Journal of Computational Design and Engineering 5 (2018) 3–18

Contents lists available at ScienceDirect

Journal of Computational Design and Engineering

journal homepage: www.elsevier.com/locate/jcde

An additive manufacturing oriented design approach to mechanical assemblies

Germain Sossou, Frédéric Demoly ⇑, Ghislain Montavon, Samuel Gomes
ICB UMR 6303, CNRS, Univ. Bourgogne Franche-Comté, UTBM, France

a r t i c l e i n f o a b s t r a c t

Article history: Firstly introduced as a prototyping process, additive manufacturing (AM) is being more and more consid-
Received 27 April 2017 ered as a fully-edged manufacturing process. The number of AM processes, along with the range of pro-
Received in revised form 21 September cessed materials are expanding. AM has made manufacturable shapes that were too difficult (or even
2017
impossible) to manufacture with conventional technologies. This has promoted a shift in engineering
Accepted 6 November 2017
Available online 15 November 2017
design, from conventional design for manufacturing and assembly to design for additive manufacturing
(DFAM). Research efforts into the DFAM field have been mostly dedicated to part’s design, which is actu-
ally a requirement for a better industrial adoption. This has given rise to topologically optimized and/or
Keywords:
Design methodologies
latticed designs. However, since AM is also capable of manufacturing fully functional assemblies requir-
Design for additive manufacturing ing a few or no assembly operations, there is a need for DFAM methodologies tackling product’s devel-
CAD opment more holistically, and which are, therefore, dedicated to assembly design. Considering all the
Assembly design manufacturing issues related to AM of assembly-free mechanisms and available post-processing capabil-
ities, this paper proposes a top-down assembly design methodology for AM in a proactive manner. Such
an approach, can be seen as the beginning of a shift from conventional design for assembly (DFA) to a new
paradigm. From a product’s concept and a selected AM technology, the approach first provides assistance
in the definition of the product architecture so that both functionality and successful manufacturing
(including post-processing) are ensured. Particularly, build-orientation and downstream processes’ char-
acteristics are taken into account early in the design process. Secondly, for the functional flow (energy,
material, signal) to be appropriately conveyed by the right amount of matter, the methodology provides
guidance into how the components can be designed in a minimalism fashion leveraging the shape com-
plexity afforded by AM. A mechanical assembly as case study is presented to illustrate the DFAM method-
ology. It is found that clearances and material (be it raw unprocessed material or support structures)
within them plays a pivotal role in a successful assembly’s design to be additively manufactured. In addi-
tion, the methodology for components’ design proves to be an efficient alternative to topology optimiza-
tion. Though, the approach can be extended by considering a strategy for part consolidation and the
possibility to manufacture the assemblies with more than one AM process. As regards components’
design, considering anisotropy can also improved the approach.
Ó 2017 Society for Computational Design and Engineering. Publishing Services by Elsevier. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction
One of the major breakthroughs in the manufacturing domain
over the last three decades is additive manufacturing (AM). Ini-
tially introduced as rapid prototyping processes, AM techniques
are now used for end-of-use products. The common core principle
of this promising technology consists in fabricating parts in a layer-
by-layer fashion, and its multiplicity (over twenty AM processes
are available so far (Bikas, Stavropoulos, & Chryssolouris, 2015;
Peer review under responsibility of Society for Computational Design and
Engineering.
⇑ Corresponding author at: ICB UMR 6303, CNRS, Univ. Bourgogne Franche-
Comté, UTBM, 90010 Belfort Cedex, France
E-mail address: frederic.demoly@utbm.fr (F. Demoly).
Gao et al., 2015; Gibson, Rosen, & Stucker, 2009)) allow virtually
any shape to be built. Moreover these techniques are becoming
more and more democratized, and there are now machines embed-
ding these technologies for a large spectrum of applications (from
hobbyists to huge companies in the aerospace industry for
instance). This manufacturing revolution leads to more freedom
and alternatives in design, and then requires either an adaptation
of current design practices or new design paradigms. This has
raised academic and industrial interests in the so-called new
design theory and methodology area: design for additive manufac-
turing (DFAM). Indeed, since conventional design approaches in
the field of design for X (DFX) have restricted the design solutions
space, it still remains efforts to be made in order to embrace the
unique capabilities offered by AM. In addition, even if AM

https://doi.org/10.1016/j.jcde.2017.11.005
2288-4300/Ó 2017 Society for Computational Design and Engineering. Publishing Services by Elsevier.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
4 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18
techniques allow more freedom, they do have their specific limita-
tions (e.g. accuracy, surface geometry, features size, building speed,
etc.) that must be considered from the earlier design stages in
order to have an additive manufacturing-friendly product defini-
tion. AM processes are uniquely characterized (and at different
extent) by the four capabilities (Gibson et al., 2009):
1. Shape complexity: the layer-by-layer wise manufacturing of AM
makes it possible to buid virtually any shape.
2. Hierarchical complexity: features of any length scale (micro-,
meso- and macroscale) can be integrated into a part’s geometry,
which has made latticed designs of unprecedented complexity
easily manufacturable (Beyer & Figueroa, 2016).
3. Material complexity: depending on the process, materials and
materials’ properties can be varied throughout a part’s volume.
This capability is fostering research into functionally graded
materials (Garland & Fadel, 2015; Wilson & Shin, 2012) or more
generally into multimaterials parts (Vaezi, Chianrabutra,
Mellor, & Yang, 2013).
4. Functional complexity: fully (or partially) functional mecha-
nisms can be manufactured. Processes such as direct writing
(DW), and even fused deposition modeling (FDM) can pattern
conductive inks, making it possible to print electronics.
While there have multiple demonstrations of the AM processes’
ability to print mechanical assemblies (Calì et al., 2012; Lipson,
Moon, Hai, & Paventi, 2004) (even without any subsequent assem-
bly operation), research efforts in design methodologies are mostly
targeted at parts design. This may, though, be explained by the fact
that the first three aforementioned capabilities are mainly related
to parts’ properties. When assemblies are viewed through a DFAM
lens, it is mainly for redesign cases (Yang, Tang, & Zhao, 2015)
where the main purpose is part count reduction in a rather ad
hoc manner (Schmelzle et al., 2015). As parts’ rarely work by them-
selves and as ‘‘assembly is the process with greatest potential to
improve product development methods” (Whitney, 2004), consid-
eration of DFAM from an assembly perspective would pave the
way to a greater industrial adoption of AM. The purpose of this
paper is to propose an approach to AM-oriented design aiming at
pushing the boundaries of DFAM beyond part’s design and a bit
nearer assembly design. Such an approach, can be seen as the
beginning of a shift from conventional design for assembly (DFA)
to a new paradigm. A paradigm in which the main concern of the
design activity is not to ensure mechanisms assemblability, but
to ensure assembled mechanisms’ ‘‘post-processability” upon
AM. By considering a product architecture and AM processes char-
acteristics the approach intents to aid designers in setting an AM-
oriented context in which detail design can be performed. In addi-
tion, a methodology for part’s design in an AM context is provided,
as an efficient alternative to topology optimization. Section 2 will
provide a brief overview of what has been achieved in the DFAM
area so far. Then the proposed approach will be described in Sec-
tion 3. To illustrate the methodology, a mechanical case study is
presented in Section 4. Conclusions are drawn and future work is
discussed in the last section.
2. Related work
In the same way that researchers focused their efforts on devel-
oping design methodologies suitable for conventional manufactur-
ing processes (design for assembly, design for manufacturing, even
both, etc.), there are more and more researches (Yang & Zhao,
2015) targeted at developing methods to fully embrace the unique
capabilities of AM. Research efforts in such an area can be divided
in three major groups, (i) design recommendations which provide
guidelines about how to make a wise use of AM when designing,
(ii) methodologies and tools to leverage the shape complexity
capability of AM through topology optimization and lattice struc-
tures and (iii) design procedures, which take the design problem
more holistically.
2.1. AM recommendations in design
To aid designers to properly design parts for AM, there are sim-
ple rules that have been derived mainly from experiments. These
rules include part shape and dimensions, manufacturing orienta-
tion, etc. A German project named ‘‘Direct Manufacturing Design
Rule” (Adam & Zimmer, 2014; Zimmer & Adam, 2011) provided
design rules (that are focused on geometry) for Selective Laser Sin-
tering (SLS), Selective Laser Melting(SLM) and FDM processes. This
led to the development of a design rules base. It states three geo-
metric standard elements’ (and their geometric attributes) whose
combination can lead to almost any shape. An experimental study
carried out with three AM machines (with specific settings) led to
the derivation of appropriate ranges’ values and guidelines. Con-
sidering these rules in design would ensure the best achievable
qualities. In a similar way, Thomas (2009) developed design rules
for SLM. Using basic geometric entities found in lightweight struc-
tures and a set of manufacturing experiment, Kranz, Herzog, and
Emmelmann (2015) produced a structured database of basic
design guidelines focused on Laser Additive Manufacturing of
TiAl6V4; the influence of part position and direction on dimension
accuracy and surface quality were studied. A list of recommenda-
tions or principles was also proposed by Becker, Grzesiak, and
Henning (2005) to free designers’ minds, so that they can take full
advantage of the freedom enabled by AM techniques. The main
point made is to save as much as possible material (and therefore
energy and time) by putting material only where it is needed
regardless of tooling.
2.2. Topology optimization and lattice structures based approaches
Since the constraints, related to achievable tooling and relevant
to conventional manufacturing techniques, are alleviated by AM,
almost any shape (whatever its complexity) is made realistic by
AM. In addition, not only are complex shapes now worth consider-
ing, but hierarchical (i.e. through scales) complexity is also achiev-
able. These capabilities have favored two major trends in the way
item to be additively manufactured are designed:

