International Journal of Computer Integrated

Manufacturing

ISSN: 0951-192X (Print) 1362-3052 (Online) Journal homepage: https://www.tandfonline.com/loi/tcim20

Additive manufacturing–integrated hybrid

manufacturing and subtractive processes:
economic model and analysis

Guha Manogharan, Richard A Wysk & Ola L.A. Harrysson

To cite this article: Guha Manogharan, Richard A Wysk & Ola L.A. Harrysson (2016) Additive
manufacturing–integrated hybrid manufacturing and subtractive processes: economic model
and analysis, International Journal of Computer Integrated Manufacturing, 29:5, 473-488, DOI:
10.1080/0951192X.2015.1067920

To link to this article: https://doi.org/10.1080/0951192X.2015.1067920

Published online: 17 Nov 2015.

Submit your article to this journal

Article views: 1865

View related articles

View Crossmark data

Citing articles: 29 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tcim20
International Journal of Computer Integrated Manufacturing, 2016
Vol. 29, No. 5, 473–488, http://dx.doi.org/10.1080/0951192X.2015.1067920

Additive manufacturing–integrated hybrid manufacturing and subtractive processes: economic

model and analysis
a
Guha Manogharan *, Richard A Wyskb and Ola L.A. Harryssonb
a
Department of Mechanical and Industrial Engineering, Youngstown State University, Youngstown, OH, USA; bDepartment of Industrial
and Systems Engineering, North Carolina State University, Raleigh, NC, USA
(Received 24 September 2014; accepted 31 May 2015)

This article presents economic models for a new hybrid method where additive manufacturing (AM) and subtractive
methods (SMs) are integrated through composite process planning. Although AM and SM offer several unique
advantages, there are technological limitations such as tolerance and surface finish requirements; tooling and fixturing,
etc. that cannot be met by a single type of manufacturing. The intent of this article is not to show a new manufacturing
method, but rather to provide economic context to additive and subtractive methods as the best practice provides, and
look at the corresponding economics of each of those methods as a function of production batch size, machinability, cost
of the material, part geometry and tolerance requirements. Basic models of fixed and variable costs associated with
additive, subtractive and hybrid methods to produce parts are also presented. An experimental design is used to study the
influence of production volume, material and operating cost, batch size, machinability of the material and impact of
reducing AM processing time. A composite response model for the unit cost is computed for the various levels associated
with such engineering requirements. The developed models provide insight into how these variables affect the
costs associated with engineering a mechanical product that will be produced using AM and SM methods. From the
results, it appears that batch size, AM processing time and AM processing cost were the major cost factors. It was shown
that the cost of producing ‘near-net’ shape through SM and AM was the decision criteria; which will be critical for
tough-to-machine alloys and at multi-batch size.
Keywords: hybrid manufacturing; EBM; CNC-RP; AIMS; economic analysis and additive manufacturing

1. Introduction economic models for the additive-based hybrid process

Today, manufacturing industries struggle with long lead and the subtractive process are developed and analysed
times and high costs for critical low volume highly in this article.
customised parts. In some cases, like weapon system Direct digital manufacturing is the process of mak-
modernisation and rebuild, product renewal is required ing a physical product directly using a computer aided
(also in the case of some custom commercial products drafting (CAD) model. Direct digital manufacturing
such as biomedical implants), where the equipment is technologies have rapidly evolved over the past decade
old and spare parts are not readily available. Traditional and are gaining applications in aerospace, functional
processes such as forming, shaping and machining have components and biomedical implant manufacturing
part-specific processing tool requirements such as (Gibson, Rosen, and Stucker 2010; Frazier 2010).
moulds, dies, jigs and fixtures. Significant lead time Some of the appealing features of direct digital manu-
and fixed costs are associated with such tooling and facturing systems include; reduced part development
also, they need to be replaced for every individual part time, lower or no process engineering time, lower mate-
design. These problems are more prevalent for new rial consumption and faster design-to-part production
product development in low volume industries where (Chiu and Yu 2008; Czajkiewicz 2008). The layer-by-
long lead-times and tooling costs significantly increase layer principle of additive manufacturing has been
the unit cost. This need for customised tooling for employed for a variety of materials including plastics,
manufacturing components associated with each indivi- ceramics and metals. Several studies have classified and
dual product design (expensive dies, moulds and fix- compared different processes (Yan. and Gu. 1996; Pham
tures for each design) and qualification requirements for and Gault 1998; Horn and Harrysson 2012). More
new processing methods is challenging. In this article, a recently, the ability to directly manufacture metal com-
hybrid methodology that integrates additive and subtrac- ponents with complex part geometries without tooling
tive methods called additive methods integrated with (e.g. moulds, dies or fixtures) has been a topic of great
subtractive methods (AIMS) is presented. Production interest. The huge investment associated with tooling

*Corresponding author. Email: [email protected]

© 2015 Taylor & Francis

474 G. Manogharan et al.

typically necessitates the production of medium to large parts without expensive fixturing and planning, espe-
batch sizes to reduce the unit cost of components and cially, for growing applications in aerospace and biome-
hence, traditional manufacturing methods are preferred dical engineering, many of which are for very small
for mass production. A study has highlighted the advan- batch sizes. With growing demand for superalloys
tage of additive methods for production volume of one, which are often difficult to machine, this results in
when compared to conventional methods (Petrick and expensive tooling and poor material utilisation.
Simpson 2013).
In additive manufacturing, a stereolithography (STL)
file of the desired part is used to identify the layered 2-D 1.2. Hybrid processes
geometry required during processing. This includes iden- The success and widespread implementation of any man-
tifying and generating support structures for overhanging ufacturing process are based on its technical viability and
surfaces. By eliminating the need for part-dependent tool- economic feasibility. According to a recent Wohler’s
ing, additive manufacturing can facilitate economical pro- report, the market for AM including all products and
duction of batch sizes as low as a single unit. This services grew by 26% (compounded annual growth rate)
coincides with the increased interest in custom design to a total of $ 2.204 billion globally (Wohler’s Report
and shorter lead times (Silveira, Borenstein, and 2013). In particular, revenues associated with metal AM
Fogliatto 2001). In addition, metal-based additive pro- grew by 38.3% to a total of $ 24.9 million which indicates
cesses provide an alternative approach to conventional tremendous interest and potential for further improving
processes for the growing demand to process special-pur- end-metal AM products. The secondary market associated
pose alloys such as Inconel and titanium alloys which are with AM (tooling produced from AM products) grew by
tough to process through traditional methods. However, 10% to $1.19 billion in 2012 (Wohler’s Report 2013).
the current additive processes are only capable of produ- This secondary market will further benefit from shorter
cing components that are near net shape and typically lead times for part post-processing through modular fin-
require secondary operations to achieve the desired accu- ishing operation.
racy and surface finish. The project, cost analysis for additive manufacturing
during product lifestyle (CoA2MPLy) is focused
on estimating the benefits of metal AM through a life-
1.1. Subtractive processes cycle-based approach (Lindemann et al. 2012, 2013). It
Conventional subtractive manufacturing methods such was noted that material cost is one of the two largest
as milling are capable of achieving relatively higher contributors to the total cost of the part along with
precision tolerances and surface finish, but frequently operation cost. In metal-AM operation, the build time
require significant investment for custom fixturing. A (speed) and personnel cost were identified as the major
recent development in subtractive rapid prototyping is factors in the high cost, particularly, in the case of parts
the subtractive process of computer numeric control with shorter product lifecycle. One of the conclusions of
(CNC)-rapid prototyping (CNC-RP). Like additive pro- this article was that part design optimisation should be
cesses, the part is fabricated in layers, however, rather verified to lower part renewal. The final cost driver
than adding materials, CNC-RP removes materials layer noted by the project was that of data-preparation due
by layer using island milling and automatic tool path to higher labour costs since skilled and experienced
planning of a CAD/computer aided manufacturing engineers are required to ‘prepare’ building plans, par-
(CAM) system (Frank, Wysk, and Joshi 2004). Tool ticularly, in the case of a higher volume of smaller parts
paths are automatically generated from an STL file (Lindemann et al. 2013). A recent study analysed the
and the part is machined from a symmetrical bar stock AM supply chain in the production of spare parts and
supported between the centres of a rotary indexer. concluded that lower machine cost and a distributed
Sacrificial fixtures are added to the part design in the production system with decentralised production loca-
CNC-RP software, so that the part can be supported tions will lead to a more economical rapid manufactur-
within the cylindrical stock throughout the process of ing system (Khajavi, Partanen, and Holmstrom 2014)
machining and access to the face features can be which shows the potential of AM post-processing cen-
obtained. The CNC-RP process eliminates any manual tres near end-users. This study focused on spare parts
refixturing of the stock since the rotary indexer is used supply chains and noted that such supply chains for
for automatic repositioning (Frank, Wysk, and Joshi capital-intensive technologies like AM will result in
2004; Yang et al. 2009; Petrzelka and Frank 2010). lower overall operation costs, down time and higher
Unfortunately, both traditional CNC milling and CNC- flexibility.
RP machining can result in poor material utilisation due Prior studies have compared the cost of producing
to excess material loss through chips and scrap. This parts through traditional methods to AM, and it was con-
also presents a challenge to economic manufacturing of sistently noted that AM is suited for small batch
 International Journal of Computer Integrated Manufacturing 475

