The Adoption of Additive Manufacturing Technology in Sweden

Procedia CIRP 40 ( 2016 ) 7 – 12
Available online at www.sciencedirect.com
2212-8271 © 2016 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the International Scientific Committee of the 13th Global Conference on Sustainable Manufacturing
doi: 10.1016/j.procir.2016.01.036
ScienceDirect
13th Global Conference on Sustainable Manufacturing - Decoupling Growth from Resource Use
The Adoption of Additive Manufacturing Technology in Sweden
Babak Kianian a,b* , Sam Tavassoli c,d , Tobias C. Larsson b and Olaf Diegel a
a Design Sciences Department, Faculty of Engineering (LTH), Lund University, Lund, Sweden
b Mechanical Engineering Department, Blekinge Institute of Technology (BTH), Karlskrona, Sweden
c Centre for Innovation, Research and Competence in the Learning Economy (CIRCLE), Lund University, Lund, Sweden
d Industrial Economics Department, Blekinge Institute of Technology (BTH), Karlskrona, Sweden
* Corresponding author. Tel.: [phone]; E-mail address: [e-mail]
Abstract
This paper analyzes the adoption of Additive Manufacturing (AM) technologies in Sweden. The dataset consists of a recent and representative
sample of Swedish AM users (companies, universities, and research institutes). The authors investigate two questions. Firstly, what are the
current applications of AM in Sweden (e.g. Rapid Prototyping (RP), production)? Secondly, what are the factors that can explain the variation
in AM adoption among the users? Using a regression analysis technique, the main findings are as follows. (i) There is a variation among users’
choice of AM application, and the majority of users are expanding their AM applications beyond RP. (ii) There are two factors that positively
affect the decision of firms to expand classical RP and also incorporate production and management. These two factors are using multiple AM
technologies (as opposed to single Fused Deposition Modeling technology) and being small companies. The authors discuss the implication of
these results.
© 2016 The Authors. Published by Elsevier B.V.
Peer-review under responsibility of the International Scientific Committee of the 13th Global Conference on Sustainable Manufacturing.
Keywords: Additive Manufacturing, 3D Printing, Application, Production Technology, Sweden
1. Introduction
Additive Manufacturing (AM) enables the fabrication of
components in a process, where slices of a virtual model are
created and produced in a layer-upon-layer additive building
process. AM thus differs radically from traditional
manufacturing which is either subtractive, where material is
removed from a block of material, or formative, in which
material is formed by a mold (which, itself, is manufactured
through a subtractive process) [1]. The technology has been
adopted, since the mid-90s by most industries involved in
product development, as it is often the best method for quickly
manufacturing prototypes. Until recently these technologies
were unable to produce components of the same strength and
quality as conventionally manufactured components.
However, some of the latest technologies have advanced to the
point where it is now possible, for certain types of
components, to produce fully functional production
components, in a fraction of the time and material needed by
conventional methods, particularly if one includes
tooling/setup times. Rapid prototyping (RP) has thus evolved
into rapid manufacturing (RM) [2,3]. AM production
capabilities have the potential to reduce the environmental
impact of manufacturing, for example by, production of
lighter, more complex and integrated parts, which require less
raw material usage in their fabrication [4]. Less raw material
usage uses less of earth’s scarce resources, which is a key
sustainable challenge relating to economic growth [4,5]. One
of the sustainable paths to continue on economic growth, via
the lens of circular economy, is decoupling economic growth
from the use of scarce resources through disruptive
technologies (e.g. AM) [5].
It is believed that AM, as a manufacturing technology, is
having an increasing role in many industries and that AM
capabilities now cover various ranges of applications. But,
according to the recent survey by Wohler’s Associates,
company’s benefit from using AM to fabricate functional parts
more than from other applications. The second most popular
application is prototypes for fit and assembly. This indicates
that the main advantage of using AM is in fostering product
© 2016 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the International Scientific Committee of the 13th Global Conference on Sustainable Manufacturing

