.

NASA Technical Memorandum 89052

Acceptance Criteria For Welds

In ASTM A106 Grade B Steel Pipe And Plate
(NBSB-Ttl-89052) ACCEFITANCE CLISERIA PO& Nt37-11e74
WELDS IN A S T B A106 G R A D E B STEEL PIPE A N 6
PLATE : N A S A ) 46 p CSCL l l F
Unclas
G3/26 44847

C. Michael Hudson, Davis B. Wright, Jr. and

Brian N. Leis

October 1986

National Aeronautics and

Space Administration
Langley Research Center
tiampton, Virginia 23665
ACCEPTANCE CRITERIA FOR WELDS IN ASTM A106 GRADE B STEEL PIPE AND PLATE

C. MICHAEL HUDSON,* DAVIS B. WRIGHT, JR.* AND

BRIAN N. LEISIHc

SUMMARY

Several National Consensus Codes contain acceptance criteria for indications in welds.

Welds containing indications which fail to meet these criteria are unacceptable and must

be repaired. The technical literature indicates that these criteria are based upon

workmanship standards rather than upon the welds' structural integrity. The findings

from NASA-Langley's RECERT Program support this indication. This Program has

identified thousands of unacceptable welds. Some of these welds have been in service for

f i f t y years without exhibiting evidence of structural degradation. Thus the current

acceptance criteria appear t o be very conservative from the standpoint of structural

integrity.

Based on the RECEKT Program findings, NASA-Langley tunded a tatigue study oi coae-

unacceptable welds. This study was conducted to aeveiop usage curves w i i i d i we1-c: baed

on the structural integrity of the welds. The details ot this study are presented in NASA

CK-1781 14. The information presented herein is a condensation and reinterpretation of

the information in NASA CH-I 781 14. This condensation and reinterpretation generated

usage curves for welds having (a) indications 0.20-inch deep by 0.40-inch long and (b)

indications 0.195-inch deep by 8.4-inches long. These curves were developed using the

procedures used in formulating the design curves in Section VIII, Division 2 of the

American Society of Mechanical Engineers Boiler and Pressure Vessel Code.

* NASA-Langley Research Center

++ Battelle's Columbus Laboratories
 INTRODUCTION

Reference 1 contains typical acceptance criteria for indications in welds. Welds

containing indications which fail to meet these criteria a r e unacceptable and must be

repaired. Reference 2 indicates that these criteria are based upon workmanship
standards rather than upon t h e welds' structural integrity. The findings from NASA-

Langley's RECERT Program support this indication. (Reference 3 describes the details

of the RECERT Program.) This program has identified thousands of unacceptable

welds. Some of the most severely unacceptable welds were probably on t h e verge of

failing. Their identification and repair may have precluded serious accidents at NASA-
Langley. The remainder of the unacceptable welds exhibited no evidence of structural

degradation. Some of these welds had been in service for fifty years. Thus the current
acceptance criteria appear to be very conservative from the standpoint of structural

integrity. Heference 2 refers to this conservatism in its discussion of these acceptance

criteria.

Based on the HECEHT Program findings, NASA-Langley tunded a fatigue study of code-

unacceptable welds. This study was conducted to develop usage curves which were based

on the structural integritv of the welds. Battelle, Columbus Division conducted this

study, and presented their findings in reference 4. The information presented herein is a
condensation and reinterpretation of t h e information in reference 4.

2
 TESTS

Specimens

The study used four specimen configurations tor the fatigue lite tests; flat plate, pipe-

wall-segment, pipe segment and service pipe. The following paragraphs describe these
configurations.

Flat plate specimens. f i g u r e 1 shows the configuration of the flat plate specimens.
(Note: Inspection of f i g u r e 1 will clarify the following discussion). To tabricate these
specimens, sections of 8-inch diameter pipe were tirst c u t t o the desired length. These
sections were then cut lengthwise to form two, equal-sized curved shells. These shells
were flattened and normalized to relieve flattening stresses. (The normalization process

involved heating the flattened shells (plates) to 16OO0f for 30 minutes and air cooling.)

The normalized plates were saw c u t a t two locations and the edges at both locations

were beveled tor welding. l h e plate pieces were then welded together t o reform the flat

plates. The welds were radiographed and found to be code-acceptable. The plates were

then normalized again and t h e specimens were machined trom these plates.

