...

HY80SteelFabrication-SubmarineConstruction.pdf

by user

on
Category: Documents
12

views

Report

Comments

Transcript

HY80SteelFabrication-SubmarineConstruction.pdf
I.,
UNCLASSIFIED
AD
258 306
ARIMED SERVICES TECHNICAL INFORMTIION AGENCY
ARLINGTON HALL STATION
ARLINGTON 12, VIRGINIA
0
II
UNCLASSIFIED
NOTICE: When goverinent or other dravings, specifications or other data are us.td for any purpose
other than in connectioA with a definitely related
goverrment procurement operatioL, the U. S.
Govermzent thereby incurs no responsibllity, nor any
oblipation whatsoever; and the fact that the Governsent may have formilated, furnished, or in any way
supplied the said dravings, specifications, or other
data is not to be regarded by implication or otherwise as in any mnner licensing the holder or any
other person or corporation, or conveying any rights
or permission to manufacture, use or sell any
patented invention that my in any way be related
thereto.
__7J
n•
qt
lh,••
•',..•
I.
_
I.ORWARDED BY THE CHIEF, BUREAU OF SulM"
r,
,
cz
BUREAU OF SHIPS
CONFERENCE
e
HY-80 STEEL FABRICATION IN
SUBMARINE CONSTRUCTION
cm
AT BUREAU OF SHIPS
21-22 MARCH 1960
LAJ
cn
THIS [)O(CUdENT MAY LM- RI,'LEASIED WITH N%
Ri.STRlGTONS ON DISSEMINATION
AST 1IA
l JUN• 2%0' ,s6'
de
Sl IMIKl
IIPo
CONFERENCE AGENDA
Investigation Of Weldability Of Navy Grade HY-80 Steel
G. Emmanual, Babcock & Wilcox Co.
.
...............
.
Flux Developments For Welding HY-80 ................
D. Martin, Battelle Memorial Institute
I
..........
2 8
....................
Development Of Crack Free Electrodes For Welding HY-80 Steel ........
J. Cahill, New York Naval Shipyard
Soime Observations On The Weldability Of Quenched And Tempered High Yield Strength Alloy Steels
W. D. Doty and G. E. Grotke, U . S. Steel Corporation
34
Fatigue Properties Of Welds In HY-80 Steel ..............................
I. L. Stern and H. V. Cordiano, Material Laboratory, New York Naval Shipyard
68
Report Of Semi-Automatic Inert Gas Metal Arc Welding HY-80 Steel Out-Of- Position .....
L. Robbins, Mare Island Naval Shipyard
The Application Of Preheat To Submarine Construction
F. Daly, Newport News Shipbuilding & Drydock Co.
10
........
133
......................
Report On The Development Of A Crack Resistant Electrode For HY-80 ......
S. IV Rnherts, Portsmouth Naval Shipyard
Explosion Test Performance Of Small Scale Submarine Hull Weldments ....
A. J. Babecki and P. P. Puzak, Naval Research Laboratory
...............
...............
Explosion Bulge Test Performance Of Experimental Stubmerged-Arc Weldmnents Of HY-80 ...
A. J. Babecki and P. P. Puzak, Naval Research Laboratory
Notch Toughness' Evaluations Of Modified HY-80 Steel In Heavy Gage Plates ......
A. J. Babecki and P. P. Puzak, Naval Research Laboratory
Effect Of Welding Variables On The .Yield Strength Of 9018 Weld Metal
W. L. Wilcox, Arcos Corp.
Quality Control In The Fabrication Of HY-80 Structures .
T. Dawson, Ingalls Mdpbuziding Corp.
Workmanship Controls In The Fabrication Of HY-80 Hulls
E. Franks, Electric Boat Co.
iv
........
....
148
.....
170
.............
184
...............
196
.
. ................
. ...
139
..............
204
215
Norfolk Naval Shipyard
W.F. Booker, Jr.
E. Elsarelli
Edward E. Reedy
A. 0. Smith Co.
D. Buerkel
J. Koss
R.C. Steveling
SUPSHIP CAMDEN
Philadel ph .i N.ia: I Shipyard
LCDR R.F. Roche
J. D ~lm
J.R. Girii
T.O. Maginnis
LT D.A. Marangiello
P.P. Santlanau
E.A. Edelmann
SUPSHIP GROTON
S.J. Secora
SUPSHIP NPTNW8
LCDR R.J. DIUikowmk
United States Steel
W.J. Baumgarten
W.B. Doty
G.E. Grotke
H.V. Joyce
Ports-mouth Naval Shipyard
C.E. Cole
F.W. Duniham
W. X,. Ha rdy
CDR S.H. Heller. Jr.
S.A. Roberts
T.L. Sheehan
ME. Watts
Reid Aver!, Cunipaii
W.A. Rohdt
Harnold E. S
S.L. Magee
D.J. Synder
J. Trutz
Engineering Experiment Station
H. Seigel
W. L. Williams
k.
,it
Here is the corrected Table of Contents for the Bureau of Ships
Conference, HY-80 Steel Fabrication in Submarine Construction,
at Bureau of Ships, 21-22 March 1960.
INTRODUCTION
A seminar, sponsored by the Bureau of Ships, on the Weldability of HY-80
Steel as applied to submarine construction was held at the Bureau of Ships,
Washington. D.C. on 21 - 22 March 1960. Technical pakpers presented and
a list of atterdees are included herein.
The boxok is intended to be a matter of interest to production personnel engaged in fabrication of submarine hulls from HY-80 steel as well as to engineers and architfects. It presents potential as well as new developments
in weld fahrnaiiozi materials and methods.
R.E. BALL
Captain, U.S.N.
Director, Hull Division
Bureau of Ships
CONFERENCE AGENDA
Page
1
Investiption Of Weldability Of Navy Grade HY-80 Steel ........................
G. Emmanual, Babcock & Wilcox Co.
Flux Developments For Welding HY-80 ...............
D. Martin, Battelle Memorial Institute
23
.................................
35
...................
Development Of Crack Free Electrodes For Welding HY-80 Steel ........
J. Cahill, New York Naval Shipyard
Some Observations On The Weldability Of Quenched And Tempered High Yield Strength Alloy Steels
W. D. Doty and G. E. Grotkc, U. S. Steel Corporation
41
Fatigue Properties Of Welds In HY-60 Steel ..........................................
I L. Stern and H. V. Cordiano, Material Laboratory, New York Naval Shipyard
86
Report Of Semi-Automatic Inert Gas Metal Arc Welding HY-80 Steel Out-Of-Position .....
L. Robbins, Mare Island Naval Shipyard
The Application Of Preheat To Submarine Construction .......
F. Daly, Newpcrt News Shipbuilding & Drydock Co.
........
................
"
Report On The Development Of A Crack Resistant Electrode Fcr HY-80
........
129
....
152
159
.........
S. I. Roberts, Portsmouth Navwd Shipyard
. . .
Explosion Test Performance Of Small Scale Submarine.flull Weldmctl.s
, ,
A. J. Babecki and P. P. Puzak, Naval Research Labratory
Explosion Bulge Test Performance Of Experimental Submerged-Arc Weldnints Of HY-80,.
"
.9
A. J. Babecki and P. 'P. .Puzak, Naval Research Laboratory
Notch Toughness Evaluations Of Modified HY-80 Steel In Heavy Gage Plates
A. J. Babecki and P. P. Puzak, Naval Research Laboratory
Workmanship Controls In The Fabrication Of HY-80 Hulls
E. Franks, Electric Boat Co
iv
.
.
191213
.
."
Effect Of Welding Vixiables On The Yield Strength Of 9018 Weld Metl
I'"
W. L. Wilcox, Arcos Corp.
Quality Control in The Fabrica, ion Of HY-80 Structures ...........
T. Dawson, Ingalls Shipbuilding Corp.
i"......
•
.
169
.
.
........
•.........
.
...
229
.o
........................
..
...........
237
250
'
INVESTIGATION OF WELDABILUTY
OF NAVY GRADE
HY-80 STEEL
by
G. Emmanual, Babcock & Wilcox Co.
INTRODUCTION:
Under BuShips Contract NObs-77064, the Babcock
& Wilcox Company has been engaged in an investigation primarily aimed at improving the weldability of
HY-80 steel under the conditions required for the
construction of submarine hulls,
This discussion is in the nature of a progress report of our findings to date. Briefly, the problem
has been approached from the following angles:
1. Behavior of the heat-affected zone with respect to its transformation characteristics
and mechanical properties.
2. Development of weldability test which would
consistently produce cracKing of commercial
plate.
3. Study of welding variables on susceptibility
of cracking.
4.
Effect of compositional variations on tendency
for cracking.
5. Misc.ellaneous Tests:
(a) Thermal expansion characteristics of
11018 weld deposit and IIY-80 plate.
(b) Submerged arc welding tests on a
limited basis.
TEST PROCEDURES:
11gat Affected Zone - One of the few aspects of the
HY-80 problem on which there is almost universal
agreement is that cracking originates and propagates
'rom the heat-affected zone. It was, therefore, decided to study this segment first.
Our original st-idles were made on a commarcial
heat of steel (Table I) using a high temper;i..jre
vacuum microscope (Figure 1) for direct observation
of the Ms and Mf temperatures. A peak austenitizing
temperature of 2230 F was attained. After holding
at this temperature for 3 minutes, the sample was
allowed to cool and observed. At a temperature of
890 F the first needles of martensite appeared.
With decreasing temperature, more and more
martensite appeared until at 830 F, the reaction was
virtually complete. The sequence of transformation
is shown in Figure 2. The cooling time from 2230 F
to 890 F was 14 minutes, and time in the martensite
transformation zone was 9 minutes.
Published information on the transformation
characteristics from an austenitizing temperature
of 1650 F indicated an Ms temperature of 680 F, and
no information relative to the Mf temperature.
Since the heat-affected zone reaches temperature
of the order of 2400 F, it was decided to determine
experimentally the transformation characteristics
of HY-80 material from that temperature. This was
done by resistance heating 4" x 1/4" diameter bar
to 2400 F in 16 seconds, followed by immediately
quenching into lead baths at various temperatures,
and holding for different periods of time and then
water quenching. By this procedure, it was possible to determine the beginning of isothermal transformation of HY-80 at various temperatures.
In order to determine the Ms temperature, a
somewhat different procedure was followed. In this
case, samples were quenched into lead baths at
various temperatures held for short times (less than
1 minute), transferred to another lead bath at 12501275 F for 1-2 minutes and water quenched. If the
temperature to which the sample had been quenched
was below the MS point, martensite will form. Upon
reheating to 1250-1275 F, the martensite is tempered
and subsequent metallographic examination differentiaies between tempered and untempered martensite.
Thus, the T-T-T diagram (Figure 3) was developed.
Examination of this diagram readily shows the
extreme sluggishness with which HY-80 will transform as well as a surprisingly high martensite
transformation temperature when the peak austenitizing temperature is 2400 F.
From these observations, the following conclusions appear to be warranted:
1. The formation of bainite in the portion of
the heat-affected zone is not likely.
2. Tb high M. and Mf temperatures do not favor
the :., mation of cold cracks since it is generally accepted that as the temperature for
martensite transformation is increased, the
susceptibility for cold cracking is decreased.
3. The present limit of 300 F on interpass temperature appears to be overly stringent for
high chemistry HY-80. This should be checked
by explosion bulge tests with higher interpass
temperatures.
Hot Ductility Tests - A test developed at the
Rensselaer Polytechnic Institute has had some
success In predicting the weldabililty characteristics
of heavy stainless pipe. Briefly, this tests consists
of heating a small specimen at controlled rates
equivalent to those e'icountered during welding. The
procedure involves heating series of samples to increasing temperatures, and breaking the samples
and measuring reduction of area. As would be expected, reduction of area values increase with temnperature until at some temperature (depending-on
the material) the ductility drops to essentially zero.
Once this te~mperature is determined, another se~ries
of samples is heated to that temperature, cold t various temperatures broken, and reduction 6f are'a-7,
determined. In the case of stainless stee'l, it was
noted that when the reduction of area on cooling was
markedly lower than that determined during heating,
the weldability was poor. When the reduction of
area curve on cooling was essentially similar to the
curve on heating, no trouble was experienced during
.
welding.
Under the welding condition employed, (30, 000
joules/inch 1/8", 11018 electrode, room temperature preheat, and 2-inch thick commercial HY-80
plate) it was found possible to consistently produce
cracks in the heat -aff ected zone.
Admittedly, these welding conditions bear little
resemblance to those employed in the shipyards,
however, it was realized that producing restraint in
the Laboratory, comparable to that experienced in
actual submarine constructi~on, would be difficult.
In an attempt to compensate for this lack, the weld'ing conditions were made unrealistically severe.
Welding' Variables - Holding all other conditions
constazql, p~r~eat_,Iid;ogen content, electrode and
Ji1'at. input were varied. rk1'as found that varying
preheat Lietween 32 and 10,50 F, hydrogen between
.l.5.and*i ppmn (F'igure 6), heat input beLwgen 30,000
and 70,000 joules. in". (Fij~re 7)'and-using lower
y~ield st.rength elcltrodes'(l68-2 and. -9018) did*not
r.educe the tL~ndcrncy.for.c?'rackingt(Figure 8).. As a
ha*j'iiait'ýJof atait was~pos~ible to produce cracking
of- heat byiptdadling the surb
A curve of this type (Figure 4) was dete;rmined e,p ý,try~eJWr$aiptlWaiqn
for the analysis shown in Table I. It will be itotc% facefwith i'terc k4!omic~hydrogen
orF
toiches (Fig-.
that there is a substantial differencelaItween the'
. 'nure 9)
1i4~*.
heating and cooling curve. We pln0a
ea
irberS
i ýn?
ape ied that wIýditig vazriabl~s were'not
of these tests made on experimental! maml. aa,
see if a correlation aatween the resud~of t4,4p1
~ e primptiu'*
i nttsgcakn,.i
a e
u ided to7
sljkvesiat~e tjwecihof cornsiioa vanic.
test and weldability exists.
a~ins~staa11w.ipeifiediavte for HY280.
A series of convential high teaiperdflV testewere
Cý*.o'.
POSITIONA LNVARIXXION!
run following the same procedure'. Th VruýOM
t&
to
11fi~p
shown in Table [I. It is interestg
A ntinilx 4 of experam~ciiar induct io'Nheats were
the peak temperature available *(2400 F) tLicre t1s no
ctedboth in a~r anid in vicuum. The nominal
apparent damage to the material as evifn~ced bm-ee
ayis~'hs
heatseal'e 'shown in Tab~le 111. It
duction of area measurements. This P41e1r~ymt*e
be raote~ith~atthese 'hea- s.were planned to hold
w~ieIwwe&l
tends to indicate that damage takes.0
..
all efetnents I~t dne'tn the approximate center of
2400 and 2510 F.
.. <
leit'specified range; Each'eiement was then varied
&.,a~
'wlitja its speelf"e limits. -The only exceptions to
WeldabiiMy_ Tests -It was felt 16at one of "r
this rule occurs'in the phosphorous and sulfur
primary problems was to find or develop a simple
vafttb.
weldability test, which would consistently prodiae
cracks under a given set of welding conditista Th
"o
The ingots- thus produced were then either forged
nature of the welding condition was felt to be rela.or vu~frd& heat treated, welded as CTS tests, and
tively immaterial, since we would vary iinhy tifttcr
factor at a time, and all results could .be judged on .
examined. The results are shown in Figures 10
through 12. It will be noted that the only sample
the same basis.
which did~not show any cracKing &Call was the one
with the lowest sulfur ccntents.
After examining a number of possible tests, it
was finally decided to use the Controlled Thermal
MISCELLANEOUS TESTS:
Severity (CTS) test developed by the British Welding
Research Association. Figure 5 shows the configExpansion rates of HY-80 and 11018 weld deposits
uration of this test. Essentially, this consists of
were determined. As shown in Table IV, maximum
fillet welding one (1) 3" x 3" block to a 4" x 7" block
difference of approximately 24% in coefficient exof HY-80 as indicated. The two welds parallel to
pansion exists between these two materials at 1400 F.
the long axis oll the bottom block are two (2) pass
welds, and are known as tile "anchor" or "restrainAs part of our efforts to investigate automatic
ing" welds. The two transverse welds are single
methods for welding HY-80 material, some experipass, and are the test welds, For a fuller descripmental wire was produced, and a submerged arc
tion of this test, the reader is referred to the Deweld produced in 1" thick lean chemistry plate. The
cember 1959 issue of "Welding Research Abroad."
results are shown in Figure 13.
2
zone of HY-80 material is of high temperature origin,
and can be initiated by the mere application of heat.
Reduction of the sulfur content to rather low levels
appears to be effective in reducing the tendency to
crack. Since the cracking observed is microscopic
in nature, it must be shown that it is the:e microscopic cracks which propagate into detectable cracks
of macroscopic size. It must, also, be shown
whether propogation occurs at relatively low temperatures, and is the result of transformation stresses
and/or the action of hydrogen, or whether the cracks
both originate and propagate at temperatures close
to fusion temperatures.
EQUIPMENT:
As a mttter of general interest, Figure 14 illustrates some of the equipment that is being used in
this program. Having the facilities to produce experimental heats, process therr into plate, rod or
wire has greatly expedited the progress of tziif
program.
SUMMARY AND CONCLUSIONS:
From the results of the foregoing tests, it is our
belief that the cracking observed in the heat-affected
G. N. Emmanuel
THE BABCOCK & WILCOX COMPANY
GNE:ds
April 14, 1960
Talk given
at the Bureau of Ships
HY-80 Conference March 21, 1960.
3
TABLE I
ANALYSIS AND) PROPERTIES OF
1.988' HY-80 MATERIAL
LUKENS STEEL HEAT No. 21228
C
Test Repe~rt
-
B&W Analy~sis-
*Mn
P
S
.024
.16
.34
.014
.18
.32
.011
K
Heat Treatment:
Ni
Cr
Mo
25
2..87
1.52
.41
.25
2.90.
1.50
.40
..
.027
1650 F
Si
-
We.Q.
~
T~~rnpered&
~
..
110ýWF
Pr.
*.T.S
psi.
x 1006
Y.S. psi.
.x1000
ELONG.FTLB
1; 2"
86.8-87.2
26..2.
hA.,,
~
910
-Test
Report
103-103.1
B&W
104.2-104..5
B&W tests run at
94-98.1
,6.9-28.5
100OF..
.
4
2~7.
66,473.3
98-102,106#-
.
TABLE II
HIGH TERMPERATURE MECHANICAL
PROPERTIES OF HY-80
TEST TEMP.
F
TENSILE
STRENGTH
PSI
YIELD STRENGTH
PSI
.2% OFFSET
E LONG. %
IN 2"
1600
9800
6400
25.0
24.6
1700
7700
5500
27.5
25.6
1800
6000
4200
29.5
30.4
1900
4700
2900
38.5
35.2
2000
4000
2600
74.5
69.3
2100
Z200
1800
103.5
Infinite
"2200
2500
120.5
Infinite
2300
1900
1100
113.5
Infinite
2400
1400
600
113.5
Infinite
2300
1900
1100
129.5
Infinite
2200
2400
1200
130. 0
Infinite
2100
3100
1700
110.5
Infinite
2000
4000
2500
123.5
99.4
1Q00
4900
2700
107.5
96.5
1800
6300
2900
75.5
64.7
1700
7900
5000
51.0
44.3
1600
10100
63.0
55.0
R.A. - %
Heated to 2400 F
and tested at:
5
TABLE m
EXPERIMENTAL COMPOSITIONS
(AIR & VACUUM MELTED)
C
M11
Si
Median Analysis
.16
.30
.25 o
C Variation
.08
.23
.30
.30
Mn Variation
.16
.16
.10
.70
Ni Variation
.16
.16
.30
.30
16
.30
.25V
.16
.16
.30
.30
725'..
.16
.30
.16
.16
.30
.30
':
.16
.30
.- *.•25
.16
.16
16
.30.
.30
.30
.25
.25
..25
.16
.30
*2i
Cr Variation*
Mo Variation*
P &S Variation
.
Ni
Cr
Mo
P
S
2.80
1.60
.45
.020
.025
•" .25
.25
.2.80
2.80
1 60.
1.60
.45
.45
.020
.020
.- .25
2.80'
. :.2.80,,.
1.60
1.60
.45
.45
.020
.020
.025
.025
•020
.020
.025
.025
.020
.025
.
.
25
u . ,•
'a'2'40'..'il.60
."3•,210
O.
42..
.
.:r
.•.
2. 80"
2.80"
""2'480
.
160g.
1.60...
,l60',•,
" 2.80. •
2.80
w
.>2.806
d:
,"2; 80
.
These analyses have not yet been tested.
6
1,
.
-,1;-,If,60Ot-Wr-•.,I45
.25
- "2, 80 '.
..... 90.-.w
• €.-25..'280.-,'%.V'.1-90%.
.".25
-. 25
.45.. 5.
"..45
It'
.2
1.60%';
.45-'6
1.60 w, .45
1.60
• e.45
."
.025
. 4 5d,.
.020,
.'45......020
., ..02o..
.60
60
.025
.45
M
.
.
.025
.025
.025
.020
020
.025
.02ý
.020W.
.025
0
,035
0
.035
*
0
0
.040
.040
TABLE IV
COMPARISON OF THERMAL EXPANSION CHARACTERISTICS
OF HY-80 AND 11018 WELD METAL
RT200F
RT400F
RT600F
RT80OF
RT1000F
RT 1200F
RT1400F
RT1600F
RT1800F
V Plate
6.3-6.4
6.7-6.7
7.0-7.0
7.2-7.2
7.3-7.6
7.4-7.7
5.5-6.1
5.6-5.8
6.2-6.4
6" Plate
6.3-6.4
6.6-6.6
6.8-6.9
7.1-7.3
7.1-7.5
7.1-7.1
5.8-5.9
5.7-5.7
5,9-6.0
11018
Weld
Deposit
6.6-6.8
7.3-7.5
8.1-8.1
8.2-8.3
8.2-8.2
7.9-8.0
7.2-7.2
6.1-6.2
6.4-6.4
CHEMICAL ANALYSES
C-
Mn
Si
Cr:
Ni
M0
1" Plate
.16
.25
.17
1.1§
2.29
6" Plate
.25
.30
.17
1.58
11018 Deposit
.03
1-.67
.3
.12
7
*Cu
P
5
.25
.012
.017
2.98
.50
-
2.38
.44
.05
.013
-
.024
N
-
Figure 1 - High Temperature Vacuum Microscope Used For Direct
Observation of Martensite Transformation in HY-80 Steel.
*
/
(a)
(d) Time - 18 Minutes
Austenite at 2230 F, Time - 0
8
0
40.
W.0o-I
- -
-
"",-.-
'
,44
%.-%
(b) Ms
-890
F, Time
-14
Minutes
(e) Time
-
-..
20 Minutes
.- 0
(c)
Time
Figure 2
-
-
(f)
16 Minutes
Mf
-
830 F, Time 23 Minutes
Martensite Transformation Occurring Over Range Of Temperatures. Original Magnification
205X, Enlarged 2.5 Times.
9
• !
1 0 ....----------
-
•+ ; a,
4,
i,t
-
1-.
...
I
+1h1T
-
VtS0
-
* A'A
c. *
"I
.'
..
?,.
...
4
-
-- •
.
"
..
+o.
....
.- ÷
..
--
÷
..
sl
-
0
V1. U41
30 -
?ufp.- 1',40 7.
t (:,-.
s
RoULI4O
hm s'•"1dOc.
(ri.
l metm~t/dzg4
.6
*
?'i~.- 2%O Pe.
o'
.~ ._2%
1.
:.-1..1+•
"IDI
"
[ 2
¶•
!9t
I
It5O¶'VRNAI rINADAJ~I
..
I*"
I.. '9
,
3•MM
?IU Wl hIYJ
Figure 3 -- TTT Curve Develo~ped From Austenitizing Temperature 01 2400 F. Note Increase In Transformation Time And Increase In Us Temperature As Compared With Curve Developed From 1650 F.
X10
100
-
t
~oo
*W
TFSTINW
2200
29W
_•
20
TEK4_RATUrE. "F
Figure 4 - Results of R.P.I. Hot Ductility Test
Conducted ON HY-80 Material. Note Substantial
Difference In Reduction Of Area Values On Cooling
As Compared With Values Obtained During Heating
Cycle. It Is Believed That This Indicates Damage
To Material As A Result Of Heating To 2510 F.
Reduced To Approximately 1/3 Size
Figure 5
•
-.
Photograph Of CTS Test.
--
..
".
3 ppm H2
1. 5 ppm H2
Magnification
Figure 6 -
-
-
50OX
Effect Of Hydrogen Content Between 1. 5 And 3 ppm On Cracking Tendency.
11
(See Figure 6 Left Side)
30, 000 Joules/ In.
70, 000 Joules/in.
Magnification -500X
Figure 7
-Effect
Of Heat Input On Cracking Tendency.
16Cr-BNi-2Mo Electrode
2018 Electrode
Magnification
Figure 8
-
-50OX
Effect Of Var~ying Electrode Composition On Cracking Tendency.
12
Magnification
50OX
-
Figure 9 - Showing Crack Produced By Puddling
Surface Of HY-80 With Atomic Hydrogen Torch.
Same Phenomenon. But To A Lesser Degree Observed When Heiiare Torch Employed.
%%
()Carbon
(b) Carbon
-0.09%
Magnification
Figure 10
-
Showing Effect
bf
-
-0.
23%
500X
Carbon And Nickel Variations On Cracking Tendency.
13
(c)
Nickel
-
0.02%
(d) Nickel
Magnification
Figure, 10
-
-
3.22%
-500X
Showing Effect Of Carbon And Nickel Variations On Cracking Tendency.
Manganese
Manganese
-0.11%
Magnification
Figure 11
-Showing
-
0.56%
-500X
Effect Of Manganese Variations On Cracking Tendency.
14
-
continued
0:
NN
II
(a)
0-. 0061,
(b)
S-O.081"
Magnification-
*
;
A.A
~ ~
~
"'
(c)
P-O.04 4%, S-O.033%
500X
~~
i•€p
(d)
p-0. 005'ý,, S-O.041ýi
Magnification-
Figure 12 -
P-0. 004'-4,
S-0. 004%
500X
Effect Of Phosphorous And Sulfur On Cracking Tendency.
15
Welding Condition - 250a, 30v, 6" min.
Flux - 80 Baked at 350 F - 12-hours prior to use.
Plate Thickness - 1-inch, Groove -5/8" root spacing, 1-1/4" wide at top
of groove.
Preheat - 200 F, Interpass - 300 F max.
C
Cr
Ni
Mo
Mn
Si
S
P
flate
Coniposi•ion
.16
1.15
2.22
.22
.26
.17
.019
.013
Deposit
Composition
.06
1.40
2.97
.51
.95
.53
.02
.013
All Weld Deposit - T.S. - 139,000 psi, Y.S. -117,000 psi, El. in 2"-15%,
R.A.-38%
Impack Properties ca-60 F. 19-24 Ft-Lbs Charpy V-notch.
Side Bends
-
OK
Transverse Tenisile Test Across Weld
-
Failed in Plate at 113, 700 psi.
Full Size
Figure 13
-
Submerged Arc Weld Made With Experimental Wire.
16
(a)
Left View Of The Work Area Showing Heating Furnace, Wire Drawing And Induction Melting Equipment.
(b)
Right View Of Work Area Showing Rolling Mill And Small Electric Furnace.
Figure 14 - a and b
17
(c)
Close-up Of Wire Drawing Equipment In Operation.
(d)
Gas Analysis Equipment For Determining H2' 0 2 and N2 *
Figure 14 -c and d
18
FLUX DEVELOPMENTS FOR WELDING HY-80
by
D.C. Martin*
*
D. C. Martin, Consultant, Metals Joining Division, Battelle Memorial Institute
19
A.
FLUX DEVELOPMENTS FOR WELDING HY-80
D. C. Martin
" D. C. Martin, Consultant, Metals Joining Division,
Battelle Memorial Institute
About a year and a half ago, the research being
done on HY-80 at Battelle Institute was described at a
Bureau of Ships Seminar. A number of commercial
fluxes and several commercial and experimentalwires
had been studied . In these experiments, the results of
vee-notch Charpy tests were used to compare weld
metals made with the commercial materials,
concerned, there is little to choose between the two
welding techniques. Welds made by either method
have about the same yield strength, ultimate
strength, elongation, and reduction in area. However, this is not so for impact properties'as can be
seen in Figure 1. When you consider that a minimum of 20 ft-lb at -80 F and 50 ft-lb at room temperature is desirable in a weld in HY-80 plate, it is
obvious that the submerged-arc weld is unacceptable.
Such impact results are not surprising in submergedarc welds. They have been accepted in the past.
The results shown in the lower curve were obtained
with the commercial flux shown to be best in our
tests and with a commercial filler wire which is
used in inert-gas-shielded welding of HY-80. In
these and subsequent tests, all impact properties
were obtained from welds having a yield strength of
more than 80,000 psi.
Throughout all of this research, the results of
submerged-arc welds have been compared with the
results obtained from argon-shielded consumableelectrode welds. The major requirements for
welds in HY-80 are a minimum yield strength of
80,000 psi and a 20-ft-lb Charpy rve-notch value at
-80 F. Table I compares the tensile strengths,
elongation and ductilities of welds made by
submerged-arc welding with commercial fluxes and
argon welding. So far as the tensile properties are
TABLE I. TENSILE PROPERTIES OF SUBMERGED-ARC AND
INERT-GAS WELDS MADE WITH SAME FILLER
WIRE AND HEAT INPUTS
Yield
Strength
(0.27 Offset),
psi
Ultimate
Tensile
Strength,
psi
Elongation
in 2 Inches,
per cent
Reduction
in area,
per cent
Submerged arc
100,000
100,000
114,250
114,250
17.1
19.8
52.6
52.3
Inert gas
98,000
102,500
107,750
117,750
18.2
21.4
51.0
68.3
Welding
Process
Since it was obvious that impact properties had
to be improved, efforts were made to do this by
modifying filler-wire compositions. However,
wire compositions which produced outstandingly
better results than the commercial wire shown in
Figure 1 were not developed. Finally, a look was
taken at some of the results obtained in metallographic examinations and analyses of submergedarc welds. This indicated that the proper way to
improve the notch-bar properties of submergetiarc welds was to modIfy the welding fluxes,
silicon content was much higher in the submergedarc weld. Second the oxygen content of the
submerged-arc weld was considerably higher.
Fractional Fas analyses of submerged-arc welds
showed that part of the extra oxygen was tied up in
silicates. However, part was dissolved in the iron
and this was felt to have a bad effect on the properties of the weld metal.
When the weld metals were examined metallographically, quite a difference was found between
the argon-shielded weld metal and the submergedarc weld metal. Figure 2 shows photomicrographs
of a polished but unetched section of inert-gasshielded weld metal and ,ubmerged-arc weld metal.
Two striking differences were found between the
compositions of welds made by'submerged-arc welding and by inert-gas-shielded welding. First, the
20
(b6
4a
A //
36
36
V\
34
32
30
26
S10O
2 1,
Figure I.
26
24
ht per cent
22
20
18
*-e7"
Comparison of Temperature Dependence of Notched-Bar
Properties of
Sulmerged-Arc and Inert-Gas Welds
Made With Same Filler Wire and Heat
Input
21
22go
.
•
••
*•
•
400
" *
S.
Fiu e2.
Coprio
Of Inlsin
InIet
heddAn
N5374N537
umre-r
Inr-GsSh0ddr
0
edMaeWt
(0,)
b
Wl
Fiue2.Cm~rsnOfIcusosInIet-•SllddAd
*
.
0-rcWl
*be
umrgdAc
tnadFu
edMaeWthSadadFu
o
The inert-gas-shielded weld metal is relatively
clean. There are a few very small globular sillcates in the structure. The block-like inclusions
have not been identified, but are probably complex
aluminates.
could be raised. Powder mixes of the standard
flux plus 5 per cent titania and 20 per cent calcium
fluoride were used first. The notch impact properties of the welds were good both at low temperature and at room temperature.
The submerged-arc weld metal contains a large
number of silicate inclusions some of which are
large. The difference
in cleanliness
in the two *impact
wt~d
obiou.
metls
s
The chemical and metallographic analyses indicated that improvements in submerged-arc weldmetal properties
bethe
obtained
by and
reducing
the
oxygen
content as might
well as
number
sizes of
inclusions. It was reasoned that oxygen along with
inclusions migtwas
dhat b r gen d b nwhich
h
the number of inclusionscalcium
creasing the fluidity of the moiten slag. The obvious way to increase the fluidity was to lower the
melting point. A number of additions were m-ide to
the standard commercial flux that was being uwed in
an effort to lower its melting points. In prelirninary tests, these additions were made by mixing
powdered additive with the ground fluxes,
Having found that a flux containing both titania
andaca properties,
floride
woul
p was
wes
with
the
next step
to find
outood
whether there were still better combinations of the
two additives. It was also decided to use fused
fluxes. Consequently, all of the flux compositions
shown by the dots on the diagram in Figure 5 were
shown by thed
wre
prepared
and used on
in teldiag
welding tests. InFir
In their preparation, the standard flux composition was used to
was added various amounts of titania aild
fluoride. Further adjustments were made
carying the s
urter aftustandard mlue
by varyin'g the silica content of the standard flux. "
The numbers on the figure at the various flux
compositions correspond to an arbitrary rating of,
the notch impact behavior of welds made with the
fluxes. Welds made with the flux having a rating
of 1 had the best impact behavior. Welds made
with the composition having a rating of 18 had the
poorest impact behavior. A standard commercial
wire was used in all experiments. This wire is the
one which was developed for inert-gas-shielded
welding of HY-80. All results discussed hereafter
were obtained using this wire, because it was
available in sufficient quantities and eliminated
from the flux experiments any variability from the
wire itself.
The first additive that was tried was calcium
fluoride. This addition was made because it is
known to lower the melting point and increase
the fluidity of neutral fluxes. As can be seen in
Figure 3, fluxes with 10 per cent and 20 per cent
calcium fluoride were tested. The larger addition was the most effective It was found that
calcium fluoride additions did not affect the lowtemperature impact properties greatly but did
raise the room-temperature properties appreciably.
In fact, the flux containing 20 per cent calcium
fluoride produced welds which had room-temperature
notch impact properties that were quite good. They
were high enough to be of interest for HY-80 weldments.
As can be seen in Figure 5, the best results
were obtained with fluxes containing around 5 per
cent titania and 20 per cent calcium fluoride. Consequently, for subsequent developments on fluxes
these amounts of titania and calcium fluoride were
used as basic additions.
The second addition made was titanium dioxide.
Here again, as shown in Figure 4, two different
amounts of titania, 5 per cent and 20 per cent, were
used. In this case, the lower quantity produced the
greatest improvement in notch toughness. The effect of the titania addition was different from that
of the calcium fluoride addition in that with titania
low-temperature properties were increased without
affecting the room-temperature properties a great
deal. The -80 F notch-toughness valuelof 30 ft-lb
obtained with the flux containing 5 per cent titania
was especially interesting from the viewpoint of requirements of weldments in HY-80 steel.
The standard flux used in preparing the experimentJl compositions is said to be neutral. Figure 5
shows that the best of the experimental fluxes not
only contained 5 per cent titania and 20 per cent calcium fluoride, but also had lower amounts of silica
than the poorer fluxes. This suggests that reducing
the acidity of the flux might also play a part in improving the notch-bar properties of submerged-arc
welds. Consequently, steps were taken to determine
the influence of still greater reductions in acidity.
A series of fluxes having different silica to calcium
oxide ratios was therefore prepared. As the ratio
of silica to lime decreased in the fluxes the welds
tended to become cleaner. From this series of experiments a flux was compounded which gave the
best results obtained thus far. Table 2 compares
the composition of this flux with that of the standard.
Its calcium oxide content is twice that of the standard and its silica content is one-half. In addition,
Up to this point, two additions to the standard
flux had been tried, one raised the low-temperature
notch impact properties and the other raised the
room-temperature notch impact properties. Cornbinations of the two were then tried to see if both
the low- and the high-temperature impact properties
23
TABLE 2. Comparison Of Nominal Compositions Of
Standard And Best Experimental Flux
Flux Number
Na 2
Calculated Compositions, parts by weight
K 2 0 MgO CaO MnO A1 2 0 3 502 TiO 2 CaF 2
Standard
2.2
0.35 12.1 20.3 7.2
47 (experimental)
2.2
0.5
10.0 40.0 0.0
titania and calcium fluoride have been added and both
alumina and manganese oxide have been eliminated,
Figure 6 shows the effects of these changes in flux
composition on weld (leanhness. The left-hand
photomicrograph is that of a submerged-arc weld
made with the standard flux and shown earlier in
Figure 2. It will refresh your memory of the quantity, size. and type of inclugions formed on using the
standard flux. This is not an unusual area but is
fairly representative of the whole weld. This weld
has an oxygen content of 900 ppm.
The right-hand photomicrograph is that of a weld
made with the best of the experimental fluxes. It is
easy to see that this weld is much freer of inclusions-,
than the weld made with the standard flux. The oxygen content of the weld metal in this case is 300 ppm.