8 Babak Kianian et al. / Procedia CIRP 40 ( 2016 ) 7 – 12
development processes by improving product quality,
reducing cost (less assembly and tooling etc.), and reducing
time to market etc. The use of AM to fabricate production
parts has also shown a continuous increase since 2003. In
2003 it was valued at 3.9% of the total global product-service
revenues from AM, while this value increased to 42.6% in
2014. Besides this, the global market growth in the part
production segment was 66% in 2014 to an estimated $1.748
billion [3].
International approaches to AM are significantly different
to that of Sweden. Governments in many countries (e.g. USA,
New Zealand, Australia, UK, Germany, Singapore and China)
are heavily investing in AM as a production technology. One
of the biggest challenges currently facing Swedish industry is
a lack of the ability to adopt AM technologies and to deploy
them to their full potential beyond just prototyping [4].
Currently Sweden lies behind its neighbours Germany, the
United Kingdom, France, and Belgium in its AM adoption
rates and is thus lagging behind on its capacity to innovate
and keep up with similar industries in other countries [3,4].
International competitors in many sectors are continuously
investing in AM, and the authors of this paper thus believe
that Sweden needs to strategically reduce the current gap on
the lack of knowledge and implementation of AM. The
authors also believe that the ‘Swedish Agenda for Research
and Innovation Within AM and 3D Printing’, which was
recently coordinated by Umeå University in Sweden is a great
start to that path. According to the agenda, Swedish industry
has adopted and utilized AM comparatively little beyond
prototyping, and the infrastructure level in Swedish
universities are well below the rest of the Europe and
worldwide [4].
The above-mentioned agenda is quite general; particularly
concerning why Sweden is lagging behind in adoption of AM
and also what factors can explain and predict the plausible
move beyond using AM solely in rapid prototyping. This can
be a problem from the perspective of users and policy makers
in Sweden, because the agenda provides little insight on how
to reduce the gap between Sweden and other advanced
countries. This paper, therefore, aims to provide a more
detailed perspective regarding the current status of AM
adoption and utilization in Sweden through investigating two
questions. Firstly, what are the current applications of AM in
Sweden (e.g. rapid prototyping, production)? And secondly,
what are the factors that can explain the variation in AM
applications among the users?
The rest of the paper is as follows: Section 2 provides a
literature review of AM state of art in particular related to AM
applications, AM technology types and raw materials. Section
3 first introduces the research design (e.g. data collection
method), and then specifies the method (empirical strategy),
which is regression analysis. Section 4 provides the results in
two sequences: description of analyses and empirical result.
Section 5 first provides a discussion on the results, and then
concludes and points out suggestions for future research.
2. Literature Review
This section provides a review of literature in order to
identify the additive manufacturing (AM) technology types,
raw materials types, and AM application areas worldwide and
in Sweden.
Innovation in production technology is viewed as a
powerful competitive weapon, which, if an industry adopts
and implements it strategically, can bring about many other
competitive advantages (e.g. superior quality, shorter delivery
cycles, lower inventories, shorter new product development
cycles) [6]. Utilizing innovation in production technologies,
such as AM, can cause improvements in market share. There
are a few cases of research and research and development
(R&D), which propose that the management of the adoption
and implementation of innovative production technologies is a
distinctive area of study in research [7]. For example, in
recent years the emergence of enterprise resource planning
systems (ERP) and radio-frequency identification technology
(RFID) has created extensive investigation and research
articles on their implementation in academia and industries.
The results of this are the creation of many process models
and frameworks to assist decision makers and managers to
implement those new innovations successfully. However, the
research and practice of AM as a production technology has
stayed relatively behind in its exploitation level [8].
Besides the Umeå Agenda, which is addressed in the
introduction, there are relatively few scholars investigating
current AM state of art or practice in Sweden. For example,
Kianian et al 2015, share the same opinion as the agenda,
concerning the level of adoption and deployment of AM in
Sweden, and they also provide a few examples as exceptions
to the lack of AM utilization in Sweden, which are excellent
examples of personalized high-tech manufacturing [9].
The AM applications worldwide are well summarized by
figure 1 below and, as noted earlier in the introduction, AM is
having an increasing role in many industries, and AM
capabilities now cover a wide ranges of applications. The
authors categorize AM applications for this paper’s analysis
into five categories in line with the Wohler’s report
classification in figure 1.
Figure 1. AM applications worldwide. Source: Wohler’s Associates, Inc.
The five categories are Rapid Prototyping (RP) including
general prototyping applications, and patterns for prototyping

