Cluster of Excellence

Integrative Production Technology

for High-Wage Countries

Perspectives
of inter-disciplinary Research
 Contents

Content

Preface
Introduction1
Individualised Production5
Direct, Mould-less Production Systems 6
Automated Optimisation for High-pressure Die Casting Dies 11
Automated Optimisation of Profile Extrusion Dies 16

Virtual Production Systems21

Virtual Production Intelligence goes Virtual Reality 22
Development of Aluminum-Free Case Hardening Steel27

Integrated Technologies33
Integrative Sheet Metal Processing Centre34
Multi-Technology Machining Centre39
Monolithic Plastics Optics44
Structural Parts49
Integrating Electrical Functionalities in Plastics Parts54

Self-optimising Production Systems59

Corporate Material and
Information Flow60
Cognitive Assembly Cell65
Self-optimising Production Planning and Control69
Self-optimising Injection Moulding74
Self-optimising Manufacturing Systems for Laser Cutting79
Model-based Modules for the Self-optimisation of Gas Metal Arc Welding Processes84
Self-optimisation of the Weaving Process89
Self-optimised Metal Cutting Processes94
Self-optimisation of the Radial Braiding Process99
Self-optimising Assembly of Large Components104
Self-optimising Assembly of Laser Systems109

Collaboration Productivity115
Theory of Production116
Scientific Cooperation Engineering 117
Scientific Cooperation Portal 118
Technology Transfer122
PROTECA Integrative Production Technology Aachen123

Preface
Manufacturing is one of the core sectors of the European
economy with more than 30 percent of all employees.
Manufacturers in high-wage earning countries are there-
fore more exposed to volatile changes in the global en-
vironment and under pressure from the financial climate
so that, a reduction in costs is only possible to a certain
extent. To remain globally competitive, it is therefore
essential to compensate cost disadvantages by providing
greater customer proximity, flexibility, adaptability and
process synchronisation. These measures in combination
with high productivity lead to locational advantages, and
need therefore to be focused on.
Since 2006, scientists from the RWTH Aachen Univer-
sity have been working together on interdisciplinary
research projects within the Cluster of Excellence (CoE)
titled, Integrative Production Technology for High-Wage
Countries. Sponsored by the Deutsche Forschungs-
gemeinschaft (DFG, German Research Foundation) 80
scientists from more than 25 institutes from the areas of
production, material sciences, economics as well as natu-
ral and social sciences, are developing in cooperation with
Prof. Dr.-Ing. Christian Brecher (CEO)
Preface
renowned companies from different branches, sustaina-
ble solutions for production in the future.
The focus of the Cluster of Excellence is primarily on
increasing productivity, not only on production in a
traditional sense, but also in the product development
process. Product development and production are
addressed using a holistic approach, and lead to a new
theory of production. Taking into consideration the
background of new breakthroughs in hard- and software
and in intelligent automation and social networks, a new
type of collaboration productivity arises, that can be seen
in the results.
This publication will provide you with an overview of the
current technical demonstrators, and research results and
analysis of different studies by the Cluster of Excellence,
Integrative Production Technology for High-Wage
Countries.
We look forward to receiving your suggestions and ideas,
and if you are interested in participating in any of our
activities, please feel free to contact us.
Cathrin Wesch-Potente (COO, 04/12 - 06/14)

Introduction
The Polylemma of Production
For production in high-wage countries, it is vital to offer
products corresponding to customer demands, at com-
petitive prices as well as to fully adapt to the market, and
social changes while still assuring an optimal usage of
resources. These contradicting objectives (scale vs. scope;
plan vs. value) form the so-called polylemma of produc-
tion. Its resolution is the key to sustainable production in
high-wage countries.
Performance indicators and models within individual
production technologies are not sufficient, a deeper
technology-spanning perception is necessary. Moreover,
it has become evident that reductionist description and
interpretation models have to be combined with a cyber-
netic approach in order to account for the unpredictable
and complex nature of production systems as a whole.
The Cluster of Excellence aims therefore, not only to
develop production technologies for the future, but
also to increase productivity in production, and product
development processes by combining deterministic and
cybernetic models.
Introduction
The rapid progress in hard- and software during the last
years has already changed the common procedures of
manufacturers. In future the strong integration of infor-
mation and communication technologies (ICT), intelligent
sensors and automation in production as well as social
networks will accelerate this trend, and lead to an incre-
ase in productivity. In retrospect, all preceding industrial
revolutions resulted in a significant productivity growth
within production. In the first industrial revolution, the
steam engine was introduced as a main source of energy.
At the same time, labour became increasingly specialised
and standardised parts were launched, allowing for an
easier exchange of broken parts. These developments
led to the establishment of central production sites, and
to a significant growth of manufacturers. Demand for
energy rose simultaneously and urbanisation around
industrial areas took place. The next boost of productivity
was triggered during the second industrial revolution
when Frederick Winslow Taylor introduced the division
of labour, and Henry Ford introduced the conveyor belt.
Standardised products, e.g. Model T and clear work task
descriptions enabled learning, while economies of scale
1
 Introduction
Individualised Production
Scope
Plan
Virtual Production Systems
allowed for rising quantities of an item to be built with
a considerable reduction of unit costs and marking the
beginning of mass production.
Computer development, as well as the invention of pro-
grammable logic controllers (PLC) and industrial robots
led to the third industrial revolution. This involved auto-
mation of all sorts of processes driven by Moores law,
which states that the number of transistors on a compu-
ter chip, and therefore the computing capacity doubles
every two years. In contrast to the three preceding indus-
trial revolutions, the current revolution has its roots, not
in the production sites but in the private domain where
social networks and mobile devices are used as means to
gather, evaluate and communicate information. This new
development of networking, interaction and cooperation
among and between employees and machines offers
greater potential for higher productivity. Not only manu-
facturing processes, but indirectly other departments can
also benefit from this higher degree of interconnection.
Therefore, two major objectives have been defined within
the Cluster Integrative Production Technology for High-
Wage Countries: Return on Production and Return on
Engineering. These objectives are addressed in the four
main research areas: Individualised Production, Virtual
Production Systems, Integrated Technologies and Self-
optimising Production Systems.
2
Self-optimising Production Systems
Value
Scale
Integrated Technologies
Return on Engineering
The Return on Engineering quantifies the growth of
productivity which is gained in the product development
process. The aim is to attain considerable cost benefits for
small lot sizes, down to lot size one. The key for indivi-
dualised production is a radical shortening of product
development processes, in order to quickly develop a
product from the beginning of the idea. The means of
prototypes and agile development on the basis of short
feedback cycles are as important as new business models
which embed customers and reduce disruptions. With the
strategic use of simulations, development times and costs
can be decreased. For being able to deploy virtual proto-
types, it is not sufficient to look at one single process step
and one single technology. Instead the entire value chain
needs to be taken into consideration. Only the creation
of complete and consistent virtual process chains allows
simulating diverse scenarios and factors that may exercise
an influence. Such simulations strengthen the basis for
making high-quality decisions in production systems.
Return on Production
Increasing productivity in manufacturing processes is
represented by the objective Return on Production. A
revolutionary shortening of the complete value chain,
and therefore a reduction of interfaces and disruptions
 Introduction

Radically Short Product Development Processes Better Performing than Engineered

Return on Engineering Return on Production

Integrative
Production Technology Integrative
Production Technology

today today

#
# #

Virtual Engineering of the Complete Value Chain Revolutionary Short Value Chains

are essential for a raise in productivity. By integrating or
substituting several process steps and technologies into
one, by parallelising process steps, or by integrating mul-
tiple functions into one product, production times and
disturbances can be reduced.
Furthermore, self-optimising production systems which
are directly adaptable and capable of making decisions
so that they are able to immediately react to failures offer
high potential. The aim is to continuously improve the
performance from a socio-technical perspective so that
even the initially anticipated level of productivity can be
exceeded. It may be possible to gain productivity growths
beyond todays foreseeable limits, by adapting the target
system and corresponding structures during the process.
Concepts and first results regarding a rise of Return on
Production and Return on Engineering will be described
in this publication in the following diverse research fields.

CHAIR FOR
COMPUTATIONAL
ANALYSIS OF
TECHNICAL
SYSTEMS

INSTITUT FR REGELUNGSTECHNIK
UNIV.-PROF. DR.-ING. DIRK ABEL

3
 Introduction
Steering Committee
Prof. Wolfgang Bleck
Department of Ferrous Metallurgy (IEHK), RWTH
Aachen University
Prof. Christian Brecher
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Prof. Gerhard Hirt
Institute of Metal Forming (IBF), RWTH Aachen
University
Prof. Christian Hopmann
Institute of Plastics Processing (IKV) at RWTH Aachen
University
Prof. Sabina Jeschke
Institute of Information Management in Mechanical En-
gineering (IMA) & Center for Learning and Knowledge
Management (ZLW) & Assoc. Institute for Management
Cybernetics e.V. (IFU), RWTH Aachen University
Prof. Christopher M. Schlick
Institute of Industrial Engineering and Ergonomics
(IAW), RWTH Aachen University
Prof. Gnther Schuh
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Prof. Wolfgang Schulz
Fraunhofer Institute for Laser Technology (ILT) &
Nonlinear Dynamics of Laser Processing (NLD), RWTH
Aachen University
Denis zdemir (since 07/2014)
Director of the Cluster of Excellence Integrative
Production Technology for High-Wage Countries
Cathrin Wesch-Potente (04/2012 06/2014)
Director of the Cluster of Excellence Integrative
Production Technology for High-Wage Countries
4
Core Team of Deputy Coordinators
Dr. Markus Bambach
Institute of Metal Forming (IBF), RWTH Aachen
University
Jennifer Btzler
Institute of Industrial Engineering and Ergonomics
(IAW), RWTH Aachen University
Dr. Reinhard Freudenberg
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Dr. Werner Herfs
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Christian Hinke
Chair of Laser Technology (LLT), RWTH Aachen
University
Denis zdemir (since 07/2014)
Director of the Cluster of Excellence Integrative
Production Technology for High-Wage Countries
Dr. Christina Reuter
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Till Potente
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Dr. Ulrich Prahl
Department of Ferrous Metallurgy (IEHK), RWTH
Aachen University
Cathrin Wesch-Potente (04/2012 06/2014)
Director of the Cluster of Excellence Integrative
Production Technology for High-Wage Countries
Christian Windeck
Institute of Plastics Processing (IKV) at RWTH Aachen
University

