Challenges in Qualifying Additive Manufacturing For Turbine - Components - A Review
Challenges in Qualifying Additive Manufacturing For Turbine - Components - A Review
Challenges in Qualifying Additive Manufacturing For Turbine - Components - A Review
https://doi.org/10.1007/s12666-021-02199-5
ORIGINAL ARTICLE
Abstract Gas turbine engine advancements have been tolerant discs, high modulus blades, integrally bladed
enabled by innovation in materials and manufacturing rotors, etc. are some of the major breakthroughs in mate-
technologies. The evolution of additive manufacturing rials that have played a key role in powering the state-of-
(AM) has changed the face of direct digital technologies the-art engine developments [1–4]. Materials continue to
for the rapid production of models, prototypes, and func- play an enabling role in turbine engine performance and
tional parts including repair and maintenance for turbine future engine capabilities that require lightweight struc-
engines. Metal 3D printing is poised to be an enabler for tures and higher temperature capability with greater dura-
the next industrial revolution in enabling advancements in bility. The design and manufacture of turbo machinery
turbine engine performance. While there has been components can be accelerated by making use of topology
tremendous efforts on research and development in utiliz- optimization (TO) and additive manufacturing (AM).
ing this versatile technology, the number of qualified Lightweight structures with improved high temperature
metallic parts that are running in the engine has not been capability can be designed and manufactured by using an
commensurate with the applied research and development optimized design for additive manufacturing (DFAM)
efforts and the benefits that the AM technology offers. This approach to ensure suitable mechanical performances and
review addresses some of the key technical issues that are incorporate new system architectures. Additive manufac-
currently limiting the prolific usage of AM as a successful turing (or 3D printing) has the potential to enable engineers
vehicle for accelerated progress in gas turbine engines. to design components, systems and shapes, once thought
impossible to make. It allows for complex design geome-
Keywords Additive manufacturing Gas turbine engines tries that can potentially make products that are lighter,
Challenges Qualified parts Gas turbine components stronger and more efficient. The performance matrices that
Structural integrity Anisotropy AM can possibly offer compared to conventional (or sub-
tractive manufacturing) include: (a) design complexity,
(b) weight reduction, (c) part simplification, (d) inventory
1 Introduction reduction, (e) enhancement of efficiency (durability), and
(f) unitization and via all of the above, can perhaps
Aerospace propulsion innovation has been enabled by accelerate the development cycle for new and improved
advanced materials and manufacturing technologies for designs [5–14] and potentially decrease costs. By defini-
decades. Directionally solidified and single-crystal turbine tion, it can transform the design into a near-net shaped
blades, cast and wrought aero engine discs, fracture product, by slicing the computer-aided design (CAD) file,
typically a standard triangulation language (STL), STEP,
or Parasolid file, and translated to machine language so that
& Dheepa Srinivasan it can be manufactured in a layer-upon-layer process
dheepa.srinivasan@prattwhitney.com
(Fig. 1, [15]). AM can open new engineering capabilities to
1
Pratt & Whitney R&D Center, Indian Institute of Science, optimize part and system designs in a way that is less
Bangalore, Bangalore, India difficult than traditional manufacturing processes. The
123
Trans Indian Inst Met
ultimate goal for AM is to improve material capabilities, Several efforts have addressed the technology develop-
minimize design trade-offs, and facilitate digital vs. analog ment process in using 3D printing to make lightweight
control which would improve quality via reduced defect parts, as well as restore damaged parts on service returned
opportunities as compared to traditional manufacturing. turbo machinery hardware [30–41]. Starting with the T25
AM typically simplifies systems by means of an optimized compressor inlet temperature sensor casing by General
design and enables more robust designs, by reducing part Electric (GE), as the first part to be certified by the Federal
counts thru the elimination or decrease of brazed/welded, Aviation Administration (FAA) for use in a commercial jet
riveted/bolted joints and assemblies. The key attribute of engine, there have been several parts that are running in
AM that is particularly relevant and attractive for turbine several engines. Table 1 lists some of the recent applica-
engines is the potential equal or inverse relationship tions in the Aerospace industry, by Boeing, Airbus, Gen-
between complexity and cost. In other words, as the design eral Electric, Pratt and Whitney, Rolls Royce, Liebherr,
sophistication increases, the manufacturing cost is mar- etc. However, the rapid translation of the technology
ginally affected or actually decreases. This is attractive for development is still not readily adopted due to various
an innovative approach towards a paradigm shift to reduce challenges that arise operationally [15, 18] and technically
fuel burn and noise, by incorporating designs that are [9, 14, 42–44]. As stated previously, this review lists some
‘‘lighter’’ and ‘‘hotter’’, starting with a design thinking of the technical issues that are limiting the full benefits of
approach. It removes traditional design and manufacturing metal AM technology to produce gas turbine components.
constraints by its ability to create complex designs and These observations are mainly based on the practical
produce lightweight, yet strong structures via TO, or as an experience of the author through several varied AM pro-
enabler of better conformal cooling, which go to achieve duct qualifications, starting with ideation through proto-
higher temperature capability in turbo machinery [16–22]. typing to production, with an emphasis on products
The technology also provides potential supply chain flex- qualified from India in the last few years.
ibility. In many ways, AM could be a natural fit for The article is organized as follows. Section 2 lists the
empowering the gas turbine technology development for operational challenges in adopting additively manufactured
improvements to existing engine designs, as well as for components in the aerospace industry and the key technical
new engine developments. This can also enable a unique challenges that accompany part qualification. Section 3
solution in the aftermarket (services) sector related to describes the challenges pertaining to microstructure and
sustainability and novel repair capabilities. Some have mechanical properties, Sect. 4 includes challenges related
even eluded to the potential component life extension for to design and post-processing, and Sect. 5 comprises
difficult to weld materials, including directionally solidified inspection and qualification challenges, with a summary in
(DS) and single-crystal (SX) blades and vanes, by virtue of Sect. 6.
the rapid solidification process [23–29].
