Optimizing Weld Repair in Cast Steels

7/25/2019 Optimizing Weld Repair in Cast Steels

http://slidepdf.com/reader/full/optimizing-weld-repair-in-cast-steels 1/7

Optimizing Repair Welding Techniques

in Cast Steels—Part

II

Hz-induced cracking can be prevented by using low

hydrogen practice if the maximum HAZ hardness is below

Re

35, and all steel

castings

should be normalized before repair w elding

BY D. K. A I D U N AND W. F. SAVAGE

of cast

for hydrogen- induced

low- and

h igh-

in simulated repair

It was shown tha t to prevent

is not possible,

of hardness in

be be low

e 35.

Metal lographic examinat ion of these

the

fus ion zone

out

and

including

the

unaffected base

are:

1 Fusion zone

2. Unmixed zone .

3. Coarse-gra ined homo geneou s HAZ.

4 . Fine-grained homogeneous HAZ.

5. Part ial ly t ransformed region.

6. Spheroidized region.

7. Unaffected base metal.

All as-cast heats with carbon contents

0.21 to 0.32 wt-%,

upon

to a 5-second arc t ime, exhib

of about

0 Re in the ne tworks of martensite

the

islands

of

unt ransformed

in the part ial ly t ransformed region.

was a t t r ibuted to the inhomogenei ty

f the as-cast structure. However, for

was

by normalizing the alloy

the we ld ing opera t ion . For cast

of abou t 0.30 wt-%, in addi t ion to

a pos tweld heat t reat

be ad visable.

It is generally recogn ized that four

be met s imultaneously for

to

occur.

can be summar ized as fo l lows:

1 . A me tallurgical requisite. For crack

in g to occur , a crack-sensitive microstruc

ture must

be

present , and,

in

genera l ,

the

susceptibility to hydrogen- induced crack

ing increases with increase in the hard

ness of the m icrost ructure.

2. A chemical requisite. For cracking to

occur , the level of dif fusible hydrogen

present in the crack-suscept ible micro-

structure must exceed a cr it ical concen

trat ion, which decreases sharply as the

yield strength of the microstructure

present increases (Ref. 1).

3.

A

mechanical requisite. A l though

cracking has been show n to init iate at

lower stress levels, it appears that crack

prop aga t ion requires localized stresses of

approx imate ly y ie ld-s t rength magni tude

(Ref. 2).

4 .

An environment requisite. Hydro

gen- induced crack ing has been show n to

result f rom the ext raord inary reduct ion in

the notched-tensile strength

of

steels

caused by the presence of a relatively

small amount of d i f fus ib le hydrog en (Ref.

3). Since this loss in strength in the pres

ence of a notch occurs only in the range

- 1 4 8 to 390°F ( -100 to 200°C), hydro

gen-induced cracking requires that the

weldments be exposed to temperatures

within this range in the presence of the

f irst three requisites.

A l t hough in theory h ydrogen - induced

cracking

can be

p reven ted

by

el iminat ion

of any one of the above requisites, in

pract ice it is usually possible to exert

cont ro l over only the first two. This

results f rom

the

fact that

the

thermally-

D. K. AIDUN

is Assistant

Professor,

Mechanical

an d

Industrial Fngineering

Department, Clark-

so n

University,

Potsdam, New

York;

a nd W F.

SA

VA

GE is Professor

Emeritus,

Rensselaer

Poly

technic Institute,

Troy,

New

York.

induced residual stresses in all but the

simplest of weldment geometr ies are

usually sufficient in magni tude to init iate

cracking if the other requisites are met.

Furthermore, unless stress relief heat

t reatment is per fo rme d af ter weld ing , the

addi t ion of service usually raises the local

ized stresses in the we ldmen t above the

crit ical level for c rack propagat ion.