topology optimization (TO) (Bendsøe & Sigmund, 2004; Rozvany,
2009): a numerical method which consists in optimizing matter
distribution within a defined design space, under a set of
boundary conditions, a set of loads and constraints. It is used
at the conceptual design stage to generate parts’ concepts that
are then altered either to improve aesthetics, performance, or
manufacturability.

lattice structures (LS) (Gibson & Ashby, 1997): these are numer-
ically generated truss-like structures. The main advantages of LS
include high strength-to-weight ratios, good energy absorption,
good thermal and acoustic insulation characteristics. LS are
used at the detail design stage on a part’s geometry (topology
optimized or not) to improve a specific performance.
Some design methods using TO have been developed to opti-
mize weight (Paz, Monzn, Gonzlez, Winter, & Ortega, 2015), stiff-
ness and strength (Lin, Wirtz, LaMarca, & Hollister, 2007; Rezaie,
Badrossamay, Ghaie, & Moosavi, 2013), and compliance (Joo,
Reich, & Westfall, 2009). It is worth mentioning a design method
integrating TO, bionics structures and guidelines related to powder
bed fusion processes, proposed by Emmelmann, Sander, Kranz, and
 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18
Wycisk (2011). Their methodology aims at aiding designers in the
development of lightweight structures in the aircraft industry.
While TO provides theoretically optimal solutions, their manufac-
ture may still be an issue. Indeed TO designs are not usually easily
manufacturable even for AM (e.g. too much support, enclosed
spaces, etc.). This has promoted a new research trend where the
goal is to integrate manufacturing constraints into the TO solver.
Zegard and Paulino (2015) have proposed an approach to make
ready for AM, designs output by TO. The presented workflow and
techniques are suitable for most of the AM techniques, and a tool
supporting their method has been developed. Support elimination
is being more and more considered as a constraint to the TO algo-
rithm, thus leaving unprintable or ‘‘un-postprocessable” designs
out of the design space. Such proactive approaches have been pro-
posed for FDM (Leary, Merli, Torti, Mazur, & Brandt, 2014) and for
SLM and Electron Beam Manufacturing (EBM) (Langelaar, 2016). In
order to account for the fact that support requirement are process
dependent, Gaynor and Guest (2016) put forth a more generic
support-free TO method, where a self-supporting angle is taken
as an input variable.
The shape complexity now affordable by AM, is though cur-
rently at the cost of accuracy and surface quality, characteristics
which make conventional substractive processes still competitive.
This has made hybrid manufacturing (combination of AM with
substractive processes) more and more researched (Flynn,
Shokrani, Newman, & Dhokia, 2016). That new trend in manufac-
turing is accompanied by a gap on the design side. Liu and To (in
press) have somehow pioneered on this research area by proposing
a TO-based design methodology for hybrid additive-substractive
manufacturing.
Another way for tapping into the shape complexity and hierar-
chical complexity is the use of lattice structures(LS). These are peri-
odic arrays of trusses, whose combination yields unique properties.
Their structural properties have been extensively investigated
(Vesenjak, Krstulovi-Opara, Ren, & Domazet, 2010; Yan et al.,
2015). The intricate geometry of these structures makes them dif-
ficult to design by conventional CAD systems. Many researches
have then been focused on developing tools to generate LS. Type
of lattices structures for metallic AM have been reviewed and used
as a basis for a computational tool to generate these structures in
Hadi, Vignat, and Villeneuve (2015). Design synthesis methods
for generating LS to achieve a specified stiffness were reviewed
in Jane, Sarah, Gregory, and David (2010). Automatic generation
of LS is mainly done within basic primitive shapes (cuboid, cylin-
der, etc.), which are likely not to be the real part’s geometry. To fill
that gap, Aremu et al. (2017) have proposed a voxel-based method
to generate uniform and graded lattices; the generated LS can con-
form to any external geometry.
In order to expand the design freedom allowed by AM, Primo,
Calabrese, Del Prete, and Anglani (2017) have proposed based on
a case study an approach combining TO and lattice structures.
While the results of their combined approach did not yield a global
optimum solution (considering criteria such as weight, stress, dis-
placement, etc.), at least it did demonstrate the feasibility of mix-
ing TO and lattice-based design.
2.3. DFAM approaches
The two types of contribution reviewed in the previous subsec-
tions are useful in that they provide knowledge raising the aware-
ness of designers in how their designs should be impacted by AM
or how to derive an optimal shape. However to, somehow,
rationalize the way AM knowledge is to to be taken into account
while designing, other more procedural contributions to DFAM
have been put forth. These are design procedures, which are basi-
5
cally characterized by the integration of AM processes’ characteris-
tics and functional requirements.
Ponche, Kerbrat, Mognol, and Hascoet (2014) developed a
methodology to design metal parts with additive laser manufactur-
ing processes. By considering functional specifications, process
characteristics, constraints and physical phenomena, the approach
generates a ‘realistic CAD model’ and manufacturing program plan
(i.e. printing paths and their sequence). Both the part’s geometry
and the manufacturing paths are optimized and deviations
between the ‘‘as designed” (CAD model) and the ‘‘as manufactured”
views are minimized. They previously designed a process-
independent global approach (Ponche, Hascoet, Kerbrat, &
Mognol, 2012) in which process characteristics (e.g. achievable
dimensions, accuracy, kinematics, physical phenomena, etc.) are
combined with functional specifications (e.g. geometry, mechani-
cal requirements) and context (what is sought to be optimized)
so as to generate an appropriate design definition. In addition,
Vayre, Vignat, and Villeneuve (2012) proposed an AM-oriented
approach for metallic parts. Their method consists in four steps
including analysis of product specifications, rough shapes proposal,
parametric optimization of the shape (according to the processes’
capabilities such as minimum printable length, achievable fillet
radius) and validation of the design solution.
Kumke, Watschke, and Vietor (2016) noticed that the existing
contributions made to the DFAM field influence different stages
of product design, and as such only local optimization are achieved.
For a better integration, they proposed a methodological DFAM
framework which combine the general procedure of systematic
design (related to the guidelines VDI 2221(Engineers)) with all
the existing types of DFAM methods and tools. The developed
framework guide the designer from design problem to a design
solution. In Rodrigue and Rivette (2010), a design methodology tar-
geted at assemblies design for AM has been proposed. The goal is to
assist the designer in the embodiment design stage, so that all the
advantages offered by AM are taken into account; these include
mainly part consolidation and multimaterial part. Surveying the
existing DFAM contribution (Rias, Bouchard, Segonds, & Abed,
2016) came up with the conclusion that concept generated for
AM were only partially original with ‘‘a maximum of 75% new-
ness”. They have then proposed a 5 stages ‘‘creative DFAM” meth-
ods fostering the generation of creative concepts exploiting the
unique capabilities of AM; the proposed method is somewhat
based on the AM design features database presented in Bin
Maidin, Campbell, and Pei (2012). The method is mostly intended
to early design stages. The 5 stages include: features discovery,
exploration, ideas evolution, concept generation and concept eval-
uation. This contribution is mainly related to part design. A similar
contribution related to assembly design was made in Yang and
Zhao (2016), where it was established that AM has no influence
on functional analysis and the recognition of desired behaviors.
The study shows that AM unique capabilities have rather a role
to play in the process of mapping desired product’s behaviors to
physical structures. Based on this finding, a design synthesis strat-
egy was proposed with the aim of helping designers who are
novice in AM to think in an ‘‘AM way”.
2.4. Motivation
Currently, the mainstream in DFAM seems to be focused on
leveraging AM capabilities to design components with complex
shapes, or equivalently optimized performance. While the afore-
mentioned contributions proof efficient in designing (or redesign-
ing) effective parts or in consolidating assemblies, none of them
seems to consider the ability of some AM processes to produce
functional assemblies. When fully fonctional assemblies are con-
sidered in an AM context, it is in an ad hoc manner for prototyping
6 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18
(Lipson et al., 2004) or benchmark purposes (Mavroidis,
DeLaurentis, Won, & Alam, 2000), or for joint analysis (Chen &
Zhezheng, 2011; Rajagopalan & Cutkosky, 2004). Besides, issues
related to the AM processes’ and post-processing constraints are
not enough considered. Here the research work is intended to fill
these gaps, by proposing a DFAM approach, to ease the design of
assembly-free mechanisms, that is, assemblies of moving parts
that can be additively manufactured in a single print and whose
moving parts are manufactured fully assembled. Illustrations of
such mechanisms include the fully assembled 3D printed 28-
geared Cube made by the maker Maund (2013) as shown in
Fig. 1, the historical mechanisms printed by Lipson et al. (2004),
or the bearings manufactured by SLM in Changhui, Yongqiang,
Ziheng, and Di (2013).
3. Additive manufacturing oriented design approach
As it is possible to manufacture fully functional assemblies with
AM processes (Calignano et al., 2014; Gibson et al., 2009), the AM-
oriented design approach is targeted at designing assemblies to be
partially or totally additively manufactured. As such it provides
assistance in:
1. Structuring the product architecture (or layout) so that it is
additively manufacturable.
2. Providing guidance into the way the product’s components
must be designed with the bare minimum matter.
Ullman (2009) has stated that: ‘‘It has been estimated that
fewer than 20% of the dimensions on most components in a device
are critical to performance. This is because most of the material in
a component is there to connect the functional interfaces and
therefore is not dimensionally critical. Once the functional inter-
faces between components have been determined, designing the
body of the component is often a sophisticated connect-the-dots
problem.” The philosophy behind the way the proposed methodol-
ogy is to design components is profoundly related to that state-
ment, in that the shape complexity allowed by AM gives way to
design components whose dimensions are closer to (or lower than)
the aforementioned 20%. Furthermore, to better suit our approach
the statement can be extended as ‘‘. . .a sophisticated connect-the
dots problem in order to assure a structural integrity and a proper
conveyance of the flows related to the product functionality”.
Fig. 1. Assemble-free 28-geared Cube (Maund, 2013).
For a product’s components to have minimal functional dimen-
sions, we posit that there are three characteristics to look at first.
Functions performed by a product occur mainly at its functional
interfaces. These also include interfaces of the product with exter-
nal elements (user, raw materials, a fluid flow, etc.) and interfaces
between the product’s components (due to their kinematics rela-
tionships). Besides, those functions are usually associated with a
flow (Pahl, Beitz, Feldhusen, & Grote, 2007) (i.e. energy, signal, or
material) which are physically routed throughout the product (or
more specifically its components). Components’ interfaces are the
‘‘gates” of these flows. Finally, each component is bounded by a
design volume in which matter forming the component can be
allocated. Any of the product’s components is therefore character-
ized by: (i) its functional interfaces, (ii) the possible flows getting
through it, and (iii) a design space. The components – and therefore
the whole product – are then designed by connecting the FIs,
allowing a proper conveyance of the flows while staying in their
specified design space and – in an AM context – abiding by the
selected AM process’ constraints. These elements (FIs, flows and
design space) will be subsequently referred to as component’ con-
trol structures (CS).
As the proposed approach aims at designing product whose
components are at their minimal dimensions, it is structured in
three main stages (see Fig. 2) consisting of:

