Available online at www.sciencedirect.

com
Available online at www.sciencedirect.com
ScienceDirect
Available ScienceDirect
online
Available atonline
Procedia CIRP
www.sciencedirect.com
at www.sciencedirect.com
00 (2019) 000–000
www.elsevier.com/locate/procedia
ScienceDirect
Procedia CIRP 00 (2019) 000–000
ScienceDirect www.elsevier.com/locate/procedia

Procedia CIRP 00 (2017)

Procedia CIRP 000–000
81 (2019) 28–33
www.elsevier.com/locate/procedia
52nd CIRP Conference on Manufacturing Systems
52nd CIRP Conference on Manufacturing Systems
Development of a simulation tool for predicting energy consumption of
Development selectiveof28th
a simulation
laser melting
CIRP Design toolbyfor
Conference, predicting
using MATLAB/Simulink
May 2018, energy
Nantes, Franceconsumption of
selective laser a,melting by using MATLAB/Simulink
A new methodology to Li analyze
Yi *, Bahram the Ravani
functional
b, Jan C.and Aurich physical
a architecture of
existing products
Institute for Manufacturing
a forandLian
Technology Yiassembly
Production Systems (FBK)oriented
a, *, Bahram Ravanib, Jan C. Auricha
product
of TU Kaiserslautern, family identification
Gottlieb-Daimler-Str.D-67663, Kaiserslautern, Germany
bDepartment of Mechanical and Aerospace Engineering of UC Davis, 1013 Academic Surge, One Shields Avenue, Davis CA 95616, USA
aInstitute for Manufacturing Technology and Production Systems (FBK) of TU Kaiserslautern, Gottlieb-Daimler-Str.D-67663, Kaiserslautern, Germany

* Corresponding author. Tel.:

bDepartment of MechanicalPaul Stief *, Jean-Yves Dantan, Alain Etienne, Ali Siadat
+49-631-205-3369; fax: +49-631-205-3304.
and Aerospace Engineering of UCE-mail
Davis, address: [email protected]
1013 Academic Surge, One Shields Avenue, Davis CA 95616, USA

* Corresponding author. Tel.: +49-631-205-3369; fax: +49-631-205-3304. E-mail address: [email protected]

École Nationale Supérieure d’Arts et Métiers, Arts et Métiers ParisTech, LCFC EA 4495, 4 Rue Augustin Fresnel, Metz 57078, France

