applied

sciences
Article
Metallurgical Aspects in the Welding of Clad
Pipelines—A Global Outlook
Ivan Bunaziv *, Vigdis Olden and Odd M. Akselsen
SINTEF Industry, P.O. Box 4760 Torgarden, NO-7465 Trondheim, Norway
* Correspondence: ivan.bunaziv@sintef.no; Tel.: +47-457-95-269




Received: 27 June 2019; Accepted: 28 July 2019; Published: 1 August 2019


Abstract: In the present work, the metallurgical changes in the welding of clad pipelines are studied.
Clad pipes consist of a complex multi-material system, with (i) the clad being stainless steel or a
nickel-based superalloy, (ii) the pipe being API X60 or X65 high-strength carbon steel, and (iii) the
welding wire being a nickel-based superalloy or stainless steel in the root and hot pass, with a nickel
or iron buffer layer, followed by filling with carbon steel wire. Alternatively, the corrosion resistant
alloy may be used only. During production of the clad pipe, at the diffusion bonding temperature,
substantial material changes may occur. These are carbon diffusion from the carbon steel to the clad,
followed by the formation of hard martensite at the interface on cooling. The solidification behavior
and microstructure evolution in the weld metal and in the heat-affected zone are further discussed for
the different material combinations. Solidification behavior was also numerically estimated to show
solidification parameters and resulting solidification modes.

Keywords: arc welding; clad steel; nickel-based superalloys; solidification; microstructure

1. Introduction
Clad or lined pipes represent cheap alternative to pipes fully made of corrosion-resistant alloys
(CRAs), which are very expensive materials including stainless steel or nickel-based alloys. As a
substrate, a carbon steel is used. Clad pipes are widely used in refineries and petrochemical industries,
fertilizer and chemical industry, power generation, environmental technology, seawater desalination,
pulp and paper manufacturing, the food industry, and shipbuilding [1]. Concerning the oil and gas
industry, clad and lined pipes, including the pipe-in-pipe method, have already been installed in the
North Sea and the Norwegian Sea. These are often made of carbon steel with an inner clad or liner to
protect the carbon steel against hydrogen embrittlement and sulfide stress corrosion cracking. The clad
and liner are most frequently made of austenitic stainless steel, often 316L, or a Ni-based superalloy,
typically Inconel 625. Clad pipes are mainly produced by hot roll bonding, explosion bonding, and weld
overlay. Hot roll bonding is by far the most widely adopted production method [2]. Nickel-based
superalloys also frequently used in the aerospace industry [3].
There are different methods of producing clad pipe, e.g., via a mechanical inner liner.
This technology is a cost-effective alternative to more expensive options such as solid CRAs or
metallurgically bonded clad pipe. The principles are the same as for metallurgically bonded clad
pipes, but involve being connected to the external carbon steel pipe through a mechanical bond.
Recent results [4] claimed that the liner separation is of minor importance and the use of the established
acceptance criterion derived for plain carbon steel pipes can be justified to apply to CRA-lined pipelines.
Both clad pipe and lined pipes can be laid by reeling. Although lined pipes are a cheaper solution than
clad pipes, the welding technology for the former case may be more challenging. Since the welding
solutions for reelable CRA-lined pipe may be considered as key enabling technologies for future
exploitation of deep-water oil and gas reserves, it will not be further addressed in the present work.

Appl. Sci. 2019, 9, 3118; doi:10.3390/app9153118 www.mdpi.com/journal/applsci

Appl. Sci. 2019, 9, 3118 2 of 24

Welding represents a very complicated joining process by the melting and subsequent solidification
of metal forming a hermetic and strong joint. During the process, all four states of matter are present
and interact. The most widely used methods are gas tungsten arc welding (GTAW) and gas metal
arc welding (GMAW). In the case of GTAW, filler wire is added externally to the weld pool where
the tungsten electrode is nonconsumable, while in GMAW, filler wire is added directly since it is a
consumable electrode. Both are very well suited to automated processes [5–7].
The welding of clad pipelines may encounter several challenges, specifically when damage or
leakage occur and the repair of clad pipes is required. Repair of a pipeline implies replacement of
the damaged part with a new pipe section to be joined at the seabed via remote control. Such subsea
operations are very expensive, which makes the robustness of the joining operation of crucial importance.
This led to the development of the welded sleeve concept for pipe repair, which was the only technique
where acceptable robustness could be achieved [8]. In a repair situation, the welded sleeve technique
involves the installation of an oversized pipe segment (sleeve) over the joint by threading the pipe
ends through the sleeve and performing a subsea fillet weld between the pipe and sleeve on each
end [9]. This technique is fully remotely controlled and has been proven by simulating conditions
in a hyperbaric chamber to work down to at least 2500 m depth, corresponding to 250 bar pressure
conditions, for the welding process. However, the sleeve solution is not acceptable for clad pipes due to
the direct exposure to the environment, as it may suffer from corrosion and/or hydrogen embrittlement
during service [10]. A comprehensive review is available for further studies [11] with cohesive zone
numerical modeling [12]. To retain the corrosion resistance of the clad pipe, the root and hot pass
should be deposited with a wire compatible with the clad material. This limits the freedom in the
selection of welding consumables.
In recognition of this situation, the development of hyperbaric butt-welding technology was
initiated with the aim to develop a fully remotely controlled dry welding technology for deep waters
(below about 200 m depth). In addition to the welding technology challenges, there are metallurgical
aspects that need to be considered. Therefore, the present paper addresses the metallurgical aspects in
the welding of clad pipes, involving clad materials such as AISI 316L SS and Inconel 625 (typically of 3
mm thickness) and the typical pipeline API X60/X65 high-strength carbon steels. With the addition
of the specific welding wire—Ni-based superalloys such as Inconel 625 and Alloy 59—a complex
multi-material system is formed that needs more comprehensive understanding. Inconel 625 filler wire
has been proven (high toughness at −30 ◦ C) for hot-tap welding operating in subsea conditions [13] in
joining X65 high-strength steel pipelines. Examples from an ongoing research project will be used
to illustrate various microstructure observations, concentrating on the clad, the heat-affected zone
(HAZ) of the pipe and the weld metal, or the fusion zone. Possibly, for onshore operations, a relatively
new method of friction drilling and form tapping can be used to reduce the harmful effect of heat on
welded joints [14].
Under hyperbaric conditions for pipe repair, from a practical standpoint, the use of multiple wires
will be very difficult, such as is frequently employed in conventional clad pipeline welding; i.e., a CRA
alloy for the root and hot pass, followed by a buffer layer of either Fe or Ni, and finally, filler passes by
a matching ferritic wire. In a fully remotely controlled hyperbaric operation, a much simpler solution
with one wire only will be employed throughout the weld. The use of backing gas is applicable since it
is purged into the chamber prior to welding to remove the seawater. The moisture and oxygen content
of the chamber gas must be carefully controlled continuously during welding. Moreover, the use of
one welding process for the root pass and another for the fillers will not be suitable. Although GTAW
is frequently used in the root pass welding of clad pipes due to the smooth surfaces, it is not applicable
in remotely controlled conditions. GMAW is therefore under development for future dry subsea pipe
welding, including the use of an inner clad.
The manuscript is organized as follows: description of the clad pipes and their production is
followed by an explanation of microstructure evolution principles (applicable for both clad pipe
production and welding), applied to the weld metal and HAZ microstructure evolution with different
Appl. Sci. 2019, 9, 3118 3 of 24

material
Appl. combinations,
Sci. 2019, 9, x a description of hyperbaric (repair in subsea level) welding conditions, and 3 of an
24
outline of the potential numerical simulation of the arc physics.
2. Production of Clad Pipes
2. Production of Clad Pipes
The metallurgical behavior of clad pipes depends on their production route during fabrication.
The metallurgical
Diffusion bonding representsbehavior of clad pipes
a frequently useddepends
technique. on their
During production
production, route duringtemperature
a certain fabrication.
Diffusion bonding represents a frequently used technique. During production,
range is required to achieve metallurgical bonding, usually from 0.5Tm to 0.8Tm (where Tm is the a certain temperature
range istemperature).
melting required to achieve metallurgical bonding,
The temperature–time cycle will usually fromthe
thus allow 0.5T m to 0.8T
diffusion (where T
ofm alloying m is the
elements
melting temperature). The temperature–time cycle will thus allow the diffusion
that may influence the structural integrity of the pipe. Thus, the interface between the two materials of alloying elements
that attain
will may influence
a gradient the
ofstructural integrity of the
chemical composition andpipe.
hence Thus, the interface
produce between the two
other microstructures materials
compared to
will attain a gradient of chemical composition and hence produce other
the two surrounding materials, as indicated in Figure 1. One example is shown in Figure 2, which microstructures compared
to the carbon
shows two surrounding
diffusion into materials,
the 316Las indicated
stainless steelin(SS)
Figure
clad 1. dueOneto theexample is shown in
higher solubility inaustenite.
Figure 2,
which
The shows
carbon carbonatdiffusion
content into the
the bimetallic 316L stainless
or clad/pipe steelon
interface (SS)
theclad
claddue sidetoisthe higher0.15
between solubility
and 0.20 in
austenite. The carbon content at the bimetallic or clad/pipe interface on the clad
wt %. A consequence is that extensive precipitation of Cr23C6 takes place [15]. It is also expected that side is between 0.15 and
0.20 wt
other %. A consequence
elements such as Ni,isCr, that
andextensive
Mo mayprecipitation
diffuse to the of interface,
Cr23 C6 takes place [15].
although It is also
to a lower expected
extent than
that other elements such as Ni, Cr, and Mo may diffuse to the interface,
carbon due to the lower diffusivity of these large-atom elements. Due to diffusion of alloying although to a lower extent
than carbon
elements, due to the lower
microsegregation maydiffusivity
facilitateof these large-atom
martensite formation elements. Due to Local
at the interface. diffusion of alloying
microhardness
elements, of
increases microsegregation
up to 350 HV may facilitate
has been found,martensite
as shown formation
in Figure at the
3, asinterface. Local microhardness
an indication of martensite
presence. Another influencing factor may be a deformation hardening during the hot-roll presence.
increases of up to 350 HV has been found, as shown in Figure 3, as an indication of martensite bonding
Another[16].
process influencing factor may be a deformation hardening during the hot-roll bonding process [16].
Measurements made
Measurements madeonon welded
weldedsamples gave gave
samples evidence that martensite
evidence was formed.
that martensite wasThe maximum
formed. The
microhardness values of 362 and 392 HV in the clad close to the interface
maximum microhardness values of 362 and 392 HV in the clad close to the interface for two different for two different macros was
reportedwas
macros [17].reported
Because[17].of the carbonofdepletion
Because the carbon of the pipelineofsteel
depletion the measured
pipeline steel by an electron by
measured probean
microanalyzer (EPMA), the hardness level on the steel side close to the interface
electron probe microanalyzer (EPMA), the hardness level on the steel side close to the interface is low is low (156 HV). As an
example, Line 3 in Figure 2 shows a carbon reduction from 0.16 to 0.03
(156 HV). As an example, Line 3 in Figure 2 shows a carbon reduction from 0.16 to 0.03 wt % C. wt % C. Moreover, some grain
growth of the
Moreover, somecarbon
grainsteel closeoftothe
growth thecarbon
interface wasclose
steel observed,
to the which
interface is also
wasconsistent
observed,with which theislower
also
hardness in this region. Therefore, remarkable gradients of hardness and
consistent with the lower hardness in this region. Therefore, remarkable gradients of hardness and microstructure occur across
the interface, which
microstructure occurmayacrosspotentially influence
the interface, which themay
integrity of theinfluence
potentially pipe. the integrity of the pipe.

Figure 1. Carbon
Carbon steel–clad
steel–clad (316L)
(316L) interface
interface microstructure
microstructure as observed by (a) optical
optical microscopy
microscopy and
(b) SEM; 1—the interface layer and 2—the width of the carbon diffusion
diffusion layer
layer [18].
[18].

In contrast to direct cladding with AISI 316, the applicationapplication of a 30–35 μm µm nickel interlayer
between the clad and the carbon steel reduces the carbon diffusion diffusion into
into the
the clad
clad substantially
substantially [19].
[19].
With the Ni-based alloy 625 as the clad material, there is a higher potential for elemental
the Ni-based alloy 625 as the clad material, there is a higher potential for elemental segregation
than with stainless
segregation steel.
than with However,
stainless steel.diffusion
However, of diffusion
Fe into theofclad is regarded
Fe into the cladasis aregarded
more critical issue.
as a more
critical issue. This is also the case in the cladding of carbon steel with Alloy 625. This alloy in
This is also the case in the cladding of carbon steel with Alloy 625. This alloy is frequently used is
the overlayused
frequently welding of overlay
in the carbon steels
weldingto improve
of carbonthesteels
corrosion resistance,
to improve where theresistance,
the corrosion dilution with
wherethe
substrate
the is kept
dilution withlow to preserve
the substrate is the
keptinitial
low tocorrosion
preserveresistance
the initialofcorrosion
the clad [20].
resistance of the clad [20].
Appl. Sci.
Appl. Sci. 2019, 9, x3118
2019, 9, 44of
of 24
24
Appl. Sci. 2019, 9, x 4 of 24

Figure 2. Variation of carbon content (wt %) across the clad/pipe interface [21]. Lines 1, 2, and 3
Variation of
Figure 2. Variation of carbon
carbon content
content (wt %) across the clad/pipe interface
interface [21]. Lines
Lines 1,
1, 2,
2, and
and 3
represent measurement distance and BM is base metal.
represent measurement distance and BM is base metal.
metal.