A functional analysis, intended to describe the product func-
tionality and to understand both the relationships between
the product and its environment, and the relationships between
the components. Since functions structures are systematic ways
of identifying the flows ‘‘processed” by the product, this stage
will also delineate all the involved flows.

The derivation of the components’ control structures.

The design of the components’ geometries.
For the sake of clarity, the first subsection is dedicated at clari-
fying some artifacts related to the proposed AM-oriented design
approach.
3.1. Concepts underpinning the approach
3.1.1. Functional interfaces
A functional interface (FI) (of a product or a component) is an
interface, that characterizes the relationship that two spatial
regions must have for a function to be fulfilled. That relationship
may require an actual physical contact – subject to a certain clear-
ance – between two components (e.g. sliding components) or it
may be contact free. A FI may be a surface (either planar or
three-dimensional), a point, or a line. A component’s FIs are basi-
cally where its geometry ‘‘grows” from, and as such they must be
carefully identified in order for the component’s dimensions to
be as much as possible close to the aforementioned 20% that are
truly critical to its performance (Ullman, 2009). There are three
kinds of FIs that can be distinguished:

Contact free FI: it may fulfill functions related to aesthetics, gas
or water tightness, heat dissipation (such as cooling fins), fluid
deviation (such as aircraft’s wings surfaces), etc.

Handling FI: it can be for the product use (lever, button, etc.) or
for the product to be assembled, transportable or maintained,
and as such the physical contact is not permanent.

Part-to-part contact relationship FI: this kind of FI may be
required to perform positioning in an assembly, to provide
guidance or for load transmission between parts. They can be
generated using the Skeleton geometry-based Assembly Con-
text Definition (SKL-ACD) approach (Demoly, Toussaint,
Eynard, Kiritsis, & Gomes, 2011). Based on the kinematic pairs
 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18 7

Fig. 2. AM-oriented design approach.

between parts, this approach can compute the generic surfaces
on which the required FIs can be laid or the required FI itself
along with geometric entities controlling its orientation and
location.
Once FIs are identified, they must be positioned. As illustrated
in Fig. 3, given a design space, the positioning of a FI – unless it
is planar – can have two effects: it can either shape it (that is, it
restrains part of the design space into a specific shape) or hollow
it (that is, it defines the shape of a void in it) (see Fig. 4).
3.1.2. Functional flows
The overall function performed by a product is associated with
flows. In their lowest level of specificity these are energy, material
or signal (Pahl et al., 2007). Conceptually these flows are closely
associated with basic functions (e.g. transfer, channel, separate,
etc.). Hirtz, Stone, McAdams, Szykman, and Wood (2002) have clas-
sified the basic functions within eight groups including branch,
channel, connect, control magnitude, convert, provision, signal
and support. The reader is advised to refer to Hirtz et al. (2002) for
a thorough list of the basic functions along with their definitions.

Fig. 3. Functional interface effect on design spaces.

8 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18

Fig. 4. Preferential printing configuration for a pair of components.