123
Trans Indian Inst Met
Table 1 Some of the publicly released additive component applications in gas turbines
S. No Company Name Component(s) Process(es) Model(s) References
1 General Electric Aviation Fuel nozzle tip PBF-LB/M GE9X [30, 31]
2 T25 sensor housing PBF-EB/M LEAP
3 Heat exchanger
4 Inducer
5 Stage 5/6 low pressure turbine (LPT) blades
6 Combustor Mixer
8 Pratt & Whitney Compressor Stators PBF-LB/M FT4000 [32]
9 Fuel system components PBF-EB/M PW4000
10 Brackets PW1500G
11 GKN Aerospace Fabricated Fan Case Mount Ring DED-LB/M Wire PW1500G [33]
12 Fan Spacer PW1900G
15 Rolls-Royce 1.5 m diameters titanium structure with 48 aerfoils PBF-EB/M Trent XWB-97 [34] [35]
16 Boeing Structural titanium part DED-PD/M 787 Dreamliner [36–38]
17 MTU Aero Engines Nickel borescope boss PBF-LB/M PW1100G-JM [39–41]
2 Operational and Technical Challenges in AM comprises nearly 40–50% of the overall usage [46]. Metal
Part Qualification AM can be categorized into further sub-categories whose
relative and typical attributes in terms of process speed,
The ASTM F42 committee via ISO/ASTM 52,900 has precision, and build volume are listed in Table 3. It should
defined seven process categories of additive manufactur- be noted that there are exceptions to this table for many
ing. These categories as shown in Table 2 include poly- applications and specific equipment. To date, cold spray is
mers, metals and multi-materials and the material forms not formally recognized by ASTM as an AM category.
they use such as powder, wire/filament, or sheet stock [45]. This is primarily due to the fact that typically most systems
Many alternative nomenclatures are in use for each of the are ‘‘cladding’’ type systems that do not create a 3D shape
AM techniques by various equipment manufacturers and from a CAD file. This may soon change as several systems
users and some that predate the adoption of the ASTM are now creating 3D shapes. The table also includes other
terminology. The relative usage of each of these techniques near-net shape manufacturing techniques such as metal
is illustrated in Fig. 2 which shows that metal AM usage injection moulding (MIM) which is truly a powder sinter-
has increased dramatically in the recent decade and ing and consolidation process. Amongst these, laser
123
Trans Indian Inst Met
Fig. 2 Relative usage of various additive manufacturing processes and Material types [46]
Powder bed Laser powder bed fusion Powder Laser Medium High Medium
Electron beam powder bed fusion Powder Electron Beam Medium High Medium
Deposition Directed energy deposition Powder Laser Medium Medium High
Cold spray* Powder Kinetic energy High Medium High
Electron beam melting Wire Electron beam High Low High
Consolidation Metal injection moulding** Powder Binder/Die High Medium Low
Post-consolidation
Binder jet Powder Binder Medium Medium Medium
Post-consolidation
powder bed fusion (PBF-LB/M), directed energy deposi- turbine blades, hybrid parts for repair and refurbishment,
tion (DED), and electron beam powder bed fusion (PBF- and possibly large near-net shapes while the PBF-LB/M
EB/M) have been used extensively for turbine engine technique is used for smaller parts with more complex
components. Table 4 shows a qualitative comparison of the geometries such as heat exchangers, turbine blades (with
three processes in terms of the energy source, build vol- internal cooling), vanes, shrouds, fuel nozzles, and others
umes, build layers, surface finish, residual stress, and with somewhat intricate geometry requirements. These
scanning method. Typically, both the DED and PBF-EB/M attributes vary based on the part geometry and material.
processes tend to be used to produce larger parts such as
Build envelope 250–500 mm (x/y) * 200 to 300 mm (rect. and round) Typically [ 1 m
Beam size 0.1–0.5 lm 0.2–1 lm Can vary from 2 to 4 lm
Layer thickness 50–100 lm 100 lm 500–1000 lm
Build rate \ 50 cc/h 55–80 cc/h 16–320 cc/h
Surface finish Ra 9–12 lm, Rz 35–40 lm Ra 25/35 lm Ra 20–50 lm, Rz 150–300 lm
Residual stress High Minimal High
Heat treatment Stress relief required HIP preferred Stress relief not required HIP preferred Stress relief required, HIP preferred
123
Trans Indian Inst Met
The operational challenges of AM that hinders wide- and solidification as they are deposited layer upon layer (as
spread quick adoption of this technology can be broadly shown in Fig. 4 [81]) and hence undergo a very complex
categorized as follows: thermal history that could lend to a very heterogeneous
microstructure. Because of the rapid solidification process,
1. Size limitations The current footprint of the AM
with cooling rates of the order of 105 to 106°/s, defects
machines is largely restricted to parts sizes below 500
inherent to the process can appear such as porosity or voids
mm3 volume for PBF-LB/M and PBF-EB/M. A whole
in the atomized powders, un-permissible oxygen in the
new range of machines are being developed by various
powders, gas porosity during laser processing, lack of
new manufacturers to address the size and productivity
fusion, solidification cracking, and unmelted powders
issue including developing hybrid machines, as well as
[81–88]. In all the metal AM techniques, the process could
exploring new technologies such as severe plastic
lend itself to a modified phase equilibrium, as well as,
deformation (cold spray) [61] and using alternate heat
preferential solidification texture, etc. [42]. All these
sources [63]. Table 5 lists some of the recent
microstructural variations will have a direct correlation
developments in metal additive machines [47–65].
with the mechanical properties that could pose a challenge
This is by no means an exhaustive list of all the latest
in developing standards. The next challenge is in design.
AM machine manufacturers.
