Available online at www.sciencedirect.

com

ScienceDirect
Procedia CIRP 60 (2017) 163 – 168

27th CIRP Design 2017

Model-based design process for the early phases of

manufacturing system planning using SysML
Chantal Steimera,*, Jan Fischerb, Jan C. Auricha
a
Institute for Manufacturing Technology and Production Systems (FBK), 67653 Kaiserslautern, Germany
b
Siemens AG, Corporate Technology, 81739 München,Germany

* Corresponding author. Tel.: +49-631-205-3369; fax: +49-631-205-3304. E-mail address: chantal.steimer@mv.uni-kl.de

Abstract

This paper presents an approach for a model-based planning process for the early phases of manufacturing system planning (MSP). The goals are
a better integration of MSP with product development (PD) in the early design phases and an improvement of the coordination of the MSP-related
planning disciplines. The presented approach is based on model-based systems engineering (MBSE) and is supported by a modeling scheme
which uses the systems modeling language (SysML). The approach consists of a process description of the MSP and of different so called SysML
views which describe the information artifacts of each MSP step. It is divided into four different modeling levels that describe the manufacturing
system’s structure and behavior from different points of view. In order to validate the approach, an exemplary cylinder head production was
modeled. Extractions from this example are shown as SysML-diagrams. To conclude the paper, the advantages and disadvantages of using a
model-based planning approach with SysML for MSP are discussed.
© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
© 2017 The Authors. Published by Elsevier B.V.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review
Peer-review under
under responsibility
responsibility of the cscientific
of the scientifi committeecommittee of the
of the 27th CIRP 27th Conference
Design CIRP Design Conference.

Keywords: Manufacturing; System; Planning; Design; Method; Modelling; MBSE

1. Introduction
The current demand for manufacturing companies is
characterized by customer specific products with a large
number of variants and shorter product lifecycles that require a
shortening of the product engineering process in order to reach
a shorter time-to-market [1]. Thus, companies are challenged to
start the manufacturing system planning (MSP) at an earlier
stage. In order to meet this challenge, one approach is the
utilization of information from the early phases of product
development (PD). This requires a consistent information base
for all engineering processes that is not available so far. In order
to reach this information base, within PD, new methods like
model-based systems engineering (MBSE) are used [2]. MBSE
is a method for designing complex systems using a central
system model that enables an easy data exchange between all
design disciplines. Yet, it is mainly used for PD and, so far, no
approaches are known that use MBSE within MSP. Hence, this
paper presents an MBSE-approach for the early phases of MSP
that uses the Systems Modeling Language (SysML).
2. State of the art
In the following, the state of the art of current techniques for
MSP is presented and an introduction to MBSE is given.
2.1. Manufacturing system planning
MSP is part of factory planning and characterized by strong
interdependencies between the involved planning disciplines,
like process-, layout-, or control-planning. For factory planning
various approaches exist (e.g. [3]-[6]). Despite the large
amount, most of them follow similar development patterns that
can be divided into the five phases of preparation, structure
planning, detail planning, implementation planning, and
execution [7]. Within the preparation phase, preliminary work
and objectives planning is done. Structure planning provides
the design of conceptual solution variants of ideal and real
layouts for the manufacturing system. After deciding on a
variant, the detail planning finalizes the specification. Within

