus Crack Depth to Weld Leg Length Ratio For Type II Cracks (2" Diameter Rods With 5/8" Fillet Welds (0.625" LL)).
4-47 on these embeds.
This simply shows the conservatism of the engineering models used in these plastic collapse and fracture mechanics models and that pre-existing cracks could be present in the A193-B7 HAZ near the weld root with no associated loss in load capacity.
Summary The result of the analyses indicate that the plastic collapse mechanism controls f or realistic flaw sizes.
This substantiates both the test results and the ductile fracture morphology observed in the broken specimens.
Composite curves for the three specimen sizes are shown in Figures 4-26 through 4-29.
t t
5-1 m
Section 5 SIGNIFICAN OF THE RESULTS i
INTRODUCTION lt is clear from the foregoing discussion that in the A193 rod weldments 1
under consideration there is no potential for low stress brittle fracture and that the f ailure is controlled by limit load of the weld and HAZ. The tested f ailure loads for the Al93 rod weldments indicate no degradation with respect to the required load capacity dictated by the original design based on A36 rod weldments.
For the one and one-hal f Inch and the two inch dianeter rods, the j
f ailure loads represented by the average test values (138 K and 235 K) as well
}
as the statistically projected lower bound values (117 K and 209 K) are distinctly' higher than the specified ultimate load capacity of 102 K and 182 K f or A36 rods of one and one-hal f Inch dianeter and two inch dianeter, J
respectively. For the one inch disneter rods, the failure load average test value (60.6 K) is higher than the ultimate load capacity, and the statistically projected lower bound (45.3K) is equal to the ultimate load capacity of 45.6 K f or one Inch diameter A36 rods.
In this Section, the implications of the established ultimate load capacities on the design margins 1
is discussed.
DESIGN BASIS
]~
The allowable design load for anchor rods welded to ombedded plates is determined from the allowable stresses prescribed by the AISC specification (1) multiplied by the corresponding cross-sectional areas of fillet weld or j
rod, whichever results in the lower design load.
,w e
m g
,--r
l 5-2 l
l The allowable load for fillet welds is specified as follows:
(0.3 x o
( 16 ) Aw (5-1)
P
=
u
- chere, A
Area of the weld
=
w 4 x (number of sixteenths of fit let weld size)
=
w 4
Length of fIl let weld ND
=
=
D Dianeter of rod
=
Ultimate tensile strength of weld metal ( = 70 ks i f or a
=
E7018 deposits) 0.7 07 Sin 45*, app!Icable for ef fcctive throat of equal-legged
=
fillet welds i
- Hence, P
0.928 A (5-2)
=
W W
This allowable load corresponds to a f actor of saf ety (FS) with respect to the ultimate strength of the weld metal as follows:
FS 3.3 (5-3)
=
=
The allowable load for rods in tension (P ) is specified as follows:
t P
0.6 F A (5-4)
=
t y y
5-3
- ahere, A
Nominal area of rod
=
Y P
F Yleid strength of rod metal
(= 36 ksi f or A36 rods )
=
Y
- Hence, 4
P 22 A (5-5)
=
t
{
This allowable load corresponds to FS with respect to the yield strength and the ultimate strength of the rod as follows:
FS 1.67, and FS
=
=
~
06 U
it is noted that for rods in shear the allowable load is lower and does not govern f or the comparison to fillet weld allowabic loads which are spect fled as the same f or tension and shear.
1 i
Thus, the FS for fillet welds is higher than that prescribed f or structural steel, which reflects the variability of strength anticipated in production welding.
The strength is sensitive to the individual workmanship, as represented by size and soundness of the fillet weld, and to the actual material strength of the deposited weld metal. These natural variations anticipated in weld strengths are evident in the test results, and according!y, the low values, as well as the high values, are not 1o be construed as representative values to be used as the generalized strength for the production welds.
For this reason, the test results have been considered as part of a natural: distribution of val ues (i.e., log normal).
'The-FS of 3.3 is generically app!!ed by the AISC Specification, as well as -the l
AWS DI.1 Welding Code, to all flilet welds regardless of whether the welds are stressed in'the longitudinal or the transverse direction.
