Within laser additive manufacturing (directed energy deposition with laser beam), processes are further distinguished by the form of the filler material. In terms of availability, storage, safety, and cost, wire is commonly the preferred filler material in comparison to powder. Despite these advantages, due to the different material transfer modes, a greater process control is required. Within this work, an experimental setup for electrical-resistance-measurement within the laser material deposition process with a coaxial wire feed and its possible use for an automated process control is investigated. The measurement is performed between a wire, a substrate, and over the melt pool. One main influencing factor on process stability is derived from the timing of the trigger sequence of the laser power, process feed, and wire feed at the start and end points of every track. Consequently, inaccurate settings of the trigger sequence can, e.g., lead to deviations in track length and part geometry. Additionally, a smooth transfer of the wire into the melt pool is imperative during part build-up to ensure a stable deposition process. Variation in laser power, wire feed, process feed, or wire transfer mode can lead to process instabilities. This can result in imperfections, bonding defects, or pores in the tracks and layers that will add up in built components and must be avoided for defectfree three-dimensional geometries. Within the experiments, it is investigated whether the resistance-measurement provides consistent results under varying conditions and potentially can be utilized to automate the trigger sequence of deposition. Furthermore, it is investigated whether different wire transfer modes can be linked to the measured resistance values during welding of single tracks.

Additive manufacturing processes are used for repairing, coating, diversification, and building of parts. The layerwise build processes are distinguished by the deployed deposition method, energy source, and feedstock form and material. In laser-based directed energy processes (DED-LB), metal powder and wire filler materials are used.1 Due to safer handling, lower procurement costs, lower machine and personnel contamination, and almost 100% material usage, wire filler materials have become more and more popular.2,3 In wire laser material deposition (WLMD), lateral and coaxial wire feed are used. The latter enables a direction-independent deposition process, which is essential for 3D applications.3,4 At Fraunhofer ILT, a processing head for the coaxial arrangement of the wire and an annular shaped beam has been developed and investigated.5 The ratio of the inner and the outer diameter of the annular beam is defined by the optical setup and the focal length of the used focusing lens. The beam diameter in the working plane is varied by defocusing and the inner beam diameter must be larger than the wire diameter. The width of the welded tracks is linked to the outer diameter of the annular laser beam and varies with decreasing or increasing working distance.6 The main parameters for WLMD are wire diameter, wire feed rate, process feed rate, and laser power. The wire to melt pool transfer can be observed via lateral welding cameras and is divided into transfer modes “weak link” or “balling,” “smooth wire transfer,” and “wire stubbing.” In this paper, “weak link” is defined as the constriction of a molten metal between the melt pool and the wire tip. If the connection is lost, the wire can burn up resulting in balling at the tip and total process aborts. Weak link and wire stubbing result in an uneven surface and may cause the process to abort.7–9 For a stable deposition process and a smooth surface finish, the wire feed rate has to be controlled and approximately match the melting rate of the filler wire.10 Deviations in single tracks add up as the number of layers increases and can lead to shape deviations or unstable process conditions due to changes in the working distance. The most common causes for deviations are start, end, or crossing points, changes in the process feed rate or deposition in corners and on edges of previously deposited layers. Deviations can be influenced and minimized by changing the switch-on and switch-off sequences, feed rates, and the main process parameters.11,12 The prerequisite for controlling the processes is sufficiently accurate process monitoring. Process monitoring is, therefore, crucial for the production of high-quality parts.13 Often used monitoring methods are melt pool monitoring with CCD cameras,14,15 thermal imaging of the melt pool or pyrometry,16 and height measurement of the deposited weld bead.10,14,17 For lateral WLMD, another possible method is the electrical resistance measurement between the wire, the substrate, and over the melt pool. The resistance and wire feed rate can be converted by a static regression model and can be used to control the distance between the tool and the substrate.18 When using the coaxial deposition head, changes in the distance between the processing head and the substrate result in a different working plane and changes in the annular beam diameter are, therefore, unwanted. The method can be transferred, and the measurement can be fed to an iterative learning control system, which then controls the wire feed rate. To enable the use of small currents that do not heat the wire, a Wheatstone bridge is used.12 

In this work, it is investigated whether contact and sure process start, or process interruptions can be detected automatically by means of resistance measurement. Furthermore, it is investigated whether the measured values of the resistance measurement can be assigned to a transfer mode to not only control the process but also obtain online data for quality assurance.

