A design methodology for parts using additive

manufacturing
Nicolas Boyard, Mickaël Rivette, Olivier Christmann, Simon Richir

To cite this version:

Nicolas Boyard, Mickaël Rivette, Olivier Christmann, Simon Richir. A design methodology for parts
using additive manufacturing. International Conference on Advanced Research in Virtual and Rapid
Prototyping (VRAP), Oct 2013, Leiria, Portugal. pp.0-6. �hal-01197463�

HAL Id: hal-01197463

https://hal.science/hal-01197463
Submitted on 11 Sep 2015

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 Science Arts & Métiers (SAM)
is an open access repository that collects the work of Arts et Métiers ParisTech
researchers and makes it freely available over the web where possible.

This is an author-deposited version published in: http://sam.ensam.eu

Handle ID: .http://hdl.handle.net/10985/9998

To cite this version :

Nicolas BOYARD, Mickaël RIVETTE, Olivier CHRISTMANN, Simon RICHIR - A design
methodology for parts using additive manufacturing - In: International Conference on Advanced
Research in Virtual and Rapid Prototyping (VRAP), Portugal, 2013-10-01 - High Value
Manufacturing: Advanced Research in Virtual and Rapid Prototyping: Proceedings of the 6th
International Conference on Advanced Research in Virtual and Rapid Prototyping, Leiria,
Portugal, 1-5 October, 2013 - 2013

Any correspondence concerning this service should be sent to the repository

Administrator : [email protected]
A design methodology for parts using additive manufacturing
N. Boyard, M. Rivette, O. Christmann & S. Richir
Arts et Métiers ParisTech, LAMPA, Angers, France

ABSTRACT: Additive Manufacturing (AM) allows designer to sidestep several design requirements and to
create free forms, hollow parts or direct assemblies. This process also allows direct recycling of plastic into
new parts, which eases the raw material supply. However, although several methodologies are used to redes-
ign products and parts, none is dedicated to a real design of parts and products in AM. At first, we will sug-
gest the base of a new design methodology for an end-user who wants to create a product or a part in AM.
Then, we will show an example of using our methodology. Finally, we will conclude on the limits of this
methodology and on our next work to validate our methodology.

1 INTRODUCTION However, although various redesign methodolo-

gies have been proposed (Rodrigue 2010; Vayre &
Additive manufacturing (AM) is a very promising Villeneuve 2012), no design methodology has been
manufacturing process in terms of possibilities of developed specifically for AM.
shapes and complexity of parts (Vayre & Villeneuve Rodrigue (2010) reveal two design methodologies
2012), (Boyard et al. 2012). It is possible to perform that may relate to the AM: The Design For Assem-
multi-material parts (Muller et al. 2010), hollow bly (DFA), which aims to design a product in order
parts (Vayre & Villeneuve 2012), large-scale parts to validate and facilitate assembly, and the Design
(Lim et al. 2012) and ready-for-use parts (Cooper et For Manufacturing (DFM), whose goal is to design
al. 2012). workable parts using one or more specific proc-
From an ecological point of view, AM also al- ess(es) (Boothroyd 1994). In the case of AM, the
lows the recycling of certain materials. For example, geometry of the part is almost no longer constrained
Fused Deposition Modelling (FDM), which is a by the manufacturing process. To optimize the parts
method to construct a part from hot wire deposition, of the product with respect to the assembly and
allows reuse of certain plastics such as acrylonitrile manufacturing, DFM and DFA can be performed di-
butadiene styrene (ABS). rectly in the design without generating additional
However, to design a product, it is necessary to constraints or changes in the initial request of the
adopt a precise and consistent methodology end-user. Rodrigue (2010) also offers a redesign
(Segonds 2011). Currently there are many methods methodology (Figure 2).
whose relevance varies with the design goals: as-
sembly, machining, cost, quality… (Tomiyama
2009). Among all these methods, there is a recurrent
pattern (Figure 1).