Abstract
* Corresponding author. Tel.: +33 3 87 37 54 30; E-mail address: [email protected]
Abstract
Additive manufacturing (AM) is the umbrella term for manufacturing processes that apply materials layer by layer to create 3D-parts. This
technology implies many revolutionary benefits. Selective laser melting (SLM) is one promising metal AM technology which uses a laser beam
Additive manufacturing (AM) is the umbrella term for manufacturing processes that apply materials layer by layer to create 3D-parts. This
to melt material on the selective area of a powder bed for creating parts. In order to ensure the sustainable value of SLM, the energy consumption
Abstract
technology implies many revolutionary benefits. Selective laser melting (SLM) is one promising metal AM technology which uses a laser beam
for manufacturing parts with SLM should be analyzed in the design phase. Addressing this issue, this paper introduces the development of a
to melt material on the selective area of a powder bed for creating parts. In order to ensure the sustainable value of SLM, the energy consumption
®/Simulink®. The development approach includes 5 phases:
Insimulation
fortoday’s
tool for environment,
business
manufacturing
energy consumption
parts with SLM trendprediction
theshould towards with SLM
more
be analyzed inproduct
by variety
the design
using MATLAB
and Addressing
phase. customizationthisisissue,
unbroken. Due to
this paper this development,
introduces the needofofa
the development
data and
agile acquisition, definition
reconfigurable of systemsystems
production parameters, modeling
emerged to cope using bond graphs, creation
andofproduct
Simulink ® models, and creation of data structure and
simulation tool for energy consumption prediction with SLM with various
by using products
MATLAB ®/Simulink families.
®. The To design
development and optimize
approach includesproduction
5 phases:
GUI. The
systems as result of to
well as thischoose
work shows that the simulation tool product
is a useful tool formethods
analyzing and evaluating themost
energy consumption of SLMaimin the
data acquisition, definition ofthe optimal
system product
parameters, matches,
modeling using bondanalysis
graphs, creationareof needed.
SimulinkIndeed,
® models, andofcreation
the known methods
of data to
structure and
design
analyze phase.
a product orthis
onework
product family
GUI. The result of shows thaton
thethesimulation
physical level.
tool isDifferent
a useful product families, however,
tool for analyzing may differ
and evaluating largelyconsumption
the energy in terms of the
ofnumber
SLM inandthe
nature
designofphase.
components. This fact impedes an efficient comparison and choice of appropriate product family combinations for the production
© 2019AThe
system. new Authors. Published
methodology by Elsevier
is proposed Ltd. This
to analyze is an open
existing access
products in article
view ofunder
theirthe CC BY-NC-ND
functional license
and physical architecture. The aim is to cluster
© 2019 The Authors. Published by Elsevier Ltd.
(http://creativecommons.org/licenses/by-nc-nd/3.0/)
these
© products
2019 The in new
Authors. assembly
Published oriented product
by Elsevier families for the
Ltd. This is license optimization of existing assembly lines and
an open(http://creativecommons.org/licenses/by-nc-nd/3.0/)
access article under the CC BY-NC-ND license the creation of future reconfigurable
This is an
Peer-review open access
underBased article
responsibility under the
of Flow CC BY-NC-ND
the scientific committee
assembly systems. on Datum
(http://creativecommons.org/licenses/by-nc-nd/3.0/)
Peer-review under responsibility Chain, the
of the scientific physicalof
committee
the 52ndofCIRP
ofstructure
the 52nd CIRP
Conference
the products on Manufacturing
is analyzed.
Conference
Systems.
FunctionalSystems.
on Manufacturing subassemblies are identified, and
aPeer-review
functional analysis is performed.
under responsibility of Moreover,
the scientifica hybrid
committeefunctional and physical
of the 52nd architecture
CIRP Conference on graph (HyFPAG)
Manufacturing is the output which depicts the
Systems.
Keywords: additive manufacturing; selective laser melting; energy simulation; MATLAB/Simulink
similarity between product families by providing design support to both, production system planners and product designers. An illustrative
example
Keywords:of additive
a nail-clipper is used selective
manufacturing; to explain themelting;
laser proposed methodology.
energy An industrial case study on two product families of steering columns of
simulation; MATLAB/Simulink
thyssenkrupp Presta France is then carried out to give a first industrial evaluation of the proposed approach.
© 2017 The Authors. Published by Elsevier B.V.
1. Introduction layers to 3D-parts [3]. SLM was considered as a solution to
Peer-review under responsibility of the scientific committee of the 28th CIRP Design Conference 2018.
1. Introduction reduce to
layers energy
3D-partsand material
[3]. SLM consumptions
was considered in theasmanufacturing
a solution to
In the past decades, many new manufacturing
Keywords: Assembly; Design method; Family identification technologies context.
reduce energy and material consumptions in the of
According to [4], the sustainable value SLM can be
manufacturing
using the mechanism of adding materials
In the past decades, many new manufacturing technologies layer by layer to regarded from the following three perspectives:
context. According to [4], the sustainable value of SLM can be improvement
create the
using 3D-parts
mechanismwere ofinvented
adding and summarized
materials layer byunder layerthe to of engineering
regarded from the performance