In the experiments, 1.4301 stainless steel (AlSl 304) substrates and Ø0.8 mm 1.4430 stainless steel (AlSl AWS 316 L Si) wire are used. All substrate surfaces are sand blasted and cleaned with ethanol prior to the welding processes. The feed direction of the filler wire and the laser beam are coaxially aligned in this work. By means of optical elements in the processing head, an annular beam is formed with a waxicon, divided into two parts, and reassembled around the wire. The annular beam is focused with a lens with a focal length of 60 mm. The annular beam can be described with an inner and outer diameter or the ring width. The diameters can be adjusted by defocussing. The laser beam is generated by a disk laser with 4 kW maximum output power and 8 mm mrad beam quality, which is coupled to a NA 0,1, and Ø200 μm laser fiber. The filler wire is supplied on a spool and conveyed through a wire hose assembly by a push pull system, with two separate motors with four rollers each. A straightening unit with two planes is used between the wire hose assembly and the master feeder to straighten the bends coming from the production and to remove wire casting. The straightened wire is fed into a self-built, water-cooled, 1.4301 stainless steel wire guide, which is mounted in the processing head and guides the wire through the split beam. In the wire guide, the wire is guided through a PTFE sheath and is, thus, insulated from the guide. At the end of the guide, the wire is fed into the melt pool through a tip-shaped contact tube made of copper (CuCrZr1) with Ø0.8 mm inner bore. One of the contact points of the resistance measurement setup is clamped on the top of the wire guide. The other one is clamped to the substrate. Wire spool, feeder, and processing head are insulated so that no electrical shunt can occur. The experimental setup is controlled with an industrial computer with computerized numerical control (CNC) functions and a programmable logic controller (PLC). With the CNC, the processing head and the camera are manipulated by means of a cartesian three-axis machine. The setup is depicted in Fig. 1.

FIG. 1.

Machine setup and equipment.

FIG. 1.

Machine setup and equipment.

Close modal

With the PLC, complex program sequences can be executed; electrical inputs and outputs (I/O) are set, read, and monitored; and data are exchanged with an open platform communication (OPC) data access (DA) interface. The OPC DA interface supports discontinuous sampling with a sampling time between 100 and 150 ms. With analog input modules, current in a range of ±20 mA or voltage in a range of ±10 V can be measured. By means of the OPC DA interface, the time-resolved electrical current values measured by the I/Os are accessed and continuously stored in the comma separated values (CSV) file format using a python script. Furthermore, the processes were recorded with a laterally mounted welding camera with a shortpass filter with 850 nm cutoff wavelength. The camera is moved with the processing head. The recorded videos are used to evaluate the process stability, examine the transfer modes, and investigate faulty conditions on edges.

The investigation is divided into four steps. First, the resistance measurement method is modified for the coaxial processing head of Fraunhofer ILT. The overall resistance and feasibility are investigated. In the second step, the experiment is connected to the PLC. Due to the measuring range of the I/Os and the large number of possible measuring variants, interfaces between the experiment and the PLC must be developed. In the third step, it is investigated whether the WLMD process can be started or stopped automatically based on the signal. In addition, the feasibility of automatic start and stop at edges is investigated. In the fourth step, the interface is enhanced so that the actual resistance value can be determined. The interface is calibrated to match the wire diameter and the material. Furthermore, it is investigated whether the measured resistance value can be related to the observed transfer modes during deposition.

All components of the system are evaluated before the experiments, and the parameters refer to the working plane on the substrate surface. In all experiments, argon with a pressure of 3 bar and a volume flow of 8–12 l/min is used as the shielding gas. The wire stickout between the wire end and the substrate is set to 4.5 mm each time, and the concentric position of the wire and the annular beam is checked. The main process parameters for all experiments are depicted in Table I.

TABLE I.

Process parameters.