One of the key performance matrices of AM technology is
2. Repeatability and Quality It is important to assess the
in design optimization to lead to topologically optimized
quality in each build layer via in situ monitoring of the
structures and novel unitization of components [17, 18].
molten pool. Computerized inspection using X-ray
While this is very attractive from being able to obtain very
tomography has proven to be of great benefit; however,
complex and efficient turbine engine parts, the translation
the cost and time involved, as well as, the applicability
to building some of these structures poses several chal-
to specific part geometries, does not enable it as a
lenges. Several of these challenges can be listed in terms of
routine quality control tool for AM parts. Many of the
achieving the dimensional tolerance and accuracy of the
current inspection technologies for metal AM systems
build, being able to avoid too many support structures,
only offer a defect detectability up to 5 lm.
removal of parts from the build plate, and addressing dis-
3. Scalability Inventory reduction is shown as a possible
tortions and warpage due to residual stresses [7]. The fin-
benefit of AM technology by not having to stock up
ished part surface roughness could pose a key challenge,
many spare parts. It is also potentially able to cater to
especially, when concerning internal channels, with very
quick changes in design, being a low-volume produc-
narrow features, where a rough surface could potentially
tion process. However, at present it is not capable of
constrict the gas flow and lead to local hot spots in the
coping with sudden demands in production typically
component. Subsequently, after the part is made success-
due to certification challenges.
fully, it must be taken through an inspection and qualifi-
4. Material availability Although several powder manu-
cation process where several challenges arise. Transferring
facturers are now available, there is still a huge gap in
the knowledge of successful prototyping trials to the scal-
terms of material availability. Getting the new chem-
ability of production parts on a build plate typically poses
istry to be reliable and amenable to the AM involves
several challenges. Challenges involved in inspection of
establishing the structure and properties and optimiz-
the parts, potentially by the way of loss of data in trans-
ing the parameters. This is usually a time-consuming
lating from the native CAD file to the STL file, could end
process and can be a deterrent to AM implementation.
up reflecting in non-conformities in the manufactured part.
In addition, AM is a ‘‘welding’’ process and many of
Inability to inspect curved regions or expensive inspection,
the aerospace alloys are typically difficult to weld.
inspection techniques not being able to detect microscopic
Table 6 lists some of the well-known commercial
defects, etc. contribute to the inability to translate from the
powder manufacturers who are qualified for AM
coupon level to full part qualification.
powders [66–77].
The technical challenges that arise in AM adoption can
be listed under three broad themes: A. Microstructure and 3 Challenges: Microstructure and Mechanical
Mechanical properties, B. Design Optimization and Post- Properties
processing, C. Qualification and Inspection. The complex
process that comprises most additive manufacturing, 3.1 Porosity and Micro-defects:
including managing the laser power/power density, tra-
verse speed, powder flow rate, deposit layer thickness, The starting point of any additive manufacturing process
substrate temperature, inert gas flow rate, etc. as shown in relates to being able to obtain dense ([ 99.5%) samples.
Fig. 3. Metal AM parts are subjected to repeated melting Several studies have documented how parameter
123
Table 5 Various metal AM machines in use/being developed [47–65]
S. Name of the Company name Process Additional Typical Location Build volume Remarks References
no. product/model process materials
123
mes description
1 M450 Meltio DED Metals Spain 200 9 150 9 450 mm Joint venture between Additec [47]
(Germany/USA)
and Sicnova (Spain) which
are both 3D printer
manufacturers
2 EZ300 Tada Electric PBF-EB/M Japan 250 9 250 9 300 mm [48]
Company
(Mitsubishi
Electric
group)
3 CERES system Exaddon DED in electrolyte bath Innovative Metal Switzerland 100 9 70 9 60 mm Opening new horizons in [49]
and energy source is additive micro- domains such as
electrochemical manufacturing microelectronics, MEMS and
(lAM) surface functionalization
4 LUMEX Avance Matsuura Hybrid PBF-LB/M and Japan 256 9 256 9 185 mm It is the world’s first 3D [50]
25 CNC machining Printing and a hybrid
industrial 3D printer. It can
build parts in 3D at a speed
of 7 cubic cm per hour, for
a maximum part weight of
90 kg
5 LSAM—Large- Thermwood Hybrid material Machines use a Reinforced US 3000 9 1500 9 1500/ Suitable for producing a wide [51]
Scale Additive extrusion and CNC two-step, near- thermoplastic 5100 mm variety of components,
Manufacturing machining net shape composite Thermwood is focusing on
production materials producing industrial tooling,
process. The masters, patterns, moulds,
part is first 3D and production fixtures for a
printed layer variety of industries
by layer, to including aerospace,
slightly larger automotive, foundry, and
than the final boating
size, and then,
it is trimmed to
its exact final
net size and
shape using a
CNC router
Trans Indian Inst Met
Table 5 continued
S. Name of the Company name Process Additional Typical Location Build volume Remarks References
no. product/model process materials
mes description
6 ATLAS GE Gantry PBF-LB Concept Laser Titanium, US 1.1 9 1.1 9 0.3 m (x, y, Build geometry will be [52]
aluminium z) customizable and scalable for
Trans Indian Inst Met
123
Table 5 continued
S. Name of the Company name Process Additional Typical Location Build volume Remarks References
no. product/model process materials
123
mes description
12 COMPOSER Anisoprint Material extrusion Material Composite Luxembourg 297 9 210 9 148 mm The second extruder can print [58]
extrusion 3D carbon fibre composite carbon fibre
printer filament filament (CCF) which
(CCF) enables you to reinforce the
part while printing. This way
you can manufacture load-
bearing structural parts with
