Process development for wire-based laser metal deposition of 5087 aluminium alloy by using fibre laser

Journal of Manufacturing Processes 34 (2018) 721–732

Contents lists available at ScienceDirect

Journal of Manufacturing Processes journal homepage: www.elsevier.com/locate/manpro

Process development for wire-based laser metal deposition of 5087 aluminium alloy by using fibre laser

T

⁎

M. Froenda,b, , S. Riekehrb, N. Kashaevb, B. Klusemanna,b, J. Enzb a b

Leuphana University of Lüneburg, Institute of Product and Process Innovation, Volgershall 1, D-21339 Lüneburg, Germany Helmholtz-Zentrum Geesthacht, Institute of Materials Research, Materials Mechanics, Joining and Assessment, Max-Planck-Straße 1, 21052 Geesthacht, Germany

A R T I C LE I N FO

A B S T R A C T

Keywords: Laser additive manufacturing Aluminium alloy Laser metal deposition Pre-heating Microstructure Tailored geometrical shape

In recent decades, laser metal deposition, as a part of additive manufacturing, developed into a promising methodology in industrial fields. In recent years, there has been an increased interest in the processability of lightweight high-strength structural materials, such as aluminium alloys. However, in terms of wire-based laser metal deposition, there is still a lack of knowledge with regard to the processability of aluminium alloys. In this research, the process development for wire-based laser metal deposition of a 5087 aluminium alloy (AlMg4.5 MnZr) has been conducted. It is observed that pre-heating is beneficial in terms of porosity and distortion reduction. Within optimized parameter ranges, it is possible to control the geometric shape, dilution, and aspect ratios of the deposited layers in a systematic way. Accordingly, defect-free layers with tailored geometrical features can be processed and adapted to specific process requirements.

1. Introduction From its development in the late 1980s, additive manufacturing (AM) has become a profitable method to generate components and structures for industrial use today. The concept of this technique consists of melting a material, usually provided as powder or wire, and adding it layer-by-layer to build components, instead of subtractive manufacturing, such as cutting or milling [1–4]. Therefore, AM enables a high degree of freedom from the construction point of view, and includes great advantages in terms of the degree of utilization, including an inherent simplicity in building three-dimensional shaped parts [5]. In comparison to conventional machining methods, AM leads to a significant reduction of material wastage from up to 90 percent for machining down to less than 10 percent for AM [6,7]. In addition, saving the tool change time per part, the costs and cycle time can be minimized [8]. From an economic point of view, the fabrication of wire material in contrast to powder is comparably cheap. Therefore, processing wire, instead of powder raw material, is very interesting for industrial applications [9]. In the field of AM, many different approaches have been developed over recent decades. With regard to the production of metallic parts, these approaches are divided into powder-and wire-based techniques [8,10,11]. Powder-based approaches have already achieved significant application in several industrial fields [8,12]. These approaches are

subdivided in powder bed and powder injection techniques [11,13]. In case of powder bed systems, such as selective laser melting (SLM), a certain amount of pulverized metallic material has to be supplied in a processing chamber that is flooded by shielding gas. This powder material is locally irradiated by a laser beam, which follows a predefined path according to a computational design. After the solidification of the molten material, a subsequent layer of powder material is automatically placed on top of the manufactured structure. This method is repeated until the desired shape of the component is manufactured. In contrast, in powder injection approaches, the pulverized material is supplied through a nozzle. After the material leaves the nozzle, it is melted by a laser beam and added to the workpiece in order to produce the desired shape of the final component. An inert gas, such as argon or helium, is used to blow the pulverized material on the workpiece, which simultaneously protects the melt pool from reacting with atmospheric gases. In wire-based approaches, the wire is fed through a nozzle on to the workpiece to add certain layers of material on top of or next to each other. Similar to powder injection approaches, the material is shielded by an inert gas during the molten state. However, the surface area of the injected wire by using wire-based AM is much lower than the accumulated surface area of the powder particles in powder-based approaches, which reduces the hazard of possible reactions with atmospheric gases. Therefore, it is not necessary to work in a closed chamber

⁎ Corresponding author at: Helmholtz-Zentrum Geesthacht, Institute of Materials Research, Materials Mechanics, Joining and Assessment, Max-Planck-Straße 1, 21052 Geesthacht, Germany. E-mail address: [email protected] (M. Froend).

https://doi.org/10.1016/j.jmapro.2018.06.033 Received 3 July 2017; Received in revised form 17 April 2018; Accepted 25 June 2018 1526-6125/ © 2018 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

of an inert gas, which drastically increases the degree of freedom for the applications of the wire-based approach. In addition, AM, by using wire-feeding systems, offers a higher usage efficiency of a material [14], improved surface quality of the deposited structures [3,9], and enables higher deposition rates [3] in comparison to powder-feeding systems. For AM, the minimization of post-processing is one of the most important aspects [15]. In this regard, milling to straighten surface roughness is primarily conducted. This results in a certain amount of waste material and additional manufacturing costs. In order to reduce the amount of waste material and production costs, near net shaped components are targeted. For this purpose, process windows have to be precisely configured in such a way that high efficiency can be achieved. In this contribution, a laser metal deposition (LMD) technique, using filler wire as raw material and laser as energy input, is investigated. This approach offers high deposition rates in a well-controllable process [15,16]. Local shielding in the melting zone is applied. Owing to high adaptability in industrial fields, a process development addressing the identification of important parameters in the LMD of the aluminium alloy 5087, with particular attention paid to the resulting macro-morphology, is investigated. Therefore, a parametric study including the examination of the possible defects and surface quality of the specimens is conducted.