9Babak Kianian et al. / Procedia CIRP 40 ( 2016 ) 7 – 12
tooling, pattern for metal castings, fit and assembly;
Production: including functional parts and tooling
components; Management: including visual aids and
presentation models; Research & Development and Education
(R&D and Edu) Including fundamental science and research
work in AM; and Rapid Prototyping & Production and
Management (RP & PROD and MNG) including all the AM
applications used by a single user (see table 1 in section 3.2).
There are many approaches to classifying AM processes,
like categorizing them according to a baseline technology
such as whether the process utilizes UV lasers, extrusion
technology, printer technology, etc. [10,11]. Another way is to
classify AM processes together based on the type of raw
materials (e.g. polymers, metals, ceramics) the system uses
[12]. As Gibson et al 2010 argued it is more appropriate to use
more than one classification method, as there are some issues
with relying merely on one AM process classification [13].
One of the issues, according to Gibson et al 2010, is that some
of the AM processes with similar fabricated parts are placed
in separate categories, and that some processes gather in the
same categories in an odd way. There are some attempts to
address these classification problems through such as the
classification method proposed by Pham 1998, which
categorizes AM processes based on both a technology
baselines and raw material types [13,14]. This paper’s authors
carefully consider these AM processes classification
approaches, and then propose six categories for AM
technology types, and four categories for AM material types.
These classifications are shown in table 1 (Variable names
and descriptions) in section 3.2 (empirical strategy).
The authors in this paper classify the Swedish AM users’
size class in line with the European Union’s definition [15].
Thus, the three categories of small firms (employment range
1-15), Medium firms (employment 16-100), and large firms
(employment range of 101 and more) are identified as
explanatory variables (see table 1 in section 3.2).
3. Method
3.1 Research Design
This paper is based on a quantitative research
methodology. The data comes from a survey of 55 Swedish
users of AM (e.g. companies, universities and research
institutes), which was carried out by 3dp.se between 2013-
2014 [16]. The authors of this paper also add data from 15
additional users, which have not been considered in the
survey. Therefore, a total sample of 70 users is obtained. The
authors conduct regression analysis as an analytical tool in
this paper. Regression analysis is a statistical process for
estimating the relationships among variables [17].
3.2. Empirical Strategy
The main aim of this paper, as noted earlier, is twofold:
First, the paper describes and maps the various applications of
AM that are actually chosen and utilized by the users
(companies and research institutions), and second, the paper
analyzes the determinants of such choices. The first part of the
aim is descriptive in nature and the authors will explain this
part in Section 4.1. The second part of the aim is analytical in
nature and regression analysis is used as the analytical tool.
Regression analysis is commonly utilized in social science
and medicine to analyze the effect of some potential
explanatory factors on the phenomenon of interest [17]. The
phenomenon of interest in this study is the choice of
application of AM by users. The authors employ a
multinomial logit model in order to investigate the
determinants of various AM applications choices by users.
Formally speaking, the probability that user i chooses AM
application j is given by:























































































Where

is the vector of explanatory variables,

are a set
of unknown parameters per each explanatory variable,
capturing the effect of each explanatory variables on the
probabilities of choosing each AM applications, and j is the
AM applications that users decide to choose. Starting from the
vector

, it is composed of several explanatory variables,
which, as discussed in the literature review section, are
expected to have significant (positive or negative) effect on
the choice of AM application. Specifically, these explanatory
variables are as follows: the AM technology types that the
users deploy, raw material types, amount of AM machine
investments, size (number of employees) of the user, and the
location of the user (for the exact definitions, see Table 1). As
noted,

are unknown parameters but they can be estimated
through a procedure called maximum likelihood estimation,
which is the output of regression analysis. And finally, ” j”
can be coded in two ways. First, in a dichotomous way, i.e. it
gets value j=0 when user i decides to use AM only and
exclusively for the rapid prototyping (RP) purposes, and it
gets value j=1 when user i decides to use AM not exclusively
for RP. Second, in an extend version, while j=0 is still the
same as dichotomous way, j=1 can be expanded as follows:
(the new) j=1 is when firms decide to still have the AM for
RP purposes but also utilize AM for production and
management purposes, j=2 is when firms decide to use AM
only in Production, j=3 is when firms decide to use AM only
for management purposes (e.g. visual aids, presentation
models), and finally j=4 is when firms decide to use AM only
in research & development and education. This implies that in
the regression analysis, there will be two alternative
dependent variables. Both of them are measuring the
phenomenon of interest, i.e. application of AM among users,
but just in two different ways. Moreover, in either of
dichotomous or extended version, the j=0 is considered to be
the “base” category (choice). This means the result of the
estimation for each j should be interpreted with reference to
this base category. It should be noted that, multinomial logit