Figure 3.
Figure Increase in
3. Increase in hardness
hardness values
values at
at the
the bimetallic
bimetallic interface
interface [21].
[21].
Figure 3. Increase in hardness values at the bimetallic interface [21].
It should be noted that the chemical composition of the weld clad/overlay must comply with
It should be noted that the chemical composition of the weld clad/overlay must comply with the
It should be noted
the specification of the that the chemical
material. composition
In the case of Alloy of thethe
625, weld clad/overlay
maximum must comply
allowable with
Fe content is the
5%
specification of the material. In the case of Alloy 625, the maximum allowable Fe content is 5%
specification
(according of the
to API 5LD material. In the caseorof10%
[22] specifications) Alloy 625, thetomaximum
(according the DNV GL allowable Fe content
[23] standard). is 5%
To achieve
(according to API 5LD [22] specifications) or 10% (according to the DNV GL [23] standard). To
(according to API particularly
such low dilution, 5LD [22] specifications) or 10% the
the 5% requirement, (according
cold metal to transfer
the DNV(CMT)GL [23] standard).
variant of the gasTo
achieve such low dilution, particularly the 5% requirement, the cold metal transfer (CMT) variant of
achieve
metal arcsuch low dilution,
welding (CMT-GMAW)particularly
is an the 5% requirement,
advantageous process the cold
due to metal
lower transfer (CMT) variant of
heat input.
the gas metal arc welding (CMT-GMAW) is an advantageous process due to lower heat input.
the gas metal arc welding (CMT-GMAW) is an advantageous process due to lower heat input.
3. Microstructure Evolution
3. Microstructure Evolution
3. Microstructure
In the welding Evolution
of clad pipes, it is common to use corrosion-resistant alloys as consumables
In the welding of clad pipes, it is common to use corrosion-resistant alloys as consumables (filler
(fillerInwire), often Alloy
the welding of clad 59pipes,
or Inconel 625, andto304L/309/316L
it is common may be used
use corrosion-resistant forasthe
alloys root and the
consumables hot
(filler
wire), often Alloy 59 or Inconel 625, and 304L/309/316L may be used for the root and the hot pass.
pass. Thus, the weld metal (WM) will be different from the pipe material, which
wire), often Alloy 59 or Inconel 625, and 304L/309/316L may be used for the root and the hot pass. is typically X60 or
Thus, the weld metal (WM) will be different from the pipe material, which is typically X60 or X65
X65
Thus, carbon steel.metal
the weld Therefore,
(WM)the willwelded joint offrom
be different clad the
pipes consists
pipe of three
material, or four
which different
is typically materials.
X60 or X65
carbon steel. Therefore, the welded joint of clad pipes consists of three or four different materials.
carbon steel. Therefore, the welded joint of clad pipes consists of three or four different materials.
Hence, the microstructure and how it influences the final properties of the weldment is complex. In
Hence, the microstructure and how it influences the final properties of the weldment is complex. In
Appl. Sci. 2019, 9, 3118 5 of 24

Hence, the microstructure and how it influences the final properties of the weldment is complex. In the
following section, microstructures of the WM, HAZ, and clad will be discussed, starting with the
solidification of WM.

3.1. Solidification

3.1.1. Fundamentals of Solidification Theory

Welding of clad pipes is most often performed with an Inconel 625 or an Alloy 59 filler wire.
Sometimes, AISI 309 wire is employed for root and hot pass deposition. Stainless steel solidifies in
different modes according to its chemical composition. In addition, the cooling rate may induce a shift
in solidification [24]. The solidification mode of stainless steel can be divided into four types according
to the ratio between ferrite-(δ) and austenite-(γ) stabilizing elements, i.e., Creq /Nieq [25]:

F (ferritic) mode: L (liquid phase) → L + δ → δ → δ + γ for Creq /Nieq > 2.0 (1a)

FA (ferritic-austenitic) mode: L → L + δ → L + δ + γ → δ + γ → γ for 1.5 < Creq /Nieq < 2.0 (1b)

AF (austenitic-ferritic) mode: L → L + γ → L + δ + γ → γ + δ → γ for 1.37 < Creq /Nieq < 1.5 (1c)

A (austenitic) mode: L → L + γ → γ for Creq /Nieq < 1.37 (1d)

The values of Creq and Nieq can be calculated by a variety of equations. Olson [26] presented
more than 15 different equations for different purposes (welding, casting) and including/excluding
some of the microalloying elements, such as Cu and Co for Nieq and Nb, Mo, Ti, Al, V, W, and Ta
for Creq . However, the authors [27] had previously used equations by Hammar and Svensson [28] to
determine the solidification of stainless steel weld metals. The Cr and Ni equivalents can be calculated
from Equation (2a) and Equation (2b), respectively. Values are represented by the wt % of the elements.
Based on these equations, the solidified weld metal can be visualized (see Figure 4).

Creq = Cr + 1.37 Mo + 1.5 Si + 2 Nb + 3 Ti (2a)

Nieq = Ni + 22 C + 14.2 N + 0.31 Mn + Cu (2b)

Thus, it is indicated by extrapolation in Figure 4 that the classical austenitic stainless steels like
Appl. Sci. 2019, 9, x 6 of 24
AISI 309 SS and AISI 316 SS primarily solidify as ferrite.

Figure 4. Possible solidification modes and nomenclature according to Equations (1a)–(1d).

Figure 4. Possible solidification modes and nomenclature according to Equations (1a-d).
For Alloy 59, very limited information is available in the literature, while there are numerous
Table 1. Comparison of nominal chemical composition (wt %) of different materials.
publications on other Ni-based alloys. However, those are mainly precipitation-hardened alloys in
the 7xx
Alloyseries of Inconel
Ni alloys,
Cr frequently
Mo used in aerospace
Fe applications.
Nb Co Another
Mn frequently
Ti Si used
C
Alloy 59 59.0 22.0–24.0 15.0–16.0 ≤1.5 - ≤0.3 ≤0.5 - ≤0.1 ≤0.01
Inconel 625 58.0 20.0–23.0 8.0–10.0 5.0 3.15–4.15 1.0 ≤0.5 0.4 0.5 0.1
X65 0.5 0.02 - 98.5 0.05 - ≤1.4 0.04 0.25 0.07
316L 10.0–14.0 16.0–18.0 3.0 63.0–64.0 - - ≤1.3 - 0.75 ≤0.03

Several experimental studies have been conducted to investigate the behavior of the
Appl. Sci. 2019, 9, 3118 6 of 24

Ni-based alloy is Inconel 625, and the nominal chemical compositions of Alloy 59 and Inconel 625 are
outlined in Table 1. The main difference between them is the Nb content of Inconel 625 and the much
higher Mo addition in Alloy 59. The mechanical properties are also comparable, with yield strengths
of ≥340 and 350 MPa for Alloy 59 and Inconel 625, respectively. Their corresponding tensile strengths
are ≥690 and 770 MPa.

Table 1. Comparison of nominal chemical composition (wt %) of different materials.

Alloy Ni Cr Mo Fe Nb Co Mn Ti Si C
Alloy 59 59.0 22.0–24.0 15.0–16.0 ≤1.5 - ≤0.3 ≤0.5 - ≤0.1 ≤0.01
Inconel 625 58.0 20.0–23.0 8.0–10.0 5.0 3.15–4.15 1.0 ≤0.5 0.4 0.5 0.1
X65 0.5 0.02 - 98.5 0.05 - ≤1.4 0.04 0.25 0.07
316L 10.0–14.0 16.0–18.0 3.0 63.0–64.0 - - ≤1.3 - 0.75 ≤0.03

Several experimental studies have been conducted to investigate the behavior of the solidification
in superalloys containing niobium (Nb). Nb-bearing superalloys (like the popular Inconel 625/718)
terminate the solidification by a eutectic-type reaction between γ- and various Nb-rich phases, such as
NbC and Laves phases [29]. The creation of a crystal from an alloy melt causes a local change in the
composition [30]. When the dendrites solidify, elements such as Mo and Nb, as well as impurities,
segregate to interdendritic regions [31]. It was shown [32] that the presence of iron in Ni-based
superalloys lowers niobium solubility in the austenite phase. When iron is in a solid solution in
the γ Ni–Cr phase, the ability of Nb to remain in solution is limited [33] and partitioning of Nb
to interdendritic regions in the WM is increased. The solidification pattern follows Equation (3),
where primary dendrites of γ solidify first with a final microstructure of γ with NbC and Laves phases
precipitated in the interdendritic regions.

L → L + γ → L + γ + NbC → L + γ + NbC + Laves → γ + NbC + Laves (3)

Niobium-free alloys, such as Alloy 59, solidify by a simple L → γ transformation without any
eutectic-like reaction and exhibit a relatively narrow solidification temperature range. Solidification
cracking is therefore not expected.
When compared with carbon and low-alloy steels, Ni-based alloys have a high temperature
gradient (in WM) due to lower thermal conductivity, which is related to the high alloying level.
Therefore, a planar solidification front may not be stable, and the primary solidification mode is
cellular-dendritic, as shown in Figure 5. Here, only the grains with a crystallographic preferential
growth orientation (along the (100) direction, Miller index, for metals having a cubic crystal lattice
structure [34]) relationship aligned with the heat flow direction will continue to grow. When the
preferential growth direction of the solid starts to deviate significantly from the heat flow direction,
the growth stops, and new grains nucleate successively with the preferential growth direction nearly
along the heat flow direction. The criterion for constitutional supercooling for plane front instability
can be mathematically estimated [35] as follows:
The plane front will be stable when:
G ∆T0
≥ (4a)
R DL
Planar instability will occur when:
G ∆T0
< (4b)
R DL
Here, G is the temperature gradient (◦ C/m) in the liquid, R denotes the solidification front
(or crystal) growth rate (µm/s), ∆T0 is the equilibrium solidification temperature range (at composition
C0 ), and DL represents the solute diffusion coefficient (m2 /s) in liquid. The frequently used G/R ratio is
a solidification morphology parameter.
 𝐺 𝛥𝑇
(4b)
𝑅 𝐷
Here, G is the temperature gradient (°C/m) in the liquid, R denotes the solidification front (or
crystal) growth rate (μm/s), ΔT0 is the equilibrium solidification temperature range (at composition
Appl. Sci. 2019, 9, 3118 7 of 24
C0), and DL represents the solute diffusion coefficient (m2/s) in liquid. The frequently used G/R ratio
is a solidification morphology parameter.
In the diagram shown in Figure 5, the the range
range of
of solidification
solidification modes
modes are
are assumed.
assumed. Carbon steel
(e.g., X65) has planar solidification in a wide range of temperature
temperature gradients due to
to low
low alloy
alloy content.
content.

Figure 5. Schematic representation (modified from [34]) of the combined effect of crystal growth rate
and melt temperature gradient on the weld metal solidification
solidification microstructure
microstructure for
for different
different alloys;
alloys;
a—cellular,
a—cellular, b—cellular-dendritic,
b—cellular-dendritic, and
and c—equiaxed-dendritic
c—equiaxed-dendriticsolidification
solidificationmode.
mode.

rate εε ((°C/s)
◦ C/s) as follows [36]:
The product GR
The product GR is
is related
related to
to the
the cooling
cooling rate as follows [36]:

ε𝜀 = GR
𝐺𝑅 (5)
(5)
This means that the dendrite spacing increases with increasing cooling time, i.e., with increasing
This means that the dendrite spacing increases with increasing cooling time, i.e., with increasing
heat input in welding.
heat input in welding.
The solidification rate parameter (R) can be estimated empirically based on the angle (α) between
The solidification rate parameter (R) can be estimated empirically based on the angle (α) between
the welding direction and solidification growth direction, or molten surface normal, as shown in
the welding direction and solidification growth direction, or molten surface normal, as shown in
Figure 5. The solidification rate (2D, the top surface) is calculated by the following equation:
Figure 5. The solidification rate (2D, the top surface) is calculated by the following equation:
𝑅 𝑣 𝑐𝑜𝑠 𝛼 (6)
R = v cos α (6)
where v is welding speed.
whereThe
v issolidification
welding speed.rate for the 3D case can be calculated according to [37]:
The solidification rate for the 3D case can be calculated according to [37]:
R 𝑣
(7)
− ∂T
∂x
R=vr (7)
 2  2  2
∂T
∂x
+ ∂T
∂y
+ ∂T
∂z

It should be noted that Equation (5) does not represent a linear relationship due to the complex
heat flow in welding. However, it is assumed to be linear for simplicity to identify solidification
parameters (ε, G, and R).
 In order to estimate solidification parameters by finite element analysis (ABAQUS 6.14 with
DFLUX subroutine programmed in Fortran), combined double Goldak’s ellipsoid and surface
Gaussian heat source were used [38]. Different heat sources are needed to simulate the typical finger
penetration geometry of the GMAW welds, according to Yang and Debroy [39], since there is an
Appl. Sci. 2019, 9, 3118 8 of 24
impingement of droplets from the electrode into the weld pool which transports the additional heat.
A detailed modeling procedure and boundary conditions can be found in [40]. Applied
thermophysical
3.1.2. Numerical parameters
Simulationfor the materials
of Solidification are outlined in Table 2. The results of numerical
Parameters
simulation, achieved
In order by a trial-and-error
to estimate solidification method,
parametersarebypresented in Figure
finite element 6. The
analysis process
(ABAQUS parameters
6.14 with
wereDFLUX
set to 7subroutine
mm/s as the welding speed and 5.5 kW (220 A; 25 V) as the arc power. A good
programmed in Fortran), combined double Goldak’s ellipsoid and surface Gaussian agreement
withheat
previous
source experimental
were used [38].results (for
Different X70
heat carbon
sources aresteel)
needed welded under
to simulate thehyperbaric
typical fingerconditions
penetrationof 10
bar [41] was found.
geometry of the GMAW welds, according to Yang and Debroy [39], since there is an impingement of
droplets from the electrode into the weld pool which transports the additional heat. A detailed modeling
Table
procedure and2.boundary
Summaryconditions
of thermophysical properties
can be found in [40].forApplied
Ni-based alloys and various
thermophysical steels. for the
parameters
materials are outlined in Table 2. The results of numerical simulation, achieved by a trial-and-error
method, are presented Thermal
in Figure 6. The process parameters wereHeat
set to 7 mm/s Latent speed
as the welding Heat of
Density Specific Solidus
Material
and 5.5 kW (220(kgA; 25 Conductivity (W m agreement
−1 Fusion (J kg−1
m−3V)
) as the arc power. A good (J kg with
−1 °C −1) previous
Point experimental
(°C) results
°C−1)
(for X70 carbon steel) welded under hyperbaric conditions of 10 bar [41] was found. °C−1)
Inconel
625/Alloy 59Table 2.8440 Summary of thermophysical
10 properties for Ni-based
410 alloys and1290 various steels. 227000

[42] Density Thermal Conductivity Specific Heat Solidus Point Latent Heat of Fusion
Material
(kg m−3 ) (W m−1 ◦ C−1 ) (J kg−1 ◦ C−1 ) (◦ C) (J kg−1 ◦ C−1 )
316L SS [43] 7200 19 712 1424 274000
Inconel 625/Alloy 59 [42] 8440 10 410 1290 227000
X65 carbon 316L SS [43] 7800 7200 3219 712726 1424 1494 274000277000
steelX65[44]
carbon steel [44] 7800 32 726 1494 277000

Figure
Figure 6. (a)
6. (a) Simulatedweld
Simulated weldofofX65
X65carbon
carbon steel,
steel, and
and (b)
(b)fusion
fusionline
linegeometries
geometriesforfor
different materials.
different materials.

