Preheating in Welding Technology

A Hot TopicDetennination of the I'Right Temperature

DipJ.-lng. Marcus von Buseh, Germanischer lloyd AG

...------------------------.,

ime and time again, welded structures are found to exhibit cracks, although the development mechanisms - and hence also the possible countermeasures - have been studied widely and are now weil understood. A very effective way of preventing cold cracks is to preheat the weid area to higher temperatures in order to delay the cooling ofthe welded joint; this promotes an increased release of hydrogen after the welding within a shorter time period than when no preheating is applied. Ifthe cooling time t8/5 (i.e. the time taken to cool down from 800C to 500C) is long enough, excessive hardening in the heat-affected zone can be avoided. Moreover, preheating can minimize the residual stress condition. Depending on the welding conditions, "the mechanical and technological properties, in particular the hardness and toughness of the heat-affected zone [... ] can be influenced to a greater or lesser degree" [1] through preheating. If preheating is dispensed with entirely or is inadequate in scope, a good way of ensuring that the joint wil! be free of cracks is neglected. Although preheating is an effective means of reliably preventing cold cracks, it is frequently omitted or not carried out properly for reasons of cost. Ultimately, the decision on the necessity and level of preheating falls within the responsibility of the manufacturer. Today's standards provide the manufacturer with a number of methods of determining the "right" preheating temperature with a view to achieving a balance between technica! necessity and commercial interests. The established determination procedures include methods A and B according to EN 10112 and the Pcm method according to AWS Dl.l. The following description explains, from a praetical standpoint, how these methods can be applied to determining the preheating temperature. The temperatures thus obtained are then compared for the same input parameters. Furthermore, the possibilities of reducing the required preheating temperature are also indicated. Using shipbuilding as an example, it is shown how many welding tasks make it necessary to consider the topic of preheating. However, the statements made in this paper in respect of preheating do not apply exclusively to
shipbuilding; they rnay also be transferred to

Preheating Where on the Ship?

Ships are structures which are subject to highly dynamic loading. As a re sult, it is crucial to take full account of crack initiation and crack propagation. In genera!, weid joints must be free of cracks. For example, this is prescribed for ship structures in the Rules for Classification and Construction issued by Germanischer Lloyd. This publication also stipulates when and how much preheating must take place to prevent the origination of cold cracks after welding: "The need to preheat ferritic steels and the preheating temperature dep end on a number of factors. These include:

the chemica I composition of the base material (carbon equivalent) and the weid metal, the thickness of the workpiece and the type of weid joint (two- or three-dimensional heat flow), the welding process and the welding parameters (energy input per unit length of weid), the shrinkage and transformation stresses, and the diffusible hydrogen content of the weid metal." [2]
On the sh ip, there

are components

for which

other sectors.

one or more of these parameters play a role.

34

nonstop 02/2010

welded in single runs using high-performance methods sueh as electro-gas welding.

Stern Tube and Rudder Hom

The drive shaft for the propeller runs through the stem tube (see Fig. 2.1), while the rudder hom (Fig. 2.2) serves to accommodate the Tud-

der blade. In general. these components consist of weldable cast steel with material thicknesses of up to 400 mrn and are welded to the hull.

Main Engine
The foundation plat es, casing and frame of diesel engmes consist of cast steel or, in the case of the welded designs occurring more frequently nowadays, of norrnal-tensile steels with plate thicknesses of up to 300 mmo

Frequently, special attention must be given to thick-walled components or matenals with high tendency towards hardness.

Propeller
Ship propellers are made of cast copper alloys, e.g. CUI-CU4, or of non-corrodlng, and usually rnartensitic, cast steel alloys, e.g. 12CrlNi, 13Cr4Ni, 16CrSNi or 19CrllNi.

Huil
The huil of a seagoing vessel comprises, besides

the keel, bottom, outer shell and side structures, also the bow and stern sections as well as decks and bulkheads. In are as of high dynamic leading, e.g. the sheer strake or hatch side coaming of a containership, higher-tensile shipbuilding steels with the GL quality grades E36 and E40 are used. These are fully killed and fine-graintreated steels with plate thicknesses ranging between 50 and 100 mmo some of which are

Steam Boiler
High-temperature steels are used for constructing steam boilers, e.g. 16Mo3, 13CrMo4-5 or 11CrMo9-10 in accordance with CR 12187.

Cranes and Laad Suspension Devices

For crane components, higher-tensile shipbuilding steels and high-tensile fine-grained ~

FIGURE 2.1. Stern

tube.

rlGIJRE 2,2. Rudder hom,

F1GURE3. Submarin2 U32 of the German Navy.

nonstep 02/2010

35

FIGURE 4. Qualitative influences on the preheating temperature.