Because most welded structures expe

rience service in the crack-sensitive t em

perature range, this requisite is usually

met . However , it is possible to take

advantage

of

this requirement

in two

ways:

1 . Postweld soaking. It has been

s h o wn by Adams (Ref. 4) and others that

an immediate postweld soak for 30

minutes (min) to one hour (h) at a temper

ature of above 390°F (200°C) will usually

prevent hydrogen - induced crack ing. This

is at t r ibuted mainly to the reduct ion in the

level of dif fusible hydrogen made possi

b le by the greatly increased diffusivity of

hydrogen at 390°F (200°C). Thus, by the

t ime the we ld m e n t is a l lowed to return to

the crack-sensit ive temperature range,

the residual level of dif fusible hydrogen is

rendered be low the critical lev el for crack

init iation.

2. Cryogenic storage. In conduct ing

laboratory studies, it is convenient to be

able to s tore welded specimens under

condit ions which preclude either loss of

hydrogen or the f o rmat ion of hyd rogen-

induced cracks unt i l controlled experi

ments can be init iated. This has been

s h o wn (Ref. 5) to be possible by transfer

o f the we ldments to l iquid nit rogen

(-320°F) ( -1 95 .8°C boi l ing point ) as

soon as they reach ambient temperature

and stor ing under l iquid nit rogen unt i l the

exper iments begin.

The crit ical hardness level is inf luenced

WELDING RESEARCH SUPPLEMENT

197-s



Table

1—Chemical

Co mpo sition, Initial Hardness (IH), and Ideal Critical Diam eter (D[) of Heat

32, 27, 1 , and 0

Composition, wt-%

C

Mn

P

S

Si

Ni

Cr

Mo

Cu

Al

IH, Re

D|, in.

32

0.32

1.15

0 017

0.017

0.54

0.13

0.22

0.04

0 078

51

3 17

27

0.31

1.57

0.013

0 013

0 50

0 03

0.14

0 10

0 04

0 054

51

3.51

1

0.21

1.24

0 018

0.031

0 51

0 07

0 20

0 03

0 04

0.11

46

2 05

0

0 26

0 74

0 021

0 028

0 40

0 04

0.31

0.02

0.03

0.075

48

1.61

by the toughness and duct i l i ty of the

microstructure of the weld heat-af fected

zone (HAZ). O f the var ious possible

microconst i tuents, untempered mar tens

ite is the most br it t le and, therefore, the

most suscept ible to hydrogen-induced

cracking.

The other microconst ituents,

listed in order of decreasing susceptibility,

include lower bainite, upper bainite,

pearl i te, spheroidite, and ferr ite.

In general, the toughness and duct i l i ty

both of ind iv idual and par t icu lar combi

nat ions of microconst ituents is decreased

by increasing the carbon content of the

prior austenite. For this reason, the sus

cept ibi l i ty to hydrogen-induced cracking

is direct ly related to the carbon content

of the steel. For a given carbon content,

increasing the percentage of martensite

in the weld heat -af fected zone lowers

the cr i t ica l hydrogen concent rat ion

required to cause hydrogen- induced

cracking.

Carbon is the least expensive of all

alloying elements. Because of this, steel

foundries tend to increase the carbon

content of their steels in order to achieve

higher strength. The increase in carbon

content increases the ideal crit ical diame

ter, D|, more than any other alloying

element; in addition, it also increases the

hardness of any martensite formed.

A few years ago, the railroad industry

began to invest igate the fai lure of cast

com ponen ts such as couplers, yokes, and

knuckles. As a result of this investigation,

i t was concluded that hydrogen- induced

cracking was a contr ibut ing cause of

fai lure. I t seemed l ikely that hydrogen-

induced cracking could result f rom

improper weld ing procedures used in

repairing casting defects in these struc

tures. It was suspected that, in the

absence of suitable control of repair-

welding procedures, crack suscept ible

heat-af fected-zone microstructures could

easily be produced.

Unfortunately, repair welds in steel

cast ings are of ten performed solely for

"cosmetic" reasons to eliminate superf i

cial surface imperfect ions. The rapid

cooling rates, which are associated with

such "cosmet ic" repairs , coupled wi th

inadequate control of sources of hydro

gen such as improper handling and stor

age of consumables, great ly increase the

probabi l i t y of hydrogen- induced crack

ing.

Underbead cracking is one of the most

insidious forms of hydrogen-induced

cracking. This is because it often is lo

cated ent irely beneath the weld and fails

to emerge at the surface where it can be

detected by nondest ruct ive inspect ion.