ParameterSymbolValueUnit
Wire diameter dw 0.8 mm 
Laser power PL 1100 
Process feed rate vp 300 mm/min 
Shielding gas flow VAr 8–12 l/min 
Stick out sw 4.5 mm 
Inner beam diameter ØL 1.1 mm 
ParameterSymbolValueUnit
Wire diameter dw 0.8 mm 
Laser power PL 1100 
Process feed rate vp 300 mm/min 
Shielding gas flow VAr 8–12 l/min 
Stick out sw 4.5 mm 
Inner beam diameter ØL 1.1 mm 

In the experiments, the wire guide with the contact tube is used as part of the measuring circuit and connected with Ø1 mm2 cable with connectors at each end. One cable is screwed to the wire guide and one cable is clamped to the substrate. The total resistance of the circuit is measured to check compatibility with I/Os measurement ranges. The setup schematic is depicted Fig. 2.

FIG. 2.

Schematic connection setup.

FIG. 2.

Schematic connection setup.

Close modal

Preliminary tests are performed to determine whether the cut and shape of the wire end affect the signal. In the following figures, the schematic connection is depicted as a switch or a resistor.

To connect the measurement circuit to the PLC, a cost-efficient and simple interface is needed. Low currents must be used in the measuring circuit to avoid heating of the wire.18 PLC and the measurement circuit are connected to an optocoupler. On the measurement side, a voltage U1 of 10 V is applied by a constant voltage supply, and on the PLC side, a voltage of 24 V is applied by connecting the signal circuit to a digital I/O. A schematic circuit diagram of the interface is depicted in Fig. 3.

FIG. 3.

Interface for contact measurement.

FIG. 3.

Interface for contact measurement.

Close modal

The function of the interface is evaluated in 10 trials. The wire feed is switched on and the response signal of the PLC is observed when the wire contacts the substrate. The interface is considered reliable if the module switches in 10 out of 10 cases.

The signal is then used to either switch on or off laser power and process feed. Ten single tracks with 30 mm length are welded. According to the previous research,11 the switching sequence of the wire feed, the laser beam, and the process feed are adjusted for the setup to deposit single tracks without height deviation in start and end points. For validation, eight single tracks of 30 mm length are deposited on 50 × 100 × 10 mm3 substrates. The deposited single tracks are measured with 3D-profilometry, and the height is compared.

Furthermore, reliability of the contact measurement and repeatability are investigated. When building 3D parts, the deposition ideally starts and ends directly at the edges. Besides this, two faulty conditions can occur that will lead to defects in the part. The wire slips off at a corner or the end edge, or the wire stickout is too large and the wire collides with the approached edge. The experiments are welded onto a 150 × 150 × 20 mm3 substrate with five parallelly milled 10 × 10 mm2 deep and wide slots. In “neutral” condition, the process is set to start and stop exactly at the web edges. The wire diameter is subtracted from the 10 mm weld length. In “forward” condition, start- and end points are shifted +1 mm to investigate the wire slipping of at the end. The last case is “rearward” condition, the start-end endpoint is shifted −1 mm to investigate possible collision and slip off at the process start. A sketch of the milled substrate is depicted in Fig. 4.

FIG. 4.

Substrate with slots and webs.

FIG. 4.

Substrate with slots and webs.

Close modal

During the deposition processes, videos are recorded with the weld camera, and the process is observed on a display. For the first 10 mm of every deposition, no edges are encountered to validate the general process setup and adjustment. “Neutral” condition is tested for 12 times. “Forward” and “rearward” conditions are tested for five times each. In total, the reliability is tested in 110 cases. The contact measurement is considered working when the deposition process starts after contact and is detecting faulty conditions, and switches off laser power, wire feed, and process feed after detection.

In the last step, the interface is adapted to measure the resistance RM during the deposition process. By means of a Wheatstone bridge, the voltage UB, which correlates with the resistance between the wire, the melt pool, and substrate, can be measured over the bridge. The bridge circuit consists of two identical resistors RB, a reference resistor RRef, and the resistance to be determined RM. RRef is chosen based on values of preliminary tests. UB is amplified by an instrumentation amplifier to match the measuring range of the analog I/O. By changing the resistor RG, the amplification factor is adjusted. The reference voltage U2 is constantly 5 V and voltage changes in the circuit are smoothed out by capacitors. With the PLC and a python script, the time-resolved voltage is saved in the CSV file format, and the resistance is calculated by using the following equation:

RM=RRefUBU2(RRef+RB)1+UBU2(RRefRB+1).
(1)

The circuit diagram of the enhanced interface is depicted in Fig. 5.