Components, as the physical embodiments of the functions, are Table 1

Samples from the reconciled functional basis and flows sets proposed in Hirtz et al.
ultimately ‘‘flows processors”. While some functions may, explic-
(2002).
itly, require a flow to get through a component, others may just
require the flow to bypass it. For instance within the signal func- Functional basis set
Class (Primary) Secondary Tertiary Correspondents
tions group – defined as ‘‘To provide information on a material,
Channel Import Form entrance, allow,
energy or signal flow as an output signal flow” – there is the basic input, capture
function detect whose definition is: ‘‘to discover information about Export Dispose, eject, emit,
a flow”. A component fulfilling such a function is not necessarily empty, remove,
crossed by the flow, and as such the matter forming it, is not crit- destroy, eliminate
Transfer Carry, deliver
ical to its performance; instead it is the component’s interface in Advance, lift, move
contact with the flow which is critical. On the other hand, a com- Conduct, convey
ponent fulfilling the transmit function – defined as ‘‘To move an Functional flow set
energy from one place to another” – is a ‘‘flow crossed” component, Class (Primary) Secondary Tertiary Correspondents
and as such has its matter critical to the performance. Neverthe- Energy Mechanical rotational
less, it is worth mentioning that the working mechanism chosen translational
Biological
to fulfill a function may still dictate whether the flow is to bypass
Chemical
or to get through the component. Given a component whose func- Electrical
tion(s) require(s) a flow to cross it, the material, its shape and its
dimensions determine how well is (are) the function(s) fulfilled.
travelling through the components (either from other
3.2. Functional analysis of the product components, or from EEs).

While the proposed approach could easily be used for a rede-
sign case, it is mainly intended for new product development. As
such it starts with a product’s need or a required functionality
statement. In addition, it is assumed that the product conceptual
design is complete, so that the product concept is also an input
of the process. This stage is intended to get a thorough understand-
ing of the product architecture and of the way it is to work. For the
sake of clarity and consistency, the following steps are proposed to
handle the input.
At the end of this stage, the derived concept is described with its
part-to-part kinematic relationships, relationships with the exter-
nal environment, and flows through its components.
3.3. Definition of the components control structures
The purpose of this stage is to derive a design context in which
both the product’s functionality is maintained and AM constraints
are proactively considered.

Step 1 – External functional analysis. The interactions between
the product and its external environment (that is, every-
thing that the product interact with and which is not part
of the product itself) are firstly clarified. The flows
between environmental elements (EEs) and the product
are also determined.
Step 2 – Functional breakdown refinement. The overall func-
tion of the product is decomposed into basic functions
using, for instance, the reconciled functional basis (Hirtz
et al., 2002) (see Table 1). The basic functions are then
arranged in a logical structure that enables the fulfilment
of the overall function. Finally the flows are routed
through these basic functions using the reconciled flows
set (Hirtz et al., 2002) (see Table 1).
Step 3 – Product architecture. The derived concept is then
decomposed into its components and is analyzed in terms
of part-to-part relationships (kinematic/technological
pairs) within the product and with external elements.
Secondly, a part-to-function mapping is made in order to
know which basic function(s) is(are) fulfilled by each com-
ponent, and most importantly to know which flows are
3.3.1. Design spaces definition
For each of the components identified in the part-to-part rela-
tionships graph a design volume is specified by the designer in the
form of a rough shape (cylinder, cube, etc.). Then design spaces are
positioned relatively to each other according to the derived concept.
Since AM is capable of complex shapes, components without rela-
tive motion, could all be consolidated. However, some components
may be subject to wear and therefore required to be usually chan-
ged, they may be required to be removable, they may be outsourced
(such as standard components), or they may be required to be man-
ufactured either in a material that is not processed by the chosen AM
technique, or by a different manufacturing process. In these cases, a
separate design space defined for these components is removed
from their mating components’ design spaces and new FIs specific
to a rigid kinematic pair, are defined on the design space of the com-
ponents they are attached to.
3.3.2. Emergence of functional interfaces
For each component, the relationships with its neighboring
components and (possibly) with external elements are analyzed
and the corresponding FIs are picked and positioned. Positioning
 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18
a FI between two components affects the design space of both
components. The process is repeated until all the relationships
are physically instantiated by the FIs in the design spaces. At this
stage, each design space (or equivalently each component to be
additively manufactured) is associated with its FIs and the whole
product is modeled in a layout form.
3.3.3. Additive manufacturing contextualization
The time required for post-processing an additively manufac-
tured assembly should not outweigh the one required to manufac-
ture (by AM or not) the assembly’s components separately and to
assemble them as it is conventionally done. Therefore care should
be given to the process specific constraints so that the mechanism
is successfully manufactured and post-processed (if needed). Once
the developed concept has been abstracted in a layout form and
functionality has been assured through the right FIs, AM specific
constraints must then be taken into account. These constraints
are considered through the steps of clearances setting, printing
configuration, printing orientation choice, and accessibility to the
clearances.
Step 1 – Clearances settings. To ensure a successful manufac-
turing (and a seamless post-processing) without imped-
ing the functionality related to the joints, care must be
given to clearances within pairs of sliding FIs. The chosen
clearance is related to the required joint performance,
and the printing resolution. Clearance between sliding
surfaces must be tight enough to ensure joint perfor-
mance (avoidance of instability, vibrations, and
unwanted DOFs). They must also be large enough to pre-
vent mating surface from merging during the manufac-
turing. Depending on the mating surfaces geometries,
clearances values differ. For flat surfaces (involved in
prismatic pair mainly), reported successful values
include 0.3 mm with a SLA and a SLS machines
(Mavroidis et al., 2000). For circular surfaces, reported
successful values are 0.5 mm (SLA and SLS (Mavroidis
et al., 2000)) - 0.2 mm with a PolyJet machine (Chen &
Chen, 2010) - 0.2 mm and 0.05 mm (with alterations to
the journal’s shape of a revolute joint) with FDM (Wei,
Tian, & Joneja, 2016) - 0.2 mm with SLM (Su, Yang,
Wang, & Chen, 2013). Because they are specific to the
AM machines (accuracy, layer thickness) and the materi-
als used, these values cannot be considered as norma-
tive; though they can be used as a rough estimate of
successful joint clearance. Besides there are some AM
services providers Shapeways (2017) which publish for
each of their processed materials, the right clearance
value. All the sliding FIs are then to be spatially con-
strained using the required clearances. This may involve
offsetting some FIs.
Step 2 – Printing configuration. As components can move rel-
atively to each other and are in situ assembled, they can
be manufactured in different configurations. To reduce as
much as possible the need of support structure (or
unprocessed raw material) within the clearances and to
allow as much as possible access to the clearances, com-
ponents – wherever possible – should be moved in such a
configuration that mating surfaces have the lowest facing
area as shown in Fig. 3. In addition when possible, the
components should be moved such that clearances
which are along a single direction (e.g. for the kinematic
pairs prismatic, cylindrical, etc.) are mostly all aligned
along one single direction. Indeed, in such configuration,
selecting a printing direction to avoid support structure
within the clearances requires less compromise. Never-
9
theless, while choosing a configuration, care must also
be given to the available printing volume so that the pro-
duct does not exceed it.
Step 3 – Printing Orientation. Printing orientation affects num-
ber of parameters of the printed item, these include
dimensional accuracy, surface finish quality, strength,
build time, support structure, and eventually cost. As
such determining a manufacturing direction is quite a
cumbersome task, which involves many competing
goals: while a direction may yield the lowest build time,
it may be the one leading to the poorest surface quality
on a surface that is critical to the item’s performance.
Many studies have tackled this task for parts (Morgan,
Cherry, Jonnalaganna, Ewing, & Sienz, 2016; Pandey,
Venkata Reddy, & Dhande, 2007) whose geometries are
explicitly defined. However when considering assem-
blies, we posit that the main concern must be about
the clearance gaps, especially about avoiding as much
as possible support structure (or unprocessed raw mate-
rial) within them, since they are critical to the joints’ per-
formance. In order to choose an optimal printing
orientation which minimizes post-processing effort, an
indicator should be used so that different orientations
can be compared quantitatively. Requirements for such
metrics include:

It must depends on all the assembly’s joints.

It must provide a measure of the trapped matter
within the joints’ clearances, and as such its minimal
value would indicate the best orientation.

As it is easier to remove loose (unprocessed) raw
material (in the form of powder or resin) than to
remove cured support structure, be it dissolvable or
not, the measure must account for support structure
and loose material differently. Support structure pres-
ence should then be penalized by a value higher than
the one yielded by loose raw material presence.