following for threedesigning novelimprovement
perspectives: geometries;
umbrella term additive manufacturing
create 3D-parts were invented and summarized under the(AM) [1]. During this reduction of “cradle-to-gate” environmental
of engineering performance for designing novel geometries; footprints by
1.initial
Introduction
phase, AM was mainly used to manufacture non- of the
avoiding product
tools,range
dies andor characteristics
material manufactured
scraps;
reduction of “cradle-to-gate” environmental footprints by enhancement and/or of
umbrella term additive manufacturing (AM) [1]. During this
structural prototypes with polymers
initial phase, AM was mainly used to manufacture non- to demonstrate the assembled
functional in
andthis system.
environmental In this context,
performance
avoiding tools, dies or material scraps; enhancement of the main
of challenge
products in in
use
Due to
potential
structural the fastfeatures
geometrical
prototypes development
with of a product.
polymers intoThrough
the domain
advancing
demonstrate of
the modelling
phase.
functional and
and analysis
Nevertheless,
environmental is nowperformance
current not onlyalso
research toofcope
shows withthat
products single
the
in use
communication
technologies, AM and
can an
now ongoing
be used trend
for of digitization
manufacturing
potential geometrical features of a product. Through advancing complexand products,
promised a limited
sustainable productvalue rangeof or existing
SLM
phase. Nevertheless, current research also shows that the shouldproduct
be families,
critically
digitalization,
metal manufacturing
parts delivering
technologies, AM beenterprises
functional
can now usedperformance.are facing
for manufacturing Forimportant
complex but
example, also to be
regarded.
promised For able to analyze
example,
sustainable [5] and
value oftoSLM
reported compare products
thatshould
the be to
specific define
energy
critically
challenges
Boeing has in today’s
manufactured market
and environments:
applied tens
metal parts delivering functional performance. For example, of a continuing
thousands of new product
consumptions families.
of AM It can
processes be observed
including that
regarded. For example, [5] reported that the specific energySLM classical
are 1 to existing
2 orders
tendency
unique towards
aircraft reduction
parts with AM of product
[2]. development
Boeing has manufactured and applied tens of thousands of times and product
of families
magnitude are
higher regrouped
than in function
conventional of clients
machining
consumptions of AM processes including SLM are 1 to 2 orders or
andfeatures.
casting
shortened
Selective
unique product
laser
aircraft lifecycles.
melting
parts with AM In[2].
(SLM) addition,
is one of therethe ismost
an increasing
promising However, processes.
of magnitude assembly
[6]higher oriented
studiedthan the product
energyfamilies
conventional are hardly
consumptions
machining andofto find.
AM
casting
demand
AMSelective of customization,
technologies for being
manufacturingat the same
metal time
laser melting (SLM) is one of the most promisingpartsin a
and global
it is On
processesthe product
including family
SLM, level,
and products
the
processes. [6] studied the energy consumptions of AM differ
calculated mainly
adiabatic in two
rate
competition
characterized with
by competitors
using a laserall over
beam tothe
AM technologies for manufacturing metal parts and it isscan world.
the This
surface trend,
of a main characteristics:
efficiencies of (i)
different the
SLM number of
machines components
processes including SLM, and the calculated adiabatic rate for steel and
powders(ii) the
are
which
powder is inducing
bed the
selectively, development
fusing the powdersfrom to
characterized by using a laser beam to scan the surface of a macro
layers to
and micro
adding concentrated
type of components between (e.g. 9% and
mechanical, 23%. [7]
electrical,
efficiencies of different SLM machines for steel powders are estimated
electronical). that
markets,
powder bed results in diminished
selectively, fusing thelotpowders
sizes due to augmenting
to layers and adding Classical methodologies
concentrated between 9%considering and 23%. mainly
[7] single
estimatedproductsthat
product varieties (high-volume to low-volume production) [1]. or solitary, already existing product families analyze the
2212-8271 © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
To cope with this augmenting variety as well as to be able to product structure on a physical level (components level) which
(http://creativecommons.org/licenses/by-nc-nd/3.0/)
identify
2212-8271 possible
© 2019 The optimization
Authors. Publishedpotentials
by Elsevier in
Ltd. the
This is existing
an open access
Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference causes on difficulties
article under the CC BY-NC-ND
Manufacturing regarding
Systems. license an efficient definition and
(http://creativecommons.org/licenses/by-nc-nd/3.0/)
production system, it is important to have a precise knowledge comparison of different product families. Addressing this
Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing Systems.
2212-8271 © 2019 The Authors. Published by Elsevier Ltd.
This is an©open
2212-8271 2017access article Published
The Authors. under theby CC BY-NC-ND
Elsevier B.V. license (http://creativecommons.org/licenses/by-nc-nd/3.0/)
Peer-review
Peer-review under
under responsibility
responsibility of scientific
of the the scientific committee
committee of the of theCIRP
28th 52ndDesign
CIRPConference
Conference2018.
on Manufacturing Systems.
10.1016/j.procir.2019.03.006
 Li Yi et al. / Procedia CIRP 81 (2019) 28–33 29
2 Li Yi et al. / Procedia CIRP 00 (2019) 000–000