2212-8271 © 2017 The Authors. 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 scientific committee of the 27th CIRP Design Conference
doi:10.1016/j.procir.2017.01.036
164 Chantal Steimer et al. / Procedia CIRP 60 (2017) 163 – 168
the implementation planning, the implementation project itself
is derived and during the execution it is coordinated until its
finalization. [3] In this paper, especially the early phases like
preparation and structure planning are focused, as during these
the basic solution of the manufacturing system is derived.
Since for the MSP usually highly detailed information about
the product to be produced is required [3], the development of
products and the corresponding manufacturing system is
commonly executed sequentially. That means that MSP only
starts at a very late stage of the PD. In addition, the PD and the
MSP are executed by separate departments that are often only
poorly connected [8]. This results, e.g. in inefficiencies, costs
for adjustments, media disruptions, and long development
periods [7]. In order to overcome these problems, many
different approaches were proposed like simultaneous
engineering [9], Digital Factory [10], the design methodology
for mechatronic systems by VDI guideline 2206 [11], the 3-
cycle-model of product engineering [12], the double helix
development of production systems and products [13] or the
integrated product and manufacturing design by Britton et al.
[14], to name but a few. However, so far, their industrial
implementation is often limited to a strategic rather than
detailed operational level, as they mostly address the mentioned
problems on a high organizational level [15]. Furthermore,
there is a lack of detailed domain-spanning procedures for the
operational process level to properly integrate PD and MSP
contents [15] during the early design phases. Moreover, no
MBSE-approaches based on the usage of SysML are known,
that target the domain-spanning design of products and
manufacturing systems or the MSP itself.
2.2. Model-based systems engineering within the product
engineering process
MBSE is an approach for the development of complex
systems by iteratively connecting the development activities of
all involved disciplines (e.g. mechanics, electronics, IT) with a
formalized central system model, beginning in the conceptual
design phase and continuing throughout development and later
life cycle phases [16]. Thereby, the central system model is the
basis for an interdisciplinary development and also substitutes
the so far commonly used document-based information
exchange [17]. Usually, MBSE-models describe a system by
requirements as well as the system’s behavior and structure
[18]. For this purpose, modeling languages are used for a
graphical, consistent, and formal description of the model [19],
two examples for modeling languages are SysML or OPM [20].
In contrast to the MSP, within early phases of PD, model-
based methods are already used more frequently. Some
examples are the model-based system development [18], the
model-based design of cybertronic systems [21], the
geometrical constraints modeling for mechatronic design [22],
the modeling and design of manufacturing machinery [23] or
the software design for manufacturing execution systems
(MES) [24]. In order to reach a better integration of PD and
MSP, it is reasonable to use MBSE for the MSP, as well.
Furthermore, since MSP necessarily requires information about
the products to produce, synergies are to be expected by an
integrated design, e.g. reusability of product information
relevant for the manufacturing system, traceability of product
changes during the development process, or an easier
coordination between the disciplines based on similar models
and common modeling language [25]. Therefore, especially in
the context of MSP, the development of new planning methods
using MBSE is required. For that reason, a model-based design
process for the early phases of MSP is developed that uses the
SysML for the modeling of manufacturing systems.
3. Model-based design process for the early phases of
manufacturing system planning
The model-based design process for the early phases of MSP
is developed as part of an integrated process framework for the
design of products and manufacturing systems, targeting a
better integration of PD and MSP within early design phases
[26]. In the following, the MBSE-approach especially for MSP
is described in detail.
3.1. Integrated process framework for the design of products
and manufacturing systems
The MBSE-approach for MSP defines all required tasks for
the product-integrated design of manufacturing systems. It has
a two-part process structure, consisting of the cross-disciplinary
early system design and the subsequent discipline-specific
design, and is described with an iterative V-model that allows
iterations within and between the two parts (Fig. 1).