production. The cost is also impacted by AM machine and for some difficult to machine materials such as nickel
labour costs (Atzeni and Salmi 2012; Hopkinson and superalloys.
Dickens 2003). The significant impact of AM machine The common theme uniting the hybrid systems
cost was shown with respect to production volume (e.g. described so far is that the additive and subtractive opera-
injection moulding) and the cost of computational tools tions are carried out on the same piece of equipment and
required in metal AM was detailed (Hopkinson and hence, interference/gouge check is critical and making the
Dickens 2003). In the case of powder-bed processes, stu- processing planning and operating sequence very com-
dies have shown that part size and packing ratio are two plex. Significant process planning is required and the
critical factors in lowering unit cost (Ruffo, Tuck, and amount of time spent in tool change operations between
Hague 2006; Ruffo and Hague 2007), particularly, when depositions and machining is considerable. It is important
simultaneously producing different part designs (Ruffo, to develop a hybrid approach that can be integrated with
Tuck, and Hague 2007), and in some cases up to a 41% any additive manufacturing system without any or signifi-
reduction has been realised (Rickenbacher, Spierings, and cant modification to the existing equipment.
Wegener 2013). The existing economic models assume a
fixed amount of time for ‘post-processing’ which varies
from manual removal of support-structures to heat-treat- 2. The engineering model – novel hybrid process
ment (annealing) in laser powder-bed processes. The cur- In the following sections, we outline the operations of a
rent literature lacks an economic model which takes into hybrid direct manufacturing system that uses an additive
account the post-process planning, machining time, etc. manufacturing process (e.g. EBM) followed by the use of
based on the part geometry, machining allowance, CNC-RP to form a hybrid direct manufacturing process.
machinability, etc. Such an economic model will better However, it should be noted that the developed AIMS
reflect the impact of ‘batch’ metal AM-production or system can be implemented using any AM process.
mixed component production in powder-bed processes Based on the existing hybrid machines detailed earlier
on overall unit cost. (Section 1.2), we desire a hybrid system that would com-
Recently published work has reviewed hybrid pro- bine ‘freeform fabrication’ capability of AM processes
cesses and has classified them based on the principle of with minimal machining and limited process planning.
integration (Zhu et al. 2013). Much of the work in this Another desired attribute is to combine the existing AM
area has focused on directed energy metal deposition processes using a CNC machine tool with minimal mod-
processes such as wire welding using metal inert gas, ification (e.g. rotary indexer) as shown in Figure 1.
metal active gas (Akula and Karunakaran 2006;
Karunakaran et al. 2010; Xiong, Zhang, and Wang
2009) and laser melting due to the relative ease of 2.1. Additive manufacturing
integration (Jeng and Lin 2001; Amine, Sparks, and A commonly used class of additive manufacturing is
Liou 2011). These hybrid systems are formulated by powder bed fusion, which selectively focuses an energy
typically retrofitting 3-axis platforms in a CNC machin- source (laser or electron beam) on a bed of metal powder
ing centre by adding the deposition head into the (ASTM F2792 REV A; ASTM Designation 2012). There
machine volume. In such processes, hybrid manufactur- are several studies that detail the specifics of various
ing is achieved by alternating between additive and sub- commercially available powder bed fusion technologies
tractive methods after every few layers. Machining is (Mahale 2009; Kruth et al. 2005; Bremen, Meiners, and
performed after the deposition or formation of relatively Diatlov 2012). Since AIMS can be implemented with
thick layers followed by subsequent addition and subtrac- any of these metal AM processes, an overview of powder
tion steps until the final part is created. Other hybrid bed fusion processes is presented. Atomised metal pow-
processes employ additional rotary axe-based laser-aided ders are spread onto a build plate (also known as raking
deposition processes in which the deposition table is since a rake is used to spread the powders) and a focused
rotated to accommodate overhanging surfaces by depos- energy source is selectively applied based on the STL
iting materials from multiple directions followed by information. In the case of laser-based systems, the beam
machining (Liou et al. 2007). In most of the hybrid is controlled optically and the entire process takes place
approaches discussed here, the ability to withstand at room temperature. In the case of electron-beam melt-
machining forces depends on the surface area of the ing (EBM) system, the beam is controlled electro-mag-
part attached to the deposition plate. This would affect netically, and the operation takes place in a high vacuum.
the selection of build orientation since the part has to be In all powder-bed fusion processes, overhanging surfaces
oriented such that the largest cross-section is always require sacrificial support materials that are manually
formed in the first layer. Furthermore, the use of cutting removed after the AM process. Due to reduced density
fluids is somewhat limited in the hybrid systems and sintering conditions of the support, different process
described here. This can lead to increased tooling cost parameters are used (e.g. beam speed, scanning pattern,
476 G. Manogharan et al.