Fifteen pertinent specimens were tested. For twelve specimens, planar notches were

electrodischarge machined into t h e welds. These notches simulated linear indications in

the welds, e.g. luck-of-fusion and lack-of-penetration. For three specimens, volumetric

notches were electrodischarge machined into t h e welds. These notches simulated three-
. dimensional indications in the welds, e.g. porosity and slag. Table I describes the depths

and widths of these semi-elliptical notches. Figure 2 describes t h e radii a t t h e base of

the notches. figure 3 shows how notch depths a r e measured trom the inside surface of

the plate.

3
Test conditions dictated that the backing rings be intact in some instances and cut in

others, Similarly, they dictated that the weld reinforcement be l e f t unground in some

instances and ground in others. Table I ulso describes the conditiori of the backing ring
and weld reinforcement tor euch specimen. The impact of the backing rings and weld
I)
reinforcement are reviewed i n the "RESULTS" section of this paper.

Pipe-wall-seqment specimens. Figure 4 shows the configuration of the pipe-wall-

segment specimens. (Note: Inspection of Figure 4 will clarify the following discussion).

To fabricate these specimens, sections of 8-inch diameter pipe were cut t o the desired

length. These sections were then cut lengthwise t o form two, equal-sized curved shells.

These shells were saw cut a t midlength to form two cwved half-shells. The saw-cut

edges of these halt-shells were beveled and the two half-shells were welded together to

reform the shells. The shells were then normalized. After normalizing, each shell was

cut lengthwise into six specimen blanks. These blanks were machined to the final

configuration.

Ten pertinent specimens were tested. For five specimens, planar notches were

electrodischarge machined into the welds. For the five other specimens, natural weld

indications were introduced during the welding process. Table !I presents the dimensions
of the semi-elliptical notches and of the indications in the welds. Table I I also describes

the nature of the natural weld indications, the condition of the backing rings, and the

status of the we13 reinforcement.

Pipe secpnent specimens. figure 5 shows the configuration for the pipe segment c

specimens. To fabricate these specimens, sections of 8-inch diameter pipe were saw cut

t o a length ot 12 inches. The saw-cut edges were beveled, and two sections were welded

together t o torm a 24-inch long pipe segment. End caps were welded t o the ends ot each

segment. The segments were normalized and one end cap was f i t t e d with a pipe nipple
for pressurizing the specimen.

4
Twelve specimens were tested. For four specimens, internal planar notches were

electrodischarge machined into the center girth welds. For four other specimens,
internal volumetric notches were electrodischarge machined into the center girth

welds. For t h e remaining four specimens, natural indications were introduced during the

welding process. Table 111 presents t h e dimensions of t h e notches and of t h e indications

in t h e welds. Table 111 also describes the condition ot t h e backing rings, and t h e status of
t h e weld reinforcement.

Service pipe specimens. Figure 6 shows t h e configurations of the service pipe

specimens. To fabricate these specimens, two sections of 3-inch diameter piping were
removed from service at NASA-Langley Research Center. End caps were welded to the
open ends of these sections. One end cap on both specimens was fitted with a pipe nipple

for pressurizing t h e specimen.

Specimen I , Figure 6u, contained a pipe tee, a pipe elbow, two lengths of pipe, three end

caps, a pipe nipple, three code-acceptable welds, and three code-unacceptable welds.
The code-unacceptable welds contained lack-ot-fusion, porosity, slag, and melt-through

indications. The dimensions of these indications were not available from reference 4.

Specimen 2, Figure 6b, contained a pipe tee, two elbows, two miter joints, two lengths ot

pipe, three end caps, a pipe nipple, three code-acceptable welds, and six code-
unacceptable girth welds. The code-unacceptable welds contained lack-of-penetration,

lack-of-fusion, porosity, slag, undercut, burn-through and melt-through. The dimensions

of these indications were not available from reference 4.

The backing rings were intact and the weld reinforcement was unground on these two

specimens.
General. All specimens were made of ASTM A106 Grade 6 Steel. The steel for t h e flat
plate, pipe-wall-segment and pipe segment specimens c a m e from t h e same mill heat.