FIG. 5.

Setup for resistance measurement.

FIG. 5.

Setup for resistance measurement.

Close modal

For the experiments, different wire feed rates for “stubbing,” “smooth,” and “weak link” transfer are set. During the observations that are characterized as “weak link,” a constriction between the meltpool and the wire tip is observed, but connection is not completely lost. The wire feed rates and corresponding process observation images are depicted in Fig. 6.

FIG. 6.

Wire feed rate and transfer modes.

FIG. 6.

Wire feed rate and transfer modes.

Close modal

The functionality of the interface and the measurement is investigated in experiments. With each of the determined wire feed rates, ten single tracks are deposited onto 150 × 150 × 10 mm3 substrates. The single tracks are 60 mm long in total. The first and last 5 mm of the track are discarded, so that no deviations caused by start or end points influence the measurements. The voltage is measured with the interface and the resistance is calculated. The obtained resistance values are filtered with a sixteenth-order moving average filter (MAF) and plotted in diagrams. At the same time, the processes are filmed and observed with the welding camera. A schematic of single tracks is depicted in Fig. 7.

FIG. 7.

Path planning on resistance measurement samples.

FIG. 7.

Path planning on resistance measurement samples.

Close modal

To investigate whether changes in transfer modes during the deposition process can be detected with the interface and measuring setup, ten single tracks with randomized sequenced transfer modes are deposited onto 150 × 150 × 10 mm substrates. The tracks are 95 mm long in total and divided into three 25 mm sections with different wire feed rates. The first and last 10 mm of the tracks are discarded. The voltage is measured with the interface and the resistance is calculated. The obtained resistance values are filtered with a sixteenth-order MAF and plotted in diagrams. A schematic of the single tracks with three sections for different wire feed rates is depicted in Fig. 7.

All tracks are measured with 3D profilometry. From the height profile, start, end point, and average track height are derived, and the deviation is calculated. At each step, welded single tracks are examined for pores, bonding defects, or imperfections using light optical microscopy of polished cross sections for all parameter sets used in the work.

With the first setup of the interface and the digital I/0, the contact between the wire and the substrate can be detected correctly 10 out of 10 times.

With the set process parameters and no sequential timing of laser power and process feed start, the deposition process starts 10 out of 10 times. Each time a single track is deposited. Start point height deviates, with a maximum at 0.82 mm and an average track height of 0.61 mm.

The adjusted sequence times are 0.3 s from contact detection until the laser beam power is switched on and another 0.3 s until the process feed is started. With the adjusted times, eight out of eight single tracks are deposited. All single tracks are free of pores or defects and reach 30 mm in total length. The contact angles are obtuse, and the average aspect ratio of height to width is 0.31. The maximum deviation in the start or endpoints is 5.3% with an average track height of 0.63 mm. The tracks’ surfaces are smooth, and no waviness or irregularities is measured. An exemplary track deposited with the parameters depicted in Table I, the measured height profile, and a cross section cut are depicted in Fig. 8.

FIG. 8.

Height profile and cross section of a single track.

FIG. 8.

Height profile and cross section of a single track.

Close modal

The reliability testing is conducted on the webs of the milled substrate. All validation sections are deposited without any defects or deviations. For the “neutral” condition, 60 of 60 track sections are deposited. For the “forward” condition, 25 of the 25 track sections are deposited. In 20 cases, the contact between the wire and the melt pool is broken in under 1 mm overhang at the end edge and the laser beam is switched off. In five cases, a material bridge between the wire and the melt pool can be observed for a distance above +1 mm. The maximum bridge length is 2.6 mm. Eventually, the bridge does collapse, and the laser beam is switched off. At the track sections deposited in the forward condition, an end point height increase can be observed and measured. The maximum height is 0.9 mm. In five cases, a balling behavior can be observed and the deposited material extends over the edge. An exemplary observation of the forward condition end edge sequence is depicted in Fig. 9.