approximately 23.8 MJ/kg energy consumption is required for

the production of Ti6AlV metal powders. All these studies lead
to the conclusion that it is not an easy question to answer as to
whether the energy performance of SLM is better than
conventional processes. The answer depends on many different
process parameters like number of parts in a build job, laser
power and scanning speed, or desired bulk density. If users
want to ensure a cleaner production using SLM with less energy
and material consumption, they will need to explore and ensure
the sustainable values of SLM in the design stage for their own
cases. At current research status, most energy-related works on
the machine level applied experimental investigations, as will
be introduced in section 2. The simulation-based approaches
are mainly used on the process level but not machine level for
analyzing the thermal properties and temperature gradients of
powder beds. The development of simulation approaches and
tools on the machine level for quantifying the energy Fig. 1. Overview of the research approach (machine picture source [8]).
consumption of SLM is currently an emerging research task.
Aiming at the research issue mentioned above, this paper scientific simulation approaches. With MATLAB®/Simulink®,
introduces the development of a simulation tool by using we can obtain a better programmability.
MATLAB®/Simulink® to enable the energy consumption
prediction with SLM in the design phase. Moreover, a use case 3. Development of the simulation tool
is carried out to test the feasibility of the simulation tool.
3.1. Overview of the approach
2. Research background
This work introduces a five-phase approach, see Fig. 1.
In the manufacturing context, energy quantification and Initially, the research object includes a real SLM-machine
evaluation is always an essential topic e.g. [9], [10], and [11]. Concept Laser Mlab Cusing R and its peripheral units like
Current research works have studied the energy consumptions vacuum cleaner and glovebox [8]. First, the basic data of this
of SLM-machines. In [12], the machine Concept Laser M3 machine is collected. Second, the system parameters related to
Linear is studied and its average power for operation and stand- its energy consumption are defined. Third, the energy
by is 3.35 kW and 0.7 kW respectively. In [13], the average consuming behavior of this machine is modeled by using bond
power of the machine Renishaw AM250 is quantified as 1.166 graphs. Fourth, Simulink® models are created, and fifth, the
kW for operation and 0.43 kW for stand-by. In [14], the data structure for managing Simulink® models and the
respective energy consumption of the machine MTT SLM 250 Graphical User Interface (GUI) for users are programmed. The
and Concept Laser M3 Linear are studied, and the average result is a simulation tool for energy consumption prediction of
power are respectively 1.09 kW and 3.33 kW. this SLM machine.
These experiment based methods and data can be used for
development of energy models of SLM by using an empirical 3.2. Data acquisition
modeling approach, e.g. [15]. The empirical energy models
describe the energy consumption of SLM as a function of The initial step to develop the simulation tool is to collect
process parameters like scanning speed, and hence, can be used the technical documents of the machine as well as the system
for energy consumption prediction for manufacturing a new components. Usually, manufacturers provide different
product with SLM. Nevertheless, the reliability and feasibility documents like manuals, circuit sheets and technical drawings,
of an empirical energy model is only associated with the and those documents can be used as the data source. In this
machine from which we extract the empirical data. If the Table 1. Clustering of system components.
machine changes, the empirical energy model will fail. As an Function System components
alternative, the physical modeling approach can also be applied
Laser processing Laser device, deflection device, CNC
for analyzing the energy consuming and transferring module
mechanism of SLM, e.g. [16]. In the previous research work
Process conditioning Power module, vacuum pump, oxygen
[17], an energy model using the physical modeling approach meter (for build platform and glovebox),
has been introduced. In accordance with [17] , this paper circulation fan, protective gas source
mainly aims at the model implementation in a simulation Powder handling Motors (for build platform, powder
software tool, and hence, the research question of this paper is container, and powder-spreading device)
defined as how to develop a simulation tool based on Powder Screening device, compressed air source,
MATLAB®/Simulink® platform. The reason why we chose preparation/recycling vacuum cleaner
MATLAB®/Simulink® as development platform is that Machine cooling Cooling device
MATLAB®/Simulink® has inbuild many fundamental electric
circuit elements and is widely used for the development of
30 Li Yi et al. / Procedia CIRP 81 (2019) 28–33
Li Yi et al. / Procedia CIRP 00 (2019) 000–000 3