Table IV presents the average chemical composition of this steel. This composition
/

satisfies t h e chemical requirements for A106 Grade B steel. Table V presents t h e

average tensile properties of this steel in both the "as-received" and normalized

conditions. These properties satisfy t h e tensile requirements for A106 Grade B steel for

both conditions. The steel for the flat plate, pipe-wall-segment and pipe segment
specimens was approximately 0.322-inch thick.

Available records for the service pipe Specimens indicate the steel was A106 Grade B.

However, no chemical composition or tensile property data could be located for this

steel. The steel for the service pipe specimens was approximately 0.2 16-inch thick.

Welding Procedures and Details of Indications

Welds were made using E7018 electrodes and backing rings. f o u r basic types of weld

indications were evaluated in this program: artificial planar, artificial volumetric,

natural planar, a d natural volumetric. The artificial indications were introduced into

code-acceptable welds using electrodischarge machining techniques.

Different approaches were used to introduce natural indications into the welds. Lack-of-

penetration (LOP) was produced by titting t h e pipe sections together tightly while
-
leaving the root face slightly larger than normal. Lack-of-fusion (LOF) was produced by

allowing t h e weld deposit to build up on one side of t h e weld joint. Painting a coat of
zinc-oxide mold-release compound on one side of t h e joint assisted this production.

~
Slag pockets were produced by not cleaning the welds between weld passes. Porosity was
iniroaucea by upsei i ing i h e weiaing paramerers auring rne gas-merai-arc iGMC\j weiaing

process.

Equipmeni

Flat plate and pipe-wall-segment tests. Load-control tests were conducted in 25 kip and

50 kip servo-hydraulic testing machines. The load cells on these machines were
calibrated using standards which were traceable to the National Bureau of Standards.
The alignment of these machines was veritied using a strain-gaged alignment specimen.

The pipe-wall-segment specimens were gripped by using conventional f l a t plate grips

with curved inserts. A reference test w a s conducted to determine whether bending

stresses were introduced by this gripping arrangement. The results of this test a r e

discussed in the "Reterence Test" portion of this report.

Pipe segment and service pipe tests. Pressure-control tests were conducted in the test

facility which Figure 7 describes. l h e pressure transducer in this facility was calibrated
using standards which were traceable t o the National Bureau of Standards.

The test specimens were mounted inside a containment vessel and were pressurized using
hydraulic oil. In the pipe segment tests, two systems were used to detect specimen

failure. The first system had a switch positioned over the centerline of the welds. If a

crack broke through the weld, the resulting j e t of oil activated the switch. The second
system included a float switch which was activated as the oil level in the containment

vessel rose. Activation ot either switch terminated the test. In the service pipe tests,
only t h e tloat switch system was used.

7
 Reference Test

A reference test was conducted t o measure bending stresses in the pipe-wall-segment

specimens. Analysis showed that bending stresses were highly dependent upon the ,

resistance of the grips t o rotation. For this reason, a strain-gaged specimen was tested

to quantify the bending stresses. Bending stresses were negative on the convex side ot

the specimen, and positive on the concave side. The ratio of bending stress t o axial

stress at midwidth on the concave side was I-to-IO or less. However at the midwidth on

the convex side, the ratio varied from 3-to-IO t o 240-5. The maximum ratio for any

point on the cross section was about 2-to-5. These data were used t o correct nominal

stresses in the tests.

Fatigue Tests

Constant-amplitude fatigue life tests were conducted on all pertinent specimens. For

the flat plate specimens, the maximum stresses ranged f r o r i ~13.0 ksi t o 33.0 ksi, and the

stress ratio, R, was 0.01. (The stress ratio is defined herein as the rotio of the minimum

stress to the muximum stress.) For the pipe-wall-segment specimens, the maximum

stresses ranged from 8.75 ksi to 33.0 ksi and the stress ratio was 0.01. For the pipe

segment specimens, the maximum longitudinal stresses ranged from 7.7 ksi to 16.0 ksi

and the stress ratio was 0. IO. For the service pipe specimens, the r n a x I ~ ~ ? !oncjI?udIna!
~'m

stress was 12.5 ksi and the stress ratio was 0.10.