Unfor tunate ly , the geometry of most

repair welds is such as to make this a

dominant form of hydrogen- induced

cracking in steel castings. If undetected,

such cracks may experience slow crack

growth in service unt i l the cr it ical crack

size is reached and catastrophic fai lure

occurs.

Materials and Specimen

Preparat ion

Four heats of cast steel were used in

this inve st igat ion. Table

I

lists the chem ical

com pos it ion, init ial hardness,

IH

(the

hard

ness of martensite), and ideal crit ical

diameter, D, (based upon 50% martens

ite),

for all four heats.

The four heats, supplied in the form of

cast coupons, had dimensions that were

approx imate ly 10Vi X4V4 X

3

/i in.

(260 X 108 X 19 mm). From these co u

pons, specimens of 10 X 2 X V2 in.

(254 X 50.8 X 13 m m) we re machined

for the hydrogen- induced crack ing test .

The sur face on which the welds were to

be deposi ted was then pol ished wi th 24 0

grit SiC paper and degreased with ace

tone.

Polishing and degreasing are crit ical

and impor tant for the hydrogen- induced

crack ing test when us ing low-hydrogen

welding pract ice. These cleaning opera

t ions mitigate hydrogen absorpt ion in the

weld b y remov ing hydrogenous mater ia ls

from the surface of the specimen.

Procedure

The e lect rode type , hydrog en c ontent ,

heat t reatments, and the heats of steel

used in this part of the investigation are

summarized in Table 2. The level of

dif fusible hydrogen in welds was mea

sured by the RPI silicone-oil extract ion

technique (Ref. 6). Two levels of dif fus

ib le hydrogen were s tudied:

1 . Low-hydrogen pract ice. Af ter re

moval f rom hermetically sealed shipping

containers, the electrodes were trans

ferred immediately to a well-vent i lated

holding oven and maintained at 350°F

(177°C) unt i l just before use. This proce

dure gave an average level dif fusible

hydrogen of a pprox imate ly 3 ppm (3.3

cc/100 g of weld metal).

2.

High-hydrogen pract ice. The elec

t rodes were removed f rom the open

shipping container, dipped in water for

30 s and used immediately. This pract ice

gave an average dif fusible hydrogen level

o f 30 ppm (33.3 cc / 1 00 g o f we ld m et

al).

The 10 X 2 X l/

2

in. (254 X 50.8 X 13

mm) bars were welded using var ious arc

t imes and weld distances with both low

and h igh-hydrogen pract ice. Af ter

coo l

ing to ambient temperature, the

w e l d

ments were immediately loaded in the

varestraint apparatus (Ref. 7) and a 2%

augmented st ra in was then appl ied.

During this process, an acoust ic-emis

s ion t ransducer was mounted on the

specimen surface adjacent to the weld

bead,

as show n schematically in Fig. 1 , to

monitor crack init iat ion and propagat ion.

Table

2—Summary

for Electrodes, Hydrogen Conte nt, Heat Treatments, and Steel Heats Used

in Investigation

Diffusible H2 in welds made with

3/16 in. (4.76) diam. electrodes:

Moist E7018(

a

>

Baked E7018

(b)

Heats 32 , 27, 1 , 0

Heats 32, 27, 1,0

Heat 27

30 ppm

3 ppm

As-received

Normalized at 1750°F (954°C)

Normalized and postweld heat treatment

at 1100

3

F(593°C)

'"'Electrode d ipped in wa te r fo r 30 seconds .

(b)

Electrode s to red con t inuous ly a t 350-F (177

C

C) a f te r remova l f rom sea led con ta ine r .

98-sl APRIL 1 985



Because of the high ductile-britt le transi

t ion temperature (DBTT) of these heats,

the hydrogen-induced cracking could not

be per form ed at l iqu id-n i trogen tem pera

tu re , - 3 2 0 ° F (-196°C).

Specimens were then cut and

mounted for meta l lographic examinat ion

and microhardness measurements.