For a single joint, it must be able to differentiate print-
ing orientations based on the volume of trapped
material. Thus, even though there must be support
(or unprocessed) material in any two orientations,
there is still the possibility to know which one is likely
to yield the least. See Fig. 5, where this requirement is
illustrated with a prismatic joint. Using the trapped
material’s volume (weighted depending on the mate-
rial’s state) would easily indicate that the orientation
on the left (orientation 1) would require less support
than the one on the right.
To account for all these requirements we propose the following
metrics. For joint i, a measure of how much in between material,
printing it in orientation ~
u, would require is:
Fig. 5. Distinguishing two build orientations requiring support.
10 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18

uÞ ¼ kuc v i;uc ð~
J i ð~ uÞ þ kss v i;ss ð~
uÞ ð1Þ Table 2
In between materials removal.
where
v i;uc :
volume of uncured material
Type of in-
between
Processes Removal
methods
Requirement for removal

v i;ss :
volume of support structures materials
In order to penalize support structures, kuc and kss are two Uncured SLA Gravity based Access for fluid flow from
coefficients such that kuc  kss . The measure that is then to resin SLA flow, suction inside
be used for the whole assembly is an increasing function of Unprocessed SLM, SLS, Gravity based Access for the powder flow:
the all the J i ð~
uÞ: powder 3DP flow, suction holes in dimensions of a few
times particles’ size
Sð~
uÞ ¼ f ðJ 1 ; J 2 ; . . . ; J n Þ ð2Þ Water FDM, Washing away in Access for fluid flow through
soluble PolyJet solution, jetted the support
The simplest ways of constructing f being by summing or support hot solution
multiplying the J i . An illustration is provided with a single structure
Rigid FDM, SLA Breaking away Access for tools
revolute joint in Fig. 6 with kuc ¼ 1 and kss ¼ 100. Once S is
polymer with hands or
evaluated for each printing orientation in the selected config- support tools
uration, the best orientation is the one yielding the lowest structures
value of S. Rigid SLM, Breaking away Access for tools
Step 4 – Accessibility to the clearances. Even though a printing metallic LENS with hands or
support tools
orientation has been selected to minimize both uncured structures
material and support structures within the clearances,
chances are that there are still clearances filled with sup-
port and/or loose materials (subsequently refered to as
in-between materials). Care must then be given to the of 1 mm which may not be long enough to reach the
clearances, so that depending on what they are filled whole clearance.
with, there is a way to clean them, especially there is - Laser cutting: lasers’ tips diameters can be in the range
an access for what is used for removal. This is the pur- of micrometer, and their power can be adjusted so
pose of this step. Ways for clearances cleaning depend that they only destroy the support structures. There
on the clearance, the kind of in-between material, and is no limitation as regards the depth the laser can
the kind of tools used for removal. Table 2 outlines the attain. Such route can be beneficial in that it can also
possible situations. The most critical in-between mate- improve the surface quality of the facing surfaces.
rial being the rigid metallic support structure which is Though, issues related to thermal cutting processes
required for powder bed fusion processes. Despite a may arise, especially the heat-affected zone (HAZ)
recent report of dissolvable metallic support structure issue.
(Lefky et al., 2017) for the LENS process, metallic support - Water jet cutting: high precision technologies using
structures are mostly removed from parts by using pliers high pressure water such as Microwater-
or chisels. Given the values of successful clearances jetÒ(Microwaterjet, 2017) have cutting accuracy in
(which are less than a millimeter), such tools may not the range of 0.01 mm, can be positioned with an accu-
be useful in cleaning the clearances. Alternative solutions racy of 3 lm, can provide a surface quality of N7 (Ra
for hard support removal may be: 1.6 lm), and can cut almost any material between
rubber and steel. Such solution for support removal
- Micromilling: high precision milling tools can have has the advantage of being ‘‘cold”, thus avoiding
tips of diameter of as low as 0.1 mm and can process issues related to thermal solution like lasers.
materials as hard as very hard steel (HRC 55) (which is
quite harder than the typical porous metallic support The identified solutions are those that allow access to tiny inter-
structures). However such solution may be limited by stices (in contrast to conventional tools used to remove support
cutter’s length. Indeed, typical the cutter’s lengths structures under overhangs) and that are still effective at breaking
found in high precision milling tools are in the range away hard materials. As long as access is provided for the tool’s

Fig. 6. Trapped material measure illustration.

 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18
energy (be it mechanical or thermal), any of these solutions can be
used to clean the clearance from possible support structures. This
is what makes the step of accessibility to the clearances, of high
importance in the proposed methodology.
The design spaces, where necessary, are then to be updated
with appropriate accesses from outside to the clearances. The
accesses depend on the clearances themselves and the capabilities
of tools to be used to remove the support. Joint’s design alteration
should, however, be kept as an option as well in case providing
access to the clearance would lead to trimming one of the parts
and/or too long and complicated post-processing.
3.4. Components geometry definition
Once both functionality and manufacturability have been
ensured by constraining the components’ design spaces and their
FIs, a strategy must be retained to design their geometries. As out-
lined in Fig. 7, a five-steps-strategy is proposed to design the geom-
etry in a minimalist way. First, the component’s FI shapes are
drawn. Second, the FIs’ shapes are thickened into functional vol-
umes (FVs), which are the physical embodiments of the FIs; the
thicknesses can be governed both by resolution of the considered
AM process and tolerances related to the considered FI. In the third
step, paths joining the FVs are defined; this can be done either to
ensure the component’s connectedness, or to ensure a way for a
specific flow (or even a combination of flows), or both. Fourthly,
the connecting elements’ shapes along the previously defined
paths are designed. To improve the performance of the component,
the section can be of different types including solid cross-section,
solid cross-section with lattice structure, hollowed cross-section,
hollowed cross-section with lattice structure, etc. To avoid sharp
corner (and equivalently, stress concentration), corners are
smoothed in the fifth step. Each of these steps leads to some geo-
metric parameters. Those are then finally optimized to generate a
component that behaves and conveys the flows appropriately.
The following scheme has been retained for the parametric
optimization:
1. Choice of each parameters bounds.
2. Specification of any constraint on the parameters. These can be
geometric constraints such as a relationships between parame-
ters, or constraints ensuring a proper behavior such as maxi-
mum stress or minimal natural frequency. The functional
flows conveyed by the part on one hand, and chosen material
Fig. 7. Steps related to the definition of parts geometry.
11
on the other hand can provide indication on what constraints
related to an appropriate behavior must be met.
3. Definition of the component’s mass as an objective function.
Choosing the component’s mass as an objective function to
minimize is consistent with the endeavor to design components
with the bare minimum matter – thus, leveraging the AM’s
capabilities. Others objective functions which are sensitive to
the chosen parameters and which are relevant to the sought
performance may also be chosen.
4. Case study
The proposed approach has been illustrated by designing a
small, light duty, homeowner’s vise whose overall dimensions
are: 200100100 mm (lwh). It must apply a clamping
strength of around 100 lb. (around 45 kg, or 450 N). The vise is to
be built out of steel, and the selected additive manufacturing pro-
cess is SLM.
Functional analysis The vise’s concept is composed of a frame
(which is fixed to the support, and which includes one of the jaws),
a lever, a moving jaw and a screw rod. Physical relationships
through kinematics pairs between these components, and with
external elements, are presented in Fig. 8, where FIs are also spec-
ified on each part-to-part relationship. There is only one flow
involved in the product: energy in the form of forces. Such forces
are described in Fig. 9.
Components’ control structures definition It is considered
that the lever is not a component to be additively manufactured.
Indeed, it is a simple cylindrical component and as a component
which is likely to bend under the torque applied to it, it is required
to be a spare part. Therefore control structures are derived for the
screw rod, the moving jaw and the frame. Fig. 10a displays the lay-
out model with the design spaces of the components according to
the derived concept; FIs are shown in Fig. 10b.
Additive manufacturing contextualization
1. Clearances setting. Two of the pairs of sliding surfaces are circu-
lar (screw pair between the screw rod and the frame, and revo-
lute pair between the screw rod and the moving jaw). There
remains one pair of sliding surface corresponding to the pris-
matic pair between the moving jaw and the frame. Clearances
have been set for all of these joints to 0.2 mm.
2. Printing configuration. The FIs materializing the prismatic pair
between the moving jaw and the frame overlap in any configu-
12 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18

Fig. 8. Product’s part-to-part relationships along with corresponding FIs.