Fig. 1. Schematic visualization of the laser metal deposition (LMD) process. The fixed process parameters where illustrated; these include the flow rate of shielding gas Q Ar , the shielding gas nozzle angle γ , the wire feeding nozzle angle β , the focal spot to wire-feeding nozzle tip distance dN , and the direction of deposition.

the substrate. Owing to the fact that the wire is supplied on a coil, it has an inherent cast that might affect the feeding accuracy by an oscillating movement of the wire tip after leaving the nozzle. The laser spot diameter was enhanced to 1.6 mm by positive defocusing to +23 mm. By this, the occurring oscillations of the wire were compensated and the substrate could also be partly molten during the deposition, thereby leading to an additional stabilization of the process, resulting in a minimization of bonding defects between the substrate and the deposited material. The distance between the focal spot and the edge of the wire-feeding nozzle was adjusted to be as minimal as possible in order to reduce the oscillating movements of the wire and to improve process accuracy. A distance of dN = 4 mm could be identified as the minimum distance to ensure that the tip of the wire nozzle was not molten during the process but also large enough to avoid remaining not molten material on the nozzle after switching off the laser. Otherwise, this remaining material might be melted again at the deposition of the next bead and could fall on the substrate, leading to defects. The same effect might occur during the process-related development of smoulder. Within the process, the table of the CNC-system was moved in relation to the z-axis in the x or y direction at a defined deposition velocity vt . Simultaneously, the wire was fed through the nozzle at a velocity vw . Using the presented process, it was possible to deposit lines of AlMg4.5 MnZr on the AlMg3 substrate and to regulate the deposition rate by adjusting the ratio between vw and vt , named k = vw / vt as used, for example, in [18]. The main characteristics of the used laser source and the features of the focused beam are shown in Fig. 2 and summarized in Table 1. In Fig. 2(a) and (b), the caustic of the focused beam and its symmetry for the used equipment has been visualized. Furthermore, the beam intensity for the focal position (c) as well as for the Rayleigh length (d) is plotted in Fig. 2. It can be seen that a top hat distribution is achieved by using the focal position, whereas a Gaussian distribution shape is formed in the Rayleigh length. Furthermore, a possible elliptical shaping of the laser spot area due to angular relations with respect to the substrate surface was balanced by the adjustment of the equipment and a linear deposition path in perpendicular orientation to the substrate surface. For complex structures requiring curved or circularly paths, the angular relations between the optical head and the surface of the structure have to be taken into account. In this study, only straightlined beads have been considered.

2. Experimental study 2.1. Materials In this study, the aluminium alloy AlMg4.5 MnZr (EN AW-5087) as wire-material and AlMg3 (EN AW-5754) as substrate-material were investigated. The 5xxx aluminium series is characterized to be the stiffest with respect to non-heat treatable aluminium alloys [17]. The wire was provided with a diameter of 1.0 mm by using a speed-controlled automatic feeding device. The substrate material was provided in 200 mm × 150 mm rolled sheets, with a thickness of 6 mm. The deposition direction was adjusted perpendicular to the rolling direction of the substrate material. Owing to the high reflectivity of aluminium and, furthermore, in order to clean the surface of the substrate material, the substrate was firstly sandblasted and cleaned with acetone before LMD. A pressure of 8 bar and a particle size between 90 and 150 μm of the blasting material was used to achieve a surface roughness of Ra = 0.20 μm of the substrate material. 2.2. Experimental setup An 8-kW continuous wave ytterbium fibre laser YLS-8000-S2-Y12 (IPG Photonics Corporation) integrated with the optical head YW52 Precitec, in a CNC-supported XYZ-machining centre (IXION Corporation), was employed in this study. The optical head was integrated along the Z-axis of the system, which was also equipped with a wire-feeding system along with a local shielding gas supply, as schematically illustrated in Fig. 1. The wire was fed through a nozzle at a fixed angle of β = 35°, relative to the surface of the substrate. The nozzle for a local shielding gas supply was installed above the wire nozzle with an angle of γ = 55°, a vertical distance of 25 mm, and horizontal spacing of 5 mm relative to the tip of the wire-feeding nozzle. The shielding gas flow rate was adjusted to 10 l/min, thus providing enough gas to protect the molten material from oxidation with atmospheric gases but not influencing the solidification process. Owing to the high viscosity of aluminium in its liquid state and in combination with a very high cooling rate within the LMD process, the shape of the deposited structure could be distorted by using a very strong gas flow. As an alternative, the deposition could be conducted in a complete shielding gas atmosphere, such as an argonfilled chamber. However, this approach is less adaptable within industrial applications due to geometric restrictions. The circular shaped laser beam irradiates the wire perpendicular to

2.3. Experimental procedure Table 2 provides an overview of the varied process parameters 722

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

Fig. 2. Main characteristics and features of the 8-kW continuous wave ytterbium fibre laser YLS-8000-S2-Y12 (IPG Photonics Corporation) used with the optical head YW52 Precitec. Overview of the caustic (a), beam symmetry (b), and beam intensity distribution in focal (c) and Rayleigh length position (d).

deposited beads that show a good surface quality, no lacks of fusion, and constant geometrical shapes are further characterized. During the LMD process, the laser heat source melts the wire in order to add the material to the substrate or the underlying structure, leading to a high bonding. Owing to the fact that high laser powers have to be used to melt the material, high temperature gradients occur, resulting in residual stresses in the substrate or the previously deposited structure. In case that these stresses exceed the yield strength of the material, plastic deformation of the substrate as well as in the deposited structure develop [19,20]. This mechanism is known as the temperature gradient mechanism and can lead to componential distortions of the processed structure. Fig. 3 depicts the possible reasons for residual stresses within the deposited structures. It is assumed that due to shrinkage effects of the deposited beads, internal compressive stresses occur, whereas high tensile stress intensities especially in the beginning and ending regions of the deposited layer result [21]. During the process, a bead of the length lhot is deposited (shown in Fig. 3(a)) and contracts during solidification to the

Table 1 Characteristics of the laser equipment. Parameter

Symbol

Value

Unit

Maximum Power Focal Length Process Fibre Diameter Focal Spot Diameter

Pmax F D df

8000 300 300 750

W mm μm μm

Centre Wavelength Beam Parameter Product Rayleigh Length

λ BPP zr

1070 9.9 12.3

nm mm ∙ mrad mm

within this study. At first, the parameter variations of laser power P , deposition and wire velocity, vt and vw , are conducted. For this purpose, common characteristic factors in wire-based LMD, such as the aforementioned k-factor as well as the specific process energy W , expressing the energy to the deposition rate ratio, were applied. Resulting Table 2 Summary of the varied process parameters investigated during LMD experiments. Parameter

Symbol

Parameter Variation

Parameter Variation

Unit

Substrate Temperature Laser Power Deposition Velocity Wire Velocity

Ts P vt vw

25 3500, 4000, 4500 1 1, 2, 3, 4, 5, 6, 7, 8, 9, 10

150 3500, 4000, 4500 1, 2, 3, 4, 5 1, 2, 3, 4, 5, 6, 7, 8, 9, 10

°C W m/min m/min

723

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

Fig. 3. Schematic visualization of possible defects induced during LMD due to thermal effects. A deposited bead with length lhot is deposited (a). The shrinkage due to temperature decrease leads to residual stresses and the deformations of the substrate for single-bead deposition (b). Multi-layer LMD with bead lengths of lhot due to subsequent deposition (c) and cooled down state, which leads to residual stresses in both layers and the substrate in the y–z plane (d) [23,22].