FIG. 9.

Melt pool bridging of 2.6 mm at the end edge of the web deposited on.

FIG. 9.

Melt pool bridging of 2.6 mm at the end edge of the web deposited on.

Close modal

For the “rearward” condition, 17 of 25 track sections are deposited. In eight cases, the wire is already stuck out below the webs’ surface and collides with the edge. For collisions, two mechanisms are observed. The laser beam is switched on, and the wire is bent and directly exposed to the laser beam. The wire tip burns up and balling is observed. The loss of contact is detected, and the laser beam is switched off. Or, the laser beam is switched on, the wire is bent but does not burn up completely and touches down again within a short distance and the process stabilizes, and the material is deposited for the remaining section length. An exemplary observation of the collision at the start edge and process abort is depicted in Fig. 10.

FIG. 10.

Collision of wire and web edge, balling of the wire tip and process abort.

FIG. 10.

Collision of wire and web edge, balling of the wire tip and process abort.

Close modal

In total, 102 of 110 track sections are deposited. The shortest observed switch-off time is 100 ms from loss of contact to the laser beam switch off. In one set (11) of the “rearward” condition experiments, all sections are faulty. The substrate with all deposited sections is depicted in Fig. 11.

FIG. 11.

Substrate with deposited sections.

FIG. 11.

Substrate with deposited sections.

Close modal

With the enhanced interface, the smallest measurable resistance is 0.05 Ω and the largest measurable resistance is 0.225 Ω. The maximum measurement error is 5%. The measuring range can be adjusted by changing RRef, which is set to 0.05 Ω. No irregularities or defects are observed when applying the 60 mm single tracks with one transfer mode each. For deposition with “smooth” transfer mode, the smallest measured resistance is 0.112 Ω, the largest measured resistance is 0.134 Ω, and the average is 0.127 Ω. For deposition with the “weak link” transfer mode, the smallest measured resistance is 0.112 Ω, the largest measured resistance is 0.150 Ω, and the average is 0.131 Ω. For deposition with the “stubbing” transfer mode, the smallest measured resistance is 0.123 Ω, the largest measured resistance is 0.169 Ω, and the average is 0.149 Ω. All measured values of the “stubbing” transfer mode are at least 0.11 Ω larger than in the “weak link” or “smooth” transfer mode. The results of the measurements are depicted in Table II and Fig. 12.

FIG. 12.

Measurement results of single tracks with one transfer mode each.

FIG. 12.

Measurement results of single tracks with one transfer mode each.

Close modal
TABLE II.

Measured resistances for three transfer modes.

Characteristic valueStubbingSmoothWeak link
Minimum (Ω) 0.123 0.112 0.112 
Lower quartile (Ω) 0.140 0.119 0.125 
Median (Ω) 0.143 0.124 0.129 
Average (Ω) 0.149 0.127 0.131 
Upper quartile (Ω) 0.167 0.130 0.138 
Maximum (Ω) 0.169 0.134 0.150 
Characteristic valueStubbingSmoothWeak link
Minimum (Ω) 0.123 0.112 0.112 
Lower quartile (Ω) 0.140 0.119 0.125 
Median (Ω) 0.143 0.124 0.129 
Average (Ω) 0.149 0.127 0.131 
Upper quartile (Ω) 0.167 0.130 0.138 
Maximum (Ω) 0.169 0.134 0.150 

In the experiments with three different transfer modes per single track, the largest measured resistances are always measured for sections deposited with “stubbing” transfer. The filtered measured values vary between 0.14 and 0.20 Ω. Each time a change is made from “stubbing” to another transfer mode, a negative gradient is observed. The filtered measured values for “smooth” transfer vary between 0.11 and 0.14 Ω. The filtered measured values for “weak link” transfer vary between 0.10 and 0.13 Ω. Each time a change is made from “smooth” or “weak link” to another transfer mode, a negative gradient is observed. At points where the transfer mode is varied, no deviations or process instabilities are observed or measured. Four exemplary graphs of the filtered measured resistance values and the set transfer modes are depicted in Fig. 13.

FIG. 13.