The system design includes all tasks that are mainly
executed during the early development phases of a system.
These tasks usually relate to more than one design discipline of
MSP but are coordinated by the systems engineer. The system
design is subdivided into four design levels that describe the
system from different points of view e.g. a behavioral or a
structural perspective. The results of an iteration of the system
design can either start another iteration or can be used within
the subsequent discipline-specific design. The latter is usually
coordinated by specialized engineers, e.g. the manufacturing
process engineer or the logistics specialist. They build on
results from the system design in order to develop a technically
specific and implementable solution or several alternatives.
As products and manufacturing systems both represent
technical systems, a mutually used description scheme based on
the SysML is defined. It is used to build the formalized central
system model that facilitates a non-redundant synchronization
and cooperation between and within PD and MSP. The
description scheme specifies a process model for the four levels
of the system design. For most process steps, one so-called
SysML view is defined, showing the result of each step in one
or more SysML diagrams. For this purpose, the preferred
diagram types and usable elements are defined in the SysML
view. In the following, the model-based design process for
MSP, including its four design levels, is described in detail.
3.2. Model-based manufacturing system planning in early
design phases
The system design for MSP is divided into four levels:
context level, manufacturing technique level, structure &
 Chantal Steimer et al. / Procedia CIRP 60 (2017) 163 – 168
Fig. 1. Iterative V-model of model-based systems engineering approach for manufacturing system planning
control level, and technical solution level (Fig. 1). The sub-
levels provide a simultaneous illustration and step-by-step
specification of the manufacturing system by different points of
view on the same model. They are repeatedly iterated until the
required level of detail is accomplished.
3.2.1. The context level
The context level defines the framework and boundaries of
the manufacturing system (Fig. 2). Here, requirements from
corporate management, product development, and other
stakeholders are described, such as products and variants,
proposed target costs, output quantity, supplier conditions, and
statutory regulations. In addition, the use-cases of the
manufacturing system are defined, which describe the product
variants and their respective amounts, i. e. comparable to a
production program (Fig. 2 - step 1a and 1b).
Afterwards, the information about the product specified by
the PD is considered. It is described in model-based product
structure diagrams using SysML block definition diagrams
Fig. 2. Activities of context level
165
(bdd) and available within the central system model. These
diagrams contain information about the product’s assembly
groups, individual parts, and product features (Fig. 2 - step 2a).
This information is especially relevant as most of the MSP
processes are directly or indirectly dependent on the product
characteristics. In this context, product features are
characteristics of a product that are relevant for the MSP, such
as a hole or a screw joint, and thus require at least one
manufacturing process for being created. Hence, they describe
the product from a manufacturing perspective. At the beginning
of the development process, product features are on a low level
of detail as they originate from the product’s system design.
However, they are refined within each iteration of the PD
process. In case of non-producible product features, feedback
in form of change requests is given from MSP to PD. Moreover,
MSP can demand additional product features that are e.g.
required for executing certain manufacturing processes.
Product features are used to define solution-independent
manufacturing processes that describe the activities to create all
product features in an abstract way, such as “add hole” or
“create screw joint”. This solution-independent formulation
allows the later assignment of different manufacturing
techniques and resources to the respective process steps. Some
manufacturing processes may be divided into several sub-
processes if the creation of a product feature requires different
manufacturing processes. Afterwards, all manufacturing
processes are arranged based on their logical sequence of
execution (Fig. 2 - step 2b, 2c, and 2d) using activity or
sequence diagrams (i.e. a preliminary form of a work plan).
In addition to the product related information, the context of
the manufacturing system is described by material-,
information-, and energy flows (Fig. 2 – steps 3a, 3b) and
further requirements are forwarded to the subsequent design
levels (Fig. 2 - step 5).