Due to different spatial temperature gradients and complex weld pool physics, the cooling rate as
Due to different spatial temperature gradients and complex weld pool physics, the cooling rate
well as solidification rate varies according to the specific location in the WM or the HAZ [43]. Therefore,
as well as solidification rate varies according to the specific location in the WM or the HAZ [43].
in the same WM, different solidification modes are common, as illustrated in Figure 5. At the WM
Therefore, in the
centerline, thesame WM, different
temperature gradient is solidification
smaller and the modes are common,
solidification as illustrated
rate faster when compared in Figure
to the5. At
the WM centerline,
area near the temperature
the FL (fusion gradient
line or liquid/solid is smaller
interface) and thematerial.
for each selected solidification
The most rate faster
widely usedwhen
compared to the
parameter to area nearcooling
describe the FLin(fusion
welding line
foror liquid/solid
carbon interface)
steel is the coolingfor each
time fromselected
800 C material.
◦ to 500 ◦ C The
most(denoted
widely asused∆t8/5parameter
), since mosttophase
describe cooling in take
transformations welding
place for carbon
within steel [45].
this range is theHowever,
cooling fortime
thefrom
800 °C to 500
crystal °C (denoted
growth rate in WMas Δtfor
8/5), since
any kindmost phase
of alloy, transformations
the cooling rate from take
theplace within this range
peak temperature to the [45].
However, for the crystal growth rate in WM for any kind of alloy, the cooling rate from the3.peak
solidus line is used. The finite element analysis (FEA) results at the centerline are compiled in Table
The centerline
temperature to theissolidus
illustrated
lineinisFigure
used.6.TheThefinite
solidification
element parameters
analysis (FEA) are outlined
results inat Table 4 for the are
the centerline
liquid/solid interface located at 1/4 of the weld pool length at the solidification
compiled in Table 3. The centerline is illustrated in Figure 6. The solidification parameters front, and in the case of are
the top plane, see Figure 7. For the mid-plane case, the angle of the G vector is more complex to find,
outlined in Table 4 for the liquid/solid interface located at 1/4 of the weld pool length at the
since it is normal to the tangential plane for a curved solid/liquid interface. Moreover, this surface is
solidification front, and in the case of the top plane, see Figure 7. For the mid-plane case, the angle of
constantly in motion because it follows the moving heat source.
interface. Moreover, this surface is constantly in motion because it follows the moving heat source.

Table 3. Estimated solidification parameters at the WM centerline on top-plane. Cooling rate is

calculated for temperature range from peak temperature to solidus point.
Appl. Sci. 2019, 9, 3118 Solidification 9 of 24
Temperature Gradient,
Material Cooling Rate, ε (°C/s) Growth Rate, R F-Factor = G/R
G (°C/mm)
(mm/s)
Table 3. Estimated solidification parameters at the WM centerline on top-plane. Cooling rate is
Inconel
1122
calculated for temperature 160 to solidus point. 7.0
range from peak temperature 23
625/Alloy 59
316L SS 925Rate, Temperature Gradient,
Cooling 132 7.0Growth
Solidification 19
Material F-Factor = G/R
X65 carbon steel ε ( C/s)
◦
874 ◦
G ( C/mm)125 7.0
Rate, R (mm/s) 18
Inconel 625/Alloy 59 1122 160 7.0 23
Table
316L4.SSEstimated solidification
925 parameters 132
at the liquid/solid interface,
7.0 i.e., solidification front.
19
X65 carbon
Cooling steel
rate is calculated874 125 from the peak temperature
for the temperature range 7.0 to solidus point.18

Solidification
Cooling
Table 4. Estimated Rate, Temperature
solidification Gradient,
parameters α Angle interface, i.e., solidification front.
at the liquid/solid
Material Growth Rate, R F-Factor = G/R
ε (°C/s) for the temperature
Cooling rate is calculated G (°C/mm) (Degree)
range from the peak temperature to solidus point.
(mm/s)
Inconel Cooling Rate, Temperature α Angle Solidification
Material 323 242 79 1.336 Growth 181 = G/R
F-Factor
625/Alloy 59 ε ( C/s)
◦ Gradient, G (◦ C/mm) (Degree) Rate, R (mm/s)
316L625/Alloy
Inconel SS 59 241 323 166
242 7978 1.455
1.336 114
181
X65 carbon
316L SS 241 166 78 1.455 114
X65 carbon steel 373 373 149
149 6969 2.509
2.509 5959
steel

Numerically simulated
Figure 7. Numerically simulated weld
weld pool
pool profiles
profiles from
from aa top-down
top-down view
view (middle
(middle plane) and
identification of the solidification angle α
α at specific points.
points.

Blecher et
Blecher et al.
al. [37]
[37] established
established aa solidification map of
solidification map of a
a Ni-based
Ni-based alloyalloy (Inconel
(Inconel 690). For the
690). For the top
top
of the
of the weld
weld pool, the RR and
pool, the and GG values
valuesare
aresimilar
similartotothe
theestimated
estimatedvalues
valuesfor forInconel
Inconel625/Alloy
625/Alloy5959(
(Table 3; Table 4). Since fiber laser beam welding in the keyhole mode was
Table 3; Table 4). Since fiber laser beam welding in the keyhole mode was applied, higher applied, higher G values
G valuesfor
the the
for deeper partpart
deeper of the weld
of the pool
weld were
pool obtained
were obtainedby by
WeiWei
et al. [46].
et al. Along
[46]. Alongthethe
weld centerline,
weld thethe
centerline, R
parameter is very similar to the welding speed, meaning that the solidification
R parameter is very similar to the welding speed, meaning that the solidification angle (α), used inangle (α), used in
Equation (6), is approximately equal to 0 (cos α = 1) at the centerline, and hence R → v. At the
centerline, the X65 steel has a smaller G value due to higher heat conduction. For the same reason,
the solidification rate for X65 is the highest for the materials studied here. It is important to note that
the R parameter is very sensitive to the α angle (based on Equation (6) and Table 4).
 Appl. Sci. 2019, 9, x 10 of 24

Equation (6), is approximately equal to 0 (cos α = 1) at the centerline, and hence R → v. At the
centerline, the X65 steel has a smaller G value due to higher heat conduction. For the same reason,
Appl.the
Sci. solidification
2019, 9, 3118 rate for X65 is the highest for the materials studied here. It is important to10note
of 24that
the R parameter is very sensitive to the α angle (based on Equation (6) and Table 4).
The primary dendrite arm spacing (λ1) is shown in Figure 8, and has a strong effect on
The primary dendrite arm spacing (λ1 ) is shown in Figure 8, and has a strong effect on mechanical
mechanical properties [36]. Based on Kurz and Fisher [47], a generalized (simplified) formula for the
properties [36]. Based on Kurz and Fisher [47], a generalized (simplified) formula for the primary
primary dendrite spacing is:
dendrite spacing is:
λ1 = 𝜆AG−m −n
𝐴𝐺R 𝑅 (8)(8)

here,here, m (=0.5)
m (=0.5) and n and n (=0.25)
(=0.25) are constants,
are constants, and A is aand A is a composition-dependent
composition-dependent (alloy)
(alloy) material material
parameter.
Since GR values are obtained and λ1 can be measured from micrographs, the A parameter can be
parameter.
estimated Since
based GRonvalues are obtained
Equation (8). and λ1 can be measured from micrographs, the A parameter can
be estimated
The secondary based on Equation
dendrite (8). (λ2 , Figure 8) may also strongly influence the mechanical
arm spacing
TheItssecondary
properties. dendrite
size will also dependarm
onspacing (λ2, Figure
the cooling 8) may also strongly influence the mechanical
rate [36]:
properties. Its size will also depend on the cooling rate [36]:
λ2 = B(GR)−n (9)
𝜆 𝐵 𝐺𝑅 (9)
where B is B
where a material-dependent parameter
is a material-dependent parameter n (=1/3)
and and is a is
n (=1/3) constant.
a constant.
Tertiary dendrites (λ
Tertiary dendrites3 (λ3, Figure 8) have minor effects, andrelated
, Figure 8) have minor effects, and some studies studies
some related can be found
can beinfound
[48]. in
Therefore, these are these
[48]. Therefore, not discussed in detail here
are not discussed due here
in detail to their
duelower importance.
to their lower importance.

(a) (b)

Figure 8. (a) SEM of cellular (columnar) dendritic structure in the WM of 304 SS [43] and (b) classification
of dendritic arc spacing (modified from [49]). λ1 , λ2 , and λ3 indicate primary, secondary, and tertiary
dendrite arm spacing, respectively.

3.2. Segregation in Welds

Figure 8. (a) SEM of cellular (columnar) dendritic structure in the WM of 304 SS [43] and (b)
Segregation is a of
classification defect in welded
dendritic joints,
arc spacing where the
(modified fromalloying
[49]). λ1,elements
λ2, and λ3are enriched
indicate within
primary, specific
secondary,
areas due totertiary
and motiondendrite
of liquid
armand can decrease
spacing, mechanical properties locally. Segregation is very
respectively.
common in stainless steels and Ni-based superalloys due to high alloy content, whereas low segregation
3.2. Segregation
tendency in low-carbonin Welds
steel is observed, except for the segregation of impurities (S, P).
TheSegregation
Scheil equations, or non-equilibrium
is a defect in welded joints,level
where rule, describe elements
the alloying the solute
areredistribution during
enriched within specific
solidification of an alloy [50]:
areas due to motion of liquid and can decrease mechanical properties locally. Segregation is very
common in stainless steels and Ni-based (1 − fS )(k−1) due to high alloy content, whereas
CS = kC0superalloys (10)low

andsegregation tendency in low-carbon steel is observed, except for the segregation of impurities (S, P).
The Scheil equations, or non-equilibrium
CL = C0 flevel
(k−1)rule, describe the solute redistribution during
(11)
L
solidification of an alloy [50]:
where CS is the element concentration in the solid, CL is the concentration in liquid, C0 represents the
initial concentration in the liquid, k is the𝐶effective
𝑘𝐶 1segregation
𝑓 (10)
coefficient (≤1.0), and fS and fL denote
the volume
and fraction of solid and liquid, respectively.
The segregation coefficient can be estimated as follows [50]:

CS
k= (12)
CL
 Appl. Sci. 2019, 9, x 11 of 24

The segregation coefficient can be estimated as follows [50]:

Appl. Sci. 2019, 9, 3118 11 of 24
𝐶
𝑘 (12)
𝐶

TheThesegregation
segregationcoefficient
coefficient kk can
can alsoalso be
be applied
applied inin determination
determinationofofmicrosegregation
microsegregation in in
solidification,i.e.,
solidification, i.e.,the
thedistribution
distribution ofof elements
elements between
between thethe core
core ofofprimary
primarydendrites
dendritesand andthethe
interdendritic
interdendritic areas.Based
areas. Basedon onFigure
Figure 9, 9, it
it is
is shown
shown thatthat Mo
Mo and
andNb Nbsegregate
segregatetotothe theinterdendritic
interdendritic
region, while Ni and to some extent Cr are enriched in the dendrites. Silva et
region, while Ni and to some extent Cr are enriched in the dendrites. Silva et al. [51] reported al. [51] reported that thethe
that
constant
constant k for
k for Ni and
Ni and Fe had
Fe had values
values just above
just above 1.0, which
1.0, which indicates
indicates a slighta slight segregation
segregation into
into the the
dendrite
dendrite core. A coefficient of 1.0 implies that no segregation occurs. In the case of Mo,
core. A coefficient of 1.0 implies that no segregation occurs. In the case of Mo, the coefficient is ~0.9, the coefficient
is ~0.9, resulting
resulting in segregation
in segregation to the metal
to the liquid liquid andmetalhence
and hence enriching
enriching the interdendritic
the interdendritic region.
region. Like
Like Mo,
Mo, Nb also segregates to the liquid, but with a greater intensity due to a lower k value of ~0.5.
Nb also segregates to the liquid, but with a greater intensity due to a lower k value of ~0.5. Therefore,
Therefore, this strong segregation of Nb has been appointed as reason for the formation of the
this strong segregation of Nb has been appointed as reason for the formation of the secondary phases
secondary phases such as the Laves phase in welding with Alloy 625 wire. For the use of Alloy 59
such as the Laves phase in welding with Alloy 625 wire. For the use of Alloy 59 wire, this will not be
wire, this will not be the case due to the absence of Nb. The extensive segregation in high-alloy
the case due to the absence of Nb. The extensive segregation in high-alloy materials can cause residual
materials can cause residual stresses during solidification and therefore reduced mechanical
stresses during solidification and therefore reduced mechanical properties [52].
properties [52].