Hy d r o g e n - I n d u c e d Cr a c k i n g T e s t s

Equations described in an earlier paper

(Part I to this paper — Ref. 8) were used to

select weld ing condi t ions which would

give maximum weld heat -af fected zone

hardness levels ranging from about 25 to

50 Re in the coarse-grained region at the

fus ion boundary. Welds were made in

heats 32, 27, 1 and 0 using the selected

welding procedures to make simulated

repair welds wi th both h igh- and low-

hydrogen pract ice.

The welde d specimens were loaded in

the varestraint test apparatus. The n, af ter

attaching the acoustic-emission transduc

er , they were subjected to a 2% aug

me nted strain for 2 h in orde r to simulate

the mechanical factor by creat ing yie ld-

strength stresses in the outer f ibers of the

specimen.

The acoust ic-emission data provided

qualitat ive informat ion on the init iat ion

and propagat ion of hydrogen- induced

cracking.

However , quant i ta t ive cor re la

t ion between the acoust ic-emission

counts and the severity of cracking was

not found. This was believed at the t ime

to result f rom extraneous noise in and

about the laboratory that caused errone

ous counts. However, for reasons that

are discussed in a later section, this may

be only partially responsible for the lack

of correlat ion.

fjk

1

- < « • •

""31

Fig.

2

— Typical underbead crack

in the coarse

grained region

of

the HAZ

of

a stationary

weld

made on heat 27 using high-hydrogen prac

tice. Arc c urrent - 220 A; arc voltage-25 V;

arc time-5 s; section thickness- 'A in. (12.7

mm); cooling rate at 900°F (482°C)-49°F/s

(27.2°C/s); hardness-52 Re (50 Re pre

dicted).

2 nital, X500 (reduced 46%, on

reproduction)

For welds made using high-hydrogen

pract ice the maximum "safe" hardness

level was found to be about 33 Re.

Welds made wi th condi t ions that

produce hardness levels of less than 35

Re did not exhibit detectable cracks in the

coarse-grained region of the heat-af

fected zone , even w i th 30 ppm of d i f fus

ible hydrogen present.

With low-hydrogen pract ice, no crack

ing was detected in the coarse-grained

region until hardness levels of approxi

mately 45

Re

were p roduced .

Figure 2 shows a typical underbead

crack, which init iated in the

coarse

grained region of the weld heat-af fected

zone .

The microstructure around the

crack is predo mina nt ly martensit ic.

Weld Bead

To Acoust ic

-»

Emission

-T^P—i

Recorder

aug

2 R

SIDE VIEW

1 - Schematic diagram of bending section of the varestraint apparatus and location of acoustic

transducer with respect to weld

:A

X

# •

ftot

®

Fig. 3 — Effect of cooling rate on microstructure

of repair w eld HAZ in heat 1. A -stationary

weld (stationary

arc);

B — linear weld. Welding

conditions: for A and

B — 220 A

arc

current,

25

V a rc voltage,

V2

in. section

thickness;

for A

only-5

s arc time, cooling rate of 39°F/s @

900 F,

51

Re

max.

hardness;

for B

only- 10

ipm travel speed, 2.0 in. weld

length,

18°F/s

cooling rate @ 900 F, 31 Re max. hardness.

2 nital etch, X500 (reduced 35% on repro

duction)

Figure 3 shows weld heat-af fected-

zone microstructures which experienced

maximum cooling rates of 18 and 39°F/s

(10 and 21 .7°C/s ), respect ively, at 900°F

(482

°C).

The structure in Fig. 3A consists

of a mixture of m artensite ( ident if ied by

the let ter M), lower bainite ( ident if ied by

the letter

B

L

)

and up per bainite ( ident if ied

by the let ter Bu).

The measured maximum hardness in

the martensitic regions of the specimen in

Fig.

3A was

Re 51 ,

whereas the calculated

value of IH for this material (heat 1) is only

45

Re.

This means that the carbon

con

tent of the pr ior austenite was probably

inhomogeneous and was higher than the

nominal carbon content in some regions,

thus explaining the higher than expected

maximum hardness. This is entirely possi

ble because the short t ime-at- tempera-

ture could prevent complete

homogen i -

zat ion f rom occur r ing even at tempera

tures just belo w the solidus. The structure

in Fig. 3B consists of upper and lower

bainites with maximum hardness level of

Re 3 1 .