Individualised Production
The research area of Individualised Production deals with
how customised products can be produced efficiently.
The aim is to achieve competitive prices, even at the lot
size of one. Thus, the focus is on a radical shortening of
the process from the idea of a product to the production.
One approach is the implementation of direct mould-less
manufacturing processes that have been advanced by the
Cluster of Excellence. One of the key technologies, in this
area is Selective Laser Melting (SLM) which is an additive
manufacturing process originally used for generating
prototypes only, but is now being qualified for small-
batch series as well. This requires however, improvement
and acceleration of processes as well as extending the
material range.
Furthermore, SLM opens up new possibilities of product
design as a result of unrestricted geometric freedom,
prototypes with a completely new geometry and unique
functional characteristics can be developed and optimi-
sed. Hence, SLM products can be highly individual and
designed with respect to application-specific require-
Individualised Production 1
ments. Additionally, SLM allows new business models 3
where the customer is more involved in the design and
development process.
A second approach within the research cluster was to
focus on mould-based manufacturing processes used in 4
mass-production. The manufacturing of moulds is time-
consuming and cost-intensive, and requires extensive
experience, as well as the practice of trial-and-error. The
objective within the cluster was therefore, to develop
processes that are competitive especially in small-scale 5
applications. Moulds needed to be designed and pro-
duced faster, and costs had to be reduced. This applied
to continuous manufacturing, e.g. profile extrusion, as
well as to discontinuous processes, e.g. metal casting.
With the method of virtual design and optimisation of
moulds, cost-intensive experiments with a real mould can
be reduced, and development times shortened. New tool
geometries lead to further challenges. The first SLM tools
are being tested.
5
 1 Individualised Production
Direct, Mould-less Production Systems
2
Introduction
Solving the dichotomy between mass-production and in-
dividualised single piece production is the aim of ICD A2.
3 Additive manufacturing methods, like Selective Laser Mel-
ting (SLM), can be a main driver for producing individua-
lised products at nearly the same costs as mass produced
products. The basic principle of Additive Manufacturing
is to manufacture a part layer by layer, utilising e.g. metal
4 powder. In contrast to sintering, during the SLM process
this metal powder is fully molten, and therefore 100%
dense parts with properties comparable to conventional
manufactured ones can be produced. Neither tools nor
moulds are required, in contrast to conventional manuf-
5 acturing (e.g. die-cast). Comparing additive manufac-
Figure 2: Helicopter part with lattice structure
6
turing with conventional manufacturing methods, cost
curves differ significantly (see figure above): and due to
lower start-up costs, small lot sizes can be produced cost
efficiently with additive manufacturing. This is a key-
enabler for new business models focusing on highly
individualised products (Individualisation for free).
Manufacturing costs of additively manufactured parts
are almost completely independent of part complexity
(Complexity for free), and since the time and effort
needed for manufacturing is only depending on compo-
nent volume, innovative products with integrated func-
tionalities can be realised using additive manufacturing
methods like SLM (see fig. 2 and 3).
Figure 3: Ressource-efficient stub axle

Figure 4: Results of phase 1 and focus of phase 2
Practical Issues
Off-the-shelf SLM machines manufactured by considera-
ble OEMs like EOS, SLM Solutions, Realizer and Concept
Laser are limited in regards to their productivity. By incre-
asing laser power from 200 W to 1 kW and developing
a double-beam concept to realize the skin-core principle,
the process efficiency could be increased by a factor of
10 in the last five years. The double-beam concept utilises
a beam switch, to shape two different beam sizes, either
200 m or 1 mm beam diameter (see fig. 4). This ability
to switch beam diameter, enables the machine to build
the contour of the additively manufactured part with
a laser spot size of 200 m and small layer thicknesses
(approx. 50 m) to achieve high accuracy, surface quality
and resolution. In contract, core sections only resemble
bulk material and therefore a beam diameter of 1 mm
and layer thicknesses of >200 m are used to increase
Figure 5: Profile extrusion mould
Individualised Production 1
2
the build-up rate without scarifying part quality. Ho-
wever, today there is not much experience yet in industri-
al applications of the skin-core principle with laser powers 3
exceeding 1 kW. Additionally, pre- and post-process work
is needed to prepare the machine results in a rather high
amount of nonproductive time and therefore resulting in
a lower efficiency of the SLM process. Also, utilisation of
the laser source itself is limited, since powder dispersion 4
and exposure to laser radiation are distinctive, non-
parallelised process steps. A 100% laser utilisation cannot
be achieved with todays SLM machines. Mechanical
properties of additively manufactured lattice structures
under load-bearing conditions are only sparsely available. 5
The huge potential of functionally optimised parts, by
integration of lattice structures (see fig. 6), today is far
from being utilisable.
Figure 6: Functional integrated rim-support
7
 1 Individualised Production

1
Figure 7: Systematical approach for designing new SLM production systems

Approach
2
A possible solution needs to be based on a holistic
analysis of the SLM process, and the development of a re-
ference architecture for mould-less production systems.
This holistic approach is necessary to harness SLM for
3 serial production. To improve the SLM machine techno-
logy, multiple concepts of each SLM machine module are
developed and evaluated in a systematic approach (see
fig. 7). Achieving nearly 100% utilisation of the laser,
and minimal non-productive time is the major guideline
4 during evaluation and creation of a complete next-
generation SLM machine (see fig. 10). The new concept
for the laser utilisation calls for using the laser while the
powder is distributed (on-the-fly). Also, using multiple
laser sources in parallel will significantly reduce primary
5 processing time.
An innovative process chamber design should increase
the handling of the SLM machine. Multiple designs were
created and evaluated (see figures 8 and 9), with the
sliding system with two scanners systems and two se-
parate lasers having the most positive impact on handling
and process efficiency.
To increase build-rate even more, High Power Selective
Laser Melting (HP-SLM) was qualified for the maraging
steel 1.2709. Compared to phase one, adapted process
management and a further increase of laser power up to
2 kW will leverage the theoretical build-rate from 14 mm
up to 23 mm.

Figure 8: Concept 1: Rail framework Figure 9: Concept 2: Jib-arm framework

8
 Individualised Production 1

1
Figure 10: Overall concept of the SLM-machine (CAD model)

Afterwards, tensile test specimens are manufactured and
tested to prove the mechanical properties. Finally, a func-
tionally optimised mould insert used for profile extrusion
is manufactured utilising HP-SLM, to show the potential
on a real application part (see fig. 5).
To fully exploit the geometric potential of SLM, carefully
chosen lattice structures are manufactured and tested
for mechanical properties (see fig. 11 and 12). Those will
form the basis of a database for mechanical properties
of a wide range of lattice structures, to help optimising
additively manufactured parts regarding weight, function,
resource efficiency and to fully utilise the geometrical
potential of SLM manufactured parts.
Technical Challenges
2
Major technical challenges are developing and shaping
innovative concepts for the different machine modules,
in order to achieve a nearly 100% utilisation of the laser
source, and minimise non-productive time. The degree
of novelty of the chosen concepts, e.g. the multi-beam 3
concept, also requires adapted process management in
the overlapping areas, and is managed on innovative
control system software, as well as innovative machine
components.
4
5

Figure 11: Lattice structures fabrication by SLM Figure 12: Mechanical tests on lattice structures

9
 1 Individualised Production

Technical Data Project Team

2
Multi-beam concept Prof. Wolfgang Bleck
On-the-fly-laser exposure Department of Ferrous Metallurgy (IEHK), RWTH
Build envelope: 400x200x300 mm Aachen University
High Power SLM (HP SLM) process with up to 2 kW Prof. Frank T. Piller
3 laser power and build-rates of up to 23 mm/s Technology and Innovation Management (TIM),
20-50% weight saving due to optimised, additively RWTH Aachen University
manufactured parts Prof. Reinhart Poprawe
Chair for Laser Technology (LLT), RWTH Aachen
University
4 Technical Infrastructure Prof. Gnther Schuh
Laboratory for Machine Tools and Production
SLM machines: Engineering (WZL), RWTH Aachen University
HP-SLM lab system S Arne Bohl
M Solutions 280 HL Laboratory for Machine Tools and Production
5 EOS EOSint M270 Engineering (WZL), RWTH Aachen University
Further R&D systems Jan Bltmann
Department of Ferrous Metallurgy (IEHK), RWTH
SLM-Software: Aachen University,
SLM Control (in-house development) Sebastian Bremen
Marcam Autofab 1.8 Fraunhofer Institute for Laser Technology (ILT)
Materialise Magics 17 Christian Hinke
Chair for Laser Technology (LLT), RWTH Aachen
Examples of testing equipment: University
Zeiss light microscope Axiophot Simon Merkt
Infinite Focus Alicona Imaging GmbH Chair for Laser Technology (LLT), RWTH Aachen
Zeiss LEO 1455 EP (REM + EDX) University
Schenck POZ 0673 (nominal force: 1000 N) Johannes Schrage
Wick Z100 (nominal force: 100 kN) Chair for Laser Technology (LLT), RWTH Aachen
University
Christian Weller
Technology and Innovation Management (TIM),
RWTH Aachen University

10

Automated Optimisation for
High-pressure Die Casting Dies
Introduction
High-pressure die casting (HPDC) is a common industri-
al mass production process. In HPDC, molten metal is
injected into a reusable steel die at high pressure. Due to
repetitive filling however, exists an intensive thermal flow
from the molten metal to the die. In order to maintain a
stable temperature level of the die, it is therefore man-
datory to cool the die, so a temperature control system
exists built up by cooling channels inside the die. The
main advantages of HPDC are its high productivity, the
good surface quality of the castings, and the capability
to produce thin-walled structures. On the other hand,
the manufacturing effort of the dies is relatively high,
so that HPDC is particularly efficient for large lot sizes.
During the last couple of years however, the demand for
individualised products has increased, so that the aim
of the demonstrator for the automated optimisation of
Figure 2: Components of a gating system
Individualised Production 1
2
3
HPDC-dies, is now to make the mass production process
of HPDC more suitable for smaller lot sizes.
Practical Issues 4
The gating and the temperature control systems are
crucial for the design of an HPDC-die. So far there are
no standardised procedures that guide the die-designer
towards an ideal die. Nowadays, the toolmaker designs 5
the die based on empirical formulas and his experience.
For complex parts, casting-simulations are used to reach a
good design for the gating system. Usually a lot of simu-
lations need to be done in order to achieve a satisfactory
filling behaviour of the die. This iterative and therefore
time-consuming approach makes the design process
Figure 3: Temperature control system of a HPDC-Die
11
 1 Individualised Production
1 Figure 4: High-pressure die casting machine
expensive and causes the main part of the total amount
of costs for the die design. As long as the HPDC-process
2 is used for mass production this is not a major issue.
The high costs of the dies are compensated for by the
large amount of parts that are produced with one die.
For individualised production however, lot sizes tend to
be much smaller, and the die costs are distributed over
3 fewer parts resulting in the newly cast part not being cost
competitive.
Approach
4 whole die including the temperature control system. In
The motivation of the automated optimisation approach
for HPDC is to find an optimal die design with minimal
effort by the die-designer, and the quality of the resulting
casting should be improved. With this improvement of
5 efficiency in the die-design process, a mass production
Figure 5: Simulation of the filling process (Image: starcast)
12
process such as HPDC can be operated economically even
for lot sizes of individualised production. The automated
optimisation approach has already been tested successful-
ly for the plastics profile extrusion process. This has also
been part of the Cluster of Excellence.
The starting point for the optimisation is the numerical
simulation of the HPDC-process. The simulation data is
used to evaluate the filling and solidification behaviour in
the die. Also, the temperature distribution is examined.
All of this can be achieved without running expensive
experiments. These simulations take into account the
addition to that, a quality measure is needed that allows
the evaluation of the overall quality, a so-called objective
function. The aim of the optimisation is the minimisation
of this objective function, which is developed in close
collaboration with casting experts. The last component
Figure 6: Simulation of the temperature control system (Image: starcast)

Figure 7: Optimisation cycle
that is needed for the optimisation is the representation
and deformation of the geometry. In this regard it is
important to determine the constraints for the geometry
deformation (in this case: for the gating- and temperature
control system). Again the experience of casting experts
must be taken into account.
The automated optimisation process is designed as
follows: Using a start-geometry a numerical simulation
of the filling process is performed. Based on the results
of this simulation, the objective function is evaluated. An
optimisation algorithm now uses the objective function
and generates a new geometry by deforming the old
one, eg., for the gating system a free-form deformation
as well as a translation is possible. The modified geo-
metry is then used as the basis for the next optimisation
cycle. This process is reiterated until the results show the
desired quality. A die, which is manufactured, based on
Figure 8: Freeform deformation of the runner
Individualised Production 1
such an optimisation, can reduce the amount of manual
simulations/trial-runs, and therefore the overall manufac-
turing costs. In this context, the application of generative 2
manufacturing processes for die manufacturing will be
investigated as well. Using an exemplary geometry, the
simulation results will be validated and the numerical mo-
dels will be improved with the help of the gained data.
3
Technical Challenges
A big challenge for the optimisation is the numerical
modelling of the filling process. This has two reasons. 4
On the one hand, the filling process lasts only fractions
of a second, and on the other hand the metal flow is
very complex. Due to the high flow-velocities, the melt is
exposed to high pressures in the range of 300-1200 bar,
and the melt front underlies turbulent movements. The 5
Figure 9: Translation of the runner
13
 1 Individualised Production