Figure
Figure 9. 9.(a)(a)Elemental
Elementalchemical
chemical profile
profile crossing
crossing dendrites,
dendrites,as
asshown
shownbybySEM,
SEM,andand(b)(b)elemental
elemental
segregation during solidification process, as shown by EDS (energy-dispersive
segregation during solidification process, as shown by EDS (energy-dispersive X-ray X-ray spectroscopy).
spectroscopy).
The
The arrows
arrows indicate
indicate theinterdendritic
the interdendriticregions
regions[51].
[51].

Carbon
Carbon diffusion
diffusion toto the
the interfacearea
interface areahas
hasalready
alreadybeen
beenshown
showntototake
takeplace
placeduring
during production
production of
of the clad pipe. Transportation of the alloying elements may change the phase transformation
the clad pipe. Transportation of the alloying elements may change the phase transformation behavior
behavior at the interface. A temperature for the martensite start transformation (Ms) is strongly
at the interface. A temperature for the martensite start transformation (Ms ) is strongly influenced by
influenced by the chemical composition of the alloy, according to the following, i.e., the Andrews
the chemical composition of the alloy, according to the following, i.e., the Andrews equation [53]:
equation [53]:
(◦sC)
Ms M = 539
(°C) − −423
= 539 423CC−−30.4 Mn −
30.4 Mn Ni−−12.1
17.7Ni
− 17.7 12.1CrCr− −
7.57.5
MoMo (13)(13)
Different Ms temperatures can be used for Fe–Mn–Ni–Cr austenitic weld metal, which can be
Different Ms temperatures can be used for Fe–Mn–Ni–Cr austenitic weld metal, which can be
found in [26].
found in [26].
With a carbon content increase from 0.03 to 0.20 wt % C, the transformation temperature is
With a carbon content increase from 0.03 to 0.20 wt % C, the transformation temperature is reduced
reduced by◦around 85 °C. With segregation of the alloying elements, being Ni and Cr in the present
by around 85 C. With segregation of the alloying elements, being Ni and Cr in the present case, the Ms
case, the Ms temperature will decrease by 37 °C per wt % Ni and 11 °C per wt % Cr. Note that for clad
◦ C per wt % Ni and 11 ◦ C per wt % Cr. Note that for clad steels,
temperature will decrease by 37
steels, this will be the segregated concentration, not that of the initial bulk.
this willInbe the segregated
addition concentration,
to the influence not that of
from martensite the initial bulk.
transformation, the enhanced carbon diffusion to
the clad side causes extensive precipitation of chromium carbides the
In addition to the influence from martensite transformation, enhanced
at austenite carbon
grain diffusion
boundaries in to
thethe
clad side
clad causes
[54]. extensive precipitation
The precipitation kinetics of Crof 23Cchromium
6 is shown incarbides at austenite
Figure 10, grain boundaries
clearly showing how carbon in
theraises
clad [54]. The precipitation
the precipitation kineticsasofindicated
temperature, Cr23 C6 isbyshown in Figure
the dotted 10, clearly
red arrow. With showing how carbon
a carbon content of
raises the precipitation temperature, as indicated by the dotted red arrow. With a carbon
0.15–0.20 wt % in the clad close to the interface, carbides will spontaneously precipitate as soon as content
of 0.15–0.20 wt % in the clad close to the interface, carbides will spontaneously precipitate as soon
as the temperature in hot rolling or diffusion bonding of the clad pipe falls to around 900–950 ◦ C.
These reactions may have a negative impact of the corrosion resistance of the clad layer, since part of
the Cr content will be consumed by the precipitates.
Appl. Sci. 2019, 9, x 12 of 24
Appl. Sci. 2019, 9, x 12 of 24
the temperature in hot rolling or diffusion bonding of the clad pipe falls to around 900–950 °C. These
the temperature
reactions may have inahot rolling impact
negative or diffusion
of thebonding of the
corrosion clad pipe
resistance offalls to around
the clad layer,900–950 °C. of
since part These
the Cr
reactions
content willmay
be have a negative
consumed by theimpact of the corrosion resistance of the clad layer, since part of the Cr
precipitates.
Appl. Sci. 2019, 9, 3118 12 of 24
content will be consumed by the precipitates.

Figure
Figure 10.
Figure Precipitation
10.10. Precipitationkinetics
Precipitation kineticsof
kinetics ofCr
of Cr
Cr2323C
23 as aa function
C666 carbide as
C functionof
ofcarbon
carbon content
carboncontent
content [55].
[55]. The
The
[55]. The dotted
dotted redred
dotted red
arrow
arrow
arrow indicates
indicates
indicates more
more
more rapidkinetics
rapid
rapid kineticsdue
kinetics dueto
due carbon diffusion
tocarbon
carbon diffusion fromthe
diffusionfrom thecarbon
carbonsteel
carbon steelto
steel the
totothe clad.
theclad.
clad.

Selection
Selection of
Selection CRAs
ofof CRAsfor
CRAs forwelding
for weldingwires,
welding wires,such
wires, such as
such asAlloy
as Alloy
Alloy59 or
59or Inconel
orInconel 625,
625,oror
Inconel625, stainless
orstainless steel,
steel,
stainless will
will
steel, form
form
will form
welds
welds without
without solid-state
solid-state phase
phase transformations.
transformations. This means
means that
thatthe
thesolidification
solidification behavior
behavior
welds without solid-state phase transformations. This means that the solidification behavior controls controls
controls
thethe
the size
size
size and
andand shapeof
shape
shape ofofgrains,
grains,microstructure,
grains, microstructure, macro-
microstructure, macro- and
andmicrosegregation,
and microsegregation,distribution
microsegregation, distribution
distribution ofof
inclusions,
of inclusions,
inclusions,
defects
defects such
such as as porosity
porosity andand hot
hot cracks,
cracks, and
and ultimately
ultimately the
the mechanical
mechanical properties
properties
defects such as porosity and hot cracks, and ultimately the mechanical properties of the WM, of
of the
the WM,
WM, as well
as well
as as
well
as strongly
strongly influencing
influencing the the corrosion
corrosion resistance.
resistance.
as strongly influencing the corrosion resistance.

3.3.3.3.
WeldWeld Metal
Metal Microstructures
Microstructures
3.3. Weld Metal Microstructures
The TheWM WM microstructureininclad
microstructure cladpipes
pipes is
is complex
complex as as itit isis aa mixture
mixtureofoftwotwodifferent
differentmaterials
materials
The WM microstructure in clad pipes is complex as it is a mixture of two different materials
(mixed
(mixed zone)
zone) forfor
the the rootpass
root pass(see
(seeFigure
Figure11a),
11a), the
the filler
filler wire,
wire, and and thetheclad
cladmaterial.
material.For the
For thehot
hotpass
pass
(mixed
(see zone) 11b),
Figure for the the root pass (see
situation can Figure
be even 11a),
more the filler wire,asand
complicated a the clad
third material.
material, the For thesteel,
carbon hot pass
is
(see Figure 11b), the situation can be even more complicated as a third material, the carbon steel, is also
(see Figure
also 11b), theThe
incorporated. situation
use of can
the be
termeven
“hotmore
pass” complicated
is due to the asweld
a third material,
being depositedthe after
carbonthe steel,
root is
incorporated. The use of the term “hot pass” is due to the weld being deposited after the root pass
also incorporated. The use of the term “hot pass” is due to the
pass within limited timing. The Schaeffler diagram [56], DeLong [57] diagram, and WRC-1992 [58] weld being deposited after the root
within limited timing. The Schaeffler diagram [56], DeLong [57] diagram, and WRC-1992 [58] diagram
pass withinhave
diagram limited
limitedtiming.
use The
in the Schaeffler
predictiondiagram
of the [56],
final DeLong
microstructure [57] diagram,
of the WM and WRC-1992
for Ni-based[58]
have limited use in the prediction of the final microstructure of the WM for Ni-based superalloys due
diagram
superalloys due to very high Nieq values (>55–60 wt %) and low dilution with the carbonfor
have limited use in the prediction of the final microstructure of the WM Ni-based
steel. The
to very high Nieq values (>55–60 wt %) and low dilution with the carbon steel. The final microstructure
superalloys due to very high Ni
final microstructure will be the Ni-rich γ-phase. In addition to the chemical composition of the WM,The
eq values (>55–60 wt %) and low dilution with the carbon steel.
will be the Ni-rich γ-phase. In addition to the chemical composition of the WM, the cooling rate must
final
themicrostructure
cooling rate must willbebe the Ni-richIt γ-phase.
considered. should beInnotedaddition
that to thethe CMT chemical
processcomposition
is often usedoffor theroot
WM,
be considered.
thepass
cooling rateIt[59]
welding
should
must as be
be noted thatItthe
considered.
a low-heat-input
CMTbe
should
welding
process
notedItis
process.
often
that
tends
used
theto CMT for
provide
root
process pass
high is
welding
often
cooling used[59]
rates
as
forthe
in
a
root
low-heat-input
WM,
pass i.e., with
welding welding
[59]Δtas process.
8/5 values
It tends
even below welding
a low-heat-input to provide high
2.5 s at atmospheric cooling
process. It pressure, rates
tends to provide in the WM,
which tends i.e., with
hightocooling
decrease∆t 8/5 values
further
rates in the
even
WM, below
(to 0.6 2.5
i.e.,s)with
withΔts at atmospheric
increasing
8/5 values pressure
pressure,
even below [60].2.5
Suchwhich
s atfast tends to
cooling rates
atmospheric decrease further
may promote
pressure, (to 0.6
whichmartensite s) with increasing
formation.
tends to decrease further
pressure
(to 0.6 s) [60].
with Such fast cooling
increasing pressure rates may
[60]. promote
Such martensite
fast cooling rates formation.
may promote martensite formation.

Figure 11. (a) Root pass with double-material mixing and (b) hot pass with triple-material mixing
in welding.
Appl. Sci. 2019, 9, 3118 13 of 24