indicated in Fig. 4(b) and predominantly determined by an adjustment of the bead spacing dc , has to be considered. To examine the resulting surface roughness for the estimated values of dc, four different bead distances, ranging from 2.15 mm to 4.1 mm, were used.

length lcool (shown in Fig. 3(b)). As a result of strong bonding between the substrate material and the deposited bead, compressive stresses in the substrate occur. These stresses may induce distortions of the structure in y–z plane, as shown in Fig. 3(b). Fig. 3(c) visualizes the deposition of a second bead on top, where the same effect occurs. This leads to compressive stresses in the first bead and the substrate, which additionally increases the development of the distortion angle φ1 (as shown in Fig. 5(d)). Owing to the so-called temperature gradient effect, additional distortions occur in the x–z plane occur [22,23]). In addition, these residual stresses may introduce cracking inside the deposited materials or between the layer and the substrate, which is also known as the delamination effect [20,24]. As already stated in [25], the pre-heating of the substrate reduces porosity and cracking owing to the reduction of the thermal gradient and the increase of time, in which the melt pool is able to gas out. In addition, in [3], [43], and [44] it was stated that constant process conditions with regard to material supply and temperature distributions are needed in order to reduce porosity and the occurrence of thermal gradients. Therefore, the substrate was pre-heated in order to provide a more homogenous heat profile within the substrate and the deposited structure during the process. For this purpose, the substrate was clamped on an IKA C-MAG HS7 heating plate with a maximum power of 1 kW (maximum temperature = 500 °C). However, also a feasibility study of wire-based LMD by using a room-tempered substrate was conducted in order to investigate the possibility of using wire-based LMD for repairing purposes as well. Through wire-based LMD, parameter-dependent bead shapes with height H and width ω are produced, as schematically shown in Fig. 4 (a). As a result of this dynamic process, the wire partly dilutes inside the substrate, which is expressed by the dilution depth h . Hence, the created bead is welded into the substrate, which results in a strong bonding between the build-up structure and the substrate or between multiple layers [15,26–29]. In relation to the chosen process parameters, the bead width, dilution depth, and bead height vary [30]. As an indicator for the ratio between the width and the height of the bead, the bead side-angle α and the width-to-height aspect ratio RB are defined according to [31]. Fig. 4(b) and (c) present two different possibilities of LMD, with multiple layers next to each other or on top of each other. In case of a surface-coating application, the overlapping ratio μc , which is

3. Characterization methods 3.1. Analysis of the bead geometry and surface defects To characterize the shape of the deposited beads, several assessment criteria were defined. In terms of constant processability, a homogenous bead shape is needed; this is defined by low surface roughness and constant width along the bead. Fig. 5(a) depicts homogenous bead shapes with constant height and width along the bead. In contrast, Fig. 5(b) indicates possible shape instabilities, such as globally varied bead widths ω1 and ω2 , locally varying widths Δω along the bead, different bead heights between the start and end points H1 and H2 , and increased surface roughness ΔH owing to a varying height along the bead. In addition, occuring lacks of fusion have to be taken into account. With regard to a structure building LMD, a bead spacing dc = 0 and a side-angle of the structure δ = 90° during the multi-layer deposition on top of the previously deposited bead in order to produce z-directional structures are required. Optimal (dc = 0 ) as well as a displaced positioning (dc ≠ 0 ) of a second layer on top of each other is shown in Fig. 6. This also illustrates the effect of an occurring tilting angle of the deposited structure ε . 3.2. Analysis of inner defects and microstructure After a surface morphology inspection of the deposited beads, a volume inspection to investigate its freedom from defects, such as pores and cracks, was conducted. The detection of inner defects was achieved by X-ray analyses. A Seifert Isovolt 320/13 X-ray tube, using an irradiation angle of 42°, a working distance of 800 mm, a tube voltage of 70 kV, and a tube current intensity of 4.2 mA, which affected an effective focal spot of 2.25 mm², was used. The deposited specimens were investigated conforming to EN DIN 35224:2016-02. To investigate the density of the deposited beads only at stationary process conditions, the lead-in and lead-out area, respectively 10 mm 724

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

Fig. 4. Schematic of the geometric parameters for bead morphologies after LMD for single beads (a), multiple beads next to each other (adopted from [30]) (b), leading to a coating structure; alternately, multiple beads on top of each other (c), resulting in a structure building process.

each, were not taken into account for the defect analyses. The lead-in and lead-out regions show an increased amount of porosity due to process instabilities, resulting from changes in the acceleration of the machine and the injection of the laser. The pores within the structure are assumed to be spherical. The pore diameters have been measured and the relative volume of pores within one deposited bead was calculated by using Eq. (3.1). The density of the bead ρbead is expressed as n

ρbead = vw π r 2t −∑ Vpore i,with Vporei = i=1

π d porei 3 6

(3.1)

where dpore, i represents the diameter of each individual pore i . In order to investigate the microstructure and geometric shape of the beads, transverse cross sections from the middle section of the beads were taken. They were mounted, grounded, and polished using an oxide polishing suspension compound (OPS). Microstructural observations were performed by using inverted optical microscopy (OM) Leica DMI 5000 M with polarized light. Owing to its relatively high magnesium content, the material is very corrosion-resistant and, consequently, common aluminium etching agents are not sufficient to visualize the microstructure. Therefore, electrolytic etching, using a 2.5-percent acidiferous tetrafluoroboric acid causticize (35%) was used. This technique, also known as the Barker method, was conducted by using 30 V and an active reaction time of 90 s.

Fig. 6. Optimal positioning of a subsequent layer in the multi-layer LMD in order to achieve a straight z-directional structure by using dc = 0 (a) and a resulting tilting angle ε of the deposited multi-layer structure in the case of dc ≠ 0 (b).

4.1. Adjustment of energy specific parameters Using the deposition velocity vt within the process time t , it is possible to determine a deposited line length l , which is expressed as

l = vt t.