Filtered resistance measurements of tracks with varying transfer modes.

FIG. 13.

Filtered resistance measurements of tracks with varying transfer modes.

Close modal

With the interfaces presented, the contact signal and contact resistance of the wire, the melt pool, and the substrate can be measured. By exchanging the resistors, the setup can be adjusted to different wire diameters or materials. To be able to make statistically reliable statements, it is necessary to repeat the experiments with different wire diameters, materials, and a larger number of tracks or when depositing complex parts.

The resistance measurement and the corresponding signal can be used to quickly detect a loss of contact between the wire and the melt pool and, therefore, a possible process abort. Without observation, a process abort would lead to uncontrolled balling and eventually to the destruction of the contact tip or worse. This can be avoided, with the reasonably cheap and simple setups presented. With the ability to automatically start a process as soon as the wire and the substrate come into contact, setting the complex timing sequence of the wire feed, the laser beam, and the process feed can be simplified and becomes easier for untrained users.

In the results, only the “stubbing” transfer mode can be reliably identified. The difference between “smooth” and “weak link” cannot be clearly identified with the setup. One possible reason for this is the use of the wire straightener. To achieve concentric feeding of the wire into the annular beam, the wire is straightened. In arc-welding, where similar contact tubes are used for power transfer, poor contact leads to process irregularities or abort. A defined three-point within the bore of the contact tube by a slight bend of the wire could improve the signal quality.

In further research, an improvement in the measurement and use of signals to control or monitor build processes could be investigated. For quality assurance, the data must be measured with equidistant time stamping. Additionally, a wider range of transfer modes and deviations like “wire slip off” or “collisions” must be investigated, assigned to measured values, and qualified.

The goal of this paper was to investigate whether contact and resistance measurement can be useful to identify process interruptions and whether the resistance can be assigned to an observed transfer mode.

With the developed interfaces, the measurement is possible and process aborts can be detected. With this knowledge, emergency stop functions can be integrated and the process start can be automated. Expensive and complex monitoring solutions as image monitoring or laser scanning can be replaced, and an operator will need less training.

In the resistance measurement, only one transfer mode can be distinguished. Further research is needed on the measurement setup and identification of the modes. Eventually, the data can be used for process control by means of machine learning or quality assurance, e.g., in a digital twin.

The authors have no conflicts to disclose.

Max Fabian Steiner: Conceptualization (lead); Data curation (lead); Formal analysis (lead); Funding acquisition (lead); Investigation (lead); Methodology (lead); Project administration (lead); Resources (lead); Supervision (lead); Validation (lead); Visualization (lead); Writing – original draft (lead); Writing – review & editing (lead). Philipp Holger Lohrer: Data curation (supporting); Investigation (equal); Software (equal); Visualization (equal). Thomas Schopphoven: Supervision (supporting); Writing – review & editing (supporting). Constantin Leon Häfner: Supervision (supporting).