3.3.1. Root Pass Welding with Ni-Based Wire and Stainless Steel Clad
Appl. Sci. 2019, 9, x 13 of 24
Electron beam welding (EBW) of Inconel 617 and 310 SS provided a complex microstructure due
to nonequilibrium conditions and fast solidification [61]. Due to higher G/R values,
Figure 11. (a) Root pass with double-material mixing and (b) hot pass with triple-material mixing in the WM consisted
of small columnar
welding. dendrites, which were located near the edge towards the Ni-alloy solid surface.
The core of the dendrites consisted of Ni, Mo, and Nb. No cracks or other defects were found. At the
3.3.1. Root Pass Welding with Ni-Based Wire and Stainless Steel Clad
weld centerline, a uniform equiaxed cellular structure was obtained. The XRD analyses showed the
same γ-Ni matrix lattice
Electron beam welding
in both(EBW)
the base of Inconel
metal617andandthe310 SS provided
WM of Inconel a complex
625. The microstructure
WM hardness due was at
to nonequilibrium conditions and fast solidification [61]. Due to higher
the same level as Inconel 625 (230 HV), but dropped sharply to 150 HV close to the 304L SS. G/R values, the WM consisted
of small columnar dendrites, which were located near the edge towards the Ni-alloy solid surface.
Inconel 617 and 310 SS as base metals were joined by applying different filler materials, such as
The core of the dendrites consisted of Ni, Mo, and Nb. No cracks or other defects were found. At the
Inconelweld
82, Inconel 617, and 310 SS. Each filler alloy provided different dilution rates by GTAW [62],
centerline, a uniform equiaxed cellular structure was obtained. The XRD analyses showed the
and hence different
same γ-Ni matrix microstructures
lattice in both the were
baseformed
metal and in the
theWM WM. of With
Inconel the use
625. TheofWM Inconel 82 wire,
hardness was atthe WM
was fully
theaustenitic
same levelwith a substantial
as Inconel amount
625 (230 HV), of dispersed
but dropped intergranular
sharply to 150 HV close precipitates
to the 304L(NbCSS. and γ-Laves
eutectic). TheInconel
304617 SS and
filler310 SS asalso
wire baseprovided
metals were joined
fully by applying
austenitic different dendrites.
columnar filler materials,
Due suchto as
the high
Inconel
Nieq from the 82, Inconel 617, superalloy,
nickel-based and 310 SS. Each no δ-ferrite
filler alloywas
provided different
obtained. Thedilution
Inconel rates
617bywire
GTAW [62],
provided fine
and hence different microstructures were formed in the WM. With the use of Inconel 82 wire, the
columnar dendrites, where Mo-rich particles were found in interdendritic regions.
WM was fully austenitic with a substantial amount of dispersed intergranular precipitates (NbC and
With deposition of Ni-based tubular filler wire on the 316L SS base metal for underwater welding
γ-Laves eutectic). The 304 SS filler wire also provided fully austenitic columnar dendrites. Due to the
purposes
high Nieq no
[63], from cracking was observed
the nickel-based superalloy,duenoto austenite
δ-ferrite microstructure
was obtained. The Inconel(5–7% of ferrite
617 wire providedis needed
to prevent cracking)dendrites,
fine columnar with equiaxed dendrites
where Mo-rich withwere
particles appropriate toughness. regions.
found in interdendritic However, the tensile test
showed a decrease in mechanical
With deposition properties
of Ni-based tubularinfiller
WM,wirecorresponding
on the 316L to SS 85%
base tensile
metal for strength of BM (base
underwater
welding purposes [63], no cracking was observed due to austenite microstructure
metal) with high elongation (up to 32%) and ductile fracture mode, possibly due to lower hardness (5–7% of ferrite is
needed to prevent cracking) with equiaxed dendrites with appropriate toughness. However, the
in the WM. A sharp increase of the hardness was seen near the FL, specifically in the unmixed zone,
tensile test showed a decrease in mechanical properties in WM, corresponding to 85% tensile strength
due to significant iron and some chromium diffusion in from the BM.
of BM (base metal) with high elongation (up to 32%) and ductile fracture mode, possibly due to lower
Alloy 59 hasinbeen
hardness the WM.usedAoffshore for hot-tapping,
sharp increase while was
of the hardness bothseenAlloynear59 the
andFL, Inconel 625 are
specifically incandidates
the
for subsea tie-ins
unmixed anddue
zone, weld repair [13].
to significant ironThese
and sometwochromium
fillers will give similar
diffusion in fromweld
the BM. metal microstructures.
The main difference
Alloy 59 has is that
beentheusedLaves phase
offshore forand NbC arewhile
hot-tapping, found bothin Inconel
Alloy 59625 andwelds
Inconeldue625toarethe high
candidates for subsea tie-ins
Nb and C contents, as indicated in Table 1. and weld repair [13]. These two fillers will give similar weld metal
microstructures. The main difference is that the Laves phase and NbC are found in Inconel 625 welds
due to
3.3.2. Root the Welding
Pass high Nb and C contents,
with StainlessasSteel
indicated
Wireinand
TableStainless
1. Steel Clad
The useRoot
3.3.2. of AISI
Pass 309L SS with
Welding wireStainless
frequently Steelused
Wire for
andthe root and
Stainless Steel hot
Cladpasses showed a solidification
microstructure containing δ-ferrite, a darker phase as illustrated
The use of AISI 309L SS wire frequently used for the root and hot passes in Figure
showed12,a with a vermicular
solidification
morphology in the root and columnar morphology in the hot pass [19]. The matrix
microstructure containing δ-ferrite, a darker phase as illustrated in Figure 12, with a vermicular was austenitic,
and thismorphology
correlates well
in thewith
root the
and fact that 309L
columnar SS is a Cr-rich
morphology in the hotalloy,
passin[19]. The δ-ferrite
which matrix was is the first phase to
austenitic,
solidify.and thisformation
The of δ-ferrite
correlates well with the (a
fact
fewthatpercent,
309L SS ise.g.,
a Cr-rich
3–20%) alloy,
andin FA
which δ-ferrite is the firstsolidification
(ferrite-austenite) phase
to solidify. The formation of δ-ferrite (a few percent, e.g., 3–20%) and FA (ferrite/austenite)
mode is strongly recommended to avoid solidification cracking in welding with austenitic fillers [64].
solidification mode is strongly recommended to avoid solidification cracking in welding with
A reduction of impurities is very effective in avoiding cracking problems [57].
austenitic fillers [64]. A reduction of impurities is very effective in avoiding cracking problems [57].

Figure 12. Microstructure in the weld metal (a) root pass and (b) hot pass centers etched with oxalic
acid [21].
 Appl. Sci. 2019, 9, x 14 of 24
Appl. Sci. 2019, 9, x 14 of 24
Appl. Sci. 2019, 9, 3118 14 of 24
Figure 12. Microstructure in the weld metal (a) root pass and (b) hot pass centers etched with oxalic
Figure 12. Microstructure in the weld metal (a) root pass and (b) hot pass centers etched with oxalic
acid [21].
acid [21].
3.3.3. Root Pass Welding with Ni-Based Filler Wire and Inconel 625 Clad
3.3.3. Root Pass Welding with Ni-Based Filler Wire and Inconel 625 Clad
3.3.3. Root Pass Welding with Ni-Based Filler Wire and Inconel 625 Clad
For the combination of a Ni-based filler wire (Alloy 59/Inconel 625) and Inconel 625 clad layer,
For the combination of a Ni-based filler wire (Alloy 59/Inconel 625) and Inconel 625 clad layer,
the weld For the combination
metal microstructure of ashould
Ni-based notfiller wire (Alloy
be much 59/Inconel
influenced by the625) and of
mixing Inconel 625 cladand
clad melting layer,filler
the weld metal microstructure should not be much influenced by the mixing of clad melting and filler
the
wire weld
due metal
to to
their microstructure
similar should not be much influenced by the mixing of clad melting and filler
wire due their similarchemical
chemicalcompositions.
compositions.
wire
Thedue to their similar chemical compositions.
The solidification modeisisinfluenced
solidification mode influenced to to some
some degree
degree byby the
thecooling
coolingrate
rate[34].
[34].There
Thereis is also
also local
local
variationThe solidification mode is influenced to some degree by the cooling rate [34]. There is also local
variation in chemical composition at the subgrain level due to the microsegregation of alloying (Fe,Cr,
in chemical composition at the subgrain level due to the microsegregation of alloying (Fe,
variation in chemical composition at the subgrain level due to the microsegregation of alloying (Fe,
Nb,Cr,Mo,
Nb,and
Mo,Si)andand impurity
Si) and impurityelements
elements (S,(S,
P, P,
O,O,andandC)C)during
during solidification, leadingtoto
solidification, leading thethe second
second
Cr, Nb, Mo, and Si) and impurity elements (S, P, O, and C) during solidification, leading to the second
phase formation (e.g., carbides and Laves
phase formation (e.g., carbides and Laves phase phase for Inconel 625). These phases form
Inconel 625). These phases form at the end at the endof of
phase formation (e.g., carbides and Laves phase for Inconel 625). These phases form at the end of
solidification
solidification in the interdendritic regions. Composition analyses show that dendrite cores consist of of
in the interdendritic regions. Composition analyses show that dendrite cores consist
solidification in the interdendritic regions. Composition analyses show that dendrite cores consist of
Ni,Ni,
whereas
whereas interdendritic
interdendritic areas
areasmostly
mostly include
include thetheprecipitation
precipitationofof NbNb(carbides,
(carbides,Laves
Lavesphase)
phase)andandMo
Ni, whereas interdendritic areas mostly include the precipitation of Nb (carbides, Laves phase) and
Mo (Laves
(Laves phase) phase)
[65]. An [65]. An example
example is shown is in
shown
Figure in13.
Figure
The 13. The
Laves Laves
phase is phase
an is an intermetallic
intermetallic compound
Mo (Laves phase) [65]. An example is shown in Figure 13. The Laves phase is an intermetallic
of compound
the type A2of
compound B,the
of the
typeAAis
where 2B, where
type A2B,Ni or CrA
where
is NiB or
Aand
is Ni oris Cr and B is constituted
constituted by Mo or Nb.
Cr and B is constituted
by Mo or Nb.
by Mo or Nb.

Figure 13.Primary
Figure13. Primaryweld
weld metal
metal microstructure withInconel
microstructure with Inconel625
625filler
fillerwire.
wire.
Figure 13. Primary weld metal microstructure with Inconel 625 filler wire.
Due
Due toto
the segregation
the segregationofofalloying
alloying andand impurity elements,there
impurity elements, thereisisaapossibility
possibility forfor liquation
liquation or or
Due to the segregation of alloying and impurity elements, there is a possibility for liquation or
solidification cracking (Figure 14a) to occur in the welding of solid-solution-strengthened
solidification cracking (Figure 14a) to occur in the welding of solid-solution-strengthened alloys, as alloys, as well
solidification cracking (Figure 14a) to occur in the welding of solid-solution-strengthened alloys, as
as well
ductility-dip cracking (DDC); see Figure 14b. DDC is particularly common
as ductility-dip cracking (DDC); see Figure 14b. DDC is particularly common for multipass for multipass welding
well as ductility-dip cracking (DDC); see Figure 14b. DDC is particularly common for multipass
due to reheating
welding due to[65]. Regarding
reheating [65]. nickel-based alloys, Renalloys,
Regarding nickel-based et al. [66]
Renidentified
et al. [66]that the constitutional
identified that the
welding due to reheating [65]. Regarding nickel-based alloys, Ren et al. [66] identified that the
constitutional
liquation of M liquation
(B,C) (Mof M
can23 (B,C)
be Mo,6 (MCr,can
orbe Mo,
Fe) Cr, or
carbides Fe)
is carbides is
responsible responsible
for the for
liquation the liquation
cracking in
constitutional23liquation
6 of M23(B,C)6 (M can be Mo, Cr, or Fe) carbides is responsible for the liquation
WM.cracking
It wasinin WM.
noted It was noted that the liquation cracking was reduced with a decrease in heat input;
cracking WM.that thenoted
It was liquation
that cracking was cracking
the liquation reduced was
withreduced
a decrease
withinaheat input;inthe
decrease heat heat input
input;
the heat
must, input must,
therefore, be therefore,
optimized be optimized
during welding. during welding.
the heat input must, therefore, be optimized during welding.

Figure 14. (a) Solidification crack: the crack path and eutectic constituents are coexisting. The eutectic
constituent forms from the terminal solute-rich liquid at the end of solidification. (b) Ductility-dip
cracking (DDC) in multipass welding [65].
Appl. Sci. 2019, 9, x 15 of 24
Appl. Sci. 2019, 9, x 15 of 24
Figure 14. (a) Solidification crack: the crack path and eutectic constituents are coexisting. The eutectic
Figure 14. (a)forms
constituent Solidification
from thecrack: the solute-rich
terminal crack path and eutectic
liquid at theconstituents are coexisting.
end of solidification. The eutectic
(b) Ductility-dip
constituent forms from the terminal
cracking (DDC) in multipass welding [65]. solute-rich liquid at the end of solidification. (b) Ductility-dip
Appl. Sci. 2019, 9,(DDC)
cracking 3118 in multipass welding [65]. 15 of 24

Recently, it was found that significant reduction (by ~90 vol %) in Laves phases in welds can be
Recently,
achieved it was found
by applying that significant
an acoustic device which reduction
induces (bythe~90 vol %) in
vibration Laves phases
pool in welds acan be
Recently, it was found that significant reduction (by ~90 vol %)ofin the weld
Laves phases through
in welds high-
can
achieved
frequency by applying an acoustic device which induces atthe
a vibration of the weldrate pool[67].
through a high-
be achievedcurrent
by applying (the ultrasonic
an acousticAmpere’s
device which force)induces specific frequency
the vibration of the weld poolHere,
throughtwo
frequency
mechanisms current
are (the
involved ultrasonic
during Ampere’s
ultrasonic force)
treatment: at a specific
reduction of frequency
the local rate [67].
concentration Here,
of Nb two
and
a high-frequency current (the ultrasonic Ampere’s force) at a specific frequency rate [67]. Here,
mechanisms
specific are involved
(distorted) duringgrowth
Ni-dendrite ultrasonic treatment: reduction of the local concentration of Nb and
directions.
two mechanisms are involved during ultrasonic treatment: reduction of the local concentration of Nb
specific (distorted) Ni-dendrite growth directions.
and specific (distorted) Ni-dendrite growth directions.
3.3.4. Hot-Pass Welding
3.3.4. Hot-Pass
3.3.4. Hot-Pass Welding
Welding
In the hot pass, the three materials are mixed (as visualized in Figure 11b), namely X60 (BM),
In the
Ni-based
In theAlloy
hot pass,
hot pass, thewire,
59 filler
the threeand
three materials are mixed
clad material
materials are mixed
(316L(as
(as visualized
SS). Some part
visualized inof
in Figure 11b), namely
the previously
Figure 11b), namely X60 (BM),
deposited
X60 (BM),
root
Ni-based
pass will Alloy
also be 59 filler
melted, wire,
forming and clad
a material
quaternary (316L
mixing SS). Some
system. part
The of the
dilution
Ni-based Alloy 59 filler wire, and clad material (316L SS). Some part of the previously deposited root previously
rate will deposited
depend on root
both
pass will parameters
welding
pass will also be
also be melted,
melted, forming
andforming aa quaternary
bevel geometry, quaternary mixing system.
particularly
mixing system.
root gapTheThe dilution
anddilution
root face rate
rate will depend
height.
will depend on both
With excessive
on both
welding
welding parameters
carbon diffusion
parameters from and
and X60bevel
bevel geometry,
or from particularly
the carbon-enriched
geometry, root gap
particularly rootinterface and root
gap andbetween face
root facethe height. With
clad With
height. and BM excessive
of the
excessive
carbon diffusion
pipe, there
carbon is a risk
diffusion from
from X60
of X60 or
embrittlement.from the carbon-enriched interface between the
or from the carbon-enriched interface between the clad and BM of the pipe, clad and BM of the
pipe, there
there The is a
chemical
is a risk risk of embrittlement.
composition of single beads was determined by EPMA to include both root and
of embrittlement.
The chemical
hot pass,
The chemical composition
as showncomposition
in Figure 15. of For
of single
single beads
a 309
beads was determined
SS was
wire, determined
there is noby by EPMA to
significant
EPMA tochange
includeinboth
include both root and
the root and
chemical
hot pass, as
composition.
hot pass, as shown
shown
When in in Figure
Figure
using 15. For
Inconel
15. aa 309
625 wire
For 309(see
SS wire,
SS wire, there
Figure there is no
16), Ni
is no significant
significant
increases change in
significantly
change in the
the root
chemical
pass,
chemical
composition.
while Fe decreases.
composition. When
When using using Inconel
This isInconel
related625625 wire
to the (see
wiremelting Figure 16),
of a 16),
(see Figure Ni
partNi increases
of increases significantly
the pipe steel (X60) which
significantly in the
in thehasrootlow
root pass,
Ni
pass,
while
content. Fe decreases.
For the same This is
reason, related
a to
slight the melting
increase in of
Cr, a part
Mo, of
and the
Nb pipe
in
while Fe decreases. This is related to the melting of a part of the pipe steel (X60) which has low Ni steel
the root(X60)
passwhich
is also has low
observed, Ni
content.
while theFor
content. Fe the
For the same
same reason,
concentration reason, aa slight
slight increase
is substantially lower
increase in Cr,
inin
Cr,theMo,
Mo, and
root
and Nb
pass, in
Nbclearly the root
in the root pass
showing passvery
is also
is alsolowobserved,
dilution
observed,
while
with the
while the Fe
thepipe. concentration
On the contrary,
Fe concentration is substantially lower
the hot passlower
is substantially in
contains the
20–30
in the root
rootwtpass, clearly
% Fe,
pass, which
clearly showing
impliesvery
showing very low
a certain dilution
dilution
low dilution
with the
with the
with pipe.
the pipe.
pipe. OnOn the the contrary,
contrary, the the hot
hot pass
pass contains
contains 20–30
20–30 wt wt %% Fe,
Fe, which
which implies
implies aa certain
certain dilution
dilution
with the pipe.
with the pipe.