Figure 2 : Redesign methodology for AM (Rodrigue 2011)

Figure 1 : Standard schema of a design methodology (Se-
gonds 2011)
 The resulting model is then optimized in two dif- Our method includes the six steps of a standard
ferent directions: preventing failures and respect for design methodology:
user requirements. Prevention of failures is based on - Functional specification (1) ;
FMECA (Failure Modes, Effects and Criticality - Conceptual design (2) ;
Analysis). It aims to optimize the product to increase - Architectural design (3) ;
its reliability to meet the specifications. Compliance - Detailed design (4) ;
with user constraints aims to improve the design of - Implementation (5).
each part to meet the design constraints with mini- Phases 3 and 4, corresponding to loops of DFM
mal compromise. The final purpose is to meet end- and DFA, are performed in parallel.
user’s needs as accurately as possible. Finally, the The framework of our research regards points 3
results of these optimizations are confronted to de- to 5 of this methodology. We will assume that the
cide the structure and shape of the final product. work relating to the collection of needs and the
This paper aims to establish the basis of a design planning tasks have already been made. In addition,
methodology that can take into account the specifici- we will not deal the implementation phase in its en-
ties of the AM. It doesn’t deal with innovation tirety. These two steps are necessary for any design
methodologies. In this paper, we deal only with the methodology and do not participate in the design it-
design of products which does not contain any inter- self. We also will assume that our methodology is
nal relative movement during use. These will be in- consistent from the moment the designer is able to
tegrated in future developments dedicated to opti- produce a manufacturable digital mock-up, corre-
mize, test and then assess our method. sponding to a prototype or to the finished product.
One of the difficulties in product design lies in
the association of manufacturing and assembly with
2 A MORE COMPLETE MEHODOLOGY respect to the functional specifications. In our meth-
odology, we propose to keep these two types of de-
sign. But rather than to perform them separately and
2.1 General overview
at different times, we try here to make them com-
In the following sections, referring to our state of plementary and simultaneous.
the art above, we present the fundamental keys to
operationalize our methodology (Figure 3).

Figure 3 : Design methodology for Rapid Manufacturing (DFRM, Design for Rapid Manufacturing)
 Since AM can manufacture any type of mechani-
In addition, the democratization of the AM cal part, we assume that for any set of functions
(Anderson 2012) and the setting up of co-design there is at least one piece that meets all the functions
phases in companies (phases with all stakeholders of this set. A part is defined here as a unitary physi-
concerned with the whole product design: engineer- cal body.
ing, stylists, end-user, designer ...) implies to de- The rules for establishing a set are:
velop a methodology that can be used by any person - Is the part corresponding to this set a wear
without any prerequisites. part?
- Can functions be grouped on the same part, or
should they be separated?
2.2 Conceptual design - Are parts movable relative to each other?
The conceptual design phase (Figure 3, step 2), - ...
which is the real starting point of our research, aims A set of functions is a collection of functions
to take each function and constraint of the functional connected by links. A set must contain at least one
specifications. function. If all functions of the functional specifica-
However, changes from the client occur fre- tions are interconnected to others, then the graph
quently throughout the product design. This is why will contain only one set. Sets are interconnected by
the extraction of functions from the functional speci- dotted line representing a fixed joint between two
fications should be flexible and modular to allow of parts.
new functions and constraints. Each function is represented as a sphere. The
Once the features extracted, they will be arranged spheres are the nodes of the graph. The functions are
in a graph of functions by the actors of the design in then linked to each other by segments. These seg-
order to provide a first iteration of the architecture of ments represent both direct connections between
the final product. The first proposal is based on a functions and spatial organization of the functions
simple set of rules that we need to define. Some of with each other. Thus, different interconnected func-
these rules concern, for example, the geometric ap- tions belong to the same part. In addition, a function
pearance. Thus a function requiring the holding in A connected to a function B itself connected to a
position of a part will be disposed downward, function C indicates that the function B will be
whereas a function requiring fast access to a part of found between or will separate functions A and C
the product is preferably placed on an outer sur- (Figure 4).
face... Similarly, the graph will be able to integrate This representation allows the user of the meth-
new functions as and when the changes of the client. odology to spatially reorganize functions with each
This graph is a modular three-dimensional graph other. The advantage is, without conducting discus-
to be drawn in CAD software and then to provide sions of technical solutions, to begin to propose ar-
support to the digital mock-up (Figure 4). chitecture of the final product based only on the
functions and constraints to be addressed.
Moreover, this graph is modifiable: each function
can be reorganized with respect to each other and
connections can be changed. This has two advan-
tages: the first is to provide a modular basis for dis-
cussion with stakeholders of the design, the second
is to allow stakeholders to easily review the results
of the conceptual design.