1.
A.
Gebhardt
,
Additive Fertigungsverfahren: Additive Manufacturing und 3D-Drucken für Prototyping—Tooling—Produktion
, 5th ed. (
Hanser
,
München
,
2016
).
2.
F.
Klocke
,
K.
Arntz
,
N.
Klingbeil
, and
M.
Schulz
, “
Wire-based laser metal deposition for additive manufacturing of TiAl6V4: Basic investigations of microstructure and mechanical properties from build up parts
,” in
Laser 3D Manufacturing IV SPIE LASE
, edited by
B.
Gu
,
H.
Helvajian
,
A.
Piqué
,
C. M.
Dunsky
, and
J.
Liu
(
SPIE
, San Francisco,
2017
), p.
100950U
.
3.
J.
Kelbassa
,
A.
Gasser
,
J.
Bremer
,
O.
Pütsch
,
R.
Poprawe
, and
J.
Henrich Schleifenbaum
, “
Equipment and process windows for laser metal deposition with coaxial wire feeding
,”
J. Laser Appl.
31
,
022320
(
2019
).
4.
N.
Klingbeil
, “
Untersuchung des Drahtsystems zur Prozessstabilisierung für das drahtbasierte Laserauftragsschweißen in der additiven Fertigung
,” Ph.D.
dissertation
,
RWTH Aachen
, 2019.
5.
J.
Kelbassa
,
O.
Pütsch
,
J.
Stollenwerk
, and
P.
Loosen
, “
Koaxiale Laserstrahlformung für die Lasermaterialbearbeitung mit Zusatzwerkstoff
,”
PHOTONIK
2017
,
80
83
.
6.
J.
Kelbassa
,
A.
Biber
,
K.
Wissenbach
, and
O.
Pütsch
, “
Influence of focal length on the laser metal deposition process with coaxial wire feeding
,” in
High-Power Laser Materials Processing Applications, Diagnostics, and Systems VIII, San Francisco, California, 5–7 February 2019
, edited by
S.
Kaierle
and
S. W.
Heinemann
(
SPIE
,
Bellingham, WA
,
2019
), p.
11
.
7.
T. E.
Abioye
,
J.
Folkes
, and
A. T.
Clare
, “
A parametric study of Inconel 625 wire laser deposition
,”
J. Mater. Process. Technol.
213
,
2145
2151
(
2013
).
8.
M.
Motta
,
A. G.
Demir
, and
B.
Previtali
, “
High-speed imaging and process characterization of coaxial laser metal wire deposition
,”
Addit. Manuf.
22
,
497
507
(
2018
).
9.
E.
Govekar
,
A.
Jeromen
,
A.
Kuznetsov
,
M.
Kotar
, and
M.
Kondo
, “
Annular laser beam based direct metal deposition
,”
Proc. CIRP
74
,
222
227
(
2018
).
10.
S.
Takushima
,
D.
Morita
,
N.
Shinohara
,
H.
Kawano
,
Y.
Mizutani
, and
Y.
Takaya
, “
Optical in-process height measurement system for process control of laser metal-wire deposition
,”
Precis. Eng.
62
,
23
29
(
2020
).
11.
M. F.
Steiner
,
M.
Speier
,
J.
Kelbassa
,
T.
Schopphoven
, and
C. L.
Häfner
, “
Influence of tool path planning on process stability and deposition accuracy in laser material deposition with coaxial wire feed
,”
J. Laser Appl.
34
,
012026
(
2022
).
12.
P.
Hagqvist
,
A.
Heralić
,
A.-K.
Christiansson
, and
B.
Lennartson
, “
Resistance based iterative learning control of additive manufacturing with wire
,”
Mechatronics
31
,
116
123
(
2015
).
13.
S. K.
Everton
,
M.
Hirsch
,
P.
Stravroulakis
,
R. K.
Leach
, and
A. T.
Clare
, “
Review of in-situ process monitoring and in-situ metrology for metal additive manufacturing
,”
Mater. Des.
95
,
431
445
(
2016
).
14.
A.
Heralić
,
A.-K.
Christiansson
,
M.
Ottosson
, and
B.
Lennartson
, “
Increased stability in laser metal wire deposition through feedback from optical measurements
,”
Opt. Lasers Eng.
48
,
478
485
(
2010
).
15.
S.
Ocylok
,
E.
Alexeev
,
S.
Mann
,
A.
Weisheit
,
K.
Wissenbach
, and
I.
Kelbassa
, “
Correlations of melt pool geometry and process parameters during laser metal deposition by coaxial process monitoring
,”
Phys. Proc.
56
,
228
338
(
2014
).
16.
Z.
Yan
,
W.
Liu
,
Z.
Tang
,
X.
Liu
,
N.
Zhang
,
M.
Li
, and
H.
Zhang
, “
Review on thermal analysis in laser-based additive manufacturing
,”
Opt. Laser Technol.
106
,
427
441
(
2018
).
17.
A.
Heralić
,
A.-K.
Christiansson
, and
B.
Lennartson
, “
Height control of laser metal-wire deposition based on iterative learning control and 3D scanning
,”
Opt. Lasers Eng.
50
,
1230
1241
(
2012
).
18.
P.
Hagqvist
,
A.
Heralić
,
A.-K.
Christiansson
, and
B.
Lennartson
, “
Resistance measurements for control of laser metal wire deposition
,”
Opt. Lasers Eng.
54
,
62
67
(
2014
).