1 Figure 10: Modular HPDC-die

complexity of the modelling is even further increased as tive functions used for the HPDC-process are a complete
2 the solidification of the molten metal also needs to be filling of the cavity, or the compliance to certain tempe-
taken into account. This leads large variations in the ma- rature-boundaries for the die. In addition, an approach
terial properties like density and viscosity leading inevitab- needs to be developed for a simple and robust way to
ly to higher computational costs. In order to perform the deform the geometry during the optimisation process.
numerical investigations in a reasonable amount of time,
3 big computing clusters have to be used, thus implying
that the simulation tools must be able to support them. In
the context of automated optimisation, the definition of
a suitable objective function is another challenge. When
developing an objective function several parameters need
4 to be taken into account. Possible ingredients for objec-

Figure 11: Principle structure Figure 12: SLM-Insert with cooling channels

14
 Individualised Production 1

Technical Data Project Team

2
P-/T-measurement at the die Prof. Marek A. Behr
Simulation results for filling and solidification Chair for Computational Analysis of Technical Systems
3D-Data of the die, the runner and the temperature (CATS), RWTH Aachen University
control unit Prof. Andreas Bhrig-Polaczek
Foundry Institute (GI), RWTH Aachen University 3
Prof. Gnther Schuh
Technical Equipment Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Experiments: Abassin Aryobsei
Laboratory for Machine Tools and Production 4
Hardware: Engineering (WZL), RWTH Aachen University
HPDC unit Bhler H-630 SC Frank Schmidt
Locking Force 7250 kN Foundry Institute (GI), RWTH Aachen University
Plunger stroke 580 mm Markus Frings
Real-time control unit Chair for Computational Analysis of Technical Systems 5
(CATS), RWTH Aachen University
Software: Yann Queudeville
Pro/ENGINEER Foundry Institute (GI), RWTH Aachen University
Henning Sauerland
Chair for Computational Analysis of Technical Systems
Simulation: (CATS), RWTH Aachen University

Hardware:
Bull Cluster, IT Centre RWTH
27880 Cores, Intel Westmere/Nehalem
94192 GB Ram
3D Visualisation (Powerwall)

Software:
STAR-CCM+, STAR-Cast
Optimisation framework in Fortran/Python
Paraview

15
 1 Individualised Production

Automated Optimisation of
1
Profile Extrusion Dies

Introduction

2 Profile extrusion is an established process for the mass

production of plastics profiles. However, the develop-
ment and manufacturing of the profile die is combined
with high efforts and costs. These are mainly caused by
elaborate running-in trials before the start of production.
3 Because of this, a cost-effective use of profile extrusion in
small series is not possible yet. Nevertheless, the customer
demands are increasingly individualised, and product life
cycles continue getting shorter resulting in smaller lot
sizes. Numerical flow simulations have the potential to
4 provide substantial time and cost savings in the design of
profile dies. However, the currently available simulation
tools lack the ability to automatically improve the flow
channel geometry based on calculation results. There-
fore, an optimisation framework for the design of profile
5 extrusion dies is being developed in the demonstrator
for the automated optimisation of profile extrusion dies.
The framework combines the flow simulation with an
Figure 2: Large variety of plastics profiles (Images: Dllken) optimisation algorithm. This combination enables an
automated optimisation of the flow channel geometry in
profile extrusion dies.

Figure 3: Design features in the flow channel

16

Figure 4: Approach in the design and manufacturing of profile extrusion dies
Practical Issues
Due to the almost unlimited complexity of possible profile
cross-sections, an analytically exact design of the profile
extrusion die is usually not possible. Instead, the design
and manufacturing often take place in form of a trial
and error process. Starting from a first draft, the die
designer changes the flow channel geometry iteratively
based on empirical knowledge and insights from the
running-in trials. This happens via deposition or removal
of material until the produced profile satisfies the previ-
ously defined quality criteria. Depending on the complexi-
ty of the geometries this can take up to 15 iterations. This
type of production is very time-consuming and costly, the
extruder is not available during the running-in trials for
the regular production and considerable costs occur for
the experimental materials. A shortening of the design
Figure 5: Running-in trials in a profile extrusion line
Individualised Production 1
1
process is offered by flow simulations with the stabilised 2
FEM (Finite-Element-Method) which transfers the iterative
design process from the machine to the computer.
Through the numerical analysis of the flow processes, the
running-in process is simplified as vulnerabilities can be
detected even before the manufacturing of the die. 3
However, in this case the experience and the knowledge
of the die maker are still necessary to decide how the
geometry has to be changed, in order to achieve the
design goals. Therefore, the design process is still domi-
nated by its iterative, time-intensive nature. 4
5
Figure 6: Flow simulation of a die
17
 1 Individualised Production
Figure 7: Automated optimisation of the flow channel with the optimisation framework
Approach
2 The fundamental approach to finding a solution to the
problem is to imitate the real process of continuous
improvement of the die. The realisation of this process is
the usage of optimisation algorithms, which minimise a
so-called objective function. The objective function mainly
3 reflects the quality of the die. In order to simulate the real
process, it is necessary to model the reality within certain
limits. Here we use several methods that replicate all the
relevant properties and the behaviour of the real process,
but are still calculable by high performance computers
4 or ideally ordinary computers within acceptable times. It
is crucial for this optimisation application, to get to the
target in as few steps as possible.
The background is, that one tries to simulate as few
5 geometries as possible, because each simulation is time
consuming and computationally intensive and, in the
Figure 8: Extrusion of a plinth profile
18
end, again costly. In this respect, this demonstrator goes
beyond the numerical and experimental components
and integrates best practices of the industry and the
manufacturing process. Another future step is the use
of SLM (Selective Laser Melting) produced components
and modules within the dies, so that previously not only
manufacturable but also rapidly deployable modules
could be used.
Technical Challenges
Due to the non-intuitive behaviour of the plastics, the
representation of reality is one of the biggest technical
challenges. The core property which complicates the
behaviour of the plastics to such a degree is called visco-
elasticity. Viscoelasticity is defined as the superposition
Figure 9: Swelling of the extruded profile
 Individualised Production 1

Figure 10: Die for the extrusion of a L-shaped profile

2

of elastic and viscous flow behaviour; its mathematically

description is very complicated. This additional complexity
also flows into the simulation. Furthermore, the definition 3
of the objective function has a decisive influence on the
optimisation process. Both the choice of the measured
quantity and the actual mathematical formulation play an
important role. Ultimately, the goal is to achieve the ma-
ximum efficiency in the implementation of the parametric 4
design and optimisation algorithms and hereby realise a
minimum number of numerical iterations. It is necessary
and also especially challenging to involve institutions into
a close work process to validate theoretical simulations
in practical experiments and vice versa, to influence the 5
models.

Figure 11: Simulation of the profile swell Figure 12: SLM in the die manufacturing (Image: Fraunhofer ILT)

19
 1 Individualised Production

Technical Equipment Project Team

2 Experiments: Prof. Marek A. Behr

Chair for Computational Analysis of Technical Systems
Hardware: (CATS), RWTH Aachen University
Brabender laboratory extruder 19/10 DW Prof. Andreas Bhrig-Polaczek
D = 19 mm Foundry Institute (GI), RWTH Aachen University
3 L = 25 D Prof. Christian Hopmann
Throughput 0,8 to 4 kg/h Institute of Plastics Processing (IKV) at RWTH Aachen
University
Prof. Gnther Schuh
Simulation: Laboratory for Machine Tools and Production
4 Engineering (WZL). RWTH Aachen University
Hardware: Abassin Aryobsei
Bull Cluster, IT Center RWTH Laboratory for Machine Tools and Production
27880 Cores, Intel West-mere/Nehalem Engineering (WZL), RWTH Aachen University
94192 GB Ram Frank Schmidt
5 3D Visualisation (Powerwall) Foundry Institute (GI), RWTH Aachen University
Markus Frings
Software: Chair for Computational Analysis of Technical Systems
XNS (CATS), RWTH Aachen University
Paraview Yann Queudeville
Foundry Institute (GI), RWTH Aachen University
Roland Siegbert
Chair for Computational Analysis of Technical Systems
(CATS), RWTH Aachen University
Nafi Yesildag
Institute of Plastics Processing (IKV) at RWTH Aachen
University

20

Virtual Production Systems
The research area Virtual Production Systems improves
and connects virtual models on all levels of the produc-
tion system and along the value chain starting with the
setup of comprehensive and consistent models. These
can then be tailored for specific applications. The aim
is to radically shorten the product development process
for gaining productivity growth already in the develop-
ment phase of production technology. The focus lies on
the support, the development and the optimisation of a
production system, as well as on the decision capability of
the planner. The virtual try-out of numerous alternatives
improves understanding of the behaviour of produc-
tion systems and its elements, and identifies key levers.
Compared to testing in reality, cost and time advantages
can be achieved by virtual simulations. Todays computing
power and new methods of data management open up
new possibilities to process and visualise data at high
speed. The main focus of vertical integration lies on data
mining, processing and visualisation of information on
all levels of the production system, from logistics on the
Virtual Production Systems 2
2
factory level to the behaviour of the work piece in the
manufacturing process, e.g. laser cutting. The key term
Virtual Production Intelligence describes the application-
specific visualisation and decision-making support of the 3
planner.
The horizontal integration on the other side places the
emphasis on consistent simulations along the value chain.
Here the behaviour throughout the product life cycle is 4
simulated and forecasted beginning with the moulding of
the material through forming and cutting to the finishing
and use of the product. All models are merged on a
standardised, modularly extensible platform Integrative
Computational Materials and Production Engineering 5
(ICMPE). In this way all influences of the material struc-
ture on further processes can be taken into consideration
in corresponding models. Furthermore, the standardisa-
tion of the simulation interfaces is pushed forward at the
global level.
21
 2 Virtual Production Systems

2
Virtual Production Intelligence goes
Virtual Reality
3
Introduction

Based on the methods of information integration and
4 visualisation explored within the first funding period of
the research area The Ontology and Design Methodo-
logy of Virtual Production Systems, the research area
Virtual Production Intelligence consistently continues
the research on virtual production.
5 The vision of this project is to reduce the plan-value
During the second funding period of the Cluster of
Excellence, the focus is on the systematic analysis of infor-
mation from the production environment. This includes
the identification, integration, extraction, analysis and
visualisation of production information. The concept of
Virtual Production Intelligence (VPI) is based on success-
ful, already well-established concepts of information
integration and aggregation as well as their evaluation
and interpretation by experts in the relevant application
domains.
dilemma by facilitating the aggregation and propagation
of heterogeneous information generated within processes
from real and virtual production environments.