Also the generalized FS of 3.3 is preserved not withstanding that the test results of i
l
5-4 Ref.. (3),' which are AISC conf irmatory tests f or the speci f ication provisions, Indicate' minimum FS ranging from 2.67 for. longitudinal to 4.06 for transverse fillet welds.
In view of the foregoing considerations, it is recognized that in order to address the adequacy of the tested ultimate loads of fillet weld with respect
~
to the design criteria, it is necessary to compare the ultimate loads with the governing allowable design loads.
The objective is to demonstrate that the achieved FS are consistent with the nominal FS of 3.3 and that the FS variations from individual tests are within acceptable bounds, rather than solely demonstrating that'the tested ultimate loads are higher than the rod ultimate load. The corresponding comparison is presented in Table 5-1.
The tabulated results indicate that the required ultimate capacity to obtain the nominal FS of 3.3 is satisfied by all of the test results except the three 2
lowest test value f-or one inch dianeter rod.
These cases af f ord FS = 2.7, 3.2, and 3.2, and are acceptable with respect to the lower bound of 2.67 f or FS from Individual test results as reported in Ref. (E) and is consistent with 4
the variations in strength that are anticipated f or fil let welds as described therein. The 905/95% statistical projection for the lower bound of available load capacity results in a FS = 2.6.
For similar reasons, pertaining to.the natural verlability expected in weld strengths, this is considered to be acceptable as a lower bound evaluation of the FS.
in summary, the statistically projected lower bound (45.3 K) for the available ultimate load capacity derived f rom tests satisfies the specified ultimate load f or A36 rods, the tested ultimate capacity average values for each rod size satisf y the required f actor of saf ety (3.5, 3.9 and 4.0 > 3.3), and all of the 30 test results have load capacities higher than the A36 rod ultimate load and af ford f actors of saf ety higher than the nominal FS of 3.3.
i A comparison of the FS for each of the tested series is shown in Table 5-1.
s i
4 l
5-5 I
Table 5-1 FACTORS OF SAFETY DETERMINED FROM TEST LOADS 1" &
lb" c 2" &
Design Allowable Load:
A36 rod 17.3 38.8 69.1 E7018 weld (fillet size) 17.5 (3/8) 34.9 (1/2) 58.2 (5/8)
P ble (g verned by) 17.3 (rod) 34.9 (weld) 58.2 (weld) allo Required Ult. Load to Develop P with a
FS = 3.3 57.1 115 192 Ultimate load of A36 rod 45.6 102 182 4
Tested Ultimate Load Capacities'and Corresponding Factor of Safety (FS) with Respect to Allowable J
Load Pmean (FS) 60.6 (3.5) 138 (3.9) 235 (4.0)
Ptest lowest (FS)-
47.3 (2.7) 125 (3.6) 222 (3.8) 90/95 (FS) 45.3 (2.E) 117 (3.4) 209 (3.6)
P
5-6 DYNAMIC LOADING CONSIDERATIONS As noted earlier, the unbedded plates with welded anchor rods perf orm a variety of functions including the support of pipe whip restraints.
Loading rates under pipe whip events from a double ended pipe break are usually rapid, and design basis force / time f unctions are used to define the applied loads.
However, the energy absorbing characteristics incorporated into the design of alI restraint structures, produce rates of loading transmitted to the anchor rods that are substantially lower than the postulated forcing f unction for the applied loads. As outlined earl ier, al l tests were perf ormed using slow (static) loadi ng con f orming to ASTM A370. The question of "what of fect wil I impact loads have?" needs to be addressed.
As the loading rate on a steel structure is increased, the properties of the materials may change.
In general, as the rate of loading increases, the yleid stress increases and the toughness decreases.
Therefore, under dynamic loading, the toughness may decrease.
This effeet is most marked in low strength structural steel and becomes less pronounced with higher strength steels.
Detailed analyses have been perf ormed to establish loading rates for dif ferent types of structures (.9).
These have been compared to the loading rates that can be expected in impact (or dynamic) fracture tests.
Loading rates are generally expressed as stress intensity f actor rates or R in units of ksIS/sec.