Figure 1. AIMS: Integrated AM and CNC-RP integrated flow illustration.

rate (MRR), depth of cut and cutting tool selection,

CNC-RP operations can be distinguished into the
following sequential stages: (1) roughing, where feature-
independent z-planar roughing operations with larger
depth of cut is used to reduce the cylinder into roughing
stock, i.e. area clearance operation, (2) semi-roughing,
where feature-dependent island milling is pursued based
on visibility analysis and reduces the part to ‘near-net’
dimensions and finally, (3) finishing, where machining
with a smaller cutting tool and small depth of cut is
pursued to produce the part to the required dimensions.
The stages are grouped into: (1) roughing, semi-roughing
and (2) finishing, where the geometric shape of the part is
enhanced progressively in each operation. The integration
Figure 2. EBM fabricated Ti6Al4 V biomedical and mechan-
of best practices of both additive (roughing through near-
ical components.
net AM) and subtractive (CNC-RP finishing) processes
could provide a solution to the rapid manufacturing of
beam power, etc.) for support and part volume. Figure 2 such complex part geometries using exotic alloys to the
shows some examples of biomedical and mechanical tolerances typically required for functional assembly
parts that have been produced using EBM. components.
Unlike, laser-based processes; EBM is a hot bed pro-
cess where the entire powder bed is maintained at elevated 2.3. Hybrid processes – EBM and CNC-RP
temperature throughout the course of the build. Initially,
after the spreading of a layer, the electron beam scans the In the AIMS hybrid system, the part is processed in two
entire bed known as ‘pre-heating’, followed by creating separate machines: (1) the near-net shape fabrication
part-specific features (‘melting’ and ‘supports’) and final through additive manufacturing (e.g. EBM or direct
scanning of the bed known as ‘post-heating’. These steps metal laser sintering (DMLS)), and then (2) a CNC-RP
are repeated onto subsequent layers until the final part is ‘finish’ machining where in a layered manner in a 4-axis
produced. In the case of laser-based systems, pre-heating CNC machining centre, the part is produced within speci-
and post-heating do not exist. fied tolerances (typically as small as 0.005 in;
Manogharan, Soundarajan, and Wysk 2011). The major
advantage of this approach is that no special tooling is
needed and minimal or no part programming is required
2.2. Subtractive manufacturing – CNC-RP which would result in elimination or reduction of fixed
CNC-RP uses a 4-axis CNC machine configuration to cost for adding a subtractive process. This approach
machine parts directly from bar stock. Based on the selec- focuses on complimenting the advantages associated
tion of machining parameters such as material removal with AM technologies along with already well-established
 International Journal of Computer Integrated Manufacturing 477

subtractive processes including CNC-RP. The hybrid man- Table 1. Nomenclature used for cost models for additive, sub-
ufacturing method shown in Figure 1 presents an inte- tractive and hybrid processes.
grated capability for the rapid manufacturing of complex Major notations Unit Comments
high precision components, in small batch sizes and with
minimal human intervention or engineering expertise. The General factors
processing environment in powder bed fusion processes Cunit $ Cost per unit
such as EBM or DMLS is not ideal to incorporate machine Pv mm3 Part volume
SPv mm3 Support volume-sacrificial supports
tools within the build volume. The part is then transferred Cprocess $/hr Operating cost for each process
to a separate CNC machine and machined using only the Cmat $ Cost of the material in each process
‘finishing’ step in subtractive CNC-RP. This approach tbuild hr Time to fabricate the part in the
expedites the hybrid production of multiple batches additive process
(depending on batch build time and only finish machining tsetup_process hr Setup time in each process
tpost_process hr Post-processing time in each
time) by increasing the availability and efficiency of both process
the machines. In other words, additive manufacturing of CNC-RP specific factors
batch ‘x’ and subtractive only finish machining of near-net Sv mm3 Volume of bar stock
shape units in batch ‘x – 1’ in CNC-RP can be simulta- thog hr Time for hogging operation
neously conducted. Since, relatively lower percentage of trough hr Time for roughing operation
tfinish hr Time for finishing operation
production time per unit part is spent on CNC-RP proces- ttool_life hr Cutting tool life duration
sing when compared to AM processing, the CNC machine ttool_change hr Time for changing tool and tool
can then be used to process multiple (or other) unit parts set-up time
when the AM machine processes the next batch. Such a
V mm 3
Total volume removed at each stage
system also ensures availability of the CNC machine for in CNC-RP
MRR mm3/hr Material removal rate at each stage
traditional subtractive machining including drilling, trap- in CNC-RP
ping of conventionally produced bar stock. With the sacri- Ctooling $/tool Cost of cutting tools
ficial location/orientation fixtures incorporated into the Nt – Number of tool changes in each
part, the STL file of the required part is ready for EBM stage
process planning. As of now, all the design modifications Ctooling $/tool Cost of cutting tools
EBM specific factors
and operations to the STL part file across the process are nEBM – Number of layers in EBM
manual (generation of CNC-RP fixtures, support struc- fabrication
tures for EBM, allowances and generation of toolpath). Ρ kg/mm3 Density of metal powder used
Further investigation into process characteristics of this tEBM hr Total build time in EBM
hybrid process based on material properties, part geometry tplate hr Time to pre-heat the start plate to
and part size can help in fully automating this process and required temperature before
fabrication
is ongoing. tcool hr Time to cool the build volume,
retrieve part and recycle unused
powder
3. Economic model
The following section describes the economic models
used to evaluate unit cost and production time as a func-
tion of part and support volume for the following systems:
However, for highly complex geometries, it is recom-
(1) Subtractive manufacturing – CNC-RP mended to include this cost. The cost model is material
(2) Hybrid manufacturing AIMS and activity-based, and the notation for all the major
(a) Additive manufacturing (e.g. EBM) cost factors is detailed in Table 1.
(b) CNC-RP (finishing)

Development of the cost model consists of material

3.1. Additive manufacturing cost model
and manufacturing cost (including setup and tool
costs) components. In this cost model, the energy For many additive manufacturing processes, unit cost can
costs of AM and CNC-RP are included in their respec- be simplified as a function of set-up time, part height (i.e.
tive manufacturing cost. The time for preparation of number of layers), summation of cross-sectional area of
process parameters in terms of NC codes (CNC-RP) each layer and post-processing time. For a part with n
and additive manufacturing process files are approxi- layers, where tbuild i is the build time for layers i = 1,
mately the same (and small typically < 60 minutes), 2,. . . n, the total manufacturing time and unit cost are
and the engineering cost is not included in the model. formulated as:
478 G. Manogharan et al.

Cunit ¼ Cmaterial þ ðCadd  tadd Þ; (1) 3.3. AIMS cost model

The developed hybrid process has two steps: AM and
X
n
‘finishing’ stage of CNC-RP. First, the cost model for
t add ¼ tsetup add þ ðtbuild  i Þ þ tpost process (2)
i¼1
AM is developed by integrating EBM-specific process
steps for the total build time in Equation (7).
The generic model presented in Equations (1) and (2)
can be adapted to any layer-by-layer manufacturing meth- X
n

ods depending on individual cost components in each ðtbuild i Þ ¼ tplate þ traking þ tpreheating þ tmelt i
i¼1
layer (e.g. different part and support generation para-
þ tsupport i þ tpostheating: (7)
meters) and the corresponding setup and post-processing
time such as annealing. In the following sections, this
model is extended to two specific systems: CNC-RP and In this process cost model, the total build time as
the AIMS hybrid process. shown in Equation (7) is further expanded to differentiate
individual operations in each layer ‘i’ namely; pre-heating
of the plate, raking of the metal powder, pre-heating the
powder bed, melting the contour (or edges) and part
3.2. CNC-RP cost model volume, support structures and finally, post-heating scan.
In the case of CNC-RP, the manufacturing cost is a Constant plate pre-heating, raking, pre-heating and post-
function of: cost of the stock, total machining time heating duration are assumed. For each layer ‘i’, the
and the tooling cost. The machining time is determined melting and support generation time are formulated
based on the total volume of metal to be removed from based on EBM process parameters as shown below in
the stock (ΔV) for a givenMRR, i.e. ΔV/MRR. Equations (8)–(9).
Furthermore, the tool geometry, feed rates and depth
!
of cut will vary in hogging, roughing and finishing Xn Xn  
Contour scan length Melt scan length
stages, leading to decreasing MRR. The setup time tmelt i ¼ þ ;
(tsetup_CNC-RP) is assumed to be uniform irrespective of i¼1
Contour speed i¼1
Melt speed
the part volume (because the stock is fixtured across (8)
two chucks). Furthermore, the material cost (Sv – stock
n  
volume) is based on part orientation (a cylinder with a
minimum diameter equal to the diagonal of the part). X Support scan length
tsupport i ¼ : (9)
The post-processing step in CNC-RP is the removal of i¼1
Support speed
sacrificial supports which takes negligible time and is
not considered in the cost model derived from
In the case of EBM, the post-processing involves cool-
Equations (3)–(6) as shown below:
ing down of the build chamber, part retrieval and recov-
ery/recycling of unused powder. Hence, the overall cost-
Cunit ¼ðSv  Cmat CNCRP Þ þ ðCCNCRP  tCNCRP Þ
(3) time of the EBM component in this hybrid process is
þ ðCtooling  nstage Þ; defined as:

tCNCRP ¼ tsetup CNCRP þ thog þ trough þ tfinish ; (4) X

n

  tEBM ¼ tsetup EBM þ ðtbuild i Þ þ tcool ; (10)