For the f l a t plate and pipe-wall-segment specimens, the loading frequency ranged from

60 cpm to 1200 cpm. For the pipe segment and service pipe specimens, the loading

frequency was 60 cprn.

a
 RESULTS

Flat Plate Specimens.

Table VI summarizes the test results for the flat plate specimens. Consider f i r s t the

specimens having planar notches, Table VI a. For specimens having an initial notch depth

of 0.l-inch, the backing rings appear t o inhibit crack growth from the notch. In two

specimens (Nos. FPA-S-3 und 41, fatigue cracks initiated at the code-acceptable toes of

the welds and not at the notches. In a third specimen (No. FPA-S-21, the crack initiatea

at the notch, however a very large number of cycles (2.4 x IO6) was required t o

propagate this crack t o failure. Because of this inhibition, the backing rings on the

rernaining flat plate specimens and the pipe-wal I-segment specimens were cut.

Similarly, the weld reinforcement was ground t o reduce stress concentrations at the weld

toe. Figure 8 shows a plot of maximum stress against cycles t o failure for the flat plate

specimens having a 0.1-inch notch depth. An S-N curve is faired through these data in

Figure 8 t o show the general trend ot the data.

For specimens having an initial notch depth of 0.2-inch, all cracks initiated at the

notch. Figure 9 shows a plot of maximum stress against cycles t o failure for these

specimens, An S-N curve (solid line) i s faired through these data. Figure 9 includes the

S-N curve (dashed line) for specimens having Q 0. I-inch deep notch. Comparison ot these

curves indicates the following. At high stress levels, specimens containing 0.1 -inch

notches had fatigue lives approximately five time longer than the specimens containiny

0.2-inch notches. A t low stress levels, the fatigue lives of the specimens with both sizes

of notch were approximately the same.

Consider next Specimens having volumetric notches, Table VI b. In two specimens (Nos.

FPA-ES-I and FPA-ES-41, cracks initiated at the code-acceptable backing ring-pipe

interface, not at the notch. In the third specimen (No. FPA-ES-21, no crack initiated.

Since the cracks did not initiate at the notches in these specimens, it is not meaningful ,
t o compure their tatigue lives with the fatigue lives of the planar notch specimens.

Pipe-Wall-Segment Specimens.

Table VI1 summarizes the test results for the pipe-wall-segment specimens. Consider

f i r s t the planar notch specimens. Cracks initiated at the notch in tour of these

specimens. No crack initiated in the f i f t h specimen. (See Table VI1 a.). Figure IO shows
a plot of maximum stress against cycles t o failure for these specimens. An S-N curve is

faired through these data t o show the general trend of the data. Figure I 1 compares the

S-N curves for the flat plate and pipe-wall-segment specimens having 0.2-inch deep

notches. Essentially no difference exists between these curves indicating that the

flattening of the pipe had l i t t l e impact on fatigue behavior.

Consider next the specimens with natural indications. Cracks inititated at the

indications in four ot these specimens. No crack initiated in the t i f t h specimen. (See

Table VI1 b.) Figure 12 shows a plot of maximum stress against cycles to failure tor

these specimens. Figure I2 includes the S-N curve for pipe-wall-segment specimens

containing planar notches. The fatigue lives ot all specimens containing natural

indications exceeded the S-N curve for the specimens containing planar notches. Thus

the natural indications were not as deleterious t o fatigue life as the planar notches were.

IO
 Pipe Segment Specimens.

Table VI11 summarizes the test results for t h e pipe segment specimens. Consider first

t h e specimens having planar notches, Table Vlll a. Fatigue cracks initiated at the

notches in all of these specimens. Figure 13 shows a plot of maximum stress against

cycles to failure for these specimens. An S-N curve is faired through the data in Figure

13 to show the general trend of the data.

Consider next t h e specimens having volumetric notches, Table Vlll b. Fatigue cracks

inititated at the code-acceptable end cap welds in two specimens (Nos. 4305 and 43Q8).

Crack initiation appeared to be associated with t h e nearby pipe nipple welds. These pipe
nipple welds were not normalized during specimen fabrication. Such normalization is not

required by the codes for branch connections on A106 Grade B piping.

Fatigue cracks initiated at the notches in the two other specimens (Nos. 4306 and

4307). The pipe nipple welds were normalized on these two specimens prior to testing.