high mechanical properties
13 HYBRID multi- Mazak DED (Hot Wire) Hot wire Metals (Repair US It combines additive [59]
tasking Corporation deposition of turbine technology, such as direct
machines and Oak blades and metal laser sintering and
Ridge other high- multi-laser HWD, with
National wear parts) subtractive manufacturing
Laboratory’s operations. The machine’s
(ORNL) hot wire laser incorporates an
automatic wire feeder system
that feeds welding wire to an
argon gas nozzle
14 FS621M Farsoon PBF-LB/M Large-frame laser Metal China 620 9 620 9 1100 mm Single 1000 W laser or four [60]
Technologies beam powder 500 W lasers
bed fusion
machine
15 High speed Essentium Material extrusion TPU, ABS, US 740 9 510 9 650 mm High speed extrusion (HSE) [61]
extrusion PEEK, and Flash Fuse technology
180-S Series ULTEM,
FlashFuse
z-materials
16 NeuBeam Wayland PBF-LB/M with PBF- New powder bed Refractory UK It is a hot part process rather [62]
process Additive EB/M fusion process metals than a hot bed process and
creates parts that are free of
residual stresses
17 Selective LED- Graz University PBF (LED energy Melting of Metals The new technology is quite [63]
based melting of source) powder using similar to SLM (Selective
(SLEDM) Technology LED light Laser Melting) and EBM
sources (Electron Beam Melting),
where the same metal
powder is melted by an
electron beam or a laser and
layered up to create the
desired finished product
Trans Indian Inst Met
Table 5 continued
S. Name of the Company name Process Additional Typical Location Build volume Remarks References
no. product/model process materials
mes description
18 Duo printer Creative 3D Material extrusion Dual extrusion Abrasive and US 910 9 460 9 675 mm The Duo printer has been [64]
technologies system flexible designed to encompass large
Trans Indian Inst Met
123
Trans Indian Inst Met
1 Sandvik, Osprey Sweden * Specialists in Maraging Steel of all kinds for AM—18Ni300, 13Ni400, HK30, etc. [66]
* Newly added Ti64 manufacturing in 2019
2 Carpenter—LPW USA-UK * Tantalum alloys, All other alloy grades [67]
3 Oerlikon USA-UK * VIGA—Ni, Fe, Co—Nitrogen and Argon [68]
* EIGA Ti grades 5 and 23—only Argon
4 Tosoh USA * EIGA process for new beta Ti chemistries [69]
* Ta-Ti powders—Patented process
5 Praxair USA * All grades [70]
6 Indo MIM India * First AM powder manufacturing from India [71]
7 Hoganas Germany All grades except Al [72]
8 Ametek USA All grades except Al [73]
9 MIMETE Italy All grades except Al [74]
10 Aubery and Duval France * Specialists in Nickel-based superalloy powders [75]
* Partnership with Pyrogenesi (Canada) for Ti powders for AM
11 H. C. Starck Germany All grades [76]
12 Valimet USA * Al alloys speciality [77]
13 Heraeus * Scalmaloy [78]
14 AP & C USA GE Additive company—TiAl [79]
15 6K USA Reclaiming from Subtractive Mfg [80]
optimization is done via a design of experiments (DOE), material. The energy density is calculated based on the
when starting with a new chemistry, to ensure that coupons formula, Ev = PL/(vs*hs*Ds), where Ev = Volume energy,
with minimal porosity are obtained [87–101]. The complex PL = Laser power, Ds = Layer thickness, Vs = Scan speed,
process shown in Fig. 4 is typically translated into an Hs = Hatch distance. Figure 5 shows an example of a large
energy density with the DOE aiming to have a dense difference in the porosity with different energy densities in
123
Trans Indian Inst Met
123
Trans Indian Inst Met
Fig. 5 DOE to optimize porosity in PBF-LB/M alloys in the ‘‘as-printed’’ condition, a, b CoCrMo [87], c, d MM247LC [88], for low and high
energy density, e, f porosity variations in Ti6Al4, for bulk (e) versus a thin wall implant (f) [99]
eliminate the pores and to prevent micro-crack nucleation whereas in several other AM samples, there was a creep
and the formation of a serrated grain boundaries, thereby debit. The non-equilibrium processes during the AM pro-
enabling better creep strength. This is definitely seen to be cesses sometimes led to precipitation of unexpected inter-
beneficial and the accepted practice for cast and wrought metallic and carbide phases during HIP process. Not all
alloys. However, for materials processed using AM, the alloys subjected to HIP and HT have responded with better
benefits tend to be less definitive and have in certain cases structural integrity in terms of better mechanical properties.
been found to have a somewhat limited effect [126–138]. Table 7 lists some of the effects of HIP HT studies in
In some cases, it was seen to improve the creep behaviour various alloys towards their mechanical properties as
123
Trans Indian Inst Met
123
Trans Indian Inst Met
Fig. 7 a, b Samples for mechanical testing showing different orientations on the build plate along with small-scale test coupons, c variation in
Vicker’s hardness and d yield strength (room temperature) with build orientation [86]
reported by various researchers. While HIP process can based on whether the test specimen is located near the
remediate some of the internal cracks, they are not very blower or the recoater, corner or centre, as indicated in
effective in closing some of the sub-surface or surface Fig. 7b. Figure 7c, d shows the variation in Vickers micro-
cracks [125]. In some alloys, heat treatment of the as- hardness and room temperature tensile strength as a func-
printed structure results in a recrystallized, fine-grained tion of build orientation for a wire arc AM build [86]. Kok
microstructure that lead to a poorer creep performance. et al [42] have carried out a comprehensive review on the
Figure 6 shows an example of recrystallized grains and a anisotropy of mechanical properties and heterogeneity of
loss in preferential texture after heat treatment [87]. The microstructure for different metal AM technologies. Maroti
heat treatment was optimized to result in the breakdown of et al. [139] studied the anisotropy by using static and
the as-printed structure and in an isotropic microstructure dynamic experiments. The differences in mechanical
in PBF-LB/M CoCrMo alloy. properties are attributed to the greater cooling time of the
individual layers along a particular orientation, hetero-
3.2.4 Mechanical Properties geneity in microstructure and phase formation, crystallo-
graphic orientation and porosity.