2.3 Detailed design

The detailed design phase, which analyzes each set
of functions present in the global graph of functions,
corresponds to a DFM phase (Figure 3, Step 4).
Figure 4 : Graph of functions Firstly, all the constraints that the part will sup-
port are added to the set of this part. These con-
The graph is a global representation of the prod- straints can be mechanical, geometrical, physical, of
uct. As stated earlier, our methodology concerns wear, ecological...
only products containing no internal movement in Constraints are layered on geometry proposed by
functioning. However, these products can be com- the global graph of functions.
posed of several parts. We therefore decompose the In a first step, the CAD software will compare to
graph into several sets, each set representing a part a database the set corresponding to the part. This da-
of the product. tabase contains all the graphs of functions already
designed by the design team and all the associated 2.5 Co-design
3D models. It can be mutualised with other design Another purpose of this methodology is to allow the
teams. This corresponds to the state of the art step. co-design of the product, involving all the stake-
Three alternatives are possible: either the proposed
holders of design (end users, designers ...) at every
graph already exists and a piece is proposed. Other- moment of the design process.
wise, if a variant of the proposed graph (dependent Thus, if parts of the detailed design and the archi-
on a minimum number of common functions) exists, tectural design are outsourced to CAD software for
it can serve as a basis for the design of the part. Ei- the searching of existing solutions, most of the de-
ther there is no corresponding or similar graph, the sign is a result of subjective decisions of the stake-
part has to be fully designed. holders, especially those of end-users. This is why
In the last two cases, the stakeholders will detail the co-design phase - including the conceptual, ar-
the geometry of the part and propose a combination chitectural and detailed design - should allow adding
of one or more materials to satisfy the constraints some constraints of appreciation:
and functions imposed. - Physical (textures, distribution and percep-
Finally, designers should specify an AM ma- tion of product weight);
chine. The geometry of the part will be confronted - Mechanical (flexibility, elasticity);
with manufacturing capabilities of the AM machine
- Geometric (grip, aesthetic);
and assembly constraints with other parts. - Architecture (type of mobile connection, dis-
If the part does not meet these criteria, the stake- assembly);
holders will then optimize part by adjusting the fol- - …
lowing parameters: part geometry, materials, func-
tions of to the part.
If the part is valid, it is locked in the graph. If a 2.6 Deliverable
function is added or removed to a part (which could The purpose of our design methodology is to allow
be validated or not), it must again validate the de- the user to manufacture his “own” product.
tailed design phase. Once the product has been approved by all the
stakeholders and the software no longer requires
2.4 Architectural design modification, parts resulting of the design will be
post-processed by the CAD software to convert digi-
tal files parts in a machine readable code.
The architectural design phase, or embodiment de- This post-treatment has been validated during the
sign phase, is realized in parallel of the detailed de- detailed design phase. Indeed, at this stage, each part
sign phase (figure 3, step 3). Thanks to the modular- was confronted with the characteristics of the ma-
ity of the graph, the structure of the product may be chine which will manufacture them. The notions of
modified whenever. surface quality, material manufacturing cost, ma-
Each time a set is created or each time a part is chinability, manufacturing strategy and precision are
validated during the detailed design, the architecture thus validated.
of the product must also be re-validated. Once the manufacturing strategy has been vali-
To proceed, as in the detailed design phase, the dated by stakeholders, the code is generated and the
CAD software compare the entirely graph (and not a part is ready to be manufactured.
specific set) with the models of products of a data-
base. This comparison could provide existing prod-
ucts which give a model for the structure of the
3 CASE STUDY: THE SALTCELLAR
product and for the connections between the sets.
This is particularly helpful to design a range of To illustrate our methodology, we propose a sim-
products or for mass customization (Zarb, F. 2012). ple case study: the saltcellar.
Otherwise, either an existing product may be fit, For this, we use a simplified functional specifica-
or, if there is no existing similar product, the archi- tion including the following nine functions:
tecture is to be entirely designed.
- F1: Contain condiments;
Once the architecture of the product is designed,
it is faced to the assembly constraints (removable or - F2: Can be filled;
irremovable parts, accessibility to parts, steps of as- - F3: Distribute the content;
sembly/disassembly…). If the product fits the con- - F4: Protect condiments in case of fall;
straints, it is validated. Else, the software indicates - F5: Indicates type of contained condiments;
that the architecture has to be redesigned, which may - F6: Indicates quantity of contained condi-
lead to the transfer of one function from a set to an- ments;
other or to the internal reorganization of a set. - F7: Do not slide;
- F8: Hold on a table;
- F9: Being comfortable in hand.
 The first task of the design stakeholders will be to
establish a graph of the functions.
We will consider that the result of the discussion Function F3 « Distribute the content » is only
between the stakeholders, according to the rules of contained in the first set. However, function F2
designed we will define, leads to a graph of two sets. « Can be filled » is contained by both set 1 and set 2
The first one contains the functions F2 and F3. The and links them. Indeed, if the final product fits to the
second one contains the other functions (Figure 5). functional specifications, it is possible to duplicate a
function, even it is redundant. In our case, this corre-
sponds to the fact that the screwtop is screwed to the
body of the saltcellar when used and unscrewed
when filled: there is a relative movement only dur-
ing the filling step. This matches with our methodol-
ogy: when used; the screwtop is related to the body
of the saltcellar by a rigid joint.
To fit exactly with the graph deduced from the
functional specifications, the graph presented on the
figure 6 may be modified and adapted.
Either functions F10 and F11 may be deleted re-
spectively from set 1 and set 2, or the graph pre-
sented on the figure 5 may include these functions
(only if the end-user agree this modification of the
functional specifications). Only the stakeholders of
Figure 5 : Graph of functions of the saltcellar the design can make this choice.
Nevertheless, to fit the graph, it is necessary that
every other function is satisfied. This involves the
Then, the CAD software compares this graph to step of constraint of machinability from the DFM
those which are contained in the database (functions loop (Figure 3, step 4). According to the machine
and structure). An approaching solution is the salt- which will be used, stakeholders may have the pos-
cellar with a metallic screwtop (Figure 6) which is a sibility to produce the body of the saltcellar in an en-
product also composed by two sets (Figure 7) which tirely transparent material or to produce it in multi-
includes two other functions: material, at least one of them is transparent.
- F10: Limit the flow when distributing (set 1); Once the graph is validated, stakeholders will be
- F11: Be washable (set 2). able to modify the position of the functions inside
the sets to start to create the geometry of the parts.
In our case, since the software proposed an exist-
ing solution, it is possible to base on the existing
numerical model of this solution to design the ge-
ometry of the parts. In every case, stakeholders will
have to design the parts on the CAD software which
allows the use of the graph of functions.
Here, the first set corresponds to the screwtop.
The second corresponds to the body of the saltcellar.
By adding constraints on the sets, stakeholders
Figure 6 : Saltcellar with metallic cap will be able to design step by step the final geometry
of the product and assign materials to the parts. In-
deed, geometry and material must answer to the
functional requirements. This will give the cost of
raw material of the product.
In the case of imposed external parts, like an elec-
tric motor-reducer, stakeholders have to know the
exact geometry of these parts. They must have, at
least, sharp diagrams to redesign and include these
parts in the mock-up based on the graph in order to
create the geometry of the product.
In our case, the body of the saltcellar may be
made in transparent glass, to allow the end-user see
the type and the remaining quantity of contained
condiment. More, a glass saltcellar is easily wash-
Figure 7 : Graph of function of the saltcellar with screwtop able. The type of glass and the geometry of the body
have to fit functions F4 « Protect condiments in case cess to cutting tool, lubrication, tool wear...). It also
of fall », F7 « Do not slide », F8 « Hold on a table » allows the design of parts satisfying both DFA and
and F9 « Being comfortable in hand ». These DFM in the earliest phase early of the design. This
choices are delegated to the stakeholders. However, possibility will limit a priori costly late changes.