.' Preheating temperature reduced is Influencing factor Preheating temperature increased Higher alloying element content High Thick Low High Low is ,.
l;

Low alloying element content

Chemical composition of the base material

Low Thin High Low High

Hydrogen content of the weid metal Thickness of the workpiece or component Heat input during welding Residual stress condition Ambient or workpiece temperature (heat dissipation)

~ structural steels are used, offering yield strengths of up to 960 MPa.

Pressure Vessels
The pressure hullof a submarine (Fig. 3) consists of quenched and tempered structuraI steels meeting the German Naval Standard, such as the grades HY80or HYIOO.

Determining the Right" Temperature

11

As already mentioned in the introduetion. the manufacturer of a welded structure is faced with the challenge of producing a component that is crack-free - or at least crack-resistant - at acceptable cost. The necessary preheating temperature depends on the complexity of the component, the welding process used, the magnitude of the component's residual stresses and the ambient temperature. Fig. 4 shows the qualitative influence of various factors on the preheating temperature level. If a quantitative estimation of the preheating temperature is needed, a range of concepts are available for diverse applications. The guidelines of Germanischer Lloyd as an example for shipbuilding as wel! as the DIN 18800-7 standard for steel construction in Germany provide the manufacturer with an according choice of methods of determining the preheating temperature. Both rulebooks, however, give preferenee to a particular method.
Gl Rules for Classification and Construction:

(minimum preheating temperature and maximum interpass temperature) for (hull) structural steels may be determined in accordance with EN 1011-2." DIN 18800-7:"The required preheat temperatures are available in SEW088." As a matter of principle, this means the manufacturer is permitted to use competing methods to optimize the preheating temperature in his best interests. Verification th at a certain preheating temperature will yield the intended outcome (lower hardness and therefore a reduced crack susceptibility) can be provided by means of e.g. a welding procedure test. The following are possible methods of determining the minimum preheating temperature:
DIN EN 1011-2:2001 - Annex C, Method A (derived trom the British Standard) graphical determination DIN EN 1011-2:2001 - Annex e, Method B (based on SEW 88) computational determination AWS D1.1 - Annex XI (American method) computationalltabular determination JIS B 8285:2003 (Japanese method on the basis of the carbon equivalent eEN).

"The operating temperature to be maintained

Fig. 5 compares the concepts according to EN and AWS Dl.l. The Iapariese method will not be considered in this paper. The basis of all methods is the determination of a carbon equivalent. Since other alloying elements besides carbon also promote cold cracking, carbon equivalents are often used to estimate the crack sensitivity. CaIculation of the carbon equivalent is defined by numerous formulae, in which the various alloying ele1011-2

36

nonstop 02/2010

ENo1011~2MeltiodA
':
, ~~'

o'
.\ ,

.'0

,-"'J

~~

_.

,~<.

1. Carbon equivalent of the base

material (CE) 2. Hydrogen content ofthe welding consumable 3. Combined thickness

1. Carbon equivalent of the base

material (CE)

1. Carbon equivalent of the base material (Pcm)

FJGURE5. Qualitative influences on the preheating temperature comparison of methods.

2. Hydrogen content of the welding

consumable

2. Hydrogen content of the welding

consumable

3. Plate thickness 4. Heat input 5. Computational determination of

the preheating temperature

3. Plate thickness

4. Heat input 5. Graphical determination of the

preheating temperature

4. Residual

stress condition

5. Tabular determination of the preheating temperature

ments are weighted differently. In general, the carbon equivalent can therefore be viewed as a measure for the susceptibiliry of a materiaJ to cold cracking as a function of its chemica] composition.

EN 1011-2 MethodA
"Method A [ ... ] is based on extensive experience and data which is mainly, but not exclusively, for carbon manganese type steels. The prehearing conditions [... 1 have been found from experience to provide a satisfactory basis for denving safe welding procedures for many welded fabrications, [... ] They should therefore be considered for gudance only," [IJ
According to DIN 18800-7, for example, Meth-

od A may "not be used, since German experts are of the opinion that it offers inadequate assurance, partieularly when welding the S355 steels and the fine-grained structural steels with
Re > 355 MPa." [3J

surements and was derived under the assumpnon that alloying elemenrs conrriburing towards the hardening promote cold cracking to the same degree. Since the carbon equivalent CE, in contrast to newer carbon equivalents, seriously underestimates the effect of the carbon, it is less useful for addressing cold cracking problerns than later models. It is unsuitable specifically in the range of shorter cooling times." [SJ After determination of tbe carbon equivalent, the "scale" for the hydrogen content of the weId metal must be obtained from Table C.2 of EN 1011-2 (seeFig. 7). Information on the hydrogen content of a particular wel ding consumable may be found in the corresponding datasheet of the manufacturer OI printed on the packaging. The next step is to ascertain the "cornbined thickness". 'luis is ~