Figure 14 chows n plot of maximum stress against cycles to failure for Specimen Nos.
4306 and 43Q7. The S-N curve for the pipe segment specimens containing planar notches

is included on this figure. The fatigue lives of the volumetric notch specimens were four

to eight times longer than t h e fatigue lives of t h e planar notch specimens.

Consider finally the specimens having natural, planar indications, Table Vlll c. Three of

these specimens (Nos. 54PI, 54P2 and 54P3) failed in t h e code-acceptable end cap

welds. The pipe nipple welds were not normalized during the fabrication of these three

specimens. A fourth specimen (No. 54P4) also failed in t h e code-acceptable end cap

weld. However, the pipe nipple on this specimen was normalized during specimen
fabricat ion.
Table 111 shows that t h e indication areas for Specimen Nos. 54P1 through 54P4 were only

20% to 50% of the desired areas. The undersizing of the indications probably contributed

to shifting t h e failure location from t h e center weld to t h e end c a p welds.

Service Pipe Specimens.

Table IX summarizes the test results for the service pipe specimens. in spite of
numerous weld indications, both service pipe specimens failed due to fatigue cracking of
the base metal in t h e piping tee. Apparently, the stress concentrations in t h e tee were

more deleterious than the numerous, code-unacceptable weld indications were.

0lscussloN

The results of lhis study indicate that some code-unacceptable weld indications a r e not

as deleterious as various code-acceptable stress concentrations, e.g. weld toes, backing

ring-pipe interfaces, branch connections and piping tees. They also explain why

thousands ot code-unacceptable welds have survived long periods ot service without

failing. Finally, they confirm the conservatism of the current code acceptance criteria.

The tests on the flat plate specimens indicated a n important aspect of using backing
rings in making weld joints. Apparentlv, backing rings reduced the stresses acting on the
notches at t h e weld root. This stress reduction shitted t h e fatigue critical area to the
weld toe. (Similar results have been found by other investigators, reference 5.) This

shift implies that lack-of-penetration indications at backing rings are not particularly
fatigue-sensitive. Additional research is needed to verify this implication.
The tests on pipe-wall-segment specimens showed the efficiency of using

electrodischarge machined notches t o simulate natural weld indications. The size, shape
and locations of these notches can be more easily controllea than t h e size, shape and
locations of natural wela indications. Further, the fatigue lives of the specimens

containing notches were consistently shorter than t h e fatigue lives of specimens

contairling both planar and volumetric natural indications. These fatigue lives were
shorter even though the notches were considerably smaller than t h e natural indications.

Thus, specimens with notches cun be readily prepared, and will provide more
conservative estimates of fatigue behavior.

USAGE CURVES

The S-N curves for the pipe-wall-segment specimens, Figure IO, and for the pipe segment

specimens, Figure 13, were used to develop usage curves for code-unacceptable welds.

The procedures outlined in reference 6 were used to complete this development. These

procedures a r e described in the following paragruphs.

First, the S-N curves in Figure I O cnd 13 were adjusted t o equivalent S-N curves tor K =

- 1 loading. T h e tollowinc, relationship from reference 6 was used to make these

adjustments.

where

Seq is the alternoting stress which produces the same fatigue damage for an H = - 1

loading as the combined alternating and mean stress, Salt ana Smean , produce at any
other H value.

13
 ~ ~ _ _ _

Salt is t h e actual alternating stress

Smean is t h e actual mean stress

Su is t h e ultimate tensile strength of t h e material

Next the adjusted S-N curves were moditied to eliminate t h e effects of mean stress. The
following relationship from reference 6 was used to make this modification.

- su - sy tor S < Sy
'eq' - 'eq s
U
- seq eq -

and

for S (2Bj
e q s' Y

where

'eq( is the adjusted alternating stress tor which mean stress effects can be ignored.

Sy is the yield strength of t h e material

Seq and Su hove been previously defined.

Finally, a safety factor of two on stress or twenty on cycles, whichever was greater, was
applied to the adjusted and modified curves. The resulting usage curves are shown in
Figure 15 and 16. The reader should note that t h e procedures used herein to cievelop the
usage curves differ somewhat from the procedures used in reference 4.