ð
V Þ 1 i¼1
nt ¼  ; (5)
MRR ttool life
C 0EBM ¼ ðη  ðPv þ SPv Þ  ρ  Ckg Þ þ ðCEBM  tEBM Þ:
ð
V Þ   (11)
tstage ¼ þ nt  ttool change : (6)
MRR
Unlike the complete CNC-RP, the subtractive stage of
At every ‘stage’ of CNC-RP (roughing, semi-rough- the hybrid system consists of only the finishing stage.
ing and finishing), there are two time components Since the sacrificial fixtures used in the subtractive stage
namely; machining time and tool change operations are considered as ‘EBM-made part volume’, there is no
Equations (5)–(6). Subsequently, the unit cost of the additional material cost associated with the subtractive
CNC-RP made part as shown in Equation (3) is derived stage. The ‘near net’ part from EBM is the stock volume
from the manufacturing time Equation (4), material cost and hence the cost-time CNC-RP component in this
and tooling cost. hybrid process is defined as:
 International Journal of Computer Integrated Manufacturing 479

CHybrid ¼ ðCCNCRP  ðtsetup CNCRP þ tfinish ÞÞ For the process variables in CNC-RP, it was assumed
 
þ Ctooling  nfinish þ C 0EBM Þ: (12) that the operating cost was $ 25/hour (Bureau of Labor
Statistics 2013) with a stock setup time of 10 minutes and
a tool change of 10 minutes including qualification of the
Currently, the cool-down time for EBM processing is a tool length. The cutting tools used were four flute carbide
significant part of the total EBM time about 20 hours, flat end mills with diameters of 25.40, 6.35 and 3.18 mm
Manogharan, Harrysson, and Wysk (2013). However, cur- for roughing, semi-roughing and finishing, respectively.
rently, work is being performed to reduce the cool-down The machining parameters were estimated for the surface
time in the case of EBM. This is considered in this study speed of 508 mm/s with a chip load of 0.05 mm in the
through sensitivity analysis of cost variables. case of roughing and semi-roughing and 0.03 mm in the
case of finishing. The layer thickness (depth of cut) con-
sidered were 5.08, 0.51 and 0.05 mm in the case of rough-
ing, semi-roughing and finishing operations, respectively.
3.4. Case study The stock volume for this study was a cylinder with a
The economic models developed are now analysed using a diameter of 63.50 mm and a length of 203.20 mm with a
case study of the part shown in Figures 3 and 4. The unit cost of $ 400 per bar. The average tool life was assumed to
cost and time are determined for manufacturing the case be 100 minutes of the machining time and the tooling cost
study part through CNC-RP and AIMS. Also, the impact of $20/tool.
of low-volume batch production is also investigated. The In the case of EBM, the operating cost was estimated
material selected for this study is Ti6Al4 V, which is one to be $ 104/hour (the charge for service work at NC State
of the more popular metal alloys processed through addi- University) and the layer thickness used during this study
tive methods. The part geometry selected in this study is was 0.07 mm (current parameter used for contract work at
typical of an additive manufacturing part that when pro- NC State University is 0.05 mm). The setup time in EBM
cessed through traditional machining would require multi- including lowering the pressure in the build chamber
ple refixturing and location qualification. down to the appropriate vacuum level and pre-heating

Figure 3. Unit and batch production of sample part in EBM with sacrificial supports for the subtractive stage.

Figure 4. (a) Schematics of EBM build slices and (b) Individual layer showing the support and melt features for the sample part in this
study.
480 G. Manogharan et al.

the EBM plate was assessed to be 90 minutes. The cost of batch production in the additive stage in some cases fol-
the EBM powder used was $ 300/kg and a 5% loss in lowed by ‘finishing’ operation of every ‘unit’ in the batch
powder during handling was also considered. Figure 4 separately in repeated CNC-RP operation.
shows the build slices (a) and several individual layers
(b). The individual layers illustrate the contour, support
and melt features. These component features will
change as a function of the build direction. Constant 3.5. Results
beam speed conditions throughout melting and support Using exclusively CNC-RP to produce the part through
generation are assumed in this study, although it is roughing, semi-roughing and finishing stages for machin-
recognised that in reality the EBM slightly adapts the ing parameters for Ti-6Al-4 V resulted in a processing
beam speed throughout the process. However, the pro- time of 23.42 hours and a unit cost of $1358.25 as
posed model can be used for varying processing con- shown in Table 2. The material chosen is a difficult to
ditions to replicate the exact physical build conditions machine material, so long machining times were not
across each layer. In this experiment, a multi-spot unexpected.
Arcam A2 was used and the effective contour speed This is representative for machinability of many of the
was 17.18 mm/s and the support speed was 50 mm/s. alloys used in aerospace applications such as Ti-6Al-4 V
In the case of melting, the beam overlap of 0.20 mm and superalloys such as Inconel 625. Such alloys have
was used and hence the effective beam travel distance desired high temperature strength when compared to other
per unit is defined by the total melt area per overlap. metals with superior machinability such as aluminium
Constant raking duration per layer (10 seconds), pre- alloys and most steels. This also leads to higher tool
heating (12.5 seconds) and post-heating duration wear resulting in the consumption of multiple cutting
(12.5 seconds) per layer are assumed. During this tools. In addition, the cost of stock for those alloys is
experiment, the melt beam speed used was 500 mm/s significantly higher than that of other commonly used
and the total number of layers was 528. The total material such as aluminium, steel and brass. When con-
cooling time for the build and part retrieval was esti- sidering the volume of parts to produce (larger batch
mated to be 1200 minutes. In the case study, the total production) using the CNC-RP process, the unit price
melt area of a unit part was calculated to be remains the same because the same process is repeated
152,255 mm2 and the total contour distance was for each part. Thus, there will be no affects with CNC-RP
found to be 21,605 mm. In the case of support struc- as batch size changes.
tures, the total support distance were calculated to be In the case of the additive stage of EBM in the
7766 mm. Also, based on the EBM wafer support and hybrid process, the unit price varies as shown in
given jaw-contact length (during CNC-RP), the total Table 3 for unit and batch production. It can be observed
volume for the sacrificial fixture in this study was that increasing batch size reduces the unit cost because
21,548.42 mm3. there is a significant fixed cost for each run. This can be
During batch production using CNC-RP, the process attributed to the significant amount of time required in
plan is repeated according to the batch size since stock is EBM for cool-down, plate pre-heating and part retrieval.
replaced after each run. In contrast, the number of parts in The material cost of alloys used in EBM (and other AM
a single build can be increased in additive processes sub- processes) is often greater than wrought stock due to the
ject to build envelope restrictions as shown in Figure 4. In preparation of materials through atomisation of the pow-
such batch runs using EBM, the processing time to set-up der. From prior studies, material cost for AM processing
the build plate, plate pre-heating, raking, layer pre-heating is typically one to two orders of magnitude higher than
and post-heating does not change (fixed costs per batch). polymer and metal material costs for most traditional
However, the layer processing time varies based on the manufacturing methods (injection moulding, casting,
area and perimeter of the cross-sectional geometry of each machining, etc.) (Manogharan, Harrysson, and Wysk
layer. Therefore, the hybrid system could benefit from 2013). The immediate implication here is that AM

Table 2. Breakdown of CNC-RP operation and tooling cost for sample parts.