; Dlffusable hydrogetf
content mI/lOOg of
. deposited metal

...
A
B

provided by determining the carbon equivalent CE:

The basis of Method A is

" i(l:

CE = C + Mn + Cr+Mo+V + Cu+Ni 6 5 15

[4J

> 15

~,

10 $15 5 :slO 3$5

i~ ~~
1:

This formula is applicable tu steels with a CE value lying in the range 0.30 to 0.70 and for the maximum weight fractions of alIoying contents stated in Fig. 6. "The CE is based primarlly on hardness mea-

C
0

h.! !i@

flGURE 7.

Selection of the hydrogen scale 11I.

FIGURE 6. Scope of

applicability for CE.

nonstopOm016 37

\v FIGURE 8. Diagram from EN 1011-2.

200 180 160 E .~ 140 --

,----'!-r--r--r.....,---.,-r-r--.----.
--+--+-+--i--l--I-+-+---j Minimum preheating temperature in
--I---j 150 125

~ 120

-i-125( +---withoutI

H-i

I I

oe

P48P
\

,'>.

100 I

75 50 20

1--1

.~ 100

;S
~ c:

I II. .......,..-+--,I~.,...~.-

801-----j-~-~~~+S/~~

zs

BV,

CL,
LR,

~ : ~---.-:.:~-::~I----.-j
O~-~-~~-~-~-~~~
0.0 0.5 1.0 1.5

-I- --_

--=3.0

GL,
FIGURE 10. Ascertaining the hydrogen content from the packaging label (manufacturer: ELGA).

!
2.0 2.5

Heat input in kJ/mm

AC/De

'.1
350-400 "C, 2h

Ta be used lor carbon equivalent not exceeding

SSSSS)SSS

~ done by adding together all the plate thicknesses of the affected joint:
tcombined

= t1

the preheating line immediately above or to the left of the intersection of the lines for heat input and combined thickness.

+ t2 + t3 EN 1011-2 Method B
This method is applicable to the are welding of steels belonging to groups 1 to 4 according to CR ISO 15608. The basis of this method is provided by comprehensive examinations of the cold cracking behaviour of steels during welding, bath through special cold cracking tests and welded joint tests. As for Method A, the carbon equivalent must first be determined. However, a different formula is used here:

For directly opposed twin fillet weids that are deposited simultaneously, the combined thickness determined in this way must be halved. Af ter that, the heat input is calculated as follows from the are voltage, the welding current amperage and the rate oftravel:

Q = k- U[V] . I [A] . welding time [sec] = -..kL

length of weid [mm] . 1000 mm
where k is a correctioin factor for the welding process. For submerged are welding, k is set to 1. For MAG and flux-cored are welding, k = 0.8. In the case of multi-wire welding, the heat input must be calculated as the sum of the results obtained for each single wire, using the individual current and voltage parameters. [1] Depending on the relevant hydrogen scale (e.g. C - see Fig. 7) and the carbon equivalent (e.g. 0.43 - see Fig. 8), the appropriate diagram is chosen from EN 1011-2. The necessary preheating temperature is obtained from the chosen diagram by reading off

CET = C + Mn + Mo + Cr+Cu + Ni 10 20 40
which applies to the alloying contents stated in Fig. 9. By analogy to Method A, the hydrogen content of the welding consumable in ml/ 100 g of weid metal is also required; this can be taken directly from the datasheet or by reading off the standard designation on the packaging (see Fig. 10).

FIGURE 9. Maximum alloy contents for

CET.

max.o.71

38

nonstop 02/2010

The heat input is determined exactly as for Method A. The calculation of the prehearing temperature eau then be carried out according 10 the following formula:

residu al stress level. A distinction is made between three levels of the stress condition:
11

TrC] ;;:;697 eET + 160 .

tanh(3~)

11

+62HDo,35+(53CET-32)O-328
Thisrelationship applies to steels with a yield point of up to 1,000 MPa and eET:::0.2% to 0.5%. plate thickness d = 10 mm to 90 mm, hydrogen content HD ::: 1 m/11009 to 20 ml/100 g, heat input Q = 0.5 kJ/mm to 4.0 kJ/m.