(4.1)

It is clear that the achieved length of the bead l is only dependent on the deposition velocity vt and the process time t but independent from the wire-feeding velocity. In contrast, the deposition rate m˙ depends on the wire-feeding velocity vw as

4. Theoretical considerations Within the LMD process, numerous parameters have to be adjusted in order to define quantifiable input values. For a better evaluation of the process parameter interlink, specific characteristic factors are used. Therefore, the following theoretical equations and assumptions are derived:

m˙ = vw π r 2ρ

(4.2)

where r is the radius of the wire and ρ is the density of the deposited material. The specific energy input is expressed as

Fig. 5. Target shapes considering the constant height and width (a) for single beads. In contrast, beads of varying heights and widths (b) represent undesirable shapes and defects, such as locally poor fusion characteristics. 725

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

WV =

P vw π r 2

In addition [33], described the overlapping ratio μc of the beads as follows:

(4.3)

or

Wg =

μc = P . vw π r 2ρ

(ω− dc ) . ω

(4.4)

In order to generate a defined irradiated area of the circular shaped laser beam, it is possible to adjust the focal diameter by positioning the optical head relative to the surface of the substrate [32]. Hence, the spot diameter d (z ) is described by Eq. (4.5) as a ratio between the Rayleigh length z r and the defocusing distance in the z-direction, which is abbreviated by z relative to the equipment-specific smallest focal spot diameter df taken from [32] as

2

⎡ ⎡ ⎡ ω ⎤2 2 ⎤ ⎤ ⎛ ⎞ ⎢ ⎢⎣ 2 ⎦ + h ⎥ ⎥ 2 hω ⎦ sin−1 ⎜ ⎟− ⎡ ⎡ ω ⎤ − h2 ⎤ ⎥ ⎢⎣ 2 hω ⎢ ⎦⎥ ⎜ ⎡ ω ⎤ + h2 ⎟ ⎣ ⎣ 2 ⎦ ⎢ ⎥ ⎝⎣ 2 ⎦ ⎠ ⎣ ⎦

2

z d (z ) = df 1 + ⎡ ⎤ ⎢ ⎣ zr ⎥ ⎦

(4.8)

As formerly stated, the beads were welded into the substrate, which results in high bonding. Owing to different irradiation times relative to the specific process energy, varying dilution ratios occur. As stated in [34], the dilution is “the variation degree of composition of cladding arising from the blend of molten substrate during the laser cladding” [34]. Assuming a circularly geometric shape of the bead as well as for the molten zone, the dilution in [34] is expressed as

(4.5)

η=

4.2. Geometrical considerations

2 2 2 ω 2 ⎡ ⎡ ω ⎤ + h2 ⎤ ⎡ ω ⎤ 2⎤ ⎡ ⎤ h⎡ ⎛ ⎞ ⎛ ⎞ ⎢ ⎡ ⎤ − Hh ⎥ [H + h] ⎢⎣ 2 ⎦ + H ⎥ ⎢⎣ 2 ⎦ ⎥ ⎣2⎦ hω Hω ⎣ ⎦ sin−1 ⎜ ⎦ sin−1 ⎜ ⎦ ⎟+ ⎣ ⎟− ⎣ 2 2 2 H hω ωH ⎜ ⎡ ω ⎤ + H2 ⎟ ⎜ ⎡ ω ⎤ + h2 ⎟ 2 2 ⎝⎣ ⎦ ⎠ ⎝⎣ ⎦ ⎠

.

(4.9)

As previously mentioned, the LMD process parameters have to be adjusted to achieve optimized bead geometries in terms of the desired application. In order to improve the surface quality in the coating processes, it is assumed in [33] that the geometry of the resulting beads can be considered as circularly shaped, which is visualized in Fig. 7(a). In contrast to the structure-building application, in case of a multi-layer LMD process depositing beads next to each other (e.g., coating application), a certain bead spacing dc > 0 and dc < ω has to be defined in order to achieve low surface roughness.

Taking only the maxima of dilution and the bead height into account, the dilution ratio can be simplified to

ηmax =

h . (h + H )

(4.10)

To express the relation between the bead width and bead height, the so-called aspect ratio RB of the bead is defined as

RB =

ω . H

(4.11)

In addition, the ratio between the deposition height and dilution depth RC is calculated as

RC =

H . h

(4.12)

Within the representation and discussion of the results, these considerations are used to calculate the geometrical features and identify the adaptable process parameters according to the qualitative requirements defined in Section 3. 5. Results and discussion 5.1. Visual inspection

Fig. 7. Geometric assumptions for the bead and molten zone geometries as circularly shaped with defined parameters, according to [33] in LMD used as a coating process.

During this experimental study, changes with regard to the bead shape and processability for different pre-heating temperatures of the substrate were observed. Deposition at room temperature often results in irregular geometries, including varying bead heights and widths, as well as lacks of fusion within one layer. The shape variation with respect to the different substrate temperatures is shown in Fig. 8(a) for a substrate at room temperature, (b) for a pre-heated substrate of 150 °C, and (c) for a substrate temperature of 300 °C. It is observed that the depositions at pre-heated substrates achieve constant bead shapes as well as the minimization of bonding defects, such as a lack of fusion. In addition, it is seen that an increased pre-heating temperature does not deteriorate the surface quality of the deposited layers but leads to a slight enlargement of the bead. The improved bead quality for pre-heated substrate temperatures is basically explained by improved fusion characteristics. By increasing the substrate temperature, less energy is required to heat the material above its melting temperature. However, in case if no pre-heating is applied, parameter adoptions, such as decreased deposition velocity or reduction of the deposition rate, might be used instead to receive better fusion characteristics and defect-free beads.