Setup Machining Number of tool Tooling Tool change Stage Operating Total cost including
CNC-RP stage time (hrs) time (hrs) changes cost ($) time (hrs) time (hrs) cost ($) material ($)

Roughing and 0.17 14.04 8 160 1.33 15.37 417.50 1358.25

hogging
Finishing 9.38 5 100 0.83 10.21 276.00
Total CNC-RP 0.17 23.42 13 260 2.17 25.59 698.25
 International Journal of Computer Integrated Manufacturing 481

Table 3. EBM cost-components in unit and batch production for sample parts.

Batch Size Setup-plate time (hrs) Build time (hrs) Cool time (hrs) Total time (hrs) Material cost ($) Total cost ($) Unit cost ($)

1 1.5 6.53 14 22.03 104.23 2395.35 2395.35

2 1.5 12.48 16 29.98 138.97 3256.89 1628.45
3 1.5 18.43 18 37.93 185.30 4130.02 1376.67
4 1.5 24.38 20 45.88 247.07 5018.59 1254.65

Table 4. Hybrid process unit and batch production for sample parts.

Hybrid process

EBM stage CNC-RP

Batch size Time (hrs) Material cost ($) EBM cost ($) Time (hrs) CNC-RP cost ($) Total cost ($) Unit cost ($)

1 22.03 104.23 2395.35 10.55 383.75 2779.10 2779.10

2 29.98 138.97 3256.89 21.10 765.50 4024.39 2012.20
3 37.93 185.30 4130.02 31.65 1151.25 5281.27 1760.42
4 45.88 247.07 5018.59 42.20 1535.00 6553.59 1638.40

processing will be limited to smaller batch sizes because be further optimised with different parts of varying
of the higher material cost. batch sizes based on the build volume and part geo-
The hybrid process results as shown in Table 4 include metries. This could lower the unit cost for each of
the setup of the ‘near-net’ part in CNC-RP, finish machin- those hybrid parts. From Figure 5, it is recognised
ing and tool change time and costs. that the batch size does not impact the unit cost
For the given part geometry and available build when producing a part through CNC-RP (including
volume of EBM, CNC-RP was more economical for roughing and semi-roughing). From Table 3, it is evi-
batch sizes of up to four parts. Furthermore, the unit cost dent that the capability to fabricate in batches can
for the analysed part reduced by 41% with increased batch reduce the unit cost in the EBM stage in this example.
sizes d of four. This indicates that with greater build This is primarily due to the time involved in raking,
volume and/or nesting of parts into a single AM build pre-heating and post-heating in each layer and, impor-
with smaller part volume would further reduce the hybrid tantly, the cool down time as shown in Figure 6. By
cost. For instance, in this case study, the unit cost of batch increasing the number of parts that can be accommo-
size of five can be lowered by two runs of three and two dated in the AM build, the significant cooling time is
parts, respectively. This indicates that the build plan could spread across all units in the EBM build-batch.

Figure 5 Unit cost through batch production in CNC-RP and hybrid process AIMS for the sample part.
482 G. Manogharan et al.

Figure 6. Time components in the AIMS production based on batch size for sample part.

4. Sensitivity analysis - influence of cost components tools. It should be noted that the MRR ratio is with respect
The previous economic analysis was conducted for only a to the previous study. Hence, the lesser the ratio, the
single part geometry made from a hard to machine metal. greater the MRR when compared to the case study.
One could argue that for other part geometries where Additional analysis showed that 15% reduction in the
CNC-RP would make swarf of a large per cent of the volume of materials used for the sacrificial fixture used
initial stock, the economics of hybrid processing would be in AIMS resulted in the lowering of unit cost by 30%.
better. The same argument could also be made for other This shows that with further studies on ‘optimisation of
independent variable cost parameters. For instance, the sacrificial fixtures’ in AIMS (e.g. machining forces vs.
cost of materials for AM processing is 1–3 orders of fixture geometry), the overall unit cost of AIMS can be
magnitudes higher than that of traditional processes. further lowered.
Similar costs are present in the processing time and cost The total AM production time which is a function of
associated with AM when compared to traditional meth- pre-heat time, rake time, melt time and cooling or heat-
ods. Furthermore, with significantly higher fixed cost of treatment time is important, since advancements in AM
the equipment (an order of magnitude greater for metal- methods have been improving rapidly. Hence, it is
AM), the unit AM cost is much greater. While, combining important to analyse the effects of reducing this time
AM with CNC-RP in the AIMS system mitigates this component. The unit EBM time component reduces
effect to some extent, metal AM is often considered to based on batch size as shown in Figure 6. Therefore,
be an infant process for production application. This is a this sensitivity analysis of single unit batches includes
consideration to the newness of metal AM (a couple of multiple units in a single build where the unit EBM time
decades) when compared to other metal processing tech- is reduced. Also, the unit material cost is a function of
nologies which have existed for more than half a century. the part volume and hence does not vary for different
As with any developing processing technology, the cost- batch sizes. Furthermore, since AM is a relatively new
driving components will be improved with further technology when compared to machining, considerable
research and development. In order to identify those cri- training and experience are required in the workforce to
tical components for AIMS, a sensitivity analysis is con- operate the equipment. As a result, the operator costs are
ducted. The following sensitivity analysis studies the almost four times on average to that of a machinist.
major factors for the case study in order to evaluate the Finally, this analysis studies the impact of these variables
impact of the cost components. The variables include: cost on the unit cost for hybrid and CNC-RP production. It
of materials, AM operating cost, MRR ratio and AM should be noted that batch production is not considered
production time. for CNC-RP since the same process plan is repeated for
The cost of materials varies significantly in EBM (and each unit with a bar stock. The analysis was performed
other AM) based on the alloys such as Ti-6Al-4 V, on the same part design detailed in Section 3.4 for the
Inconel, etc. and also, the method and quantity of produc- following conditions shown in Table 5 with optimistic
tion through atomisation. In the case of CNC-RP, the cost conditions (50% improvement) and most optimistic con-
of machining stock also varies significantly based on the ditions (100%) in the cost factors.
materials such as aluminium and Ti-6Al-4 V. The price of atomised Ti-6Al-4 V metal powder for
Machinability is an important factor for the unit cost, EBM is approximately $ 300/kg, and it is expected that
since selection of machining parameters can vary based with gaining popularity of metal AM, the overall produc-
on the materials (aluminium, Ti-6Al-4 V), available cut- tion volume will increase and hence the powder cost
ting tools (high speed steel, coated carbides) and machine would be lowered. It should be noted that the cost of the
 International Journal of Computer Integrated Manufacturing 483