work, the documents are collected from the department of parameters are assigned to corresponding physical properties
metal workshop where the machine is located. for respective function as well as system components. Taking
the function powder handling for example, see Fig. 2.
3.3. Definition of system parameters According to the circuit data and machine drawing, the function
powder handling is provided by three electrical motors. Motor
After the data acquisition, the system components and 1 and motor 2 are responsible for the motion of the powder
parameters are defined. In this work, the system components container and the build platform respectively, while motor 3 is
refer to electrical or pneumatic devices of the machine and used to drive the powder-spreading device. For the powder
peripheral units. For clustering these system components, five container and the build platform, their linear motion is
basic functions of SLM are defined [17]. The functions and converted by a transmission system from the rotary motion
identified system components are shown in table 1. output of the electrical motors, and hence, the physical
The physical properties of the system components associated properties related to those motions like weights, torques, linear
with forces and movements are analyzed and the system and angular velocities are defined. For motor 3, the movement

Interpretation of system parameters Interpretation of colors:

System parameters for function powder handling
Pi: Power consumption of motor i Red: Effort quantity
Ri: Resistance of motor i Blue: Flow quantity P3 , R3 Motor 3
P(Ri): Power loss of motor i Green: Power consumption ?3
Ti: Torque output of motor i Yellow: Constant parameter Powder-Spreading T3 , ? 3 r3
Guide rod ?m v3
? i: Angular velocity of motor i device F3 , v3
µgr
ri: Transmission ratio for motor i
Fi: Mechanical load for motor i
m0 Powder µ0 mps
vi: Moving speed for motor i
mbp: Mass of build platform and piston 2 mpc mbp
mps: Mass of power-spreading device
mpc: Mass of powder container and piston 1 Piston 1 ?1 Piston 2 ?2
r1 r2
?m: Mass of powders for each spreading process v1 v2
n: Number of the finished scanning loop Collected Data F 1 , v1 F2 , v2
µgr: Friction coefficient of guid rod
µ0 : Friction coefficient between powders and platform T1 , ? 1 T2 , ? 2
y
g: Standard gravity Motor 1 P1 , R1 Motor 2 P2 , R2
z x
1 Powder container Build platform

Modeling of the power flow using Bond Graphs

Powders in
(m0-?m*n)g
Power loss 1 P(R1) container
T1 Transmission F1 v1
Motor 1 Load 1
?1 system 1 v1 mpc*g
Friction of µ0*?m*g P1 Piston 1 and
v1
powders container
v3 F3 Transmission T3 P3 Electricity
Load 3 Motor 3
µgr*mps*g v3 system 3 ?3 supply
Friction of Powders and
v3 P(R3) ?m*n*g
guid rod P2 part on platform
Power loss 3 T2 Transmission F2 v1
Motor 2 Load 2
?2 system 2 v2 mbp *g
Power loss 2 Piston 2 and
2 A B: Power flow from A to B
P(R2 ) v1
platform

Architecture of the simulation tool

Processing program + Conrol window

(.m file) (.mlapp file)
3 Simulink ® models (.slx file) GUI

FBK034/027

Fig. 2. Example for definition of system parameters (1) and creation of bond graphs (2) with respect to the function powder handling and architecture of the
simulation tool (3).
 Li Yi et al. / Procedia CIRP 81 (2019) 28–33 31
4 Li Yi et al. / Procedia CIRP 00 (2019) 000–000