From Fig. 7 and according to [33], the radius R of the deposited bead is calculated by considering the deposited height H and width of the bead ω . Following [33], the radius can be expressed as

R=

ω2 + 4 H 2 . 8H

(4.6)

It is assumed that the smoothest surface is achieved when the areas A1 and A2 are equal (see Fig. 7). Only in this case, the surface transition between the neighbouring beads is close to being a plane. As shown by [33], for x > 0 , the two areas are determined by d

A1 = R ∙ dc− R2 sin−1 ( 2Rc )− ω A2 = ⎡R2 sin−1 ( 2R )+ ⎢ ⎣

ω 2

2

dc 2

d

2

R2− ⎡ 2c ⎤ ⎣ ⎦

ω R2− ⎡ 2 ⎤ ⎤− ⎡R2 sin−1 ⎣ ⎦ ⎥ ⎣ ⎦ ⎢

( )+ dc 2R

dc 2

2

d R2− ⎡ 2c ⎤ ⎤ ⎣ ⎦ ⎦ ⎥

+ [H −R][ω− dc ]. (4.7)

726

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

Fig. 8. Side and top views of bead shape deviations, resulting from different (pre-heated) substrate temperatures as follows: TS = 25 °C (a), TS = 150 °C (b), TS = 300 °C (c) deposited, with P = 4000 W, vt = 1 m/min, and vw = 8 m/min.

structures. It is assumed that this possible energy reduction is attributed to the development of a large melt pool during wire-based LMD. After this melt pool is developed in the early stages of the process, the wire is fed into it and receives additional heat input. This contributes to the wire melting and reduces the necessary laser energy input. By this, it is also assumed that the laser power can be reduced along the deposition layer to a certain extent. Taking this melt pool effect into account, the reason for the shown porosity trend in Fig. 9 can be explained. Once the melt pool has developed, the process becomes more stable; this is supported by the implementation of the pre-heating of the substrate. As previously known from LBW, porosity and hot cracking behaviour can increase with higher cooling rates [40,41]. In this case, sufficient outgassing of the melt pool is prevented and the remaining gases, such as evaporations or shielding gas, are embedded in the resolidified material. As LMD is very similar to the fusion-welding processes, this phenomenon might also occur during LMD. By this phenomenon, the increasing porosity of the non-pre-heated substrate material can be explained. By using lower deposition velocities and specific energy inputs, the homogeneous heating and melting of the wire as well as the substrate material is obtained. Therefore, the temperature globally increases in such a way that the cooling rates are slow enough to allow the sufficient outgassing of the melt pool. In the case of a very large process energy in relation to the deposition velocity or wire-fed material, the laser mostly penetrates the underlying (substrate) material. Therefore, conduction-welding conditions are exceeded. This

5.2. Radiographic inspection The results from the radiographic inspections also show that it is beneficial to use pre-heated substrate material with respect to achieving defect-free dense beads. Generally, in order to achieve dense beads ( ρbead > 99.99%), the process-window range significantly increases for a pre-heated substrate. In the absence of pre-heating (TS = 25 °C), only the deposition velocity of 1 m/min, in combination with a specific energy input of 10 to12 kJ/g, achieves dense beads with adequate surface qualities (as shown in Fig. 9(a)). By using these parameters, the substrate material uniformly heats up during deposition in such a way that the material in later stages of the bead is assumed to be deposited on an already pre-heated substrate material. Therefore, the focus in the following passage is only on a pre-heated substrate. Fig. 9(b) shows the identified process parameter range for a preheated substrate of TS = 150 °C, which results in dense defect-free beads with good surface quality. The process parameter window, in terms of specific energy, is identified from 10 to 16.5 kJ/g, combined with k-factors between 1 and 9. It is observed that energy inputs above 16.5 kJ/g show decreased surface quality and a specific process energy down to 26 J/mm³ and 10 kJ/g, respectively, is required in order to achieve defect-free beads. Compared to powder-based approaches such as SLM, this is an immense reduction of the necessary energy input. In [35] and [36], the necessary energy input for SLM of aluminium alloys was identified to be around 50–100 J/mm³ in order to achieve dense

Fig. 9. Process parameter window in terms of specific energy Wg and k -factor to achieve dense defect-free structures for a substrate at different initial temperatures, room temperature TS = 25 °C (a), and a pre-heated substrate TS = 150 °C (b). 727

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

Fig. 10. Micrograph of a LMD-produced structure by using a specific energy input Wg of 12.8 kJ/g and a k -factor of 1.3 on a substrate at initial room temperature, showing typical metallographic characteristics, such as changing grain shape and orientation along the height of the layer, as well as inter-layer segregation and defects, such as cracking, along the grain boundaries.

material using non-optimized process parameters, as indicated in Fig. 11(b). The results show that the shapes of the pores are spherical in singlelayer LMD, as assumed in Section 4. Cracks are a challenging problem in LMD. They are found to mostly proceed along the grain boundaries (see Figs. 10 and 11), or along the transition zones between the beads or the substrate (see Fig. 11 (b)). During AM, cracking mostly occurs due to high cooling rates, grain-boundary segregation, and non-homogeneous temperature fields, thus leading to internal residual stresses [24,37–39]. Stress concentrations in the lead-in and lead-out areas support crack initiation up to a total separation of two beads or a bead and substrate material, which is known as delamination [41]. However, cracking within the deposited material was detected infrequently using these Al-Mg alloys in LMD. Pores were found to be the main type of internal defect. This is in good agreement with the literature were AlMg alloys were stated to frequently yield in poor density by the effect of magnesium content on the melt viscosity and reduction of wettability [42]. From the observations regarding grain shape and orientation, a nonhomogeneous cooling condition is assumed. It is assumed that the heat transfer near the substrate is extensively higher than on top of the structure, which is in good agreement with the literature [43–45]. Nonhomogenous heat transfer conditions, resulting in changing grain sizes with increasing distance to the substrate, are assumed. However, for a

would result in high penetration depth but poor surface qualities of the deposited material. For larger pre-heating temperatures, TS = 300 °C, random radiographic investigations of the samples documented that the observations for TS = 150 °C are valid for larger temperatures as well. However, the specific energy input, in case of TS = 300 °C, to achieve dense structures can be additionally reduced. In this experimental study, the process parameter window resulting in dense beads for TS = 150 °C were also found suitable for TS = 300 °C. 5.3. Metallography Fig. 10 shows a resulting micrograph of one bead, illustrating, in particular, a z-directional grain growth, which is typical for LMD-processed metallic structures [45–46]. The general morphology of the microstructure is coarse-grained. Near the top surface, the grains are globular, which is related to the thermal gradient between the solidifying material and the surrounding atmosphere. Interestingly, a transverse grain growth from the prior deposit into the newly added bead is observed. During multi-layer LMD, the globular grains in the upper regions are melted once again and form the z-directional grain growth through the transition zone of the beads; these have been plotted in Fig. 11(a) and (b). In the majority of the cases, defects such as pores and cracks are detected in the transition zones between the beads or the substrate

Fig. 11. Typical grain shape and z-directional grain growth behaviour (a), pore and crack occurrence and growth direction (b), observed in a deposited wall structure by using a specific energy input Wg of 12.8 kJ/g and a k - factor of 1.3 on a substrate, initially at room temperature. 728