Figure 2: Interactive exploration of data Figure 3: Trajectories-provided navigation

22

Figure 4: Factory planning shown in the plan view
Practical Issues
In established factory planning processes, an inspection of
the factory is not possible until its completion. Due to this
limitation, natural effects, such as the visibility between
workstations, cannot be considered during the planning
phase. In many projects, such issues have negative effects
on the planning result.
Another important aspect is the ideal configuration of
machines for the planning of a factory. The variety of
adjustment possibilities that have impact on the produc-
tion result leads to a large parameter space which can
no longer be intuitively grasped. As a consequence, ideal
configurations for individual processes often cannot be
found and are therefore substituted by approximations
based on experience or technology tables
Figure 5: Workstation at machinery within the aixCAVE
Virtual Production Systems 2
2
Approach
The first scenario examines the inspection on the factory
floor in a fully immersive environment. Firstly, a factory 3
model that has been developed by means of the factory
planning tool visTABLE is loaded into the aixCAVE and
made available for exploration. The user can observe his
immediate surrounding from different perspectives by
physically walking through the aixCAVE, while remote 4
areas can be reached by intuitive navigation techniques.
Furthermore, the model can be analysed for compliance
with relevant criteria: the stereoscopic representation and
natural mobility enable, for example, the visual accessibi-
lity of workplaces or the estimation of their layout. These 5
possibilities of immersive visualisation represent a signifi-
cant enhancement compared to previous planning tools.
Figure 6: The factory from a different perspective
23
 2 Virtual Production Systems
Figure 7: Hyper slice-visualisation of a meta-model for oxygen cutting
1
The second scenario is an approach to visualising so
called meta-models for the examination of production
processes. These meta-models map multiple machining
2 parameters to certain criteria, for example, the impact of
gas pressure or cutting speed on the roughness of the cut
in laser cutting. To explore the resulting multi-dimensio-
nal parameter space, a visualisation application based on
HyperSlice has been developed. The technique represents
3 the parameter space as a matrix of several two-dimensio-
nal slices, which illustrate every possible pairwise combi-
nation of the axes of the parameter space.
Additionally, a three-dimensional volume visualisation
4 containing a three-dimensional slice through the parame-
ter space that is individually chosen by the user is shown.
The immersive presentation of the application allows the
exploration of the parameter space for identifying ideal
configuration parameters for the desired process result.
5 Moreover, this presentation provides the possibility to
identify interdependencies between output variables and
Figure 8: Visual axes in the factory
24
individual input parameters. Such analyses can be used in
research as well as in education.
Technical Challenges
The first step of information integration in both scenarios
consists of information modelling. Subsequently, informa-
tion is integrated into the Virtual Production Intelligence
platform. This platform provides, among others, the inte-
gration and provision of information by using a service-
oriented architecture enabling, e.g. the connection of
various analysis-methods and appropriate visualisation.
Simultaneously to the creation of the factory model
through visTABLE touch, the virtual access to the model
shall be possible. It is therefore a requirement that the
factory model is not just loaded into the immersive envi-
ronment but will also be continuously synchronised with
modifications made in visTABLE touch. This requires a
Figure 9: Virtual conference room

Figure 10: The view out of the warehouse
communication channel between the immersive environ-
ment and visTABLE touch that transmits layout changes
such as creating, deleting or moving of machine models.
Therefore, a transmission with only short delay needs to
be realised allowing for a workflow without unintended
interruptions.
Another challenge is the navigation through the factory
model. While it is possible to overcome short distances
by real walking within the aixCAVE of RWTH Aachen
University, greater distances have to be covered by using
appropriate navigation techniques due to limited physical
extents of the immersive system. Their implementation
calls for special input devices allowing for an easy and
quick operation.
The biggest technical challenge for the HyperSlice-based
visualisation of multi-dimensional metamodels is to
guarantee a sufficient interactivity of the application.
The creation of slices through the parameter space is
extremely CPU-intensive and can take several minutes
Figure 11: aixCAVE at RWTH Aachen University
Virtual Production Systems 2
1
for high resolutions. To address this issue, methods like
progressive rendering and task-based parallelisation
with user-centered scheduling are used. This allows an
immediate display of visualisations in low-resolution while 2
the simultaneous interaction with the application remains
possible. This allows the user to get a first impression of
the current situation while navigating through the para-
meter space. As soon as the user stops navigating, the
resolution of the visualisation is continually increased. 3
4
5
Figure 12: Another scenario within the aixCAVE
25
 2 Virtual Production Systems
1
Technical Data
Factory planning
2 VRML-based geometrical data
Transformation data via network for synchronisati-
on with visTABLE touch
Metamodels
Radial Basis Function-based multi-dimensional
3 metamodels
VPI platform
Platform for collaborative factory planning and
analysis of laser cutting processes
Provision of information integration und domain-
4 dependent processes for data analyses
Based on an implementation of a service-oriented
architecture
5 Technical Equipment
aixCAVE
Five projection surfaces with rear projection, (four
walls and floor) with 24 active-Stereo projectors
each with 1920 x 1200 pixels resolution and
12.000 ANSI lumen brightness, dimensions:
5,25 m x 5,25 m x 3,30 m
Visualisation cluster with 24 nodes, each with two
NVIDIA Quadro 6000 GPUs and two Intel Xeon
Westmere CPUs (six cores with 2,7 GHz), connected
via Infiniband QDR
Project Team
Prof. Sabina Jeschke
Institute of Information Management in mechanical En-
gineering (IMA) & Center for Learning and Knowledge
Management (ZLW) & Associated Institute for Manage-
ment Cybernetics e.V. (IfU), RWTH Aachen University
26
Prof. Torsten Kuhlen
Virtual Reality Group (VRG), RWTH Aachen University
Prof. Daniel Schilberg
Institute of Information Management in mechanical En-
gineering (IMA) & Center for Learning and Knowledge
Management (ZLW) & Associated Institute for Manage-
ment Cybernetics e.V. (IfU), RWTH Aachen University
Prof. Wolfgang Schulz
Fraunhofer Institute for Laser Technology (ILT) &
Chair of Non-linear Dynamics of Laser Processing (NLD),
RWTH Aachen University
Toufik Al-Khawli
Chair of Non-linear Dynamics of Laser Processing (NLD),
RWTH Aachen University
Christian Bscher
Institute of Information Management in mechanical En-
gineering (IMA) & Center for Learning and Knowledge
Management (ZLW) & Associated Institute for Manage-
ment Cybernetics e.V. (IfU), RWTH Aachen University
Urs Eppelt
Chair of Non-linear Dynamics of Laser Processing (NLD),
RWTH Aachen University
Sascha Gebhardt
Virtual Reality Group (VRG), RWTH Aachen University
Sebastian Pick
Virtual Reality Group (VRG), RWTH Aachen University
Rudolf Reinhard
Institute of Information Management in mechanical En-
gineering (IMA) & Center for Learning and Knowledge
Management (ZLW) & Associated Institute for Manage-
ment Cybernetics e.V. (IfU), RWTH Aachen University
Hanno Voet
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Julian Utsch
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University

Development of Aluminum-Free Case
Hardening Steel
Introduction
In this project, a simulation platform for the integrative
development of materials and process chains is introdu-
ced. This simulation platform is applied to the numerical
development of aluminum-free case hardening steel with
improved cleanliness for large gears in the wind industry.
In order to validate the new numerical based combined
material and process concept, new gear steel developed
Figure 2: Gear damage of a wind turbine
Virtual Production Systems 2
3
by simulation will be produced on a laboratory scale, 4
forged to a bevel gear and carburised to demonstrate its
suitability for high-temperature carburisation. In future,
the industrial implementation will be established with a
consortium of industrial partners along the entire process
chain for large gear components. 5
Figure 3: Tooth fracture of a large gear wheel
27
 2 Virtual Production Systems
1
Figure 4: AixViPMaP Aachen (Aix) Virtual Platform for Materials Processing
2 Practical Issues
Improving the service life of carburized gear components
is a high priority for wind turbine manufacturers. Only a
reliable prevention of early failures ensures economic
3 operation of wind turbines. The service life of gear
components depends on many factors which are set
during material production, forming or heat treatment.
Important influencing parameters such as phase fraction,
grain size and grain size distribution can be reset again
4 by an adapted heat treatment at any point in the process
chain. However, microscopic oxide purity is an important
endurance limiting issue, which is set during the
metallurgical production of steel and is irreversible from
then on in the solid material. The improvement of the
5 level of purity without deterioration of the fine grain
stability is a high priority for steel manufacturers and
wind turbine operators.
Figure 5: Time savings in HT-carburising
28
Approach
A material based approach for improving the oxide
purity level represents the reduction of the aluminum
content, which decreases the likelihood of lifetime-
reducing aluminum oxides in the material and thus a
premature failure of the wind turbine. However,
aluminum is used in the further course of the process
chain in combination with a defined addition of
nitrogen also for ensuring the fine-grain resistance,
which in turn is responsible for setting the tooth
strength. Metallurgically this effect is based on
aluminum nitrides that are finely dispersed in the
microstructure which exist as finely dispersed
precipitates on the nanometer scale and reduce the
grain boundary mobility.
Figure 6: Integrative simulation of precipitation

Figure 7: Numerical alloy design for a micro-alloyed Al-free case hardening steel
Technical Challenges
Purity grade-related early failures are to be avoided for
the economic use of wind turbines. Aluminum-free
microalloyed case hardening steels show an improved
microscopic oxide purity level and promise a reliable use
of large gears. Moreover, they are produced time and
energy reduced by the so-called high-temperature carbu-
rising process allowing an efficient manufacturing route.
The development of the current aluminum-free alloy
concept is based on a combination of modern numerical
simulation methods and aims at the substitution of alumi-
num nitrides by niobium carbonitrides. The basis for the
combination of different model approaches is a simulati-
on platform for the calculation of process chains, which
allows the calculation at all relevant levels of observation
at the nanoscale of the precipitates, at the micro scale of
the microstructure and at the macro scale of the compo-
nents. Based on this multiscale approach the optimisation
of alloys and process parameters can be realised in a
holistic manner.
Figure 8: Process chains of gear manufacturing
Virtual Production Systems 2
1
In order to implement an integrative, simulation-based
development of materials and processes, several different 2
simulation approaches and programmes are necessary.
The efficient and effective use of various programmes is
ensured by developing an Internet-based platform for the
simulation of process chains. In doing so, uniform formats
for data and visualisation were defined. This approach al- 3
lows simultaneous simulation and analysis on all relevant
scales from nano- to component-scale.
The test case Al-free case hardening steel deals with
quantification of the fraction and size evolution of 4
microalloying precipitates along different alternative
process chains. On this basis it allows to predict the fine-
grain stability during carburisation as a function of the
process chain, the process parameters and the chemical
composition. 5
Figure 9: Possible grain size distributions
29
 2 Virtual Production Systems
Figure 10: Detection of the improved level of purity and the shortened process chain on a laboratory scale
2
For validation of the purely numerically developed alloy
concept, a laboratory melt was created, processed into
a bar and examined with respect to purity and fine grain
3 resistance at high carburising temperatures. Eventually,
good fine-grain stability has been documented for the
modified variant after blank-hardening at 1050 C for
12 hours or at 1100 C for 1 hour for a process chain
with FP-annealing.
4
Future Work
In order to achieve the fastest possible industrial imple-
5 mentation of the new combined material and process
concept in large gear components, an AVIF project is
Figure 11: Detection of fine grain resistance
30
started together with FVA and FOSTA. The consortium is
composed of various steel manufacturers, a ring rolling
company and a wind turbine manufacturer and two
research departments, namely the Steel Institute (IEHK)
at RWTH Aachen University and the Institute for Material
Technology (IWT) at University Bremen. In this combinati-
on, the entire production chain for large gear componen-
ts including component application is investigated.
Internationalisation Strategy
As part of an EU-funded Coordinated Support Action
(CSA), an expert group for ICME is formed: ICMEg. This
group of experts will develop a future data standard for
ICME-Tools during two international workshops with
non-European participation. For sustainable maintenance
of this ICMEg standard, an Association ICMEg e.V. is for-
med, where research institutes, software developers and
ICME-User from industry are involved.
Figure 12: Demonstrator of a bevel gear
 Virtual Production Systems 2