11

low - simple fillet and butt weids with suffident possibility of shrinkage medium - weids with limited possibility of shrinkage that are already connected to larger components high - walds without any possibility of shrinkage (e.g. very thck plates or rep air weids).

lf the plate thickness (for the thickest plate of the joint) and the hydrogen seale are known, the prehearing temperature can be taken from the table in Fig. 12. The heat introduced by the welding process is not considered by the Pcm metlied.

Procedure According to AWS D1.1 -

Pem Methad
Here toa, a carbon equivalent is calculated as the input quantity. However, the formula for Pcm involves more elements than for CE or eET. For example, it also considers the element boron, which is a fine-gram agentthat ensures nuclei at high temperatures but at the same time has a powerful hardening effect. Baron is given 5 times the weighting of carbon in the calculation. The Pcm method is especiaJly suitable for short cooling times and for root welding appJications.

Comparison of the Methods

An example (Fig. 13) will be used to show what resuJts are obtained by the three methods for the sarne input quantities. A higher-tensile shipbuildtng steel of the grade GL-E36 (similar to S355 NL according to DIN EN 10025-3) with a thickness of 50 mm is chosen as the material. Various specimens are to be butt-welded with different heat inputs (l or 2.5 kJ/mm) and differentwelding consumables (hydrogen content: 5 or 10 mi/IOO g weld rnetal). The Iadle analysis yields the corresponding carbon equivalents (Fig. 14). By considering the plate thicknesses (Fig. 15), the prehearing ternperatures are determined according to the different methods
(Fig. 16). The benefit of reducing the available hydrogen is clearly recognizable for both of the methods according to EN 1011-2. The preheating tem-

Pan=C+~+Mn+ 30 20

CU+M+ Cr+Mo+V +5.8 20 60 20 15 10

The susceptibility index describing the sensitivity towards hydrogen-Induced cracks arising from the hydrogen content of the welding consumable ean then be determined (see Fig. 11). Finally, and here this approach differs from the methods described above, the preheating temperature is determined as a function of the

perature is reduced considerably. The same effect could also he expected ofthe Pcm meth- ~

Susceptibility Index _
Hl = 5 mll100 9 weid metal H2 = 10 ml/100 9 weid metal H3 30 mll100 9 weid (aU consumables not induded
in Hl or H2)

.
A

'-

Pem < 0.28

Prm < 0.33

Pcm,< 0.38

c
D

o
E

E F

metal

B C

FIGURE 11. Table tor determining the index tor ti: sU5ceptibility to ~.~, hydrogen-induced

J
--

cracks according

to

AW D1.1.

nonstop 02l261G39

FIGURE 12.lable for determining the minimum preheating temperature according toAWS Dl.l.

Restraint Ilevel

low
Describes comman fillet and groove welded joints in which areasonable freedom of movement of members exists.

Medium Describes fillet and groove welded joints in which because of members already attached to structural work, a reduced freedom of work exists. High Describes weids in which there is almast no freedom of movement for members joined (such as repair weids, especially in thick material).
!ij'. "",,, ..
"

<10 10 -20 20-38 38-75 > 75 < 10 10-20 20 - 38 38 -75 > 75 <10 10-20 20 - 38 38 -75 > 75

< 20 <20 < 20 20 20 < 20 < 20 20 20 95 < 20 < 20 20 115 115

< 20 < 20 < 20 20 20 < 20 < 20 20 80 120 < 20 20 85 130 130

!!l., .

< 20 20 20 40 40 < 20 20 75 110 140 20 35 115 150 150

< 20 60 80 95 95 < 20 80 110 130 150 40 105 140 150 150

60 100 110 120 120 70 115 140 150 160 110 140 150 160 160

150 140 140 140 140 140 145 150 150 160 150 160 160 160 160

150 150 150 150 150 160 160 160 160 160 160 160 160 160 160
~.,;#.lli!!

!'

"

~,

~:
"

.....
nalyse
Si
tC10-C!I9J

!T~~~~"lli~"'.

.-, ..~,~~"ll;",~""~;,",.;;,,,.

FIGURE 13.Ladle analysis of shipbuilding steel grade GL-E36.

Schmelzena
HIl"lteUeflngahen Schrneften-Nr.

toon
32726
S<:hmelzen-Nr. IB(7)

C
'Is
SO,J'

..

Mn
% O~.-I.liO

P
%
SO)IJS

S %
$0,.(135

% 0,004 2)
%

sMil

Al %
>D)l2O

Cu
%
<0)5

Cr
% :co,m 0,06

Ni %
~O,40

0,16 V
%
~O.IO

0,35 Nb
%

1,47 Me %
.s0,08

0,014

0,004 EV1 %
:!iIO.t2

0,041

0,09

0,06

snos 0,02

s,..,
%

Ti

1) EV2

32726
11 EV1: V,Nb-tTi

0,10

0,01

0,00

0,12
2J

0,45
EV2: CEV-=CtMn/bMt)'S+~!/IS+Cr/5+VIS+Ctr/1S

Ersehmelzungsverfahren: Ic/o)

Sauersteffaufblasverfahren

Slghted ~~ and ae ~:" 0"

t!lj ..........