The major dif ference between cast

and wrought steel products l ies in the

nature and severity of the composit ional

heterogenei ty . In both types of product ,


199-s



.*--

E F

«*•

4 £ C 0

fig.

4 — Microstructure

of 5

s stationary

repair weld

in

heat 27 w ith KHN

(50

g

load) hardnesses as

follows:

A- 168;

6 - 2 8 0 ;

C - 4 0 3 ; D-403 p rior pearlite, 197 prior ferrite; £ -6 1 5 prior pearlite, 84

prior ferrite;

F—

163 prior pearlite; C- (n o t stated). Nital etch, X50 (reduced 49% on reproduc

tion)

the init ial solidif ication process is accom

panied by dendrit ic segregat ion with the

dendrite interst ices being enr iched in

those elements which depress the l iquid

us tempe rature of the a l loy. H ow eve r ,

dur ing soaking at high temperature in

preparat ion for hot -work ing, par t ia l

homogenizat ion occurs, and the pat tern

of the residual segregation is altered by

the hot -work ing operat ions. Thus, in the

case of hot-rolled plate products, the

combinat ion of thermal and mechanical

t reatment forms alternat ing bands of

sol

ute-rich and solute-lean material parallel

to the rolled surface.

In the case of steel castings, the p atte rn

of segregat ion is controlled by the

den

drit ic-arm spacing, which increases as the

section thickness increases. Since no

mechanical work in g is involve d, hom oge

nizat ion of cast ings can only be achieved

by dif fusion-controlled thermal t reat

ments. When the dendrite-arm spacings

are large, the increased diffusion dis

tances require longer t imes and higher

temperatures to achieve a reasonable

degree of homogenizat ion. Thus, it is

customary to normalize cast steels prior

to the f inal heat- t reatment operat ions

which are designed to control the

mechanical propert ies of the cast ings.

Figure 4 shows the microstructural

changes from the fusion zone at the lef t

outward to the unaf fected base meta l a t

the r ight. The panoramic view of the

weldment shows seven dist inct types of

microstructures labelled A t h rough C in

Fig.

4 :

A

—

fusion zone. Figure 5 show s the

microstructure of this region as revealed

by a 2% nital etch at X500. The micro-

structure consists of Widmanstatten fer

r ite and f inely laminated pearl i te. The

hardness of this low-carbon weld metal

was only 168 KHN.

B

—

unmixed zone (Ref. 9). The top

section of Fig. 6 shows this zone at X500.

Because mechanical mixing is incomplete

in this zone, the substitutional alloy con

tent approximates that of the base metal.

Carbon, on the other hand, dif fuses so

rapidly that the carbon content is inter

mediate between that of the weld (0.07

wt-%) and that of the base metal (0.31

wt-%). The region consists of bainite,

Widmanstat ten fer r i te and very f ine

pearl i te structures with measured hard

ness of only 280 KHN.

C—homogenized coarsed-grain re

gion. Figure 7 shows the microstructure

of this region at X500. The uniform

response to the etchant within this region

in Fig. 4 indicates it to be reasonably

hom ogen eou s as a result of the high pea k

temperatures experienced in this region.

The micro structure is ful ly martensit ic and

exhibited a hardness of 403 KHN.

D

—

inhomogeneou s fully transformed

region. Within this region the response to

the etchant in Fig. 4 is uneven and the

- >

*

Fig.

5-Region

A in

Fig.

4 at X500 (reduced

46 on reproduction). 2 nital etch

Fig. 6 -

Region

B in Fig. 4 at X500 (reduced


mott led appearance is indicat ive of inho-

mogeneity. Figure 8 shows a port ion of

this region at X500. The light gray regions

are ful ly martensit ic and exhibited a hard

ness of 403 KHN. The intermediate

regions which consist of a mixture of

e levated- temperature t ransformat ion

products exhibited a hardness of 197

KHN.

Apparent ly region D exper ienced peak

temperatures abo ve the ef fect ive A3 long

enough to complete the t ransformat ion

to austenite but for too short a t ime t o

exper ience homogenizat ion of even the

interstit ial carbon. The austenite grain size

in this region ranges from slightly coars

ened to ful ly ref ined and the hardenabil i

ty var ies accordingly.