3.2.2. The manufacturing technique level
On the manufacturing technique level, the manufacturing
processes, which can be divided into primary and secondary
processes, are technically specified and structured in a material
166 Chantal Steimer et al. / Procedia CIRP 60 (2017) 163 – 168
flow oriented form. Primary processes are all value adding
processes and secondary processes are mainly logistics
processes like transport or material handling. For this purpose,
the previously solution-independent manufacturing processes
are allocated to manufacturing techniques, like the general
technical solution principles for manufacturing, such as drilling
or milling, as e.g. specified in DIN 8580 [27]. Each
manufacturing process can be allocated to one or several
manufacturing techniques in order to show the possible solution
space for the realization of each manufacturing process. In a
second step, process groups are created by selecting a
manufacturing technique for each manufacturing process while
also considering the process’ position in the process sequence
(compare Fig. 6). Hence, process groups represent a first
abstract form of factory resources. By arranging the process
groups within an internal block diagram (ibd) and by
connecting them by ports representing the object flows between
them (structure of primary process groups), interdependencies
between manufacturing processes regarding material flows
become visible (Fig. 3 - step 1, 2, and 3; compare also Fig. 7).
Afterwards, secondary manufacturing processes such as
transport or handling processes are added and concretized in the
same way as the primary processes (Fig. 3 - steps 4a-7). The
illustration of the structure of secondary process groups is
comparable to the visualization of material flows e.g. in a value
stream map.
3.2.3. The structure & control level
The structure & control level fulfills two purposes - the
design of the manufacturing system’s structure and the
definition of its control logic. For the structural design, the
process groups from the manufacturing technique level are
allocated to resource types that are able to provide the required
manufacturing techniques. Thereafter, at least one concrete
resource for each process group is preselected for being utilized
in the manufacturing system. Analogous to the manufacturing
process description at the manufacturing technique level, the
resources are specified by manufacturing process capabilities,
thus allowing an automated matching between the required
manufacturing processes and the capabilities of the
manufacturing resources, e.g. by using constraint matching
Fig. 3. Activities of manufacturing technique level
technologies. All selected resources are then arranged within an
ibd and linked to each other based on the possible material flow
routes, leading to a first manufacturing system structure with
material flows. This illustrates a qualitative and preliminary
form of the manufacturing system’s layout (Fig. 4 - step 2a, 2b
and 2c).
Furthermore, the control logic of the manufacturing system,
in the sense of production planning and control (PPC), is
specified. Therefore, in a first step (Fig. 4 - step 3a), all logical
control techniques are defined in a bdd as well as allocated and
parameterized (e.g. FIFO, Kanban, negotiation-procedures). In
a second step, the behavior of these control elements is
specified by using behavioral diagrams such as activity,
sequence, or state diagrams (Fig. 4 - step 3b). Especially
modern types of manufacturing systems that are often
controlled by agent based negotiations can be effectively
modeled in such a way (Fig. 4 - step 3a and 3b). Based on the
manufacturing system’s structure and control logic, material
flow simulations can be executed in order to evaluate or
optimize different system designs (Fig. 4 -step 5 and 6). For this
purpose, SysML test cases can be defined (Fig. 4 - step 4) and
the results of the simulation runs can be transferred back into
the system model (Fig. 4 - step 7).
3.2.4. The technical solution level
On the technical solution level, all technical sub-systems of
the manufacturing system are identified that need to be further
detailed in the following planning phases. Therefore, three sub-
system-types are differentiated: (i) sub-systems that require
another iteration of the manufacturing system’s system design
because they are, so far, only considered as a black-box in the
current design iteration, (ii) sub-systems that can be described
as products, (e.g. special purpose machines) and therefore have
to be developed within a product development process, starting
with the product’s system design, and (iii) sub-systems whose
development is continued in the subsequent discipline-specific
design processes of the manufacturing system as e.g. the
Fig. 4. Activities of structure & control level
 Chantal Steimer et al. / Procedia CIRP 60 (2017) 163 – 168 167