10.6
38.0 0.0
7.0
0.6
20.0 5.0
20.0
experimental batches were prepared by different
production methods. The notch impact properties
of welds made with the flux prepared in various
ways are compared in Figure 8. The top curve is
for welds using a flux prepared from commercially
pure materials melted in graphite. The same materials melted in a fire-clay crucible gave welds with
impact properties that were considerably lower.
Even welds prepared with flux prepared from highpurity materials melted in a fire-clay crucible had
impact properties that were not so good as those
made with the flux described above melted in graphite.
Figure 8 indicates only one of the Variables that
needs to be taken into account in producing the experimental flux commercially, if maximum notch
impact properties are to be achieved in the welds.
To sum up, both filler wire and flux modifications
have been tried to overcome the problem of -producing
submerged-arc welds with acceptable properties in
It can be secn that the best experimental flux has
a high calcium oxide content, contains calciumcomrilwesanbuedfpoerlxsae
HY-80. The results obtained indicate that available
tat tese fluoes shuld be
tis biev
used
fluoride and titania, and has a low silica content.
chemically basic in character, should contain addiIf it were an open-hearth slag, it would be called
should
fri
cion o l aniu an caci
a basic slag because of its high ratio of calcium
tions of titanium and calcium fluoride, and should
oxide to silica.
not contain manganese dioxide. Using experimental
fluxes made according to these premises, vee-notch
The impact properties of welds produced with
impact values of 55 ft-lb at -80 F and as high as
the best experimental flux are shown in the upper
90 ft-lb at room temperature hav.ý been obtained.
curve in Figure 7. They are compared with the
It is believed that commercially produced fluxes
impact properties of welds made with the same
made to the same composition will make it possible
filler wire by inert-gas-shielded welding and with
to have submerged-arc welds in HY-80 steel that
the standard flux. Although these same resulls
have impact values of between 30 to 40 ft-lb at 80 F
may not be obtainable with this flux if produced
and between 60 to 70 ft-lb at room temperature.
commercially yet they do show the latitude in nqtch
These Charpy values can be obtained in welds havimpact properties that can be obtained in welds
ing minimum yield strengths of 80, 000 psi.
prepared by the submerged-arc process by varying
the flux composition.
At present, further studies of the effects of wire.
compositions are being made to see if changes in
composition will develop even further imprqvements
It is entirely possible that the wire used may not
in notch-bar behavior of submerged-arc welds. A
have been the best for submerged-arc welding. R
few of our own composition modifications have been
was specifically developed for inert-gas-shielded
studied and others who are working with submergedwelding. Wires of other compositions may produce
arc welding have sent samples of wires which they
better results in submerged-arc wc4ding. This poshave developed for us to try. So far, results are
sibility had not been investigated.
not sufficient to draw conclusions.
To indicate what might be expected from the experimental flux when produced commercially, some
In the future, it is hoped that some large batches
of the best experimental fluxes will be made. If
24
.
be the final step in the development of method4 of
making acceptable submerged-arc welds in HY-80
steel.
this is done, samples will be furnished to various
people so that they can try them out under shipyard
welding conditions. It is considered that this will
80
I
a
T----
Experimental Flux-
Inert-Gas Weld Metal
70
-,
..
60 ..
..-
60 .......
...
80
T
Inert-Gas'
410
S20J
0
0
I20
gr0
~30-
0
or-120
z
,
-
-80
Figure 4.
20-
Stcd Flux-
*I
-40
0
Temperature, F
80
40
Pd6635
Effect Of Titanium DioxideAddition To
Standard Flux On Notched-Bar Properties of Sbmerged-Arc Weld Metal
Submerged-Arc Weld Metal
_120
__
-20
Figure 3.
0__
-80
S00
40_0_40_go
0
--40
Temperature, F
40
Commercial-Purity Addi tinsGraphite Crucible
8'0
46630o
Effect Of Additons Of Calcium Fluoride
To Standard Flux On Notched-BBar Propertios Of Submerged-Arc Weld Metal
60 "
L
-
H igh-Purity AdiditionsFireclay Crucible
-
0o
20
-
-
o0
120
-
-2
------------A
Commercial Purity AdditionsFirecloy Crucible
1
-80
0
-40
Temperature, F
1-
40
I
80
N6634?
Figure 5. Titanium, Fluoride, Silica Compositionz Of Experimental Fused Fluxes
25
A.
"0
*
.
.
0
0
-*
"
*
0
*
•
0•
g
"*26
60
-
_-
4
td Flux +200/aCa
F2
_
Sd 10 % Ca 2
S | Flux
'40
C
'-30
-Std Flux
20
....-..
0
z
0
4
-120
-80
40
0
-40
Temperature, F
N"6349
80
Figure 7. Comparison Of Impact Properties Of
Welds Made With Standard SubmergedArc Flux, Inert-Gas- Sielded Arc
Welds, And Welds Made With Experimental Flux
T
!Std Flux +5
n4 0
I
--
J
3
TiO
5edFlux +u20%1,
W
20
-120
M
F
W
E
rStd Flux
---
es'
-80
I
T
-T -
50-
-4
0
c
4-8
0
10--
-120
80
40
0
-40
0066340
Temperature, F
O ImactProperties Of
Figue
Comarion
8
Welds Made With Experimental Fluxes
Made In Various Types Of Crucibles
-80
27
V
NOTES ON
DEVELOPMENT OF "CRACK FREE"
ELECTRODES FOR WELDING HY80 STEEL.
JOHN L. CAHILL
WELDING ENGINEER
NEW YORK NAVAL SHIPYARD
BROOKLYN, N.Y.
BUSHIPS CONFERENCE
March 21-22 1960
Washington, D.C.
28
'
NOTES ON DEVELOPMENT OF "CRACK FREE" ELECTRODES FOR HY-80 STEEL,
Obviously the stipulation as to preheat cannot be
observed in fillet welding and as in the case of
special treatment steel, a different type of electrode
is required for such work performed with no preheat. The type 310 electrode used for STS fillets
also is applicable on HY-80 with no preheat but the
Bureau has not sanctioned its use. Therefore it was
necessary to set up a "crash" program with the
electrode manufacturers to find out what they could
provide at once in the way of a suitable electrode
for use in making attachment welds with no preheat.
It was felt that a lower yield strength weld metal
would shift the load from the heat affected zone to be
distributed and carried partly by the weld metal thus
avoiding or minimizing cracking in the heat affected
zone. The test noted in this first slide was set up
to simulate actual service requirements as a basis
of weeding out the crack sensitive electrodes from
those less apt to show such cracks. Note the remarks
on this sketch as to ambient temperature as well as
the provisions for magnetic particle inspection after
grinding off the weld deposit. False impressions
may be obtained because the HY-80 retains magnetization. This is extremely important in interpretation of results and experience is probably the only
teacher in such interpretation.
About two years ap we had a meeting at the Naval
Research Laboratory to discuss the fabricating of
HY-80 steel in submarine construction. At that meeting several pros aaad cons were brought out relative
to the problems involved in such fabrication. The
various elements entering into the proper use of
Grade 11018 electrodes for welding HY-80 were
thoroughly discussed and the required precautions
spelled out at this meeting and Ln subsequent Bureau
of Ships Notice 9110.
Since that time it has become evident that the 11018
type has proven to be the best electrode available for
the welding of grooves in HY-80. It is superior to the
Grade 260 in low temperature impact strength. The
existing specifications MIL-E-19322 appear to be
adequate for purchase requirements aad this type
electrode is being used currently in several locations with very good success.
However some of the building yards feel that
more latitude in preheat application would be desirable in the welding of HY-80. The ultimate objective of course would be an electrode which could be
used under any and all conditions without preheating
or babying in any way. This UTOPIA will never be
attained particularly in heavy sections of tempered
and quenched steels welded under shipyard conditions but we can strive towards this goal and possibly reach a plateau closer to the utopian objective
than-where we now are. In this effort to arrive at
such a point it is essential that we advance one step
at a time and not try to take too big a jump suddenly
lest we find that the "cure is worse than the disease".
We therefore proposed to the electrode manufacturers
that they try to develop an electrode for use in making
temporary attachments to HY-80 with no preheat.
These kind of welds are prevalent in all shipbuilding
and despite all efforts by Buhips and others to
eliminate them we must live with them in our building work. It is not possible to construct vessels by
welding without having twme of these, temrpurary
attachments.
This next slide shows data on the HY-80 test plate
used in these tests. Note that it is of the high
chemistry type. In order to check the impact properties at 120 F Charpy Vees were made in both directions, that is parallel to the rolling direction.
These tests were repeated after stress relieving at •
1150' F for 2 hours followed by furnace cooling.
Please note that effect of the direction of rolling
on the impact strength at low temperatures. Also
note that the stress relieving treatment did not
materially affect the impact properties.
We obtained data on the various types of electrodes
forwai-ded by the different manufacturers as shown in
this next slide. Four out of eleven types did not crack
on this clip test. A few hundred pounds of each type
were obtained and sent to the submarine building
yards and, in the case of two of them, the yards had
good luck. The other two types have just been forwarded to the yards and it is still too early to have
any data on their performance out in the field. It is
to be stressed that these electrodes are not netessarily commercially available electrodes but have
been developed basically to meet the requirements
set up in our meetings with the electrode manufacturers.
In all cases where such attachments are welded
with Grade 260 or 401B electrodes with no preheat
some cracking is evident after removal of the attachment and subsequent grinding. Repairs to such
cracked areas must be made because such cracks
being in the outside fibres are bound to propagate
under loading. This repair can be very expensive
because of the great number of areas encountered.
It is therefore of the utmost interest that, if possible, an electrode be developed for makmng such
welds. Strength or impact level requirements are
not as rigid as those required for the 11018 type.
As a matter of fact it is generally felt that the yield
strength of this "crack free" electrode might well
be considerably lower than'that of the 11018 type.
It must be understood that this original series of
tests was conducted In order to try to get an electrode out into the yards as quickly as possible so
that they would be enabled to keep on working with
a minimum of cracking troubles. The rather good
29
reports from the first two electrodes sent out
hydraulic pressure rather than by impact loading.
prompted a second look as to why this "crack free"
electrode was less susceptible to cracking than
the 11018 electrodes. From our tests, chemistry,
yield, impact strengths do not follow any pattern.
I personally prefer the impact loading as that is
more nearly what happens in service when clips
are removed. Other types of tests are being considered so that we may wind up with a specification
requirement for "check free" electrodes.
It was then decided to try out modifications of
some of these electrodes in an effort to ascertain,
if possible, the effect of various elements on this
"crack free" property of weld metal of this class,
This next slide shows some of the variations in the
electrodes tested. For example No. 15 is the same
as No. 14 except for the slightly higher manganese
content of the latter. Likewise No. 18 is the same
as No. 15 except that the latter has no molybdenum.
By changing one variant at a time we hope to be
able to come up with the reason for the difference
in susceptibility to cracking between v4rious
analyses of welding deposits. The effect of manganese, nickel, and molybdenum variations is being
studied. At the same time the limitations of the
clip test itself are also being examined. For
instance the impact loading in the removal of c-lips
is a factor in the evaluation of the clip test. We
are just completing a series of tests in which some
clips were removed by hammering with the weld in
compression instead of in tension. It is proposed
to try a series in which loading to break off the
clips will be accomplished by means of a constant
It is recognized that this clip test employed in
the testing of these electrodes differs somewhat
from the more conventional methods used for other
types of electrodes.* However it is our belief that
the tests described herein served a very useful
purpose by providing electrodes which were less
crack sensitive in making attachment welds with no
preheat than were the 11018 type. As noted above
we are continuing with this program as we feel that
even better electrodes can be provided by the electrode manufacturers if we tell them what our targets
are. In a few short years we have progressed
from weld metal with little or no low temperature
impact strength (as comparud with thcse of the
base metal) to weld metal having many times more
impact strength. Our basic target of course Is to
get weld metal equal to the base metal in low
temperature impact strength by use of fabrications
procedures involving use of little or no preheat
and 1, for one, feel extremely confident of the
ability of our suppliers to furnish such material.
30
TESTS OF HY-80 PLATE USED
1-3/8" Thick (Lukens Steel Co.)
Chemistry
. 16
.56
3.08
1. 63
Carbon
Moly
Nickel
Chrome
Impact at - 120' F (As received)
Direction Rolling.
114
114
114
113
113
Impact at -120
ill
111
103
99
90
Trans. Direction Rolling.
62
62
62
62
61
F (After Stress Relief at 1150"F ± 25'F for 2 hours F.C.)
50
47
47
45
44
31
0
C.
0.
0)
03
C,
UMW
n
8l
C4
L'
U
L
0
en
N
0
Nnt
r-
0
0
0
c
o
-i -i
070
01D1
Iin
0n
tN
N
Nn
0
N
Vl
0))
IU
oo
N%
IC..
N
-n
23
'm0
f
0)
cor
C.)
!I
v.)
;
0)
0)
to
-n
,
i
N
Ln
4C
r7( r-lU
uC)
to0
to
t-
LI LO Uin
D
00'
Vco
m
I
tN
co
00
i0
O
-
~~~
CO
IV
t
q
v C.,
ti
c
I
i~ I.U.
IfU)
1
t-.
CO 0
4
AV
UM
w
-.
Ic.
un
to
-
c
to
N
COMPARISON OF 2nd BATCH
-
TABLE U
"CRACK FREE" ELECTRODES (5/32 Size)
CLIP
C
MN
S1
NI
CR
MO
12
.04
1.03
.24
1.75
.29
.32
13
11.05
1.31
.30
2.23
-
.37
14
.05
1.00
.30
2.00
-
.40
15
.05
1.20
.30
2.00
-
.40
16
.05
1.40
.30
2.00
17
.05
1.00
.30
2.00
-
.00
18
.05
1.20
.30
2.00
-
.00
-
.00
TEST
TYPE
.40
REMARKS
Modification of 14
..
..
..
Modification of 17
..
..
..
19
.05
t .40
.30
2.00
20
.05
1.20
.30
1.80
21
.05
1.20
.30
2.50
22
.05
1.40
.30
1.80
23
.05
1.40
.30
2.50
-
.00
24
.06
1.50
.50
1.85
-
.00
25
.03
1.50
.50
1.85-
.00
26
.06
1.50
.50
1.85--
.00
27
.06
1.35
.50
1.85
.30
28
.06
.97
.37
1.58
.22
.05
1.08
.33
1.66
.35
.
.
.05
1.24
.41
1.99
.27
.
.
.
.05
..
..
..
.05
"
"
"
29
30
31
32
_
_
_
.05
1.32
.23
1.92
.05
1.20
.28
3.01
33
.00
. 00
'.00
-
Modification of 20 Higher Nickel
Modification of 20 Higher Mn
Modification of 22 Higher Nickel
Same as 24 with different coating
Modification of 3
.
SOME OBSERVATIONS ON THE WELDABILITY OF QUENCHED
AND TEMPERED HIGH-YIELD-STRENGTH ALLOY STEELS
By W. D. Doty and G. E. Grotke
(Prepared for presentation at a Navy Department Welding
Conference to be held in Washington, D. C., on March 21,
1960, and for publication by the Navy Department.)
Abstract
Navy HY-80 steel and USS "T-1" constructional
alloy steel have opened the door to new opportunities
in the design of engineering structures. However,
as in the introduction of many other new engineering
materials, the use of HY-80 steel and USS "T-l"
constructional alloy steel has required designers
and fabricators to depart from conventional practices
for structural carbon steels.
cracking in welds. It is suggested that a preheat to
insure plate dryness, the use of adequately dried
low-hydrogen electrodes, and proper contouring of
fillet welds would assist in eliminating the welding
difficulties.
Fillet-welded test specimens made using dry
low-hydrogen electrodes showed that heavy-gagecompoqition HY-80 steel was very susceptible to
heat-affected-zone root cracking in contrast to
light-gage-composition HY-80 steel, which was not
susceptible. "T-I" steel was moderately susceptible
to this type of cracking in stress-relieved joints.
Studies indicated that root cracking might be minimized or eliminated in double-fillet-welded tee
joints by depositing the fillet welds in a manner
which would provide a uniform and simultaneously
equal input for each fillet.
In the spring of 1958, several shipyards rn.
countered weld-cracking difficulties in the fatri..,:tion of HY-80 steel for the pressure hull oe subnii
rines, and for nearly two years U. S. Stet.' nae, been
collaborating with the Navy Department and submarine yards in seeking a solution to the problem
This report points out some observations from
studies on cracked welds removed from a submarine, from studies to produce toe cracking in
fillet-welded specimens of HY-80 steel and "T- I
steel, and from studies of the microstructure of
welds in these steels.
The microstructure of the maximum-graincoarsened heat-affected zone in 1/2-inch-thick
heavy-gage-composition HY-80 steel welded at
70 F consisted essentially of untempered and selftempered martensite, whereas the microstructure
at a similar location iA "T-I" steel consisted
essentially of untempered and self-tempered
martensite and balnite. Both steels had, at the
prior austenite grain boundaries, a constituent
believed to be high-carbon, alloy-enriched
martensite. Further metallographic studies are
in progress.
Metallographic studies on specimens from a
frame-to-hull tee joint and from an auxiliaryballast-tank tee joint showed extensive evidence
that intergranular weld-metal cracking and .intergranular heat-affected-zone cracking had occurred
in heavy-gage-composition HY-80 steel and in lightgage-composition HY-80 steel, p~irticularly at the
toes of fillet welds. The type of cracking appeared
to be consistent with the theory of "underbead"
34
Introduction
that weld-metal and heat-affected-zone cracking
had been observed in some highly restrained joints
in "T- " steel. For example, Arnoldl)* has described experiences in the welding of "T-l" steel
spheres in Japan. Although weld-metal-cracking
difficulties were encountered in the early stages of
fabrication, these difficulties were eliminated
when care was taken to insure that the steel was
dry (obtained by 150 F preheat) and to insure that
the low-hydrogen electrodes were essentially
moisture free. Arnold's experiences were with
welded joints not subsequently stress- relieved. In
a few other cases, cracking has been observed in
the heat-affected zone of welded joints in "T-I"
steel, primarily only after the joints had been
stress-relieved. Such difficulties at the toes of
welds have been overcome by properly contouring
the welds to minimize points of stress concentration, by peening at the toes of the welds, or by depositing weld metal having strength lower than that
of the steel being welded and having sufficient
ductility to adjust locally and relieve the stress.
In 1945 U. S. Steel joined a Navy Department,
Bureau of Ships program to develop an improved
submarine hull steel possessing high yield strength
and good notch toughness, together with good
formability and weldability. One of the results of
this program was a low-carbon quenched and
tempered alloy steel of 80,000 psi mininimu yield
strength. This steel was accepted in 1951 by the
Navy Department as HY-80 under the Navy Specification MIL-S-16216.
Concurrent with this undertaking, U. S. Steel
developed a high-yield-strength constructional alloy
steel to meet the needs of industry. This steelUSS "T-1" constructional alloy steel-was designed
to have a yield strength of 90,000 to 100,000 psi,
together with the desirable characteristics of good
low-temperature notch toughness, good weldaLitity,
and sufficient ductility to undergo bending to
reasonable radii,
These two quenched and tempered high-yield-.
strength alloy steels have opened the door to new
opportunities in the design of engineering structures, both military and industrial. The steels
have met with outstanding success Testimony to
this fact may be found in the many military applications in which HY-80 steel has been used, and in
the earth-moving equipment, pressure vessels,
bridges. penstocks, scroll cases, and steel mill
equipment in which "T- 1" steel has been used.
Since the introduction of quenched and tempered
h.gh-yield-strength sti.tvs nearly ten years ago,
much has been published on the welding of these
steels. The purpose of this report is to point out
some additional observations on their weldability.
The observations stem from studies on cracked
welds removed front the pressure hull of a submarine, from studies to develop a weld-test specimen
that will reproduce the type of cracking encountered
by fabricators, and from metallographic studies of
welds.
As in the introduction of many other new engineering materials, the use of ltY-80 steel and "T-l"
steel has required designers and fabricators to depart from conventional practices used with structural 6arbon steels. Thus, it is not surprising that
some fabrication difficulties have been encountered.
In most, if not all these cases, it was possible to
minimize or eliminate the difficulties when recognition was given to the importance of using fabrication practices tailored to meet the nee~ds of highyield-strength alloy steels rather than patterned
after procedures acceptable only for steels which
have moderate strength
Materials and Experimental Work
Samples of Welds From a Submarine
Two samples, each cut from a different multiplepass-welded tee joint made from plates of HY-80
steel were supplied by a submarine shipyard to the
United States Steel Applied Research Laborabory.
One sample, Figure 1, had been part of a frameto-hull tee joint, Figure 2, removed from the
pressure hull. The hull plate was 2-1/4-inchthick HY-80 steel (heavy-gage composition) and the
"frameplate was 1-3, 8-inch-thick HY-80 steel .
(light-gage composition). The root passes in the
joint reportedly had been deposited by the twin-arc,
inert-gas-shielded metal-arc welding process using
Ni-Mo-V steel wire (A632). The remaining passes
in the joint reportedly had been deposited by the
manual metal-arc welding process using either
ELI018 (Mn-Ni-Cr-Mo) electrodes or El0015
(Ni-Mo-V) electrodes. The shipyard conditions
prevailing at the time the joint was welded suggest
the possibility that the joint and the electrodes
were not adequately dry. The second sample
In the spring of 1958, several shipyards involved
in the Navy submarine program encountered weldcracking difficulties in the fabrication of HY-80
steel. These cracking difficulties were encountered
during fillet welding of the frame members to the
hu!l plates. Transverse cracks formed in the weld
metal and toe cracks formed in the heat-affected
zone adjacent to the weld metal. These observations prompted much investigation, and for nearly
two years U. S. Steel has been collaborating with
the Navy Department 7--d submarine yards in seeking a full solution to the problem. Significant to
U. S. Steel's approach to the problem is the fact
See References.
35
supplied by the shipyard, Figure 3, had been part
of an auxiliary ballast tank. In this sample, heavygage-composition HY-80 steel (1-1/2 inch) was
welded to light-gage-composition HY-80 steel
(1-3/8 inch), reportedly by the same procedure
described for the frame-to-hull joint,
of cracking was expressed as a percentage of the
length of the specimen.
Figure 7 gives the details of a multiple-passwelded cruciform-shaped specimen also used for
preliminary studies undertaken in an attempt to reproduce the fillet-weld toe cracking reported by
fabricators of HY-80 steel. One such specimen was prepared as desc rlbed in Table 11,and with the HY -80 steel
identified as "B" in Table I. The specimen was
examined by magnetic-particle inspection to determine whether cracks were present at or near the
weld surfaces.
The chemical composition was determined for
each of the plates and weld metals in the two
samples. However. tensile properties were determined only for the plates in the frame-to-hull
tee joint. Both of the welded joints were examined
by magnetic-particle inspection and then sectioned
and examined metallographically at many locations
to determine the nature and extent of cracking.
After the above-described preliminary studies
were made with multiple-pass-welded tee-shaped
specimens and with a multiple- pass-welded
cruciform-shaped specimen, extensive work was
undertaken using a single-pass-welded cruciformshaped specimen, Figaure 8, similar to that developed by Watertown Arsenal to evaluate the weldability of armor steels. The details of the specimen
are given in Figure 9. Specimens in groups of approximately eight were prepared from the following
1, 2-inch-thick plate materials described in Table
ILl: one light-gage-composition HY-80 steel plate,
two heavy-gage-composition HY-80 steel plates*,
and three "T-l' steel plates. Each group contained
approximately eight specimens because the preliminary studies with the multiple-pass-welded
specimens had raised doubts concerning the reproducibility of cracking encountered in tee-shaped
and cruciform-shaped specimens. The welding
procedure for the preparation of the single-passwelded cruciform-shaped specimen is given in
Table IV. Dry low-hydrogen electrodes were used
in all cases. It will be noted that a preheat was
not used. The selected welding conditions provided
a heat input of 69,000 joules per inch, a value near
the suggested maximum heat input (70,000 joules
per inch) for welds made in 1,2-inch-thick "T-1"
steel plate at 70 F. It will also be noted that specimens, stress-relieved after welding, were prepared
from all the plate steels described in Table I1I,
whereas as-weloed specimens were prepared only
from one heavy-gage-composition HY-80 steel plate.
As-welded specimens of light-gage-composition
HY-80 steel and of "T-l" steel were not included
in the present study because previous experience
with laboratory-size filet-welded specimens prepared from these steels and not stress-relieved
after welding had indicated that these steels when
welded with dry low-hydrogen electrodes were not
susceptible to heat-affected-zone cracking.
Restraint-Cracklng Studies
Soon after U. S. Steel was informed by the Navy
Department in 1958 tlmt weld-crackng ddficulthv•
tlud been encountered in the fabrication of HY-80
steel, preliminary studies were undertaken to reproduce in a laboratory-size welded specimen the
fillet-weld toe cracking reportedly encountered,
Figure 4 shows two views of a multiple-pass-welded
tee-shaped specimen used for these preliminary
studies and Figure 5 gives the details of the specimen. It may be noted that a "stem" plate of 1-inchthick HY-80 steel (light-gage composition) was
fillet-welded to a 'base" plate of 1-Ir 2-inch-thick
HY-80 steel (heavy-gage composition) previously
stiffened by another plate of 1-1, 2-inch-thick HY-80
steel,
Three multiple-pass-welded tee-shaped specimens were prepared using the HY-80 steels described in Table I and the welding procedures described in Table [I. Dry low-hydrogen electrodes
were used in all cases. It may be noted in Table II
that two different preheat temperatures (70 and
200 F) and two different welding sequences were
used. These different conditions were employed in
order to determine the effect of such changes in
welding procedure on crack susceptibility. After
the specimens were prepared, all the welds in the
specimens (including the welds in the large stiffener
plate) were carefully inspected by magnetic-particle
methods to determine whether cracks were present
at or near the weld surfaces. The large stiffener
plate was then removed and the remaining teeshaped portion of the specimen was sectioned to
provide three tee-bend specimens, two metallographic specimens, and three "underbead" cracking
specimens, all as shown in Figure 6. The tee-bend
specimens were tested as simple beams with a concentrated load at the center opposite the stem so
that maximum stress occurred at the toe of the
fillet welds. Bend performance was judged quailtatively rather than quantitatively. The "underbead" cracking specimens were examined by
magnetic -particle-inspection metJ1ds and the extent
Although heavy-gage-composition HY-80 steel is
not normally produced in thicknesses less than
1-1/4 inches, the Laboratory arranged to have
1/2-inch-thick plates (Items D and E in Table m)
of this material produced for welding tests.
36
THIS
PAGE
is
MISSING
IN
ORIGINAL
DOuuIvlm±NI
150 F when plate dryness is not tertain and the use
of adequately dried low-hydrogen electrodes would
assist greatly in reduicing the weld-nietal and heataffected-zone cracking difficulties by eliminating
the sources of hydrogen,
Therefore, no further studies were made with this
specimen and consideration was given to a singlepass-welded cruciform-shaped specimen since
such a specimen was less time consuming to prepare and, therefore, more suited for extensive
studies to determine the reproducibility of heataffected-zone cracking.
Another observation made from examination of
the two tee-joint samples from a submarine concerns the geometry of the fillet welds. The contour
of the face of each of the fillet welds, Figures 14
and 17, was such that an abrupt change occurred at
the toes of the fillets, particularly at the toes adja-
cent to each through plate (or hull plate).
Magnetic-particle inspection of thc weld surfaces
of single-pass-welded cruciform-shaped specimens
in both the as-welded and the stress-relieved conditions revealed no toe cracks or weld-metal cracks.
However, inspection of thc "underb.ead" tracking
specimens prepared from the fillet-welded specimen
revealed extensive amounts of root cracking for
so•me of the steels studied. The results are given
in Tables IX, X, and XI. It may be noted from the
data in Table IX thlat heavy-gage-composition HY80 steel was very susceptible to heat-affected-zone
root crackinig in contrast to lig,.ht-g-age-composition
HY-80 steel which showed no susceptibility. The
amount of root cracking in the weldea and stressrelheved spec imens of heavy-gage-compo~sition
HY-80 steel was aixiut the same as that in similar
specimens in the as-welded condition. In both
cases, the cracking was grqeatest at t~he root of the
"C" fillet. This observation will be discussed later
in regard t(o the effect of welding sequence on the
distribution of cracking. Metallograpiiic examination of a typical root crack in an as-welded specimen of heaiy-gage-composition HY-80 steel
showed that the crack was intergranular, as
illustrated in Figure 21.
These
abrupt changes in contour provided points of stress
concentration when the weld was stressed during
contraction on cooling or during the fabrication of
adjacent members. Therefore, any plastic defor-
mation resulting from high stress was concentrated
at the toes of such poorly co~ntuured fillet welds.
Gsx~d fillet-weld contouring in which the fillets are
faired into the adjacent plates has long been recogniized to be a desirable welding practice, regardless
of the type of material being joined. In the welding
of high-yield-strength materials, refinements in
design, includinlg weld contour, are a must if advantage is to be taken of the high strength of these
materials. Figure 19 shows a view of the auxiliaryballast-tank tee-joint sample •,t a section about 2
inches away from that shown in Figure 17. Note
the very abrupt change in contour at the toe of one
of the fillet welds and the suggested desirable contour.
Results of Restraint-Crackimig Studies
It may be noted from the data in Table X that
"T- 1' steel was moderately susceptible to rut~o
cracking iin contrast to the susceptibtility, shlown
in Table DX, for heavy-gage-compo~sition HY-80
steel. It is also significant that heat-affectedzone root cracks have been observed only in welded
Slpcimens of "T-l' steel, stress-relieved after
welding. Also, note in Table X the rare occurrence
of a toe crack (•qteel H-fillet "A" in specimen No.
2). Present indications from studies in progress to
determipe the cause of crackiilg in thet heat-affected
zone of welds in "T- 1" steel, when such welds are
stress-relieved, suggest that failure occurs by
stress rupture ir. grain-coarsened regions of high
residual tensile stress resulting from welding.
Failure is belheved to occur in the early stage of
the stress-relief trteatmlent before the restidual
stress from welding is signlificantly reduced.
It may be recalled that studies were undertaken
with multiple-pass-welded tee-shaped specimens
of HY-80 steel in an attempt to reproduce the filletweld toe cracking encountered by shipyards. It is
significant that all the welding: wais done with dry
electrodes. From the results summarized in Table
VIII, it may be nioted that no toe cracks or weldmetal cracks were found in any of the specimens,
Hfowever, root cracks were detected in thle two
specimens (No. 1 and 2) in which the first fillet
weld was completed before the second fillet weld
was started. The extent of roo~t cracking was
tower (l6'L. vs 43%1) in the specimen welded alter a
200 F preheat rather than after nio preheat. A
typical root crack is shown in Figure 20. No
cracks were found in the specimen (No. 3) made by
alternately depositing beads on eachl side of the
stem plate. The results of the bend tests on specimens from each of the multiple-pass-welded tee
joints indicated that toe cracking occurred when
each of the specimens was bent a moderate
amount-approximately 20 degrees (included angle).
The results of the tests of thre single-passwelded cruciform-shaped specimens, presented in
Tab~les DCand X, showed that the extent of root
cracking was greatest at thle root of the "C" fillet
and was least at the "D" fillet. The effect of
welding sequence on the distribution of the root
cracks is illustrated by the 'data in Table XI, the
upper half oh which gives the results for fillet
In studies undertaken with a multiple-passwelded cruciform-shaped HY-80 steel specimen,
no toe cracks or weld-metal cracks were detected,
38
•
'
affected zone, would provide information on the
cause of cracking in these high-yield-strength alloy
steels. Bead welds rather than fillet welds were
used for this comparative study, since it was desired to use the simplest weld configuration. The
specimens were welded with a heat input of 47, 000
joules per inch and at initial plate temperatures of
70, 300, and 500 F. To date, the comparative
metallographic study has been completed only for
the specimens welded without a preheat (70 F).
Therefore, only a limited number of observations
can be reported at this time.
welds deposited progressively in a clockwise sequence (standard for this report), and the lower
half of which gives the results for fillet welds deposited progressively in a counterclockwise sequence. It may be noted in Table XI that cracking,
when the welding sequence was counterclockwise,
was greatest at the "A" fillet, instead of the "C"
fillet, and was least at the "B" fillet, instead of
the "D" fillet. Since it was observed that cracking
occurred only in the through plate at the root of the
first fillet weld in double-fillet-welded tee joints
(see Table VIII) and that the cracks were present
only after the second fillet weld was deposited, it
is believed that cracking was greatest at "C" fillet
(clockwise sequence) and at "A" fillet (counterclockwise sequence) because each of these fillets was
the first of a pair of fillet welds at a joint in which
the second weld was made only after the through
plate had been adIfened by the welds on the opposite
side. The results suggest that root cracking might
be minimized or eliminated in double-fillet-welded
tee-joints by depositing the fillet welds uniformly in
a manner which would provide a simultaneously
equal heat input for each fillet. It is known that this
technique has been effective in the past in avoiding
cracking in joints in structural carbon steel welded
under conditions of high restraint,
Figure 22 is a photomacrograph of a bead weld
in "T- i" steel. It is also typical of the appearance,
at a magnification of X7, of a similar type of weld
in heavy-gage-composition HY-80 steel. The numbered circles in Figure 22 indicate the locations in
the specimen that have been given detailed metallographic examination. Typical photomicrographs of
the structure revealed at X1500 by the light microscope and at X10, 000 by the electron microscope
for selected zohes in the HY-80 steel specimens
(heavy-gage-composition) and in the "T-I" steel
specimen are shown in Figures 23 through 25.
Photomicrographs of the structures for comparable
zones in the "T-l".steel specimens are shown in
Figures 26 through 28. The base-metal microstructures for HY-80 steel (heavy-gage composition) and for "T-l" steel are illustrated in Figures
23 and 26, respe,:tively. The microstructure of
the HY-80 steel and the "T-P"steel consisted of
tempered martunsite and tempered bainite.
In summary, the results of the restraint-cracking
studies to develop a laboratory-size welded specimen which would produce fillet-weld toe cracking
showed that such cracks were not produced in
multiple-pass-welded tee-shaped specimens or in a
multiple-pass-welded crucilorm-shaped specimen,
and were rarely produced in A single-pass-welded
cruciform-shaped specimen, all welded with dry
low-hydrogen electrodes. However, the studies
showed that heavy-gage-composition HY-80 steel
wz•s very susceptible to heat-affected-zone root
cracking in contrast to light- gage-compouition
HY-80 steel, which was not susceptible. The
amount of cracking in welded ,ind stress-relieved
joints in heavy-gage-composition HY-80 steel was
about the same as that in as-welded joints. Therefore, stress relieving did not contribute to crack
susceptibility. The "T-I" constructional alloy
steel was moderately susceptible to heat-affectedzone root cracking in welded and stress-relieved
joints. The results of a study of the distribution
of the root cracks indicated that the root cracking
might be minimized or eliminated in double-filletwelded tee joints by depositing the fillet welds in a
manner. which would provide a uniform and simultaneously equal heat input for each fillet.
The microstructures of the maximum-graincoarsened heat-affected zone in the as-welded
HY-80 steel specimen, Figure 24, consisted essentially of untempered and self-tempered
martensite, the constituent at the prior grain
boundaries probably being alloy-enriched areas.
Tempering of the alloy-enriched areas during a
stress-relief treatment resulted in the formation of
agglomerated carbides, Figure 25.
The microstructure of the maximum-graincoarsened heat-affected zone in the as-welded "T-l"
steel specimen, Figure 27, consisted essentially of
untempered and self-tempered martensite and
bainite. Alloy-enriched areas are believed to be
present at the prior austenite grain boundaries and
also adjacent to the bainite needles. Tempering of
these alloy-enriched regions during a stress-relief
treatment resulted in the formation of many agglomerated carbides, Figure 28.
Though the metallographic studies have provided
much information on the microstructural changes
that occur in the weld heat-affected zone of both
HY-80 steel and "T-l" steel, these studies have
not, as yet, clarified the mechanism of crack
susceptibility. Further metallographic studies are
Results of Metallographic Studies
A comparative study of the microstructures of
bead welds in heavy-gage-composition HY-80 steel
and "T-l" steel was undertaken to see whether the
microstructures, particularly those in the heat39
in progress on specimens welded at 300 F and at
500 F, and it is hoped that the information from
these studies, together with that from the specimens welded at 70 F, will be valuable in determining the cause of cracking in the high-yieldstrength alloy steels.
welded cruciform-shaped specimen and
were rarely produced in a single-passwelded cruciform-shaped specimen, all
welded with dry low-hydrogen electrodes.
5.
Fillet-welded test specimens made using
dry low-hydrogen electrodes showed that
heavy-gage-composition HY-80 steel was
very susceptible to heat-affected-zone root
cracking in contrast to light-gagecomposition HY-80 steel, which was not
susceptible. "T-l" constructional alloy
steel was moderately susceptible to heataffected-zone root cracking. The root
cracking in "T-1" steel occurred in welded
and stress-relieved joints, whereas the
root cracking in heavy-gage-composition
HY-80 steel occurred on welding.
6.
A study of the distribution of the root
cracks indicated that the root cracking
might be minimized or eliminated in
double-fillet-welded tee joints by depositing the fillet welds in a manner which
would provide a uniform and simultaneously
equal heat input for each fillet.
7.
Metallographic studies on bead welds made
in 1 2-inch-thick plate at 70 F and with a
heat input of 47,000 joules per inch showed
that the microstructure of the maximumgrain-coarsened heat-affected zone in
heavy- age-composttion HY-80 steel and
"T- I" steel consisted essentially of untempered and self-tempered martensite.
l3th steels had, at the prior austenite grain
boundaries, a constituent believed to be
high-carbon alloy-enriched martensite.