The factors contributing to the anisotropy of mechanical Carter et al [91] have documented the effect of aniso-
properties of AM materials are: a. build and crystallo- tropic gliding in different load directions to the build
graphic orientation, b. heterogeneity of microstructure and direction by a detailed analysis of the slip planes in small-
micro-defects, c. location specific residual stresses, and d. scale testing done using the SEM where it has be observed
preferential texture or orientation. Mechanical properties of that two sliding systems are activated along the build
additively built coupons are known to show variations as a direction (BD) and at 90° to the BD, whereas only one
function of build orientation (Fig. 7a [86]). Several reports sliding system can be seen activated at 45° to the BD.
have indicated the mechanical property anisotropy for Saboori et al. [146] have done systematic comparisons of
different turbine engine materials such as Ti6Al4V, IN718, the mechanical behaviour including high temperature
IN625, Hastealloy X, MM247LC, and AlSi10Mg fatigue and creep strengths of AM Ti6Al4V alloy that is
[139–156]. The anisotropy can tend to be location specific, used extensively in turbine engine compressor components.
123
Trans Indian Inst Met
Particularly in titanium alloys, a key issue is the role of which are comparable to the geometry of the part with
oxygen and its sensitivity to cooling rates, resulting in comparable microstructures and phase equilibria. There are
varied microstructure in different locations of the part and limited reports and practically few standards for testing at
hence mechanical properties, which may become geometry small length scales [88, 157–163]. Figure 8a shows an
specific. Understanding the variations in the room tem- example of micro-tensile specimen geometry vs regular
perature mechanical (tensile) properties may not be ade- tensile (ASTM E8) [163] which is particularly important to
quate if the application is for high temperature. For assess the mechanical behaviour related to heterogeneity in
example, there may be a poor fatigue at elevated temper- microstructure along a build, as shown in Fig. 8b [86]. This
atures or a creep debit or poor notch sensitivity at elevated is an area of research and development as there are cur-
temperatures that poses a big challenge to making use of rently no published, directly applicable standards available
the technology reliably for hot gas path components. for evaluating small-scale mechanical properties, espe-
Another important factor to be considered is the necessity cially for fatigue and creep, at elevated temperatures that
to evaluate mechanical properties at small length scales can truly represent the AM material behaviour.
123
Trans Indian Inst Met
123
Trans Indian Inst Met
The layer-upon-layer melting and solidification process element analysis of the AM process to predict the magni-
for most AM techniques lead to residual stresses that are tude and distribution of residual stresses in geometries of
tensile in nature [178–184]. Dimensional accuracy, dis- interest [180, 181]. Substrate heating is one way to over-
tortions, warpage, and shrinkage are some of the factors come residual stresses. An accurate understanding and
affected by residual stresses. Warpage occurs if the ther- modelling of thermal stresses and in situ heat transfer can
mally induced stresses exceed the substrate yield strength. facilitate optimal processing parameters and placement of
Disparities in residual stresses can sometimes cause sacrificial heat dissipation supports attached to a compo-
microstructural gradients. This becomes important espe- nent. The residual stresses can be significantly minimized
cially in the layers close to the build plate, where roughness by reducing the layer thickness, as well as, doubling the
coupled with residual stresses can lead to delamination at heat source; however, this may end up enhancing the part
the bottom surface of the unsupported melt pool making it distortion.
difficult for part removal. The curling of unsupported sur-
faces occurs when new layers on a build solidify and
contract upon cooling, warping previously deposited lay- 5 Challenges: Inspection and Qualification
ers. Residual stress measurements can be carried out using
X-ray diffraction, neutron diffraction, etc.; however, these Production appropriate inspection technologies for AM
are line of sight techniques, and it is a challenge to processes are somewhat underdeveloped or costly and is a
experimentally establish the residual stress, in different key challenge in going from the coupon level, to the part
locations of the part with precision. Some efforts have been level, and then full-scale production. In situ process mon-
made to perform transient thermo-mechanical finite itoring for layer-wise defects or partwise quality is critical,
123
Trans Indian Inst Met
Fig. 12 Technical challenges in additive manufacturing adoption for gas turbine components
Develop Machine, material, processes, including heat treatment, post-processing, inspection, structure–property correlations
Design Utilize the geometrical freedom of topology, post-processing treatments, support generation,
Qualify Establish structure–property correlations, small-scale testing suitable to design requirements, establish specifications, first piece
qualifications, and scaling up
Validate Performance of the new part and engine requirements, coupling with analytics and component testing
Certify Reproducible component level testing and engine level certification
ideally with precise closed-loop feedback systems, to drives extraneous and unimportant data to be collected and
[185–195] to ensure a quick turnaround time for obtaining possibly mislead outcomes. The information that the cur-
robust quality control. Sensors and advanced analytics are rent methods offer is typically not optimized and may not
being developed for anomaly detection. Figure 9a, b shows have supporting industry adopted specifications. Majumder
an example of thermal monitoring of the melt pool using [191] developed an in situ smart optical measurement
on-axis photodiodes, and optical detection of powder system (SOMS) to detect inconsistencies in composition,
spreading, which are under development. Even small detect pin holes, tiny cracks, porosity and segregation to
defects in fabricated parts can have a large impact on help achieve ‘‘certify as you build’’ to enable corrections
performance and safety, especially in turbine engines. Most for a better quality. A critical aspect of certification
of the current methods involve destructive testing to cap- requires a combination of non-destructive and destructive
ture fine details such as micro-scale cracks, porosity, and evaluation methods that account for process variabilities