Figure
Figure 15.15. Electron
Electron probe
probe microanalyzer
microanalyzer (EPMA)
(EPMA) quantitative
quantitative analysis
analysis of X60
of X60 steel
steel pipe
pipe with
with AISI
AISI 309
Figure 15. Electron probe microanalyzer 309(EPMA)
SS fillerquantitative
wire [19]. analysis of X60 steel pipe with AISI
SS filler wire [19].
309 SS filler wire [19].

Figure 16. EPMA quantitative

Figure 16. quantitative analysis
analysis of
of X60
X60 steel
steel pipe
pipe with
with Inconel
Inconel 625
625 filler
filler wire
wire[19].
[19].
Figure 16. EPMA quantitative analysis of X60 steel pipe with Inconel 625 filler wire [19].
Appl. Sci. 2019, 9, x 16 of 24
Appl. Sci. 2019, 9, 3118 16 of 24

3.4. Heat-Affected Zone

3.4. Heat-Affected
The HAZ ofZone
the pipeline steel is another important aspect to consider, due to the grain growth
near the FL. As a result, the coarse-grained HAZ (CGHAZ) should be considered. The major
The HAZ of the pipeline steel is another important aspect to consider, due to the grain growth near
parameters which may control the final microstructure evolution of the HAZ are the weld cooling
the FL. As a result, the coarse-grained HAZ (CGHAZ) should be considered. The major parameters
rate, the clad pipe production method (e.g., TMCP or QT steel) [68], and the chemical composition of
which may control the final microstructure evolution of the HAZ are the weld cooling rate, the clad pipe
the material [65].
production method (e.g., TMCP or QT steel) [68], and the chemical composition of the material [65].

3.4.1. High-Strength,
3.4.1. Low-AlloySteel
High-Strength Low-Alloy SteelHAZ
HAZ
Modern pipeline
Modern pipeline steel
steel normally
normally has has low
low carbon
carbon content
content (≤0.06
(≤0.06 wt
wt % C). The
% C). The final
final HAZ
HAZ
microstructure can be predicted by a CCT (continuous cooling transformation) diagram,
microstructure can be predicted by a CCT (continuous cooling transformation) diagram, e.g., for X70 e.g., for X70
steel (see
steel (see Figure
Figure 17),
17), where
where it
it can
can be
be seen
seen that
that aa fully
fully martensitic
martensitic microstructure
microstructure is
is formed
formed upon
upon fast
fast
cooling (Δt < 1.4 s) with a hardness level of ~340 HV. At slower cooling rates (>20 s), coarse
cooling (∆t8/5 < 1.4 s) with a hardness level of ~340 HV. At slower cooling rates (>20 s), coarse bainite
8/5 bainite
(~220 HV)
(~220 HV) is
is formed.
formed. This
This type
type ofof microstructure
microstructure may may give
give low
low toughness
toughness at
at subzero
subzero temperatures.
temperatures.

Figure 17. CCT

Figure 17. CCT (continuous
(continuous cooling
cooling transformation)
transformation) diagram
diagram for
for X70
X70 steel
steel [69].
[69].

In the case of multipass welding (filling passes), reheating of the CGHAZ may induce the formation
In the case of multipass welding (filling passes), reheating of the CGHAZ may induce the
of brittle martensite–austenite (M-A) islands, i.e., the intercritically reheated CGHAZ (ICCGHAZ).
formation of brittle martensite–austenite (M-A) islands, i.e., the intercritically reheated CGHAZ
Hence, very low toughness at subzero temperatures may occur [70,71]. Afterwards, a comprehensive
(ICCGHAZ). Hence, very low toughness at subzero temperatures may occur [70,71]. Afterwards, a
study by Mohseni et al. [72–75] and Haugen et al. [76] on brittle crack initiation in the ICCGHAZ
comprehensive study by Mohseni et al. [72–75] and Haugen et al. [76] on brittle crack initiation in the
confirmed earlier results [77].
ICCGHAZ confirmed earlier results [77].
3.4.2. HAZ of Welded 316L Stainless Steel Clad
The HAZ of austenitic stainless steels (e.g., 304L/309/316L) consists of an austenitic matrix after
solidification with almost negligible grain growth when compared to carbon steels. The grain size of
the 316L SS base metal is ~45–75 µm, where the variation highly depends on manufacturing route [78].
When compared to fine-grained carbon steel having grain size smaller than 10 µm [79], there is a much
lower driving force for grain growth [80]. In addition, upon cooling, some delta ferrite may form at
the HAZ grain boundaries, suppressing grain growth [57]. This results in a minor hardness increase
near the FL [63]. However, despite the low susceptibility for grain growth, high heat input should
be avoided, especially when the BM has smaller grains that can cause significant grain growth even
in austenitic steels with subsequent softening [81]. In addition, due to high heat input, there is the
formation of chromium carbides on the grain boundaries of austenite. As a result, the area around the
carbides has lower chromium content compared to the bulk material. This can lead to the sensitization
phenomenon, which reduces corrosion resistance significantly [57].
Due to dissolved precipitates, some carbides and nitrides can form upon cooling along grain
boundaries in the HAZ or at the ferrite–austenite interface, especially when the stainless steel contains
Appl. Sci. 2019, 9, 3118 17 of 24

Ti and Nb. This can cause grain boundary liquation by lowering the melting temperature, in addition
to the segregation of impurities (primarily P and S), causing liquid films upon cooling. Notably,
between the FZ and the HAZ, there is the partially melted zone (PMZ) due to temperature gradients,
which has low mechanical properties [82]. Grain growth in this zone is not significant and the grain
size is similar to the as-received BM [83]. The thickness of the UMZ (unmixed zone) (<0.5 mm) can be
reduced by applying mechanical vibrations during welding [84].

3.4.3. HAZ of Ni-Based Superalloy Clad

Ni-based superalloys are divided into solid-solution-strengthened (substitutional alloying
elements: Cr, Fe, W, Cu, Mo, Co, Ta, and Re; among them are 59, 617, 625, 690, 800, 825,
and 902 alloys [65]) and precipitation-strengthened (precipitation formation elements: Ti, Al,
and Nb; among them are 300, 263, 713, 718, 725, 751, 925, 945, and Rene 41 alloys [65]) categories,
the latter of which are not frequently used for cladding. Therefore, further metallurgical aspects for
solid-solution-strengthened alloys only are discussed.
Due to the high alloy level of Inconel 625 (see Table 1), additional metallurgical reactions may take
place when it is used as a clad. In the HAZ, these may consist of the dissolution and reprecipitation of
NbC, grain boundary segregation, and liquation phenomena [65]. A segregation mechanism of some
elements (S, P, Pb, and B) has the most profound effect on liquation in the HAZ, and the principal
Appl. Sci. 2019, 9,
mechanisms arex illustrated in Figure 18. The same mechanisms are valid for stainless steels. 18 of 24

Figure 18. Solute segregation and localized melting along grain boundaries in HAZ during different
Figure 18. Solute segregation and localized melting along grain boundaries in HAZ during different
temperature phases in welding [65]; 1, 2, and 3 define progression phases of liquation cracking.
temperature phases in welding [65]; 1, 2, and 3 define progression phases of liquation cracking.

Grain growth in welding of these alloys does not present any problems, since the initial grain
3.4.4. Bimetallic Interface After Welding
size is coarse (~20–40 µm [85]). Accordingly, the driving force for grain growth is very low and
similarThetointerface
stainlessbetween
steels. Athe 316L SS clad
secondary factorand high-strength,
is the interaction low-alloy (X60/X65)
of secondary phasesbase
(e.g.,metal has
borides,
already been
sulfides, discussed
and carbides in Section
such 2, as
as TiC and the majority
NbC) with grain of boundaries
the elemental
by diffusion and segregation
a constitutional has
liquation [86];
taken
this is place
calledduring the clad
penetration pipe production.
mechanism In hot-roll
[87]. During welding(metallurgical) bonding
(rapid heating), of the phases
secondary clad pipe,
do
carbon diffuses at elevated temperatures (>A point in the Fe–C diagram) from the carbon-rich
not dissolve within the matrix, and upon heating, they react with the matrix by forming interfacial
3 ferrite
to the austenite, with subsequent carbide precipitation of M23C6, mainly Cr23C6 [90]. This diffusion
results in a softer, carbon-depleted zone in the ferrite (decarburized layer in the pipe steel close to the
interface), and a hard, carbon-enriched zone (carburized layer) in the 316 SS (austenite) clad [91].
Moreover, this phenomenon makes the clad sensitized, and the existence of Cr-depleted zones may
promote intergranular corrosion [92].
Appl. Sci. 2019, 9, 3118 18 of 24

liquid films [65]. However, small-scale M23 C6 particles dissolve at high temperatures during GTAW
welding [88]. NbC and TiC can also be responsible for hardness and strength increases, based on
experimental observations made by GTAW of 2 mm Inconel 625 [89]. As a result, a high-heat-input
process must be avoided.
The liquation cracking may occur in the partially melted zone (PMZ) in the HAZ adjacent to the FL.
The crack length has been observed to increase with a decrease in heat input [66]. Preheating suppressed
the crack length, possibly by reducing the cooling rate. Therefore, the cooling rate must be adjusted to
avoid the liquation cracking.
Another factor contributing to crack forming is stresses built up during solidification and further
cooling [65].

3.4.4. Bimetallic Interface After Welding

The interface between the 316L SS clad and high-strength, low-alloy (X60/X65) base metal has
already been discussed in Section 2, as the majority of the elemental diffusion and segregation has
taken place during the clad pipe production. In hot-roll (metallurgical) bonding of the clad pipe,
carbon diffuses at elevated temperatures (>A3 point in the Fe–C diagram) from the carbon-rich ferrite
to the austenite, with subsequent carbide precipitation of M23 C6 , mainly Cr23 C6 [90]. This diffusion
results in a softer, carbon-depleted zone in the ferrite (decarburized layer in the pipe steel close to
the interface), and a hard, carbon-enriched zone (carburized layer) in the 316 SS (austenite) clad [91].
Moreover, this phenomenon makes the clad sensitized, and the existence of Cr-depleted zones may
promote intergranular corrosion [92].
After welding, interface cracking has been found between the BM (X60) and the 316L SS clad
material, as shown in Figure 19a. It was observed that the cracking was usually initiated in the HAZ.
Upon further inspection, it was found that there is a grain boundary precipitation on the clad side
near the interface (see Figure 19b) [93]. The electron microprobe analysis revealed high carbon content
in the clad near the interface, which increases the risk of cracking [21]. This possibly takes place
due to additional stresses caused by welding; i.e., residual stresses added to production thermal
stresses because of thermal expansion mismatch between the 316L stainless steel clad and the pipe
steel. Fortunately, the cracks occurred in the direction following the pipe length, where the load is
low, and were not regarded as representing any risk of pipeline failure. The cracking problem can
be prevented using a Ni interlayer [17]. To the authors’ knowledge, similar cracking problems with
Inconel
Appl. 625 have
Sci. 2019, 9, x not been reported. 19 of 24

19. (a) Cracking

Figure 19.
Figure Cracking near
near the
thebimetallic
bimetallicinterface
interfaceininthe
theclad
cladand
and(b)(b)
grain boundary
grain precipitates,
boundary as
precipitates,
indicated
as by by
indicated thethe
redred
arrow, near
arrow, thethe
near interface [21].
interface [21].

4. Potential Application of Hyperbaric Weld Repair

Repair welding of clad and lined pipes requires advanced and robust hyperbaric welding
processes. It is thus critical to understand the physical phenomena of the welding process as well as
optimize the welding parameters to obtain high-quality welds. During the last decade, significant
progress has been made in the field of modeling of the heat and mass transfer during GMAW. Models
Appl. Sci. 2019, 9, 3118 19 of 24

4. Potential Application of Hyperbaric Weld Repair

Repair welding of clad and lined pipes requires advanced and robust hyperbaric welding processes.
It is thus critical to understand the physical phenomena of the welding process as well as optimize the
welding parameters to obtain high-quality welds. During the last decade, significant progress has
been made in the field of modeling of the heat and mass transfer during GMAW. Models including
the computation of current density, temperature, and momentum transfer, including the transfer of
individual droplets, have been presented [94–105].
The complex interaction of the shielding gas and current with weld bead appearance and
temperature distribution has been studied. However, no detailed modeling of the arc physics and
metal transfer under hyperbaric conditions has been found in the literature. One could expect that both
the arc pressure and the effect of droplet impingement would be stronger in the case of a more confined
or constricted arc with higher shielding gas density. There are five main flow-driving forces during
arc welding processes that should be investigated: the gradient of surface tension (Marangoni force),
the gas drag force at the surface, the buoyancy, the electromagnetic (Lorentz) force, and the moment
transferred by impinging droplets in the weld pool. In the case of CMT, the droplet momentum is
strongly reduced, but there could be other challenging effects from the large variations in current and
changes in the boundary conditions for the plasma arc. So far, no papers with numerical simulation of
the CMT arc mode physics have been published, which is probably related to the high complexity of
the process.