Technical Data Dr. Markus Apel

ACCESS e.V. 2
Aluminium oxides are responsible for 80 % of purity Dr. Markus Bambach
level-related gear failure Institute of Metal Forming (IBF), RWTH
For numerical material and process developments 2/3 Aachen University
of the time are consumed for data conversion Jens Dierdorf
High temperature carburising allows 60 % time Institute of Metal Forming (IBF), RWTH 3
reduction in case hardening Aachen University
Patrick Fayek
Department of Ferrous Metallurgy (IEHK), RWTH
Technical Equipment Aachen University
Jannik Henser 4
Several workstations and material modelling tools Laboratory for Machine Tools and Production
3D-VR-ICME-Center Engineering (WZL), RWTH Aachen University
Vacuum furnace for the production of lab-sized steel Viktor Kripak
casts Department of Ferrous Metallurgy (IEHK), RWTH
Semi-Product-Simulation-Center Aachen University 5
Forging robot Dr. Gottfried Laschet
Various laboratory furnaces (vacuum and protective gas ACCESS e.V.
atmosphere) Hamidreza Farivar
Metallography and quantitative image analysis software Department of Ferrous Metallurgy (IEHK), RWTH
Scanning electron microscope Aachen University
Gearbox test rig Dr. Ulrich Prahl
Department of Ferrous Metallurgy (IEHK), RWTH
Aachen University
Project Team Dr. Georg J. Schmitz
ACCESS e.V.
Prof. Wolfgang Bleck Andre Messias Teixeira
Department of Ferrous Metallurgy (IEHK), RWTH Fraunhofer Institute for Production Technology (IPT)
Aachen University
Prof. Gerhard Hirt
Institute of Metal Forming (IBF), RWTH
Aachen University
Prof. Fritz Klocke
Fraunhofer Institute for Production Technology (IPT)
Prof. Christian Brecher
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University

31
32

Integrated Technologies
The research area Integrated Technologies aims at
productivity growth within production by a drastic shor-
tening of the value chain. Integrating and substituting
process steps along the value chain allow eliminating
interfaces that would otherwise lead to disruptions. The
development of multi-technology production systems, ho-
wever, is challenging since more interdependencies have
to be considered. Examples of multi-technology produc-
tion systems that have been developed within the cluster
are platforms of machining and sheet metal processing.
Within the cluster both parallel and combined processing
on one single machine are being considered, e.g. by inte-
grating a robot into the machining process.
Furthermore, individualisation of mass-production
components is being researched by combining mass-pro-
duction processes, e.g. stretch forming, with prototype
Integrated Technologies 3
processes, e.g. incremental sheet forming. Moreover,
process integration can enable or facilitate the produc-
tion of multi-technology products that integrate different 3
functionalities and materials. Manufacturing processes for
these products are being developed within the cluster,
e.g. processing of metal and plastics within the same
mould. By integrating functionalities, e.g. structuring
the surfaces of foils, materials and products enter new 4
application areas.
Also special treatment of surfaces like coating opens up
new possibilities, e.g. regarding conductivity. In addition,
the systematic development and evaluation of new tech- 5
nologies is pushed forward, in consideration of economic,
ecologic and social challenges.
33
 3 Integrated Technologies

1. Image: Thilo Vogel

Integrative Sheet Metal

3 Processing Centre

Introduction

4 Flexibility is a key issue in modern production technolo-

gies. In recent years, most industries experience a rising
demand for customised products. The manufacturing of
customised products creates opportunities for economic
growth and secures international competitiveness. Mo-
5 dern production systems have to be sustainable and ensu-
re a high product quality. At the same time, they have to
allow for an economic and flexible modification of the
process. Hence, concepts that meet the target First Time Figure 2: Tailgate Audi A4
Rightneed to be developed in order to save time, money
and resources during the development and production of
the product.

The rapid development of industrial sectors processing

sheet metal parts, such as the automotive industry,
requires flexible production processes adapted to their
needs. That way sheet metal parts which are adjusted at
its best to their functionality could be produced cost-
efficiently.

Figure 3: Door element Airbus A320

34

Figure 4: Process variants of Incremental Sheet Forming (ISF)
Challenges in Application
The establishment of a both flexible and resource-effici-
ent production represents a major challenge for produc-
tion engineers. Customised sheet metal parts are usually
characterised by a high complexity and low production
volumes. Common sheet metal forming processes, e. g.
deep drawing and stretch drawing, are in many cases
not capable to meet the requirements of the customers.
Furthermore, a specific forming tool is needed for every
single geometric variety. The production of forming tools,
however, is expensive and time-consuming.
Another issue is the processing of different material
grades. Especially the aerospace industry uses high
strength metals, such as titanium alloys. The possibilities
of forming these materials with common sheet forming
processes at room temperature are limited. Combining
different technologies renders the manufacturing of com-
ponents made of high strength materials with complex
geometries possible.
Figure 5: Forming tool with hemispherical tip
Integrated Technologies 3
Approach
In Incremental Sheet Forming (ISF) the final part is formed
by a sum of localised plastic deformations. A generic 1
forming tool moves along the contours of the desired
part. The forming process can be supported by a full or
partial die. Since the forming tool is independent of the
produced geometry, this production method is excellent
for the manufacturing of prototypes and small production 2
volumes.
A combination of stretch drawing and ISF expands the
production possibilities of the Integrative Sheet Metal Pro-
cessing Centre. The first process step is stretch drawing 3
followed by the incremental shaping of geometric details.
Due to the combination of these two processes the com-
paratively large process time of pure ISF can be reduced
significantly. Additional advantages in the process combi-
nation of stretch drawing and ISF are the homogenisation 4
of the sheet thickness distribution and the increased
geometric accuracy of the entire part. Hence, more com-
plex geometries can be produced within a shorter time. In
general, ISF is characterised by low forming forces due to
the locally limited forming area and a low tooling effort. 5
These features turn ISF into an economic alternative for
industrial sectors with specific product requirements,
Figure 6: Combination of ISF and stretch drawing
35
 3 Integrated Technologies

1
Figure 7: Processing centre for hybrid Incremental Sheet Forming from EIMA Corp. (Image: Thilo Vogel)

such as the aerospace industry. Many geometric varieties, are produced in consecutive processes by combining dif-
2 small production volumes, large parts and the use of high ferent technologies in one single production unit. That al-
strength materials are characteristic for this industry. Ty- lows the customised production of geometrically accurate
pically, titanium alloys which are difficult to form at room components made of sophisticated materials. Simultane-
temperature, e. g. TiAl6V4, are used. Their formability ously, stock costs and set-up times can be reduced.
can be increased by a local heating of the forming area
3 during ISF. To realise ISF of high strength alloys, a laser
optics is integrated in the Integrative Sheet Metal Pro-
cessing Centre. The two introduced hybrid technologies
enable a flexible production of sheet metal parts with an
extended component and material spectrum. The parts
4

Figure 8: Combination of ISF and laser-heating Figure 9: Temperature distribution on the back side of the plate

36
 Integrated Technologies 3

2
Figure 10: Self-supporting sheet metal structure manufactured by Incremental Sheet Forming

Technical Challenges
The integration of different technologies and manufactu-
ring processes is challenging. The different technologies
have to be mounted in a limited workspace and linked
within the control unit. Furthermore, complex forming
strategies have to be implemented to apply the combi-
nation of stretch drawing and ISF. Another task is the
temperature regulation of laser-assisted ISF. During the
3
process neither blank nor forming tool should overheat.
The optimal process temperature for a specific material
has to be analysed and kept constant during the forming
process. Thus, the temperature of the locally formed area
has to be measured and controlled precisely. Suitable 4
solutions have been developed and tested on industry-
related components.

Figure 11: Sheet metal part manufactured with ISF and stretch Figure 12: Laser-assisted ISF
drawing

37
 3 Integrated Technologies

Technical Data Technical Equipment

3 Integrative Sheet Metal Processing Centre from Laser-assisted ISF

EIMA Corp. Combination of ISF and stretch drawing
Combination of ISF and laser-cutting (in process of
Available workspace: planning)
2.0 x 1.5 x 0.7 m (l x w x h) Portal milling machine
4 Max. velocity of the forming tool:
40 m/min for x-, y- and z-direction
Max. force of the forming tool: Project Team
4 kN in x-, y- and z-direction
Max. force of the stretch drawing unit: Prof. Gerhard Hirt
5 200 kN in x- and y-direction Institute of Metal Forming (IBF), RWTH Aachen
100 kN in z-direction University
David Bailly
Diode laser from Laserline Institute of Metal Forming (IBF), RWTH Aachen
University
Wavelenght: 978 1025 10 nm Laura Conrads
Laserpower: 2200 W Institute of Metal Forming (IBF), RWTH Aachen
University
Markus Gottschalk
Institute of Metal Forming (IBF), RWTH Aachen
University
Chris Mertin
Institute of Metal Forming (IBF), RWTH Aachen
University
Holger Voswinckel
Institute of Metal Forming (IBF), RWTH Aachen
University

38

Multi-Technology Machining Centre
Introduction
Rapid changes in global markets require high flexibility
in production. However, the demands on the achievable
product complexity and production profitability should
not be reduced. Therefore, one focus for the design and
selection of the necessary means of production and ma-
chinery lies on maximum productivity. In contrast, further
focus is on the production of any complex components.
These commonly contrary poles scale and scope
are an essential part of the research within the Cluster
of Excellence and can therefore be found in the mission
statement, the polylemma of production technology.
Figure 2: Machine structure
Integrated Technologies 3
3
Practical Issues
The project Development of Multi-Technology Plat-
forms from the interdisciplinary cluster domain Integra-
ted Technologies examines how various manufacturing 4
technologies of traditional individual machines can be in-
tegrated into so-called Multi-Technology Platforms (MTP).
In order to evaluate the success potential of an MTP, the
comparison of an MTP and a concatenation of individual
machine tools of the given production technologies can 5
be performed. Important criteria for a distinction between
MTPs and single machines are accuracy, lead times and
idle times. Since in principal an MTP can perform several
manufacturing processes in a single setup, the operator
of an MTP can achieve higher accuracies and prevent
waiting and setup times.
Figure 3: Decreasing costs per piece for small lots
39
 3 Integrated Technologies
Figure 4: Different laser based manufacturing technologies on a sample workpiece
1
However, individual machines can also operate in parallel,
and thus for a certain minimum number of pieces - at
least theoretically - an occupancy rate of 100% can be
2 achieved. The possibility of utilising all technologies inte-
grated in an MTP in contrast decreases with an increasing
number of integrated production technologies.
3 Approach
To make the benefits of MTP available and to be able to
operate in an economically viable way, a promising soluti-
on is the integration of additional, simultaneously usable
4 workspaces in the MTP. This approach has already been
successfully used for single technology machine tools to
reduce downtimes. To evaluate the technical possibilities
and limitations of an MTP with multiple workspaces, as
well as to evaluate its operation, such an MTP, the Multi-
5 Technology Machining Centre was built up. The basic
machine consists of the 5-axis milling centre Mill 2000
from CHIRON-WERKE GmbH & Co. KG, which already
Figure 5: Motion axes of the robot
40
is equipped with a milling spindle and two equally and
simultaneously usable rotary swivel tables. Within the
project the machine has been extended with an indus-
trial robot and two laser processing units, enabling the
machine to perform several laser processes besides the
main process of milling. One machining head allows wire-
based cladding, hardening and material removal using a
fibre laser. The second unit consists of a short-pulse laser
and a scanner for laser structuring. For the execution of
different laser processes and for the work piece handling
the robot has been fully integrated into the machine
structure and control. Through flexible combination
strategies of the modules spindle, robot and both lasers
with the two workspaces a high utilisation of all system
components can be achieved.
Bild 6: bersicht der Arbeitsrume
Figure 6: Overview of workspaces
 Integrated Technologies 3

Technical Challenges

One central issue during development for the mechanical

construction, as well as for the control conception and
commissioning was the integration of the industrial robot
into the machine. The robot was placed mechanically
on a platform between the two working spaces, which
enabled it to reach both working spaces directly. The
selected control concept is based on the idea to control
both the robot and the spindle from the same numerical
controller (NC). Axes values are generated for the robot
and then sent to the robot controller for transformation
and application. Due to the different signal propagation
times between the NC and the robot controller, it is ne-
cessary to adjust the run times by using control extensi- Figure 7: Machine and active laser protection housing
ons (NC kernel extensions).