small voids. Not only are these generally time-consuming that need to be carefully built into the specification for
and expensive, but may also lack the required resolution, qualification. Smart optical and thermal measurement
precision, and accuracy, for expected properties/component systems are being developed for in situ monitoring of AM
life. The lack of a layer-by-layer inspection process that is fabrication of 3D structures; however, the resolution of
robust and accurate is a challenge. In addition, the lack of a many of these techniques including micro-computer
full understanding of the key process inputs and outputs tomography is not yet ready to capture the micro-defects or
123
Trans Indian Inst Met
anomalies during the fabrication of the part. The current technology to turbine engine components. All of these need
methods we have just do not give us enough information to be considered as part of the qualification processes for
fast enough or cost effectively, to cover the geometric AM. Figure 12 depicts the list of technical challenges that
complexity of AM components and adds significant chal- were highlighted in this review. Although the challenges
lenges to accurate measurement [192–195]. One must not have been grouped under three different themes, the chal-
forget that the final inspection of the part is based on the lenges that relate the processing–structure–properties to the
part requirements, not necessarily the smallest defect that performance are inter-related and the cross-effects of these
can be created. The ability to define the manufacturing must be considered as an application moves towards part
process windows that establish material property design qualification. Being able to predict some of these via
curves is critical and enabling for part design application material and numerical modelling could serve to accelerate
and associated process control requirements. The non- the process. Table 8 lists an additive application strategy
availability of first principle models to optimize process that is essential in order to take the technology from a
parameters accurately to predict the layer-upon-layer concept, to feasibility, through the process development
heterogeneity is currently limiting the ability to reach mass and maturation stage, and ultimately accelerated adoption
production with ease. Figure 10a, b shows an example of a for turbine engine components. This process is comple-
physics-based defect prediction model which needs to be mentary to the industry accepted technology and manu-
integrated with the processing of the part. Figure 11 dis- facturing readiness planning. Comprehension of the
plays a flowchart for inspection and qualification and various challenges beforehand and at every stage of the
potential areas that need development. It is also important development will certainly assist in addressing them in a
to consider the thermal history of an AM component when scientific manner and accelerate the developmental cycle
designing for critical applications. For example, if ten and enable faster adoption of AM technology and to enable
tensile test specimens are fabricated one-by-one sequen- 3D printed parts to fly.
tially and removed from the machine when complete, the
material properties may be different when all ten speci- Acknowledgements The author would like to thank Dayananda
Narayana, Pratt & Whitney R&D Center, Bangalore, for several
mens are fabricated simultaneously without any post-heat discussions and help with the references and Jesse Boyer, Pratt &
treatment. In the former case, the component will be built Whitney, East Hartford, for reviewing the article and providing
comparatively quickly and so experience a higher mean valuable inputs.
temperature. In the latter case, the time interval between
layers for a given specimen will be greater due to the time
spent melting other specimens, and so, more thermal References
energy will dissipate between layers. Batch-wise variability
[1] Schafrik R, and Sprague R, Adv. Mater. Proc. 162 (2004).
of poorly controlled or understood processes has been seen [2] Roger C R, The Superalloys: Fundamentals and Applications,
in the actual part performance, which is difficult to detect Cambridge University Press (2006).
via microstructural or porosity variations. Because of many [3] Pollock T, and Tin S, J Prop Power 22 (2006) 361.
of the challenges discussed previously and some of the [4] Huron E S, Bain K R, Mourer D P, and Gabb T, Superalloys
(2008) 181.
historic bias created in early application shortcomings, [5] Dutta B, Babu S, and Jared B, Science, Technology and
process-based qualification and certification is yet to be Applications of Metals in Additive Manufacturing, Elsevier
readily and fully integrated. (2019).
[6] Wimpenny D I, Pandey P M, and Kumar L J, Advances in 3D
Printing and Additive Manufacturing Technologies. Springer
US, (2017).
6 Summary [7] Milewski J O, Additive Manufacturing of Metals: From Fun-
damental Technology to Rocket Nozzles, Medical Implants,
Metal additive manufacturing has offered some fascinating and Custom Jewelry, Springer, (2017).
[8] Gibson I, Rosen D W, and Stucker B, Development of Additive
paths to manufacture and realize the next generation Manufacturing Technology, Springer, (2015).
advancements to gas turbine engine designs and enable the [9] Simpson J, Haley J, Cramer C, Shafer O, Elliot A, and William
engineering of new alloys with superior performance at P, ORNL/TM-2019/1190 M3CT-19OR06090123, (2019).
short development cycle times and costs. However, more [10] Frazier W E, J. Mater. Eng. Perform, 23 (2014) 1917.
[11] Gao W, Zhang Y, Ramanujana D, Ramani K, Chenc Y, Wil-
can be realized and challenges still exist in controlling the liams C B , Wang C C L, Shina Y C, Zhang S, and Zavattieri P
unique microstructures, defects and properties that are D, Computer-Aided Design 69 (2015) 65.
created through this process to further ensure the delivery [12] Herzog D, Seyda V, Wycisk E, and Emmelmann C, Acta
of repeatable and reliable components. Several technical Mater. 117 (2016) 371.
challenges were discussed as being the key to accelerated
qualification of AM parts and faster adoption of the AM
123
Trans Indian Inst Met
123
Trans Indian Inst Met
[100] Srinivas, P, Srinivasan D, Dutta B, and Banerjee D, Unpub- [134] DebRoy T, Wei H L, Zuback J S, Elmer J W, Milewski J O,
lished work, Eurosuperalloys 2018. Besse A M, Heid A W, De A, and Zhang W, Prog. Mater. Sci.
[101] Harrison N J, Todd I, and Mumtaz K, Acta Mater, 94 (2015) 92 (2018) 112.
59. [135] 135. Helmer H E, Körner C, and Singer R F, J. Mater. Res. 29
[102] Carter L N, Essa K, and Attallah M M, Rapid Prototyping (2014) 1987.
Journal 21 (2015) 1. [136] Elmer J W, Vaja J, Carlton H D, and Pong R, Weld. J. 94
[103] Dutta B, and Froes F, The AMMTIAC Quarterly 6 (2011) 5. (2015) 313.