Fig. 9. Functional flow in the form of mechanical energy.

Fig. 10. Vise layout model of the case study including (a) volume-envelopes and (b) FIs.
 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18
Fig. 11. Altered revolute pair between screw rod (orange part) and moving jaw
(green part).
ration, as do the ones materializing the screw pair. However, to
reduce regions where material can be trapped, the revolute pair
has been somehow altered in such a way that the degree of
freedom along the ~ x direction is not totally restrained. Indeed
as shown in Fig. 11 when moving in the ~ x direction, the screw
rod translates the moving jaw by contacting its Interface A, and
when moving the in the þ~ x direction, the motion transmission
is made through a contact with Interface B. Therefore the print-
ing configuration has been chosen such that the screw rod con-
tacting surface is between Interface A and Interface B. The
clearances’ directions are all (functionally) aligned along a sin-
gle direction, which will ease the printing orientation choice.
3. Printing orientation. Fig. 12 shows the evaluation of the three
main printing orientations. As one could have intuitively
asserted, the orientation requiring the less effort to post process
is the one along the ~ x direction. Indeed in this orientation, all
the clearances are parallel to the printing direction, which mean
there is no (hard metallic) support structures within them. It is
worth noticing that without the revolute joint alteration, there
would be support structure required within its clearances for
Fig. 12. Printing orientation choice.
13
any of the three orientations. The selected printing orientation
is then ~x.
4. Accessibility to the clearances. With the chosen printing orienta-
tion, the only one material within the clearances is loose unpro-
cessed powder, and these clearances are already opened to the
outside, therefore there is no need to provide a specific access to
them. The powder can be removed by suction or by blowing it
away.
Components’ design. Once the components’ design spaces have
been positioned, their geometries can be defined and optimized. A
proprietary tool (Pegasus CAD Assistant (Demoly et al., 2011)),
whose a screenshot is provided in Fig. 13, has been used to semi-
automatize the generation of the functional interfaces.
Fig. 15 shows the steps for the frame’s design; parameters
were defined for FIs’ geometries, the thicknesses of the functional
volumes, the profiles of the connecting elements, the spatial
curves defining their paths, and for the fillet radii. The connecting
elements have been generated using a lofting function. Using
COMSOL MultiphysicsÓ, parametric optimization was run to min-
imize the frame’s weight while maintaining the maximum stress
below the chosen steel’s yield stress. A simulation was first run to
know where are the most critical regions. This suggests what
parameters must be included in the parametric optimization,
and what other parameters may be ignored. For the frame, up
to 21 geometric parameters could be used to optimize the geom-
etry. As can be seen in Fig. 16, the parameters (particularly the fil-
let radii) defining the connecting elements between the frame
and the jaw must be included in the optimization. As shown in
Fig. 14, only 9 parameters were finally retained for the optimiza-
tion. Only the vise’s weight was used as objective function, and
the materials’ yield stress was used to constrain the maximum
Von Mises stress. Since the location of the peak stress can jump
from a place to another one, this constraint is non-
differentiable; therefore a gradient-free optimization algorithm
has been chosen. Using the Constrained optimization by linear
approximation (COBYLA) method – implemented in COMSOL
MultiphysicsÓ– on a laptop with Intel Core i5 Processor
(2  1:9 GHz) and 8 GB RAM, the optimization took 40 min 32 s.
14 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18

Fig. 13. Generation of the assembly skeletons for functional interfaces definition within Pegasus CAD Assistant and CATIA v5.

Fig. 14. Retained parameters for the parametric optimization.

The frame’s initial mass is 1.57 kg, and the optimization led to a
mass of 0.71 kg. Fig. 17 shows the design before and after
optimization.
The whole vise’s design is presented in Fig. 18.
5. Conclusions and future work
In this paper, a twofold contribution has been made to the field
of DFAM. Firstly, a procedural approach has been proposed to help
to embrace the AM capability to manufacture assemblies without
any assembly operations. Taking as input a product’s concept, it
provides guidance into how to derive a functional architecture
which is additively manufacturable. Secondly, an alternative
approach to how to leverage the AM shape complexity capability
– other than through topology optimization – has been proposed.
It consists in identifying the component’s functional interfaces –
that is, interfaces required for the component’s functionality –
along with their positions within a design space, and the possible
functional flows required for the functionality. Then, the geometry
is designed by connecting the FIs so that the necessary flows are
conveyed appropriately and/or structural integrity is assured;
parametric optimization is considered to allow a further minimiza-
tion of the required matter. A case study has been done to illustrate
and convey the main ideas of the AM-oriented design approach.
To the best of our knowledge, it is the first design methodology
in the DFAM field aimed both at designing assemblies (to be man-
ufactured with a few or no assembly operations) and parts. And as
such, a comparative study cannot be conducted to objectively eval-
uate the proposed approach. Nevertheless, a few comments can be
made, and there is still room for perfecting it.
As regards the way components are to be designed, while TO is
run by iteratively removing matter until a specific criterion (or a
set of criteria) is met, our approach prompts to design a component
 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18 15

Fig. 15. Frame’s geometry definition: (a) FIs shapes within the design space - (b) FIs thickening into functional volumes - (c) connecting elements profiles and guide curves
definition - (d) connecting elements - (e) corners smoothing.

Fig. 16. First simulation run with initial values of the parameters.