5. Conclusions
The present work has addressed metallurgical changes in the welding of clad pipes, and several
conclusions may be drawn:

• Complex metallurgical phenomena present in the welding of clad pipes due to multi-material
systems, especially concerning the weld metal. Clad pipes represent a solid economic alternative
to pipes made of stainless steel which is vital to the oil and gas industry.
• Welding process significantly alters the base metal near the weld metal, i.e., the heat-affected zone.
Therefore, in complex clad pipes, utilizing expensive materials must be considered in detail.
• Solidification parameters have very high importance on solidification behavior and resulting
mechanical properties. Moreover, they can be efficiently estimated by numerical simulation at
any point in welded joints.
• The subsea clad pipe network requires an emergency method for pipe repair to be developed.
Therefore, the current manuscript represents an important contribution to the knowledge on
microstructural changes and diffusion that may occur in welding, and how these may influence
the pipeline integrity.
• During repair under hyperbaric conditions, the use of multiple filler wires is very complicated
for clad pipes. Therefore, the suitability of filler wire for multi-material systems, both for
carbon and nickel-based/stainless steel materials, must be carefully investigated for single filler
wire applications.

Author Contributions: I.B. wrote the manuscript and performed numerical simulations. V.O. and O.M.A.
supervised the further improvement of the manuscript.
Funding: The authors wish to thank the Research Council of Norway for funding through the Petromaks 2
Programme, Contract No. 234110/E30. The financial support from Statoil Petroleum AS, Gassco AS, Technip Norge
AS, EDF Induction AS, and Pohang Iron and Steel Company Posco is also acknowledged.
Conflicts of Interest: The authors declare no conflict of interest.
Appl. Sci. 2019, 9, 3118 20 of 24

References
1. Lipp, W.; Racu, A.; Moll, M. The World Market for Clad Pipe; SMI (Steel Market Intelligence) GmbH: Pflach,
Austria, 2016; p. 256.
2. Smith, L.M.; Celant, M. CASTI Handbook of Cladding Technology, 2nd ed.; CASTI Pub: Edmonton, AB,
Canada, 2000.
3. Calleja, A.; Tabernero, I.; Ealo, J.A.; Campa, F.J.; Lamikiz, A.; López de Lacalle, L.N. Feed rate calculation
algorithm for the homogeneous material deposition of blisk blades by 5-axis laser cladding. Int. J. Adv.
Manuf. Technol. 2014, 74, 1219–1228. [CrossRef]
4. Odina, L.; Hardjanto, F.; Walker, A. Effects of impact loads on CRA-Lined pipelines. Ocean Eng. 2018, 166,
117–134. [CrossRef]
5. Blondeau, R. Metallurgy and Mechanics of Welding; Wiley-ISTE: Hoboken, NJ, USA, 2008.
6. Weman, K. MIG Welding Guide; CRC Press: Boca Raton, FL, USA, 2006.
7. Khan, I. Welding Science and Technology; New Age International Pvt Ltd Publishers: New Delhi, India, 2009.
8. Woodward, N.J.; Yapp, D.; Blackman, S.A.; Richardson, I.M.; Widgery, D.; Armstrong, M.A.P.; Verley, R.L.P.;
Berge, J.O. Diverless underwater GMA welding for pipeline repair using a fillet welded sleeve. In International
Pipeline Conference; American Society of Mechanical Engineers: Calgary, AB, Canada, 2004; pp. 1475–1484.
9. Berge, J.O.; Armstrong, M.; Verley, R. Deep Water Remote Pipeline Repair Using Welded Sleeve Technique.
In Proceedings of the International Pipeline Conference, Calgary, AB, Canada, 4–8 October 2004.
10. Jemblie, L.; Olden, V.; Akselsen, O.M. A coupled diffusion and cohesive zone modelling approach for
numerically assessing hydrogen embrittlement of steel structures. Int. J. Hydrog. Energy 2017, 42, 11980–11995.
[CrossRef]
11. Jemblie, L.; Olden, V.; Akselsen, O.M. A review of cohesive zone modelling as an approach for numerically
assessing hydrogen embrittlement of steel structures. Philos. Trans. R. Soc. A Math. Phys. Eng. Sci. 2017, 375,
20160411. [CrossRef] [PubMed]
12. Jemblie, L.; Olden, V.; Mainçon, P.; Akselsen, O.M. Cohesive zone modelling of hydrogen induced cracking
on the interface of clad steel pipes. Int. J. Hydrog. Energy 2017, 42, 28622–28634. [CrossRef]
13. Woodward, N.J.; Fostervoll, H.; Akselsen, O.M.; Ahlen, C.H.; Berge, J.O.; Armstrong, M. Inconel 625
Performance as Hyperbaric GMA Welding Consumable for Diverless Retrofit Tee Hot Tap Applications.
In Proceedings of the International Society of Offshore and Polar Engineers (ISOPE), Lisbon, Portugal,
1–6 July 2007; pp. 3444–3451.
14. Urbikain, G.; Perez, J.M.; López de Lacalle, L.N.; Andueza, A. Combination of friction drilling and form
tapping processes on dissimilar materials for making nutless joints. Proc. Inst. Mech. Eng. B J. Eng. Manuf.
2018, 232, 1007–1020. [CrossRef]
15. Dhib, Z.; Guermazi, N.; Ktari, A.; Gasperini, M.; Haddar, N. Mechanical bonding properties and interfacial
morphologies of austenitic stainless steel clad plates. Mater. Sci. Eng. A 2017, 696, 374–386. [CrossRef]
16. Dhib, Z.; Guermazi, N.; Gaspérini, M.; Haddar, N. Cladding of low-carbon steel to austenitic stainless steel
by hot-roll bonding: Microstructure and mechanical properties before and after welding. Mater. Sci. Eng. A
2016, 656, 130–141. [CrossRef]
17. Jemblie, L.; Bjaaland, H.; Nyhus, B.; Olden, V.; Akselsen, O.M. Fracture toughness and hydrogen embrittlement
susceptibility on the interface of clad steel pipes with and without a Ni-interlayer. Mater. Sci. Eng. A 2017,
685, 87–94. [CrossRef]
18. Bjaaland, H. Evaluation of Welded Clad Pipe—Microstructures and Properties. Master’s Thesis, Norwegian
University of Science and Technology (NTNU), Trondheim, Norway, 2015.
19. Bjaaland, H. Metallurgical Reactions in Welding of Clad X60/X65 Pipelines; Norwegian University of Science and
Technology (NTNU): Gjøvik, Norway, 2014.
20. Koshy, P. Alloy 625 Weld Cladding of Wellheads and Valves: Review of Dilution-Control Techniques and
Weld Process Development. In Proceedings of the Offshore Technology Conference, Houston, TX, USA,
6–9 May 1985; p. 10.
21. Bjaaland, H.; Akselsen, O.M.; Olden, V.; Nyhus, B.; Karlsen, M.; Hjelen, J. Metallurgical reactions in welding
of clad X60/X65 pipelines. In Proceedings of the Twenty-fifth International Ocean and Polar Engineering
Conference—In International Offshore and Polar Engineering Conference, Kona, HI, USA, 21–26 June 2015;
pp. 61–66.
Appl. Sci. 2019, 9, 3118 21 of 24

22. Specification, A.P.I. 5LD, Specification for CRA Clad or Lined Steel Pipe, 3rd ed.; American Petroleum Institute:
Washington, DC, USA, 2009.
23. Veritas, D.N. Offshore Standard DNV-OS-F101, Submarine Pipeline Systems; Det, DNV; Norske Veritas: Høvik,
Norway, 2012.
24. Wu, C.; Li, S.; Zhang, C.; Wang, X. Microstructural evolution in 316LN austenitic stainless steel during
solidification process under different cooling rates. J. Mater. Sci. 2016, 51, 2529–2539. [CrossRef]
25. Suutala, N. Effect of solidification conditions on the solidification mode in austenitic stainless steels.
Metall. Trans. A 1983, 14, 191–197. [CrossRef]
26. Olson, D.L. Prediction of austenitic weld metal microstructure and properties. Weld. J. 1985, 64, 281–295.
27. Akselsen, O.M.; Rørvik, G.; van der Eijk, C.; Kvaale, P.E. Mechanical properties of experimental 13% Cr
stainless steel weld deposits. In Proceedings of the Nordic welding conference, Reykjavik, Iceland, 20–22
September 2000.
28. Hammar, O.; Svensson, U. Solidification and Casting of Metals: Proceedings of an International Conference on
Solidification; The Metals Societ: London, UK, 1979.
29. Cieslak, M.J.; Headley, T.J.; Romig, A.D.; Kollie, T. A melting and solidification study of alloy 625. Metall. Trans.
A 1988, 19, 2319–2331. [CrossRef]
30. Knorovsky, G.A.; Cieslak, M.J.; Headley, T.J.; Romig, A.D.; Hammetter, W.F. INCONEL 718: A solidification
diagram. Metall. Trans. A 1989, 20, 2149–2158. [CrossRef]
31. DuPont, J.N.; Marder, A.R.; Notis, M.R.; Robino, C.V. Solidification of Nb-bearing superalloys: Part II.
Pseudoternary solidification surfaces. Metall. Mater. Trans. A 1998, 29, 2797–2806. [CrossRef]
32. DuPont, J.N.; Robino, C.V. The influence of Nb and C on the solidification microstructures of Fe-Ni-Cr alloys.
Scr. Mater. 1999, 41, 449–454. [CrossRef]
33. DuPont, J.N.; Banovic, S.W.; Marder, A.R. Microstructural Evolution and Weldability of Dissimilar Welds
between a Super Austenitic Stainless Steel and Nickel Base Alloys. Weld. J. 2003, 82, 125–135.
34. DuPont, J.N. Fundamentals of Weld Solidification. In ASM Handbook—Welding Fundamentals and Processes;
Lienert, T., Siewert, T., Babu, S., Acoff, V., Eds.; ASM International: Almere, The Netherlands, 2011; Volume 6,
pp. 96–114.
35. David, S.A.; Babu, S.S.; Vitek, J.M. Welding: Solidification and microstructure. JOM 2003, 55, 14–20.
[CrossRef]
36. Lienert, T.; Siewert, T.; Babu, S.; Acoff, V. ASM Handbook; Welding Fundamentals and Processes (ASM):
Almere, The Netherlands, 2011; Volume 6.
37. Blecher, J.J.; Palmer, T.A.; DebRoy, T. Solidification Map of a Nickel-Base Alloy. Metall. Mater. Trans. A 2014,
45, 2142–2151. [CrossRef]
38. Aarbogh, H.M.; Hamide, M.; Fjær, H.G.; Mo, A.; Bellet, M. Experimental validation of finite element codes
for welding deformations. J. Mater. Process. Technol. 2010, 210, 1681–1689. [CrossRef]
39. Yang, Z.; Debroy, T. Modeling macro-and microstructures of Gas-Metal-Arc Welded HSLA-100 steel.
Metall. Mater. Trans. B 1999, 30, 483–493. [CrossRef]
40. Bunaziv, I.; Akselsen, O.M.; Frostevarg, J.; Kaplan, A.F.H. Laser-arc hybrid welding of thick HSLA steel.
J. Mater. Process. Technol. 2018, 259, 75–87. [CrossRef]
41. Azar, A.S.; Ås, S.K.; Akselsen, O.M. Analytical Modeling of Weld Bead Shape in Dry Hyperbaric GMAW
Using Ar-He Chamber Gas Mixtures. J. Mater. Eng. Perform. 2013, 22, 673–680. [CrossRef]
42. Heigel, J.C.; Gouge, M.F.; Michaleris, P.; Palmer, T.A. Selection of powder or wire feedstock material for the
laser cladding of Inconel® 625. J. Mater. Process. Technol. 2016, 231, 357–365. [CrossRef]
43. Tan, W.; Shin, Y.C. Multi-scale modeling of solidification and microstructure development in laser keyhole
welding process for austenitic stainless steel. Comput. Mater. Sci. 2015, 98, 446–458. [CrossRef]
44. Zhang, L.J.; Zhang, J.X.; Gumenyuk, A.; Rethmeier, M.; Na, S.J. Numerical simulation of full penetration
laser welding of thick steel plate with high power high brightness laser. J. Mater. Process. Technol. 2014, 214,
1710–1720. [CrossRef]
45. Easterling, K. Introduction to the Physical Metallurgy of Welding; Butterworth-Heinemann: Oxford, UK, 1983.
46. Wei, H.L.; Elmer, J.W.; DebRoy, T. Crystal growth during keyhole mode laser welding. Acta Mater. 2017, 133,
10–20. [CrossRef]
47. Kurz, W.; Fisher, D.J. Dendrite growth at the limit of stability: Tip radius and spacing. Acta Metall. 1981, 29,
11–20. [CrossRef]
Appl. Sci. 2019, 9, 3118 22 of 24