Fracture test samples experience much higher loading rates 5
($ NIO ks tS/sec) than structures, including bridges, under most types of loading. Drop forging presses and ships in collisions are two cases that will generate loading rates similar to those experienced in impact tests.
In the present case (i.e., anchor rods) that may see attenuated rapid loading from pipe whip events, the highest (bounding) loading rates may be calculated from the force / time f unctions used in the design analysis.
This analysis has been done for the stainless U-bar type of pipe whip restraints, and peak values of d are less than those experienced in impact tests.
This analysis was also perf ormed for the energy absorbing material (EAM) type of pipe whip
5-7 restraints, and the force / time response determined f rom tests of EAM indicate loading rates with k values less than these experienced in impact tests.
In the fracture analysis, a lower bound value of 30 kst/Iii was used for the toughness of the A193 weld region.
This is a worst case toughness and would not be expected to change with Impact loading.
In any event, the load capacity is controlled by limit load analyses, which under dynamic loading, imply increased strength properties.
d
.~..
i' 6-1 1
)
h i
Section 6 CONCLUSIONS AND RECOPENDATIONS l
The program of work described in this report was aimed at establishing the integrity of embedded plates that may contain A193-87 anchor rods welded in place of A36 rods.
The fof lowing conct usions were reached as a resul t of the 4
j telding and load capacity tests, analysis of metallurgy, and f racture j.
predictions:
e For the A193-B7 material, which is readily hardenable by heat treat-ment (or welding), peak HAZ hardness values are similar irrespective of ^ preheat levels, thickness, weld Interpass temperatures, type of weld, and arc energy levels.
In other words, the material is readily j
hardened under all welding conditions used in the test matrix.
l o Peak hardness values measured in the A193-87 HAZ were of the ceder of 750 HY10.
e The welding procedures selected resulted in hard martensitic HAZs in j
which some cracks occurred.
e The welding procedures used in the field f abrication of test specimens j
bounded the field welding parameters f rom the field welding procedures i
and are, theref cre, representative welding conditions.
j e All tests were in accordance with APTE01 requirements.
DI> placement j
raeasurements were made for each test as a f unction of Icad.
l e in spite of the f act that the potentially commingled A193 rods have been welded with a nonconforming welding procedure, there is no 1
l i
i
- -~.
6-2 j
[
degradation in load capacity.
The tested ultimate capacity average value for each rod size satisfies the required f actor of saf ety, and all of the test results exhibit load capacities higher than the A36 i
rod ultimate load.
l e The fracture and limit load analysis indicated that there was little potential for brittle fracture with the observed flaw sizes.
This was confirmed by the test results.
. o in view of the above conclusions, it is proposed that the potential' i
use of A193-87 rods welded to A36 plate, using a welding procedure specified f or A36 material, does not represent a degraded condition, and the enbedded plates can be used "as is."
]
l 4
t t.
)
i i
i i
i 3
1 1
1
)
4
R-1 REFERENCES 1.
American Society For Metal s, Mntal s Handhock, 9th Edition, Vol. 1,
" Properties and Selection:
Irons and Steel s" (1979).
2.
Interof fice Memorandum, G. R. Schmidt to R. W. Strat ton (Bechtel Materials and Quality Service Department), " Anchor Bolt Evaluation, South Texas Project," GRS-074-13 (June 27,1984).
3.
Interoffice Manorandum, G. R. Schmidt to R. W. Stratton (Bechtel Materials and Quality Service Department), " Anchor Bol t Evaluation, South Texas Project," GRS-094-03 (September 10, 1984).
4.
- Tada, H., Stress Intensity ni Cracks Handbook, Del Research Corporation, St. Louis, Missouri (1973), Pp. 229.
5.
Air Force Materials Laboratory, Damage Tolerant Design Handbook, MCIC-HB-01 (January 1975).
6.
Steigerwald, E.
A., " Plane Strain Fracture Toughness of High Strength i
Steel Materials," Engineering Fracture Mechanics, Vol. 1 (1969),
Pp. 473-494.
7.
American Institute of Steel Construction, Inc., Manual ni Steel Constr uct i on, 7th Edition, New York, NY.