it is not possible to add a stick on the saltcellar to in- Finally, this methodology falls within the scope
dicate the type of condiment. Otherwise, the graph of environmental concerns by providing, for certain
would have been composed by three sets: one for the processes, the ability to produce parts by using only
body, one for the screwtop and one for the stick. the amount of material necessary to the final volume
As one goes along the design, the software veri- (Boyard et al. 2012).
fies the respect of assembly and machinability con-
straints, indicating the possible mistakes or incom-
patibility. In this way, the mock-up resulting of the 6 REFERENCES
design is ready to be converted in a file which can be
read by the additive manufacturing machines chosen Boothroyd, G. 1994. Product design for manufacture and as-
to manufacture the product. sembly. Computer-Aided Design 26(7): 505-520.
Boyard, N. & Rivette, M. & Christmann, O. & Richir, S. 2012.
Méthodologie de prise en compte de la Fabrication Addi-
tive lors de la phase de conception. Proc. AIP PRIMECA.
4 LIMITS Le Mont Dore: France.
Cooper, D. & Stanford, M. & Kibble, K. & Gibbons, G. 2012,
The methodology presented in this article is the sub- Additive manufacturing for product improvement at Red
ject of a thesis, which can explain why some points Bull Technology. Materials and design 41: 226-230.
Hascoet, J.Y. & Ponche, R. & Kerbrat, O. & Mognol P. 2011.
have not yet been tested. Thus, the overall method- From functional specifications to optimized CAD model:
ology remains to be validated in an experiment in- proposition of a new DFAM methodology. Proc. 5th Con-
volving different stakeholders (ergonomists, engi- ference on Advanced Research in Virtual and Rapid Proto-
neers, manufacturers, designers, end users...). typing, Lieira: Portugal.
Moreover, it is difficult to estimate the cost of Kadri, A. 2007. Contribution de la réalité virtuelle à l'évalua-
some developments which have to be done on the tion de produits, dans les phases amonts du processus de
conception. PhD thesis. Université d’Angers: France.
CAD software. If the technical feasibility of our Kou, X. 2005, Computer-aided design of heterogeneous ob-
methodology has been validated through the use of jects, PhD thesis, HKU, China.
SolidWorks software (Kou X. 2005), which allows Lim, S. & Buswell, R.A. & Le, T. & Austin, S.A. & Gibb,
the addition of features such as the integration of a A.G.F., Thorpe T. 2012. Developments in construction-
comparative database, the length and complexity of scale additive manufacturing processes, Automation in con-
these developments has not yet been quantified. struction 21: 262-268.
Muller, P. & Mognol, P. & Hascoet, J.-Y. 2010. Modeling and
Thus, the methodology proposed in this paper is control of a direct laser powder deposition process for
still at an ideal level. However, it can make a precise Functionally Graded Materials (FGM) parts manufacturing.
statement of requirements and necessary develop- Journal of Materials Processing Technology 213(5): 685-
ments. These are all avenues of research to explore. 692. Elsevier.
Another point to develop is how stakeholders will Rodrigue, H. 2010. Méthodologie de conception et
d’optimisation de mécanismes fabriqués par fabrication ra-
interact with to the function graph. There are differ- pide. Master thesis. Ecole Polytechnique de Montréal: Ca-
ent technologies for multiuser interaction (Kadri nada.
2007). However, the challenge is to propose a suit- Segonds, F. 2011. Contribution à l'intégration d'un environne-
able solution for all stakeholders, for their vision of ment collaboratif en conception amont de produit. PhD the-
the design vision, their vocabulary or to their field of sis. LCPI & LSIS, ENSAM. France.
competence (Segonds 2011). Tomiyama, T. & Gu, P. & Jin, Y. & Lutters, D. & Kind, Ch. &
Kimura, F. 2009. CIRP Annals - Manufacturing Technolo-
gy 58(2): 543-565. Elsevier.
Vayre, B. & Vignat, F. & Villeneuve, F. 2012. Designing for
5 CONCLUSION Additive Manufacturing. Proc. 45th CIRP Conference on
Manufacturing Systems. Athens : Greece.
The methodology we propose will therefore allow to Vayre, B. & Vignat, F. & Villeneuve, F. 2012. Metallic addi-
tive manufacturing: state-of-the-art review and prospects.
design parts made using AM, with the most accurate Proc. AIP PRIMECA. Le Mont Dore: France.
respect of the functional specifications. Zarb, F. 2012. 3D printing: The new industrial revolution,
More, it differs from other methodologies (Has- Business Horizons 55: 156-162.
coet et al. 2011) by the use of the graph of function
which transcribes the will of designing the geometry
of the product the later as possible.
In addition, this methodology allows the design
of parts with complex geometries by abstracting a
large number of manufacturing constraints associ-
ated with conventional methods (cutting angles, ac-