,~
~

r"::\

~U\

e _~

'<_\
~

e)

-__- _.:::'::: -: _- \&67

.., -

~ od, but in this particular case it is overshadowed by the influence of the restraint condition (high for thick plat es), so that there is no reduction and the preheating temperature remains unchanged. A similar effect is to be seen with the heat input. High heat inputs tend to proAWS 01.1

duce lower preheating temperatures. With the Pcm method, the heat input through the welding process has, by definition, no influence on the result. The greatest impact, however, is caused by a reduction in the carbon equivalent or carbon content, because the preheating tempera-

FIGURE 14.
Comparison of the

CE 0.449

CET

Pcm

carbon equivalents.

0.317

40

non stop 02/2010

,.

Heat input

EN 1011-2 Method A

1.0 kJ/ 5 ml/ 100 9

mm

. ft:

'

FIGURE 16. Comparison of the preheating temperatures.

50

0 (

129(

2.5 kJ/
mm

No preheating

106(

150

0 (

1.0 kJ/ 10 ml/ 100 9

mm

125(
No preheating

159(

150

0 (

2.5 kJ/
mm

136(

150

0 (

ture exhibits a linear dependency on the carbon content of the base material (see Fig. 17). A decrease in the CETby 0.01 % reduces the required temperature by as much as 7.5 K! Through the steel order, the manufacturer therefore can exert the largest influence on the preheating temperature that is needed during production. By changing the material procurement from normalized plate to thermo-rnechanically rolled steel grades, the CETcan easily be reduced by 0.05% and thus the preheating temperature by approx. 40 oe. Iudging by Fig. 16, Methad A according to EN 1011-2 must be preferred for the ehosen example from the viewpoint of the welding contractor, since it demands the lowest preheating temperature. Depending on the conditions at hand, it mayalso be possible to reduce the temperature further by applying the measures deseribed above.

hardenabIe material or large dimensions, many eomponents suffer the risk of cold craeking. This was demonstrated by means of examples from the shipbuilding industry. A number of standardized methods with different approaches are available to determine the preheating temperature. In methods A and B according to EN 1011-2, the heat input during welding is taken into account, whereas this is neglected in the Pcm method aecording to AWSDl.l. Here the estimation of the preheating temperature is conducted in relation to the residual stress condition. The various methods may be used in competition with each other. The results obtained can then be verified before the start of produetion, e.g. with the aid of welding procedure tests. For economie reasons, the goal will be to keep the preheating temperature as low as possible. Through the welding process (heat input), the welding consumable (hydrogen supply) and the chemical composition of the material (carbon equivalent), welding production contractors have several effective ways of influencing the preheating temperature.

Literature
[11 DIN EN 1011-2 [2] Germanischer [3] Schmidt, Lloyd, Rules for Classification and Welding, and Construc1999) tion, 11- Materials 3 - Welding (Edition Zwtz, Br, Schulze - Execution [in German] of Steel Structur-

Summary
Preheating is an effeetive way of preventing the occurrenee of eold cracks. Owing to their

es I Notes on DIN 18800-7 [4] Technical Report

1967, IIW Doe. IX-535-67 of the cold

[5] Uwer, D. und H. Hhne - Characterization cracking behaviour of steels

during weid ing. SchweiBen [in German]

und Schneiden

43 (1991),

No. 4, p. 195-199

and IIW Doe. IX-1630-91 250 200 :,; 150 [6]lto, Y. und K. Bessyo - Weldability Steels, Related Formula of High Zone Crack ing.

Strength Sumitomo 631-69

to Heat-Affected

Search,

1 (1969), H. 5, p. 59-70, and IIW Doe. IX-

(1969),

1-18

,.!'"

.s Ei 100
50 0 0.15

Secondary Literature
[7] Jahrbuch Preheating 0.55 [8] TWI website drogen SchweiBtechnik [in German] - Job Knowledge for Weiders - Defects 1988 - 4.8 Welding Production -

I Hy-

Cracks in Steel and Cures Sonderdruck 41/01 - Preheating [in German] by Flame in the

[9] Kobelco Welding Today - Cold Cracks: Causes

FIGURE 17. Preheating temperature

as a function of the

[10] linde

carbon equivalent according to EN 1011-2.

Welding Workshop

nonstop 0212010 41