E—partially

transformed region (PTR).

This region is character ized by regions of

near ly eutecto id carbon which cor re

spond to the or iginal networks of pearl i te

Fig. 7—Region C in Fig. 4 at

X500

(reduced

52 on reproduction)

Fig.

8 —Region D in

Fig.

4 at X500 (reduced


Fig.

9-Region

E in

Fig.

4 at X500 (reduced


100-s|APRIL 1 9 8 5



*f

>

• ^

-.

v

*^

• - - % ^

•r tf

%*

jr

-.

S *

(

-

-

E

'

:,

'A

'

*

*.,*

.»r»-«

"

"

" '

...IIP TR * -

' ' 'tjf

» *P >

•" "

.»: :,. •£ - £•

% . 10-Region F in Fig. 4 at X500 (reduced

46 on reproduction).

2%,

nital etch

• • •

..J .̂

*i* '

*

•

M

1

*

-.

j

-

^

|

j

'• 4 0 »

f /g .

11-Region

G in Fig. 4 at X500 (reduced


in the as-cast microstructure. This region

transforms to high-carbon austenite

whenever the peak temperature exceeds

the ef fect ive

A-i,

and the short t imes-

at- temperature prevent signif icant redis

t r ibut ion of the carbon. Therefore, the

cont inuous

prior-pearlite

ne tworks f o rm

cont inuous networks of h igh-carbon aus

tenite; upon cooling, these have suff icient

hardenabil i ty to t ransform to cont inuous

networks of h igh-carbon mar tens i te . The

hardness of 615 KHN is in the h igh-

carbon martensit ic regions and the low

values of 84 KHN are encou ntered in the

: , ( ,

— m * •

•Ai

:

'.:~3A

Fig.

14—Microstructure of

5 s stationary repair weld in heat 0. Nital etch, X75 (reduced 42 on

reproduction)

untransformed ferr ite islands. Figure 9

show s the loc at ion of E at X500.

F—spheroidized

region. In this region,

the peak temperature exper ienced

be tween the e f f ec t ive Ai and about

900-1000°F (482-538°C). The lamellar

carbides within the pearl i te nodules

spheroidize while the or iginal ferr ite

grains remain unchanged. Figure 10

shows this microstructure at X500. The

average hardness is 163 KHN in the pr ior

pearl i te and about 84 KHN in the pr ior

ferr ite.

C—unaffected base metal. This micro-

structure consists of cont inuous networks

of pearl i te nodules at the dendrite inter

stices as a result of the segregation of

carbon and alloying elements to the last

mater ial to solidify. This means, in ef fect ,

that the as-cast base metal, which has a

nominal carbon content of 0 . 3 1 % , b e

haves l ike a composite mater ial. As a

composite material, it consists of islands

of low carbon fer r i te gra ins (~0.025% C)

complete ly sur rounded by near ly cont in

uous networks of mater ia l wi th carbon

contents ranging f rom approx imate ly

0.5% (the carbon c onten t of l iquid steel at

the per itect ic temperature) to nearly

eu tec to id (~0.8%) composi t ion. F igure

<

o

_i

Q IOOO

Q

CC

<

I

o

o

UNTRANSFORMED

REGION OF

S C ST STRUCTURE

TYPIC L H RDNESS OF

PE RLITE NODULES

FfRJITS : " . » _ _ _ _

_L_

Q

CC

<

I

0 0 050 0 I00 0 I50

DISTANCE FROM FUSION BOUNDARY-inches

Fig. 12 -Results of hardness traverse across regions show n in Fig. 4

11

shows the m icrostruc ture of this zone

at X500. Note that the pearl i te lamina

t ions can be resolved in some areas at

this magnif icat ion.

Figure 12 summarizes the results of a

microhardness traverse across the region

shown in Fig. 4. Attent ion is drawn to the

large spread between the maximum and

minimum hardness values observed in

the part ial ly- t ransformed region. This

ref lects the nonuni form carbon content

of the martensite formed in this region.