Since the machine uses high power laser components,

a sophisticated safety concept is necessary. As even the 1
scattered reflections of laser radiation can already cause
serious eye damage, the entire machine has been sur-
rounded with a double-walled active protective enclosure.
Damages to this enclosure immediately trigger an alarm
which stops all machine resources. 2
Due to the close integration of the described compon-
ents, processes and mechanisms within a small space,
interactions are to be expected. This can e. g. be the ge-
neration of vibrations from a milling process, the thermal- 3
ly caused displacement of the machine bed due to laser
power, or heat accumulation due to the safety enclosure.
Thus, another research focus is the investigation of these Figure 8: Laser deposition welding (Laser off)
interactions on a mechanical, thermal and organisational
level. 4

Figure 9: Laser deposition welding (Laser on)

41
 3 Integrated Technologies

Figure 10: Laser deposition welding experiments

2
Regarding the further development of the demonstra-
tor the integration of additional tools and processes is
planned for the near future. In particular a cooperative
3 processing of robot and spindle for different processes
such as handling operations or the robot-based deburring
of components shall be achieved.

Displacement
Sensors

X-Direction

Y-Direction

Z-Direction

Figure 11: Measuring displacement with a laser tracker Figure 12: Investigation of thermal displacement

42
 Integrated Technologies 3

2
Technical Data

Milling, drilling, laser deposition welding and laser Structured laser

structuring in a single machine Highly pulsed 3
Control of all components by a single NC controller High-precision microstructuring / deburring
All processes can be used in one clamping Led by milling spindle
Two workspaces for optimal use of resources (parallel Peripherals (chip conveyor, cooling and ventilation
processing) systems, hydraulics, etc.)
Several camera and test computers 4
Laser radiation protective enclosure
Technical Equipment
Software:
Hardware: CAM-System
Machines & components: Siemens PLM NX 5
Base machine: Chiron Mill 2000 Extensions for simulation and laser cladding
5-axis milling centre (self-developed)
Two identical workspaces, each with a swivel and Laser control
rotary axis Lasernet
Spindle power: 25 kW Inscript
Siemens 840D sl with own NCK extensions
Robot: Kuka KR-16
6-axis industrial robot Project Team
Payload: 16 kg
Deposition welding laser Prof. Christian Brecher
Laser source vendor: IPG Laboratory for Machine Tools and Production
Laser scanner vendor: Arges Engineering (WZL), RWTH Aachen University
Type: fiber laser Yb:YAG Frdric du Bois-Reymond
Power: 2.4 kW Laboratory for Machine Tools and Production
Welding head with scanner and media supply (Ad- Engineering (WZL), RWTH Aachen University
ditional wire and shielding gas) Wolfram Lohse
Led by robot Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Johannes Nittinger
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University

43
 3 Integrated Technologies
Monolithic Plastics Optics
1 nology, and miniaturised sensors.
Introduction
The properties of plastic components are not only defined
by their base material but also to a higher degree by the
2 design of the surface. In many cases, surface structures
with a size of a few micrometres are sufficient to drasti-
cally alter the surface properties, for example to avoid the
adhesion of dust particles or water droplets (Fig. 2) or to
create an optically functional surface (Fig. 3). Caused by
3 the progressing miniaturisation in sensor systems, image
and data processing, conventionally used macroscopic
optical systems are increasingly replaced by micro-optical
systems. Additionally, the general demands for compact
designs in the consumer market as well as the gene-
4 ral challenge in the adjustment of miniaturised optical
systems contribute to the increasing integration of micro
optical systems into monolithic components.
Due to the size of typical optical components, they are
5 well suited for mass production by injection and extrusion
moulding. Such monolithic, multifunctional plastics optics
Figure 2: Superhydrophobic microstructures (Image: Thilo Vogel)
44
are used in new imaging applications, such as light-field
photography and microscopy, display and security tech-
Practical Issues
Drawbacks of established design and production methods
are:
Insufficient flexibility relating to the design geometry of
individual micro-optical elements
Limited compactness of optical components due to
required functionality
Moulding precision and deviations for optical elements
For a future increase in efficiency and functionality of pro-
ducts based on hybrid optics, it is necessary to improve
the whole production chain by a holistic approach which
improves not only design and production methods itself,
but also views each step in respect to the other ones.
An example for non-imaging micro-optical systems that
is suitable for monolithic integration is the backlight of
liquid crystal displays that in conventional designs consists
Figure 3: Microstructured optic
 Integrated Technologies 3

2
Figure 4: Division of the production chain into injection and extrusion moulding

of several components. Other areas of application are

beam shaping (laser technology), security features (holo- 3
grams) or warning labels (reflexion).

Approach
4
In the test case Optical components, we are working
on an optimised and holistic process chain (Fig. 1) for
the design and production of optically functional plastics
components. The process chain begins with the design of
the hybrid micro-optical components. We are developing 5
a design method that allows for the optimisation of an
optical surface with a high degree of freedom, optionally
including deviations from the target geometry during pro- Figure 5: Prism under production constraints
duction (Fig. 5).

In a second processing step the structures resulting from

the design process are transferred into a moulding tool
with an ultra-short pulse (USP) laser (Fig. 8). Through this
so called cold ablation by USP laser structure sizes in
the low m regime can be produced, which largely occurs
without any melting (Fig. 6).

For the production of complex functional 3D surfaces

with foils and plates a variothermal extrusion or injection
moulding process is investigated (Fig. 4). The utilised
moulding tool is therefore either a laser structured,
coated embossing roller (Fig. 11) or a laser structured and
coated injection tool. Figure 6: Height profile of the negative structures

45
 3 Integrated Technologies

Figure 7: Variations of the surface geometry by the adjustment of the variothermal moulding process
1

To improve the moulding precision and release properties

of tools, a coating is applied. These coatings are applied
2 by Physical Vapour Desposition (PVD) and reduce the
adhesion between the surface of the tool and the moul-
ded part to increase the precision and wear resistance
of the tool. A coupling of the design and production
process is realised by an iterative process, which consists
3 in measuring the deviation of the produced component in
relation to its simulated counterpart. These deviations are
then used as an additional constraint in the design phase.

The functional component or foil (Fig. 9) is a demonstra-

4 tor application for test case optical components. The
function of surface structures is checked and designed Figure 8: Laser structuring of the moulding sleeve
by the Chair of Optical Systems (TOS). The structure is
generated by the Chair of Laser Technology (LLT) and
afterwards coated by the Institute for Surface Enginee-
5 ring (IOT). The moulding process was carried out by the
Institute of Plastics Processing (IKV).

Figure 9: Moulded microstructure

46
 Integrated Technologies 3

Figure 10: PVD coating of structured tools or laser structured PVD coatings

2
Technical Challenges

The currently used design routines for hybrid micro-opti-

cal systems need to be replaced by a less computationally 3
intensive approach that also allows the incorporation of
restrictions and deviations arising from production.

For the transfer of simulated structures into a moulding

tool, new ablation processes that utilise USP laser radia- 4
tion need to be established and validated, especially for
the structuring of free-form surfaces. Depending on the
size of the functional features, the structuring is applied
to the surface of the moulding tool or to the coating (Fig.
10). 5
Challenges in the coating process consist in the deve- Figure 11: Structured tool for extrusion moulding
lopment of coating with low adhesion and high wear
resistance, while also reaching a high uniformity of
the coating layer on complex surface geometries. The
variothermal moulding process is ideal for cost efficient
reproduction of optical surfaces. Here, controlling the va-
riothermal moulding process through a choice of suitable
process parameters is a major challenge.

To this end, detailed knowledge of the various factors

that influence the process is essential, while the effects
of any changes to parameters should be considered in
advance. In principle, variations in process control can
be utilised to create a multitude of different moulded
designs from a single negative (Fig. 7).

47
 3 Integrated Technologies

2
Technical Data Project Team

Minimal structure size: 5 m Prof. Kirsten Bobzin

3 Thickness of the PVD-layer: 2 - 6 m Surface Engineering Institute (IOT), RWTH Aachen
University
Prof. Christian Hopmann
Technical Equipment Institute of Plastics Processing (IKV) at RWTH Aachen
University
4 Multi-axis working stations with USP laser Prof. Peter Loosen
Atomic force microscope (Bruker Rados N8) Chair for Technology of Optical Systems (TOS), RWTH
Scanning electron microscope (Zeiss Leo) Aachen University
UV-NIR-spectrometer (Perkin Elmer Lambda 1050) Prof. Reinhart Poprawe
Mono-extrusion line for plastic optics Chair for Laser Technology (LLT), RWTH Aachen
5 Fully hydraulic injection moulding machine with embos- University
sing option for processing transparent plastics Michael Berens
Diode laser and induction system for variothermal Chair for Technology of Optical Systems (TOS), RWTH
temperature control of injection moulding tools and Aachen University
embossing rollers Stephan Eilbracht
Industrial PVD system CC800/9 HPPMS Institute of Plastics Processing (IKV) at RWTH Aachen
Industrial PVD system CC800/9 Sinox University
Digital laser scanning confocal microscope Keyence Tobias Mnstermann
VK-X200 Surface Engineering Institute (IOT), RWTH Aachen
Workstation (Intel Xeon X5675, 96GB RAM, NVidia University
Tesla C2070 Mehmet te
Surface Engineering Institute (IOT), RWTH Aachen
University
Maximilian Schngart
Institute of Plastics Processing (IKV) at RWTH Aachen
University
Michael Steger
Chair for Laser Technology (LLT), RWTH Aachen
University
Mona Naderi
Surface Engineering Institute (IOT), RWTH Aachen
University

48

Structural Parts
Introduction
Lightweight construction is a crucial challenge for todays
industrial production. It is considered to be one of the key
technologies of the 21st century to counteract the advan-
cing climate change and preserve natural resources.
The concept of lightweight construction does not only in-
volve the use of lightweight materials, it also includes the
purposeful combination of different materials. The aim of
these composites is to combine the specific advantages
of different materials in order to achieve a functional
and also lightweight design. Considering these aspects,
the used joining technologies have a massive impact on
target-oriented lightweight constructions.
Two possible approaches for a purposeful lightweight-de-
sign are the integration between metal alloys and techno-
Figure 2: Machine for multi-component high pressure die-casting
Integrated Technologies 3
1
polymers, and the combination of steel- and aluminium-
alloys. The aim of using a material mix is to realise certain
functions within a part or component with the most
suitable material. In case of the metal-polymer-materials,
the metal could be used to withstand mechanical stress, 2
while the polymer-components, which distinguish oneself
through their excellent formability and their high freedom
of design, could take over the part of functional integra-
tion. In the case of steel-aluminium-hybrids, steel parts
could be applied to the highly stressed areas of a product, 3
while lesser strained areas would be manufactured from
the much lighter aluminium. Both options offer new ways
of target-orientated lightweight construction.
4
Figure 3: Plastic-metal joint in one step
49
 3 Integrated Technologies

Figure 4: Shortening of process chains via thermal joining

1
Practical Issues

Polymer-metal-composite-materials have been used on an

2 industrial basis for years. The manufacturing techniques
used for those material combinations are usually divided
into In-Mould and Post-Mould-Assembly techniques. The
In-Mould-technique joining operation and the polymer-
moulding process are combined. This is done by placing
3 the metallic component in the injection mould, where it
is joined with the injected polymer. The main challenges
with these processes are the adhesion between metal
and polymer as well as the limited freedom of design. The
Post-Mould-Assembly techniques are applied either if the
4 desired component geometry cannot be produced by the
In-Mould technique or its usage is unfeasible due to eco- Figure 5: Shear test steel with PA66
nomic reasons. The most common Post-Mould-Assembly-
techniques are rivet- and gluing-methods.