[104] Dinda G P, Dasgupta A K, and Mazumder J, Scripta Mater 67 [137] Seifi M, Salem A A, Satko D P, Grylls R, and Lewandowski J
(2012) 503. J, Proc. 9th Int. Symp. Superalloy 718 & Deriv. (2018) 515.
[105] Dinda G P, Dasgupta A K, and Mazumder J, Mat Sci Engg. A [138] Seifi M, Salem A A, Satko D P, Ackelid U, Semiatin S L, and
509 (2009) 98. Lewandowski J J, J Alloys Compd 729 (2017) 1118.
[106] Sames W J, Unocic K A, Dehoff R R, Lolla T, and Babu S S, [139] Maroti P, Varga P, Abraham H, Falk G, Zsebe T, Meiszterics
J. Mater. Res. 29 (2014) 1920. . Z, Mano S, Csernatony Z, Rendeki S, and Nyitrai M, Mater
[107] Wei H L, Mazumder J, and DebbRoy T, Nature Sci reports 5 Res Express 6 (2019) 035403.
(2015) 16446. [140] Nia M, Chena C, Wangb X, W P, Lia R, Zhanga X, and Zhoua
[108] Martin J H, Yahata B D, Hundley J M, and Mayer J A, Nature, K, Mater Sci Eng A 701 (2017) 344.
549 (2017) 365. [141] Tian Z, Zhang C, Wang D, Liu W, Fang X, Wellmann D, Zhao
[109] Tana Q , Zhang J, Suna Q, Fana Z, Li G, Yina Y, Liua Y, and Y, and Tian Y, Appl Sci, 10 (2020) 81.
Zhang M X, Acta Mater. (2020). [142] Kunze K, Etter T, Grässlin J, and Shklover V, Mater Sci Eng
[110] Smith T M, Thompson A C, Gabb T P, Bowman C L, and A, 620 (2014) 213.
Kantzos C A, Nature Research Scientific Reports (2020). [143] Etter T, Kunzer K, Geiger F, and Meidani H, IOP Conf. Series:
[111] Yanyan Z, Xiangjun T, Jia L, and Huaming W, Mater Des 67 Mater Sci Eng A 82 (2015) 012097.
(2015) 538. [144] Liu S, and Shin Y C, Mater Des. 164 (2019) 107552.
[112] Awd M, Stern F, Kampmann A, Kotzem D, Tenkamp J, and [145] Wang P, Nai M L S, Tan X, Sin W Jk, Tor S B, and Wei J,
Walther F, Metals 8 (2018) 825. Proc. TMS2016.
[113] Wilson-Heid A E, Wang Z, McCornac B, and Beese A M, [146] Saboori A, Gallo D, Biamino S, Fino P, and Lombardi M, Appl
Mater Sci Eng A 706 (2017) 287. Sci. 7 (2017) 883.
[114] Carroll B E, Palmer T A, and Beese A M, Acta Mater\ 87 [147] Carroll B E, Palmer T A, and Beese A M, Sci Dir Acta Mater
(2015) 309. 87 (2015) 309.
[115] Deng D, PhD Thesis, Linköping Univ. (2019). [148] Antonysamy A A, PhD Thesis, Univ. of Manchester (2012).
[116] Amato K N, Gaytan S M, Murr L E, Martinez Em Shindo P [149] Tomus D, Tian T, Rometsch P A, Heilmaier M, and Wu X,
Wm Hernandez J, Collins S, and Medina F, Acta Mater., 60 Mater Sci. Engg. A 667 (2016) 42.
(2012) 2229. [150] Kunze K, Etter T, Grasslin J, and Shkloyer V, Mat. Sci. Engg.
[117] Karlsson J, Snis A, Engqvist H, and Lausmaa J, J. Mater. A 620 (2015) 213.
Process. Technol. 213 (2013) 2109. [151] Divya V.D., Munroz-Moreno R, Messe O M D M, Barnard J S,
[118] Leuders S, Thone N, Riemer A, Niendorf T, Troster T, Richard Baker S, Illston T, and Stone H, J Mater Charac. 114 (2016)
H A, and Maier H J, Int. J. Fatigue 48 (2013) 300 62.
[119] Xu W, Brandt M, Sun S, Elambasseril J, Liu Q, Latham K, Xia [152] Helmer H E, Korner C, and Singer R F, J. Mater. Res. 29
K, and Qian M, Acta Mater. 85 (2015) 74. (2014) 1987.
[120] Murr L E, Martinez E, Amato K N, Gaytan S M, Hernandez J, [153] Korner C, Ramsperger M, Meid C, Burger D, Wollgramm P,
Ramirez D A, Shindo P W, Medina F, and Wicker R B, J. Bartsch M, and Eggeler G, Metall. Met Trans A 49 (2018)
Mater. Sci. Res. 1 (2012) 3. 3792.
[121] Deffley R J, PhD Thesis, Univ. of Sheffield, 2018. [154] Geiger F, Kunze K, and Etter T, Mater Sci Engg. A 661 (2016)
[122] Parimi L L, Ravi G A, Clark D, and Attallah M M, Mater 240.
Charac. 89 (2014) 102. [155] Muñoz-Moreno R, Divya D, Driver S L, Messe O M D, Illston
[123] Pinkerton A J, Karadge M, Syed W U H, and Li L, J Laser T, Baker S, Carpenter M A, and Stone H J, Mater Sci Engg A
Appl 18 (2006) 216. 674 (2016) 529.
[124] Brenne F, Taube A, Pröbstle M, Neumeier S, Schwarze D, [156] Sames W J, Unocic K A, Dehoff R R, Lolla Tapasvi, and Babu
Schaper M, and Niendorf T, Prog in Additive Mfg 1 (2016) S S, J. Mater. Res. 29 (2014) 1920.
141. [157] Howard C B, PhD Thesis, Univ. California, Berkeley (2018).
[125] Mathias S, Srinivasan D, Jayaprakash S, Ahmed S, and Ban- [158] Haghshenas M, Totuk O, Masoomi M, Thompson S M, and
erjee D, Submitted to Trans. INAE. Shamsaei N, Proc.Solid Freeform Fabrication 2017.