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

of 4 kJ/g. Processing higher specific energies results in more heat input into the materials. Therefore, the temperature of the substrate increases, leading to an increased melting pool and, consequently, deeper thermal penetration. Higher dilution depths and widths, expressed by the increased aspect ratio RB and dilution ratio η, represent this increase in the heat input. From the previous observations, it can be concluded that defect-free structures are achieved by using a range of specific energy inputs. By using these inputs, the resulting bead geometry can be influenced and adapted for the required application. In addition, the penetration depth into the underlying structure can be influenced. By using the identified process parameter window for T S = 150 °C, two different bead shapes, depicted in Fig. 14, were deposited using high- or low-process energies and low or high deposition rates, respectively. Fig. 14(a) shows a bead deposited at a low specific energy input but at a high deposition rate. The shape of the upper part of the bead is similar to a top-hat distribution; this is advantageous for subsequent layer deposition. By using a specific process energy of Wg = 12.6 kJ/g combined with a deposition rate of m˙ = 21 g/min high-angled, low-diluted, z -directional oriented beads can be processed. An unetched cross-section of a manufactured bead is plotted in Fig. 14(c). In contrast, a high-diluted, low side-angled

detailed conclusion with respect to the grain orientation and size, additional texture analyses are required, representing future work.

5.4. Geometrical analysis As shown in Fig. 9, a wide range of specific energy inputs by varying the k-factor can be used to process defect free-beads, which show a high surface quality. As shown in the following section, these process parameter ranges can be adjusted to tailor specific geometrical features of the beads. Figs. 12 and 13 summarize the resulting aspect and dilution ratios, RB , RC , and η, as well as the bead side-angle as a function of the specific energy input Wg and the k-factor for a pre-heated substrate of TS = 150°C . It is shown that the aspect ratio RB increases by around 50 percent, whereas RC decreases by around 40 percent for a specific energy increase from 12.5 to 16.5 kJ/g. By this increase, the bead geometry becomes wider and flattens, which is expressed by the decreasing side-angle of −2.5° per kJ/g, which shown in Fig. 13(a). With regard to the dilution ratio, Eqs. (4.9) and (4.10) are used. However, Eq. (4.9) is more complex and requires more input data. For comparison, both equations are plotted in Fig. 13(b). It is observed that Eq. (4.10), which is a simplified form of Eq. (4.9), only slightly overestimates the dilution rate. The dilution ratio increases to around 30 percent with an increase

Fig. 12. Changes in the geometrical shape of the deposited beads by using the identified process parameter window at a pre-heated substrate of TS = 150 °C. An increase of the aspect ratio RB for higher specific energies Wg and lower k-factors expressing an enlargement of the deposited beads (a) as well as a decreased aspect ratio R c , expressing a flatting of the beads for higher specific energies Wg and k-factors (b) can be seen.

Fig. 13. Changes in the geometrical shape of the deposited beads by using the identified process parameter window at a pre-heated substrate of TS = 150 °C. A decrease of the bead side-angle α with an increase of the specific process energy Wg and decreasing k-factors of around −2.5° per kJ/g is shown (a). In addition, the comparison of Eq. (4.9) and Eq. (4.10), which show a good agreement but a slight overestimation using Eq. (4.10) of the dilution ratio η. An increasing dilution ratio η for the increased specific process energies Wg and decreasing k-factors of around 30 percent per 4 kJ/g can be seen (b). 729

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

Fig. 14. Theoretical top hat-distribution (a) and Gaussian-distribution shapes (b) as well as the unetched cross-sections of the deposited beads showing top-hat (c) and Gaussian (d) shapes deposited on TS = 150 °C.

flat bead geometry is deposited by using the specific process energy of Wg = 16.4 kJ/g and a deposition rate of m˙ = 15 g/min (Fig. 14(b) and (d)). These results illustrate that it is possible to process specific bead shapes and dilution ratios by using different parameter sets, thereby allowing the tailoring of the desired geometric features. During the deposition of a given componential geometry, the position of one layer next to a subsequent layer has to be precisely prearranged in order to achieve high surface quality. Therefore, the overlapping ratio μc and the optimum bead spacing dc, as shown in Fig. 4(b), needs to be precisely determined. In the following section, the optimum bead spacing based on the geometrical assumptions from Fig. 7 are analysed by using the process parameters of Wg = 16.4 kJ/g and a kfactor of 1.4 on the pre-heated substrates of TS = 150 °C and 300 °C. Through geometric examinations of the beads, the values of H and ω for the specific process parameters were determined. By using Eq. (4.7), an optimum bead spacing of dc = 2.45 mm in the case of a pre-

heated substrate of TS = 150 °C has been calculated. According to Eq. (4.8), this leads to an overlapping of μc =18 percent. The surface shapes of the conducted experiments for different bead spacing are shown in Fig. 15. A surface quality of Rt = 130 μm is achieved. A further decrease of bead spacing to dc = 2.15 mm did not improve surface roughness. Hence, Fig. 15(c) depicts the optimal bead spacing conditions for a preheated substrate of 150 °C for the used process parameters. Assuming that the substrate temperature exceeds 150 °C in multi-layer LMD, the bead shape and the optimization of bead spacing has to be adjusted accordingly. Therefore, these are investigated for a pre-heated substrate at TS = 300 °C as well (see Fig. 15(e)–(h)). The increased substrate temperature leads to an improvement of the surface quality. For this substrate temperature, a bead spacing of dc = 2.8 mm represents an optimum, resulting in a surface roughness of Rt = 80 μm and an overlapping ratio of μc = 45 percent. In summary, homogenous heat distribution is a very important factor in terms of process stability and the

Fig. 15. Results from metallographic investigations concerning different bead spacings of (a) dc = 3.4 mm, (b) dc = 2.9 mm, (c) dc = 2.45 mm and (d) dc = 2.15 mm for TS = 150 °C and bead spacings of (e) dc = 4.1 mm, (f) dc = 3.6 mm, (g) dc = 3.1 mm and (h) dc = 2.8 mm for Ts = 300 °C. 730

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

reproducibility of constant bead shape quality. [9]

6. Conclusion

[10]