5 Even though the difference in the physical properties of

steel and aluminium forbid a combined fusion-welding
of these two metals, this material combination opens up
new opportunities for lightweight-construction. In this
context, joining-technologies are a key factor for success.

Another challenge is the ongoing demand for shorter

development-times. In most cases, the only way to com-
ply with this demand is the use of numerical simulation-
processes. The issue is that there is still a lack of suitable
methods to calculate compound-strengths and material-
transitions, especially in the multi-material-domain.

Figure 6: Failure of steel and TP-FRP

50
 Integrated Technologies 3

Figure 7: Process variants of thermal joining of plastic-metal joints

1
Approach

One approach to expand the In-Mould-Assemblys range

of application is the combination of metal die-casting and 2
plastics injection moulding (Fig. 2 and 3). During the first
step, the metallic component is injected into the die. The
second step consists of the polymer being injected into
a cavity which has been closed with a slide in the run-up
before. The result is the formation of a composite inside 3
one tool.

Thermal joining is a promising approach for the sub-

sequent joining of polymers and metals. It utilises the
adhesion of a polymer-melt with solid metal to create a 4
high-strength connection (Fig. 5 and 6). Therefore, the Figure 8: Potential of thermal joining
metallic component is first heated above the polymers
melting temperature. After that the materials are joined
under a continuous pressure. The advantages are a short
cycle time (Fig. 4) and a high initial-strength without any 5
additional weight caused by mechanical elements (Fig. 8).
The heat required can be provided by a heating plunger
or conventional resistance-welding equipment, for exam-
ple (Fig. 7).

Friction stir welding is a common procedure to join alumi-

nium. Its advantages lie in the fact, that the materials are
joined without reaching their melting-temperatures, so
that the metals are stirred in a malleable condition. This
principle can also be transferred to a steel-aluminium-
compound (Fig. 10 and 11). Figure 9: Surface pre-treatment

51
 3 Integrated Technologies
Figure 10: Procedural principles and parameters for friction stir welding of aluminium and steel
1 Technical Challenges
Due to the technical complexity of the dies used for the
combined pressure-injection-casting-process, they need
to be very precisely machined. To maintain a good bondi-
2 ng performance between the metal and the polymer, an
exact control of the die temperature during the process is
necessary.
An accurate thermal processing is the key to a good
3 metal-polymer-adhesion quality. A homogenous tempe-
rature distribution should be achieved before starting the
joining process. Unfortunately, a homogenous heating of
complex work pieces cannot always be ensured, thus the
process control needs constant adaption. Another chal-
4 lenge is the correct design of the joints. If the mounting
areas are too small or not well placed by constructive me-
Figure 11: Multi-material friction stir welding joint
52
ans, an additional surface-treatment needs to be applied
to increase the composite-strength. Examples for such a
treatment are CMT-Pin-welding, Electron-Beam-Structu-
ring or Laser-structuring, used to generate undercuts
(Fig. 9 fltr).
The joining-mechanism during friction stir welding is a
result of temperature input and material-deformation.
The task is to create a composite which can still be
formed into a complex structure after being joined (Fig.
12). Furthermore, friction stir welding tools need to be
wear-resistant.
Figure 12: Manufacturing of light weight structures
 Integrated Technologies 3

2
Technical Data Project Team

Aluminium-steel joints with 1 mm sheet thickness Prof. Andreas Bhrig-Polaczek

Joints of thermoplastics and thin-walled metal sheets Foundry Institute (GI), RWTH Aachen University 3
Erlanger-bearer: weight ca. 77 g plastic (PA6 GF30), ca. Prof. Edmund Haberstroh
200 g metal (AlMg5Si2) Chair of Rubber Technology (LEK), RWTH Aachen
University
Prof. Gerhard Hirt
Technical Equipment Institute of Metal Forming (IBF), RWTH Aachen 4
University
High pressure die-casting machine: Bhler H 630 SC Prof. Christian Hopmann
with 700 t locking force Institute of Plastics Processing (IKV) at RWTH Aachen
Injection unit: MaicoPresse, 45 mm screw diameter, University
1600 bar max. injection pressure Prof. Uwe Reisgen 5
In house constructed heat contact joining machine Welding and Joining Institute, RWTH Aachen University
IR camera Flir SC 305 Jan Klein
Power source: medium frequency continuous current Institute of Plastics Processing (IKV) at RWTH Aachen
machine with a control system by company Harms und University
Wende Chris Mertin
Friction stir welding machine: Institute of Metal Forming (IBF), RWTH Aachen
3-axis system University
Max. spindle performance 7.5 kW Andreas Naumov
Welding speed up to 3600 mm/min Welding and Joining Institute, RWTH Aachen University
Gabriel Schenke
Foundry Institute (GI), RWTH Aachen University
Johannes Schnberger
Welding and Joining Institute, RWTH Aachen University
Michael Sickert
Chair of Rubber Technology (LEK), RWTH Aachen
University

53
 3 Integrated Technologies

Integrating Electrical Functionalities

in Plastics Parts

1
Introduction

The combination of plastics and metal offers a wide
range of properties, especially concerning the possibility
2 of weight reduction compared to an all-metal solution.
Case components, for example, will oftentimes be made
of metal in order to increase the electromagnetic tole-
rance of enclosed electronics. Frequently however a thin
metal layer on the case suffices to ensure electromagnetic
3 tolerance. The combined utilisation of plastics and metal
allows for an application-specific use of the materials.
The newly developed In-Mould-Metal-Spraying (IMMS) is
supposed to enable the integration of metallic areas on
plastics parts. Therefore it incorporates the processes of
4 thermal spraying of metal and plastic injection moul-
ding in one advanced moulding tool. In the first step of
the IMMS a metallic layer is applied to specific areas of
the moulds surface using wire arc spraying or cold gas
spraying, which is immediately followed by the plastic in-
5 jection. During this second step the metallic layer is trans-
ferred onto the plastics part similarly to the In-Mould-
Labelling process. The plastics part with the transferred
partial metallic surface is then removed from the mould.
This new production technology for metal-plastics hybrids
benefits from a shorter process chain in comparison to
current methods. A compact manufacturing cell consi-
sting of an injection moulding machine, a robotic thermal
spraying unit and an appropriate mould is to render
possible the highly integrated process and to enable the
production-oriented fabrication of metal-plastics hybrids
in shortest possible cycle times. A strong, permanent
connection between the plastics and the metallic layer
and an exact outline replication of the metallic layer in
the pre-assigned areas are to be realised. With the help
of different test parts the electromagnetic tolerance, the
electric functionality and the accuracy to size are evalua-
ted. The feasibility of the innovative approach has already
been proven.

Figure 2: Zinc-coated plastics part Figure 3: In-Mould-Metal-Spraying (IMMS)

54
 Integrated Technologies 3

Figure 4: IMMS process steps

1
Practical Problems

In order to achieve an optimal quality of the metallic coa-
ting, the layer has to stick to the moulds surface during
the plastic injection moulding and may not be damaged
by the melt flow. During demoulding of the part on the
other hand, it has to peel off the mould easily and to
adhere to the plastics. Additionally there are no previous
experiences concerning the ability to reproduce more
complex geometries of relevance for the application, e.g.
thickness changes, cut-outs or ribs. Especially the minimal
flanging radius that has to be provided at such construc- 2
tion elements must be experimentally determined to
ensure an error-free transfer of the metallic layer from the
mould to the plastics part.
3

Figure 5: Schematic drawing of wire arc spraying process

Figure 6: Schematic drawing of cold gas spraying process

55
 3 Integrated Technologies

Figure 7: Variation of part surface via sole change of variothermal temperature control

Approach

To solve the conflict of objectives between adherence
and peeling off, it is investigated which surface qualities
the mould has to possess to enable the peeling off of
1 an applied metallic layer. The research parameters are
the kind of tool steel, the surface roughness, the metal
coatings and varying geometrical elements. Since the pro-
duction parameters of the injection moulding influence
greatly the quality and the transferability of the metallic
2 layer, the injection moulding process has to be adapted
to the IMMS process. An adapted mould using changing
inserts is developed to be able to examine the mentioned
parameters. Additional evaluations relating to component
design and process management are to be made with
3 the help of a second test part which possesses a struc-
tured surface. The structure is repeated along the radius.
Along the circumference the structure differs in shape,
depth and flanging radius. The achieved adherence
between tool steel and metallic layer can be determined
4 via tensile tests before the injection moulding process.
In order to represent the whole product development
chain simulative tools will be developed for the product
and the process. On the product side the morphological
simulation of the sprayed metallic layers is to predict the
theoretical electrical conductivity of final metal coating.
On the process side, based on the simulation of the
injection moulding, the probability of a successful transfer
of the metallic layer is to be calculated depending on
the components geometry and the process parameter.
These simulative tools are to enable short development
cycles and to reduce the number of necessary iterations in
mould manufacturing.

Figure 8: Wire arc spraying of zinc Figure 9: 2nd demonstrator

56

Figure 10: Micrograph for validation
Technical Challenges
The main question is which combination of tool steel,
sprayed metallic layer and spraying procedure provide for
a reliable transfer of the metallic layer to the plastics part
while also ensuring a long lifetime of the mould. For exa-
mple one challenge is the reproducible adhesion of cold
gas sprayed conductive tracks. The high kinetic energy of
the particles during cold gas spraying causes a roughe-
ning of the mould surface. Therefore the conditions chan-
Figure 11: Comparison of two-color presentation and reconstructed
3D-representation
Integrated Technologies 3
1
ge with every subsequent spraying process. Furthermore
the thin metallic layer can be washed away or displaced
by the melt flow during the injection phase of the injec-
tion moulding process. To prevent this, the roughness of 2
the mould surface is increased in order to provide anchor
points for the metallic layer. Additionally depressions can
be placed in the mould surface to enhance this effect.
3
Figure 12: Simulated shear stress distribution
57
 3 Integrated Technologies

2
Technical Data Project Team

Layer thickness: 100-500 m Prof. Dr.-Ing.Kirsten Bobzin

3 Spraying output: 8-10 kg/h Surface Engineering Institute (IOT), RWTH Aachen
University
Prof. Dr.-Ing. Christian Hopmann
Technical Equipment Institute of Plastics Processing (IKV) at RWTH Aachen
University
4 Thermal spraying units: Xifang Liao
Wire arc spraying OSU LD/U2 Surface Engineering Institute (IOT), RWTH Aachen
Cold gas spraying, Oerlikon Metco, Kinetic 3000 University
Plasma spraying, Oerlikon Metco, F4MB-XL Mehmet te
Analytics Surface Engineering Institute (IOT), RWTH Aachen
5 Confocal Laser Microscope University
Optical Microscope Christiane Schulz
Tensile test machine Surface Engineering Institute (IOT), RWTH Aachen
Manufacturing cell University
Injection moulding machine Johannes Wunderle
Kuka six-axis robot Institute of Plastics Processing (IKV) at RWTH Aachen
University
Philipp Ochotta
Institute of Plastics Processing (IKV) at RWTH Aachen
University

58

Self-optimising Production Systems
The research area of Self-optimising Production Systems
deals with the development of technical and socio-tech-
nical production systems that achieve better performance
than initially planned and forecasted by self-optimisation.
Productivity can be increased on the one side by impro-
ving local decision capability of employees on the shop-
floor and in supporting areas and on the other side of
cognitive and adaptive systems and networks.
The object of research covers all levels of a production
system from supply-chain-level through order processing
to process control and machine control level. This inclu-
des data mining and information retrieval by adequate
sensors, preparation and disposition of information, and
simulation of alternative solutions and their visualisation
to enable a faster decision-making and adaption on-site.
On the basis of the Viable System Model information
is managed across all levels of production systems and
cybernetic control methods are established.
Self-optimising Production Systems 4
At the supply-chain and production control level decision
support for employees is in the focus. That requires the
design of suitable interfaces between the human and the 3
machine and/or the system.
For automated but flexible and robust manufacturing
and assembly it is important to integrate intelligence and
cognition capability into the control of the system. Thus, 4
not only controlled variables but also control structures
can be adopted during the process without the interven-
tion of employees. Essential is on the one hand sensors
that perceive and detect environment conditions and
pre-process information. On the other hand mechanisms 5
that enable self-regulation and -learning need to be
established.
59
 4 Self-optimising Production Systems

Image: ante4C GmbH

Corporate Material and

Information Flow

1 Introduction

A lack of connectivity between different solutions often
prevents an efficient added value in production and
logistic networks in as much as the necessary informa-
2 tion is not available on time in an adequate quality. The
optimisation of logistic performance of manufacturing
companies through the application of innovative concepts
and solutions in logistics and IT is researched by the FIR
at the RWTH Aachen. As part of the Cluster of Excellence
3 at RWTH Aachen University, the FIR works with manufac-
turing and service companies, associations and software
companies to develop new solutions to further improve
the logistics performance.
asingly complex logistical coherences and concepts more
tangible and explorable in the real production and IT envi-
ronments. The Campus-Cluster Logistics offers a develop-
ment and testing environment in order to decipher the
complex relationships in logistics and to finally illustrate
them in a real production and IT environment.
The centre of the RWTH Aachen Campus-Cluster Logistics
has three Innovation Labs: the ERP Innovation Lab, the
Smart Objects lnnovation Lab and Service Science lnnova-
tion Lab.