Fig. 17. Frame’s geometry: (a) before optimization - (b) after optimization.
16 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18
Fig. 18. The designed vise assembly.
Fig. 19. Ordinary vises obtained by topology optimization.
rather by adding matter until all the requirements are met; as such
the approach could be coined as ‘additive design for additive man-
ufacturing’. Fig. 19 shows what vises, designed using topology opti-
mization, could look like. In addition, the proposed approach
provides a control over how the FIs are linked; indeed various
kinds of linking elements can be used and optimized either for
physical requirements (be them for the a single physical problem
or a multiphysics one), for aesthetic reasons, or both. Ways for
improve the proposed method include: taking into account the ani-
sotropy related to the chosen AM process, in the FE simulations run
for the parametric optimization, as this could improve the quality
of the result. Finally, lattices structures could be considered in the
component final shape optimization.
Besides, concerning the assembly design aspect, since compo-
nents which are being built can theoretically move, there may
not be anything preventing them from moving during the manu-
facturing depending on the AM process. As such, owing to their
increasing weight, gravity could move them, leading thus to a high
risk of manufacturing failure. Another way of extending the pro-
posed methodology would then consist in taking into account an
approximate weight of the components in the printing direction
choice. Furthermore, a limitation of the proposed approach is the
choice of a printing orientation through geometrical hand calcula-
tions of the in-between material. This could be improved by a com-
putational tool taking into account a specific AM process
characteristics (such as angle by which support structures are
required within a clearance and maximum length of self-
supporting overhangs) and operating conditions (powder particle
size, laser power, layer height, etc.). Such tool would compute in-
between material more accurately to generate an optimal printing
orientation. In fact, the whole stage of AM contextualization is
highly dependent on the designer’s knowledge of the used AM pro-
cess characteristics; what is not necessarily guaranteed. As such,
this stage of the design process could be fully automatized. A com-
putational tool, storing the various AM processes’ assembly related
characteristics, would then take as input the assembly’s layout
model including each component design space, the FIs related to
all the kinematic pairs, along with some motion constraints on
the design spaces. The output of the tool would be the layout
model with the successful clearances set between the FIs, an opti-
mal printing configuration, an optimal printing direction, and (if
required) proposal of ways to provide access to the clearances.
The latter would then be chosen depending on the tools to be used
for clearances’ supports removal. Finally, another promising area of
future work, is the possibility to consider part consolidation (up to
the selected AM process(es) capabilities) through an architecture
minimization stage.
Conflicts of Interest
Authors declare that there is no conflict of interest.
Acknowledgment
This research activity is part of much larger project in the field
of design for 4D printing (additive manufacturing with smart
materials). The authors would like to thank the Ministère de
l’Enseignement Supérieure et de la Recherche as main financial
support of this research program, and S.mart Franche-Comté clus-
ter for their participation.
References
Adam, G. A. O., & Zimmer, D. (2014). Design for additive manufacturing–element
transitions and aggregated structures. CIRP Journal of Manufacturing Science and
Technology, 7(1), 20–28. https://doi.org/10.1016/j.cirpj.2013.10.001.
Aremu, A., Brennan-Craddock, J., Panesar, A., Ashcroft, I., Hague, R., Wildman, R., &
Tuck, T. (2017). A voxel-based method of constructing and skinning conformal
and functionally graded lattice structures suitable for additive manufacturing.
Additive Manufacturing, 13, 1–13. https://doi.org/10.1016/j.addma.2016.10.006.
Becker, R., Grzesiak, A., & Henning, A. (2005). Rethink assembly design. Assembly
Automation, 25(4), 262–266. https://doi.org/10.1108/01445150510626370.
Bendsøe, M. P., & Sigmund, O. (2004). Topology optimization: Theory, methods, and
applications (2nd ed.). Berlin, Heidelberg: Springer-Verlag. https://doi.org/
10.1007/978-3-662-05086-6.
Beyer, C., & Figueroa, D. (2016). Design and analysis of lattice structures for additive
manufacturing. Journal of Manufacturing Science and Engineering, 138(12).
https://doi.org/10.1115/1.4033957. 121014–121014–15.
Bikas, H., Stavropoulos, P., & Chryssolouris, G. (2015). Additive manufacturing
methods and modelling approaches: a critical review. The International Journal
of Advanced Manufacturing Technology, 83(1), 389–405. https://doi.org/10.1007/
s00170-015-7576-2.
Bin Maidin, S., Campbell, I., & Pei, E. (2012). Development of a design feature
database to support design for additive manufacturing. Assembly Automation, 32
(3), 235–244. https://doi.org/10.1108/01445151211244375 <http://www.
emeraldinsight.com/doi/abs/10.1108/01445151211244375>.
Calì, J., Calian, D. A., Amati, C., Kleinberger, R., Steed, A., Kautz, J., & Weyrich, T.
(2012). 3d-printing of non-assembly, articulated models. ACM Trans. Graph., 31
(6), 1–8. https://doi.org/10.1145/2366145.2366149.
Calignano, F., Manfredi, D., Ambrosio, E. P., Biamino, S., Pavese, M., & Fino, P. (2014).
Direct fabrication of joints based on direct metal laser sintering in aluminum
and titanium alloys. Procedia CIRP, 21, 129–132. https://doi.org/10.1016/j.
procir.2014.03.155.
 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18
Changhui, S., Yongqiang, Y., Ziheng, Y., & Di, W. (2013). Digital design and direct
manufacturing of non-assembly mechanisms by selective laser melting. In 2013
IEEE International Symposium on Assembly and Manufacturing (ISAM) (pp. 142–
144). doi:https://doi.org/10.1109/ISAM.2013.6643511.
Chen, Y. H., & Chen, Z. Z. (2010). Major factors in rapid prototyping of mechanisms.
Key Engineering Materials, 443, 516–521. https://doi.org/10.4028/
www.scientific.net/KEM.443.516.
Chen, Y., & Zhezheng, C. (2011). Joint analysis in rapid fabrication of non-assembly
mechanisms. Rapid Prototyping Journal, 17(6), 408–417. https://doi.org/10.1108/
13552541111184134 <http://www.emeraldinsight.com/doi/abs/10.1108/
13552541111184134>.
Demoly, F., Toussaint, L., Eynard, B., Kiritsis, D., & Gomes, S. (2011). Geometric
skeleton computation enabling concurrent product engineering and assembly
sequence planning. Computer-Aided Design, 43(12), 1654–1673. https://doi.org/
10.1016/j.cad.2011.09.006.
Emmelmann, C., Sander, P., Kranz, J., & Wycisk, E. (2011). Laser additive
manufacturing and bionics: Redefining lightweight design. Physics Procedia,
12, 364–368. https://doi.org/10.1016/j.phpro.2011.03.046.
Engineers, V.D.I.A.o.G. Vdi 2221 - systematic approach to the development and
design of technical systems and products.
Flynn, J. M., Shokrani, A., Newman, S. T., & Dhokia, V. (2016). Hybrid additive and
subtractive machine tools research and industrial developments. International
Journal of Machine Tools and Manufacture, 101, 79–101. https://doi.org/10.1016/
j.ijmachtools.2015.11.007.
Gao, W., Zhang, Y., Ramanujan, D., Ramani, K., Chen, Y., Williams, C. B., . . ., Zavattieri,
P. D. (2015). The status, challenges, and future of additive manufacturing in
engineering. Computer-Aided Design. doi:https://doi.org/10.1016/j.cad.2015.04.
001.
Garland, A., & Fadel, G. (2015). Design and manufacturing functionally gradient
material objects with an off the shelf three-dimensional printer: Challenges and
solutions. Journal of Mechanical Design, 137(11). https://doi.org/10.1115/
1.4031097. 111407–111407–11.
Gaynor, A. T., & Guest, J. K. (2016). Topology optimization considering overhang
constraints: Eliminating sacrificial support material in additive manufacturing
through design. Structural and Multidisciplinary Optimization, 54(5), 1157–1172.
https://doi.org/10.1007/s00158-016-1551-x.
Gibson, L. J., & Ashby, M. F. (1997). Cellular solids: Structure and properties. Cambridge
solid state science series (2nd ed., ). Cambridge University Press. https://doi.org/
10.1017/CBO9781139878326.
Gibson, I., Rosen, D. W., & Stucker, B. (2009). Additive manufacturing technologies:
Rapid prototyping to direct digital manufacturing. Springer Publishing Company,
Incorporated. doi:https://doi.org/10.1007/978-1-4419-1120-9.
Hadi, A., Vignat, F., & Villeneuve, F. (2015). Design configurations and creation of
lattice structures for metallic additive manufacturing. In 14ème Colloque
National AIP PRIMECA, La Plagne, France.
Hirtz, J., Stone, R. B., McAdams, D. A., Szykman, S., & Wood, K. L. (2002). A functional
basis for engineering design: Reconciling and evolving previous efforts. Research
in Engineering Design, 13(2), 65–82. https://doi.org/10.1007/s00163-001-0008-