Within this study, a process of wire-based laser metal deposition of the AlMg alloys EN AW-5087 as wire and EN AW-5754 as substrate material was successfully developed and investigated. The relationships between the specific process energy, k-factor, porosity, and the resulting geometric shape of the deposited beads were shown using a systematic parametric study. Pre-heating of the substrate turned out to be beneficial in terms of the reduction of porosity and cracking. By using no pre-heating of the substrate, only a deposition velocity of 1 m/min in combination with a specific energy input of 10–12 kJ/g achieved dense beads with adequate surface qualities. In contrast, dense structures, using specific process energies between 10 to 16.5 kJ/g, by using deposition velocities from 1–5 m/min on a pre-heated substrate of 150 °C were achieved. It turned out that the necessary specific energy input is reduced by around 50 percent in comparison to powder-based approaches, such as SLM of aluminium alloys. It was shown that it is possible to tailor the geometrical features of the deposited beads by a specific adjustment of the process parameters within the identified process window. Through this adjustment, high diluted, broad expanded, and flat beads, showing a low side-angle, were deposited by using high specific energy densities and low k-factors and vice versa. As post-processing is an expensive (but still necessary) part in AM, controlled layer geometries contribute to its reduction. The results contribute to the possible implementation of wire-based LMD of aluminium alloys in automated industrial processes. High deposition rates and controllable layer geometries to process either tall structures or to coat surfaces can be processed. Regarding temperature distributions, cooling rates, and the resulting features attributed to heat transfer mechanisms, only assumptions were possible. More research using controlled temperature measurements are required in order to identify temperature-related effects. In addition, future research to investigate the detailed microstructure of the specimens, mechanical properties (e.g., stress-strain behaviour, microhardness, fatigue crack initiation, and crack propagation) is necessary. A further study could also assess the identification of process parameters to reduce residual stresses and componential distortions.

[11] [12] [13]

[14]

[15]

[16]

[17] [18]

[19]

[20]

[21]

[22] [23] [24]

[25]

[26]

Acknowledgements

[27]

The authors would like to thank Mr. R. Dinse, Mr. P. Haack, and Mr. F. Dorn from Helmholtz-Zentrum Geesthacht for their valuable technical support.

[28]

[29]

References

[30]

[1] Williams SW, Martina F, Addison AC, Ding J, Pardal G, Colegrove P. Wire + Arc additive manufacturing. Mater Sci Technol 2016;32:641–7. https://doi.org/10. 1179/1743284715Y.0000000073. [2] Gu J, Cong B, Ding J, Williams SW, Zhai Y. Wire+Arc additive manufacturing of aluminium. USA, Texas: SFF Symp Austin; 2014. p. 451–8. [3] Syed WUH, Pinkerton AJ, Li L. A comparative study of wire feeding and powder feeding in direct diode laser deposition for rapid prototyping. Appl Surf Sci 2005;247:268–76. https://doi.org/10.1016/j.apsusc.2005.01.138. [4] Dubourg L, Archambeault J. Technological and scientific landscape of laser cladding process in 2007. Surf Coatings Technol 2008;202:5863–9. https://doi.org/10.1016/ j.surfcoat.2008.06.122. [5] Fessler J, Merz R, Nickel A, Prinz F, Weiss L. Laser deposition of metals for shape deposition manufacturing. Univ Texas Austin, USA: Solid Free Fabr Symp Proceedings; 1996. p. 117–24. [6] Addison A, Ding J, Martina F, Lockett H, Williams S, Zhang X. manufacture of complex titanium parts using Wire+Arc additive manufacture. Hilt Birmingham Metrop Hotel; 2015. [7] Murr LE, Gaytan SM, Ramirez DA, Martinez E, Hernandez J, Amato KN, et al. Metal fabrication by additive manufacturing using laser and electron beam melting technologies. J Mater Sci Technol 2012;28:1–14. https://doi.org/10.1016/S10050302(12)60016-4. [8] Olakanmi EO, Cochrane RF, Dalgarno KW. A review on selective laser sintering/ melting (SLS/SLM) of aluminium alloy powders: processing, microstructure, and

[31] [32] [33]

[34]

[35]

[36]

[37]

[38]