Further metallographic studies are in
progress.
§!mmary
The results of studies on samples of welds from
a submarine, of studies to produce toe cracking in
fillet-welded specimens of HY-80 steel and "T-l"
steel, and of studies of the microstructure of welds
in these steels are summarized as follows:
I.
Metallographic studies of a specimen from
a welded frame-to-hull tee joint and of a
specimen from a welded auxiliary-ballasttank tee joint, both removed from a submarine, showed extensive evidence that
intergranular weld-metal cracking and
intergranular heat-affected-zone cracking
had occurred, particularly at the toes of
the fillet welds. The heat-affected-zone
cracking was observed in heavy-gagecomposition HY-80 steel and in lightgage-composition HY-80 steel.
2.
The type of cracking appeared to be consistent with the theory of "underbead"
cracking in welds. Thus, the use of a
preheat to a temperature of approximately
150 F, when plate dryness is not certain,
and the use of adequately dried lowhydrogen electrodes would assist greatly
in eliminating the weld-metal and the heataffected-zone cra.ýzing difficulties with
HY-80 steel plates.
3.
Abrupt changes in fillet-weld contour provided points of stress concentration in the
frame-to-hull tee joint and in the auxiliaryballast-tank tee joint. This condition
probably contributed to the cause of the toe
cracking in the tee joints removed from
the submarine. Good weld contouring is a
necessity if full advantage is to be taken of
the properties of welded high-yieldstrength alloy steels.
4.
References
Studies to develop a laboratory-size welded
specimen which would produce fillet-weld
toe cracking showed that such cracks were
not produced in a multiple-pass-welded
tee-shaped specimen and in a multiple-pass
40
I.
Arnold, P. C. "Problems Associated With
the Welding of 'T- l' Material," The Welding
Journal, 36 (8), Research Supplement 373-s
to 381-s, August, 1957.
2.
R. D. Stout and W. D. Doty, "Weldability of
Steels," Welding Research Council, 1953.
0
;
i
010
;
-i
0
,-44
,
IV
El
GoE
C.3
'
l izI
oC
-
-10.
0
0C
-ý
V
8
NI
N
I
o.
~
41~6
0
.aI
C;
c
0
0
414
0
p
41
4112
IL~J~
401
ji0
-I
.s
2
III
ef
~:
*~~
~-
I
tv
.
-m
7EC
"1. an
0,
4.
IQ.
V2
C4
C
.A29
N
'
C
4JEDr
-
(42
ZIr
00
g
1
w
04Ci 0
C
CI
@n
ll
0
B
c
inn1
0
II
0
3
C4
0U
.0
0,
n
C4
C-
D
00CI0
-
c
0-i
J4
q
-
0
aga
~
8 ~CJ
43
TABLE IV
Summary of Welding Conditions and Stress - Relief Treatment Used in
Preparation of Single-Pass-Welded Cruciform-Shaped Specimens
Manual Shielded-Metal-Arc Welding
Welding Process:
Initial Plate Temperature:
70 F
*
Interpass Temperature:
70 F
Electrode*:
AWS Class El1018, 5/32-inch diameter
Arc Voltage, volt•
23
Current, amperes:
180
Travel Speed, ipm:
6
Energy Input, joules per inch:
69.000
Fillet Size, inch:
1/4
Stress-Relief Treatment":
The specimens were charged into a furnace heated
to 400 F maximum, heated to 1100 F at a rate of
400 to 500 F per hour, held at 1100 F for one
hour, cooled to 400 1 at a rate of 50 F per hourfollowed by air-cooling.
*
"*
All electrodes immediately upon removal from hermetically sealed containers were stored in
a drying oven at 225 F. The electrodes were used within 30 minutes after removal from the
oven .
As-welded LbpecLmens were prepared from one heavy-gage-composition HY-80 steel plate
(Item D, Table MI)whereas stress-relieved specimens were prepared from all the plate
materials described in Table MI
44
0
0
TABLE V
Summary of Welding Conditions Used in
Preparation of Bead-Welded Specimens
Welding Process:
Automatic Slelded Metal-Arc Welding
Initial Plate Temperatureo:
70, 300, and 500 F
Electrode*:
AWS Class E11018, 3/16-inch diameter
Arc Voltage, volts:
22
Current, amperes:
260
Travel Speed, 1pm:
7.3
Energy Input, joules per inch:
47,000
"OAll electrodes
immediately upon removal from hermetically sealed containers were stored
in a drying oven at 225 F. The electrodes were used within 30 minutes after removal from
the oven.
45
98
C4-
N
44
' ei
00
i
44
St
2t
eq
an
00
-%
-0
0)
0
r
4
O
0cc4~.
-C
ou)I
4.
nO
00
0
0
os
0cl
0D
0
0t
0
0
0
4)1
-
~
0
0
0
--
~~4
04)
0
0
0
0
#U
-~4
0
co
c
-
0
an
cc Adc
Vn
-l4
-4
CCL
&I
co
0C
j
46
IV
90
ev 0
0
6z
z
M L.
0
Qo84IA
o.
04)
v
-1 .2
.2)
o
4)
g4L
A
06
M
4)
cm
0Q)
4)
00
0!)4
c~cj
447
.
0VcO~o-C0000000
Qj
0unO
,
0000000000
8
.
-'
r, ---
'
1100 0
-
-W
-
0000000
u0OtC
In 00
C4 C
*
n
0
Q0 00C0000D
0Q
..
-
000)
Iw t
0000
In
C-
In
)
ci
::u
4:A
Io~~~~-
,Th i
m0In
'D
u)c
vCII-
C13
4 C.
~
-~Cc
~
L
fz
C:. X
o~~~
c
V
U
t-7
C! U
U
48)
L.
TABLE X
Results of Tests of Single- Pass- Welded Cruciform-Shaped Specimens
of 1, 2-Inch-Thick "T-1" Steel
Steel*
Specimen
No.
"Underbead" Cracking, per cent of weld length*
St ress- Relieved Corndition**
Fillet
Fillet
Fillet
Fillet
D
C
B
A
Specimen
Average
Item
Type
G
USS "T- I"
1
2
3
4
5
6
7
8
0
0
0
0
0
0
0
0
0
0
0
3
0
0
0
0
25
0
0
0
0
1
10
0
0
0
0
0
0
0
0
0
6
0
0
1
0
0
3
0
H
USS "T-I"
1
2
3
4
5
6
7
8
0
0+
25
0
0
0
0
0
0
0
0
15
0
0
0
0
45
0
0
15
2
23
27
43
0
0
0
0
0
0
0
0
11
0
6
8
1
6
9
11
USS "T- 1"
1
2
3
4
5
6
7
8
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
15
0
15
90
60
0
20
45
0
0
0
0
0
0
0
0
4
0
4
23
15
0
5
11
See Table.Ill for complete identity of material.
Values shown. indicate extent of root cracking.
wise indicated by note
"*"
{
No toe cracks were observed unless other-
Previous experience with as-welded specimens (different type of fillet-welded specimen
of "T-1" steel welded with dry low-hydrogen electrodes) indicated no susceptibility to
heat-affected-zone cracking,
A 0. 10-inch-long toe crack was detected.
49
TABLE Xl
Effect of Welding Sequence on the Distribution
of Underbead Cracks in Single- Pass-Welded
Cruciform-Shaped Specimens
Welding Sequence
Spcimen
for Fillet Welds
No
Clockwise Sequence
(As Shown in Fig. 9)
j
1
K
1 AýjB\2
\ ,)
L
Cuntur _ý:kwise
Sequence
S•
.I
"Underbead" Cracking, per cent of weld length- .
Fillet
Fillet
Fillet
Fillet
Specimen
A
B
C
D
Average
1
2
3
4
5
6
7
8
9
10
37
50
0
0
0
60
61
0
15
25
0
10
0
0
0
0
0
0
0
0
50
40
27
55
25
83
87
79
35
30
23
0
0
0
0
0
0
0
0
0
28
25
7
14
6
36
37
18
13
14
1
2
3
4
100
100
90
90
0
0
0
0
0
0
0
0
50
87
50
30
38
47
35
30
85
0
0
0
21
6
100
0
0
45
36
.45
2d
TC3
'\D~
Stress-relieved specimens of 1, 2-inch-thick HY-80 steel (heavy-gagecomposition Steel D in Table UI). Values shown indicate extent of
root cracking. No toe cracks were observed.
50
CRACKS
PLATE
HULL.
PLATE
Figure I.
Sanirple of franic-to-hull tee-joint remnoved fromi the prtessurehiul of :i zubmariine.. View zhaws
side of joint where craicks were observed.-
51
FRAME
HlIt L
PLATE
PI A'f
Fikure 2. Framim-to-huUl tee-Joint, reinoved~from the pre'nuure hull of a submarine.
STEW
-BASF
T.
"MO4Ch
Fipi..rv 3.
T'AýHF1
&iniplc of auxdiarv -blxla~st-tjiik
ronioved fr~ini a1 siubiarinw
52
t('t-j(iflt
CA
ROUGH9
A
Side view
53!
,41
B.
Figure 4.
End view.
Multiple-pass-welded tee-shaped specimen.
53
2
_
_
_
I
_
4S K(ftti 1d•.• VD(
SO-WO
-II
LAND)
_t
S(LD 110 OF
MOPaO
TO
JOAMAO"
|
IJ
ZZ•<
LIZIL
"'Ul
1L&Im0L.A
Details (of Preparation of Multiple- PussWelded Tee - Shaped Spec imen
Figure 5.
I
I
I
I
!
II
____
\-
I--
-_
_ _
BEAD"
CRACK
CRACKj
lrdCI
,
,
I•.
"
I_
I_ _ u
_
I
BEAD-"
I
|
'i- I
,NcBoIl BEAD"
CRACK
1
-____
METALLOGRAPHIIC
STEP I
SPECIMEN SAW CUT AS SHOWN "UNDERBEAD"
CRACK SECTIONS WERE THEN PREPARED AS
SHOWN IN STEPS
11 AND
M
TYPICAL
I
AFTER
THE "UNDERBEAD"
CRACK SECTIONS WERE
SAW CUT AS SHOWN, THE
ROOT CRACK
-0CA
i\
:!:
WELD DEPOSITS WERE
CAREFULLY GROUND FLUSH It
I
WITH THE SURFACE OF
APPEARANCE
L._Jl
L
._J
INSPECTION
DILL..
Figure 6.
OF"UNDERBEAD"
CRACK SECTION AFTER ETCHING
AND MAGNETIC-PARTICLE
Details of Sectioning and Inspection Procedure for Multiple- Pass-Welded Tee-Shaped Specimens
54
tU
VINK MAGuNONi
ShR&iNIT
AND IOVA0i POWe
PI.LIT UILJ* WM NAD
BY TH9 SIMLhAN90Ug
m'suIOsNi OF Porno I *1010.
3 AND". SAMO6. IT*
V£L&MU S(OUINCA
Figure 7.
Details of Preparation of Multiple-Pass-
Welded Cruc iform- Shaped Specimen
Figure 8. Single -pass -welded., c ruciform -shaped
spec imen
55
VI
SCCAlt•v'
.OMlt
IV L
I
tlin SIp'a~tI
iO,.?a
WOLD-N6
IO941i Cf Slim
Of
SIOUISC-
OFSNtG. -ftSS(P&ftPRATIOI
WELOEDCRUCIFORMS-o.KA SlPtC.OIN
9TALtS OF
F'igure 9.
Detailsof Preparation of Single-Pass
Welded Crucdiorm- Shaped Specimen
SPECIMENS WERE
SAW CUT AS SHOWN
gI
AFTER THE STEMS WERE
THE WELD DEPOSITS
WERE CAREFULLY GROUND
/REMOVED,
FLUSH WITH THE SURFACES
.. ,,OF
THE THROUGH PLATE
I,
SI
D A3 SOWN KEOW.
CIOWEELD
WELD
SECTIONS Ci) AO
RITAINTO FOR
I'
IU
STEP MI, SECTIONS (4 AND C,) RETAINED
,FOR METALLOGRAPHIC EXAMINATION
DPREPARATION
A
TO ETCHING
0
OF SECTIONS
AND (
AND MAGNAFLUXING
PRIOR
TYPICAL
CRACK
SROOT
I
L
---- J
.EP
_J
Dgt
MARO~
T1__
ITLP
APPEARANCE OF SECTIONS ()
AND (•) AFTER
ETCHING AND MAGNETIC PARTICLE INSPECTION
Figure WO. Details of the Sectioning and Inspection Procedure for Single-Pass-Welded Cruciform-Shaped
Spec linens
56
0.
"KO
A071
ROD
"ýOr
SpwMma
MILOO APII
wLOg
.&LF OFSpac~wift
si
106.90
Pon
.
*eALGR01
CIG*Ia
Figure 11.
Details of Preparation Bead-Welded
Specimen
57
CRCACKS
'
,
.
•ME
TAI
CRACK
Figure 12.
Locations of weld-metal cracks and a toe crack found by magnetic-particle inspection of the
frame-to-hull tee-joint sample.
58
Ait
414;
lb
X1500.
S.
FigureI3./
Tyia
negau
serve in
he
E che in-prPca
~*.59
wl-ea
f
amet-
rc,
ltejitsml
b
Figure 14.
Photomacrograph showing heataffected-zone cracks in the heavy-gagecomposition HY-8O steel hull plate in the
frame-to-hull tee-joint sample. Etched
in Ammzonium Persulfate. X6.
60
A, Aj ZOWJE
A.
Region where craick propaLgated through weld metal
adjacent to the bond. X500.
4W(Rt
Figure 16.
Te rm i nu s of a heat-affected- zone
crack in the light -gage-composit ion
HY-80 steel frame plate of the
frame-to-hull tee-joint sample.
Etcheed in Super Picral. X250.
"TuE.
STET
Cu
M
?Of Cfl
ANI
B.
TD aW_
t
lntergranular crack in the grain-coarsened heataff'acted-zone. X1000.
Figure 15.
Photomicrographs of atoe crackin heavygage-composItion HY-80 steel hull plate
of frame-to-huUl tee-joint sample. Etched
in Super Picral.
Figure 17. Photomacrograph of the auxiliaryballast-tank tee-joint sample showing a
slag inclusion, a weld-metal crack and
several heat-affec ted zone cracks.
Etched in Ammonium Persulfate. XL.
61
U'
t
A.
B
.-
flW,.Arnc'IC
Heavy- gage -composition HY-80 steel stem plate.
X250.
Lgt-pe-cmoiin
B.Figuret8 gagecompositio
Figure 19.
H-0sel.aepae
cr-ackste baste plate.
coarsened heat-affected zones of the
auxil iary- ballast -tank tee -joint skmple.
Etched in Super Picral.
82
A view of the auxil iary-ballast -tank teejoint sample showing abrupt change in
contour at the toes of the fillet welds and
suggested desirable contour. Ammonium
Persulfate Etch. Xl.
'a.
"
.
..
"
,WELD
,'. ,%
METAL
°
"
: ",-
. . "/
"HEAT-AFFECTED
ZONE
S.
°
"
,
Figure 21.
"
Typical heat-af ected-zone root crack
in single-pass-welded c ruciformshaped specimen of heavy-gagecomposition HY-80 steel. As-welded
condition. EtchedinSuper Picral.
X500.
-
Figure 22.
~Aý
Phiotornicrograph of bead weld in USS "T- 1" steel plate welded
at 70 F. Numbered circles identify the zones deicribed below.
Etched in Nital -X7.
Location
Description
I
2
Weld Metal
Weld Metal and Maxxmnuni-Grain-Coarsened
Heat-Affected Zone
Moderately Grain-Coarsened Heat-Affected Zone
GrAin-Refined Heat-Affected Zone
Grain- Refined Heat -Affected Zone
AI-A 3 Heat-Affected Zone
Base Metal
3
4
5
6
7
64
tj.
A.
Ligh
phti
o grap
X10.A-ih
htmcorp.X50
A,~
AB
Eleghtro photomicrograph.
Figure 23.
0
X 10,00.
A. lightro photomicrograph. X 10,00.
Microstructure of heavy-gagecomposition HY-80 steel plate. Unwelded. Etched in Super Picral.
Figure 24.
65
M icr osatr u ct u re of mimum..prain..
coarsened heat -affected- zone in heavygage-composition HY-80 steel plate
welded at 70 F. As-welded condition.
Etched in S9uper Picral.
lbi
k Or
14-
~
A
2
htmcorp.
~ ~
Lih
b~
hto
cr
rp.X50
~~
A.Eleghtro photomnicrograph.
Figure 25.
ih
10.A
'~A
X15,00.
A.Elgctro photomicrograph.
Micros~.ructure of maximum-graincoarsened heat -affected- zone in heavygage-composition HY-80 steel plate
welded at 70 F. Stress-relieved condition. Etched In Super Picral.
Figure 26.
66
X15,00.
Mficrostructure of USS"T-1"1 steel plate.
Unwelded. Etched in Super Picral.
Pt
A
Ligh
.,
'•-°'
"
B
E .
.
"
.
"k
5
,'.
•
t
A.
B.
Light photomicrogr'aph.
Electron photomicrograph.
Figure 27.
XI500.
X
0#."
p,
Fgr*
,•"
p
f
S
X.10, 000.
1
27.um
I
,
*,•
A. Light photomicrograph.
Xl0,000.
. 28
M
0#
c
I
X15O0.
B. Electron photomicrograph. X10,000.
Microstructure of maximum-grain-
Figure 28.
Microstrueture of maximum-grain-
coarsened heat-affected zone in USS
coarsened heat-affected zone in USS
"T-1" steel plate welded at 70 F. Aswelded condition. Etched in Super
Picral.
"T-" steel plate weldedat 70F. Stressrelieve( condition. Etched in Diper
Picral.
67
FATIGUE PROPERTIES OF WELDS IN 111-80 STEEL
I. L. STERN
(Head, Welding Development Section)
by
and
H. V. CORDIANO
(Acting Head, Mechanics Branch)
MATERIAL LABORATORY
NEW YORK NAVAL SHIPYARD
*
INTRODUCTION
In view of the above, it was decided to develop a
suitable specimen for the available Material Laboratory beam fatigue testing machine and, in addition
to develop a new fatigue machine to apply repeated
uniform loading to plate type specimens.
Background
In view of the fact that HY-80 alloy steel is used
extensively in the Naval service, information is required for its application as a structural material.
One of the more important considerations is the effects of welding on the HY-80 base metal structure
and properties. While a background.of information
has been accumulated for static properties of welds
in HY-80 steel, there have been little data available
relative to fatigue properties.
Large Scale Beam Type Fatigue Tests
A sketch of the large scale beam type fatigue
specimen developed for this investigation is given in
Figure 1. The upper portion of the specimen is made
from HY-80 steel and is 5 inches wide by I inch thick
at the critical section. The specimen is proportioned
so as to permit developing the desired high stresses
at the weld in the largest welded joint practicable
within the capacity of the machine. The lower portion was grade HT plate.
In January 1959, the Bureau of Ships authorized
the Material Laboratory to obtain data relative to
the fatigue properties of welds in rich chemistry
HY-80 steel. The immediate need wa! for information relative to the fatigue lives of fillet assemblies;
however, it was indicated that similar data for butt
welds would also be required. It was also known
that the long range requirements of the program,
which could involve other joints and materials
would require apparatus which could operate near
the yield strength of HY-80 material (approximately
80,000 psi).
The machine used for dynamically loading the
beam type specimens has been described in detail
in another paper. * The machine is of the vibratory
type and may be classified as a constant-load
repeated-bending machine. The repeated bending
moment which is generated by rotating eccentric
discs is impressed on the specimen in such a manner as to be uniform along the span. The cross
section of the specimen is so designed that the maximum stress occurs in the outer fibers at the weld.
The machine applies completely reversed stresses
to the specimen at a frequency depending on the
stiffnesa of the assembly. This frequency which is
approximately 430 cycles per minute is maintained,
constant by the machine until failure occurs, at
which time the safety cut-off switch is automatically
actuated. The impressed repeated bending moment
is readily calculated from the weight, eccentricity,
speed of rotation and shaft spacing of the eccentric
discs, thereby permitting a simple calculation of
stress at the critical section of the specimen. Figure 2 shows a specimen set up in the machine for
test.
The related field problems which had required
the development of fatigue data involved complex
weldments in comparatively heavy thicknesses.
Consequentiy, small scale tests were not considered suitable because they would not take into
account such factors as residual stress, mass effects or variations in contour and quality normally
encountered in a weldment. On the other hand, it
was realized that an excessively large specimen
was objectionable for the following reasons:
a. Fewer specimens would be obtained for a
given effort and amount of material. In
view of the normal variations encountered
in fatigue testing, conclusions based on an
insufficient number of samples could be
Large Scale Plate Type Fatigue Tests
A sketch 6( the large scale
plate type fatigue
specimen developed for uniform loading under pneumatic pressure is given in Figure 3. This plate type
specimen, which was used in most plate tests
misleading.
b. In very large scale specimens consistency
and reproducibility become more difficult
to obtain. In some cases failure may be indoced by a random flaw in base or weld
metal.
*A Unique Machine for Large Scale Fatigue Testing,
ASTd STP 216, 1957
66
e8A
conducted to date, was designed to have a thlcknzess
of one inch, so that tests could be conducted with the
available building compressed air supply of 100 psi.
However, compressed air at higher pressures has
since become available in sufficient quantity, and
current tests are being conducted on full thickness
(1-1/2") assemblies.
PROCEDURE
The apparatus for applying repeated uniform
loading to these plate type specimens was specifically designed and constructed for this work. A
photograph of the complete assembly Is given in
Figure 4. The apparatus consists of an air supply,
a reservoir to minimize pressure drop in the supply line, a loading frame, a system of solenoid
valves and pressuIle switches, a deflection recorder, a pressure recorder, and other incidental
pieces of equipment. The strain gage indicator and
recorder employed for stress measurement are not
shown in the photo. The plate type specimen is
simply supported at the two edges parallel to the
weld at a span of 28 inches, nominal, and is free at
the other two edges. A sheet of neoprene gasket
material is cemented to the top edge-surfaces of the
lower frame to seal the loading chamber. The system was originally designed for operation with water
but more rapid cyclical loading, approximately 20
cycles per minute, was obtained by operating with
compressed air. The cyclical stresses obtainedai-e
not reversed but range from zero to a maximum.
Since two edges of the plate are free, stresses may
be calculated by means of the simple beam formula.
A pressure recorder and deflection recorder are
connected to the test equipment in order to provide
continuous records of pressure and deflection throughout the test. In addition, two dial indicators arefixed
to the stationary frame so as to indicate deflections
of the plate at two points along the web symmetrically
located with respect to the center. An electrical
cycle counter is also included in the solenoid valve
control circuit to indicate the total number of loading cycles at any time.
Observations made during the course of the testt,
reveal that failure of the plate type specimens starts
long before ultimate failure occurs. This failure
parallels the increase in deflection recorded during
the test. In the case of tests of welds from 0 to
maximum tension, the increase In deflection starts
before the first indication of a crack has been observed. It appears therefore that in these :ases,
the increase in deflection may be taken as a measure of the overall structural damage of the plate.
In order to arrive at some rational basis for determining the fatigue life of the plate type specimen,
an increase in deflection of 10 percent over the
initial deflection has been tentatively chosen as the
criterion of fatigue failure. This value has lbeen
chosen because it has represented substantial failure of the specimen and the start of rapid increase
in deflection accompanying the increase in crack
propagation near ultimate failure.
All HY-80 material used ranged from 1-7/16" to
1-3/4" in thickness and was in conformity with the
requirements of specification MIL-S-16216D whose
chemical and mechanical requirements are shown in
Table I. Details of typical welding procedures,
joint designs and pass sequences used are shown in
figures 5 through 8 and table 2. In the case of fillet
assemblies, web and flange were of equal thickness.
While it was realized that in actual application the
web member is usually thinner than the flange, the
relatively heavier web thickness was used to emphasize the effects of the weldability factor.
For the dynamically loaded beam tests, a 24" x
8" (Min.) web was welded to a 66" x 24" flange (See
Figure 9). As indicated therein, the welds were
made parallel to the final direction of rolling of the
flange and opposite in direction to each other. The
welded assembly was machined into 4 specimens as
shown. The start of weld and the finishing weld
crater were not included in the fatigue specimens.
Each machined specimen was then welded to the high
tensile (Grade HT, MIL-S-16113B) bottom plate.
For the plate fatigue fillet specimens, a 35" x 8"
(Min.) web was welded to a 35" x 30" flange, and
then machined as was shown in Figure 3. A procedure was developed to minimize the distortion resuiting from the extensive welding and machining
involved. In this test, the stress imposed at a
given pressure varies with the thickness of the plate.
A uniform thickness adiacent to the length of the
weld is required in order to yield a uniform stress
distribution in the heat affected zone. Maximum
thickness at the edges would tend to produce initial
failure where edge effects are present, thereby
complicating the interpretation of results. Since a
plate of absolute uniformity is not possible, it was
decided to utilize an asxembly wherein the maximum
thickness was at the center and the thickness variation in the weld area did not exceed 0. 032". The
method described in Figure 6 resulted in a surface
with a longitudinal convexity approximating 1/32".
This degree of flatness was maintained throughout
the subsequent machining operations.
The increased plate thickness at the center, which
represents an area of slightly greater stress under
the conditions of test, is not considered objectionable and may even be considered desirable for the
following reasons:
a. The effect of thickness variatloo of this magnitude is small compared to variations due to
other effects introduced in the course of fabrication of the assembly.
b. The fracture tends fo be initiated away from
the edges, thereby minimizing any complications which might be introduced due to edge
effects.
69
c. Since the approximate zone of fracture can
be localized, this area can be explored fully
in the course of test.
The distortion encountered (approximately 3/8")
in the course of machining the butt assemblies from
1-11/16" to the hlesired 1" thickness was appreciably
greater than that observed with the fillet assemblies.
This was due to the fact that, in the case of the fillet
assemblies, the 2" web section stiffened the assembly and minimized movement longitudinal to the weld
during machining. The lastAree of the eight butt
welded specimens did not require reduction to a I"
thickness, since the available air supply and apparatus were able to accommodate specimens of full
plate thickness.
The welding procure for plate fatigue butt samplea is shown in Figure 8. As indicated therein, all
assemblies were restrained by welding two 30" x 6"
x I" bars to the assembly. In addition, all layers
prior to the last two on each side were welded in the
horizontal position (plate vertical). This procedure
provided the following advantages:
a. The moderate degree of restraint employed
more closely approximated field conditions
than unrestrained assemblies,
Beam fatigue and plate fatigue specimens were
then subjected to fatigue loading in their respective
types of testing machines.
b. Welding in the horizontal position (plate
vertical) enabled two operators to work stmultaneously. This shortened the time of
fabrication and -qualized heat input and
thermal stresses imposed.
After test, each assembly was examined by magnetic particle inspection to determine the extent and
location of the fatigue cracks produced. A hardness
survey was made on representative assemblies.
Several specimens which had .,mcotested were fractured to study the fracture pattern. Reprcscntativc
sections of welds were examined metallographically
at lOOX magnification.
c. The restraining bars maintained flatne.s
throughout the welding operation.
d. The equalization of heat inppt in turn resuited in specimens which were essentially
flat (within 1/'16") after the restraining side
bars had been removed,
Using the techniques and tests described above,
the following information relative to fatigue loading
in HY-80 weldments was developed:
Type
Assembly
Test
Stress
Fillet
Fillet
Fillet
Beam
Plate
Beam
Alternating
0 to tension
Alternating
Butt
Beam
Alternating
Butt
Fillet
Plate
Plate
Beam
0 to tension
0 to compression
0 to compression
Objective
.
In addition a total of eight beam type specimens of
fillet welds (Mil-8016 weld deposits) in HTS base
metal were tested at 32,000 and 37,000 psi to compare the properties of fillet veldv in HTS and HY-80.
S-N Curve
S-N Curve
Investigation effects of grinding
and shot peening
-Comparison Butt vs. Fillet
[assemblies; ultimately to
Ldevelop S-N curve
S-N curve
S-N curve (currently underway)
S-N curve (currently underway)
Fillet Welds - Plate Type Specimens - 0 to maximum
tension
Prior to testing, six SR-4 type strain gages were
applied to each specimen in order to measure strains
in the base plate at the toe of the fillet weld on each
side of the web. Three gages were cemented on each
side, one at the center and one near each edge of the
base plate. The first step In the testing of each plate
was to obtain static load (uniform pressure) versus
deflection and static load versus strain data for equal
increments of load from zero to the maximum to be
applied during the cyclical loading. The plate was
then subjected to cyclical loading antil ultimate
failure occurred. Continuous records of cyclical
pressure and cyclical deflection were obtained for
the entire test of each plate. During the course of
the test, minimum and maximum dial indicator
readings were recorded at intervals of approximately
Fillet Welds - Beam Type Specimens - Alternating
Stress
Eight beam type specimens (PB series) were subjected to cyclical stress on the dynamic vibratory
machine. Stress levels required to give a life in the
range of 10,000 to 100,000 cycles were ;stimated for
the first few specimens and these results used as a
guide to select stresses for the remaining specimens
so as to provide S-N data over the desired range of
lives. The preliminary results obtained in the above
manner were used as a guide in selecting the desired
stresses for the conduct of fatigue tests on the major
series of specimens (LB series),
70
500 loading cycles. In addition, observations were
made to determine the approximate number of loading cycles at which a crack started and progressed
to various lengths throughout the test.
in a specimen with a straightness deviation of less
than 1/16".
Fillet Welds - Comparison of the Effects of Grinding
and Shot Peening - Beam Type Specimens
As was indicated above, the first five specimens
were tested with thickness reduced from 1-11/16"
to 1". With the subsequent availability of sufficient
high pressure compressed-air, it was possible to
test the last three specimens in full plate thickness.
Butt Welds - Plate rMT
All weld assemblies for these tests were derived
from a single 1. 427" thick plate in accordance with
Figure 10. Shot peening was done under contract in
accordance with laboratory specifications. The apparatus employed is shown diagrammatically in
Figure 11 along with details of the peening procedure.
A macro representing adjacent section was shot
peened simultaneously with each specimen. The
peened surface and macrosection are shown in Figure 12. Intensity of shot peening was verified by
using an individual Almen strip for each specinien
and by inspecting each peened surface at lOX magnification. The method of measurement of intensity
of shot peening by means of Almen strips is shown
in Figure 13. In brief, the principle involved is
that shot peening depresses the metal on the surface,
and in the case of the Almen strip, results in an
elongated surface. This in turr produces the curvature shown, the curvature being proportional to the
amount of cold work or intensity of shot peening.
An example of the extent to which shot peening deforms the surface of metal is shown in Figure 14.
Examples of shot peened surfaces, viewed at lOX
magnification, are shown in Figure 15. Figures 16
through 18, which show sections at 250X, indicate
the extent of cold work induced. As indicated therein, the observable deformed metal appears to be
less than .008" deep. Shot peening was accomplished
on each specimen as a final operation prior to testing,
imens
Fillet Welds - Compression - Plate Type Specimens
Plate type specimens are being tested in a manner
similar to that previously described for the plate type
0 to maximum tension tests, except the pressure is
imposed on the welded side of the assembly. Full
plate thickness (1-1/2") specimens are being used.
Since the weld is inside the pressurized chamber, it
is not possible to observe the formation of cracks.
Therefore, it became necessary to develop a system
of inspection from the unwelded side for detecting
cracks at the toe of a fillet. This was accomplished
by ultrasonic methods using a shear type crystal on
the exposed face, at a distance approximating the
plate thickness under test. The application of this
technique is shown in Figure •2.
RESULTS
Fillet Welds - LBam Trpe Tests
Results of fatigue tests of fillet welds on beam
type specimens are shown plotted on semi-logarithm-ic
paper in Figure 23. The curve is filted by eye through
the median points, the intermediate point of each set
of three, at the six stress levels. The straight line
representing the S-N data for a preliminary (PB)
series is also shown for comparison.
Grinding operations were conducted with an AR
6020, 600 taper cup wheel on a full 24 inch width of
assembly prior to sectioning. This eliminated end
effects. Photographs of welds in the "as ground",
and ground and shot peened conditions, are shown in
Figures 19 and 20. As indicated therein, the desired contour is readily obtainable. While initial
undercutting of the flange was encountered, this dif-
With due consideration for the scatter of results,
the S-N curve for the LB series of beam type specimens, Figure 23, indicates approximate fatigue
strengths for various lives as follows:
Fatigue Life.
ficulty was overcome by intentionally clogging the
lower face of the wheel by running it over an asphalt
tile. (See Figure 21.)
All assemblies were tested in the beam type machine ii the manner previously described.
Approximate Fatigue Strength
Cycles to Failure
psi
10,000
100,000
1,000,000
52,000
20,000
11,000
The results represent ultimate failure of the
specimens, or the point at which the deflection, or
amplitude of vibration, increased sufficiently to
actuate the cut-off switch. In most cases the crack
at the toe of the fillet was found to have progressed
across the full 5 inches of width of the specimen but
not through the depth. The failures started anywhere
along the toe of the fillet weld and did not favor the
edges of the specimen or any other location. The
beginning of failure was observedto start shortly
Butt Welds - Beam Type Specimens
Butt weld assemblies were sectioned and tested
in a manner similar to the fillet welds described
above. In the course of reducing the area under test
to the required 1" thickness, it was found that some
distortion was encountered as a result of welding,
However, machining the convex side of the assembly
to reduce the thickness from 1-1/2" to V"resulted
71
number of inches of crack on one side of the web corresponding to the indicated deflection.
before ultimate failure occurred. In general. failures occurred on only one side, but in a number of
instances a crack had also started on the opposite
side.
Measurements of stress by means of SR-4 strain
gages reveal that the stress developed is not uniform
across the plate near the toe of the weld. The measured stres:ses are highest near the center of the plate
where the thickness is greatest. In addition, the calculated stress appears to be representative of the
average of the measured stresses.
I
Typical macrospecimens before and after testing,
as well as typical paths of fracture, are indicated in
Figures 24 and 25. Photomicrographs of a failed
section, and the adjacent section which had not been
subjected to fatigue tests are shown in Figures 26
and 27. A hardness survey on typical macrospecimens is shown in Figure 24. It should be noted that
the hardness results on the web member were influenced by flame cutting, as well as subsequent
welding. Accordingly, these results have not been
included in the summary shown below. A summary
of the results of Rockwell C hardness tests showed
the following:
We'd Metal
Ran _ AveraPie
24-34
29
The fatigue results (or olate type specimens cannot be compared directly with the results shown for
beam type specimens for the following reasons:
a. The stress cycle for a plate type specimen
is 0 to maximum tension whereas for a beam
type specimen it is a completely reversed
stress cycle. This condition alone should
result in a higher fatigue strength for the
plate type specimen.
Heat Affected Zone
ilFlarnt)
Base Metal
!.•:_
Average
Range Av era
28-41
36
17-21
b. The fatigue cracks in a plate specimen must
traverse a larger section and as a result the
specimen will probably show a longer life.
19
The particular side of the tee weld where failure
originated did -ot appear to be influenced by the variations of range in fillet contour or welder technique
associated with fabrication. Each fracture originated
at the toe of the weld.
c. When failure starts at any point, the stress
level in the undamaged portion of the plate is
not raised as high on the plate type specimen
as on the beam type specimen.
Failure of the plate type specimens start long before ultimate failure occurs. Figure 29. which indicates the total length of crack on one side of the
web corresponding to various deflections. shows
that the cracks are quite extensive much before ultiniate failure occurs. It appears, therefore, that
some criterion of fatigue failure for the plate type
specimeli is required, before a fatigue strength for
a given life can be established and before a comparison can be made with the fatigue results on beam
type specimens.
Figure 23 also indicates the results of the cornparative tests conducted with HTS assemblies. As
indicated therein, at the levels investigated, the resuits for HY-80 and HTS were similar,
Fillet Welds - Plate Teats - 0-TenMs.i
Figure 28 shows an S-N curve for plate type
specimens using the 10 percent increase in deflection criterion of failure. The curve was fitted by
the method of least squares. The 95 percent confidence envelope is also shown. On the basis of
these results, the fatigue strength of the fillet welds
is 60,000 psi for a life of 10,000 cycles. This result should be taken with caution since the life represents almost complete failure of the plate and
start of failure actually occurs much earlier.
Based on the data therein, the 10,000 and 100,000
cycle lives of these assemblies approximate
60,000 psi and 30,000 psi, respectively.
The curves of deflection versus number of cycles.
similar to Figure 29 provide a means for establishing the fatigue life under a given stress. Eachcurve
rises very slowly at first during initiation and propagatlon of a crack, then rises more rapidly as the
crack becomes more extensive, and finally rises
very rapidly as impending ultimate failure is approached. The complete curve to failure represents
an increase in deflection of approximately 100 percent over the initial deflection. It may be observed
from Figure 29 where the numbers on the curve indicate length of crack, that an increase in deflection
of 10 percent corresponds to a length of crack of over
16 inches or 50 percent of the full width of specimen.
The increase of deflection, which reflects the depth
of crack as well as length, is considered to be a
better measure of fatigue failure than size of crack
which is not completely visible, especially the depth.