Table 5. Variable conditions for AIMS hybrid process. had lesser effect on the unit cost where 50% reduction in
material cost resulted only in 8% reduction in unit cost.
Current Most-
Variables level Optimistic optimistic Furthermore, increasing the material removal rate
(machinability) of the material during finish-machining
EBM material cost ($/kg) 300 150 30 by 100% reduced the unit cost by 10%. This can be
CNC-RP stock cost ($/unit) 400 200 40 attributed to the fabrication of ‘rough stock’ through AM
Ratio: MRR/MRR-case study 1 0.5 0.1 leading to minimal machining.
EBM production time 20 10 2
(hrs/part) Increasing the machinability (MRR) by 50% and
EBM operating cost ($/hrs) 104 66 33 100% in CNC-RP positively influenced the CNC-RP
unit cost by 35% and 52%, respectively, as shown in
Figures 7 and 8. However, reducing the material cost by
similar orders had a lesser effect on the unit cost (13% and
metal powder varies based on the material and process 25%), and when compared to the case study, this can be
selection (laser vs. EBM). The cost of round stock for attributed to the significant amount of machining time
machining is higher for superalloys such as Ti-6Al-4 V leading to high tool costs. In other words, machining
and lower for steel and lowest for aluminium. Hence, the time is more dominant on unit cost than the material cost
CNC-RP stock cost is considered to be 50% and 10% of because increased machining time results in higher pro-
Ti-6Al-4 V cost to simulate the effects of processing steel duction time as well as more tool wear. The values for cost
and aluminium. Similarly, the machinability of steel and and rates used in the case study were typical costs for the
aluminium is of the same order in terms of MRR. With effect of these rates should illustrate typical responses for
improvements to current processing capabilities and tech- changes in the manufacturing efficiencies and material
niques (e.g. cool down external to the EBM machine), the costs.
production time would be reduced. Finally, with growing It can be identified from this analysis that the cost of
workforce and skilled operators, the operating cost of the hybrid process can be greatly reduced by reducing the
metal AM such as EBM would reduce in the future. production time and operator cost in EBM (or AM). In the
The influence of the major factors on hybrid unit cost case of CNC-RP, the unit cost can be greatly reduced by
is shown in Figure 7. In Figure 7, it was observed that increasing the material removal rate significantly and
EBM (AM) production time tremendously affects the unit lower the starting material cost. This can be attributed to
hybrid cost along with production cost. It was found that a the amount of additional machining time required to
50% reduction in EBM production time resulted in low- machine alloys such as Ti-6Al-4 V (MRR ratio = 1) and
ering the unit cost by 40% and when EBM production brass (MRR ratio > 0.33). In many ways, this study
time was reduced to 90%, the overall unit cost was reflects the current state of additive and machining-based
reduced by 72%. Similarly, a 36% reduction in operation manufacturing processes, where expensive alloys with part
cost reduced the unit cost by about 30%. For the analysed designs requiring multiple fixtures are preferred to be
part design and volume, it was found that the material cost processed through AM (particularly, for low volume

Figure 7. Influences of cost-factors on single-batch production through hybrid AIMS process for sample part.
484 G. Manogharan et al.

Figure 8. Influences of cost-factors on single-batch production through CNC-RP for sample part.

batches). The mass production of easier-to-machine alloys by broadening the AM-experienced talent pool through
like aluminium, brass, etc. can be easily processed through appropriate workforce training and education. So far, the
machining with a higher material removal rate. From this noted findings have been for a single-batch production as
study, it was analysed that EBM production time (which shown in Figure 9.
includes cool-down time) and EBM operator cost are the It can be identified that the production cost signifi-
major levers for lowering the hybrid unit cost. In the case cantly affects the hybrid unit cost and its effects are greater
of AIMS, the analysis above identified the interaction at larger production time (in the case of EBM→ cooling
between the important variables on unit cost for the case time). The impact of higher EBM production cost and
study for single-unit and batch production. The number of time is mitigated by increasing the batch size. This also
parts that can be included in a single EBM build (and shows that by ‘nesting’ multiple unique parts (of single-
other AM processes) is limited by the available AM build batch), the overall unit cost of each part can also be
volume and build orientation. Furthermore, based on the lowered. For instance, combining single-order parts of
case study, it was identified that the batch size does not different geometries during the AM stage would lower
impact the CNC-RP unit cost. Hence, in this analysis, the the overall unit cost of each part.
batch size of one that can be accommodated along the
shortest build height is only considered for CNC-RP.
Based on findings from the first order influence of 5. Discussion
independent variables (Figure 7), the interaction between From this study, it is observed that the machining duration
EBM production time and cost appears to be critical. In of CNC-RP only production (including roughing and
order to compare AIMS to CNC-RP, that interaction is semi-roughing) is significantly longer than AIMS (only
extended to single-batch production and is presented in the finishing stage of CNC-RP). This is attributed to the
Figure 8. It was found that the unit cost of AIMS was machining parameters such as feed and depth of cut
lesser than that of base-study CNC-RP ($1358.25), when employed for milling alloys such as Ti-6Al-4 V. For
either the production time was lowered to 2 hours or when instance, the machining time would be drastically lower
the production cost was $33/hour and also, when the in the case of processing relatively softer materials such as
production time was lowered to 10 hours (for $66/hour). aluminium or brass. However, since the operation cost of
Based on the economic information provided, it appears CNC-RP is much lower than the operation cost of rela-
that reducing the total time on the AM machine is a tively newer technologies such as EBM (and other metal
critical variable. If the time can be reduced by 75%, the AM), the CNC-RP unit cost is lower for a single unit
economics of AM only or hybrid processing will improve when compared to AIMS. It was also observed that long
significantly. This is also a current focus of the next cooling time (and/or heat treatments) has a major influence
generation of AM (e.g. EBM machines), where post-AM on the unit hybrid cost. However, material utilisation in
processing time including cool-down/annealing time will terms of part-stock volume of expensive, tough to
go down significantly. Also of significance is lowering the machine materials is a critical factor based on part geo-
production cost of AM (e.g. EBM). This can be achieved metry while processing solely in CNC-RP. This is
 International Journal of Computer Integrated Manufacturing 485

Figure 9. Influences of EBM production time and cost on single-batch production through AIMS.

important in the case of expensive superalloys used in and if:

aerospace and mechanical applications.
The hybrid process demonstrated higher material uti- CRough stock ¼ðCmachining  tprefinishing Þ þ CtoolðH Þ
lisation of expensive alloys in EBM along with the ability (14)
þ CtoolðRÞ þ CCNCRPstock ;
to ‘finish’ machine through CNC-RP. If machining can be
limited to only the critically tolerance surfaces as shown in
Figure 10, the CNC-RP time will be reduced significantly. and:
It can also be noted that the ability to batch produce CRough stock  CEBM : (15)
‘near-net’ shaped parts in a single build lowered the EBM-
component in the total cost. Another critical factor is the Hence, it is considered to be optimal to choose the
part design limitations with subtractive methods. For CNC-RP process over the hybrid process (neglecting any
instance, it is not feasible to produce non-conventional geometric constraints) when the cost of rough stock is
features such as mesh or lattice structures through CNC- lesser than that of EBM. Otherwise, the hybrid process
RP. By integrating such part design within surfaces as provides the least costly solution. Any situation that
shown in Figure 10 and finish machining through CNC- increases the pre-finishing time i.e. costs of roughing or
RP, the hybrid process enables an increased strength-to- semi-roughing tooling, or stock costs will shift the deci-
weight ratio along with precision feature and surface sion in the direction of the hybrid process and ‘near-net’
accuracy. This is particularly critical in the case of expen- shaping using additive processes. This is intuitive and
sive alloys. explains why significant research thrusts in the additive
In addition, the following relationship can be derived manufacturing arena are towards nickel-based superalloys
from above: and other difficult to machine materials. It is also some-
what instructive as to why (again neglecting any geo-
metric constraints) aluminium research in these AM
tPrefinishing ¼ tsetupCNCRP þ thog þ trough ; (13)
processes is somewhat muted, as the machinability of

Figure 10. Functional surfaces and CNC-RP finish machining.