731

properties. Prog Mater Sci 2015;74:401–77. https://doi.org/10.1016/j.pmatsci. 2015.03.002. Abioye TE, Folkes J, Clare AT. A parametric study of inconel 625 wire laser deposition. J Mater Process Technol 2013;213:2145–51. https://doi.org/10.1016/j. jmatprotec.2013.06.007. Schmidt M, Merklein M, Bourell D, Dimitrov D, Hausotte T, Wegener K, et al. Laser based additive manufacturing in industry and academia. CIRP Ann 2017;66:561–83. https://doi.org/10.1016/j.cirp.2017.05.011. Frazier WE. Metal additive manufacturing: a review. J Mater Eng Perform 2014;23:1917–28. https://doi.org/10.1007/s11665-014-0958-z. Kathuria YP. Some aspects of laser surface cladding in the turbine industry. Surf Coat Technol 2000;132:262–9. https://doi.org/10.1016/S0257-8972(00)00735-0. Kaierle S, Barroi A, Noelke C, Hermsdorf J, Overmeyer L, Haferkamp H. Review on laser deposition welding: from micro to macro. Phys Procedia 2012;39:336–45. https://doi.org/10.1016/j.phpro.2012.10.046. Syed WUH, Li L. Effects of wire feeding direction and location in multiple layer diode laser direct metal deposition. Appl Surf Sci 2005;248:518–24. https://doi. org/10.1016/j.apsusc.2005.03.039. Nowotny S, Scharek S, Beyer E, Richter KH. Laser beam build-up welding: precision in repair, surface cladding, and direct 3D metal deposition. J Therm Spray Technol 2007;16:344–8. https://doi.org/10.1007/s11666-007-9028-5. Huang SW, Samandi M, Brandt M. Abrasive wear performance and microstructure of laser clad WC/Ni layers. Wear 2004;256:1095–105. https://doi.org/10.1016/ S0043-1648(03)00526-X. Davis JR. Aluminum and aluminum alloys. ASM International; 1993. Williams SW, Martina F, Addison AC, Ding J, Pardal G, Colegrove P. Wire + Arc additive manufacturing. Mater Sci Technol 2016;32:641–7. https://doi.org/10. 1179/1743284715Y.0000000073. Froend M, Fomin F, Riekehr S, Alvarez P, Zubiri F, Bauer S, et al. Fiber laser welding of dissimilar titanium (Ti-6Al-4V / cp-Ti) T-joints and their laser forming process for aircraft application. Opt Laser Technol 2017;96:123–31. https://doi.org/10.1016/j. optlastec.2017.05.017. Mukherjee T, Zhang W, DebRoy T. An improved prediction of residual stresses and distortion in additive manufacturing. Comput Mater Sci 2017;126:360–72. https:// doi.org/10.1016/j.commatsci.2016.10.003. Senthilkumaran K, Pandey PM, Rao PVM. Influence of building strategies on the accuracy of parts in selective laser sintering. Mater Des 2009;30:2946–54. https:// doi.org/10.1016/j.matdes.2009.01.009. Incropera FP, DeWitt DP, Bergman TL, Lavine AS. Fundamentals of heat and mass transfer. 2007. https://doi.org/10.1073/pnas.0703993104. Jung H. A study on laser forming processes with finite element analysis. Dr Thesis. Univ Canterbury, New Zeal; 2006http://hdl.handle.net/10092/1099. Bose S, Ke D, Sahasrabudhe H, Bandyopadhyay A. Additive manufacturing of biomaterials. Prog Mater Sci 2018;93:45–111. https://doi.org/10.1016/j.pmatsci. 2017.08.003. Maliutina IN, Si-Mohand H, Piolet R, Missemer F, Popelyukh AI, Belousova NS, et al. Laser cladding of gamma-TiAl intermetallic alloy on titanium alloy substrates. Metall Mater Trans A: Phys Metall Mater Sci 2015;47:378–87. https://doi.org/10. 1007/s11661-015-3205-9. Syed WUH, Pinkerton AJ, Li L. Combining wire and coaxial powder feeding in laser direct metal deposition for rapid prototyping. Appl Surf Sci 2006;252:4803–8. https://doi.org/10.1016/j.apsusc.2005.08.118. Huang J, Li Z, Cui H, Yao C, Wu Y. Laser welding and laser cladding of high performance materials. Phys Procedia 2010;5:1–8. https://doi.org/10.1016/j.phpro. 2010.08.023. Xu J, Liu W, Kan Y, Zhong M. Microstructure and wear properties of laser cladding Ti–Al–Fe–B coatings on AA2024 aluminum alloy. Mater Des 2006;27:405–10. https://doi.org/10.1016/j.matdes.2004.11.011. Carcel B, Sampedro J, Ruescas A, Toneu X. Corrosion and wear resistance improvement of magnesium alloys by laser cladding with Al-Si. Phys Procedia 2011;12:353–63. https://doi.org/10.1016/j.phpro.2011.03.045. Cao X, Jahazi M, Fournier J, Alain M. Optimization of bead spacing during laser cladding of ZE41A-T5 magnesium alloy castings. J Mater Process Technol 2008;205:322–31. https://doi.org/10.1016/j.jmatprotec.2007.11.120. Toyserkani E, Khajepour A, Corbin S. Laser cladding. London, New York, Washingt DC: CRC Press Boca Raton; 2005. p. 280. Steen WM, Mazumder J. Basic laser optics. Laser mater. Process. Springer Verlag; 2010. p. 106–7. https://doi.org/10.1007/978-1-84996-062-5_3. Zhang K, Wang S, Liu W, Shang X. Characterization of stainless steel parts by laser metal deposition shaping. Mater Des 2014;55:104–19. https://doi.org/10.1016/j. matdes.2013.09.006. Zhang K, Zhang XM, Liu WJ. Influences of processing parameters on dilution ratio of laser cladding layer during laser metal deposition shaping. Adv Mater Res 2012;549:785–9. https://doi.org/10.4028/www.scientific.net/AMR.549.785. Olakanmi EO. Selective laser sintering/melting (SLS/SLM) of pure Al, Al-Mg, and Al-Si powders: effect of processing conditions and powder properties. J Mater Process Technol 2013;213:1387–405. https://doi.org/10.1016/j.jmatprotec.2013. 03.009. Kimura T, Nakamoto T. Microstructures and mechanical properties of A356 (AlSi7Mg0.3) aluminum alloy fabricated by selective laser melting. Mater Des 2016;89:1294–301. https://doi.org/10.1016/j.matdes.2015.10.065. Hekmatjou H, Naffakh-Moosavy H. Hot cracking in pulsed Nd:YAG laser welding of AA5456. Opt Laser Technol 2018;103:22–32. https://doi.org/10.1016/j.optlastec. 2018.01.020. Ojo OO, Taban E, Kaluc E. Loop travel-path of fibre laser welded alclad AA2219-O alloy. J Mater Process Technol 2018;251:118–26. https://doi.org/10.1016/j.

Journal of Manufacturing Processes 34 (2018) 721–732

M. Froend et al.

mechanical property considerations in additive manufacturing of aluminum alloys. MRS Bull 2016;41:745–51. https://doi.org/10.1557/mrs.2016.214. [43] Cheng XM, Morris JG. Texture, microstructure and formability of SC and DC cast AlMg alloys. Mater Sci Eng A 2002;323:32–41. https://doi.org/10.1016/S09215093(01)01393-4. [44] Gu J, Wang X, Bai J, Ding J, Williams S, Zhai Y, et al. Deformation microstructures and strengthening mechanisms for the wire+arc additively manufactured AlMg4.5Mn alloy with inter-layer rolling. Mater Sci Eng A 2018;712:292–301. https://doi.org/10.1016/j.msea.2017.11.113. [45] Zhang C, Li Y, Gao M, Zeng X. Wire arc additive manufacturing of Al-6Mg alloy using variable polarity cold metal transfer arc as power source. Mater Sci Eng A 2018;711:415–23. https://doi.org/10.1016/j.msea.2017.11.084.

jmatprotec.2017.08.026. [39] Chen Y, Lu F, Zhang K, Nie P, Reza S, Hosseini E, et al. Investigation of dendritic growth and liquation cracking in laser melting deposited inconel 718 at different laser input angles. Mater Des 2016;105:133–41. https://doi.org/10.1016/j.matdes. 2016.05.034. [40] Sergio M, De Lima F, Sankaré S. Microstructure and mechanical behavior of laser additive manufactured AISI 316 stainless steel stringers. Mater Des 2014;55:526–32. https://doi.org/10.1016/j.matdes.2013.10.016. [41] Mukherjee T, Zuback JS, Zhang W, DebRoy T. Residual stresses and distortion in additively manufactured compositionally graded and dissimilar joints. Comput Mater Sci 2018;143:325–37. https://doi.org/10.1016/j.commatsci.2017.11.026. [42] Ding Y, Muñiz-Lerma JA, Trask M, Chou S, Walker A, Brochu M. Microstructure and

732