Although an increase in deflection of 10 percent
In general failures started in the heat affected
zone away from the edges of the plate. In many
cases, cracks opened at more than one location and
progressed until they met or reached the edge.
A typical deflection record for a plate subjected
to cyclical stress Is shown in Figure 29 on a reduced
abscissa scale. This curve shows a rapid increase
in deflection as ultimate failure is reached. The
numbers along the deflection curve indicate the total
72
:
appears to be a reasonable basis for discussion of
fatigue failure, other criteria based on different degrees of deflection or on length of crack might also
be used.
The combination of grinding and shot peening
generally yielded the greatest improvement. In some
cases failure occurred away from the weld in the
HY--80 steel. Figure 33 illustrates a failure of HY80 base plate neat the area of curvature of the speclinen. The limitations of this procedure are:
Fillet Welds - Comrison of Effects of Grinding
and Shot Peening - Beam Type Specimens
a. Its relatively high cost.
Results of beam tests comparing as welded,
ground, shot peened and the combination of ground
and shot peened fillet welds are shown in Figure 30.
Figures 31 and 32 indicate sample macrosections of
the fatigue failures. Since study of most oi the
failures has not as yet been accomplished, the photographs are being offered for information purposes
without comment at this time.
b. The question of permanency of the full effect.
c. The limitation of knowledge ini respect to the
effects of peening on properties other than
fastigue.
The limitation as to permanency of effect is not as
critical as that in the case of welds which are only
peened. since we may reasonably expect that at the
wir>;t. ri ult; w,)uId revert axick to the as ground
condition.
A review of the summary data of Figure 30 in the
light of the information contained herein w-,i pcrmit
us to draw various conclusions and implications,
Considering the lines representing the as welded
condition. we see that weld quality is reproducible.
This was achieved through close control of the welding operation. In view of the above, it is reasonable
to assume that the welds under discussion represent
a quality which is better than the average obtained
in the field. Any deficiencies, such as undercut.
porosity at the fillet toe, overheating if the heat
affected zone. etc. which aIre conditions occasionally
encountered in the field, could very likely havc procduced poorer results.
Results of tht, comnbined grinding and shot peening
confirm one important fact. Since the first fatigue
results on fillet welds in IIY-80 were obtained, their
rtelatively poor fatigue properties has raised a question ais to whethher incipient submicroscopic discontinuit.i u wert, prt-sent throughout the welds, prior to
tsst. The fact !ha! the fatigue life can be improved
to the txtent noted, indicates that the welds were
sound.
Butt Neid& - Beain ,id
The improvement as a result of grinding is significant and readily explainable on the biasts of the
reduction of the acuity of notch effect at the toe of
the fillet, with some possible added benefit derived
from the removal of weld metal containing locked Ut)
stresses. It is reasonable to assume that the benefits are permanent and predictable.
Plate Type Tests
Results ,f tests of bea:im and plate type butt weld
specimens have Ibuen included in Figutires 23 and 28.
As indicated therein, the lives of these welds were
tuch longer than the comnipirable tee fillet welds.
Some if the iwam specimens tested to date failed
through the welds. IL-diographs illustrating the extent and type of weld cracking encountered are shown
in Figure 34. All failures of plate type specimens
observed to dcate. iccurred in the heat affected zone
adjacent to the weld. Work on this phase is currently
underway and a more complete study of the butt weld
fatigue fractures will be made.
The results of shot peening indicate an improvement over the ground condition. However, in assessing shot peening. the process should be con,sidered in relation to the overall application. ýhot
peening represents a surface effect of imposed cornpressive stress. Related work on other steels has
shown that upon removal of the surface layer of
"peened material (such as may occur as a result of
corrosion), the beneficiating effects are lost and
fatigue properties revert to the original condition.
In addition, the effects of shot peening on some of
the other properties which formed a basis for selection of HY-80 steel,- have not as yet been assessed.
The preceding statements are not intended to imply
that shot peening. should be ruled out as a method of
improving the fatigue properties of weldments in
HY-80 steel. However, the facts are presented so
that consideration of the method should be based on
its limitations as well as its apparent advantages. -
Compression Tests - Beam and Plate T2Rpe Tests
Some results of fatigue tests on beam type specimens in which the stress was always compression are
tabulated below. The mean initial compressive
stress was applied in setting up the specimen by
forcing a shim between the specimen end and specimen holder. The holder consisted of a large base
plate in which a recess was cut for the specimen
which was held in place by steel straps which permitted the imposition of additional initial compression. Results obtained to date are indicated below:
"73
Initial Mean
Stress, psi
Applied Alternating
Stress, psi
Stress
Range,_psi
Cycles to
Ultimate
Failure
Cycles to
First Indication
of Failure
20, OOc
20,400
300T 40,500c
2,565,000
29,200c
28,500
700c 57,700c
*
36, 700c
33.700
3000c 70,400
*First indication not observed; cut-off switch stopped machine at 790,000 cycles.
3,005,200
790,000
354,000
During the
course of these tests the initial oompression relaxed from 10 to 15 per cent.
Three results have been obtained un plate type specimens in which the fillet weld was always in compression. For these tests the specimen was mounted with the web on the loaded side of the plate. The full
thickness of plate, 1-1 2 inches, was tested with results, as follows:
Cycles to First $ 0 *
Indication of Failure
Stress Range
0-70,000 c
0-70, 00 c
0-60,00j c
*There was no clear indication of
Cycles to
Ultimate Failure*
485010
8180
2000
4660
4000
ultimate failure; crack progressed along full
length of fillet.
"-Failure was
noted at this poi:it but may have started earlier.
'*Asevidenced by ultrasonic method.
DISCUSSION
b. Fillet WVelds - Plate Type - to 10% Deflection
0 to max. tension
It should be noted that the results obtained are
readily explainable on the basis of the mechanical
and metallurgical characteristics of HY-80 weldments as well as the geometrical effects involved.
-
10,000 cycle level - 60,000 psi
100,000 cycle level - 30. 000 psi
c.
Fillet Welds - Beam Type Tests - Effects of
The improvement effected by grinding can be attributed to a reduction of the geometrical notch
acuity; the improvement by peening, the superimposition of a residual compressive stress in the critical area.
grinding and shot peening - Alternating Stress
Both grinding and shot peening improved fatigue
properties, grinding being somewhat less effective
than shot peening. The combination of grinding and
shot peening was the most effective. However, shot
The better fatigue life of butt welds as compared
to fillets can be attributed to the reduction of the
acuity of the geometrical notch; in addition there is
a possibility that the degree of locked in stress at
the toe of the butts may have been significantly less
than at the toe of the fillets,
peening represents a surface effect which could be
destroyed by removal of the surface layers of metal.
d. Fillet Welds - Beam Type Tests - Comparison
HTS and HY-80 - Results obtained with HTS
assemblies at 37, 100 and 32, 000 psi were
similar to these obtained with comparable
welds in HY-80 base metal.
e. Butt Welds - On the basis of both plate and
beam type specimens, the fatigue properties
direrbutt eldsloadditrnsv the
SUMMARY
for the various
In general, the results obtained
butt welds loaded transverse to the direcad
nvetigtedwer
asseblis
seciens
assemblies and specimens
investigated
were asof
astonfwedgaparbtrthncm
rbl
tion of welding appear better than comparable
follows:
fillet assemblies.
a. Fillet Welds - Beam Type Tests - To Total
Failure - Alternating Stress
f. Compression - Limited results obtained with
the beam type machine indicate that longer
life may be expected when the stress range is
all compression. Results on plate type specimens also indicate that the life under compression is longer. However, results are too few
to draw conclusions at this time.
10,000 cycle level - 52, 000 psi
100,000 cycle level - 20,000 psi
1,000,000 cycle level - 11,000 psi
74
In addition to the above, the beam type apparatus
is being redesigned to accommodate full thickness
(1-1,2") specimens. The modifications will permit
imposing any desired proportion of stress ranging
from complete compression to complete tension.
FUTURE WORK
It is apparent that there are many facets of the
overall problem of evaluating the effects of welding
on the fatigue characteristics of high strength steels
of the HY-80 type. Some aspects relative to fatigue
properties of HY-80 weldments which are currently
being investigated are:
ACKNOWLEDGEMENT
The authors wish to acknowledge the advice and
effort of Mr. T.J. Griffin of the Bureau of Ships
who initiated the program and was present during
the overall planning and the participation of their
many colleagues in the Material Laboratory, especially Cdr. R. Hall, Project Officer, Messrs.
E.A. Imbembo, Metallurgy Branch Head, H.
Nagler, V.A. Di Gighlo and K. Pon, Welding
Engineers, and R. Wolfe, P. Abramov and E.
Lewis Mechanical Engineers.
a. Compression loading
b. Welds containing imperfections commonly
encountered in the field (undercut, embrittlement, lack of fusion)
c. Repair welds in HY-80 plate
*
d. Comparison of welded plate with base metal
properties.
Table 1
Mechanical and Chemical Requirements - Rich Chemistry HY-80 Steel Plate
(MI L-S- 16216D)
Chemical Analysis
0.23 (Max.)
0.10-0.40
0.35 (Max.)
0.040 (Max.)
0.15-0.35
2.50-3.25
1.35-1.85
0.30-0.60
Carbon
Manganese
Phosphorus
Sullur
Silicon
Nickel
Chromium
Molybdenum
Mechanical Propert ies
Yield Strength (psi)
(0.2N Offset)
80,000-95,000
Tensile Strength (psi)
Not specified*
Elongation (cf)
2" Gage
20 (min.)
Charpy V Notch Impact
50 (Min. Av.) - 1-1, 2" thick
(-120°F) ft. tbs.
30 (Min. Av.) - 1-3/4" thick
Tensile strength of plates used ranged from 99, 400 psi to 105,800 psi.
75
Table 2
TYPICAL PROCEDURE FOR WELDING OF HY-80 PLATES
1. All surfaces in area of weld were ground prior to w,'lding.
2. Electrode - Code A Brand - Type MIL-110-18
a. 5 32" diameter was used for the first three layers on each side;
3'16" diameter for remaining layers.
b. Electrodes were conditioned prior to welding by aiking at 800°F for 2 hours and then
stored at 200*F-300*F until used.
3.
a. Preheat temperature 150'F
b. Interpass Temperature 200 F
4.
a. String bead technique.
b. After the optimum operating current and voltage was determined, welding speed was set
to maintain heat input below 50,000 joules per inch. Typical conditions used were as
follows:
Electrode
Dia.
Volts
(Av.)
Amperes
(Av.)
Polarity
Range'
5, 32"
23
165
D.C. (Rev.)
31,000
-
48,200
3,16"
24
185
D.C. (Rev.)
25,900
-
45,900
Joules (In.)
Joules in. were calculated for each assembly as shown below:
Joules
In. -
V x AxT
-
V AT In -
Volts
Amperes
Time (sec.) for welding one side
Length of weld (In.)
Number of beads on one side
c.
Tempering the Heat Affected Zone - The first or outside beads of each
layer of weld and particularly of the outside or last layer of weld metal
were deposited against and connecting to the base plate with the following
beads deposited in such a manner as to overlap the first bead without
touching the base plate.
5. Inspection - After the third layer was completed on each side with the 5/32" size electrode, the
welds were inspected for cracks by magnetic particle iispection while the assembly was maintained at 150- -0°F. All assemblies were examined by magnetic particle inspection after completion.
76
I
--
I-V2"
I"R
66"
12"
.
I8
1I0
Figure I
-
Large Scale Beam Type Fatigue Specimen
77
ii
78
g~c~
-F
2,J
S~.,0
U
t
1o
.q
S"
!
2
[
U
°
r2
•
•
t•
.
C,.
78-
N
\8
4,a
IL
ci..
Ca.
!
/
,'
/
/i
/
/
/
79
,/
,
4,.
AIM
AM
Figure 4
-
Hydrostatic Loading Equipment
so
CYt
w
p
00
1010-
ft)ICz
el~
iU
Figure 6 - Final Procedure for Welding Plate Fatigue Assembly
DIRECTION OF FINAL ROLL
DIRECTION OF WELDING AFTER IST THREE LAYER•S
Al
B
6
8
ICI
7
4'4504
E
8- ---
-- +3+3----
6
DIRECTION OF FINAL ROLL
DIRECTION OF WELDING AFTER IST THREE LAYERS
--
6 -.- I--5j---,42
SI
S2
-6-.--421--
S3
5- -41- 1
S_4
6-A
Notes to Figure 6
Entire top surface of plate should be ground flat prior to welding and machining.
Locate spacers (Si to S4) as shown. Spacers should be of hardened steel.
Deposit the 1st three layers in block sequence as shown using a 5/32" diameter electrode.
Retain spacers in position until the respective block is to be welded. Remove prior to welding.
Complete as shown with 3/16" dia. electrode using continuous welds.
Examine welds by magnetic particle inspection after the third and final layers.
"7.See Table 2, Notes 2,3,5 and 7.
1.
2.
3.
4.
5.
6.
82
Figure 7 - Typical Procedure For Welding Of HY-80 Plates
Joint Design and Typical Pass Sequence - Double-V Butt Joint
600
The last two layers on each side were w~elded in flat position, all other
layers were welded in horizontal position.
83
Figure 8 -
Final Procedure For Welding Butt-Joint Plate Fatigue Assembly
7-
31"e FILLET WELDS
44
43
___
__
__
__
___
_1__
___
DIRECTION OF FINAL ROLL AIN
_
___
4"
41'
44!
GRADE M RESTRAINING BAR
Notes
0
1. The assembly was properly aligned and 3/4" fillet welds were deposited as shown.
2. The first 4 layers were welded in the horizontal position with 5,/32" diameter electrodes, in block
sequence as follows:
a. Sections I and 2 of both sides of the double-V butt joint were welded simultaneously until the
completion of the 4th layer.
b. The roots on both sides were chipped and inspected for cracks by magnetic particle inspection.
c. Sections 3 and 4 were welded similarly to Sections I and 2.
3. The 5th and 6th layers were welded in the horizontal position using 3/16" diameter electrodes, simultaneously on both sides, in straight progression and reversed layers.
4. The 7th and 8th layers were welded in the flat position in straight progression and reversed layers.
84
t.
•
,,!
( DNmO'llv ino lIim 1)
ON1013M O0 NO.13MIA/
"o"o
IVI .4NIl
8NIl3
z
NO±~A
LM
IDIRECION
Dt1
La
85
0i
Iz
WLING
WI4 IC
1Oii
:1
--
99 -
99
Xd064A
*
Oae.i
D9
X V6.4
NVi
:
4
_._
1.
Nao~i.4a
N9301.4
d
NXOIJi
NGLA
L
9
x a 14
x 311T1
_
0.w
.49iC
_________
____
8VI
aZAd1
--
w
:1
d 86 A
XVelid
I-
4U
0
z
-m
2
w
9
8
UlX0
-
-w--
F-~~XGI
-
XSIG
LXV~X~bi
a
d rL V---2_
I-i,
Na aj
>
dZl
___
N66 Z
Za
)(W 0
w
5 c Z
V
-
'.
o
04
00
jI-
r-
z
U)-)
w o
0u.
MovQowwo
w
w
o
o
acZ4v
.0~~
-:)
_j
0
Z
00
W
00
LJ
00
0:00
am
Apr-
Figure 12
-
Peened Specimen
88
WA
0
cm
0
-IF
I
acD
o
OC/O
[*iA
W4.
;.S
U.
89
2
w-
w
w.
z
0,
U
C.
W)
Z
U(AOpTmJU
Figure 14
-
Deformation Of Surface After Shot Peening
90
04*
911
Figure 16 - Photomicrographs of Peened and
Unpeened Areas HY-80 Base Metal - 25OX
Figure 17 -Photomicrographs of Peened and
Unpeened Areas Heat Affected Zones - 250X
92
mP
vk
IFI
Figure 18 - Photomicrog~raphs of Peened and
Unpeened Type RGIL- 100- 18 Weld Depositsi- 250X
93
4'..
Figure 19 - Ground Specimens
94
Figure 20
-
Ground and Peened Specimens
95
LLIl.w
(D'
o-'W~
im
OCO
11 jr
0ý X
gZ
03
a
UU
N0
ICL
00
--
11.1
0
U
974
-
coJ
0
0
I--
0
*U).
0
0
2
0
N
00
T
U).
.
8
i
-
0
-
CU
Cd
93
Fracture patlis in
101gitUudiii
SCCjl)IrS
particle indidcations..
VFigý,rv 25
-T
rpic.-ilI Pahs Of Fr~ac~tures After Dewm Fatiguie
Best Availabl Cp
TdswtE-
Macrospecimen (LB4MDP)
(Magn ';X)
wu
SReduced,
'
Magn 100X
6X
Itchant: Vilella's Etch
Figure 26 - Microstructure of Area Near Toe of Weld Double Beveled Tee Joint,
Fillet Reinforced - 60# HY 80 Plate
101
I o'
<
Macrospecirnen (LB4DP 2)
(Magn 5X)
OReduced .65X)
Etchant: Vilella's Etch
Figure 27
-
Photomicrograph of Fatigue Fracture
102
-
As Welded
100
-J
2.
-
2
0..
0
3
0..
-.
,
w
WQ
00u
'
39<
0.
I-
0-
0
In
0
v
(ISd) SS3M.LS IVNIV40N
103
a
0o
0z
!,
NUMBERS ON DEFLECTION CURVE REPRESENT LENGTH OF CRACK IN INCHES
.4'
.5
.0
0
4,000
8,000
12,000
16,000
20,000
24,000
NUMBER OF CYCLES
Figure 29 - Curves Showing Progressive Increase In Deflection As Crack Propagated Through
Critical Section Of Plate Type Specimen LH-7 Stress Cycle 0-56,600 PSI
104
7--
0
0
27
J
4-
.1
*
105A
F0
91
0
F7.4 P
Figure 31
-Macroaections
o( Unground Fatigue Fractures
106
-
As Welded and Shot Peened
Figure 32 - Macrosections of Ground Fatigue Fracture
107
I
Figure 33 - Ground and Shot Peened Section - Fracture in BY-80 Steel
108
*
1O
-;
Figure 34 - Radiographs Of Butt Welds After Beam Fatigue Tests
109
REPORT OFV SEMI-AUTOMATIC INERT
GAS METAL ARC WELDING
HY-S0 STEEL OUT-OF-POSITION
by
L. ROBBINS, MARE ISLAND NAVAL SHIPYARD
I.
ABSTRACT
Numerous techniques for out-of-position welding
with the Inert Gas Metallic Welding Process have been
proposed and evaluated in the past for application to
HY-80 submarine structures. Early work was confined
to use of small diameter wire 0.030 to 0.035-rinch diameter, using an extremely short arc, commonly
called dip transfer or short arc technique. ArgonCarbon Dioxide gas mixtures were used with the process. The efforts to date, have been unsuccessful, due
to the inability of the welded joints to produce adequate
toughness under explosion bulge testing. Cold shuts
were believed to be the principle cause of failure. This
investigation involved the evaluation of the same process using larger diameter wire to overcome the propensity for depositing welds with cold shuts.
Three conditions or welding procedures, as outlined below, were selected for evaluation on one-inch
thick HY-80 plating:
Condition No. I - Four beads to complete the
weld using 1/16-inch diameter electrode.
dition No. I to evaluate the combined performance
of weld, heat affected and fusion zone and base metal.
Tensile and bend tests were satisfactory. Drop
weight and charpy "V" notch tests of the weld metal
were also satisfactory, although somewhat lower
than normally expected. Explosion bulge tests revealed unsatisfactory performance of the joint. Brittle HAZ failure was experienced with most of the
plates when tested at 0°F.
II.
INTRODUCTION
The object of this investigation is to evaluate
new procedures and techniques developed by Air Reduction Company for semi-automatic inert gas (Metallic Inert Gas) welding processes for making welds on
HY-80 steel in the vertical position.
Welding was accomplished working in conjunction with Air Reduction Company representatives and
welding techniques employed were those recommended
by Air Reduction Company.
Condition No. 2 - Four beads to complete the
weld using 1/16-inch diameter electrode in the root
and 0.045-inch diameter electrode for the cover
passes.
Condition No. 3 - Six passes to complete the
weld using 1/16-inch diameter electrode in the root
and 0.045-inch diameter electrode for the remain- •
Ing passes.
One butt welded plate was prepared using each
condition described above. These plates were examined by radiography and then dissected and prepared into drop weight test specimens. Nil-ductility
transition temperature of the deposited weld metal
resulting from each condition was determined by
drop weight testing,
Condition No. I which produced the most desirable nil-ductility transition was used in preparing a butt welded joint for transverse full thickness,
reduced section tensile tests and charpy "V" notch
impact tests.
This novel technique was intended to overcome
the propensity for depositing welds with small cold
shuts, experienced with the small diameter electrode.
The evaluation was to include a determination of the
soundness of the weld by radiographic examination
and the physical conditions of the welded joint as revealed by the various methods of destructive testing.
Side bend tests were to be made along with tensile
tests to determine the adequacy of the welded joints
to meet strength and ductility requirements for submarine applications. Charpy "V" notch impact
properties would be required for the as-deposited
weld metal for evaluation of notch toughness. The
weld metal would also be evaluated for the transition
temperature from ductile to brittle behavior in the
presence of a notch by drop weight tests. The final
phase of the investigation would employ the explosion
bulge and explosion bulge crack starter tests to evaluate the combined performance of the weld metal,
fusion and heat affected zone and base metal.
MI. MATERIALS
The HY-80 steel plate for this investigation was
Evaluation of the results of the tests described
above indicated that the welding procedure, Condition No. 1, was adaptable to submarine fabrication
and producedBY-S0
weldg in
plating that were within
the general Bureau of Ships requirements.
obtained from Shipyard stock and represented standard plating for submarine construction conforming
to the requirements of Military Specification MILS-16216. A chemical analysis was made on each
heat for the plates used for all tests and found to
conform to the specification requirements. Table
Six butt welded plates suitable for explosion testing were prepared using welding procedure Con-
I includes the analysis of the plates used for these
tests.
110
I'
TABL
-$ t
pits.
ELEzMENT
PERCENT COMPOSITION
Carbon
Manganese
Phosphorus
Sulfur
Silicon
*
Chromium
Nickel
Molybdenum
0.13
0.17
0.008
0.024
0.19
0.22
1.23
2.20
0.30
1.11
2.07
0.34
Welding Torch - Aircomatic Model AH35-A Pull
Gun and Controls.
Cooling
Torch Gas Nozzle - 5/8-inch diameter.
_Argon Gas Regulator - Two stage with dual
range flow meter.
Carbon Dioxide Gas Regulator - Single stage
with d
rane flow meter.
0.07
1.33
-,0.005
0.012
0.52
1.31
0.17
Gas Mixin - Shielding gas mixing was accomplished by a "Y" connection in the line.
Bevel Preparation - Bevel preparation was accomplished by machine planing and planed surfaces
were cleaned prior to welding by vacu-blasting with
steel g, it.
Carbon Dioxide gas welding grade was used for
all tests. Argon gas was obtained from standard
An analysis confirmed conformity to the re-
stock.
Airco water circulator.
Torch Contact Tubes - 1/16-inch and 0.045-inch
standard four-inch long.
TABLE 2
PERCENT CONTENT
Carbon
Manganese
Phosphorus
Sulfur
Silicon
Nickel
Vanadium
Plates B-i. 8-2, J-1 & C-2
PERCENT COMPOSITION
0.15
0.20
0.011
0.011
The electrode used for the tests was obtained
from standard stock and conformed to the requirements of Military Specification MIL-E- 19822.
the electrode is
The chemical analysis for
shown in Table 2.
ELEMENT
IA
quirements of Military Specification MIL-A- 18455.
Tack Welding - Tack welds were deposited using
a bridge technique and were ground out during welding so as not to be incorporated in the final weld.
Plates were preheated to 150°F minimum prior to
TV. PROCEDURE
Preliminary Test~in and Test Results.
A number of practice butt joints were welded
to familiarize the Mare Island welders with the Air
Reduction Company recommended welding techniques
the arc characteristics of the 1/16-inch diameter
and
electrode.
tack welding and tacks were approximately one-inch
long and spaced on approximately six-inch centers.
Preheating - All preheating was accomplished
using
oxygen-acetylene torch with a multi-flame
ut-lm
sn annoye-ctlnetrhwt
heating tip.
The equipment and data listed below was common to all welds made in this test program.
Temperature Control - Preheat and interpass
temperature determination was accomplished by use
Welding Power Source - 450 Amp Aircomatic
Fillerarc full wave rectifier conditions:
of surface pyrometers and temperature indicating
crayons.
Physical Testing - All physical tests reported
herein were accomplished on plates in the as-welded
condition. No stress relieving was done.
1. Reactor Position - Low
2.
Hot Start - On
3.
Rise Selector - Off
4. Arc Length Setting
5.
-
Backgrinding - The back side or root of the first
pass was background to sound metal prior todepositing the first pass on the second side of the joint.
3-4
Current - D.C.R.P.
III
V.
RESULTS OF TEST
After sufficient practice three 35-inch long butt
joints were prepared using joint design as shown in
Figure 1 and welding procedures as shown in Figures
2, 3 and 4. On completion of welding the three
welded butt joints were examined by radiography
and met the density requirenaunts of Group 2 of NAVSHIPS 250-692-2. With satisfactory radiographic
results each plate was dissected into nine 3-1, 2" x
14" x 1" drop weight specimens. Results of drop
weight testing are shown on Figures 5. 6 and 7.
Welding procedure, Condition No. 1, had the
lowest nil-ductility transition temperature (see
Figure 5). Therefore, it %%As decided to prepare an
additional plate for physical tests using welding procedure, Condition No. 1. as outlined on Figure 2.
One 18-inch long butt jolt was welded and examined by radiography and met the density requirements of Group 2 of NAVSHIPS 250-692-2. Transverse, full thickness, reduced section tensile specimens were prepared from this plate and test results
were as shown in Table 3. The tensile properties
are transverse to the weld. The reduced cross section for the specilens was one-inch thick by oneone-half-inches wide. The fractures occurred in the
base metal, revealing overmatch of the strength
level bý' the weld. There were no abnormalties in
the specimens and the welds revealed no surface defects or cracks after fracture. The necking down
and fracture was located in the unaffected base metal.
Side bend tests were removed from the same plate
and bent 180 degrees over a 3 4-inch radiused mandrel. Examination of the bent specimens did not reveal any unacceptable flaws. The charpy -V- notch
impact properties of the weld met the 20-foot pound
minimnum criterion at -60 F. Based on these results,
plates of suitable size were prepared for explosion
bulge ttests, Two comparison test plates were also
made with the 11018 and B-88 electrodes using standard Shipyard practices. This was considered advisable to assure that explosion bulge test practices
agreed with the standards offered by Naval Research
Laboratory. The conditions of welding vertically for
for three elates prepared for the explosion bulge test
are shown in Table 4.
The preheat and interpass temperatures are within the maximum limits specified for welding HY-80
structures. The heat input as measured by joules
per inch exceeds the 60,000 maximum specified for
Shipyard practices for one-inch low chemistry plate.
The conditions of welding for the three plates
prepared for the explosion bulge crack starter tests
are shown in Table 5. Here again it is noted that the
heatex
heat inputs are high.
Table 6 shows the condition of welding for the
11018 and B-88 electrode welded plates prepared for
explosion bulge tests. The welding conditions for
112
these plates for preheat, interpass temperature and
heat input were within the specified tolerances for
Shipyard practices.
All plates for explosion bulge testing were undercooled to -3 'F and held for one hour at temperature.
Cooling was done in a tank with the plate completely
immersed in the liquid solvent. All the plates were
blasted at O'F, using a four-pound C-3 composition.
The elapsed time between removal of the plates from
the cooling tank- to blast was 30 to 40 seconds.
The photograph of Plate BI shows brittle failure
occurred in the HAZ with a single blast for nearly
the entire length of the weld. This failure is typical
for HAZ degradation caused by improper thermal
quenching.
The photograph of Plate B2 shows the results
of the explosion bulge test with brittle failure occurring in the HAZ on both sides of the weld. Tearing
also occurred in the base plate on either side of the
weld for a short distance. The tearing in the base
plate exhibited shear ductile behavior. Two blasts
were necessary to produce failure.
The photograph of Plate B3 shows failure again
occurring in the HAZ. However, in this case the
failure was located in the area of least strain. The
longest crack was about two-inches and the shorter
crack about 1 4-inch in length and located on either
side of the bulge apex. Two blasts were necessary
to produce the first evidence of failure.
The photograph of Plate C2 shows the results
of three blasts %ith the plate showing no evidence
of failure. In this case, the level of strain as
measured by thickness reduction at the peak of the
bulge was a little over eight percent. This is indicative of the excellent performance of properly welded
HY-80 plating. This is the plate welded with 11018
electrodes.
The photograph of Plate Jl is another example
of excellent performance of welded HY-80 plate.
This plate was also blasted three times with the
weldment showing no failure. The measured level
of strain at the peak of the bulge for this plate was
over nine percent. In comparison, with plates B1
to B3 inclusive, it is significant to note that the
strain level was not nearly as great and in one plate
no measurable strain was observed. Both of the
vertically welded plates, failure occurred at the
measured strain level.
The next series of explosion tests were conductseinputsxposoarest whigh.dct
ed on Plates B4 to B6, inclusive, using the crack
starter for crack initiation.
The photograph of Plate B4 shows the iresults
of crack starter test for the first plate in this
series. Failure occurred to almost the full depth of
the weld and stopped on either side near the edge of
the weld in the base metal.
The photograph of Plate B-5 shows failure again
extended for the full depth of the weld under the notch,
In this case, however, the crack propagated in an ap-
previous techniques for out-of-posi' ion Metallic
Inert Gas Welding has been achieved. The tendency for depositing welds with lack of fusion defects has been overcome with this novel welding
technique.
static strength 6f welded joints
satisfies theThe
requirements for joining HY-80
steel
by this method for the low chemistry plate.
It
HAZ on the opposite side was noted and extended a
fraction of an inch on both sides of the transverse
fracture.
o
be ept
th
a similar res
wldtbe
Degradation of the heat affected zone is very pronounced and is exhibited in most of the plates
tested
by explosion bulging. From the later results it is
concluded that the process will not produce welds
which will satisfy the performance requirements for
submarine applications. However, there appears to
be two avenues open for further investigation. It is
entirely possible that satisfactory HAZ performance
may be obtained in the high chemistry plating due to
the greater hardenability. By preparing two-inch
thick HY-80 plates and using the same series of
tests reported herein, the use of this process for the
high chemistry material can be evaluated by comparison with existing standards. The second proposal, would be to investigate the use of 0.045 to
0. 050- inch diameter wire by endeavoring to reduce
the heat input by use of smaller weld sizes. Preliminarv. but limited work indicates this approach
holds excellent possibilities.
parent weak HAZ. The beginning of a failure inl the
paret wak AZ.The
eginin ofa falur inthe
b hsmto
o h
o
hmsr
lt.I
would be expected that
similar results would be
obandfrtehgcemsyorevirptn.
The photograph of Plate B-6 shows the results
of the last plate in this series of tests. Failure again
occurred almost to the full depth of the weld. The
crack propagated into the base metal on either side
for about 1, 4-inch beyond the toe of the weld. As
pointed out in previous papers by Puzak and Pellen.
superposing a crack-starter weld at the center of the
bulge test sample, the test is made selective to the
plate. The severity of the test conditions for the weld
and HAZ aredecreasedand those of the plate increased.
This factor must be considered in evaluating the performanceof the welded plates with the crack-starter
welds.
Reviewing the results reported herein, it is significant to note that a good deal of improvement over
113
WELDING ENGINEERING BRANCH
MARE ISLAND NAVAL SHIPYARD
JOINT DESIGN
FOR
VERTICAL METALLIC INERT GAS WELDING
USING
ARGON AND CARBON DIOXIDE SHIELDING GAS
60
\
r
. . ..
/
\
r
. . . . -_
.
/
__\__-
L__
. ....
3*.....
60°
FIGURE I
114
1
..
0 to
_
16
_•
_ _ __-.
WELDING ENGINEERING BRANCH
MARE ISLAND NAVAL SHIPYARD
DROP WEIGHT TEST PLATE NO. I
WELDING PROCEDURE
CONDITION NO. 1
3
2
4
PASS SEQUENCE
Base Metal: HY-80 Lukens Ht. No. 22291-4
Welding Electrode: Airco Ht. No. 8X1313
Preheat Temperature: 200'F
Welding Interpass Temperature: 200 F - 300-F
OPERATIONAL DATA
"F
•T
--FI
Energy Input
Electrode - Electrode
Travel
Welding Current
Pass _
..
Speed. : Feed Speed
Diameter IJoules/inch
No.
Amps
j VoltInArc) 1
.
1
chh__
hPM
--- - -:=
97. iL --,
ST
- -- ------1
175
19
1
92"_
1/16
111,000
92
175
19
19
.8
3
20
92
4
175
19
1.85
175i
FIGURE 2
115
1/16
108,000
1/16
108,000
99,500
WELDING ENGINEERING BRANCH
MARE ISLAfND NAVAL SHIPYARD
DROP WEIGHT TEST PLATE NO. 3
WELDING PROCEDURE
CONDITION NO. 2
PASS SEQUENCE
Base Metal:
HY-80 Lukens Ht. No. 22291-4
Welding Electrode: Airco I/ 16-inch Diameter Ht. No. 8X1313,
0.045-inch Diameter Ht. No. X10161
Preheat Temperature:
200°F
Welding Interpass Temperature:
200'F - 300'F
OPERATIONAL DATA
Welding Current
Pass
No.
,1
2
3
L
4-17
Amps
175
175
175
Travel
Speed
Volts(Arc) IIPM
I
Electrode
Feed Speed
Electrode
Diameter
Energy Input
Juules/inch
Inch
1PM
1,/16
I
142,000
19
1.4
94
19
18.5
1.9
2.4
94
200
1,'(1
104,000
82,000
18.5
2 .3
200
045
84,000
FIGURE 3
116
-
WELDING ENGINEERING BRANCH
MARE ISLAND NAVAL SHIPYARD
DROP WEIGHT TEST PLATE NO. 3
WELDING PROCEDURE
CONDITION No. 3
6
4
2
PASS SEQUENCE
Base Metal: HY-80 Lukens Ht. No. 22291-4
Welding Electrode:
Airco I 16-inch Diameter Ht. No. 8X1313,
0.045-inch Diameter fit. No. X10161
Preheat Temperature
200 F
Welding Interpass Temperature:
200 F - 300 F
OPERATIONAL DATA
Pas"
Pass
No.
Welding Current
Amps
Volts(Arc)
Travel
Speed
IPM
Electrode
Feed
Speed
IPM
Electrode T Energy Input
Diameter
Joules/Inch
Inch
1
145
19
2.1
84
1 16
79,000
2
155
18.5
3.3
84
1 16
53,000
3
155
18.5
2.9
84
1 16
59,500
4
155
18.5
3.2
84
1 16
54,000
5
175
18.5
2.5
200
0.045
78,000
6
175
18.5
3.4
200
0.045
82,000
FIGURE 4
117
r
WELDING ENGINEERING BRANCH
MARE ISLAND NAVAL SHIPYARD
DROP WEIGHT TEST RESULTS
PLATE NO. 1
WELDING PROCEDURE
CONDITION NO. I
Drop WeLght Specimen Size: I" x 3-l"2" x 14"
Cooling Bath: Michie S!udge Test Solvent
Coolant: Carbon Dioxide (dry ice) and Liquid Nitrogen
Hammer Drop Height:
13-feet
Impact Load: 1,140 foot pounds
Le~gend
X - Break
0 - No break
Note: Number over -X" or "0" indicates drop No.
TEST TEMPERATURE -'F
-90
-80
-70
-60
5, X
6;X
3/0
1,0
7/0
4/0
2,10
8/X
Nil-Ductility Transition Temperature: -80°F
FIGURE 5
118
-50
WELDING ENGINEERING BRANCH
MARE ISLAND NAVAL SHIPYARD
DROP WEIGHT TEST RESULTS
PLATE NO. 2
WELDING PROCEDURE
CONDITION No. 2
Drop Weight Specimen Size:
Cooling Bath:
Coolant:
Michie Sludge Test Solvent
Carbon Dioxide (dry ice) and Liquid Nitrogen
Hammer Drop Height:
Impact LA.ad:
X-
I" x 3-1 2" x 14"
13-fct
1, 140 foot pounds
Break
0 - No Iweak
Note: Numlbr over -X" or -0- indicates drop No.
TEST TEMPERATURE -' F
-90
-80
-70
1 X
-60
-50
2X
30
5X
40
Nil-Duclilit y Transit ion Temperatu re: -60 F
FIGURE 6
119
WELDING ENGINEERING BRANCH
MARE ISLAND NAVAL SHIPYARD
DROP WEIGHT TEST RESULTS
PLATE NO. 3
WELDING PROCEDbJRE
CONDITION NO. 3
Drop Weight Spucimen Size:
Cooling Math:
I" x 3-1 2" x 14"
Michic' Sludge Test Solvent
Coolant: Carbon Dioxldt' (dry ice) and Liquid Nitrogen
Hammer D)rop Htight:
13-feet
1, 140 foo)t pounds
Impact Lad:
Leg•-nd
X - Brtuak
0 - N
break
Note: Numit.betr over "X" or "0" indicates drop No.