486 G. Manogharan et al.

aluminium is much greater than that of cobalt chrome and and subtractive processes; in this case EBM and rapid
titanium alloys. Other situations may not be as intuitive. CNC machining (CNC-RP).
Parts that require relatively little roughing or semi-
roughing time, such as parts where part volume
approaches that of the rough stock size in CNC-RP will 6. Conclusion
favour CNC-RP, and parts where the majority of the mate- This article presented a model to determine the economics
rial removal is done by the relatively large and rigid rough- of additive and subtractive processing of mechanical parts.
ing tool will again favour CNC-RP. As shown, parts with a It also identified the critical cost components using sensi-
lower aspect ratio (with a lower ratio of material removal to tivity analysis by varying the variables of the cost models.
stock volume) are preferred to be processed through CNC- Through integration of additive manufacturing processes
RP. Parts with a higher aspect ratio and thin walls, greater with a subtractive manufacturing-based CNC-RP, the
concave areas and small internal features (internal to the inherent economic advantages were demonstrated for
convex hull) will favour the hybrid process. each individual system through near-net shape part pro-
The impact of the batch size on the hybrid process is duction, enabling the processing of otherwise difficult to
significant because, it is near-net shaping ‘n’ units of machine materials, geometric flexibility, and rapid deploy-
rough stock for finish CNC-RP with a single set-up in ment of additive manufacturing. Such a hybrid approach
EBM and amortising processing time of raking, pre-heat- results in simplified fixturing and reduced process plan-
ing and post-heating of each layer (which consists of ‘n’ ning for subtractive manufacturing. The proposed hybrid
units). Hence, the selection criteria of economical process system was demonstrated through a case study of a single
can further extended as: part using EBM and CNC-RP where the part is a func-
tional load-bearing assembly part made of difficult to
machine titanium alloy, Ti-6Al-4 V. Also presented is an
CRough stock  ðCEBM Þ=n: (16)
economic model of the AIMS system. The need for lower
AM production time (50% reduction leads to 40% lower
If the cost of roughing and semi-roughing through unit cost) and cost (30% lower unit cost at 36% lower
CNC-RP solely for a single unit is greater than processing operating cost) were identified as the critical cost compo-
‘near-net’ shape EBM-made rough stock in a batch, AIMS nent. This is not a surprising finding since materials cost
is more economical. This criterion should be considered for additive materials are close to two orders of magnitude
simultaneously for: (1) batch size, (2) build orientation greater than the materials costs for materials used in tradi-
and (3) part shape. For example, if a larger batch size tional processes, and processing time for additive manu-
(~10) of a part with a higher aspect ratio is required, it facturing is also close to two orders of magnitude greater
would be efficient to select the build orientation along the than for traditional processes like injection moulding
part length to accommodate the batch size in a single AM Also, it was shown that the decision to use CNC-
build. For commercial AM production, if even one part is RP versus the hybrid system is impacted by the
being manufactured in an AM, it could be paired with machinability and material cost, along with geometric
other parts in a single AM build so that n in Equation (16) and size considerations of the part. The hybrid
will be kept as high as possible and the resulting cost will approach is more economically attractive for more
be as low as possible. expensive and harder to machine materials while
The ability to address fixturing for subtractive opera- CNC-RP is favoured for less expensive and easier to
tions prior to fabricating the part provides a unique machine materials (10% improvement in unit cost with
advantage to analyse the location and geometry of fix- 100% increased machinability). The hybrid system also
tures on non-functional or desired surfaces, the orienta- becomes less expensive when multiple parts can be
tion of part fabrication in the additive process can also be produced in the AM process simultaneously (either
adapted to surfaces requiring precision finish (e.g. lot size increases beyond a single unit or pairing with
upward facing surfaces have better finish than downward other parts or orders in a single EBM build) subject to
facing or overhanging surface). Successful implementa- build volume-orientation constraints. It was found that
tion of the integrated hybrid system will improve mate- AM processing time (EBM cooling time, heat-treat-
rial utilisation and eliminate manual finishing processes ment of other powder-bed AM processes) and AM
and the requirement for multiple fixtures. The AIMS manufacturing cost greatly affect the cost of hybrid
system can be employed through any AM process such processing.
as selective laser sintering (SLS), selective laser melting This sequential hybrid approach utilising built-in
(SLM), etc. One of the highlights of this system is the sacrificial fixtures can be adapted with any existing
requirement for a single STL/CAD to generate process AM techniques including powder-bed fusion processes
plan for the entire hybrid system. Such approach can be such as EBM, DMLS and binder jetting processes
employed in combining advantages of similar additive where the processing environment is not suited for
 International Journal of Computer Integrated Manufacturing 487