14
The curves shown in Figures 15 and 6 were generated using a limited amount of data.

These curves provide a starting p i n for t h e rational development of usage curves for

existing code-unacceptable welds. Additional data should be generated to verify these

usage curves.

The usage curves were developed for girth welds in A106 Grade B steel piping. They may

be applicable to other weld configurations and to other materials. However, this

applicability should be verified before these curves are used.

Further, the usage curves were developed for specific pipe thicknesses. l h e curves
should be applicable to this thickness and to thicker pipe sections. However, t h e
applicability of these curvt s to thinner pipe sections should be experimentally verified.

The usage curve in Figure 5 was developed using the S-N curve for specimens containing
0.20-inch-deep planar notches. The lives of specimens containing volumetric and planar

natural indications exceeded this S-N curve. Thus the usage curve should be applicable

to both volumetric and planar indications, other than cracks.

The usage curve in Figure 16 should onlv be used for 8-inch diameter pi,m and larger.

The large flaw size for this curve (0.195-inch deep by 8.4-inches long) will significantly

reduce the remaining cross-section area of smaller pipe sizes. This reduction will muke

the resulting net section stresses much higher.

 coNcLusloNs

A series of fatigue tests were conducted on specimens made of A I06 Grade €3steel. Four

, specimen configurations were tested; flat plate, pipe-wall-segment, pipe segment and

service pipe. Based on t h e test results, the following conclusions were reached:

1. Usage curves were developed for welds having indications both 0.20-inch deep
by 0.40-inch long and 0.195-inch deep by 8.4-inches long. These curves were

developed using t h e procedures used in formulating t h e design curves in Section

VIII, Division 2 of the American Society of Mechanical Engineers Boiler and

Pressure Vessel Code.

2. Some code-unacceptable weld indications a r e not as deleterious as various

code-acceptable stress concentrations; e.g. weld toes, backing ring-pipe

interfaces, branch connections and piping tees.

3. Backing rings appear to reduce the stresses acting across indications at weld

roots. This reduction can shift the fatigue critical a r e a trom t h e weld

indication t o the weld toe.

4. The tatigue lives of specimens containing electrodischarge machined notches

were consistently shorter than the fatigue lives of specimens containing both
planar and volumetric nutural inaications, excluding crocks. These fatigue lives

were shorter even though the notches were considerably smaller than the
natural indications. The impact of weld toes, backing rings and other c

discontinuities must, of course, be considered when evaluating fatigue

resistance.

16
 REFERENCES

1. Rules for Construction of Pressure Vessels. Division I , Section VIII, ANSIIASNiE

Boiler and Pressure Vessel Code, July I , 1983.

2. Moore, C. V.: An Introduction to Section VIII, Division I of the ASME Boiler and
Pressure Vessel Code, J. Pressure Vessel Technology, August 1976.
3. Hudson, C. NL and Ramsey, J. W., Jr.: Pressure System Recertification at NASA-
Langley Research Center, NASA TM-85733, December 1983.

4. Leis, 6. N.; Goetz, 0.P. and Scott, P. M.: The lntluence of Defects on t h e Fatigue
Resistance of Butt and Girth Welds in A1066 Steel, NASA CR-178114, October 1985.

5. Yung, Y. Y. and Lawrence, F. V.: Analytical and Graphical Aids for the Fatigue
Design of Weldments. Proceedings of t h e Fitness-for-Purpose in Welded
Construction Conference, May 14-1 6, 1985.

6. Anon: Criteria of t n e ASME Boiler and Pressure Vessel Code for Oesign by Analysis
in Sections 111 and VIII, Division 2, The American Society ot Mechanical Engineers,
1969.