of the powder-spreading device is represented as a simplified 4. Use case

object which slides on a guide rod and pushes powders forward.
Hence, the physical properties like the mass of the powder- 4.1. Introduction of a lattice ring
spreading device, force for pushing powders, and friction
coefficients are defined. The part used to demonstrate the application of the
simulation tool is a ring with an outer diameter of 22 mm and
3.4. Modeling by using Bond Graphs an inner diameter of 19.5 mm. On the outer side of the ring, the
abbreviation of this research project "IRTG2057" is
Since the system parameters are defined, their interactions manufactured. In the middle of the outer side, many quad-
are described with respect to physical laws. Hence, the diametral lattices with a rod diameter of 0.3 mm and an edge
modeling approach bond graphs is applied for mapping them. length of 1.2 mm are embedded, as shown in Fig. 3. Since the
Bond graphs emphasize a power flow based point of view that lattices are micro structures and are difficult to manufacture
the power transmission (P) between connected engineering with conventional technologies, SLM is more suitable for
subsystems can be decomposed by two power variables, effort manufacturing this lattice ring. To apply SLM, an appropriate
(e) and flow (f), defined by [18]: parameter set for the build job solution should be designed. For
different solutions, the energy consumption can vary, and
(1)
hence, the simulation tool can be used to predict the energy
In different physical systems, effort and flow quantities may consumption for each possible solution and to support the
have different interpretations. For example, the effort and flow decision-making.
quantities can be force and velocity in the mechanical domain,
pressure and flow rate in the hydraulic domain, or voltage and 4.2. CAD-model preparation
current in the electrical domain. The concept for the
generalization of effort and flow quantities enables a unified In this use case, the workflow to prepare the CAD-model as
terminology to describe multi-physical subsystems in a well as to define the parameter set for a build job includes
common language [18]. Considering that this work mainly mainly four steps, see Fig. 3.
aims at energy and power issues, bond graphs are useful for The first step is to create a CAD-model of the lattice ring,
mapping the complex energy and power flow between different which is then converted into the STL-file in the second step.
system components of SLM. Moreover, the lattice ring is placed at a 45-degree angle to the
In line with the previous step for defining system platform, and the support structure is added to the bottom of the
parameters, the power flow modeling by using bond graphs ring. In the third step, an appropriate layer thickness is defined,
starts with the identification of all effort and flow quantities, according to which the model of the lattice ring including the
which are respectively indicated with red and blue colors in support structure is sliced equally to layers. In the last step, the
Fig. 2. Afterward, the other constant parameters are defined as scanning mode for each layer is defined. Fig. 3 shows three
well, and the power flow of this system is analyzed and mapped different basic scanning mode which also indicates the laser
in a word bond graph. scanning path on each layer. Last, the laser scanning speed is
defined.
3.5. Creation of Simulink® models
4.3. Energy efficient-oriented process configuration
After the power flow is modeled, the models are
Simulink®
created. Initially, the Simulink® models for all system The first step to design an appropriate parameter set for the
components are selected from the library Simscape which build job is to propose different possible alternative solutions,
provides comprehensive models for fundamental electrical see table 2. Solution 1 is regarded as the baseline solution. In
components and circuits. Then, the parameters of the selected solution 2, the laser speed is adjusted to 400 mm/s, and then the
Simulink® models like motor constants and resistances are layer thickness is correspondingly reduced to 100 µm to ensure
configured according to the collected technical documents. that the powders can absorb enough heat for melting. In
Finally, the configured Simulink® models are assembled solution 3, the hatch mode is changed to stripe hatching, and
according to the mapped energy flows, see Fig. 2. the other parameters are consistent with solution 1. In solution
4, the part count of the build job is adjusted to 4, and the
3.6. Creation of data structure and GUI remaining parameters are consistent with solution 1. The build
time (tbuild) is directly read from a CAE software. Similar to G-
Since the Simulink® models are created and configured, the code in the conventional manufacturing, SLM has its own laser
data structure and a GUI are programmed to manage the scanning files that contain all layer information including the
Simulink® models and to process the data. The data structure is build time. In addition, the time for preparing (tpreparation) and
implemented in .m files and the GUI is programmed in .mlapp recycling powders (trecycling) before and after build job is
file by using MATLAB® App Designer. In the GUI, users can assumed as 10 minutes, and hence, the entire process duration
define the scanning parameters and import the time parameters, (tprocess) is expressed by:
and then generate the power curve. Fig. 2. shows the structure
of the simulation tool.
32 Li Yi et al. / Procedia CIRP 81 (2019) 28–33
Li Yi et al. / Procedia CIRP 00 (2019) 000–000 5