TEST TEMPERATURE -'F
-90
-80
-70
-60
-50
IX
5X
3,0
2X
6 X
4,0
Nil-Ductility Transition Temperatu.,c: -60°F
FIGURE 7
120
WELDING ENGINEERING BRANCH
MARE ISLAND NAVAL SHIPYARD
PHYSICAL PROPERTIES
OF
WELDED BUTT JOINT
TRANSVERSE FULL THICKNESS REDUCED SECTION TENSIL5 TESTS.
(WELD
ORIENTED AT THE MID POINT OF THE REDUCED SECTION
PHYSICAL PROPERTIES
TEST NO.
YIELD STRENGTH (DROP OF BEAM), PSI
ULTIMATE TENSILE STRENGTH, PSI
ELONGATION IN 2-INCH, %
LOCATION OF FRACTURE
SIDE BENDS.
(1800
1E
UL
NO
8ý,200
86,000
98,900
42.5
PARENT METAL
9b OO
45.O
PARENT METAL
OVER A 3/4-INCH RADIUS MANDREL)
SPECIMEN NUMBER
BEND TEST RESULTS
I
2
NO CRACKS
NO CRACKS
NO CRACKS
Nn CRACKS
CHARPY "V" NOTCH IMPACT TEST.
(NOTCHES LOCATED ON THE WELD CENTER
LINE, ORIENTED PARALLEL TO THE FINISHED SURFACE OF THE WELD)
SPECIMEN NUMBER
FRACTURE TEMP.
I
2
3
4_
5
6
7
5
9
OF
ENERGY ABSORBED FOOT POUNDS
+70
62
+10
+30
+30
0
63
57
60
52
-30
-30
_60
5
34
38
31
10
II
-60
-0oo
29
26
12
-100
27
_
TABLE
121
3
SEQUENCE AND OPERATIONAL DATA
BASE METAL PLATING:
AIRCO HEAT NO.
ELECTRODE:
2 0 0 9F - 300OF
INTERPASS TEMPERATURE:
PASS
NO.
8X1313
200OF
PREHEAT TEMPERATURE:
SHIELDING GAS:
20 C.F.H. ARGON,
5 C.F.H.
CARBON DIOXIDE
PLATE B-1
OPERATIONAL DATA
ELECTRODE ENERGYINPUT
TRAVEL ELECTRODE
JOULES/INCH
FEED SPEED DIAMETER
SPEED
WELDING CU RENT
INCH
IPM
VOLTS(ARC I IPM
AMPS
i
2
175
19
1.7
looi0
175
19
2.3
IOO
1/16
1/16
3
175
19
2.1
IOU
r/16
L.
175
19
1.7
O00
016
WELDING CURRENT
AMPS
NO.
TRAVEL
SPEED
I VOLTS(ARCj IPM
IPM
1/06
175
1oO
100
175
175
19
19
2.0
2.2
ioo
100
1/16
1/16
2
3
4
1
2175
3
4
TRAVEL
WELDING CURRENT
VOLTS(ARCj JED
AMPS
94uuu
116,000
ENERGYINPUT
JDULES/INCH
117,000
8
1/16
89o
PLATE B-3
OPERATIONAL DATA
PASS
NO.
7,OOO
INCH
1.7
2.3
175
'000
.
ELECTRODE
ELECTRODE
FEED SPEED DIAMETER
19
19
I
I
PLATE B-2
OPERATIONAL DATA
PASS
NO.
22291-14
ONE-INCH THICK HY-80, LUKENS HEAT NO.
ELECTRODE ELECTRODE
FEJMSPEED DI IjfTER
ENERGYINPU
JDULES/INCC
19
1.71
100
i!/i 6
11,800
1
2,.05
100
0/
96,000
175
19
2.2
1Ob
I/I6
91,o00
175
19
2.1
100
1/16
95,000
175
TABLE 14
122
SEQUENCE AND OPERATIONAL DATA
BASE METAL PLATING:
ELECTRODE:
ONE-INCH THICK HY-80, LUKENS HEAT NO.
AIRCO HEAT NO.
PREHEAT TEMPERATURE:
WELDING CURRENT
jVOLTS(ARC)
. AMPS
3
175
I-5
175
4
175
2
200OF - 300OF
20 C.F.H. ARGON, 5 C.F.H. CARBON DIOXIDE
PLATE B-4
OPERATIONAL DATA
SHIELDING CiS:
1
8X1313
200F
INTERPASS TEMPERATURE:
PASS
NO.
22291-1•
TRAVEL
SPEED
ELECTRODE
FEED SPEED
INCH
IPM
IPM
ENERGYINPUI
JOULES/INCý
ELECTRODE
DIAMETER
19
19
1.5
2.35
2.05
100
100
100
0/16
0/16
1/16
Io000
6,000
97,0
19
1.80
100
0/i6
111,200
"OPERATIONAL DATA
PASS
PS
NO.
WELDING CURRENT
WSPEED
AMPS
VOLTS(ARC)
TRAVEL
PLATE B-5
ELECTRODE
ENERGYIIPUT
FEED SPEED DIAMETER
JOULES/INCh
IPM
INCH
IELECTRODE
IPM
I
175
19
1.7
0oo
0/6
111,800
2
3
175
175
19
19
2.
2.05
100
100
1/16
1/16
100000
97,000
175
19
2.25
1o0
1/16
OPERATIONAL DATA
PASS
S
NO.
WELDING CORRENT
W
N CSPEED
AMPS
VOLTS (ARC)
TRAVEL
IPM
ELECTRODE
FEED SPEED
IPM
PLATE
s-6
ELECTRODE
DIAMETER
INCH
ENERGYINPU
JOULES/INC
-I
1
175
19
1.89
O00
1/16
103,500
2
175
19
2.49
100
1/16
81,000
17
9
1/16
92,000
175
19
2.15
1.82
100
4
0oo
il/6
14,0eo
TABLE
123
5
PROCESS:
METALLIC
INERT GAS AUTOMATIC
BASE METAL PLATING:
ELECTRODE:
ONE-INCH THICK HY-80,
AIRCO HEAT NO.
LUKENS HEAT NO.
23546-63
8XI312
0
PREHEAT TEMPERATURE: * 15OF
INTERPASS TEMPERATURE:
SHIELDING GAS:
200OF
50 C.F.H. ARGON,
2% OXYGEN
OPERATIONAL DATA
PASS
NO.
WELDING CURRENT
i-6
34o
AMPS
PROCESS:
VOLTS(ARC)
PLATE NO.
ELECTROOE
DIAMETER
INCH
ENERGY INPUT
JOULES/INCH
1/16
50,000
30
MANUAL ARC
BASE METAL PLATING:
ELECTRODE:
ONE-INCH THICK HY-80,
LUKENS HEAT NO. 23253-6
MIL-IIO8
PREHEAT TEMPERATURE:
15•OF
INTERPASS TEMPERATURE:
300OF MAXIMUM
OPERATIONAL DATA
PASS
WELDING CURRENT
NO.
AMPS
1-14
J-I
I5-195
PLATE NO.
ELECTRODE
DIAMETER
C-2
ENERGY
INPUT
VOLTS(ARC)
INCH
JOULES/INCH
22-I 2
5/ 2
50,000
TABLE 6
124
-
60,000
Owr
t..
......
Equi)1-L'ioi ni BIlugft
ac
M 1.I.IG. Pu Il Gu n V
Ilaitc N,
H- I ont, Shot.
t. PoIN.
125
Falure
062- Dia.138
Figure 9 -Explosion
Bulge Plate No. B-2 Two Shot, Failure .062' Dia.
M.1. G. Pull Gun Vert. Pos.
126
-B-88
Airco
AI
E
Q
ULG
PLATE O 13-
Figure 10 - Explosion Bulge Plate No. B-3 Two Shots. Failure .062 Dia. B-88 Airco
M.I.G. Pull Gun Vertical Pos.
127
Figure 11
Flat Pos.
-Explosion
Bulge Plate No. C-2 Three Shot,
128
-
No Failure 5/32" Dia. bOL-l1O-l8
44
Figure 12
-Explosion
Bulge Plate No. -J.- 1 Three Shot, No Failure .062" Dia . B-88 M.1. G.
'129
.I
VE Rr k.
Pos
130
A
Figure 15
Vert. Pos.
*b
-
Crackstarter Plate No. B-5 One Shot, .062" Dia. B-88 Airco M.I.G. Pull Gun
131
Al"*
V'4.
4-i
Figure 16
Vert. Pos.
-Crackstarter
PlateŽ No. B-6 One Shot, .062" Dia. B-88 Airco M. I.G. Pull Gun
132
THE APPLICATION OF PREHEAT TO SUBMARINE CONSTRUCTION
by
F. Daly,
Newport News Shipbuilding I Drydock Co.
The need for preheat in the welding of alloy ateels
is a well recognireit fact, familiar to all fabricators
who have dealt with such materials in the pressure
vessel or Naval marine construction fields.
The heating equipment which appears to meet these
requirements for the greatest variety of job conditions
encountered in HY-80 submarine hull construction is
'the electrical ipsistance strip heater.
The level of preheat required for individual applications of the various alloy steels is a matter upon
which agreement is much less general. This is particularly true of the quenched and tempered steels
which are usually employed in the as-welded condition and thus are highly dependent upon proper therreal treatment during welding for the degree to which
their original toughness remains after welding.
This heater is used in a number of forms by the
various building activities.
In the application of preheat to submarine construction using flY-80 steel, we are faced with a very
basic problem. Stated simply it is this' We must
preheat to a level sufficiently high to avoid cracking
without exceeding an empirically established maximum preheat value beyond which the toughness of the
HY-80 steel will be impaired.
At the outset, it should boe understood that this
problem is surmountable although it is difficult to
live with.
The nature of the problem may best be illustrated
by stating that the generally accepted minimum preheat temperature for welding HY-80 steel, over Ithick with MIL-11018 electrodes, is 200 F. The maximum allowable preheat temperature is 300- F. Since
the maximum allowable interpass temperature is also
300c F, it is evident that the temperature rise due to
welding must be carefully watched to avoid encroaching on the maximum interpass temperature requirement when the preheat temperature is at or much in
excess of 200' F. To clarify the concept of interpass
temperature, it may be defined as: The temperature
of the previously deposited weld metal at the point
where a succeeding weld bead will start. With this
background it becumes apparent that the fabricator is
working within narrow limits indeed.
Admittedly
the lower limit, that of preheat, is one he has chosen
but his choice has been forced upon him by the requirements of the material with which he isworking.
Experience has shown us that, not only must we
maintain a certain level of , "eheatduring welding,
but that the preheat temperature must be uniform
and stable during tlhe entire welding operation. Thus,
care must be exercised in the choice of the heating
method, to insure that ample energy is furnished in
a readily controllable form, and further, controls
must be instituted to insure that the chosen method
is properly applied and policed during the welding
operation to assure that the desired temperature is
being realized continuously,
The method of attaching strip heaters to the work
is a much discussed matter. In the interest of avoiding the potentially harmful effects of attaching the
heaters by welding, as is usually done on other
classes of work, submarine building activities have
used a number of devices aimed at avoiding welded
attachments. The most successful of these devices
is the cemented stud, since, unlike wire cables,
slings and various types of brackets, it can be used
on any weldment configuration.
Newport News method uses a percussive or capacitor discharge stud weld for heater attachment.
This method requires grinding of the weld site, to a
depth of about .015" to remove all base material affected by the weld.
The photographs, figures 1-5, show various applications of one type of heater. In each instance the
elements are 1000 watt capacity and the plating 2"or
more thick.
While strip heaters are the most widely accepted
and probably the least fallible of the heating methods
employed in submarine construction, it is difficult to
conceive of an entire hull being fabricated without
some use of torch heating.
Torch heating should generally be confined to tackwelding operations or to those applications involving
the welding of small items within a limited area.
133
oA
From an abstract viewpoint, all types of ,trip
heater must be considered equally acceptable if they
satisfy the criterion of providing the minimum required preheat while not creating temperatures above
300'F; regardless of heater type chosen, the fabricator must make a decision as to the watt density
desired in a heater. For greatest flexibility, the
watt density of each heater should be such that it is
capable of satisfying the previously mentioned criterion when the heater is applied in single rows on
the thinnest material normally heated by this method.
On heavier material or under more severe ambient
temperature conditions, auplication of heater units
will satisfy the need for increased heat. Alternatives to this method of determining strip heater type
and capacity are: To stock heaters in a variety of
watt densities or to furnish a variable power supply
for the heater system.
When torch heating is used, care should be taken that
a generous area of the surrounding base material, is
slowly brought up to preheat temperature with sufficient time allowed for the heat to soak through the
thickness of the parts being welded.
lowering of preheat temperature (to favor
distortion or stress control.) The weld
overlay itself will be deposited prior to
joint welding using a high preheat. Use of
the overlay will offer assurance of excellent
base material heat-affected zone properties
and will virtually assure freedom from cracking upon subsequent joint welding while allowing the use of greatly reduced preheat for
the joint welding.
Exceptions to the above will be:
(1) Those instances in which torch heating is
used as an accessory device to decrease the
time required for reaching preheat temperature on material which is being heated with
strip heaters.
(2)
The following statements pertain to the application of preheat in general, but are particularly pertinent in the case of HY-80 submarine hull construction:
Those instances in which an element of a
weld joint provides insufficient heat sink capacity
theincrease
use of strip
heaters
becausetoofwarrant
the rapid
of interpass
wecau fheaich
wilocure
whnwelg
o
temperature which willoccur when welding
is initiated. Examplesof the latter are face
plates or coamings on lightening and access
openings, flanges on light stiffening members, etc. In these latter instances the use
of torch heating will be preferred to strip
heating,
I.' Responsibility for control of preheat, including assuranceythat •,t has reached the specified level, continues at this level throughout
the welding operation and does not exceed
the maximum allowable temperature, must
be charged toa specified groupof production
or inspection personnel. Temperature idicating crayons are ecrnsidered an adequate
tool for this purpose.
The application of preheat to weldments of cornplex shape and those containing stiffeners, bulkheads
or other plate elements in more than one plane or direction must be carefully controlled to avoid distortion of the structure or the creation of transient
stresses due to non-uniform expansion caused by
preheat. Some methods of controlling these effects
are:
(a) Preheat and weld only widely separated elements simultaneously within a single weldment.
(b)
Preheat and weld intercostal elements on a
random basis prior to preheating and welding through elements in a given area.
(c)
When possible, apply preheat such that the
completed weld will be in compression. Examplcs: When welding internal shell frames
to the shell, the preheat temperature of the
shell plating should exceed that of the frame
web; when welding external frames to the
shell the converse of the above will apply.
When welding inserts into hull structure the
temperature of the hull plating should be
above that of the insert plate, tube or weldment.
2.
Provision should be made to protect all heated assemblies from inclement weather.
3.
It is highly advisable to plan the welding operation in such a manner that all elements of
welding will proceed to completion once
started. The continuation of preheat during
an intermittent welding operation is considered a necessity but will not by itself
assure freedom from cracking.
4.
In the event of loss of preheat either locally
or on an entire assembly due to power or
equipment failure, welding should cease until proper preheat is again established. If
the weldment has cooled to ambient temperature while in a partially complete condition,
a magnetic particle inspection should precede re-institution of preheat.
5.
Injudicious placement of heaters should be
avoided.
To the extent that the heaters interfere with the
welding or allied operations there is a danger that
careless mechanics will remove or disconnect them
thereby increasing the likelihood of cracking. At
best, preheat is a necessary evil, to the production
workman. Education will temper this attitude but
only evidence of careful consideration of his problem will make it bearable.
(d) The use of weld overlay in weldments of
heavy mass and high restraint to permit the
134
Figure 1I Strip Heaters un Submarine Hull Section for Internal Frame to Shell Attachment V~eld.
Note Heater Power Distribuztion System.
135
Figure 2
-Close-Up
of View in Figure #1
136
Figure 3
-
Strip-Heaters for Vertical Butt Weld in 2-1/4"
137
HY-80.
Figure 4 - ideaters on Flange Plate for Internal Frame
Pre- Assembly Weld.
Figure 5 - Heaters on Overhead Portion of Circumferential Butt Weld on Shipway.
136
A REEPORT ON THE
DEVELOPMENT OF A CRACK RESISTANT
ELECTRODE FOR HY-80
By
S. 1. Roberts - Head Welding Engineer
Portsmouth Naval Shipyard
Portsmouth, N. H.
March 22, 1960
Abstract - This is a discourse of one phase, of
developing, proving by test, and approving by controlled production use the crack sensitive nature of some
coated electrodes. The electrode involved is the iron
powder coated type in the range of above 80,000 psi
yield strength of the AWS classification, 9018. This
covers the selection of the electrode most suitable
from several samples, and a method of grading or
predicting whirh will perform better in production use.
Because of the complicated heat treatment from
welding, further confused by many variables of production and by extremely high rigidity and restraint,
it is easy and prudent to say "we don't know", as to
the cause of cracks. Certainly there is a combination
of many causes that result in cracks in the welds, and
in the HAZ.
History: The performance of ship plate in World War
II, demonstrated the need for a higher strength material, particularly a more tough material thaf would
be resistant to brittle fracture. The added requirement would be that the material would have good weldability. The HY-80ai,d its cousin the commercial T-1
steel is the result.
We began this test on the free admission that we just
don't know what causes the cracks and why some
electrodes do have a less propensity toward producing
cracks. Again we do not know what properties in the
electrode make the difference. We simply try to
predict the type that will perform better in production
wielding. Eventually we hope to isolate the factors
that contribute to the better performance. We reasoned that if many different crack producing configurations could be placed on a plate, reducing the variable, some electrodes may prove better than others.
We therefore made different welds on a plate as
shown in sketch No. 480, and 523. It is immaterial
what kind of welds are made on the plate if identical
conditions exist in the vertical columns except for the
electrodes. However, we have found that a variety of
welds as shown in the sketch 480, will result in a
more reliable reading, as it reveals weakness in
many different areas of the weld metal or HAZ.
Comparative electrode test
This HY -80is a quenched and tempered steel of
high toughness and high resistance to crack propagation, especially at low temperatures. Forourthinking
in fabrication we might broadly say that this steel
receives its strength from its quenched and tempered
mntrasted to the strength from
heat treatment, as
chemistry of our strength steels used in the past.
This obviously is a rather loose description and not
a full description of this material, but it is sufficient
for our need at this writing. This being a quenched
and tempered steel requires somewhat more flexibility in our thinking when we have cracking problems.
For we do not have a steel designed to good weldability as proven by many laboratory tests and, indeed
still can be. Yet major problems of production use
have been, and still are encountered.
-
To obtain a comparison, the welds are flushed off
smooth with the plate surface and the plate bent to
reveal any weakness. Using a known crack producing procedure, with the wide variety of conditions,
gives us a good norm to pred!ct which electrode can
be expected to give better production welding.
The part of this quenched and tempered steel in
the heat effected zone is changed greatly in different
areas and spread over a wide range of reheat treatment. A portion may be quenched relatively severe
while another area may be heat treated differently.
Part may have a reheat similar to a temper, the degree depending on many factors.
Repeating, we purposely try to create weld metal
or HAZ micro-fissuring weakness and then subject
the plate to abnormal fiber stress to reveal this or
any weakness. Different electrodes do have a varying effect on the plate.
The indications as shown in sketch 539 are obtained by coating the plate surface with liquid penetrant white developer, then magnetic particle inspect
the plate, using red ferro powder. This enables a
photograph of the area to show the number and extent
It remains for the field activities to acquire the
know how by trial and error to adapt this steel to successful weldink. The laboratories have done their
work well, and now it is a problem of the many variables in the shop and in the field,
139
of all cracks, in true detail. A close-up photograph
is made of separate areas to obtain exact crack detail
to transfer to a sketch. This enables one to read the
comparison quickly and accurately.
The Bureau of Ships (Code 637) furnished a sampie of type S580 electrode and also a 3 1/2% nickel
type. These did not satisfactorily compare with the
11018 which we use as a standard. After several different brands and types showed no comparative advantage, a brand "R' repeatedly produced much less
cracking. This brand even compared favorably with
type 310 (25-20) for comparative cracking. The physicals were satisfactory to the design engineers, so
a small order (5000 lbs) was procured, conforming to
the mudinum requirements of specification Mil-E19322, type 9018, for the physical characteristics.
This electrode was used in production for noncritical, troublesome areas such as BLHD penetrations, attachments, foundations, and similar applications that had caused trouble when using 11018. This
aspect was important as it verified the difference as
revealed in our comparative test. In other words we
needed to verify that the superior performance would
also be apparent in actual production. To prove beyond doubt the superiority of this brand, we did not
use the 9018 until the 11018 was tried and cracked
at least once, and usually several times. Only then
did we repair with the 9018.
brands of 9018. The crack resistant characteristics
as revealed by the comparison bend test, were not
quite as goodas the first order, yet better than other
brands and the 11018. The sketch (No. 539) shows the
approximate comparison. This is only approximate,
due to the fact that the indications of cracks on the
sketch were not scaled from the photo enlargement.
However, it is the electrode "DO in the sketch, and
it revealed better performance.
In production there were isolated reports of a minor number of cracks although the full resulti of production use is not available at this writing.
There are two other brands that have proven superior in the bend test. Procurement of 5000 lbs of
each of these brands has been completed. The comparative bend test will be conducted on these and if
the reaction is similar to the sample tested, they will
be placed in production. This limited controlled use
will be in accordance with instructions outlined in
sketch No. 462-B. Because of the limited quantity
they will be used mainly for repairs. Especially in
the "tough" areas to get a production comparison
with previous work.
It is believed that the specific characteristics can
be isolated from this work and the electrode can then
be manufactured to a predictable crack resistant
characteristic.
The results were very gool. Areas that were extremely troublesome were repaired satisfactorily.
Repeated inspection reaffirmed the quality and the
absence of cracks at the time of repair and later.
The chemical and physical data was compiled by
Portsmouth on several different heat numbers of
Brand "Rl". (stress relieved and as welded)
After this demonstration of superiority, the use
for repair of cracks in the Main Ballast Tank 4A and
4B, was authorized. The isometric sketch No. 469,
and ore typical detail on Sketch No. 469-6-I, indicates
the extremely high restraint and rigidity of this
structure. Most all of the members shown are HIY-80
over I" thick.
t
The summary list indicates the extent of the troubleand lists the thirty areas, their size and number
of repair attempts with 11018. Using only moderate
preheat (70°F) not one crack appeared in the 9018 repair. Nor were there any related cracks in the area
near the repair often associated with preheated welding in rigid structures such as this. An important
item concerning preheat was, it is believed, that the
entire tank was uniformly warm. This item was behind schedule, and so much "round the clock" work
had heated it to an even heat. This supported the belief of many Welding Engineers that an even. the low,
preheat is better than higher local preheat in areas
of strong rigidity.
To test the supplier's ability to reproduce this
electrode, we placed another order (10,000 lbs). This
was tested by comparing it with 11018 and other
The range of Portsmouth's and those supplied by
Cbmsfy
Low
High.
C
Mn
p
S
S
Mo
Ni
0.06
0.60
0.009
0.025
0.34
0.26
1.42
0.09
1.15
0.02
0.028
0.48
0.32
1.69
PhylIka.ID
Low
High
Yield
Tensile
85,000
99,000
100,000
106,000
Percent elongation in 21 - 22% - 26%
Charpy V at - 60"F - 21 ft Tbe - 45 ft lbs.
140
1
908
o
31
2
3
4
A
PLATE BENT
AFTER REMO0VEAL
OF ATTACHMENTS
32
AIR AR C
GOUGE]
SECTION A-A
141
5
6
7
8
I
2
"I
iL
3
:
4
5
6
7
8
9
I LA- I/.L.L
I.
10
0
MJ
I ML
2
ML
I Al
90" APPROX.
-
Description of attachments and simulated repairs
1. Columns 1 and F
2. Columns 2 and 7
-
-
2" X 2"ang:le iron clips welded with one pass.
.Xrc-air ,ouged Tep -.roove in plate surface filled with
non beaA.
S('IColumnns .- and 8 - 2" X 2" ,aiiil, iron clips wi,lded with 2
;l0ad;.
4. Columns 4 and 9 - one be.id on plate surface about i n,.
v
5. Columns 5 mnd 10 - Arc -ar gouged Tee groove in plate surface fillhd with
3 Ix-ads thv !;ast At which e(ers 1 2 of the first 2 bvads-.
PROCEDURIE
1.
2.
?.
4.
Plat,. surface was ground free (A paint and scale.
Arc-air g•uging accomplished in areas desiglinite i.
Wld attachments and simulated repairs as indic.itvI.
M•.gnetic .prticle insp•ct each weld when plate c',)li, I t(o anihnt-t Iimcorni tur,,.
5,eni
tv anvlh (lips and grind welds smooth and flush w'th plate surfacc.
6 Magnetic particle inspect welded areas after rolling pl:ite to a.bout ain
1$" radius with the welded surface in tensi,)i.
142
Y€
I.'
NOTES:
I. Each area to be worked was cooled to 40°F immediately before welding was started.
2. Each vertical row of attachments and simulated repairs was welded on the same day.
?. All welding was done by the same operator.
4. The voltage and amperage range used was 24-26 and 180-200 respectively.
5. Current and travel speeds were maintained as nearly equal as possible for welds in
the same vertical column.
6. Chilling of the plate before welding was done to induce weld cracks as a medium for
comparison and is not an indication of production welding practice in use at this shipyard.
7. Variables have been minimized in conducting this test in order that each electrode receive identical treatment.
This test is designed to reveal weakness or micro-fissuring in the weld metal, diluted weld,
or the HAZ of HY8O.
The variety of welds applied evaluates the different conditions caused by electrodes. This
number also gives a good norm to evaluate the results. To repeat the above procedure instructions,
details must be maintained identical in the Vertical columns except for the change of electrodes.
Only by this leveling off of the variables can a true comparative evaluation be made.
Sheet 2. of 2.
WE Sketch No. 523
143
2
1
It
ELECTRODE
ELECTRODE
ELECTRODE
3
4
6
5
8
7
1S '-T
I
-j
ELECTRODE
9018
110
-18
RECORDED CRACKS
Ai
BENT PLATE
2
9
110-18
WELD
GROUND
B
9018D
144
9
10
SUMMARY REPORT OF 9018 ELECTRODE
-.
. ....
...
.
.
-
........-
Sketch No.
MLN-_._TA4I
M~ft.A~i!K
469-1
4A_
Type of Crack
Nwith
-SSNLN)6O2
-A&4BLong. & Trans
-
-
-
-
-
-
-
-
-
Size of Re•ir
.......
-
-.
9018
--......-
12xi
6
3
1 1/2X3/4" x 1/2"
5
5
3
2. 3" X 1/2" X 1/2"
3. 3" X 1/2" X 1/2"
3" X 1" X 1 1/2"
.
..
-3
-4
"
-5
"
.6
"
-7
-8
"
-9
-10
-11
"
"
-12
-13
-14
-15
-16
-
-
'-
'
"
"
.
Long.
-25
Long. & Trans.
"
-26
-27
Long.
Long. & Trans.
"
',
-28
1.
5" X 1" X 1/2"
"3.
18" X 2" X 2..
6" X I" X 1 I/2"
4 1/2"
9" X I" X 3/4"
7" X 2" X 1/2"
4 1/4" X I" X 3/4"
2.
"
-29
Long.
-30
"
Repeatedly
5
2
--
Z
Repeatedly
4
4
..
'
.
..
t
-
..
Repeatedly
1
3
4
1
7
2
3
Repeatedly
.
.
.
4
u
l
-
W6
0
Z
o
4j
{
1 1/2" X1 1/'4" X 1/2".
36" X 2" X 1
4" X I" X 1/2".
..
-22
-23
-24
6
X 3/4" X 1/2"
2" X 1/2" X 1/2"
2"X3/4"
,3"X
6" Xl" X 1/4"
3" X 1" X 1/2"
6"X 1" X 1
Long. & Trans.
-21
4
X 1 1/2"
4" X 31/2" X 1/2"
3" X /4" X 1/2"
4" X 2" X 3/4"
4" X 1/2" X 1/2"
3 1/4" X 1/2" X 1/2"
Lng.
Long.
Repeatedly
X 2" X 1 1/4"
14" Long
3" X 1" X 3/4"
71/2" X1 1/2" X1
"
-18
X 2" X 2"
X 2" X 2"..
3" X 2" X 1"
4" X 3/4" X 1/2"
Long. & Trans.
Lung.
Long. & Trans.
-17
-19
-20
1. 40"
"2. 11"
1. 4"
"2. 5"
1. 4"
'
"
inal Results
No. of Re pir
.
1108
1.
.
..
-2
-
4 1/'2- X 1/2 X I"..
4
7
3
2
1
NOTES
I. Original weld material was 11018.
2. Various preheat from 70' F. to 200 F. and different sequences were used without success.
3. Final repair was made using preheat of 70*F. with 9018 electrode.
4. Magnetic Particle Inspection after five dayq showed no cracks.
9018 ELECTRODE
PORTSMOUT4 NAVAL SHIPYARD
Welding Eng. Branch
Abraham Lincoln SSB(N)602
145
WHO
SEE(
NOTE
D~
%0
,2
ALWAST
I~N
US
IRD
WIH01404
N4 ARA IS
0TROBLEt"
S.
I
46SE OE 2
469 1
PO*T~wu;,".
;;o
TO
T
44
TANI
tLICER
CA,L
..
rIO
ft
TSUBLI:IEEIARDTIO9 3(10
0
44
~vALqN,146-
lY
46%-S
PURPOSE. To provide revised local instructions for the handling, stowage, and use of 9018 type
electrodes. These are experimental electrodes suppliel from different manufacturers
for controlled production use to develop information foi" specification writing.
ELECTR!QE._: Description - Electrode, Welding, 5/32" dla. Type AWS 9018
Color Coding; Primary
(end)
Red
Secondary
(spot)
Orange
Group
Green
STOCK DESIGNATION:
PNS local stock No. GL-3432-LOO-7580.
released by Code 355.
Manufacturers brands as
CONTROL. Experimental Type 9018 electrodes will be used only on controlled production. Electrode
will be held in locked stowage ovens, the key held by the responsible supervisor. Electrode shall be
issued by signed IBM card retording the area where the electrode is to be used. Dry the electilode in
an oven at 250'F. to 350 F. a minimum of four (4) hours prior to use. Electrodes exposed to atmosphere for four hours are to be returned to the ovens for reheating.
USAGE: This 9018 is released for controlled production use per PNS Itr 355/250 10310 of 30 Dec 1959
to BUSHIPS. Its use will be limited and governed by the following rules:
A.
B.
C.
D.
E.
F.
The physicals of any electrode psed will meet or exceed the specification
Mil-E-19322 for type 9018 electrode.
Record will be made of the location and history of all 9018 electrode used
on HY-80.
Each lot of electrode will be tested for satisfactory performance prior
to release for us( in productihn.
No welding will be permitted on plates that are at less than 70 F temperature.
No 9018 electrode will be used on butts or seams in the pressure hull,
either in the shell plate or btilkheaas.
No 9018 electrode will be used for the attachment of external frames to
shell plate.
It is requested that these rules be closely followed and that any cracking or unusual r
2t,
be confirmed and reported immediately to Code 355 so that reliable specifications can be developed.
CODE 355
SHIPYARD WELDING ENGINEER:
_ELDING ENGINEER'S- PORTSMOUTH NAVAL SHIPYARD, PORTSMOUTH, N. H. WELDING PROJECT
SKETCH NO. 462-B
PRODUCTION DEPARTMENT
NO. WE-
-W
esINSTRUCTIONS
DATE: 13 Jan 196U
|
DZT~.]e -I•
I rio..]-
~,3
,
,'
SEN)593
FOR USE OF 9018 ELECTRODES
SSBN)602
Ul_
--
9nISI
%n•l
&
.I
111J
T
I I I I
-
J.O.
SHEET 1 OF 1
OENLS6URE:
ESTIMATOR:
DATE ISSUED:
147
.
.8
EXPLOSION TEST PERFORMANCE OF
SMALL-SCALE SUBMARINE HULL WELDMENTS
by
A.J. Babecki and P.P. Puzak
Welding Metallurgy Branch
Metallurgy Division
U.S. Naval Research Laboratory
Washington, D.C. .
ABSTRACT
Underwater explosion tests of full-size submarinehull sectors appeared to show the new hull materials
might not be as effective as at first anticipated.
Analysis of Iocalized ruptures indicated the cause to
be a too-rigid framing design for the high-yieldstrength steel HY-80 rather than a deficiency of the
material. Simulated hull-structure components have
been used in explosion tests to ascertain if design
was responsible for the HY-80 submarine-hullsector failures and to demonstrate, if possible, thevalidity of structural performance prediction based
upon explosion-bulge test results of materials.
Comparison tests were made between structure
components fabricated with the high-tensile-strength
steel, HTS, and with HIY-80. The HTS weldmeift
samples were prepared by two welding procedurop
authorized for HTS (manual-arc. Mil E-8016 electrode and submerged-arc, Mil B-3 wire). The HI80 weldment samples were prepared by two authorized welding procedures (maesual-arc, Mil E-11018
electrode and automatic inert-gas-shielded metal-arc,
Mil B- 88 wire, A + 02 cover gas) and two proposed procedures, (manual-arc. Mil E-310-15 electrode and
semiautomatic inert -gas- shielded metal-arc, Mil
B-88 wire, A+C02 cover gas, vertical welding).
metals (Mil E- 11018 and Mil B-88) and an austenitic
weld metal (Mil E-310-15). The test results showed
that the Mil E-310-15 weld metal develops unique
and extensive fusion-line failures. The two approved
weld metals developed no failures despite the increased straining.
These tests duplicated the explosion-tbulge test
results of materials by themselves, which indicated
the HTS to be brittle at 0°F and 304F and that the
Mil E-310-15 weld metal develops a %%eakfusionline which serves as a ready path of failure. These
duplications support the concept that structure performaneo roild lv- predited from tests of materials.
The structure test results also showed that under
identical framing design, the HY-80 is superior to
the HTS.
INTRODUCTION
In explosion tests conducted at 30'F and 0°F with
simulated highly restrained joints, 'he -ITS samples
developed extensive brittle plate failures, whereas
the HY-80 samples developed relatively short ductile
tears, indicating superior plate toughness of the HY80 under identical joiat-design conditions as for the
HTS. No differentiation could be made in the performance of the weld metals, because the rigid joint
design protected the weld from high strain by concentrating the strain in the plate adjacent to the
weld seam.
Past efforts of the U.S. Naval Research Laboratory's long-term specialized welding-research program have been based on establishing the fundamental
factors which determine the fracturing performance
of weldments (1,2). These studies have entailed extensive tests of available structural materials conducted over a range of temperatures. This research
has demonstrated the importance of test temperature upon resulting fracture performance and has
resulted in the formulation of new concepts (3, 4)
which provide for a more exact definition of the
properties required for the weld, the heat-affected
zone, and the parent material of weldments used in
ship and submarine construction. The concepts derived from these investigations have served as the
basis for the procurement of plate (HY-80, Mil
S-16216 C) and weld materials (B 88, Mil E-19822
and E-11018, Mil E-19322) currently specified by
the Bureau of Ships for new submarine hulls.
The structure component and explosion test die
were modified to permit greater strain development
in the weld metal. Tests were made with HY-80
components fabricated with the two approved weld
The use of explosives for performance evaluations
of welded joints and prime plate materials Is a longestablished Navy procedure. For many years, large
open-ended cylinders containing welded joints were
148
€a
immersed in water with a charge which was exploded either in the center of or removed from the
cylinders. On a still larger scale, full-size
submarine-hull sectors have been similarly tested
for purposes of evaluating fabrication procedures
and optimum joint configurations. Generally, however, these full-scale sector tests were conducted
only in the summer months, when the water temperature was relatively high.
weldments were fabricated by PNSY under normal
shipyard welding conditions; however, a few weldment. were fabricated under laioratory conditions
by C Company with plate material supplied from
PNSY,
TEST PROCEDURES AND RESULTS WITH HTS
STRUCTURAL JOINTS
Figure 1 gives the dimensions and joint configuration of the simulated structural components considered for initial tests of HTS material. Several
components were fabricated with the two weld metals
authorized for the welding of HTS (manual stick
electrode, Mil E-9816, and automatic submergedarc electrode wire, Mil B-3). These components
were explosion-bulge tested on the rectangular test
die illustrated in Fig. 1. The spacing between
T-frames and test-die radii was chosen to provide
support so as to simulate the high degree of restraint
developed in the submarine at design positions which
feature external framing opposite internal bulkheads
or tank compartments. The die radii precluded
gross bulging opposite the T-frames, so that only 1
percent or less thickness reduction was developed
in the hull plate adjacent to the fillet-weld regions..
The explosive chsrge used for all specimens was
7 lbs of pentolite; however, the offset distance of
charge to plate was varied from 15 to 18 in. for
some tests in order to facilitate comparisons between HTS and HY-80 specimens at the same relative level of deformation developed at the apex of
the bulge. Complete test data are given in Table I.