[126] Kuo Y L, Nagahari T, and Kakehi K, Maters 11 (2018) 996. [159] Benson LL, Marshall L A Benson, Weston N S, Mellor I, and
[127] Zhang F, Levine L E, Allen A J, Stoudt M R, Lindwall G, Lass Jackson M, Metall. Mater. Trans A 48 (2017) 5228.
E A, Williams M E, Idell Y, and Campbell C E, Acta Mater. [160] Muhammad M, BS. Thesis, Univ. North Dakota (2018).
2018; 152. [161] Alghamdi1 F, Verma D and Haghshenas M, Proc. Solid
[128] Anam Md Ashabul, PhD Thesis, Univ. of Louisville (2018). Freeform Fabrication 2017.
[129] Carter L N, Ph D Thesis, Univ. of Birmingham (2013). [162] Aboulkhair N T, Simonelli M, Parry L, Ashcroft I, Tuck C,
[130] Zhang B, Meng W J, Shao S, Phan N and Shamsaei N, Mat and Hague R, Prog. in Mater Sci 106 (2019) 100578.
Des Process Comm. (2019). [163] Reddy A S, and Srinivasan D, Proceedia Struct Integrity 14
[131] Benzing J, Hrabe N, Quinn T, White R, Rentz R and Ahlfors (2019) 449.
M, Mater Letters 257 (2019). [164] Junk S, Klerch B, Nasdala L, and Hochberg U, 28th CIRP Des.
[132] Qi H, Azer M, and Ritter A, Metall. Mater. Trans. A 40 (2009) Conf., May 2018, Nantes, France.
2410. [165] Orme M, Madera I, Gschweitl M, and Ferrari M, Designs 2
[133] Wu A, LeBlanc M M, Kumar M, Gallegos G F, Brown D W, (2018) 51.
and King W E, TMS annual meeting exhibition (2014).
123
Trans Indian Inst Met
[166] Allaire G, Dapogny C, Estevez R, Faure A and Michailidis G, [181] Pant P, PhD Thesis, Linköping Univ. Sweden (2020).
J. Comput. Phys. 351 (2017) 295. [182] Chimmat M and Srinivasan D, Procedia Struct Integrity 14
[167] Hällgrena S, Pejryd L, and Ekengren J, Procedia CIRP 50 (2019) 746.
(2016 ) 518. [183] Sundaram H, Srinivasan D, and Baummer J, Proc. of the
[168] Decker N, and Huang Q, Proc. ASME 2019 14th Int. Mfg. ASME (2019).
Sci.Engg. Conf. (2019). [184] Mercelis P, and Kruth J P, Rapid Prototyping Journal 12
[169] Speranza D, Citro D, Padula F, Moty B, Marcolin F, Calı̀ M, (2006) 254.
and Martorelli M, Appl. Bionics and Biomechanics, 2017, [185] Everton SK, Hirsch M, Stavroulakis P I, Leach R K, and Clare
Article ID 9701762. A T, Mater. Des. 95 (2016) 431.
[170] Klar V, Pere J, Turpeinen T, Kärki P, Orelma H, and Kuos- [186] Tapia G, and Elwany A, J. Manuf. Sci. Eng. 136 2014.
manen P, Sci. Rep. 9 (2019) 3822. [187] Berumen S, Bechmann F, Lindner S, Kruth J P, and Craeghs T,
[171] Jiménez M, Romero L, Dom-nguez I A, Espinosa M D M, and Phys. Procedia 5 (2010) 617.
Dom-nguez M, Hindawi Complexity, 2019, Article ID [188] Mani M, Lane B M, Donmez M A, Feng S C, and Moylan S P,
9656938. Int. J. Prod. Res. 55 (2017) 1400.
[172] Stimpson C, Snyder J, and Thole K A, J Turbo 138 (2016) [189] Koester L W, Taheri H, Bigelow T A, Collins P C, and Bond L
051008. J, Mater. Eval. 76 (2018) 386.
[173] Snyder J, and Thole K A, J Turbo 142 (2020) 051007. [190] DunbarA J, Denlinger E R, Heigel J, Michaleris P, Guerrier P,
[174] Klein E, Ling J, Aute V C, Hwang Y, and Radermacher R, Martukanitz R, and Simpson T W, Addit. Manuf. 12 (2016) 25.
Proc.17th Int Ref. and Air Cond. Conf. (2018). [191] Mazumder J, Procedia CIRP 36 (2015 ) 187.
[175] Ghani S A C, Zakaria1 M H, Harun W S W, and Zaulkafilai Z, [192] Salehi D, and Brandt M, Int. J. Adv. Manuf. Technol. 29 (2006)
MATEC Web of Conferences 90 (2017). 273.
[176] Tan C, Wang D, Ma W, Chen Y, Chen S, Yang Y, and Zhou [193] Davis T A, and Shin Y C, Mach. Vis. Appl. 22 (2011) 129.
K, Mater Des 196 (2020) 109147. [194] Rodriguez E, Mireles J, Terrazas C A, Espalin D, Perez M A,
[177] Snyder J C, PhD Thesis, The Pennsylvania State Univ. (2019). and Wicker R W, Addit. Manuf. 5 (2015) 3.
[178] Li C, Liu Z Y, Fang X Y, and Guo Y B, Procedia CIRP 71 [195] Bi G, Sun C N, and Gasser A, J. Mater. Process. Technol. 213
(2018) 348. (2013) 463.
[179] Mukherjee T, Zhang W, and DebRoy T, Compu Mater Sci 126
(2017) 360. Publisher’s Note Springer Nature remains neutral with regard to
[180] Ali H, Ghadbeigi H, and Mumtaz K, J Mater Engg Perf 27 jurisdictional claims in published maps and institutional affiliations.
(2018) 4059.
123