1. Creation of the CAD model 2. Conversion of STL-file and 3. Definition of layer thickness 4. Definition of the scanning
of the lattice ring definition of support structures and slice of the model strategy
Close-up of lattices CAD model Close-up of bottom Oblique view
Scanning modes

Sliced STL-file
Close-up of Quad Simple Stripe
Flat view of one layer islands hatching hatching
Unit lattice IRTG2057 Close-up of top Front view

FBK034 /028

Fig. 3. CAD model preparation and build configuration.

ambient condition to melting point (ΔT), and latent heat (h0),

(2) expressed by [17]:

After the alternative solutions are defined, the energy (5)

consumptions of them are analyzed using the simulation tool.
For the assessment, four indicators are used: average power The ECPL value describes ratio between Etotal and
(Paverage), specific energy consumption (SEC), total/required manufactured layers (nlayers), given by:
energy ratio (TRER), and energy consumption per layer
(ECPL). The average power (Paverage) indicates the proportion (6)
between the total energy consumption (Etotal) and process
duration (tprocess), given by: Finally, the simulation result and calculated indicators are
shown in table 2. Besides, the difference ratio of each solution
(3) in comparison to solution 1 is calculated and shown in Fig. 4.
For each energy indicator, the value of solution 1 is regarded
The SEC value describes the ratio of Etotal to part mass (m0), as reference value. For solution 2, 3, and 4, the difference ratio
r(n) for an indicator is calculated by:
given by [17] :
(7)
(4)

In equation 7, aref and asolution are, respectively, the value of

The TRER value indicates the ratio of the Etotal to the an indicator for solution 1 and solution n. As shown that the
theoretical required energy demand (Etheoretical) to melt this part average power of each solution is nearly the same. The SEC
from ambient condition. The Etheoretical is further a function of and TRER values of solution 3 and 4 are significantly lower
m0, specific heat capacity (c0), increased temperature from than solution 1. The ECPL values of solution 2 and 3 are lower
than solution 1, while the ECPL value of solution 4 is 83%

Table 2. Alternative solutions and simulation results for the energy efficiency-oriented process configuration.
Parameter Solution 1 (baseline) Solution 2 Solution 3 Solution 4
Part count 1 1 1 4
Laser power (W) 100 100 100 100
Laser speed (mm/s) 2000a; 300b 2000a; 400b 2000a; 300b 2000a; 300b
Hatch mode Quad island Quad island Stripe hatching Quad island
Layer thickness (µm) 150 100 150 150
Layer number 135 202 135 135
Result
Build time (h) 0.53 0.62 0.27 1.25
Process duration (h) 0.86 0.95 0.60 1.58
Total energy consumption (MJ) 3.03 3.23 2.11 5.54
Average power (kW) 0.98 0.95 0.97 0.98
SEC (MJ/kg) 1375 1469 959 629
TRER (J/J) 1386 1481 967 634
ECPL (kJ/layer) 22.41 16 15.63 41.01
a: value for support structure; b: value for part
 Li Yi et al. / Procedia CIRP 81 (2019) 28–33 33
6 Li Yi et al. / Procedia CIRP 00 (2019) 000–000