17
 cc Uc
Q
’73
> > >
222
(3 ‘3c1

t t t
6
m
t t t
Ir. f $55
8
c i-

??c???u???c??u?? c?c??
000000000000 d 000

779
300

18
F

c
9
d)

19
 -??-??
> > > >
0 0 0 0
L L L L
mmom
c c c c
3333

Q)
f

0
c
t t t t t t t t
> > > > > > > >
U U W U QUUU
2
t x
b
c
f
3
z
E <<".<
n
f
E
B
6
5<5<
cocococo cocococo
<<5<
cocococo
3

-P

****
--Nm4.
44 44
Lorn Lnv)
E

21
 - c?
c;
LA
0
LA

I-
 + c
0 0
Q) Q)
0
t
0
t

r c z z .-C
t
u 0 u u .-
3 e
t t

t t
3
t
0
C
t
L

9
0 0 0 0 Q)
C W -0
Q) Q)
t t

.-c .-C .-C .-.-C

.-
0 0
.-C . t t t
.- t .- .- .- .-
3C
2

b
Lc

c
0
IA d .t
5
9
o
s m
o
9
N h l -

a b
- b
I I o ; ’ ? -
? ? ? ? +
a a a a a
a a a a a
L L L C L L L L L

23
 .-t
Y
0
n0
-8

9
9
a

t
0
iz
2
t

b
rc
(I)
t

3
U

9
t -
m
9 c3

.
.I
.I

s 9
m
9
0
L?
b
m N N

24
-.
c
7a -c
u
x
u L
0
.-0
t
0
0 .-u
t
0
C
0
t
C
.-0
0
.--a0
5
n
t
5 5 e
t t
t

t t
.-
.- 2
C
t
.-c
t

I
0
0
U
Q)
t
.-.-
0
t
0
U
Q)
c
.-
0
3
t
0
.-C .-E
t
0 0
-z
t
.- .-
0
.-
t
.-
0
U
Q)
t
0
c
.- .-
0
.-
U
t

t
0
Q)

.- .- .-
c
t .-C .-C .-C .-c t
3 2
Y
0
0
Y
u
E
Y
0
Y
0
e
3
Y
u
2
Y
0
E
Y
0
0
Y
0 0
s
u 3 u v U

2
c
b
cc
v)
c

3
a
-r: 5 - ?
co
c!
Q\ v)
0
0
4
m
N
N
d N cv m I-

f
Y
a a
Y
4
+
L n
T
Q
W
Q
W 4

25
 u u u u U
0 a , at , at, a t, a,
I t

.-.-0 .-.-0 .-
.-0 .-.-0
t
d .-
0
II
t t
.-C .-C .-C .-C
t t
.-
t

U .-C
Y Y
0 0
0
u"

I
B L
t
0
I-.

ii
2
cs
co
a
's
N
N
cv
s b
fi
0
cv
9
fi r"
t f
m
>o
b
* s
t
v)
5 ir'
3
a
5
c
i
6
s
Q\
p:
fi
d
c
2

Y
2

26
 n n n U

0
d
II
a=

gl
h
f
t
1
h
b
r
cn
t

1
U

t
t

3 i

1
a
c

27
 -0
t
-
U
t
Q
E E
d
A
Q
13
.-c
-a
-d
0
Q,
t
.-0
.-
t

II .-E
U Y Y
u 0
0 0
6

8 a,
8 0)
t t

f
t

Lc
b

3
-.I

t N
0
IA

28
 z
0

t
t

29
 m
W
n I
0
8 - E
z
H

U
n

U
H

2
I
0
I-
O
z
LL
0 0
z v,
L -J
Iv H
S <
I
G I-
W
a n
-
Iv
(u

H
LL
 z

-
E
I-
z
H
LL
I . ,

'r I
I-
n
w
a
I
V
I-
O
z

31
 v,
W
I
V
z
w
z
U

v,
Z
0
w

no
H &
nLL

32
 z
2
w
0
W
a
m
I-
z
W

5
w
m
w
a
H
a

33
 7506-6

a. Specimen No. 1

7 506- 3

b. Specimen No. 2

FIGURE 6. SERVICE P I P E SPECIMEN CONFIGURATION. FROM REFERENCE 4.