simultaneous subtractive machining. This hybrid manu- Czajkiewicz, Z. J. 2008. “Direct Digital Manufacturing, New
facturing system has not been previously demonstrated. Product Development and Production Technology.”
Economics and Organization of Enterprise 2: 29–37.
It can be used to process a wide variety of geometries
doi:10.2478/v10061-008-0016-8.
and materials while reducing engineering and material Frank, M. C., R. A. Wysk, and S. Joshi. 2004. “Rapid Planning
costs related to fixtures and tooling. Furthermore, this for CNC Milling-A new Approach for Rapid Prototyping.”
system can be readily implemented with no modifica- Journal of Manufacturing Systems 23 (3): 242–255.
tions to their currently deployed systems and only doi:10.1016/S0278-6125(04)80037-2.
Frazier, W. E. 2010. “Direct Digital Manufacturing of Metallic
requires a rotary indexing CNC capability. Future
Components: Vision and Roadman.” In 21sd Solid Freeform
work for this study will focus on analysing costs for a Fabrication Symposium. Austin: University of Texas.
wide range of materials, AM technologies (SLM, Gibson, I., D. Rosen, and B. Stucker. 2010. “Direct Digital
Binder-Jetting), part designs (conventional prismatic Manufacturing.” Chap. 14 in Additive Manufacturing
parts, AM-friendly lattice structures, etc.). In addition, Technologies: Rapid Prototyping to Direct Digital
Manufacturing. New York, NY: Springer.
a detailed energy-consumption factor will be included in
Hopkinson, N., and P. Dickens. 2003. “Analysis of Rapid
additional studies to better understand the sustainability Manufacturing-Using Layer Manufacturing Processes for
and scalability of this hybrid approach. Production.” Proceedings of the Institution of Mechanical
Engineers, Part C: Journal of Mechanical Engineering
Science 217: 31–39.
Disclosure statement Horn, T. J., and O. L. A. Harrysson. 2012. “Overview of Current
Additive Manufacturing Technologies and Selected
No potential conflict of interest was reported by the authors.
Applications.” Science Progress 95 (3): 255–282.
doi:10.3184/003685012X13420984463047.
Jeng, J.-Y., and M.-C. Lin. 2001. “Mold Fabrication and
Funding Modification Using Hybrid Processes of Selective Laser
This work was supported by the National Science Foundation Cladding and Milling.” Journal of Materials Processing
[grant number CMMI −1161926]. Technology 110 (1): 98–103. doi:10.1016/S0924-0136(00)
00850-5.
Karunakaran, K. P., S. Suryakumar, V. Pushpa, and S. Akula.
2010. “Low Cost Integration of Additive and Subtractive
ORCID
Processes for Hybrid Layered Manufacturing.” Robotics
Guha Manogharan http://orcid.org/0000-0002-9756-1220 and Computer-Integrated Manufacturing 26 (5): 490–499.
doi:10.1016/j.rcim.2010.03.008.
Khajavi, S. H., J. Partanen, and J. Holmstrom. 2014. “Additive
Manufacturing in the Spare Parts Supply Chain.” Computers
References in Industry 65 (1): 50–63. doi:10.1016/j.compind.2013.07.008.
Akula, S., and K. P. Karunakaran. 2006. “Hybrid Adaptive Kruth, J.-P., P. Mercelis, J. V. Vaerenbergh, L. Froyen, and M.
Layer Manufacturing: An Intelligent Art of Direct Metal Rombouts. 2005. “Binding Mechanisms in Selective Laser
Rapid Tooling Process.” Robotics and Computer Sintering and Selective Laser Melting.” Rapid Prototyping
Integrated Manufacturing 22 (2): 113–123. doi:10.1016/j. Journal 11 (1): 26–36. doi:10.1108/13552540510573365.
rcim.2005.02.006. Lindemann, C., U. Jahnke, M. Moi, and R. Koch. 2012.
Amine, T. A., T. E. Sparks, and F. Liou. 2011. “A Strategy for “Analyzing Product Lifecycle Costs for a Better
Fabricating Complex Structures via a Hybrid Manufacturing Understanding of Cost Drivers in Additive Manufacturing.”
Process.” In 22nd Solid Freeform Fabrication Symposium. In Proceedings of the Twenty-third Annual International Solid
Austin: University of Texas. Freeform Fabrication Symposium, 177–188. Austin:
ASTM Designation. 2012. Standard Terminology for Additive University of Texas.
Manufacturing Technologies, 1–3. Conshohocken, PA: Lindemann, C., U. Jahnke, M. Moi, and R. Koch. 2013. “Impact
ASTM International, West. F2792-12a. www.astm.org and Influence Factors of Additive Manufacturing on Product
Atzeni, E., and A. Salmi. 2012. “Economics of Additive Lifecycle Costs.” In Proceedings of the Twenty-fourth
Manufacturing for End-Usable Metal Parts.” The International Annual International Solid Freeform Fabrication
Journal of Advanced Manufacturing Technology 62 (9–12): Symposium, 998–1008. Austin: University of Texas.
1147–1155. doi:10.1007/s00170-011-3878-1. Liou, F., K. Slattery, M. Kinsella, J. Newkirk, H. Chou, and R.
Bremen, S., W. Meiners, and A. Diatlov. 2012. “Selective Laser Landers. 2007. “Applications of a Hybrid Manufacturing
Melting: A Manufacturing Technology for the Future?” Process for Fabrication of Metallic Structures.” Rapid
Laser Technik Journal 9 (2): 33–38. doi:10.1002/latj.v9.2. Prototyping Journal 13 (4): 236–244. doi:10.1108/
Bureau of Labor Statistics. 2013. Occupational Employment and 13552540710776188.
Wages, 51-4041 Machinist. Accessed 15 July 2015. http:// Mahale, T. R. 2009. “Electron Beam Melting of Advanced
www.bls.gov/oes/current/oes514041.htm Materials and Structures.” Published Phd thesis, North
Chiu, W. K., and K. M. Yu. 2008. “Direct Digital Manufacturing Carolina State University, Raleigh, NC.
of Three-Dimensional Functionally Graded Material Manogharan, G. P., O. L. A. Harrysson, and R. A. Wysk. 2013.
Objects.” Computer-Aided Design 40 (12): 1080–1093. Challenges and Benefits of Hybrid Additive and Subtractive
doi:10.1016/j.cad.2008.10.002. Manufacturing. San Juan, Puerto Rico: IERC.
488 G. Manogharan et al.

Manogharan, G. P., D. K. Soundarajan, and R. A. Wysk. 2011. Ruffo, M., C. Tuck, and R. Hague. 2007. “Make or Buy Analysis
Study of Energy efficiencies in Rapid Prototyping, May 21– for Rapid Manufacturing.” Rapid Prototyping Journal 13
25. Reno, NV: IERC. (1): 23–29. doi:10.1108/13552540710719181.
Petrick, I. J., and T. W. Simpson. 2013. “Point of View: 3D Printing Silveira, G. D., D. Borenstein, and F. S. Fogliatto. 2001. “Mass
Disrupts Manufacturing: How Economies of One Create New Customization: Literature Review and Research Directions.”
Rules of Competition.” Research-Technology Management 56 International Journal of Production Economics 72 (1): 1–13.
(6): 12–16. doi:10.5437/08956308X5606193. doi:10.1016/S0925-5273(00)00079-7.
Petrzelka, J., and M. Frank. 2010. “Advanced Process Planning Wohler’s Report. 2013. Additive Manufacturing and 3D Printing
for Subtractive Rapid Prototyping.” Rapid Prototyping State of the Industry: Annual Worldwide Progress Report.
Journal 16 (3): 216–224. doi:10.1108/13552541011034898. Wohlers Associates.
Pham, D. T., and R. S. Gault. 1998. “A Comparison of Rapid Xiong, X., H. Zhang, and G. Wang. 2009. “Metal Direct
Prototyping Technologies.” International Journal of Prototyping by Using Hybrid Plasma Deposition
Machine Tools and Manufacture 38 (10–11): 1257–1287. and Milling.” Journal of Materials Processing
doi:10.1016/S0890-6955(97)00137-5. Technology 209 (1): 124–130. doi:10.1016/j.
Rickenbacher, L., A. Spierings, and K. Wegener. 2013. “An jmatprotec.2008.01.059.
Integrated Cost-Model for Selective Laser Melting (SLM).” Yan., X., and P. Gu. 1996. “A Review of Rapid
Rapid Prototyping Journal 19 (3): 208–214. doi:10.1108/ Prototyping Technologies and Systems.” Computer
13552541311312201. Aided Design 28 (4): 307–318. doi:10.1016/0010-4485
Ruffo, M., and R. J. M. Hague. 2007. “Cost Estimation for Rapid (95)00035-6.
Manufacturing – Simultaneous Production of Mixed Yang, Z., R. Wysk, S. Joshi, M. C. Frank, and J. E. Petrzelka.
Components Using Laser Sintering.” Proceedings of the 2009. “Conventional Machining Methods for Rapid
Institution of Mechanical Engineers, Part B: Journal of Prototyping and Direct Manufacturing.” International
Engineering Manufacture 221 (11): 1585–1591. Journal of Rapid Manufacturing 1 (1): 41–64. doi:10.1504/
doi:10.1243/09544054JEM894. IJRAPIDM.2009.028931.
Ruffo, M., C. Tuck, and R. Hague. 2006. “Cost Estimation for Rapid Zhu, Z., V. Dhokia, A. Nassehi, and S. T. Newman. 2013. “A
Manufacturing - Laser Sintering Production for Low to Medium Review of Hybrid Manufacturing Processes—State of the
Volumes.” Proceedings of the Institution of Mechanical Art and Future Perspectives.” International Journal of
Engineers, Part B: Journal of Engineering Manufacture 220 Computer Integrated Manufacturing 26 (7): 596–615.
(9): 1417–1427. doi:10.1243/09544054JEM517. doi:10.1080/0951192X.2012.749530.