Acknowledgements
Assessment of difference ratio
83%
80% Funded by the Deutsche Forschungsgemeinschaft (DFG,
German Research Foundation) - 252408385 - IRTG 2057.
40% Average
SEC TRER ECPL
power
-1% 7% 7% -29%
References
0%
-3% 0%
[1] ISO. Additive manufacturing: General principles. Geneva: ISO; 2015.
-40% -30% -30% -30% [2] Wohlers T, Campbell RI. Wohlers report 2017: 3D printing and additive
-54% -54% manufacturing state of the industry annual worldwide progress report. Fort
-80% Collins, Colo.: Wohlers Associates; 2017.
Solution 1(base line) Solution 2 Solution 3 Solution 4 [3] Kruth J-P, Leu MC, Nakagawa T. Progress in Additive Manufacturing and
Rapid Prototyping 1998; 47: 525–540.
[4] Huang R, Riddle M, Graziano D, Warren J, Das S, Nimbalkar S, Cresko J,
Fig. 4. Assessment of difference ratio between solutions. Masanet E. Energy and emissions saving potential of additive
manufacturing: The case of lightweight aircraft components. J. Clean. Prod.
higher than solution 1. The reason is that solution 4 enables the 2016; 135: 1559–1570.
[5] Kellens K, Mertens R, Paraskevas D, Dewulf W, Duflou JR. Environmental
build of four parts and the scanning area of each layer in
Impact of Additive Manufacturing Processes: Does AM Contribute to a
solution 4 is 4 times than solution 1. Although the total energy More Sustainable Way of Part Manufacturing? 2017; 61: 582–587.
consumption of solution 4 is nearly 2 times than solution 1, the [6] Gutowski T, Jiang S, Cooper D, Corman G, Hausmann M, Manson J-A,
SEC and TRER values of solution 4 are nearly half of solution Schudeleit T, Wegener K, Sabelle M, Ramos-Grez J, Sekulic DP. Note on
1 because that, in solution 4, the energy consumption of powder the Rate and Energy Efficiency Limits for Additive Manufacturing. J. Ind.
preparing/recycling as well as motors can be amortized by 4 Ecol. 2017; 21: S69-S79.
parts, while in solution 1 only by a single part. This implicates [7] Paris H, Mokhtarian H, Coatanéa E, Museau M, Ituarte IF. Comparative
that the more parts are built in one job, the higher energy environmental impacts of additive and subtractive manufacturing
efficiency can be achieved. technologies 2016; 65: 29–32.
In addition, according to the experiment [5], the SEC value [8] Concept Laser. Mlab Cusing R. https://www.concept-laser.de/en/-
products/machines.html.
of this SLM-machine varies from 309 MJ/kg to 533 MJ/kg. The
[9] Herrmann C, Thiede S, Kara S, Hesselbach J. Energy oriented simulation
simulation result of this case is significant higher. The reason of manufacturing systems – Concept and application 2011; 60: 45–48.
may be two parts: First, this work also considers the energy [10] Linke B. Life Cycle and Sustainability of Abrasive Tools. 1st ed. Cham:
consumption of peripheral units, while [5] just studied the Springer International Publishing; 2016.
single machine. Second, the part manufactured in this work is [11] Kara S, Li W. Unit process energy consumption models for material
very small and leads to an increased SEC value. Nevertheless, removal processes 2011; 60: 37–40.
the reliability and accuracy of this simulation model requires [12] K. Kellens, E. Yasa, R. Renaldi, W. Dewulf, J.P. Kruth, J.R. Duflou.
further study and confirmation. Energy and resource effiency of SLS/SLM process 2011.
[13] Faludi J, Baumers M, Maskery I, Hague R. Environmental Impacts of
Selective Laser Melting: Do Printer, Powder, Or Power Dominate? J. Ind.
5. Summary and outlook Ecol. 2017; 21: S144-S156.
[14] Baumers M, Truck C, Wildman R, Ashcroft I, Hague R. Energy inputs to
In this paper, the development of a simulation tool for additive manufacturing: Does capacity utilization matter? 2011.
energy consumption prediction is introduced. Based on data [15] Baumers M, Tuck C, Wildman R, Ashcroft I, Rosamond E, Hague R.
Transparency Built-in. J. Ind. Ecol. 2013; 17: 418–431.
acquisition of a real SLM machine, the system parameters
[16] Gobal A, Ravani B. Physical Modeling for Selective Laser Sintering
related to the energy consumption of SLM are defined and Process. J. Comput. Inf. Sci. Eng. 2017; 17: 21002.
modeled by using the bond graphs approach. A simulation tool [17] Yi L, Krenkel N, Aurich JC. An energy model of machine tools for
is then programmed using MATLAB®/Simulink®. Moreover, a selective laser melting 2018; 78: 67–72.
use case is carried out to test the feasibility of this simulation [18] Karnopp DC, Margolis DL, Rosenberg RC. System dynamics: Modeling
and simulation of mechatronic systems. 5th ed. Hoboken, NJ: Wiley; 2012.
tool. Here, the energy demand for manufacturing a small lattice
ring with different build solutions is simulated and assessed.
The result of the use case shows that the simulation tool is
feasible for analyzing and evaluating energy consumption of
SLM in the design phase. Nevertheless, the further work of an
experimental comparison to analyze the simulation reliability
is still called for.