Note that the areas of untransformed

ferr ite in this region exhibited hardness

values ranging from 77 to 94 KHN. An

approximate scale of

Re

hardness is

included at the right side of Fig. 12 for

comparison with calculated hardness

data presented ear l ier.

Homogenizat ion Stud ies

The results summa rized in the previous

sect ion clear ly indicate the dangers of

repair welding as-cast steel structures. It is

Fig. 13 — As-cast and normalized heat 0 micro-

structures: A

—

as-cast;

B

—normalized. 2 nital

etch, X250 (reduced 44 on reproduction)


1101-S



IOOO

o

o

o

o

m

9 5 0 0 -

cn

a

tr

<

I

a.

o

o

' *

0 0 . 0 5 0

0.100

D I S T A N C E F ROM F US I ON B OU NDA RY - in c h e s

Fig. 15 —Results of hardness traverse across the regions in Fig. 14

Fig. 18 — Tempered martensite in Fig. 17 at

X500

(reduced 52 on reproduction). Nital

etch

Fig. 16 —M icrostructure of 5 s stationary repair weld in normalized heat

27. Nital etch, X75 (reduced 55 on reproduction)

Fig. 17 — Microstructure of 5 s stationary repair weld in normalized heat

27 with postweld tempering at IIOVF (593°C) for 1 h. Nital etch, X75

(reduced 55 on reproduction)

vir tually impossible to choose welding

condi t ions which prov ide cool ing rates

suff icient ly slow to prevent the format ion

of cont inuous networks of crack-sensi

t ive microstructures in the regions which

experience part ial t ransformat ion to aus

tenite.

The hardness levels identif ied in these

regions are far above those shown to be

acceptable for avoid ing hydrogen-

induced crack ing even wi th good low-

hydrogen pract ice. Thus, i t would be

dif f icult , i f not impossible, to make repair

welds in as-cast steels without the risk of

forming microcracks in the high-carbon

mar tens i t ic networks. In fact, hindsight

n ow suggests that the lack of qu ant itat ive

cor re lat ion between the acoust ic-emis

sion studies and ident if iable hydrogen-

induced cracks is at t r ibutable to this phe

nomenon.

Unfortunately, microcracks in these

hard br it t le ne two rks in the part ial ly t rans

formed region can serve as nuclei for

s low crack growth even af ter postweld

heat t reatments. Therefore, i t appears

that the only sure way to avoid this

problem would be to e l iminate the net

works of interdendrit ic pearl i te nodules

by a homoge niz ing t reatment pr ior to

weld ing. Thus, 10 X 4 X V2 in. (254 X

102 X 13 mm) specimens from each heat

we re no rmalize d at 1750°F (954°C) for 2

h.

Figure 13 compares the as-cast and

normalized microstructures at X250. Note

the complete e l iminat ion of the cont inu

ous networks of pearl i te that results f rom

CC

F-

Q.

M

<

n

UJv

5

UJ /

r /

Q.

CO

Q

UJ

r

O UJ

w z

r

0

<

Z.

z>

->

CQ

~ I 0 0 0 ° F

Fig. 19—Schematic of various regions produced in steel casting repair welds: FZ—fusion zone;

UM Z — unmixed zone; PMZ—partially me/ted zone; HAR — homogeneous austenite region; HA Z—

heat-affected zone; CG—coarse grained; E.G.—fine grained; PTR — partially transformed

region

102-s | APRIL 1 985



Figure 14 shows a panorama of a weld

Figure 15 is a plot of the microhardness

se run on the w eld sho wn in Fig.

4. Comparison with Fig. 12 clear ly indi

Although these plots are for welds

1200 KHN in the weld m ade in the

Figures 16 and 17 show microstruc

wi t h and wi th out a postw eld temper

1100°F (593°C) for 1 h. The we ld in

16 was no t t empered and showed

hard

Re. However ,

Re.

Figure 18 shows this

The var ious zones, which are pro

lding of steel castings, are sum marized

gen -induc ed cracking is the p art ially

It is apparent from the above results

of hydro gen- induced crack ing

reases in bo th the alloying con tent and

Conclusions

The following conclusions pertain to

the hydrogen-induced cracking studies of

repair welds in steel castings.

1 .