New submarines are designed for much deeper
submergence than has ever been considered before;
stricter specifications for hull' requirements, ineluding material strength, toughness, and scantling
design, were necessary to combat the greater water
pressure. An apparent difficulty developed when
initial explosion tests of full-scale sectors of the
new HY-80 hull structures indicated that they might
not be as effective as anticipated. Localized tears
were developed at certain locations in the framing
members which did not penetrate the Sector's pressure hull. Although the nature and locations of the
tears suggested that the difficulty was related to the
design of the framing rather than to deficiencies of
the material, questions were raised concerning
possible differences in fracture performance which
might result from tests of materials by themselves.
as contrasted with actual tests of highly restrained
structural joints,
If, as suspected, the framing design, w-as responsible for the localized tears devcl.)ped in the ItY-80
sector tests, it was essential to investigate the behavior of various weld-joint configurations in order
to develop improved design details. It was also considered necessary to demonstrate the valdity of predicting structural performance on tht- basis of test
results of materials exclusively. Because of the
cost and the impracticability of controlling test ternperatures for the large-scale sector tests, there
was a need for the development of smaller-scale
tests suitable for screening studies. This report
describes new developments in the procedures for
NRL's explosion-bulge test (5), which now permit
the testing of structural components, such as the
stiffeners for the pressure hull. The new procedures allow for the use of relatively small seetions (approximately 2 x 2 ft) which are readily
fabricated and permit rapid investigation of a wide
variety of design methods.
Figures 2-4 illustrate the results obtained with
the HTS specimens manually welded with Mil E-8016
electrodes. In the tests conducted at 30 F (Figs. 2
and 3), the fractures started in the HTS at the toe of
the fillet welds, and comp!ete penetration of the
hull plates was developed on the first shot. At 0 F
(Fig. 4), similar fractures were developed; however. propagation of these fractures was sustained
through the hold-down regions of the plate and resuited in complete separation of the specimen. The
brittle fractures which developed in these HTS structural components are in complete accord with predictions based unon prpvious 0 and 30 F tests of
average-quality HTS (6).
For the HTS s.ampies automatically welded with
submerged-arc Mil B-3 electrode wire, complete
penetration fractures of the HTS hull plates were
also obtained on the first shot (Figs. 5-7). In addition, two of these specimens also developed secondary fractures in the submerged-arc fillet welds.
One of the T-frames was partly blown off in the 30'F
test (Fig. 5), and one T-frame was completely
blown off in the 0'F test (Fig. 7). As shown by the
photomacrograph inserts, these secondary fractures
started in the HTS web section at the toe of the fillet
weld but propagated almost exclusively in the weld
metal. Previous tests of submerged-arc weldments
The materials considered for this investigation
included the conventional high-tensile steel (HTS,
Mil S-16113) previously employed for submarine
construction and the new-tough, high-strength HY80 steel currently specified for all new submarine
construction. Several different plates of each steel
were ubed for the various weldments to be described,
These were selected at random from stock at the
Philadelphia Naval Shipyard and used after checking
to insure that each complied with respective specification requirements. The majority of the
149
(7, 8) have shown this weld metal to be susceptible to
brittle fracture at 0' and 30'F. Thus, in addition
to the fractures developed in the HTS, the weldmetal fractures developed in these structural cornponents are also in agreement with predictions
based upon materials tests exclusively.
RESULTS WITH HY-80 STRUCTURAL JOINTS
Simulated structurkl components similar to the
configuration shown in Fig. 1 were also fabricated
with HY-80 steel. In addition to the Bureau-approved
weld metals (manual stick electrode Mil 11018 and
automatic, inert-gas, metal-arc Mil B-88), sampies were welded with the manual stick electrode,
Mil 310-15 (austenitic- 25 percent Cr, 20 percent
Ni), and with semiautomatic, inert-gas, metal-arc
electrode (vertical position welds with Mil B-88
electrode using experimental cover gas of 80 percent argon + 20 percent C0 2 ) and included for
comparison purposes. These components were all
explosion tested with conditions similar to those
described previously for the HTS samples.
Thickness-reduction measurements indicated that the
restraint afforded by the die radii also precludedthe
development of more than 1 percent thickness reduction in the HY-80 hull plate at locations adjacent
to the fillet welds. For comparison with HTS resuits, it was established that a 15-in. standoff of
the 7 lb charge resulted in approximately the same
relative level of deformation (5 to 6 percent thickness reduction) at the apex of the HY-80 bulge as
was developed by an 18 -in. standoff for HTS samples.
The appearances of the various HY-80 samples
after one explosive shot are typified in Figs. 8 and
9. Test data for these samples are included in
Table I. In each case. ruptures were started in at
least one of the T-sections in these HY-80 samples
at the positions of high restraint adjacent to the toe
of the fillet weld. However, in only two samples
(Nos. 7 and 14) did the resulting rupture penetrate
through the 1-in. -thick hull plate. This penetration
was relatively small and barely visible on the tension side of the sample. All other HY-80 specimens
exhibited only small tears which only partially penetrated the 1-in. -thick hull plate.
Irrespective of the weld metal employed for these
difference could be
HY-80 specimens, no significant frctue
dvelped
he
f initial
intia fracture developed
attrn of
observed in theobsevedin
pattern
at the first shot. A second explosive shot was given
to each of these HY-80 specimens. Figure 10 is
representative of the fractures which developed after
the second shot in the various HY-80 samples welded
In each case where the
with
first different
shot had electrodes.
developed a partial penetration crack
firscentshothad dWevweloped,aatihen
wasextenOnly
crack
adjacent to the fillet weld, the crack was extended
weld was developed by the second shot (compare Fig.
8 with Fig. 10).
Macrosections were cut from the penptrationfailure areas of these samples for examination.
Irrespective of the weld metal employed, the fracture pattern was similar for all HY-80 samples. As
shown in Fig. 11, the partial-penetration failures,
denoted by Area A, developed as tensile ruptures.
The extension of these failures to full penetration on
a subsequent shot, area B, occurred in a punching
or true-shear manner. Thus, design positions
which feature external framing opposite internal
bulkheads, or tank compartments are shown to
represent a particularly unfavorable design "hard
spot. " The high restraint and restricted movement
incurred at these positions precludes the development of much deformation in the HY-80 and is conducive to the formation of shear-type failures via
the punching mechanism described above. It should
be noted, however, that the high-energy-absorption
shearing characteristics of the HY-80 at these ternperatures precludes brittle fractures and extensive
prop'igation such as that developed in HTS. Thus
the resulting fractures in simulated structural joints
of HY,-80 are limited to relatively short shear-tears.
RESULTS WITH MODIFIED HY-80 STRUCTURAL
JOINTS
In submarine fabrication, some of the highly restrained fillet welds of framing members to the
pressure shell occasionally require ccostly weldrepair operations of small fabrication cracks occurring at the toe of the fillet weld. To alleviate such
cracking tendencies arnd to reduce o0erall production
costs, some of the shipyards have requested Bureau
permission to make frame-to-shell welds with
austenitic Mil 310-15 electrodes. However, a performance evaluation of the Mil E-310-15 weld metal
in such a joint as compared to that of the approved
ferritic weld metals had to be made prior to such
approval. The explosion tests described previously
for structural joints, simulating conditions of external framing opposite internal bulkheads, failed
to discriminate in fracture performance between the
various weld metals employed in the HY-80 samples.
The reason for this failure was that the test conditions at
concentrated
most
of the
strain
in theinsuffihull
plate
the toe of the
fillet
weld,
thereby
as ahole. The
straining the wet
pae
The the
as a whole. over
the weld
ciently straining
solution
was to spread
out joint
the deformation
and full penetration of the hull plate at that area was
entire weld joint and prevent excessive localization
at this critical region. Accordingly, the test die
tere
. 1h
at spcimegion According
and specimens of the type shown in Fig. 12 were
prepared to study possible differences in fracture
performance of different weldment combinations.
the Bureau-approved weld metals (Mil 11018
and Mil B-88) and the proposed austenitic Mil 310-15
electrodes
developed by the second shot. In addition, for the
T-sections which were not cracked on the first shot,
a partial penetration crack at the toe of the fillet
The primary difference between the conditions of
the modified T-sector test and that of the original
were considered for this evaluation.
150
II
test described earlier is that related to the support
rendered by the die at the web-to-hull fillet joint. In
the modified test, the support opposite the T-section
is not nearly so great. Explosion tests of these
samples could develop more tensile strains on the
weld metal and permit discrimination in the performance of various weld metals and/or welding techniques. As was expected, the fracture behavior
was similar for both specimens welded with the
Bureau-approved weld metals. Figure 13 illustrates the appearances of those specimens after a
second explosive shot of 7 lb at a 15-in. standoff
distance. The fractures, which started in the HY80 at the toe of the fillet weld, did not penetrate
through the 1-in. hull plates. A third shot was required for these samples before the fractures were
observed to penetrate through the hull plates. As
shown in Figs. 14 and 15, however, the penetrations art, short tears barely visible on the tension
side of the specimens; this is indicative of the high
fracture resistance of HY-80.
The pattern of failure which developed in the
specim•n welded with the austenitic Mil 310-15
electrodes, however, was significantly different
from that described above. As snown in Fig. 16,
extensive tearing in the web section of the specimen
was developed by two shots of explosive at a 15-in.
offset distance. After the third shot, the T-frame
was completely separated from the hull plate (Fig.
17). The source and path of propagation of fracture
were uniquely c,.fined to the fusion line between
the austenitic weld and ferritic plate. Such extensive failures have previously been shown to he typical of failures to be expected at all ambient temperatures in austenitic-ferritic weldments which are
subjected to deformation loadings (9). AMcordingly,
the use of austrnitic welds for framing attachments
to the pressure shell at locations which may be
subjected to deformation is considered to be especially hazardous.
the materials themselves as prime plate or nonrestrained butt weldments.
The similar fracture performance shown by these
materials, whether in unrestrained or highly restrained weldments, implies that the performance is
dependent upon material properties and not upon
structural design. That is, if one material behaves
more poorly than another in an unrestrained test, it
will be poorer in a restrained test as well.
Structural design is of particular importance,
however, in determining the procedure by which
failure may develop in notch-tough materials at particular design locations. For the test specimens
which simulated the conditions of external framing
opposite internal ioulkheads, the ultimate failures
which deve-l!o.ped in the notch-tough HY-80 samples
were identica; irrvspective of the weld metal employed. The high restraint developed at such locatwons precludes the development of deformation in
the weld region, so that under explosive loadings,
failures are ultimately developed via a punching or
shearinv mechanism when the material is overloaded. tmproved pe,'rformance can be expected at
these "hard-spot" reg:ons by design details which
are aimed ait spreading out the deformation and
prevent iii excessive localizat ioit of strain at
critical regions.
REFERENCES
1. Hartbower, C. E., and Pellini, W. S. ,
"Investigation of Factors Which Determine the
Performance of Weldments, " Welding J.
30:499-s (1951)
2. Puzak, P. P. , Eschbacher, E. W. , and Pellini,
W.S., "Initiation and Propagation of Brittle
Fracture in Structural Steels, "Welding J.
31:561-s (1952)
3. Puzak, P. P. , and Pellini, W. S., "Evaluation
of the Significance of Charpy Tests for Quenched
and Tempered Steels, " Welding J. 35:275-s
(1956)
4. Pellini, W.S. , "Notch Ductility of Weld Metal,"
Welding J. 35:217-s (1956)
5. Pellini, W.S. , "Use and Interpretation of the
NRL Explosion Bulge Test," NRL Report 4034,
Sep. 1952
6. Puzak, P.P., "Notch Ductility of Normalized
HTS Steels," NRL Report 5007, Sept. 1957
7. Puzak, P.P., "Explosion-Bulge Test Performance of Machine Welded 1 Inch Thick HY-80
Steel," NRL Memo. Rept. 691, Apr. 1957
8. Puzak, P.P., and Babecki, A.J., "ExplosionBulge Test Performance of HY-80 Weldments,"
NRL Memo. Rept. 878, Dec. 1958
9. Pellini, W.S., and Eschbacher, E.W., "Investigation of the Performance of 1 Inch STS Weldments of G260 and 25-20 Types," NRL Memo.
Rept. 191, July 1953
SUMMARY AND CONCLUSIONS
In previous investigations, all deductions and
conclusions have been based upon the results of
explosion-oulge tests of materials, as prime plate
or nonrestrained butt welds. In those studies, the
HTS plate and the submerged-arc weld materials
weie found to be characterized by partial or high
brittleness at temperatures of 00 to 30° F.
Specification-quality HY-80 materials, on the other
hand, were shown to be highly notch tough and resistant to brittle fractures at these temperatures.
In the present investigation, which featured tests of
highly restrained structural joints, the fracture
performance and extent of failures which developed
in the various samples were in complete agreement
with results of previous studies. Thus it is concluded that the fracture performance which may be
developed in these materials in highli restrained
structures is predictable from results of tests on
151
a
a.
A.C
L0 c
.9..
.W W Ug
I
--
- -
cA.l
Nmc
z
-~a
V
.
co o
.
F-F- H
Lo
~
~
-
-
Z-
oý
-
A. M. MA.M>
>0
-4
- ao MC 0
>4
>4 >4>
0
Wi
1520.
-t
4>
4>
oO
"a
.4e
0.
EXPLOSION TEST SPECIMEN
EXPLOSION TEST DIE
II
II
II
II!
II-
(•-
I'
~x"
II
Ii
II
ROLLII
OF
I!I
ROLLII"
Ii
ii;
II!1 ,
25"
F-
-
T-32 JOINT DESIGN
"/
,/
/
/,"/
-
.
1r R"
2"R
%"[/
,,520 t,,
/
,"
Figure 1
"
--
MAX
/
7
25"
/
"-MIN.
-/'.-
/
.,///
/ i//"/
Simulated structural component used as explosion test specimens and weld joint design (left);
configuration of explosion test die (right)
153
~4.
S1
NO
"I
•Y
Figure 2
-
Test specimen No. I alter one expliusive shot.
154
IH - MiL 8016
SHOT 300F 7 LB
15-IN
Arrows indicate extent of ruptures.
ff02 HTS -M11-8016
Wt SHOTOF~ 11LB 18.1W
Figure 3
Test specinict No. 2 after one- explosive shot.
155
Arrows indicate extent of ruptures.
Figure 4
-Test
specimeh No. 3 after one explosive shot.
158
Complete brittle fracture at 00 F.
1104 HTS-SURVERCED ARC-NIL 0-4
IV SHOT301 118B 15-IN
Figure 5
specimen No. 4 after one explosive shot.
Insert shows submerged-arc weld failure.
-Test
157
Arrows indicate extent of ruptures.
HUFPTS-SUSIIRC(0 ARC-OIL 1-3
IV3 SH41 WfI L68VII-
OUT-10
77T.
Figure 6~ - Test specimen No. 5 after one explosive shot.
158
Arrows indicate extent of ruptures.
1110 HIS
-SWhNRCED
II SHO7 O*¶
I16
-NIL
ARC
-3-
5-1I.
Figure 7 - Test specimen No. 6 after one explosive shot. Arrows indicate extent of ruptuares.
Insert shows submerged-arc weld failure.
159
~9NT-GOU-II 11018
ILI SHTOfI B1-N
Figure 8
-Test
one explosive shot.
specimen No. 9 alter
160
Arrow~s indicate extent
ot
ruptures.
Is, $NOT
04F T'U
is-IN.
K1 NY"-,-I-
Figure 9 - Test specimen No. 14 after one
explosive shot.
indicate extent of ruptures.
161
Weld made vertical up.
Arrows
9OD
HY-80 - NIL 1101i
229 SHOT 0lf ILI 15-IN,
Figure 10 - Test specimen No. 9 after two explosive shots.
162
Arrows indicate extent of ruptures.
Figure 11
-
Typical fracture of HY-80 test specimens showing
.tensuie rupture (A) and shear
rupture (B) in failure process.
163
4
II
I
_
___
T
__
__
__
__
--------------------
W~
,rKi
$W
SPECINEM
TEST
-U-TEST
Figure 12
-
K
re
Modified explosion'3test specimen and test die designed to
produce greater distribution ofi strain over weld joint.
Note diminished support of die beneath tee web.
164
A
two explosive
19 and 2()after
edmtl
peciens140
~oe
est
13
weided with 1
Both specimens j., tenlt of ruptulres.Arrws~ indicate
Shotsa
NO. 19-HY-80O-INERT CAS- NIL B-88
3' SNOT S0°F ILI IS-IN.
Figure 14 - Test specimen No- 19 after three explosive shots. Arrow
indicates extent of rupture.
166
NO. 20- NY-SO- NIL- 11018
3*SHOT 3061 7LI-15-IN.
Figure 15
-
Test specimen No. 20 after three explosive
shots. Arrows
indicate extent of rupture.
187
NO. 21-HY-80- NIL- 310-15
2"1 SHOT SO*F ILI 15-M.
Figure 16 - Test specimen No. 21 after two explosive shots. Arrows
indicate extent of rupture in web section.
168
10.21 N- S~O-MNIL- 310-15
3 SHOT- 301 T" 15-Il.'
S
Figure 17
-
Test specimen No. 21 after three explosive shots showing complete fusion-line
separation of web section.
169
ID
r,
EXPLO6ION- RlLGKE TEST PERFORMANCE
OF EXPERIMENTAL SUBMERGED-ARC
WELDMENTS OF HY-80
by
A.J. Babeckl and P.P. Puzak
Welding Metallurgy Branch
Metallurgy Division
U. S. NAVAL RESEARCH LABORATORY
Washington, D.C.
Abstract
The materials and welding processes currently
used for fabricating critical Navy structures, such
as submarines, have received Bureau approval because they have been shown to produce weldments
which possess maximum resistance to fracture at
cold water temperatures. The demands of the current HY-80 sE-bmarine construction program make
it desirable to have additional, approved automatic
welding procedures to supplement the existing man-,
ual and automatic welding processes. To date, no
submerged-arc welding process has been approved
for this application because of g,.nerally poor performance of such weldments in explosion-lbulge tests.
ihis report presents recent test results for new
submerged-arc weldi•ig developments.
case" structures (submarine hull, torpedo defense
system, etc.) than normalized medium-carbon or
low-alloy steels. With Q&T steels, it is possible to
develop the optimum combination of strength and
notch toughness required to best withstand at cold
water temperatures the massive structural deformations expected under explosive attack. Of all commercially available Q&T alloy steels, it has been
demonstrated that weldments of the alloy designated
as HY-80 (Mil-S-16216C) could provide the greatest
assurance of complete resistance to brittle fracture
at these low temperatures (1, 2, 3, 4, 5). During
recent years, the use of HY-80 has steadily increased
for all critical BuShips submarine construction.
To date, the submerged-arc welding procedures
have not been approved for the welding of HY-80
Test results described herein showed that certain
structures because explosion-bulge test results have
combinations of submerged-are materials and tecd shown that the commercially available high strength
niques produce weldments which do not meet minimum materials and submerged-arc techniques produce
fracture performance requirements necessary for
either brittle weld deposits or inferior HAZ regions
submarine service. The weldments produced in
at submarine operating temperatures (3,5). Certain
these cases are considered unsatisfactory because
features, such as the magnitude of the HY-80 conthey exhibit potential paths of low resistance to fracstruction program, the shipyard availability of equipture propagation in welds of either brittle or lowment and trained operators, and the reduction of eyeenergy-absorption shear features. Other test resafety hazards inherent to submerged-arc welding
suits are presented for experimental materials which
techniques, have made it desirable to develop and
demonstrate that it is ppssible to produce highly
qualify submerged-arc welding materials and pronotch-tough, high-strength submerged-arc weld decedures suitable for automatic welding of HY-80.
Accordingly, a continuing development and test proposits. It is believed that such materials and techniques will be suitable for automatic submerged-arc
gram involving NRL, Naval and industrial shipyards,
weldments with weld metal toughness characteristics
and industrial welding equipment companies was esequal to or superior to those currently obtained with
tablished by BuShips (Code 637). This report preBureau-approved electrodes.
sents recent explosion-bulge test results for new
INTRODUCTION
submerged-arc welding developments. Specific details concerning procedures and materials used for
all weldments to be described herein are given in
Explosion bulge studies have shown that the Q&T
high alloy steels are much more suitable for "military Table 1.
170
The explosion-bulge test conditions employed in
this investigation duplicated those used previously in
other investigations of HY-SO weldments. All explosion tests were conducted at 0°F test temperature
to permit comparisons with previously reported tests
of various Q&T steel (4-8). In conventional explosionbulge tests, explosive shots which develop small deformations (2-4% thickness reduction) are repeated
until the first visible signs of failure are obtained,
Generally, testing of any oie sample is discontinued
if no visible failure is observed after the 3rd shot.
Such techniques permit a differentiation between
weldments by delineating the critical regions and the
level of deformation at which failures may start and
subsequently propagate in a given weldment. For
screening purposes aimed at a reduction of testing
time and costs, occasional weldments were mtdificd
by the addition of crack-starter welds and tested with
only one explosive shot*.
were found to be characterized by high brittleness
of the weld or the HAZ (3). Subsequently, similar
weldments made with lower J,'in. welding conditions
(approximately 45,000 J/in. ) were shown to develop
explosion test failures which were predominantly
brittle fractures within the weld metal (5). For this
investigation, the commercially available submergedarc materials were used by two different wxclding
equipment manufacturers to fabricate additional
HY-80 weldments. Relatively low J/in. welding conditions (32,400 Jim.) were used by one company with
its commercial electrode and an experimental Nialloy flux to prepare samples 1-3. The low shelf
values obtained in weld metal Charpy V tests (Fig.
1, Top) indicate this weld deposit to be characterized by low-energy shear features. In addition, the
impact values obtained at -60"F test temperature
(11 to 25 ft-lb) are significantly inferioi to those obtained with welds that are Bureau approved for the
welding of HY-80 (20 ft lb minimum at -60'F). Previous investigations of other low-energy shear materials (welds and plate) have shown them to be unsatisfactory for "military-case" structures (3,4).
The majority of 1-in. thick weldments received for
this investigation were generally larger than the 20-ii.
The appearances of samples 1-3 after one shot in
the explosion test are shown in Fig. 2. The num-
PERFORMANCE OF HY-80 WELDEMNTS OF
COMMERCIALLY-AVAILABLE SUBMERGED-ARC
MATERIALS
square size required fr
explosion tests.
Prior to
txrs shown on each weldment represent sample No.
bulge testing, the surplus material was re:nindv,1 for
and test temperature (top) and total number of shots
NRL studies of plate and weld metal properties. For
those weldments with sufficient material. the N[)'r
temperatures of the HY-80 als,) were establishted by
drop-weight tests of s-::-size specimens cut from the
plate surfaces. Table 1 lists the NDT temj)eratures
for 10 of the HY-80 plates investigated herein. As
expected for specificatiou quality HY-80, thu range
of NDT temperatures (-100' to -200 F) indivates that
all of these plates would exhibit 100if shear fracture
characteristics (i.e. highest possible resistance to
fracture) at cold water temperatures.
(lower rgLht). Of particular significance is the fact
that tach of these low J in. samples developed failures which were uniquely confined to the fusion line
of the weldments, as shown by the photomacrograph
in Fig. 2, B•ottom right. Similar failures in HY-80
weldments have been observed only whin austenitic
type electrodes (Mil 310-15) were used (6). The
extensive fusion-line failure which developed in the
sample modified with a crack-starter weld (Fig. 2,
Bottom left) indicates exceptionally low resistance
to fracture propagation, and, therefore, that the
materials and low J in. welding conditions which
were used arc conducive to the development of an
undesirable condition ini'HY-80 weldments.
The submerged-arc welding process is particularly
amenable to the use of welding conditions which result
in the rapid deposition of a large amount of weld metal.
However, in early tests HY-80 weldments made with
commercially ava.lable submerged-arc materials and
high J/in. " welding conditions (70,000 to 108,000.J ii.)
*The purpose is to develop a crack which results in
the catastrophic propagation of a fracture if the weld,
HAZ, or fusion line have tendencies for low energy
propagation of this crack. In the absence of such a
condition of weakness, short tears result indicating
desirable performance.
Samples 4-6 consisted of single V butt welds that
were described as having been fabricated with a
"special composition" of the company's commercial
submerged-arc products. Weld metal Charpy V
tests (Fig. 1. Bottom) indicated this deposit to be
brittle at 0°F and lower test temperatures. Brittleness of this weld deposit was also indicated by the
development of numerous, transverse (square-break)
weld cracks in th,.: explosion-bulge tests conducted
at 0OF (Fig. 3). The high resistance to fracture
(-170"F NDT) of the HY-80 plate is shown by the
complete refusal of the plate to propagate any of the
brittle weld cracks. Complete failure of these bulge
test samples occurred in each case on the Ist explosive shot principally via weld paths and separation
of the brittle weld at or near the fusion line.
'J/in. represents energy units per inch of weld expressed as Joules/in. and calculated from the welding conditions (Amps x Volts x 60 sec/rmin divided
by in./min travel speed). The J/in. factor is indicative of the heat generating condit!ons in welding
because 1 Joule is approximately equal to 0.24
calorie.
171
SUBMERGED-ARC WELDMENTS MODIFIED WITH
NOTCH TOUGH SURFACE WELDS
Earlier investigations demonstrated the feasibility
of using overlays of notch-tough welds to prevent the
initiation and propagation of brittle fracture in otherwise brittle, structural mild steels (7). In those
studies, crack-starter tests were used to indicate
that notch-tough overlays could be valuable in the
prevention of initiation or in stopping the propagation following short runs of the fracture. However,
these studies pointed out that practica! application
of the principles required consideration of the load
characteristics of the structure. Although the previous crack-arresting studies were not considered
for "military-case" applications, the principles involved were investigated by one industrial welding
equipment manufacturer and the Mare Island Naval
Shipyard (MINSY) as an alternate procedure to improve the explosion-bulge test performance of
submerged-arc weldments. Their techniques involved welding of the major portion of the weldjoint by various submerged-arc proct.dures and
completing the final surface layers of the weld
deposit with Bureau-approved, notch-tough eiectrodes (manual Mi!-11018 or automatic Mil-B-88).
Only two basic compositions were used for the
submerged-arc portion of the weld deposits in the
modified samples 7-22, inclusive. The submergedarc materials and welding conditions used for sampies 7-10, 2nclusive, were similar to hose which
developed low-energy-absorption Charpy V features
for the weld metal as reported in Fig. 1, Top, for
samples 1-3. Figure 4 depicts results of Charpy V
impact tests conducted with specimens cut from the
modified submerged-are weidments olbained from
MINSY. Low impact values were oliaimnd with the
specimens that were niachined so as to test the
submerged-arc (center) portion of the wvld (Fig. 4,
Bottom). In comparison, a considerable increase
in Charpy impact resistance was obtained with the
specimen-i
machined so that one side of the specimen
consisted of notch tough Mil-13-88 weld metal (Fig.
4, Top). Examination of the fractured Charpy specimens disclosed that shear-hps were developed at
various test :emperatures on the side of the specimen containing the notch-tough cover weld, Wut
were not developed on the side containing the
submerged-arc weld material. Similar Charpy V
test results were olained with the samples containing Mil-11018 weld metal overlays.
Table 2 summarizes the results oLKained in
explosion-bulge tests for the modified submergedarc weldments. Samples 7 and 9 were submitted for
test with the weld-crowns ground flush with the plate
surfaces. Failures did not develop until the 4th shot
for the ground samples, whereas failures occurred
on the 2nd shot for the companion samples, Nos. 8
and 10, tested with weld-crown intact. Such performance is as expected and is attributable to the
172
presence or absence of the geometrical notch at the
toe of the weld crown. However, the general mode
of failure was the same in all these samples (Nos.
7 to 10) which developed as fusion-line ruptures,
(Fig. 5, Top and Bottom left), similar to those previously described for samples 1-3.
The details provided in Table 2 indicate that 10
out of 12 modified submerged-arc samples showed
no visible failure signs after 3 shots in the explosion
test. In addition, the failures which developed on
the 4th shot on 4 out of 8 samples fabricated by
MINSY were generally minor weld-tears which did
not penetrate through the thickness, as shown in
Fig. 5, Bottom right. Such performance in the explosion bulge test would normally be considered
excellent, and these results appear to be superior
to the one-shot, generally extensive failures obtained
in previous tests uf almost all other submerged-arc
weldments.
In order to determine the type of fracture which
could be forced in such modified weldmentb, sample
No. 7 was subjected to an additional explosive shot
after the development of 'he initial, limited failure
along the fusion-line as shown in Fig. 6, Top. As
described previously, the submerged-arc portion of
this sample was characterized by low-energy shear
(i.e. not brittle) features. Figure 6, Bottom, indicates that extensive weld-metal fracture was developed under the explosive load conditions which
forced propagation of fracture in this sample. Accordingly, modified weldments which utilize materials that produce weld deposits of brittle or lowenergy shear characteristics are not desirable for
'military-case" structures, which require maximum
resistance to fracture.
PERFORMANCE OF EXPERIMENTAL SUBMERGED
ARC WELDS
In addition to independent efforts of the equipment
manufacturers, some of the shipyard laboratories
have been conducting Bureau-sponsored, extensive
development investigations aimed at improving
submerged-arc weld properties. Initial shipyard
approaches to the development of suitable submergedarc weldments involved the use of various flux materials with the Bureau-approved electrode (Mil-B-88)
that was developed for the consumable-electrode
inert-gas-shielded welding process. Because Charpy
V test results of such weld deposits were found to be
considerably inferior to those obtained with the inertgas-shielded process, subsequent investigations have
been coicerned with the development of new electrode
compositions in addition to studies of the effects of
new combinations of flux materials. Screening of
different combinations of submerged-arc materials
has been based upon Charpy V or drop-weight resuits of the various weld deposits.
Several combinations of different submerged-arc
materials have been reported to develop weld deposits
shot are illustrated in Fig. 8, left. Sample No. 26
shown in Fig. 8, top right, is representative of the
type of 3rd shot HAZ failures which developed in
three of the 40,500 J/in. weldments. The failures
in the above samples were limited in each case to
the plastically deformed (bulge) areas of the samples.
Out of thetwo 51,300 J/in. weldments which were
bulge tested, one sample had not failed after the 3rd
explosive shot while the other, after only 2 shots,
developed a HAZ failure which propagated even
through the hold down region, as shown in Fig. 8,
Bottom right.
that display nil-ductility-transition (NDT) temperatures below -100°F. These welds, however, have
invariably displayed such low Charpy V shelf characteristics as to be unsuitable for "military-case"
structures. Recent reports by an industrial research laboratory on a Bureau sponsored investigalion have indicated the development of new flux
materials which are successful in producing notchtough submerged-arc welds with high-energyabsorption Charpy V features (8). Another similarly
successful development of a notch-tough submergedarc weld deposit by an industrial shipyard is to be
described later. Bureau permission to investigate
the explosion-bulge test performance of submergedarc weldments made with two experimental electrodes
developed by an industrial and a naval shipyard was
received after shipyard results in drop-weight and
Charpy V tests indicated that these weld deposits
would be notch tough at cold water temperatures
(NDT values for weld metal lower than -60 F or
Charpy V higher than 20 ft-lb at -60'F).
Figures 9 and 10 illustrate the results of explosion tests of the 60,000 and 80,000 maximum J/in.
submerged-arc we~dments made with the experimental D'nawire 80S electrode and Gr. 80 flux.
Each of these samples developed complete failures
with the application of one shot. Fractures in these
samples appeared to start in the HAZ portion of the
weldments but were propagated predominantly in the
weld metal. The reason for modifying weldments
with a crack starter weld is shown in Figs. 9 and 10
by the results obtained with one modified sample in
each of these groups (weldments 34 and 39). The
disclIsure of such extensive paths of low resistance
to fracture propagation in a crack-starter modified
weldment should ordinarily preclude the necessity
or expense of conducting additional conventional
bulge tests with sim.lar weldments. As seen in Figs.
9 and 10, identical failures along similar paths of
low fracture resistance were obtained in conventional
bulge tests of the other 60,000 and 80,000 J in.
weldments.
For the electrode developed by the industril
shipyard (tentatively named Dynawire 80S), a cornmercially available flux material (GR. 80) was used
to study changes in fracture performance of a series
of weldments made with maximum .1 in. welding conditions which varied from approxomately 40, 000 to
80,000 J in. Representative weld-metal Charpy V
and drop-weight test results for these samples are
presented in Fig. 7.' The weid-metal Charpy spectmens of the 80,000 J in. welds did not develop fullshear fractures in tests at 60 F. The dotted portion
of this curve has been extended to the temperature
at which 100% shear fractures might be expected in
Charpy V tests of this weld deposit. Generally, the
Charpy. V results of Fig. 7 indicate that low-energyabsorption features were developed by each of these
weld deposits. Further, it should be noted that with
increasing Jjin. conditions, a small but perceptible
decrease in weld metal impact resistance was developed at all test temperatures.
The results described atxbe for weldments made
with similar materials are believed to show that a
progressive degradation of HAZ toughness properties
can be expected in HY-80 welded with increasing J/in.
conditions. In each respective series, failures appeared to start in the HAZ of the samples. It should
be recognized that extensive propagation of the fractures in the 60,000 and 80,000 J/in. samples was
sustained principally in the low-energy shear weld
metal. Htowever, the samples fabricated with the
higher Jjin. conditions developed ruptures involving
the HAZ on fewer shots, and, therefore, at lower
l,,vels of plastic deformation in the bulge test. Of
particular significance with respect to the early
ruptures in the latter samples was the fact that
numerous transverse fissures were revealed in
macroscopic examination of the HAZ portions of the
various bulge specimen fractures. The photomacrograph insert in Fig. 10, Bottom right, shows a
condition typical of that observed in the HAZ portions
of the fractures developed in all of the 60,000 and
80, OOU J/in. weldments. The weldments fabricated
with lower J/in. conditions were returned to the industrial shipyard before they could be sectioned and
examined for similar transverse crack indications.
The presence of such flaws would facilitate the development of early ruptures in explosion bulge tests.
Of the various explosion test samples investigated
herein, the prime plate areas of all except No. 28
(Fig. 8, Bottom left) exhibited high resistance to
propagation of fractures at 0•F. The HY-80 plate
material used for weldment No. 28 was shown in
Table I to display an NDT ot -100°F. The limited
tearing of this plate is indicative of borderline FTP
performance at 0'F. NRL experience and tests of
similar materials (9) have indicated that this plate
would display complete resistance to brittle fracture
(i.e. 100% shear) at all temperatures of 30'F or
higher.
Table 3 summarizes the results obtained in
explosion-bulge tests for the submerged-arc weldments made with experimental materials. Out of
six 40,500 J/in. weldments, one sample withstood
3 explosive shots with no visible failure indications,
and two of these samples which failed after the 2nd
173
-
A separate report is to be issued by the Naval
When little or no preheat is used in the welding
fabrication of high-tensile steels, moisture in electrode coatings, or on joint surfaces, or hydrogen
produced in the welding-arc atmosphere have been
shown to be important factors in the development of
HAZ fissuring similar to that observed in these
weldments (10, 11). It is possible that precautions
had not been taken in storage or during use to preelude moisture absorption in the flux materials used
to fabricate the above weldments. Because these
weldments were fabricated without preheat, it is
believed that the materials and procedures used resuited in the development of similar HAZ flaws irrespective of the J in. involved. However, the
rethermal effects of welding with the lower Jin.
suited in less degradation in HAZ properties and,
therefore, greater resistance to the irnitiation oi
failure in the bulge test than that developed in the
high J in. weldments.
As a result of the abxove test results, additional
submerged-arc weldments were not submitted for
NRL appraisal until research conducted by the industrial shipyard indicated that the same electrode
used with a new flux material developed considerably
higher Charpy V energy absorption values at all test
temperatures. Specific details concerning the weld
metal, flux compositions, and welding techniques
used were not provided by the shipyard. A small
sample received for weld metal evaluation purpose's
was described as having been fabricated with the
Dynawire 80S electrode and flux 'A" undef conditions
which developed a maximum 40,500 J in. and with a
200'F preheat and. interpass temperature control.
Concurrently. weld metal Charpy V and drop-weight
test results of air experimental electrode and flux
developed at the MINSY suggestedthat their materials
also might prove to be satisfactory for submergedarc welding of HY-80. Accordingly, samples involving three flux combinations and two J in. conditions,
as shown in Table 1, were received from MINSY for
weld metal and explosion bulge test evaluations.
The Charpy V properties of these experimental
submerged-arc weld deposits are summarized
graphically in Fig. 11. In each case, the fracture
appearances of the Charpy specimens indicated that
these weld deposits were not brittle at cold water
temperatures. Drop-weight tests conducted for
samples 41 and 44 revealed NDT temperatures for
these weld deposits of -140' and -160'F, respectively. ThuS, test results obtained with these experimental materials are superior to those of all
other previously tested submerged-arc welds. It
should be noted, however, that the MINSY weld deposits are still characterized by relatively lowenergy shear features (maximum Charpy V shelf
values of approximately 35 to 45 ft-lb). Weld metal
Charpy V and drop-weight test results of the industrial shipyard weld (No. 44), on the other hand,
are considered to be the best of any submerged-arc
weld metal previously tested at this Laboratory.
Weapons Laboratory concerning the generally excellent bulge test performance of 2-in. thick weldments prepared with materials and conditions similar to those reported herein for sample No. 44. The
appearances of the MINSY experimental weldments
after completion of the explosion bulge tests are
illustrated in Fig. 12. With the exception of one
transverse weld-tear developed in specimen No. 40,
the failures in each of these samples were confined
to the HAZ areas of the weldments. For the samples
welded with approximately 43,000 J/in. (Fig. 12 Top)
the ruptures in the HAZ developed on either the 2nd
or 3rd shot; however, propagation of these failures
were limited in nature In that they were confined to
the plastically deformed bulge area and did not penetrate through the full thickness of the plate. For the
53,000 J in. samples, complete rupturing occurred
on the 2nd and 3rd shot as shown in Fig. 12, Bottom.