3.
Jane, C., Sarah, E., Gregory, G., & David, W. R. (2010). A comparison of synthesis
methods for cellular structures with application to additive manufacturing.
Rapid Prototyping Journal, 16(4), 275–283. https://doi.org/10.1108/
13552541011049298.
Joo, J. J., Reich, G. W., & Westfall, J. T. (2009). Flexible skin development for
morphing aircraft applications via topology optimization. Journal of Intelligent
Material Systems and Structures, 20(16), 1969–1985. https://doi.org/10.1177/
1045389x09343026.
Kranz, J., Herzog, D., & Emmelmann, C. (2015). Design guidelines for laser additive
manufacturing of lightweight structures in tial6v4. Journal of Laser Applications,
27(S1), S14001. https://doi.org/10.2351/1.4885235.
Kumke, M., Watschke, H., & Vietor, T. (2016). A new methodological framework for
design for additive manufacturing. Virtual and Physical Prototyping, 11(1), 3–19.
https://doi.org/10.1080/17452759.2016.1139377.
Langelaar, M. (2016). An additive manufacturing filter for topology optimization of
print-ready designs. Structural and Multidisciplinary Optimization, 55(3),
871–883. https://doi.org/10.1007/s00158-016-1522-2.
Leary, M., Merli, L., Torti, F., Mazur, M., & Brandt, M. (2014). Optimal topology for
additive manufacture: A method for enabling additive manufacture of support-
free optimal structures. Materials & Design, 63, 678–690. https://doi.org/
10.1016/j.matdes.2014.06.015.
Lefky, C. S., Zucker, B., Wright, D., Nassar, A. R., Simpson, T. W., & Hildreth, O. J.
(2017). Dissolvable supports in powder bed fusion-printed stainless steel. 3D
Printing and Additive Manufacturing, 4(1), 3–11. https://doi.org/10.1089/
3dp.2016.0043.
Lin, C.-Y., Wirtz, T., LaMarca, F., & Hollister, S. J. (2007). Structural and mechanical
evaluations of a topology optimized titanium interbody fusion cage fabricated
by selective laser melting process. Journal of Biomedical Materials Research Part
A, 83A(2), 272–279. https://doi.org/10.1002/jbm.a.31231.
Lipson, H., Moon, F. C., Hai, J., & Paventi, C. (2004). 3-d printing the history of
mechanisms. Journal of Mechanical Design, 127(5), 1029–1033. https://doi.org/
10.1115/1.1902999.
Liu, J. & To, A. C. (in press). Topology optimization for hybrid additive-subtractive
manufacturing. Structural and Multidisciplinary Optimization 1–19. doi:
https://doi.org/10.1007/s00158-016-1565-4.
17
Maund, A. (2013). 3d printing moving parts fully assembled - 28-geared cube.
<http://www.instructables.com/id/3D-printing-moving-parts-28-Geared-
Cube/>.
Mavroidis, C., DeLaurentis, K. J., Won, J., & Alam, M. (2000). Fabrication of non-
assembly mechanisms and robotic systems using rapid prototyping. Journal of
Mechanical Design, 123(4), 516–524. https://doi.org/10.1115/1.1415034.
Microwaterjet. (2017). Precision water jet cutting - waterjet cutting services.
<http://www.microwaterjet.com/>.
Morgan, H. D., Cherry, J. A., Jonnalaganna, S., Ewing, D., & Sienz, J. (2016). Part
orientation optimisation for the additive layer manufacture of metal
components. The International Journal of Advanced Manufacturing Technology,
86(5), 1679–1687.
Pahl, G., Beitz, W., Feldhusen, J., & Grote, K.-H. (2007). Engineering design: A
systematic approach (3rd ed., Vol. 157). London: Springer-Verlag. https://doi.
org/10.1007/978-1-84628-319-2.
Pandey, P. M., Venkata Reddy, N., & Dhande, S. G. (2007). Part deposition orientation
studies in layered manufacturing. Journal of Materials Processing Technology, 185
(1-3), 125–131.
Paz, R., Monzn, M., Gonzlez, B., Winter, G., & Ortega, F. (2015). Lightweight
optimization for additive manufacturing parts based on genetic algorithms. In
Metamodels and finite element analysis. Computational methods in applied sciences
(Vol. 39, pp. 67–82). Springer International Publishing. https://doi.org/10.1007/
978-3-319-20406-2_5. Ch. 5.
Ponche, R., Hascoet, J. Y., Kerbrat, O., & Mognol, P. (2012). A new global approach to
design for additive manufacturing. Virtual and Physical Prototyping, 7(2),
93–105. https://doi.org/10.1080/17452759.2012.679499.
Ponche, R., Kerbrat, O., Mognol, P., & Hascoet, J.-Y. (2014). A novel methodology of
design for additive manufacturing applied to additive laser manufacturing
process. Robotics and Computer-Integrated Manufacturing, 30(4), 389–398.
https://doi.org/10.1016/j.rcim.2013.12.001.
Primo, T., Calabrese, M., Del Prete, A., & Anglani, A. (2017). Additive manufacturing
integration with topology optimization methodology for innovative product
design. The International Journal of Advanced Manufacturing Technology, 1–13.
https://doi.org/10.1007/s00170-017-0112-9.
Rajagopalan, S., & Cutkosky, M. (2004). Error analysis for the in-situ fabrication of
mechanisms. Journal of Mechanical Design, 125(4), 809–822. https://doi.org/
10.1115/1.1631577.
Rezaie, R., Badrossamay, M., Ghaie, A., & Moosavi, H. (2013). Topology optimization
for fused deposition modeling process. Procedia CIRP, 6, 521–526. https://doi.
org/10.1016/j.procir.2013.03.098.
Rias, A. -l., Bouchard, C., Segonds, F., & Abed, S. (2016). Design for additive
manufacturing: a creative approach. In DS 84: Proceedings of the DESIGN 2016
14th international design conference.
Rodrigue, H. & Rivette, M. (2010). An assembly-level design for additive
manufacturing methodology. In IDMME-virtual concept. doi:https://doi.org/10.
1007/978-2-8178-0169-8. <https://hal.archives-ouvertes.fr/hal-01099485>.
Rozvany, G. I. N. (2009). A critical review of established methods of structural
topology optimization. Structural and Multidisciplinary Optimization, 37(3),
217–237. https://doi.org/10.1007/s00158-007-0217-0.
Schmelzle, J., Kline, E. V., Dickman, C. J., Reutzel, E. W., Jones, G., & Simpson, T. W.
(2015). (re)designing for part consolidation: Understanding the challenges of
metal additive manufacturing. Journal of Mechanical Design, 137(11). https://doi.
org/10.1115/1.4031156. 111404–111404–12.
Shapeways. (2017). From metals to porcelain, plastics to sandstone, and everything
in-between. <https://www.shapeways.com/materials>.
Su, X., Yang, Y., Wang, D., & Chen, Y. (2013). Digital assembly and direct fabrication
of mechanism based on selective laser melting. Rapid Prototyping Journal, 19(3),
166–172. https://doi.org/10.1108/13552541311312157.
Thomas, D. (2009). The development of design rules for selective laser melting, Ph.
D. thesis.
Ullman, D. (2009). The mechanical design process. McGraw-Hill series in mechanical
engineering (4th ed., ). McGraw-Hill Education <https://books.google.fr/books?
id=dQcoAQAAMAAJ>.
Vaezi, M., Chianrabutra, S., Mellor, B., & Yang, S. (2013). Multiple material additive
manufacturing Part 1: A review. Virtual and Physical Prototyping, 8(1), 19–50.
https://doi.org/10.1080/17452759.2013.778175.
Vayre, B., Vignat, F., & Villeneuve, F. (2012). Designing for additive manufacturing.
Procedia CIRP, 3, 632–637. https://doi.org/10.1016/j.procir.2012.07.108.
Vesenjak, M., Krstulovi-Opara, L., Ren, Z., & Domazet, eljko (2010). Cell shape effect
evaluation of polyamide cellular structures. Polymer Testing, 29(8), 991–994.
https://doi.org/10.1016/j.polymertesting.2010.09.001.
Wei, X., Tian, Y., & Joneja, A. (2016). A study on revolute joints in 3d-printed non-
assembly mechanisms. Rapid Prototyping Journal, 22(6), 901–933. https://doi.
org/10.1108/rpj-10-2014-0146.
Whitney, D. E. (2004). Mechanical assemblies their design, manufacture, and role in
product development, Oxford series on advanced manufacturing. New York:
Oxford University Press.
Wilson, J. M., & Shin, Y. C. (2012). Microstructure and wear properties of laser-
deposited functionally graded inconel 690 reinforced with tic. Surface and
Coatings Technology, 207, 517–522. https://doi.org/10.1016/
j.surfcoat.2012.07.058.
Yang, S. & Zhao, Y. F. (2016). Conceptual design for assembly in the context of
additive manufacturing. In Solid freeform fabrication 2016: Proceedings of the
26th annual international.
18 G. Sossou et al. / Journal of Computational Design and Engineering 5 (2018) 3–18

Yang, S., Tang, Y., & Zhao, Y. F. (2015). A new part consolidation method to
embrace the design freedom of additive manufacturing. Journal of
Manufacturing Processes, 20(Part 3), 444–449. https://doi.org/10.1016/j.
jmapro.2015.06.024.
Yang, S., & Zhao, Y. F. (2015). Additive manufacturing-enabled design theory and
methodology: a critical review. The International Journal of Advanced
Manufacturing Technology, 80(1–4), 327–342. https://doi.org/10.1007/s00170-
015-6994-5.
Yan, C., Hao, L., Hussein, A., Young, P., Huang, J., & Zhu, W. (2015). Microstructure
and mechanical properties of aluminium alloy cellular lattice structures
manufactured by direct metal laser sintering. Materials Science and
Engineering: A, 628, 238–246. https://doi.org/10.1016/j.msea.2015.01.063.
Zegard, T., & Paulino, G. H. (2015). Bridging topology optimization and additive
manufacturing. Structural and Multidisciplinary Optimization, 53(1), 175–192.
https://doi.org/10.1007/s00158-015-1274-4.
Zimmer, D., & Adam, G. (2011). Direct manufacturing design rules. CRC Press.