34
 W
a
U
a
n
z
<
-I
W
m
m
W
>
w
I
I-
LL

I
I

1
Q !
fg
I
I

I
0
V
l--l

35
 0
z
U
I
l+

f
0
v,
~

I
k
w
n
I
u
I-
O
z
J
z
I-
U
z
H

v,
z
w
E
U
V
w
a
v,
W
I-

0
5
a
0
c I-
c
u
c
0 a 4
LL
c
c c of
0 0
0
U U LL
0 0
c c
.- .-
0 0
I-
.- .-
c

-
c .E
c

c
a
s
x w
0 u 0
c LL
E f
0
a
K
U
-1.
0 d
w o
I I I I
3 -
-0
a 0
t U
I- II
2-
0 0 In - a
l
In
0

H
LL

36
 I
V
;L
W
I
cu
0
v)
W

I
l-
a
w
n

I
Y
L It
z,
I
0
I
I
0
n
1
I
0
1
w
n m
0

H
LL

37
 4
0
0

c
0 v,
w
I
0
I-
O
z
m
z
H
z
U
Q
I-
z
0
0
v,
z
W
D E:
0 H

-.-f
9
0
c
m
-
0
0
h
0

D
0

0
0 0
n 4
 W
w
n
cn
L

0 0
0 Y)

LL

39
 v,
z
0
U

a
z
H

H
0
W
a
v,
I-
z
w
5
W
v,
I
-I

I
W
a
U
n

==c
I-
s
w
LL
U
-I.
7-i
W O
3 .
a0
U

0 0
0 Y)
 m
W
I
0
I-
O
z

i
0
W
[L
v,
c

7 W
v,

7I
W
a
H
a
nc
0
LL

P
0 W
LL
H
J .
0
W 4

0 w
Y) nc
3
W

41
 cn
z
W
r:
H
V
W
n
Ln

4
w .
no
H
a II

42
 W
I
d-
0
LL
0

0
LL
0
I
-I
n
W
n
<

H
>

H
!-
25
H
n
z
W

Y
0
LL
W
>
CY
3
0
 0

a
z
Q
I
V
z
H
I
n In
-
0

0
L
0

0 v,
z
0

wco
>
C C L
30
0 -
I
n
-
0
Wl-
aa
d Z
CAW
3-l

..

44
 Standard Bibliographic Page
. Report No. 12. Government Accession No. 13. Recipient's Catalog No.
~ n r nTU nnnr-
tinan II~I-UJUJL I I
I I
. Title and Subtitle Acceptance C r i t e r i a t o r bJeIds 1n ASTM 15. Report Date
106 Grade B S t e e l ' P i p e and P l a t e October 1986
6. Performing Organization Code
______ 323-53-71-01
c.
~

. Author(s)
i c h a e l Hudson, Davis B. Wright, J r . and 8. Performing Organization Report No.
B r i a n N. L e i s
. Performing Organization Name and Address
10. Work Unit No. .
IASA Langley Research Center
11. Contract or Grant No.
lampton, VA 23665
____ 13. Type of Report and Period Covered
2. Sponsoring Agency Name and Address
Technical Memorandum
l a t i o n a l Aeronautics and Space A d m i n i s t r a t i o n
14. Sponsoring Agency Code
las h i ng ton, DC 20546
-
5. Supplementary Notes
' h i s r e p o r t i s an abridgement o f NASA CR-178114.
:. Michael Hudson and Davis B. Wright, Jr.: Langley Research Center, Hampton, VA
lrian N. L e i s : B a t t e l 'le Col umbus L a b o r a t o r i e s , Columbus, OH

6 . Abstract

Based on the RECEKT Program tindings, NASA-Langley funded a fatigue study of code-
unacceptable welds. This study was conducted to developed usage curves which were
based on the structural integrity of the welds. The details of this study are presented in
NASA CH-1781 14. The information presented herein is a condensation and
reinterpretution of the information in NASA CR-1781 14. This Condensation and
reinterpretation generated usage curves for welds having (a) indications 0.20-inch deep
by 0.40-inch long and (b) indications 0.195-inch deep by 8.4-inches long. These curves
were developed using the procedures used in formulating the design curves in Section
V111, Division 2 of the American Society of Mechanical Engineers Boiler and Pressure
Vessel Code.

-_____.
17. Key Words (Suggested by Authors(s))

Jelds, f a t i g u e , acceptance c r i t e r i a , Uncl a s s i f ied - Unl i m i t e d

Dorosity, slag, lack-of-fusion, lack- S u b j e c t Category 26
Df-penetration, A106 Grade B S t e e l

___--
19. Security Cl,lssif.(of this report)
UNCLASSIFIED 45 A03 --
_-.__ __

For sale by the National Technical Information Service, Springfield, Virginia 22161
N A S A L a n g l e y Form 63 ( J u n e 1985)