Fig. 6. Simplified extraction of the sequence of manufacturing processes,

depicted in an activity diagram on the context level

manufacturing resources, the building of groups is mainly

based on the position of a process in the process sequence (see
Fig. 6) and the allocated manufacturing techniques (previously
conducted in a bdd or allocation matrix in step 2).
Fig. 5. Activities of technical solution level
technical configuration of the manufacturing system’s control
logic or the logistics system (Fig. 5 - step 2a, 2b, 2c). For each 4.2. Advantages and Disadvantages of using a SysML-based
technical sub-system, requirements and interfaces are specified MBSE-approach
according to the results of the system design.
The SysML-based MBSE-approach provides several
4. Validation benefits for the MSP that can be often related to those that arise
for PD or general technical systems (compare e.g. [19], [2]):
The description scheme was validated based on an (1) Due to the central system model, trace links between
exemplary manufacturing system for cylinder heads. various modeling levels exist, e.g. product-properties, value-
stream/process view, and structural view that allow the direct
interpretation of system or element dependencies, even related
4.1. Modeling examples to the product design. (2) Decisions and conceptual planning
steps during the design process are formally documented, as
Fig. 6 shows a simplified extraction of the required process e.g. the forming of process groups that leads to the selection of
steps to produce and assemble a cylinder head. Those are, as far factory resources. (3) One common non-redundant data basis
as possible, described independent from the technical solution allows the collaboration between several design disciplines and
e.g. “add_hole” and not “drill_hole”. The processes are roles during the design process. (4) SysML provides a
modeled in an activity diagram allowing the description of formalized and visually interpretable description of the
sequential, simultaneous, alternative, or even case dependant manufacturing system, and allows the easy interpretation by
process orders. For the manufacturing of the cylinder head, developers. Furthermore, this supports the extension of the
several sequences of process steps are possible, e.g. the upper presented greenfield approach to the industrially more relevant
chamfer can be added before or after the holes. However, the brownfield case, where legacy systems of an existing factory
production of the cylinder head must be finished before the
assembly can start (e.g. the assembly of glow plugs). In the
previous step on the context level, the processes were allocated
to certain product features of the cylinder head (e.g. a cooling
hole or a plane surface) and thus the dependency to the product
characteristics can be traced. Hence, eventual product changes
during MSP can directly be linked to the related manufacturing
processes.
Fig. 7 depicts a simplified extraction of the system model
(process group structure) from the manufacturing technique
level. In this view the manufacturing processes from Fig. 6 are
arranged in process groups (comp. Fig. 3 - step 3) representing
a first, abstract draft of the structure of the manufacturing
system, including the material flow of product components
between the groups. Here, the cylinder head flows from the
drilling to the turning and milling, the screwing, and finally the Fig. 7. Simplified extraction of the material flow between process groups,
testing processes. As process groups represent abstract depicted in an ibd on the manufacturing technique level
168 Chantal Steimer et al. / Procedia CIRP 60 (2017) 163 – 168
have to be considered. Therefore, the existing plant would be
initially modeled based on the presented approach and then
iteratively refined now also including the brownfield-related
design constraints and requirements. Consequently, this would
require some additional alterations of the presented approach.
(5) SysML diagrams can be aligned to other models and forms
of manufacturing systems such as value streams and material
flow simulations. Partially, these can also be automatically
generated from the system model, thus reducing the modeling
effort. (6) Due to the object oriented approach, model elements
can be easily reused during the design process.
However, as the presented approach is at a relatively young
development stage, some disadvantages to deal with exists:
(1) The approach strongly deviates from traditional
representations of manufacturing systems and procedures of
MSP, so the initial acceptance by MSP-engineers may be low.
This is also amplified, since (2) the use of MBSE or the SysML
requires expert knowledge for modeling and partially also for
interpreting the system models, and (3) mostly abstract
graphical representations of MSP elements are used, such as
rectangles, circles, and lines. SysML has no functionality for
modeling true-scaled manufacturing layouts. (4) The modeling
and evaluation of different design variants in the system model
is not explicitly covered by SysML and still part of ongoing
research (e.g. [28]). (5) Models with a bigger scope tend to
become confusing, especially when they are not modeled by
SysML experts. Models built by different persons will usually