Since similar conditions of preheat and interpas:-ý
temperature control were reported for all of these
samples, it is assumed that the higher J'in. conditions
contributed to the development of the more extensive
failures in the latter samples. It should be recognized,
nevertheless, that the performance obtained with
these bulge test specimens represents a considerable
improvement over that of previously tested submergedarc weldments which invariably developed complete
ruptures on one explosive shot.
SUMMARY AND CONCLUSIONS:
As is common for all high strength, Q&T steels,
the thermal effects of welding can develop microstructural conditions in the HAZ of HY-80 which display inferior properties to those of the unaffected plate
material. Conventional lalx)ratory test specimens
are unable to Indicate the degree of HAZ degradation
which may result from welding variables. Such degracation, however, is readily apparent in explosionbulge tests because the weld metal, heat-affected
zone, fusion line and base metal are all subjected to
an essentially uniform load field. Thus, variables
in welding conditions, or materials which develop
inferior regions in a given weldment can be measured
qualitatively by (1) the level of deformation (1,2,3, or
more shots) required for initiation of failure and (2)
by the extent of subsequent fracture propagation in
that region.
Test results obtained in this and previous investigations have consistently demonstrated that HY-80
weldments fabricated with welding conditions of approximately 40,000 to 50,000 J/in. develop an explosion test performance superior to that obtained with
similar materials welded with considerably higher
(60,000 to 80,000) J/in. In all of these investigations,
however, quantitative evaluations of the degree of HAZ
degradation developed by specific J/in. levels have
never been fully established because of complications
involving materials or other variables. In this investigation, for example, the complexities involving
174
!.0
It
HAZ fissures coupled with extensive fracture propagation in low-energy shear weld metal precluded
quantitative evaluation of the'degree of HAZ degradation developed in 60,000 and 80,000 J, in.
samples welded without preheat. Accordingly, it
should be recognized that the cumulative results
from shipyard fabrication experience and various
Bureau sponsored HY-80 weldability investigations
conducted by NRL and other shipyard or industrial
laboratories have been considered i;-. Oe preparatlion of the BuShips summary instructions of the
basic rules for welding of HY-80 (12). These instructions are aimed not Aily at alleviating problems concerning shipyard fabrication cracking in
restrained welds, but also aimed to minimize HAZ
degradation and provide maximum structural integrity of the end-product
explo.tive attack.
However, such techniques provide areas of low resistance to fracture propagation under conditions of
forced explosive loadings, and therefore, cannot be
considered satisfactory for use in "military-case"
structures.
(4) Results obtained for a limited number of samples in one test series corroborate previous data
concerning the influence exerted by the weld crown
which showed a detrimental effect of !he mechanical
notch present at the toe of a weld crown. With the
weld crown ground off, the samples required 4 shots
before failure developed; with the weld crown intact,
the samples developed failures on 2 shots. Even in
HY-80 welded with Bureau-approved electrodes, the
toe of the weld crown is a stress concentration point.
Although it would be impractical, if not impo-sible,
to completely grind flush all weld joints in a submarine hull, the influetnce of the geonmetry of the
weld crown should point out the desirability of training or cautioning the shipyard welders to "blend in"
the weld crown as much as possible, and especially
to avoid undercuts and highly convex contours.
ven under coniditions of
With respect to the various submerged-arc weldments investigated herein. the following general
conclusions are warranted:
(1) Test results presented herein indicate that
weld deposits of inferior notch toughness (can be developed by certain combinations of subinerg..d-arc
materials and techniques. A potential path of low
resistance to fracture propagation is afforded by
welds of either brittle or low-energy-absorption
shear characteristics. Such weld characteristics
at any service temperature precludes theni froni
being suitable for "military-case" applications,
where high resistance to fracture is required.
(5) Test results presented herein fo: experimental
materials have demonstrated the possiblity of producing high-strength !•tub-'rged-arc weld deposits
which are highly notch tough. Providing equivalent
results can be olbained consistently with production
heats of similar electrode arid flux materials, it is
believed that the toughness characteristics of weld
metal deposited by automatic submerged-arc processes will be equal to those currently obtained with
Bureau-approved electrodes.
(2) Specification quality HY-80 employed throughout this investigation was demonstrated to be highly
resistant to the initiation of fracture at 0 F test temperature (i.e. NDT of -100'F or lower for all plates).
Complete resistance to fracture propagation at 04F
was shown by all except one plate which was considered borderline at 0 F. Because of thermal effects of welding, the fracture resistance of HAZ
1. Pellin, W.S. and Eschbacher, E.W., "Investigation of the Performance of 1-In. S.T.S. Weldments of G260 and 25-20 Types, " NRL Memo
Report 191, July 1953.
areas of HY-80 plates was shown to decrease progressively as the J, in. were increased from approximately 40,000 to 80,000. Weldments fabricated
with very low J/in. (approximately 32,000) on the
2. Puzak, P.P. and Pellini, W.S. "ExplosionBulge Test Performance of 1-In. S.T.S. SemiAutomatic Inert-Gas Metal-Arc 'Weldments",
RKFEJREN"
NRL Memo Report 391, Nov.
other hand, developed extensive failures which were
uniquely confined to the fusion line of the weldment.
At least in part, the fusion line may have proved to
1954.
3. Puzak, P.P., "Explosion-Bulge Test Performance of Machine Welded 1 Inch Thick HY-80
Steel", NRL Memo Report 691, April 1957.
be the "weak-link" in these weldments because the
weld deposit displayed low-energy-absorption features and welding conditions resulted in little or no
degradation of tne HAZ.
(3) Test results obtained herein have demonstrated
conclusively that the notch-toughness properties of the
surface material is the controlling factor concerning
the initiation of failure in explosion-bulge tests. Accordingly, the use of notch-tough surface welds to
cover otherwise brittle or low-toughness weld deposits
can be expected to show increased resistance to fracture initiation in explosion tests of such weldments.
175
4. Puzak, P.P., "Explosion-Bulge Test Performance of Low Carbon Ni-Cr-Mo-B Quenched and
Tempered Steel", NRL Report 4919, May 1957.
5. Puzak, P.P. and Babecki, A.J., "ExplosionBulgeTest Performance of HY-80 Weldments",
NRL Memo Report 878, Dec. 1958.
6. Babecki, A.J. and Puzak, P. P., "Explosion Test
Performance of Small-Scale Submarine Huss Weldments", NRL Memo Report 996, Dec. 1959.
r
7. Puzak, P.P. and Pellini, W.S., "Crack Arresting by Overlays of Notch-Tough Weld Metal",
WELDING JOURNAL Research Suppl., Dec. 1955.
10. Arnold, P.C. "Problems Associated with the
Welding of T-I Material", WELDING JOURNAL
Research Suppl., Aug. 1957.
8. Martin, D.C. "Developments in Submerged-Arc
Welding Fluxes", Bureau of Ships Conference on
HY-80 Fabrication in Submarine Construction.
March 1960.
11. Smith, D.C. "Development Properties and
Usability of Low-Hydrogen Electrodes", WELDING JOURNAL Research Suppl., Sept. 1959.
9. Babecki. A.J. and Puzak. P.P. "ExplosionBulge Test Performance of 1-inch Thick Expertmental Low-Carbon S.T.S. Weldnients", NRL
Memo Report 883. Jan. 1959.
12.
Bureau of Ships Notice 9110. Ser 637-735
"Welding of HY-80 - General Rules for", 2
July 1958.
176
4
08
004Z
8::
*fl~~
n
'n
'n
'AU
00
0
000
.00
Go
c4
Lnv t-
C: C;
C:
-w-
ot
wC4r
o
4
n
'o
to
00
G)
in
c-4
*
coq
4 C. cm
I?
to ao
U4
w
U
0.
.
~v
r
0w-r
-. 0
0a
to
.jL
-
L)
I
7
W w
c
qr
I
w
-'
J
?c
w
a 61
SWW
W0
.nn
!'a.
to t
t-
U)I
k
qr Un 400
N
0n
c'~'n
'ome
D
o.0
NL
In
in
oýc
eD
8Ul
'
Go~
IV
In
1A
4. 4.
a*.
zo
OOOO~
000
00
00107
W
0
C
~O
2
3
n1I-
L.~
~~q-
Os
CuW.
z
.
-zzzz
U)
r
zw 'a.zraa2.zzra
.ca
ell
t- C-L
-C
-t-c
-t
-t
U
to
to Z~.
-4
"t.~.
A
O
.0
-cc
Cu
Cu10
Ads
Os .~S
0
z
Os178
S.
5.o
41
0.-
u
4
Ia.
AP
41)
00
4 Z0 .I.
h.ZZ~ a.
Ca.Z~
cNZZWZZW4aZa9N 4. z
.
Z~
I-
I
041
MW l
1.
4
'O4
S
41
a.
~
ol~~~~
Z
o- N
417
--
0 N-mNm-N
~
n
-w0N0
v
0
"
tc
-C-5
M
-
1
x.
O°F
OF N5
NQ4
s
WELD METAL QIARPY V
SUB-ARC NO& 1-3
40-
.20
No.6
WELD METAL QIHARPY V
OF
SUB-ARC NOS. 4-6
/
Z40
W
200
-80
I
-40
I
.
I
120
80
40
0
TEMPERATURE (OF)
160
Figure 3 - Explosion-bulge test samples showing
poor HAZ toughness and brittle weld performance.
Figure I - Charpy V notch test resullt of %elds of
commercially-available submerged-arc materials
exhibiting low-energy shear (top) and brittle fracture (bottom) at cold water temperatures.
OFNO. 2
NO. I
,'
OT
WEDMETAL CIIARPYV
40-
f
NWO
NO.3
WELD
-
F
Ox.G
CENTER WL.o.e
"°'•
,•
FL.
COVER WELO.MIL-0-88
0_J200
20-
1
-00
0
40
-40
TEMPERATURE (F)
60
Figure 4 - Charpy V curves of capped submergedarc welds. Results with duplex specimens machined
to include both weld metals (top) and with submergedarc weld metal only (bottom).
Figure 2 - Explosion-bulge test samples showing
fusion-line fracture at low levels of deformation,
180
-I
-
<
.-
3
0
0
/INL;,
O
60
40
WELD
•.... -
4D.00T~/j
-NDT
IIF)
4
• ... • "1 (-Soo-, VIN WELDo
-
20o-
-J
>'~60
-.
8O.OOOj/lOh)
WA4f
7--
0
0
-40
TEMPERATURErF)
Li20 -80
40
80
NO.3
CharPY V notch test results of expertat three
weld metals depOsited
levels Of heat input
mental sub-arc
Fi~re 7
N0.24
NO28
0i:
-
OFN0.26
O°F
F N 32Figure
••
2-in.
3 - ExplosiOn -bulge test samples of
one
from
sub-arc weldments
thick experimental
cracka
with
weld,
No. 39 modified
starter
source. Sample
samples frofl one
Explosion-bulge test tom
8
Figure
metals
rigt,./n
weld
sub
uremental
with expert .ae nlasub-arc
so)urce
/
g
51,300 bo
welded at 4(0,500 and
'I
181
1t
F
..
L8
9
-4
Figure 5 -Explosionbulge test samnples wvitn modified sub-arc welds (capped with Bureau-approved
weld metal). Sample No. 9 was tested with weld
crown removed.
Figure 6 - Explosion-Uul~g test sample va ntlmodified sub-arc weld. Weld crown removed and tested
to visible signs of failure (top), and under conditions
of forced propagation (bottom).
182
A
'
N
F
NQ42
Figure 10 - Explosion-bulge test samples of 2-in
thick experimental sub-arc weldments fronm one
source. Sample No. 34 modified with crack-starter
weld. Photomacrograph insert on Sample No. 37
shows numerous transverse fissures in HAZ.
Figure 12 - Explosion-bulge test samples with experimental sub-arc weld metals from MINSY prepared with 43,000 (top) and 53,000 (bottom) J/in.
SUBMERGED ARC WELO MTAL. CHARPY V RESULTS
0-
40
[3*
-*.•
-
NO 42-44
20
-I
-,20
-S0
40
0
T'1 P(RATURE
40
O
,*FI
Figure 11 - Charpy V impact test results of experimental sub-arc weld metals received from MINSY
(Nos. 40-43) and an industrial shipyard (No. 44).
183
I
...
NOTCH TOUGHNESS EVALUATIONS OF MODIFIED HY-S0 STEEL
IN HEAVY GAGE PLATES
by
A.J. Babeckl and P.P 'Puzak
ABSTRACT
Welding fabrication experience with low-alloy
H-Y-80 in plate thicknesses of 1-1,4-in, and under
has been good even without strict welding controls.
The higher alloy content of heavier gage HY-80 required strict compliance with stipulated welding
procedures for trouble-free fabrication. The added
expense and production rate reduction resulting
from such tighter welding controls prompted a request for special Bureau of Ships permission to
specify the low-alloy composition for HY-80 in
thicknesses of 2-1, 2-in. or more. The request
was denied, but it led to a cooperative test program,
involving the Naval Research Laboratory, the Naval
Proving Ground, and the Philadelphia Naval Shipyard, which was to examine more rigorously the
question of low notch toughness in heavy-gage lowalloy HY-80.
Low-alloy HY-80 in plate thicknesses of 1-1/4,
1-3,4, 2-1/2, and 3-in. was procured from two
suppliers and was to be given extensive evaluation.
However, preliminary drop-weight test results
showed that the notch toughness of the low-alloy
HY-80 was seriously and progressively impaired as
plate thickness increased. This conclusion was
unequivocally confirmed by the results of crackstarter explosion bulge tests of full-thickness butt
weldments which showed the 2-1/2 and 3-in. plates
to be brittle at 0' F and, therefore, unsuitable for
submarine application.
184
*
"S
INTRODUCTION
MATERIALS AND TEST PROGRAM
Of all high strength structural steels which have
been investigated, the low-carbon quenched and
tempered Ni-Cr-Mo steel known as HY-80 has exhibited the best optimum combination of weldability,
notch toughness and high energy absorption characteristics. Because of these characteristics, HY-80
has found increasing applications in many critical
welded structures. In order to provide the hardenability required to develop the desired level of
toughness equivalent to that of 1-1/4" thick plates
in heavy section plate, the composition limits specified for Ni, Cr, and Mo were increased as specified
in Table 1. The composition ranges for what is now
commonly referred to as "low-chemistry" HY-80
and "high-chemistry" HY-80 relate specifically to
the requirements for plates below 1-1, 4-in. and
above 1-3/8-in. respectively. The fabrication experience with low-chemistry HY-80 structures has
proved to be relatively free of welding difficulties.
In many cases where welding fabrication of a given
structure with conventional steels was known to have
required numerous weld crack repairs, it has been
reported that the same structure made with lowchemistry HY-80 was invariably found to be crackfree. Precautions concerning moderate preheat,
electrode moisture control, etc. stipulated in the
Bureau welding instructions for HY-80 are given in
reference (1).
Mill inmpection reports and laboratory check
analyses have invariably shown that 1-in. thick
HY-80 plates generally conform in chemical composition to the very low side of the composition
ranges specified for Ni-Cr-Mo contents. Accordingly, for this investigation, the suppliers were
requested to provide low-chemistry HY-80 material
into plates of 1-1/4, 1-3/4, 2-1/2, and 3-in. thicknesses. The Ni-Cr-Mo contents for all of these
plates were requested to be aimed at the lean side
of the specified low-chemistry HY-80 limits for
these elements (defined as between the mean and
lowest value). One full size plate of each i.iiii:kness
was obtained from each of two steel mills (coded
Y-Company and Z-Company). Although described
herein as nominally 3-in. thick, the heaviest plate
from Y-Co. measured 2-13/16-in. because
machining from one side was allowed to remove
surface defects and to comply with flatness requirements. As can be seen from the data given in Table
2, only the Mo contents of the Y-Co. plates fell on
the low side of the lean analysis range. With the
above exception, the Ni-Cr-Mo contents of all
plates generally fell near or above the requested
maximum aim-value. Thus, test results are considered to be representative of the properties one
can expect in heavy section plate for the meancomposition of tow-chemistry HY-80.
Experience with transition from low-chemistry
HY-80 construction to high-chemistry HY-80 construction required by the use of increased thicknesses of material (1-in. and under to 2 to 3 in.)
indicated that deviations from stipulated welding
procedures resulted in cracking and uther difficulties.
Time and cost considerations oased on fabrication
difficulties with 2 and 3-in. thick HY-80 at a major
industrial shipyard resulted in a request for special
Bureau permission to specify only low-chemistry
HY-80 for use in a new construction in which the
minimum plate thickness was 2-1/2-in. This request was denied by the Bureau after a review of all
available data for light and heavy gage HY-80 indicated that the proposed change was potentially
dangerous for submarine service in that notch
toughness properties would be sacrificed with no
assurances that fabrication would in fact be less
difficult with heavy-gage, low-chemistry HY-80
material. In order to examine this question more
rigorously BuShips procured material to investigate
properties and weldability characteristics of lowchemistry HY-80 in heavy gage plates. A cooperative test program between the Naval Research
Laboratory, Naval Proving Ground and Philadelphia
Naval Shipyard was established by BuShips (Code
637). NRL was requested to participate in and
coordinate this test program.
DROP-WEIGHT AND EXPLOSION BULGE TEST
RESULTS
The drop-weight test (2,3) represents .Asimple,
laboratory impact-bend test which determines the
highest temperatures at which even minute amounts
of deformation in the presence of a sharp crack
cannot be tolerated by a given steel without incurring
brittle failure. This temperature is defined as the
nil-ductility transition (NDT) temperature. The
significance of NDT to possible initiation of brittle
service failures, and the validation of NDT concepts by correlations with numerous ship and nonship service failures have been previously described (4, 5). The Charpy V-notch test requirements for specification quality HY-80 were
established by correlations with NDT and are aimed
at insuring that minimum acceptable quality HY-80
displays an NDT of -100' F or lower. Tests of
numerous HY-80 plates to date have indicated that
the NDT temperatures of normal mill production
ranges from -100' to -150'F for low chemistry and
from -130'F to -180°F for high chemistry HY-80.
Figure I illustrates the severe reduction in
notch-ductility properties which occurs in these
lean analyses HY-80 materials with increasing
thicknesses. The NDT temperatures of the 2-1/2
"185
4
and 3-inch thick plates are raised approximately 80'
to 100" F above the values obtained for 1-1/4-in,
thick plates. With the exception of the low Mo content for the Y-Co. plate, the 1-1/4-in. plates conform to low-chemistry HY-80 composition requirements, and NDT determinations of approximately
-100' F were developed in these plates. At each
thickness level, the moderately lower NDT's of the
Z-Co. plates than that for the Y-Co. plates are
considered to reflect the differences in chemistry
between both materials. In general, the data appear
to indicate that 1-12-in. and heavier plates which
are melted to the low side of Ni, Cr and Mo contents
for low chemistry HY-80 could be expected to exhibit NDT temperatures higher than - 100 'F. It is
also suggested that plates with compositions at the
high-end of the low-chemistry HY-80 specification
ranges for Ni, Cr, and Mo would result in NDT
temperatures of - 100 F or lower in plate thicknesses
of 1-1, 2 to 1-3, 4-in.
The extensive experience in explosion bulge (6)
techniques gained by NRL has generally been limited
to plates of approximately I-in, thickness. In the
conduct of these tests, either the explosive charge
or the standoff distance is adjusted so that approximately 3 to 4 percent thickness reduction of the
plate at the apex of the bulge is developed irrespective of the specimen thickness or yield strength of
the material being tested. Prior to this investigation, only one carbon-steel of 1-3/4-in. thickness
had ever been tested; 2-1,2 and 3-in. thick plates
had never previously been explosion bulge tested.
In the absence of established test conditions for
these heavy gage specimens it was necessary to
resolve questions concerning test conditions by a
"best-effort" approach. One of the 2- 1/2-in,
specimens prepared with steel from each mill
was used for initial tests and the data obtained
from the two shots given these samples were used
to approximate the conditions necessary for all
other specimens.
Proving Ground, Dahlgren, Va. Table 3 gives the
test conditions employed herein for all samples.
To compare results of these specimens with
previous HY-80 bulge tests, it was attempted to
conduct all tests at 0° F. However, difficulties
involved in the handling of these large specimens,
and the time-lapse between removal from the
temperature conditioning room and firing of each
plate, resulted in some plates being slightly higher
than 0' F. Figures 2 and 3 illustrate the appearances of the 1-1/4-in, samples after tests at 100
and 20' F. Complete resistance to fracture with
no evidence of HAZ deficiencies were obtained at
these test temperatures for these weldments. Significant differences in notch-ductility between
the Y-Co. and Z-Co. plates of 1-3/4-in, thickness
were indicated by drop-weight test results. As
expected, more extensive plate metal cracking was
developed in the Y-Co. plates which were explosion
tested at 10' and 20" F, as shown in Figs. 4 and 5.
It should be noted, however, that these fractureb are
only partially brittle with significant surface shearlips (1/4-in, or more). The smaller amount of
cracking developed at the higher (20' F) test temperature indicates that these specimens were tested
at temperatures very close to their FTP temperatures where complete resistance to brittle fractures
could be expected. In view of the notch-ductility
characteristics as determined by the NDT's for
these plates, such fracture performance is in complete agreement with expected results.
For the initial tests conducted at 20' F with the
2-1/2-in. thick weldments, the charge weight and
offset distance used did not deform the plates adequately to permit proper functioning of the crackstarter weld. Figure 6 illustrates the appearances
of these plates uoon completion of the second shot
given at a 0"F test temperature, Of special significance, it should be noted, is the fact that failure
is developed predominantly via plate metal fractures. Essentially all square breaks with no
visible shear-lips were developed in sample Y-6,
however, small surface shear lips were visible in
the plate metal fractures of sample Z-5. Similar
observations apply equally as well to the duplicate
samples of 2-1/2-in. thickness which were given a
single shot at 00 F, Fig. 7; however, the chargeweight used for the latter did not develop the
expected thickness reductions. Future tests of
80,000 psi yield strength plates of 2-1/2-in. thickness should be tested with 36 to 40 lb of explosive
and a 15-in. standoff distance.
As contrasted with conventionad bulge tests which
require a repetition of shots until failure is developed, the use of a crack-starter weld reduces
the overall time and costs since only one shot is
used to complete the test. Such techniques are
useful for screening purposes and have proved
successful in the past to delineate the possible
fracture paths of least resistance in a weldment.
Duplicate weldments with steel of each thickness
from each mill were prepared by the Philadelphia
Naval Shipyard. These were all welded under
normal shipyard conditions with the Mil-110-18
stick electrodes that are approved for welding of
HY-80. To expedite the tcst program and to provide the maximum amount of information with the
fewest number of full thickness weldments, crackstarter welds were added to the center of each specimen. The bulge tests were conducted by Naval.
The test conditions used herein for the 3-in.
plates also happened to be lighter than that considered standard for bulge tests. However, the
inferior notch-ductility properties of 3-in. plates
from both mills (NDT of -10' F and +20' F) resulted
186
2. Explosion bulge tests of weldments corroborated the results of drop-weight studies. In the
bulge test samples which were thicker than 1-1/4-in.,
failure always occurred via fractures of the plate.
Specification quality Mil- 110-18 electrodes were
employed, and no evidence of weld fracturing was
observed for the test temperatures studied (0' to
20" F). The 2-1/2 and 3-in. t! ck weidments
exhibited extensive brittle plate i actures
in extensive f,,',lures via brittle plate-metal fractures, as shown in Figs. 8 and 9. In all cases, the
fractures were essentially square-breaks with no
visible indications of surface shear lips. The
proximity of some of the cracks to the HAZ is believed to be only the result of general brittleness of
these plates rather than HAZ deficiencies developed
by welding.
SUMMARY AND CONCLUSIONS
3. Test results with the Y-Co. plates which were
low only in Mo content suggest that material on the
low side of the low-chemistry HY-80 composition
range would probably be suspect in thickness of
1-1/2 or 1-3,4-in, because of inferior toughness
properties. However, results obtained for 1-3/4in. thick p!ates indicate that consideratin of a
chemistry range intermediate to those presently
specified for HY-80 would be warranted for plate
thicknesses between 1-3 8 and 2-in. Providing
future HY-80 construction involves these thicknesses to such an extent as to warrant an intermediate chemistry HY-80, it is believed that a
composition range for Ni. Cr and Mo between 'he
mean-values now specified for light-gage and
heavy-grage HY-80 would develop adequate notch
toughness in thicknesses up to 2-in. Prior to
general acceptance of the above, it is believed
necessary to confirm these deductions by tests of
material specifically melted to such compositions.
It should also be noted that normal mill production
of low chemistry HY-80 has generally been found
to confirm the low-side of the composition ranges
specified for Ni, Cr, and Mo contents. It is believed, therefore, that to obtain plates with composition at the high-end of the chemistry range, a
special procurement order such as was necessary
to obtain the subject lean-analyses plates will be
required.
The principal aim of the investigation was to
explore the possibilities of altering HY-80 plate
chemical composition in thick sections without
materially reducing specification requirements for
physical, mechanical, toughness and weldability
characteristics. An extensive test evaluation program was originally planned for these materials;
in addition to drop-weight tests of surface and
centerline thickness specimens to evaluate hardenability characteristics, and explosion bulge tests
of full thickness weldments, conventional tensile
and Charpy V-notch test evaluations, and explosion
crack-starter tests of plate were also planned.
However, the results obtained in drop-weight and
explosion bulge tests were sufficiently informative
as to indicate that extensive testing was mit warranted for these particular heavy-sectiuo lean
analysis plates. These results are summarized
briefly as follows:
I. Drop-weight test results indicate that a
significant decrease in notch toughness is developed by the low-chemistry HY-80 as plate
thickness is increased from 1-1, 4 to 3-in. For
2-1/2 and 3-in. thick plates, the loss in toughness
is so great that brittle plate fractures would be
developed by explosion loadings at water temperatures. Consequently, such material is not suitable
for submarine hull structures.
187
REFERENCES
I.
Bureau of Ships Notice 9110, Ser 637-735, "Welding of HY-80 - General Rules for," Jul 1958
2.
PNllini, W.S., Puzak, P. P., and Eschbacher, E .W.,
Memo. Rept. 316, Jun 1954
3.
Puzak, P P, aind Babecki, A.J., "Normalization Procedures for NRL Drop-Weight Test," NRL Report
5220, Nov 1958
4.
Puzak, P. P., and Pellini, W.S., "Evaluation of the Significance of Charpy Tests for Quenched and
Tempered Steels," Welding J., 35:275-s (1956)
5.
Puzak, P. P., Babecki, A.J., and Pellini, W.S., "Correlations of Brittle Fracture Service Failures
with Laboratnry Notch Ductility Tests," Welding J., 37:391-s (1958)
6.
Pellini, W.S.,
"Procedures for NRL Drop-Weight Test," NRL
"Use and Interpretation of the NRL Explosion Bulge Test," NRL Report 4034, Sep 1952
188
TABLE I
Chemical Composition of HY-80 Steel (Mii-S-16216C)
C
Nominal Thickness*~
max.
Mn
.Mx
P
S
Si
a. Max. %,• max.
%,
It
-
-
Ni
Cr
__5
%
___
%-
Mo
%
Up to l-3/8-in.
inclusive
0.22
0,10-0.40
0.035
0.040
0.15-0.35
2.00-2.75
0.90-1.40
0.23-0.35
Over 1-1/4-in.
.23
.10-0.40
.035
.040
.15-0.35
2.50-3.25
1.35-1.85
.30-0.60
• For thickness between 1-1, 4 and 1-3 8 inches, either comrmxsition may be applied.
TABLE 2
Chemical Composition of Lxan Analysis HY-80 Plate Materials
Mill
Y-Co.
Z-Co.
S
Si
Ni
Cr
Mo
(,)
(()
(,p)
(6)
(--.)
2.00-2.38
.90-1.25
.23-29
C
Thickness
(f•)
Requested Aim
0.22 Max.
.10- .40
.035 Max.
1-1,4"
0.14
0.21
0.014
0.022
0.14
2.32
1.24
0.18
"1-3 4"
0.16
0.23
0.614
0.025
0.14
2.41
1.46
0.19
0.14
"2-1/2"
0 19
0.013
0.019
0.13
2.25
1.18
0.16
0.19
0.21
0.018
0.020
0.14
2.21
1.33
0.23
0.17
0.18
0.006
0.034
0.17
2.26
1.13
0.30
0.14
0.18
0.18
0.004
0.019
0.17
2.24
1.16
0.30
0.006
0.028
0.19
2.34
1.17
0.32
"2-3,4
Mn
(r
(,r)
P
-
Plate
(
.040 Max.
15-35
1-1,4"
"1-314"
"2-1/2"
"2-3/4"
0.18
189
TABLE 3
LEAN Analysis HY-80 Explosion Bulge Test Data
Specimen
No.
Plate
Thickness
Est
Temp
Chge.
Size
Stand
Off
No. of
Shots
Ck. St.
Orient.
NDT Temperature
Surface
Centerlinr,
Y- 1
1- 1/4"
+ 1OF
12#
19"
1
L
-90
Y-2
1-1, 4"
+ 20F
12#
19"
1
T
-90
Y-3
1-3,'4"
+ IOF
240
17"
1
L
-80
-70
Y-4
1-3, 4"
+ 20F
240
17"
1
T
-80
-70
Y-5
Y-6
2-1, 2"
2-1, 2"
OF
+20F
32#
32#
15"
20"
1
T
-10
-30
1
L
-10
-30
OF
32#
18"
1
L
Y-7
2-13, 16"(3")
OF
360
15"
1
L
+30
+10
Y-8
2-13, 16"(3")
OF
36#
15"
1
T
+30
+10
Z-1
1-1, 4"
+20F
12#
19"
1
T
-110
Z-2
Z-3
1-1, 4"
+1OF
120
19"
1
L
-110
1-3, 4"
+20F
24#
17"
1
T
-140
-110
Z-4
1-3, 4"
+ 1OF
24#
17"
1
L
-140
-110
Z-5
2-1, 2"
+20F
32#
18"
1
L
-50
0
OF
32#
18"
1
L
Z-6
2-1/2"
OF
32#
15"
1
T
-50
0
Z-7
3"
OF
36#
15"
1
L
-10
-20
Z-8
3"
OF
36#
15"
1
T
-10
-20
T - transverse to weld
L - parallel with weld
190
A
-izoV
FmN~I____
Figure -Variation of drop-weight NDT temperature with plate thickness of low-alloy HY-80 steel
procured from two sources. An NDT temperature
above -100°F indicates a notch ductility too inferior
for submarine applications.
20F
II-Z
1L 19"SlO.
,12
Figure 2 -- Crack-starter explosion-bulge test plates of 1-1/4-in, thick low-alloy BIY-SO slowing
good resistance to brittle fracture at 20* F.
191
II7"SO.
24 LB 1711S.O0.24B1so
Figure 4 - Crack-starter explosion bulge test plates of 1-3/4-in, thick low-alloy HY-BO. Plate
on left exhibit~s extensive brittle fracture at 10*F; plate on right shows superior performance.
192
13W-Y4
20OF 13'4-Z3
24 LB 17ISQO
2O
~j7"S..O 24
Figure 5 -Duplicate
1-3/4-in. thick crack-starter explosion-bulge test plates.
plate again possesses the better notch toughness.
B 1
The Z-Ccmpany
21/2"-Y620OF 21/2"-Z5
32LB 21 S..3LB
32LB 1^".O.
18 SOQ.
3LB l8"S.Q
Figure 6
explouion-bulge test plates of 2-1/2-in, thick low-alloy HY-BO given
two test shots. Both plates developed extensive brittle fractures in the plate metal on the second.
shot.
-Crack-starter
193
00 F 2.V2"-Z6r
ZW-Y5
32 LB 15"1SO.a
Figure 7
-
2L
5IS
Duplicate 2-1/ 2-in, thick crack-starter explosion- Olage test platesI exhibiting
brittle fracture at 0OF.
136 LB 15"S.O.
'36 L.B 15S.
Figule a - Ciki&~arctz dA~iLItulca-IAAIge tobt plates of 3-In. thick luw-alloy HY-80 which
developed extensive brittle fractures at 07F.
194
3'Y
0
F
0
3"-Z80
36 LB 15"SO
~6
Figure 9 - Duplicate 3-in, thick crack-starter explosion-bulge test plates.
brittle fracture.
195
L
15"SO
Failure again is by
EFFECT OF WELDING VARIABLES ON THE YIELD STRENGTH
OF MIL-9018 WELD METAL
by
Wayne L. Wilcox
Arcos Corporation
Philadelphia
tained in accordance with Federal Test Method Standard No. 151, July 17, 1956.
Abstract
The results of this investigation, conducted to determine the causes of yield strength variations between yard acceptance tests and the manufacturer's
quality control tests of MIL-g018 electrodes, show
the effect of the use of stringer beads and low preheat and interpass temperatures on increasing the
strengthof the weld metal in both the as-welded and
stress relieved conditions,
Selection of electrodes
The particular lots of MIL-9018 electrodes used
were selected on the basis of availability and the fact
that considerable information about them had been
obtained from previously conducted standard quality
control tests. This information, shown in Tables I
and II, was required to establish the electrode's conformance with the speification and was further useful for correlation with the present test results. Except for nickel, the weld metal chemistry of the two
brands of 9018 electrodes tested was strikingly similar. This, however, was not a consideration in their
splection.
Introduction
Yield strength values for MIL-9018 weld metal in
the as-welded condition have differed by as much as
20,000 psi between yard acceptance tests and the
manufacturer's lot inspection tests conducted in accordance with the provisions of the MIL-E-19322A
(SHIPS) specification.
The MIL-11018 electrodes used were selected
also on the basis of availability; they had been recently produced and, again, previoustest information
was available for comparison purposes. Unfortunately, 5/32" electrodes, with the same comprehensive background information, could not be obtained
in time for these tests and it was necessary to substitute the 3/16" diameter.
It is generally recognized within the welding industry, although sometimes overlooked, that variations in welding conditions and procedures can alter
the properties of the resulting weld metal - particularly low alloy steel weld metal. This investigation
was conducted in an attempt to determine, on the
yield strength of MIL-9018 weld metal, the magnitude of the effects of welding variables within the
range of those currently encountered in electrode
testing and subsequent fabrication use.
The weld metal chemistry for each of the lots
of electrodes tested is shown below in Table I and
the mechanical properties in Table 11.
Selection of base metal and Joint type
TEST PROCEDURE
T-Steel was selected as the base metal because
of its strength and the fact that it is, for these electrodes, permitted for the inspection plate test in the
electrode specification.
Selection of test conditions
Eleven test plates were welded with two different brands of 5/32" MIL-9018"electrodes and one
brand of 3/16" MIL-11018 electrodes, using procedures encountered in actual submarine hull construction and those permitted for the lot inspection
tests of electrodes required by specification MILE-19322A(SHIPS) for quality assurance purposes,
The selection of the joint design and dimensions
for the test model was based on the desire to provide
some restraint and, with a minimum of welding, to
obtain, in addition to sufficient weld metal for the
required tensile and charpy specimens, some indication of electrode usability. The latter was the
initial purpose of the "Inspection Plate Test" in the
specification. Each joint was intended to provide
two tensile bars and ten charpy.Xwei/menis from
which both as-welded and heat treated mechanical
properties could be determined. The test model is
shown in Figure L A cross section through the
groove, in a test plate ready for welding, would be
In the present tests all electrods were used
just as they came from their containers without
rebakfg. Mechanical testing of the weld metal was
conducted on .505" tensile bars in both the as-welded
and heat treated conditions. Specimens for the "asWelded" tests were machined and tested immediately upon completion of welding and were not given the
benefit of aging. The tensile data reported were ob196
TABlLE I
9O1801)
C
Mn
Si
S
P
Cr
Ni
Mo
V
Weld Metal Chemisatry
9old(2)
.046
1l018 (3)
.05
,-,8
.37
.015
.016
.12
2.16
.37
.00
.052
.96
.30
.012
.013
.12
1.67
.29
-. 01
1.78
.4;j
.010
.023
.16
2.20
.40
.00
(1) Lot OOBI1A, Weld Pad
(2) Lot 0501, HT 79L432, Outside Report - details unknown to us.
(3) Lot #OA30B, Weld Pad
TABLE H1
Weld Metal Mechanical Properties
Material
Y.S., Ksi
T.S., Ksi
Elongation
%2!nches
Reduction of
Area- %
9018(l)
A.W.
83.7
96.6
24.0
63.8
9018(2)
A.W.
91.0
100.5
23.0
59.8
11018(3)
A.W.
H.T.
108.1
103.0
122.7
111.6
21.0
24.0
63.5
65.8
(1) Lot #0BlIA - MIL-E-19322A Restrained Groove
(2) Lot #501, HT 791,432, Outside Report - specific test details unknown to us
(3) Lot #OA30B - MIL-E-19322A Restrained Groove
A.W.:
H. T.:
As-welded
1150° F 2 hrs. F.C.
--
01~
---
-T
FIGURE I
Combination Restrained
Groove-Inspection Plate Test
197
very much like that of the inspection plite in MIL-E19222A(SHIPS) and would be of proper root dimension, including the back-up strip, for a 7/32" electrode. This cross section and tensile bar location is
shown in Figure II.