The metallurgical factor was

con

t rol led by using the method descr ibed

previously to select combinat ions of

weld ing condi t ions which produced both

crack-sensitive and crack-insensitive mi

crostructures in typical case steel speci

mens.

2. The chemical factor was controlled

by modifying the moisture content of the

E7018 electrode coat ings.

3. The mechanical factor was synthe

sized by applying a controlled

aug

mented st ra in to SMA welds deposi ted

on simple rectangular specimens.

4 .

Acoust ic-emission techniques were

employed to monitor the init iat ion and

growth of hydrogen- induced crack ing.

5. The results conf irm the fact that

proper low-hydrogen pract ice is manda

tory for al l repair welding operat ions if

hydrogen-induced cracking in the coarse

grained heat-af fected-zone is to be

avo ided.

6. The suscept ibil i ty to hyd roge n-

induce d cracking is direct ly related to

both the maximum hardness in the heat-

af fected zone and the level of dif fusible

hydrogen .

7. Hydrogen-induced cracking can be

prevented by us ing good low-hydrogen

pract ice if the maximum hardness in the

heat -af fected zone is kept belo w Re 35.

8. Since the maximum hardness in the

heat-affected_ zone is direct ly related to

the carbon content , reduct ion in the

carb on con tent is the simplest way to

improve the weldabil i ty of steel cast

ings.

The following conclusions pertain to

the results of microstructural studies of

repair welds in steel castings:

1 . For cast steel repair welds one can

identify seven distinct regions by a nital

etch.

2. The part ial ly- t ransformed region of

the heat-af fected zone of repair welds in

as-cast steel was found to have the

microstructure most suscept ible to

hydrogen- induced crack ing.

3. Cont inuous n etworks wi th hard

ness levels ranging from 600 to 1200 KHN

(50 to 70 Re) were ident if ied in the

part ial ly- t ransformed region.

4 . The problem in the part ial ly- t rans

formed region ar ises from the interden

drit ic segregation present in the as-cast

material.

5. Normalizing eliminates the problem

in the part ial ly- t ransformed region by

homogenizat ion.

6. All steel castings should be normal

ized before repair welding to minimize

problems in the part ial ly- t ransformed

region.

7. As the alloy content is increased,

postweld-heat t reatment may become

desirable.

Acknowledgments

The authors wish to thank Mr. Dan

Stone and the Associat ion of American

Railroads (AAR) for its f inancial support of

this project .

References

1.

Rogers, H. C. 1956. The influence of

hydrogen on the yield point in iron. Ada Met.

4(2):114-117.

2. Savage, W . F., Nippes, E. F., and Homm a,

hi. 1976. Hydrogen-induced cracking in HY-80

steel w eldments. l/Ve/c//r?g,/oiv/7?,3/55(11):368-s

to 376-s.

3. Frohmberg, R. P., Barnett,

W.

)., and

Troiano, A. R. 1955. Delayed failure and

hydrogen embrittlement in steel. Trans. ASM

47:892.

4. Adams, C. M., |r. 1958. Cooling rates and

peak temperatures in fusion welding. Welding

lournal 37(5):210-s to 215-s.

5. Savage, W. F., Nippes, E. F., and Sze-

keres, E. S. 1976. Hydrogen-induced cold

cracking in a low-alloy steel. Welding lournal

55(9):276-s to 283-s.

6. Ball, D. )., Cestal, W . J., Jr., and Nipp es,

E. F. 1981. Determination of diffusible hydro

gen in weldments by the RPI silicone-oil extrac

tion method. Welding Journal 60(3):50-s to

56-s.

7. Savage, W. F„ and Lundin, C. D. 1965.

The varestraint test. Welding lournal

44(10):433-s to 442-s.

8. Aidun, D. K„ and Savage, W. F. 1984.

Optimizing repair welding techniques in cast

steels-Part

I.

Welding lournal63(11):345-s to

353-s.

9. Savage, W. F., Nippes, E. F., and Sze-

keres, E. S. 1976. A study of weld interface

phenomena in a low alloy steel. Welding

lournal 55(9):260-s to 268-s.

Optimizing Weld Repair in Cast Steels

Documents

Transcript of Optimizing Weld Repair in Cast Steels