not look the same. (6) Libraries for standard manufacturing
entities do not exist, yet and thus make it effortful to model
manufacturing system elements. (7) The approach requires a
relatively high level of abstract thinking especially within early
design phases as many model elements do not represent real
(physical) elements or properties of the manufacturing system.
(8) The alignment of system modeling processes with business
processes of an enterprise usually requires the administration of
the system model with PLM systems. However, such
integrations are currently still researched/ developed (e.g. [29]).
5. Summary and outlook
This paper introduced a model-based design process for the
early phases of manufacturing system planning using the
modeling language SysML. It consists of a description scheme
with four description levels – context level, manufacturing
technique level, structure & control level, and technical solution
level – that allow the modeling of manufacturing systems from
different perspectives and are iteratively applied during the
system design process.
Several opportunities for further research exist, such as the
use of SysML also in later phases of MSP or the extension of
the interface specification between PD and MSP.
Acknowledgements
This work was funded by the German Federal Ministry of
Education and Research (BMBF) within the Framework
Concept “Research for Tomorrow’s Production” (funding
number 02PJ2573) and managed by the Project Management
Agency Karlsruhe (PTKA).
References
[1] Brettel M, et al. How Virtualization, Decentralization and Network
Building Change the Manufacturing Landscape: An Industry 4.0
Perspective. International Journal of Mechanical, Aerospace, Industrial
and Mechatronics Engineering 2014;8(1):37-44.
[2] Eigner M, et al. An Initial Approach for the Application of Product
Assembly Information in the Early Phases of the Product Development
Process by Using Methods of Model Based Systems Engineering.
Proceedings of the NordDesign 2014;724-733.
[3] Grundig C-G. Fabrikplanung. 5th ed. München: Hanser; 2015.
[4] Kettner H, Schmidt J, Greim H-R. Leitfaden der systematischen
Fabrikplanung. München: Carl Hanser; 1984.
[5] Schenk M, Wirth S, Müller E. Factory Planning manual.
Berlin,Heidelberg: Springer; 2009.
[6] Wiendahl H-P, Reichardt J, Nyhuis P. Handbook Factory Planning and
Design. Heidelberg: Springer; 2015.
[7] Steimer C, Aurich JC. Analysis of information interdependencies between
product development and manufacturing system planning in early design
phases. Proceedia CIRP 2016;50:460-465.
[8] Bracht U, Geckler D, Wenzel S. Digitale Fabrik. Heidelberg; 2011.
[9] Eversheim W, et al. Simultaneous Enigneering approach to an integrated
design and process planning. European Journal of Operations Research
1997;100:327-337.
[10] VDI Guideline 4499. Digital Factory - Fundamentals. Berlin: Beuth; 2008.
[11] VDI Guideline 2206. Design methodology for mechatronic systems.
Berlin: Beuth; 2004.
[12] Gausemeier J et al. Integrative development of product and production
system for mechatronic products. Robotics and Computer-Integrated
Manufacturing 2011; 27: pp. 772-778.
[13] Wiktorsson M. Consideration of Legacy Structures enabling a Double
Helix Development of Production Systems and Products. In: Henriques, et
al., editors. Technology and Manufacturing Process Selection. London:
Springer, 2014.
[14] Britton G, Trovinen S. Design Synthesis: Integrated Product and
Manufacturing Design. Boca Raton, Fla.: CRC Press; 2014.
[15] Stoffels P, Vielhaber, M. Energy Efficiency oriented Development of
Production Systems. in: Proceedings of the 20th International Conference
on Engineering Design (ICED) 2015;4:323-332.
[16] Walden DD, et al., editors. Systems engineering handbook. 4th ed. New
Jersey: Wiley; 2015.
[17] INCOSE. Systems Engineering Vision 2020. San Diego: INCOSE; 2007.
[18] Alt O. Modellbasierte Systementwicklung mit SysML. München: Carl
Hanser; 2012.
[19] Friedenthal S, Steiner R, Moore A. A Practical Guide to SysML. San
Francisco: Morgan Kaufmann; 2009.
[20] Grobshtein Y, et al. Systems modeling languages: OPM versus SysML.
IEEE ICSEM 2007;102-109.
[21] Eigner M. et al. An Interdiscilinary Model-Based Design Approach for
Developing Cybertronic Systems. Proceedings of DESIGN 2016, 1647-
1656.
[22] Barbedienne et al. Introduction of Geometrical Constraints Modeling in
SysML for Mechatronic Design. IEEE Mechatronics 2014;145-150.
[23] Bassi L et al. A SysML-Based Methodology for Manufacturing Machinery
Modeling and Design. IEEE/ASME transactions on mechatronics
2011;16(6):1049-1062.
[24] Piétrac L et al. On the use of SysML for manufacturing execution system
design. IEEE ETFA: 2011;1-8.
[25] Weidig C et al. Process planning and decentralized process control for
cybertronic production systems. Applied Mechanics and Materials 794
2015;486-493.
[26] Meissner H et al. Model-Based Development Process of Cybertronic
Products and Production Systems. Advanced Materials Research 1018
2014; 539-546.
[27] DIN 8580. Manufacturing processes. Berlin: Beuth; 2003.
[28] Weilkiens, T. Variant Modeling with SysML; MBSE4U; 2016
[29] Müller, P. et al. SysML-Modellverwaltung im PDM/PLM Umfeld. (to be
presented at) TdSE - Tag des Systems Engineering; 2016.