8.
" Proposed Working Stresses Fcr Fillet Wolds in Building Construction,"
AlSC Ennineering Journal (January 1969).
9.
IlW Conmission, "Some Proposals For Dynamic Toughness Measurements,"
UK Briefing Group On Dynamic Testing, Welding Institute Conference (1978).
1
[
\\
A-1 Appendix A REW4ENDATl0NS FOR STATISTICALLY-BASED LOWER BOUNDS I
l A-2
~
l 1
\\
l lNTRODUCTION The strategy used to estabt Ish the load capacity of a single embed red employed an analysis of tensile load data of exemplar materials and f abrication.
The uncertainty in expected load capacity was conservatively assessed by the following assumptions:
o The welding parameters were purposely selected in the f abrication of the test samples to promote a worst case situation with regard to ductility in the weldment.
e The test results were evaluated by statistical analysis to establish a lower bound estimate of load capacity.
By adhering to the above conditions, a conservative estimate of load capacity wilI be estabtIshed.
The statistical treatment of the data provides the assurance that the population of unbeds will have at least the established capacity level given a stated probability of occurrence and a stated level of confidence.
The criteria for probability and confidence levels are developed in this Appendix.
STATISTICALLY-BASED BOUNDS A statistically-based bound is more appealing than a simple lower bound treatment of data because it provides a rational basis f or estabilshing a conservative limit to both present and f uture observations.
In the statistical treatment of the test data, a minimum or one-sided tolerance limit for the population of unbed welds was established based on standard statistical methods and suitable criteria.
The load capacity where ihere is X$ probability that past and f uture observations will f all at or above some minimum level with a confidence of Y% is expressed as PlPlF 2 F
} > X$ l
= Y$
(A-1)
A-3 Where, P { } is the symbolic representation for the probability statement, F is the load, and F is the load capacity. Equation (A-1) is a definition for c
load capacity in statistical terms. The values selected for X and Y are, theref ore, the parameters that constitute the criteria in the determination of-load capacity.
STATISTICAL (RITERIA Although the criteria selected f or use in statistical analysis can be arbitrary and usually dictated by the judgements of the analyst, an attmpt was made to select probability and confidence levels that are both reasonable (not overly restrictive) and consistent with Industry practice.
In the computations of load capacity, a 90%/95% (Xf/Y%) and a 95%/95% were both determined as part of a sensitivity study.
It is recommended that the 90%/95%
criteria be used on the basis of precedents established in the Industry.
A probability level for occurrence was selected on the basis of the criterion used in NUREG-0577 (A._l) for determining a bound on toughness data for component support materials.
The materials and applications of component supports are similar to those used in embedded plate systems.
In support of this criterien are the cesign allowables established in MIL-HOBK-5 (A-2) for use in arriving at design values for aerospace structures and elements.
The A-basis and B-basis properties in MIL-HDBK-5 have ascribed to them a 99%
and 905 probability level, respectively. The B-basis criterion establishes the material design values for components or elarents where no single f ailures l
of that component cc element wilI cause failure of the complete system to perform its function.
It has been judged that this criterion is consistent with the application of embed systems.
With regard to the level of conf idence that is acceptable f or bounding material properties, a 95% confidence level is used in both A-basis and B-basis allowables.
Hence, in the analysts, a 95%
confidence level is used for computing the tolerance limits.
L
A-4
SUMMARY
A statistical treatment of the test data wilI provide a rational basis for establishina a conservative estimate for load capacity.
It is recom:nended that a 90%/95% criterion be used in establishing the load capacity of an embed rod on the basis of industry practice.
i e
r A-5 REFERENCES A-1 Snalder, R.
P., et al., " Potential For Low Fracture Toughness and Lamellar Tearing on PWR Steam Generator and Reactor Coolant Pump j
Supports," U.S. Nuclear Regulatory Commission Report, NUREG-0577, Rev.1 (Octobe 1983).
A-2 Moon, D. P., and W. S. Hyler, "Guidel ines For the Presentation of Data,"
MIL-HDBK-5, AFML-lR-66-386.
9 l
1
-