D.C. Motor generatorsand rectifiers withdruoping characteristics were used to supply the welding
current. Amperages were checked periodically with
a tong meter and near the completion of the tests individual voltage and amperage readings were taken
for each weldor, on recording Esterline-Angus meters.
The selection of the actual welding currents used was
based on weldor preference within the requirements
of the specifications. Travel speed readings were
taken during the tests to form a basis on which to
compute heat input.
All plates were allowed to air cool to the required interpass temperature and to room temperature on the completion of welding.
3
Each joint was saw cut in half to provide the
specimens required for testing in the as-welded and
heat treated conditions. The heat treatment applied
was that specified in MIL-E-19322A(SHIPS). The
specimens requiring heat treatment were all treated
at the same time in the same oven.
,
I0S,
A /2
The .505" tensile specimens were prepared from
Yield
plate by saw cutting and machining.
strength determinations were made in accordance
the extension-under-load procedure of Federal
Test Method Standard No. 151.
4each
4with
Test Results
In both the as-welded and heat treated conditions,
considerable variation was encountered in weld metal
tensile properties as a result of changes in welding
position, wiith of weave and preheat and interpass
temperature. The overall range being 17,000 psi in
tensile and 2!,000 psi in yield strength in the aswelded condition for the MIL-9018 electrodes of Lot
-01B3A. In the heat treated condition the range was
about 15,000 and 14,000 psi respectively.
FIGURE II
CrossSectionof Test Joint and locationof Tensile Bar
Preparation of test joint and weld specimens
A comparison of the results of tests B and C ir,
Table IIIshows that some reduction in weld strength
does accompany the use of higher preheat and interpass temperatures. The trend is continued in test
Q4533-D of Table IV. Test B was considered an
example of an accepted fabrication procedure whereas C and Q4513-D, in addition to test H represent
procedures permitted in welding the "Inspection plate
test" ini MIL-E-19322A(SHIPS) from which the weld
mital charpy values are obtained.
in presawing were
cutting and
Flame
the various
components
of theemployed
test model.
All
paring
mating flame cut surfaces and those on which weld
metal would be deposited were hand ground to remeta
deositd
wold b wre hnd roun toremove the oxide layer prior to fitting and tacking.
the results of tests B, D, and
examination
III and C of
of Table IV with E and G of Table
F of An
Table
III and A and B of Table IV shows clearly the higher
strength of stringer beaded weld metal irrespective
of the position of weldingand the amperage used.
Preheat was applied with an oxyacetylene flame
and throughout the tests tempil sticks were used as
temperature indicators,
Tests H and I of Table III, to which lot #501 MIL9018 electrodes (a different brand from lot OOBIIA)
were subjected, involved only a difference of 750 F
198
1ADJEJ&IL
Table
/
in
~~A:
0 ~t
Hr.
,~~,r
!
/
/
'
47
0$/
,
41r
B
21-22
160
80r1
P
'
4-
-
150
6.2,9.0 33.2,23.0
11
5 32"
(2)
S
125
H
5 32"
(2)
S
20C 500
21-22 160
9.6
21.5
V
5 32"
(2)
S
125
150
21-23
120
4.3
36.8
V
5 32"
(2)
WAD
125
150
22-23 130
1.7
103.0
F
5321"
(2)
S
125
F
5 32"
(2)
W- 4D
200 250
22-24
175
H
5 32"
(3)
S
200 500
21-22
165
H
5 32"
(3)
S
125
21--22
160
H
3 16'
(4)
S
200
!,9-20
205
F
3 16"
(4)
W2-I 2D
125
19-21
2415
V
3 16"- (4)
W-4Di
200
19-20 155
Ht.
D
lt.
E
Ht.
F
!502
165
0-25
4.4 31.5
Ht.
6
5.1
47.2
H
lit.
1
1-11
.1
fit.
K
Ht.
L
(1)
(2)
(3)
(4)
(5)
94.0
88.
92.4
88.7
85.5
99.7
92.9
86.2
88.7
91.8
89.0
78.3
79.6
Hit.
92.7
92.2
7.1,6.0 29.1, 33.4 106.2
100.2
126.8
30.4
7. 7
116.2
109.5
46.0
6.4
107.1
111.9
56.6
3.2
7.9
27.0
'
4V
0,--
Hr.O
Ht.
C
/
,
/
/
494.4
*
.
• • ,
N
101.5 21.0 63.6'
987 2•,057.
98.0'23.5
61.9
93.7 28.0 69.8
92.5 28.0 68.2
108.5 19.0 56.4
101.5 23.0 63.5
101.6 23.5 63.5
98.7 26.0 57.1
97.2 23.5 65.6
91.0 25.5 65.1
91.2 24.0 65.7
86.2 28.0 69.8
100.1 26.0 67.8
99.4 26.0 63.6
113.2 20.0 50.7
109.2 24.0 62.1
130.3 13.0 55,0
121.8 22.0 57.6
126.4 16.5 37.7
114.7 24.0 64.5
125.8 4.0; 19.8
Base Metal: T-Steel
MIL-9018 Lot #0B1IA
MIL-9018 Lot #501 HT #79 L 432
MIL-11018 " #OA30B
Welded without regard for interpass temperature values given indicate actual plate temperature
Table III
199
TABLE IV
COMPARISON OF WELD PROPERTIES( 1 )
-//----
f
/
Q4533
A Restrained
Groove
B
C
125
125
W-2-1/2D
83.7
96.6
Y.;
63.8
125
125
W-2-1,2D1
84.2
96.2
25
66.4
200
125
S
97.0
103.9
21
61.6
200
(2)
S
79.0
93.0
27
64.3
Inspection
Plate
D
12758
'
B
(3)
125
150
S
94.0
101.5
21
63.6
D
(3)
125
150
S
99.7
108.5
19
56.4
G
(3)
200
250 •W
78.3
91.2
24
65.7
1D
(1) All electrodes 5 32" MIL-9018 Lot No. OBILIA
welds are in the as-welded condition.
(2) Unrestricted. Welding progressed without regard for
interpass temperature. Much of the time the plate was
above 600 F - never below 500 F once reaching that
temperature.
(3)
Combination plate nut in specification.
200
Test Results .cont'd)
in preheat temperature but as much as 375° F in interpass temperature, produced strength level differences
of about 13,000 psi in both tensile and yield in the aswelded condition and 8,000 and 10,000 psi respectively
in the heat treated condition.
local conditions, particularly in out of position work.
For this reason average heat input computations for
a heavy joint based on a few travel speed readings
may be considerably in error.
Another factor affecting the accuracy of calculated heat input values is the need to shift to a volume nr area basis rather than deal with a linear con:tderation only, for the width of weave increases as
the speed of travel decreases, hence the need to include the lateral dimension.
The purpose of these two tests was to provide a
basis for comparison of the properties of two different brands of MIL-9018 electrodes. A comparison
of the results of tests B and C with those of I and H
respectively, reflect the characteristically higher
weld metal strength level produced by the stringer
bead technique.
The decrease in yield strength resulting from
heat treatment was surprisingly small in the MIL9018 weld metal tested. The reduction encountered in test I and J (MIL-10I18) was the greatest,
6,000 and 10,000 psi respectively. Even the decrease
in tensile strength was not generally substantial.
The MIL-11018 tt-sts, J, K. and L. were included
only for informational purposes. The weld pad chemistry for these clectrodes, as can be .een from Table
I. differs from that of lot v0BI IA MIL-9018 electrodes
only by an addition of approximately 1i of manganese
and, except for the adjustment required to produce
.this alloy difference. the coating composition of the
two is identical. A direct comparison between these
results and those of the MIL-9018 electrodes cannot
be made because of the difference in electrode diameter. However, the significance of these tests can
be seen by comparing the results with those reported
in Table 11 (electrode -3) which would appear on a
certification of conformance with the military specification for these electrodes,
The tensile bars, by virtue of their location in
the joint, (Figure 11' contained very little if any base
metal dilution or enrichment. In view of this, the
magnitude of the variation in yield strength here encountered should generate concern over the actual
strength level of weld metal, especially that of MIL11018 electrodes, and its effect on joint soundness in
highly diluted (root passes, for instance) restrained
joints in higher carbon, higher alloy base metal such
as HY-80.
An examination of the fractured surfaces of the
tensile bars indicated sqA•cnd, ductilc f. actures foi
all heat treated bars. As-welded bars D, E, G, H.
and I contained small fisheyes. The L bars were
both tested in the as-welded condition and produced
approximately the same results- -premature yielding and fracture resulting from small fisheye-like
defects,
Conclusions
This study of the effect o" welding variables on
the yield strength of certain low alloy steel weld
metal indicates that:
1.
The strength of a welded joint depends, in
addition to other factors, on the specific de"tails surrounding the manner in which the
weld metal is actually deposited in that joint.
2.
A given electrode, classified as MIL-9018 in
accordance with the provisions of existing
specifications may deposit weld metal during fabrication, which, in the as-welded condition, may have tensile properties of a MIL9018, a MIL-10019 or possibly even MIL11018 electrode, depending on how the electrode is actually applied.
A 11018 electrode, similarly, may deposit
weld metal conforming to the strength requirements of a 12018 type and even a 13018
type, should it exist.
Specification procedures and techniques
governing manufacturing tests of electrodes,
users inupection tests, and subsequent fabrication procedures must be identical if reasonable reproduction of results is to be
expected.
Discussion
Width of weave and position of welding (which
undoubtedly is related to amperage that can be used)
appear to be more influential than preheat and interpass temperature in their effect on the resulting
weld metal strength. In all cases, regardless of position of welding, the use of a stringer bead technique
resulted in the highest strength weld metal.
The effect of heat input is more difficult to interpret since the highest computed heat input did not
result in the lowest strength levels nor, conversely,
did the lowest heat input result in the highest weld
metal strength. A macro and micro structure study
would be in order, for the effect of factors such as
bead thickness resulting from position of welding
should be considered in addition to heat input.
Heat input is, in itself, difficult to handle since
a weldor very often changes his travel speed to suit
3.
201
Recommendations
V). Moreover, in single pass fillets and*
root pass work the effect of dilution, in
addition to the effect of welding technique,
should not be overlooked.
The amount or location of fabrication cracking difficulties when correlated with weld
metal mechanical properties must be made
on the basis of actual joint weld metal properties and not on the manufacturer's lot inspection test results unless the two are actually mide in the same fashion. (see Table
2.
202
Consideration should be given to the use of
stringer '-,ad techniques only whenever possible, in fabrication, and lot acceptance test
requirements set up on the same basis.
t
L
~E
tn
:10
E
a-
0
E'
E
E
E
Zi n
tr
ozCl)
0!
c
CCd)
a.
-
203
-
QUALITY CONTROL
IN THE
FABRICATION OF HY-80 STRUCTURES
Delhiered by: Thomas J. Dawson
Quality Control Superintendent
and Chief Metallurgist
The Ingalls Shipbuilding Corp
Pascagoula, Mississippi
Before:
Bureau of Ships Symposium
Washington, D. C.
21. 22 March 1960
'hat is quality control? It has been stated that quality control in•tntions are all the nivans by which frequency of difficulties is kept down It--ta) Minimizes the number dfdeficts
(b) Catches ths
duction as possibleh
that di ,oc-cur as early in pro-
every joint, and every phase of a fabrication operation would, of course, cost a great deal more than
could be justified. On the other hand, quality control relying solely on the faith of the steel manufacturer, the rod manufacturer, shipfitter, and
welder to produce automaticaily each and every
time satisfactorily would result in no expense. All
of the shipyards are somewhere between these two
extremes.
(c) Eliminates the cause o)f those that do occur
It will be noted from this that it definitelydiffers from
inspection which is to check .n end product to assure
its being satisfactory for use
Traditionally and rightfully so, the quality of any shipyard work has been thought of as the province of the
artisan and his immediate supervisor Bureaucratic
organizations make this prerogative seem greater in
the minds of supervisors. It is an old truism that
The definition is so simple, it minht be nmsleading
" To mninmile the number of difficulties" is a .far-
pride of workmanship differentiates the workman
from the journeyman, or we might say the artisan
from the helper This is as it should be, as pride of
workmanship will produce more quality at less cost
than rigid inspection or quality control.
It is unfortunate but quality does not beginand end in
the shop or on the ship. It begins long before our
workmen have an opportunity to demonstrate their
capabilities. Many times they unknowingly face problems, and this is especially true in the HY-80 materials, which require rigid procedural control and material control from the raw material through to the
steel fitting to such technical operations as accurate
control of the moisture content of the welding electrodes tc be used. They are faced with the following
problems:
reaching statement. and intentionally applies farreaching efforts. These must h, from original design
through procurement, process, installation, and final
testing. This is a large area, so large in fact that in
any industry, such a quality control p;rogram, to be
effective, requires strong managerial interest and
control.
Quality control in the fabrication of HY-80 falls into
the category of apparently a relatively new material,
at least in the thicknesses and uses to which it is now
being put in submarine construction. It is a material
that appears to require a reasonably strong quality
control program: in order to have real assurance of
a reliable end product. For the past two years, problems in welding HY-80 have been uncovered by very
costly experiences, and late in the fabrication sched-
1. Are the plans and the job instructions such
that they make the job feasible to perform?
2. Do the drawings and specifications incorporate
the required control that must be maintained in order to meet the required quality?
ules causing serious delays.
As in any other phase of work, the question arises-How far should we go? This in the end is governed
to a large part by how much It costs. Complete quality control over any operation like complete control
in any other area would, of course, give a complete
history. Complete quality control in every step involved, every move, proof of quality of every piece
of HY-80 material, every welding rod, every weld,
3. Is the base material and the welding rod unmistakably that which is right for the job?
4. Have those actions dr processes that have gone
on in processing and fitting the material before the
204
Ji
conscientious welder starts his weld helped or provented him from achieving the quality of weld desired?
If we have negative answers to any of these questions,
they will cost many dollarr and delays in spite of the
best welders in the world of which we have many good
welders in the shipbuilding iWdustry.
Quality control to be really effective embraces many
areas and must start at the beginning of the job to be
effective. The following requ',ements are some
ideas of the minimum fa-,urs necessary for an economic quality control program for the fabrication of
HY-80 which will result ir tuoi- most economiical job
and the production of a reu.*bkI
end product of predictable behavior:
(4) Job Instruction Control: It is iin this area that
quality control cm!n function effectively by reviewing
procedures and suppiei;!enting these as necessary.
The supplement st,,)uld define areas of responsibility
and dis±seminate the necessary information to all
parties responsible so that they have sufficient inforniation to effectively perform their jobs. As an example, the following, areas of responsibility in the
welding of HY-80 material should be established with
feed back information to the Quality Control Department:
(a.) Base Material control.
(I) Design: Some of the designs today have much
to be desired in order to ease the fabrication problenrs of HY-80 steel, The moiwierous geometrical
forms described as "crucif )rms" are very difficult
to weld due to the high restraint they impose on the
welded joints. Square corners with welds emanating
from 3 or more directions present high restraint and
multiaxial stress concentrations Through members
and intersecting members welded to shell or tank
tops create problems at snipes. By this I refer to
the snipe in one member toallow the weld to progress
uninterrupted on the other member. These present
difficulties when the snipes are welded in materials
less sensitive to weld than HY-80 steel. Effective
qualitycontrol procedures should point upareas such
as this which are difficult areas for fabrication. If
proper records are kept, they will be effectively
shown by comparison
(b) Electrode control.
(c) Root opening and joint preparation control.
(d) Preheat control
(e) Inprocess inspection control such as
1 M;naflux
-ay
2
2. X-ray
3. Control and release for other operations.
(2) Procurement: All materials should be ordered as specified Positive identification must be
maintained with full information on material identidto all the workmen. An
ficaton
coes
mnite loss of back reflection. We have established a
limit for rejection of 25%, loss of back reflection.
This has not been expensive and has not resulted in
a major percentage rejection of plate material, but
has eliminated plates of questionable quality for all
of the follow-up operations that must be performed
upon them. Records of receipt inspection will serve
as an excellent indication of whether or not the specifications are adequate, and good information on the
suppliers' reliability if reviewed.
(5) Inspection of Work in
ress: The earliest
detection of errors or faults pays the biggest dividends and results in the shortest delays. Regardless
who perfouriis the inspection, this inspection should
dssemiatedof
effective quality control procedure will, as a minimum,
spot check material specifications and material identification
to assure that it is being rigidly
carried out
report miinimal inforniation to the Quality Control
Department for assimilation and review, and the developnitnt of statistical information that will predict
(3) Receipt Inspection: Effective quality control
will encompass receipt inspection of material. HY-80
material requires a specific receipt inspection by
the Bureau specifications from which a great deal of
quality control information can be obtained and
through which rigid quality control procedures can
be employed. Laminar inclusions are a characteristic of the HY-80 material. The effect of these inclusions, which have been identified largely as "alumni", upon weldability has not been definitely established since it is commonly recognized that inclusions
normally contain a variety of impurities. They unquestionably exert an influence upon weldability and
should be controlled. Many of these are extremely
small inclusions but numerous throughout the thickness of the plate in certain areas. These arc demonstrable by ultrasonic examination not always by a
definite "pip" as the large laminationsbut bya def-
205
problei areas and a dependable eid product.
(6) Humtian Factors: As in any other area in dealing with peop)le, people are the direct insurers of
quality. As stated previously, pride of workmanship
will prodluce more quality at less cost than any of the
other means of quality control or rigid inspection.
All quality control programs should be arranged so
that they point up the good quality of workmanship to
encourage the individuals responsible for it.
Quality control can actually be thought of, after
o;lainirig the above information, in 3 stages--diagnosis. remedy, and maintenance. By maintenance,
we mean the holding of the remedies once they are
developed. The information gathered above is essential for the diagnosis and a rational approach to the
development of remedial procedures. The development of statistical data condenses the information to
electrodes at the prescribed temperature Xrior to releasing them in small quantities to the weller. It is
the welder's responsibility to return all electrodes
not uuied with a 4-hour period of time.
be fed back to managerial personnel relative to prnbtem areas to obtain the proper assistance to develop
remedial procedures. The maintenance of the remedial procedures is the biggest problem in quality control. This requires the cooperation of the entire supervision and inspection personnel concerned with operations. Factual data collected as outlined above
serves as a positive control over maintenance as it
will furnish data not only on problems but furnish information on good quality, and establish confidence
in procedures in use.
3. The Quality Control Department makes
periodic checks of the welding electrode moisture
content on incoming electrodes, after rebaking, and
from the welder's electrode can on the job site.
4. Records of electrode baking, moisture
contelnt, etc. are fed to the Quality Control Department Discrepancies from established procedures
and prescribed limits are taken up with the responsible parties for immediate correction.
An example of quality control as outlined above
practiced at the Ingalls Shipbuilding Corporation for
control of fabrication and welding o," HY-80 steel encompasses the following:
Electrode Moisture Content
Sampnile Source
(a) The operational procedures are reviewed by
the staff engineers of the Quality Control Department for completeness.areas of responsibility, standards of acceptability, and questionable areas are
discussed with responsible parties in an attempt to
assure full understanding and practicability of their
application.
From electrode can
opened at oven prior
to baking
After baking at 800
f,)r one (1) hour
(b) Receipt inspection of plate material is carried out by the Quality Control Department. This
consists of the following as a minimum:
I. Surface inspection,
2. Hardness measurements at diagonal corners of the plates as a rough cheek
against chemistry and heat treatment.
3 Micrometer measurements for plate
gage
4. Ultrasonic measurements at the corners
of a 2' grid for information as to uniformity of thickness.
5 Ultrasonic measurement of plate soundnesson thesame 2' grid withareas showing indications of discontinuities searched out in their entirety. Attachment 1
shows scgreg, tions resulting in 251 or
greater loss of ba-k reflection
F
0.00Tj
Electrode from holding
oven 72 hours
0.06%
Random samples from
welders' cans on the job
site
(1)
(2)
(3)
(4)
0.06%
0.19-T
0.09%r
0.14%
(d) Preheat control: It is the joint responsibility of the Quality Control Department, Welding Engineering, and Welding Department to see that the
proper preheat is maintained on all structures being
welded from HY-80 material. The Metallurgical
Processing Division of the Quality Control Department installs the preheaters and regulates the ternperature on the assemblies Hourly temperature
rcadmngs arc recorded as well as current weather
conditions, and whether or not welding was being performed or the assembly (Attachment 2). This information is fed back to the Quality Control Department. It is the responsibility of the Welding Engineer to establish the preheat levels. It is the responsibility of the Welding Engineering Department to
check welding supervisors and welders to see that
the correct preheat is maintained and the proper
heat imput is being followed for the welding being
preformed.
This information is accumulated by the Quality
Control Dcpartment, and the plates are assigned for
specific lcrat.ions in the submarines Defective or
questionable plates are positively identified bya different identification system from those plates of
quality satisfactory for use.
(c)
0.233%
Moisture control in electrode coatings:
I. TheQuality Control Department rebakes
all electrodes as received from the manufacturer
and distributes them to welding rod distribution
rooms.
(e) Magnaflux: As magnafluxoperations are performed, written records are made by the magnaflux
operators on the form enclosed as Attachment 3.
This information is fed back to the Quality Control
Department where it is accumulated bya statistician
2. The Welding Department personnel operate the welding rod distribution rooms keeping the
206
on work forms for IBM punch card operations (Attachment 4). At periodic intervals, a recap is made
of this information by the IBM system. This recap
is then reviewed by the Quality Control Department
and Welding Engineer for problem areas. An example of the value, this information can serve as illustrated by the following:
1. From a graph (Attachment 5) which in
typical of an HY-80 fabrication, the total cracks discoveued by magnaflux are plotted against the manhours of welding applied to the job. From this it can
be seen that the cracks were reasonably proportioned
to the man-hours applied to it which varied with the
cracks for several months until true control allowed
the manpower to continue to increase and the cracking rate to decrease. The latter end of the graph is
welding under very high restraint which accounts for
the rise in the number of cracks per man-hour.
The small difference between the two curves ind:eates
the finished weld cracking that was observed both by
delayed cracking and that found immediately upon
completion of the weld.
2. Attachment 6 graphically demonstrates
the transverse cracks compared to the longitudinal
cracks for a better realization of the basic cracking
problem involved in welding HY-80 material under
restraint,
tem which from a mechanism for feed back of information on corrections or anticipated problems, we
feel, have proven potential as a cost reducer in our
program. We constantly guard, however, against
personnel drifting too far into a straight forward inspection group. Inspection by its nature is primarily
destructive criticism and we attempt to keep our
quality control in the area of constructive criticism.
If we can continuously continue our aim of constructive criticism, we will continue to enjoy the cooperation of all of the other departments with quality control services as a focal point from which we can initiate action on our HY-80 welding problems for the
control of costs as well as quality.
---
4. In order to differentiate between these
two problems and to get a truer picture of the amoutri
of delayed cracking observed, graph (Attachment 8)
shows finished weld cracking for comparison with
the delayed cracking. It will be obServed from these
curves that sufficient delayed cracking has been observed to justify the "7-dayrecheck on all of the primary strength welding performed on NY-0o strucmarestr pldnt.
tures at our plant
(f) The railiographic results are likewise
tabulated.
In summary, those parts of the quality control sys-
3. Attachment 7 graphically demonstrates
the total cracks plotted against the root weld cracks
in an attempt to evaluate the source of the cracking
problem in HY-80 steel. From this it can be observed that to a very great extent, the cracking problem
exists at the root of the weld from our experiences.
It is from this information that we, todate,have been
unable to develop a procedure that we felt was satisfactory for twin arcing or for welding over tacks.
207
ATTACHMENT I
LAMINAR PLATE DEFECTS RESULTING IN MORE THAN 25% LOSS OF BACK
REFLECTION
TYPICAL MAGNAFLUX
---
INDICATION OF SECTION AFTER DETECTED BY
ULTRASONIC INDICATIONS
208
ATTACHMENT 2
Form No. 125921
THE INGALLS SHIPBUILDING CORPORATION
PASCAGOULA, MISSISSIPPI
QUALITY CONTROL DEPARTMENT
METALLURGICAL PROCESSING DIVISION
RECORD OF STRIP HEATING
Hull
Date-
Frame No.
Assembly No.
No.
Heaters:
800 watts
Weather Conditions:
.
1100 Watts_
1st Shift
------
Ma•ual
Weld
Aircomatic
Weld
Material
2000 WattsHeated__
2nd Shift---Ambient Temp.
Time
Volts
7:00 AM
8:00 AM
9:00 AM
10:00 AM
11:00 AM
12:00- N•oo
1:00
2: 00
3:00
4:00
5: 00
6:00
PM
PM
PM
PM
PM
PM
F
Temp.n
...
_-_--TT.0
.....
Time
F_..
_
.....
Temp.
Volts
-OW__---8:00
.T PM
PM
10:00 PM
__7
_T
Shift__
___3rd
..
.
. ..
-
. .
11:00 PM
.. -_o ,
1:00 AM
2:00 AM
- AM
4:00 AM
5:00 AM
6:00 AM
_
All temperature recordings are greater than the degree indicated but below the next 50
Reading shall be taken from 8" to 12" ahead of welders.
Temple Stick.
Remarks:
Observers:
1st Shift
2nd Shift
3rd Shift
209
F
Form 11597
ATTACHMENT 3
THE INGALLS SHIPBUILDING CORPORATION
PASCAGOULA, MISSISSIPPI
QUALITY CONTROL DEPARTMENT
METALLURGICAL PROCESSING DIVISION
MAGNAFLUX REPORT
Hull
Date_.
...........
.
PART TESTED:
Frame No.
Inside Hull
Assembly No....
-Outside
Hull
.-
Forward of Frame
Aft of Fr.
-
Scars (No.)
Vertical Butt --
Flange Et'Itt
Horizontal Butt
Webb Butt
Circumferential Butt .
Webb to Shell
Frame Butt
Webb to Flange.
Tie-in
-
--
Block Ends
--
--
X-Ray Repair.-
Miscellaneoud Parts-(wriv out)
TYPE OF CHECK:
Root Pass
PROCEDURE:
Magnetic Yuok,
Finish Weld
Re-Check
Other___
(write out)
OBSERVATION:
Number of Blocks.
....-
Length of Blocks_
Numbcr of Cracks:
Longitudinal
-.
Transverse .....
PORT
Total Distance Magnafluxed:
Ft. . ..
.
In - Total Inches of cracks
S...
-
Note: Location of M/fluxed area indicated on sketch.
Horizontal seams are indicated by symbol
Remarks
REQUESTED BY:
Welding Dept.
ST60
Navy-
-
NPD
t...._Other
Reported By:
Magnaflux Technician
210
(write out)
-~
U
al.
hh
II
iH
'
,
I
I
I,
_V
0~
i
*
2-11
211
STRUCTURE CODE LIBT
STRUCTURE
CODE NO.
1
Web to shell
2
Web to flange
3
Frame butts
4
Bhd. to shell
5
Shell seams & butts
6
Sea Chent
7
Inserts (shell)
8
Misc. Inserts & Penetrations
9
High Pressure Tanks
10
Torp. rnmpulse Tanks
11 (escape hatehes)
Trunks to shell
12
Shielding tank
13
Lead Shielding
14
Foundations
15
Butt welds in bhds. & tank tops
Subject to D.S.P.
16
Scars
17
Misc.
212
,I
HYBO WELDING
NQ OF
CRACKS
HOURS
11000
22
ATTACHMENT 5
100
20
800
.
.
700
.
.
.1
.
.
.
.
.
.
.
TOTAL CRACKS
-8
MAN HOURS
....
6909..
16
.96.
14
5oo.
O
200
"V
Jo
506.7.... ..... 0
1958
8
12
2
.
.....
4
3
5T
6
7R 8
9
1959
1000
-...
6
10
1R 12N
ATACMET
2S
1960
8
90.......................TOTAL.RACKS4
HY80 WELDING
NO. OF
CRACKS
I000
ATTACHMENT 6
goo -,
TOTAL CRACKS
TRANSVERSE
800
700
600
.
.
..
500
300 -
S4
~LONGITUDINAL
i
-
CRACKS
CRACKS
.
.
--
.
.....
. .
. ..
.
..
.
.
.
,
5 -6
7
8 910111
1958
12 1
2
3
4
5
6 7
1959
8
9
1'011
12 1
2
1960
213
4
,t
HYBO WELDING
NO.OF
CRACKS
CRACKS
-ATTACHMENT
900
7
,TOTAL
,
,
,!
CRACKS-
ROOT WELD a
.SCAR CRACKS
800
,
700
1960
1959
1958
HY80 WELDING
-
NO OF
C000
900
ATTA.HMENT
.
......
800 0.
FINISH WELD CRACKS
RE-CHECK CRACKS
......
.......
..........
. .I
700..............
2000
.
.. ..
...
.
.
.
....
.
......
.TAHMN
.} .
6001
100
400
200.........................................*/
I
I00
I
4
5
6
7
8
9
10
11 121 1 2
3
*
4
5
6
1959
1958
4
,$•
*
7
8
9
10
II
12
I
2
1960
214
J1
!•
March 22, 1960
!!Y- 80•Symposium
Workmanship Controls in the Fabrication of HY-80 Hulls
E. Franks, Electric Boat Co.
The fabrication and welding of HY-80 steel struetural components. or necessity, must be closely controlled operations in order t, assure satisfactory
performance of the end product. Judicious inspection throughout !fith various stages of fabrication is
the method enplo•yed at EBDe+,,. to control the workmanship and provide this assurance.
shown in Figure 6 burned edges are checked for
bevel angles, plate laminations, gouges and uneven
burned surfaces. This data is logged ipto Inspection
data bsooks at this point.
The receipt inspection of electrodes used for welding HY-80 steel is based on the manulacturer's certification of compliance with Military Specification
together with results of a practical usability test conducted by UhloWelding Engineer's Office. Release of
electrMdes for oroduction use is based on satisfactory
complet ion of lseth requrtenment S and the results of
[his in-spection are recorded by the, Welding Engineer's
Office. In order to provide and maintain low moisture
content in electrode coatings, all Types MIL- 11018,
i!L- 10015 and MIL-260- 15 electrodes are baked
prior t4 issue to the welders. As shown in Figure 7,
this aiking is conducted a! a centrally located point
by Welding Department personnel.
Electrodes are
transferred to holding ovens strategically located
throu; :i out the plant for direct issue to the welders.
The Welding Engineer's Office makes a daily surveillance check of electrodes in the holding ovens as
shown in Figure 8. Random samples of electrodes
are selected from the holding ovens, the coating removed and tie sample tested for moisture content
by the R & D Laboratory. Results are reportod daily
to the Welding Engineer's Office
It is essential that the quality of the materials
en? er mg into tiit, fabricat o n I,. ci ciubined wits work-anshaip control so that end prodtuct deficiencies, it
they occur, can be properly tevaluated. Froem re.ceipt inspection of HY--80 plate material to final tyispection of tht, submarine structure, all inspections
form a part of worknianship control. Receipt inspectiOtn of the HY-80 plate ntateria! is conducted by the
IlnspctinrY De|artnicitt upfon arrival (of the plate in
the plant. A.; shwnmin Figure 1. Brinell Hardness
is checked at three points o)n each plate. The paint
is removed from the three areas pr,or to taking the
readings. The edge thickness is checked as shown
in Figure 2, at 10 points using a mir,moeter. For
all plates subject to sea pressure ais designated (in
plate schedules. ultrasonic tests are made for detect ,•nof laminations as shown in Figure 3. Theset
UT readings are taken along Ilies of a 24" grid over
the entire surface. This inspection is combined
with thickness measurements taken by means of
Vidigage ultrasonic equipment as shown in Figure 4.
The Vidigage !hickness readings are taken at the
intersecting points of the 24" grid mentioned above,
At this time visual surface inspection for scars,
scabs or other surface defects is made and such
areas marked on the plate surface. All this datae is
recorded on chits as shown in Figure 5. Two types
of forms are used for the receipt inspection operation. The first designated Type A inspection is used
for plates which require UT and second designated
Type B inspection for plates which do not require
ultrasonic inspection. Upon completion of the re-
The fabrication sequence now brings the plate material in the hands of the shipfitter and the welding
electrodes in the hands of the welder together to start
the erection operation. When the shipfitter supervist haF the structure ieady for welding, he requests
a "cleaied for work" chit from the structural inspector. The structural inspector checks the area for
joint preparation, fit-up tolerances and preheat temperature. All fits must be satisfactory with the required preheat on the joint. When the inspector determines that all requirements are satisfied he obtains
the signature of the shlpfitter supervisor, the welding
quired tests, the plates are marked satisfactory or
unsatisfactory, stored and the records forwarded
to the Inspection Department records section.
supervisor and certifies by his signature that the joint
is satisfactory to start welding. The welding supervisor's signature on this chit as shown in Figure 9
indicates that he accepts the joint presented by the
shipfitter as satisfactory for welding. This chit as
shown on Figure 10 is then distributed by the inspecir to the various offices for record. It should be
noted that this chit may also be used for other operations which will be described later.
When plates are drawn from storage for fabrication, they proceed to the Layout Department where
the Inspection Department checks layout in accordance with blueprints, makes visual surface inspection for scars, scabs or other surface defects. As
215
kr
)
During the course of all welding, the structural
ilspector makes surveillance cl'•eks for compliance
with established welding procedures, techniques and
preheat and interpass temperature. Figure 11 shows
an inspector checking preheat during progress of
welding an insert. If at any time the inspector considers the work is not completely satisfactory, it
is his prerogative to obtain corrective action through
the cognizant supervisor,
When. in the course of erection, the welding
supervisor determines an application is suitable for
twin arc welding as shown in Figure 12 he presents
a twin arc chit to the Inspection Department who
check to tleir satisfaction that all alplicable requirements art, met. The Inspection Departmeint then
obtains approval from the Navy Inspu-ction Office and
distributes copies of the chit which is s!io~r on
Fig'ure 13.
For full penetration welds, arc air gouging is
used for preparation of the back weld. After gouging. the back weld artm is ground and an MT inspection of the biackgouged surface conducted by the
structural inspector. The results of this inspection
are entered on a chit similar to the final MT chit and
copies distributed,
When the welding supervisor has completed the
fabrication to his satisfaction he notifies the structural inspector, who visually inspects the work
while still hot for weld contour, size or undercut.
At this point the Inspector makes certain that in the
preheated area all temporary erection clips and
brackets not required for subsequent fabrication are
removed, all surface scars are repaired and the
surface in the weld area is properly prepared for
final magnetic particle inspection. When considered
satisfactory by riie inspector he prepAres t chit bor
final MT inspection, which the welding supervisor
signs, certifying that the weld quality is in accordance with all requirements. At this point the welding supervisor authorizes the preheat to be turned
off. Final MT inspection is then made by the structural inspector 24 hours or more after cooldown.
Figure 14 shows a structural inspector making the
final MT inspection of a butt weld. Upon satisfactory MT results, the chit is so marked and signed
by the inspector and duplicate chits distributed,
This chit is same as that used for fit up.
and Immediately forwarded to the main inspection
office. The structurai inspector then gives verbal
authority to the Welding Department to procede with
removal of the defective area. When the inspector
ascertains by MT inspection that the defect has been
completely removed he signs the chit noting that the
crack w.s removed and the area is ,eady for repair
welding. This chit is now completed and duplicates
are distributed. After repair welding is corr!eted
to the satisfaction of the welding supervisor the requests the structural inspector to initiate a new chit
marked "repair weld" for final MT inspection as
with an unrepaired weld. Final MT inspection of a
repair weld is made at least 10 days after (ooldown.
Upon completion of all butts and seams, the structural inspector initiates a request to the X-Ray Department to radiograph the required area. When
the X-Ray Department radiographic inspector ascertains that the X-Ray quality mi ets all requirements
he pre.ients the radiographs to the Inspectihn Department. As shown in Figure 15 the Inspection Department reviews the radiographs and presents them to
the Navy Inspection Office fur certification. At each
review a radiography report, which accompanies the
radiographs, is initialled by the reviewer signifying
certification and the chits returned to the X-Ray
Department for file.
The data obtanmedatrom the final MT chits is
transferred to master plans by the Inspection Office
as shown on Figure 16. By this method an uprtodate record is available for ready reference on the
status of completed work. After entry on the master
plans the chits are filed in the Inspection Office as
shown in Figure 17. This file is available to all inspectors for review and reference at any time.
The use of this chit system at Electric Boat Division for workmanship control in the fabrication of
HY-80 hulls provides a constant check and record
throughout the production sequence. A flow chart
as shown on Figure 18 indicates the sequence of
major operations throughout construction. When
the various inspection operations discussed above
are factored into this flow chart we see that an inspection is performed between essentially all major
operations. By this method quality is assured
throughout the fabrication operation. The continuous
surveillance by inspection personnel and their contact with trade personnel increases the trade cognizance of quality. In this manner, workmanship is
used to build quality into rather than inspect quality
into the end product.
When unsatisfactory MT results are obtained the
chit is so marked, and a repair weld certification
procedure follows. A complete description of the
unsatisfactory area is prepared by the Inspector
216
FIGURE 1
FIGURE 3
FIGURE 2
FIGURE 4
217
FIGURE 5
FIGURE 7
FIGURE 6
FIGURE 8
218
r
FIGURE •
-n
.
,.
oi
FIGURE 11
a
FIGURE 10
FIGURE 12
219
MITN
FIGURE
'A
B ICA T I%
A N
A k r2
71
N % ; f2
FIGURE 18
221
,
Fly UP