48. Franke, M.M.; Singer, R.F.; Steinbach, I. Tertiary dendritic instability in late stage solidification of Ni-based
superalloys. Model. Simul. Mater. Sci. Eng. 2014, 22, 025026. [CrossRef]
49. Kurz, W.; Fisher, D.J. Fundamentals of Solidification, 3rd ed.; Trans Tech Publ: Aedermannsdorf, Switzerland,
1989; p. 305.
50. Porter, D.A.; Easterling, K.E. Phase Transformations in Metals and Alloys, 2nd ed.; Chapman & Hall (CRC Press):
Boca Raton, FL, USA, 1992.
51. Silva, C.C.; Miranda, H.C.D.; Motta, M.F.; Farias, J.P.; Afonso, C.R.M.; Ramirez, A.J. New insight on the
solidification path of an alloy 625 weld overlay. J. Mater. Res. Technol. 2013, 2, 228–237. [CrossRef]
52. Epishin, A.; Link, T.; Brückner, U.; Fedelich, B.; Portella, P. Effects of segregation in nickel-base superalloys:
Dendritic stresses. Superalloys 2004, 2004, 537–543.
53. Andrews, K. Empirical formulae for the calculation of some transformation temperatures. J. Iron Steel Inst.
1965, 203, 721–727.
54. Shushan, S.M.; Charles, E.A.; Congleton, J. The environment assisted cracking of diffusion bonded stainless
to carbon steel joints in an aqueous chloride solution. Corros. Sci. 1996, 38, 673–686. [CrossRef]
55. McGuire, M. Austenitic Stainless Steels. In Stainless Steels for Design Engineers; ASM International: Almere,
The Netherlands, 2008.
56. Guiraldenq, P.; Duparc, O.H. The genesis of the Schaeffler diagram in the history of stainless steel. Metall. Res.
Technol. 2017, 114, 613. [CrossRef]
57. Lippold, J.C.; Kotecki, D.J. Welding Metallurgy and Weldability of Stainless Steels; Wiley-VCH: Weinheim,
Germany, 2005.
58. Kotecki, D.J.; Siewert, T.A. WRC-1992 Constitution Diagram for Stainless Steel Weld Metals: A Modification
of the WRC-1988 Diagram. Weld. J. 1992, 71, 171–178.
59. Feng, J.; Zhang, H.; He, P. The CMT short-circuiting metal transfer process and its use in thin aluminium
sheets welding. Mater. Des. 2009, 30, 1850–1852. [CrossRef]
60. Azar, A.S.; Akselsen, O.M.; Fostervoll, H. Prediction of the Thermal Cycles in Dry Hyperbaric GMA Welding
Using Partial Differential Heat Transfer Equations. In Proceedings of the 9th International Conference,
Trends in Welding Research, Chicago, IL, USA, 4–8 June 2012; pp. 738–745.
61. Shakil, M.; Ahmad, M.; Tariq, N.H.; Hasan, B.A.; Akhter, J.I.; Ahmed, E.; Mehmood, M.; Choudhry, M.A.;
Iqbal, M. Microstructure and hardness studies of electron beam welded Inconel 625 and stainless steel 304L.
Vacuum 2014, 110, 121–126. [CrossRef]
62. Shah Hosseini, H.; Shamanian, M.; Kermanpur, A. Microstructural and weldability analysis of Inconel617/AISI
310 stainless steel dissimilar welds. Int. J. Press. Vessel. Pip. 2016, 144, 18–24. [CrossRef]
63. Li, H.L.; Liu, D.; Yan, Y.T.; Guo, N.; Feng, J.C. Microstructural characteristics and mechanical properties
of underwater wet flux-cored wire welded 316L stainless steel joints. J. Mater. Process. Technol. 2016, 238,
423–430. [CrossRef]
64. Suutala, N.; Takalo, T.; Moisio, T. Ferritic-austenitic solidification mode in austenitic stainless steel welds.
Metall. Trans. A 1980, 11, 717–725. [CrossRef]
65. DuPont, J.N.; Lippold, J.C.; Kiser, S.D. Welding Metallurgy and Weldability of Nickel-Base Alloys; John Wiley &
Sons, Inc.: Hoboken, NJ, USA, 2009.
66. Ren, W.; Lu, F.; Yang, R.; Liu, X.; Li, Z. Liquation cracking in fiber laser welded joints of inconel 617. J. Mater.
Process. Technol. 2015, 226, 214–220. [CrossRef]
67. Hua, C.; Lu, H.; Yu, C.; Chen, J.-M.; Zhang, M.-L.; Li, D.-Y. Reduction of Laves phase in nickel-alloy welding
process under ultrasonic Ampère’s force. J. Mater. Process. Technol. 2018, 252, 389–397. [CrossRef]
68. Hochhauser, F.; Ernst, W.; Rauch, R.; Vallant, R.; Enzinger, N. Influence of the Soft Zone on The Strength of
Welded Modern Hsla Steels. Weld. World 2012, 56, 77–85. [CrossRef]
69. Onsøien, M.; M’Hamdi, M.; Mo, A. A CCT Diagram for an Offshore Pipeline Steel of X70 Type. Weld. J. 2009,
88, 1–6.
70. Akselsen, O.M.; Solberg, J.K.; Grong, Ø. Effects of martensite-austenite (M-A) islands on intercritical
heat-affected zone toughness of low carbon microalloyed steels. Scand. J. Metall. 1988, 17, 194–200.
71. Akselsen, O.M.; Grong, Ø.; Solberg, J.K. Structure–property relationships in intercritical heat affected zone of
low-carbon microalloyed steels. Mater. Sci. Technol. 1987, 3, 649–655. [CrossRef]
Appl. Sci. 2019, 9, 3118 23 of 24

72. Mohseni, P.; Solberg, J.K.; Østby, E.; Akselsen, O.M. Application of Electron Backscatter Diffraction (EBSD) On
Facet Crystallographic Orientation Studies In Arctic Steels. In Proceedings of the Twenty-first International
Offshore and Polar Engineering Conference, Maui, HI, USA, 16–21 June 2011; p. 5.
73. Mohseni, P.; Solberg, J.K.; Karlsen, M.; Akselsen, O.M.; Østby, E. Application of combined EBSD and
3D-SEM technique on crystallographic facet analysis of steel at low temperature. J. Microsc. 2013, 251, 45–56.
[CrossRef]
74. Mohseni, P.; Solberg, J.K.; Karlsen, M.; Akselsen, O.M.; Østby, E. Cleavage Fracture Initiation at M–A
Constituents in Intercritically Coarse-Grained Heat-Affected Zone of a HSLA Steel. Metall. Mater. Trans. A
2014, 45, 384–394. [CrossRef]
75. Mohseni, P.; Solberg, J.K.; Karlsen, M.; Akselsen, O.M.; Østby, E. Investigation of mechanism of cleavage
fracture initiation in intercritically coarse grained heat affected zone of HSLA steel. Mater. Sci. Technol. 2012,
28, 1261–1268. [CrossRef]
76. Haugen, V.G.; Rogne, B.R.S.; Akselsen, O.M.; Thaulow, C.; Østby, E. Local mechanical properties of
intercritically reheated coarse grained heat affected zone in low alloy steel. Mater. Des. 2014, 59, 135–140.
[CrossRef]
77. Davis, C.L.; King, J.E. Cleavage initiation in the intercritically reheated coarse-grained heat-affected zone:
Part I. Fractographic evidence. Metall. Mater. Trans. A 1994, 25, 563–573. [CrossRef]
78. Lo, K.H.; Shek, C.H.; Lai, J.K.L. Recent developments in stainless steels. Mater. Sci. Eng. R Rep. 2009, 65,
39–104. [CrossRef]
79. Bhadeshia, H.K.D.H.; Honeycombe, R.W.K. Steels: Microstructure and Properties, 3rd ed.; Butterworth-Heinemann:
Oxford, UK, 2006.
80. Kou, S. Welding Metallurgy, 2nd ed.; Wiley: Hoboken, NJ, USA, 2003; p. 480.
81. Lippold, J.C. Welding Metallurgy and Weldability; Wiley: Hoboken, NJ, USA, 2015.
82. Baeslack III, W.A.; Lippold, J.C.; Savage, W.F. Unmixed zone formation in austenitic stainless steel weldments.
Weld. J. 1979, 58, 168.
83. Feng, Y.; Luo, Z.; Liu, Z.; Li, Y.; Luo, Y.; Huang, Y. Keyhole gas tungsten arc welding of AISI 316L stainless
steel. Mater. Des. 2015, 85, 24–31. [CrossRef]
84. Sabzi, M.; Dezfuli, S.M. Drastic improvement in mechanical properties and weldability of 316L stainless
steel weld joints by using electromagnetic vibration during GTAW process. J. Manuf. Process. 2018, 33, 74–85.
[CrossRef]
85. Kikuchi, H.; Takahashi, H.; Yanagiwara, H.; Murakami, T. Relationship between ferromagnetic properties
and grain size of Inconel alloy 600. J. Magn. Magn. Mater. 2015, 381, 56–64. [CrossRef]
86. Duvall, D.S.; Owczarski, W.A. Further heat affected zone studies in heat-resistant nickel alloys. Weld. J. 1967,
46, 423–432.
87. Lippold, J.C.; Beaslack III, W.A.; Varol, I. Heat-Affected Zone Liquation Cracking in Austenitic and Duplex
Stainless Steels. Weld. J. 1992, 71, 1–14.
88. Li, S.; Li, K.; Cai, Z.; Pan, J. Behavior of M23C6 phase in Inconel 617B superalloy during welding. J. Mater.
Process. Technol. 2018, 258, 38–46. [CrossRef]
89. Korrapati, P.K.; Avasarala, V.K.; Bhushan, M.; Ramkumar, K.D.; Arivazhagan, N.N.; Narayanan, S. Assessment
of Mechanical Properties of PCGTA Weldments of Inconel 625. Procedia Eng. 2014, 75, 9–13. [CrossRef]
90. Vigraman, T.; Ravindran, D.; Narayanasamy, R. Diffusion bonding of AISI 304L steel to low-carbon steel
with AISI 304L steel interlayer. Mater. Des. 2012, 34, 594–602. [CrossRef]
91. Wang, S.; Liu, B.X.; Chen, C.X.; Feng, J.H.; Yin, F.X. Microstructure, mechanical properties and interface
bonding mechanism of hot-rolled stainless steel clad plates at different rolling reduction ratios. J. Alloy
Compd. 2018, 766, 517–526. [CrossRef]
92. Lindholm, D. Carbon Diffusion across the Bimetallic Interface of Welded Clad Pipes. In Proceedings of the
28th International Journal of Offshore and Polar Engineering, Sapporo, Japan, 10–15 June 2018; pp. 165–172.
93. Olden, V.; Jemblie, L.; Akselsen, O.M.; Utvær, A. Hydrogen embrittlement susceptibility on the interface
of clad steel pipes, before and after welding. In Proceedings of the International Conference on Metals &
Hydrogen, Ghent, Belgium, 28–30 May 2018.
94. Wang, Y.; Tsai, H.L. Impingement of filler droplets and weld pool dynamics during gas metal arc welding
process. Int. J. Heat Mass Transf. 2001, 44, 2067–2080. [CrossRef]
Appl. Sci. 2019, 9, 3118 24 of 24

95. Hu, J.; Tsai, H.L. Effects of current on droplet generation and arc plasma in gas metal arc welding. J. Appl.
Phys. 2006, 100, 053304. [CrossRef]
96. Hu, J.; Tsai, H.L. Heat and mass transfer in gas metal arc welding. Part I: The arc. Int. J. Heat Mass Transf.
2007, 50, 833–846. [CrossRef]
97. Rao, Z.H.; Hu, J.; Liao, S.M.; Tsai, H.L. Modeling of the transport phenomena in GMAW using argon–helium
mixtures. Part I—The arc. Int. J. Heat Mass Transf. 2010, 53, 5707–5721. [CrossRef]
98. Rao, Z.H.; Hu, J.; Liao, S.M.; Tsai, H.L. Modeling of the transport phenomena in GMAW using argon–helium
mixtures. Part II—The metal. Int. J. Heat Mass Transf. 2010, 53, 5722–5732. [CrossRef]
99. Rao, Z.H.; Zhou, J.; Tsai, H.L. Determination of equilibrium wire-feed-speeds for stable gas metal arc welding.
Int. J. Heat Mass Transf. 2012, 55, 6651–6664. [CrossRef]
100. Zhao, Y.; Chung, H. Numerical simulation of droplet transfer behavior in variable polarity gas metal arc
welding. Int. J. Heat Mass Transf. 2017, 111, 1129–1141. [CrossRef]
101. Cao, Z.; Yang, Z.; Chen, X.L. Three-Dimensional Simulation of Transient GMA Weld Pool with Free Surface.
Weld. J. 2004, 83, 169–176.
102. Chen, J.; Schwenk, C.; Wu, C.S.; Rethmeier, M. Predicting the influence of groove angle on heat transfer and
fluid flow for new gas metal arc welding processes. Int. J. Heat Mass Transf. 2012, 55, 102–111. [CrossRef]
103. Hertel, M.; Spille-Kohoff, A.; Füssel, U.; Schnick, M. Numerical simulation of droplet detachment in pulsed
gas–metal arc welding including the influence of metal vapour. J. Phys. D Appl. Phys. 2013, 46, 224003.
[CrossRef]
104. Cheon, J.; Kiran, D.V.; Na, S.-J. CFD based visualization of the finger shaped evolution in the gas metal arc
welding process. Int. J. Heat Mass Transf. 2016, 97, 1–14. [CrossRef]
105. Ogino, Y.; Hirata, Y.; Murphy, A.B. Numerical simulation of GMAW process using Ar and an Ar–CO2 gas
mixture. Weld. World 2016, 60, 345–353. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).