4 For this reason university and industry combined in the

Campus-Cluster Logistics to work together to develop
new approaches in production, logistics and services. The
aim of the Campus-Cluster Logistics is to make the incre-

Figure 2: Campus-Cluster Logistics (Image: rha reicher haase associierte GmbH)

60

Figure 3: The logistics understanding in the Cluster
Practical Issues
Manufacturing companies need to design processes of
production planning and order processing in an efficient
way but they are simultaneously facing challenges such as
lack of data availability, low information transparency and
inadequate integration of information systems or their
absence. Therefore, actors in the supply chain can only
work with their local information, so not seeing the big
figure which can lead to economically more favourable
decisions. Hence, the consideration of the horizontal and
vertical integration is of enormous importance.
The vertical integration (within the company) is defined
as the harmonisation and integration of the companys
internal information technology from the automated
acquisition of high-resolution movement data (eg. RFID)
till the use of this data in higher-mounted planning and
control systems (eg. Manufacturing Execution System,
MES for short). A cross-company exchange of relevant
information along the supply chain is defined as horizon-
tal integration. It enables companies to share electronic
messages such as orders, order confirmations, delivery
notes etc. paperless and without large time delays.
Figure 4: Barcodescan at the stock receipt
Self-optimising Production Systems 4
1
At the same time the horizontal integration enables the
efficient usage of logistical cooperation concepts such as
Collaborative Planning, Forecasting and Replenishment
(CPFR) and hereby minimises the risk of building up 2
stocks and order quantities (the so-called bullwhip effect).
Especially in the inter-company context the benefits of
the Electronic Product Code Information Service (EPCIS)
are noticeable. In particular through the interaction of
EDI and EPCIS optimisation potentials can be realised in 3
company production and supplier networks.
The interaction of vertical and horizontal integration pro-
vides a stable base for flexible supply chains. The ability
to react opens up the possibility of targeted cost savings 4
through transparent process design. Cost-intensive
sectors, such as stocks, can be identified and sustainab-
ly improved. As a requirement towards a reactive and
sustainable supply chain, the real-time capability of the
systems and the communication between the systems 5
need to be established. Real-time capability refers to the
ability of a system to provide needed information and to
be able to process the information within a stipulated
time frame.
Figure 5: Processes in the logistics demonstrator
61
 4 Self-optimising Production Systems
Figure 6: Definition of terms
1 Approach
The logistics demonstrator which was developed in col-
laboration with partners of the RWTH Aachen Campus-
Cluster Logistics and the FIR illustrates possibilities and
2 potentials of horizontal and vertical integration. The aim
of the demonstrator is to highlight complex relationships
in logistics and make the potential of vertical and hori-
zontal integration understandable. The logistics demons-
trator displays the essential steps in order processing. Due
3 to the given transparency various scenarios can be tested
and all effects of any change can be disclosed. Through
this, the members of the supply chain can see the impacts
of their decisions on the entire supply chain and the ways
to optimise it.
4 serialised item number that uniquely identifies the pro-
Despite the different ERP systems along the supply chain,
order processing can be carried out smoothly via the
myOpenFactory . Working with real software and
hardware makes the logistics demonstrator unique com-
5 pared to other demonstrators that usually work only with
virtual simulations. Traceability of material flows through
Figure 7: Control center of the logistics demonstrator
62
the use of RFID tags at the item level is possible any time.
For that purpose the logistics demonstrator uses GS1
standards. Hereby the real-world challenge of worldwide
zero overlapping identifiability is solved. The USB stick
blanks shipped by the supplier are distinguished with a
standardised shipping label and a Serial Shipping Contai-
ner Code (SSCC) on the package. After an automated re-
ceipt from the producer and a comparison with the prior
to shipping electronically distributed dispatch notification,
the blanks are engraved with laser lettering and refined
with customised documents on the stick.
In this specific example the products are provided with a
duct through the following process flow. This serialised
item number stored on both the RFID tag and in a GS1
Data Matrix allows for a clear assignment and traceability
of products and shipment units. In this way the serialised
item number can be used to match orders with deliveries
and with the aid of pick-by-voice solutions to support
stock piling and removal.
Figure 8: Horizontal and vertical integration
 Self-optimising Production Systems 4

Figure 9: Overview logistics demonstrator

1
The demonstrator shows how processes can be made
more efficient. This enables companies to quickly access
information and to respond to problems such as the
threat of delivery delay. The impacts of a 20%-increase in 2
orders from a merchant can become directly visible to all
participants. The combination of horizontal and vertical
integration makes it possible for all the players to adapt
themselves to changes in the ordering behaviour of the
customer or other changes. This case of an USB stick 3
production can also be applied to other industries.

Figure 10: Manufacturing cell Figure 11: USB-Stick laser engraving

63
 4 Self-optimising Production Systems
3 Technical Data
Server consisting of 3 Blades Type Dell Power Edge
M620.
Each equipped with: 2 x CPU Xeon E5-2660
4 With 8x 2,2 GHZ, 128 GB Ram cores each
5 Niklas Hering
64
Project Team
Prof. Gnther Schuh
FIR e.V. at the RWTH Aachen University
Prof. Volker Stich
FIR e.V. at the RWTH Aachen University
Matthias Blum
FIR e.V. at the RWTH Aachen University
Marcel Groten
FIR e.V. at the RWTH Aachen University
FIR e.V. at the RWTH Aachen University
Simone Korall
FIR e.V. at the RWTH Aachen University

Cognitive Assembly Cell
Introduction
The degree and the scope of automation of the technolo-
gy used increased with the growing number of quantities.
One reason was the limited flexibility of the automated
systems. The effort to design, program and operate such
a system safely were high in the case of frequent changes
in the production process. Todays automation systems
are characterised by rigid rules and have no intelligence
nor the skills to develop intelligence. An automation
technology, that limits the variety of products and the
flexibility of production processes, also hinders the indivi-
Figure 2: Planning cycle of the cognitive architecture
Self-optimising Production Systems 4
3
dualisation of production. The ability to adapt to chan-
ging demands and environmental conditions is therefore
a central requirement for future automation systems. 4
However, the automatic decision-making requires a much
broader knowledge within the automation systems. This
expansion of knowledge demands an increase of the pos-
sible transparency of internal processes of the automation
technology, so that in case of a problem, the operator 5
can, for example, also be involved in the solution process.
Figure 3: Individual assembly of a switch cabinet
65
 4 Self-optimising Production Systems
Figure 4: Design and processes of the assembly cell
Practical Issues
The increasing demands on the automation of processes
within production lead to a disproportionate increase
1 in planning and preparation activities. An increasing
automation is only useful if the required adaptability can
be compensated by a reduction of the necessary effort.
These costs are in particular reasoned in the field of non-
value added planning, design, programming, commissio-
2 ning and testing. Even the close integration and networ-
king of real automation technology with the virtually
generated models opens up potential. Finally, the degree
of automation in carrying and supporting of planning
tasks can be further increased.
3 robots and a transfer system consisting of six indepen-
Approach
The approach aims at shifting the planning process into
4 the execution level. This will require an intelligent
control platform capable of planning and executing the
process autonomously with regard to the target and the
current situation and to interact with the human expert.
Thereby, the (given) generic structure is always distingu-
5 ished from the (acquirable) problem specific content. The
sum of both is finally observable as intelligent behaviour.
The structure is also named cognitive architecture and
has deep interdependencies with the representation of
the contents. In the following application, a methodical
approach for the integration of a cognitive control archi-
Figure 6: Cooperating robots
66
Figure 5: Virtual image of the assembly cell
tecture into production systems is demonstrated using a
robotic assembly cell: the operator of the assembly cell
specifies an assembly task by composing simple basic
components to a complex assembly group by means of a
graphical interface. In order to realise high transparency,
variability and scalability and to keep the complexity of
the actual assembly process as low as possible, Lego-
Duplo components have been used. The task of the co-
gnitive control unit is to find a possible assembly sequen-
ce of the components at runtime, to plan the assembly
autonomously with respect to the available resources and
to control the process. The assembly cell comprises two
dently controllable conveyor belts with gates and light
barriers. The assembly robot uses a camera for identifying
the components and a flexible gripper for handling the
components. Beside these active components, also exists
different areas which can be used for the assembly and
buffering. The assembly groups that are finished are then
checked with regard to correctness of the geometry and
colour by means of 3D image processing by the second
robot.
Technical Challenges
The most relevant technical challenges are the deve-
lopment of the necessary algorithms for planning and
Figure 7: Cognitive control architecture
 Self-optimising Production Systems 4

Figure 8: Assembly sequences Figure 9: CAD model of the assembly group

controlling the assembly process. These are planning Furthermore, the variety of technical components and
algorithms such as for generating the possible assembly their interfaces are additional technical challenges. The
sequences as well as technical algorithms for instance for decisions made in the cognitive control unit must be for
2D and 3D image processing. instance, transformed into control commands for the
robot and programmable controllers. 1
The possible assembly sequences are determined by
means of graph-based algorithms. Therefore, first the
possible assembly steps are determined using a so-called
Assembly-by-Disassembly strategy. The assembly group
is disassembled stepwise until the final component. The 2
intermediate states describe in the reverse direction all
the possibilities to perform the assembly. And afterwards,
the resulting valid states are transformed into an assem-
bly graph by combining equal states, and weighted for
the further planning process. 3
During the process of grasping an object 2D image pro-
cessing enables the identification of moving components.
Due to the connection to the control unit of an industrial
robot, their velocity can additionally be determined, so 4
that the movement of the robot be synchronised. By cal-
culating the relative pose, the component can be grasped
without any interception of the movement.

In order to verify the correct build-up, the colour and 5

depth data are recorded from four sides. By combining
these four images, the 3D representation of the assembly
group is calculated. This representation is then compared
with the predefined assembly group and deviations are
communicated to the operator this communication also
has to be transparent. Figure 11: Geometric grip planning

Figure 10: Gripping process Figure 12: 3D assembly group analysis

67
 4 Self-optimising Production Systems

Technical Data Project Team

2
Cognitive Architecture with extensions regarding the Prof. Christian Brecher
utilisation of algorithms from the field of Artificial Laboratory for Machine Tools and Production
Intelligence Engineering (WZL), RWTH Aachen University
Image recognition and processing based on Halcon Prof. Christopher M. Schlick
3 Gripping processes utilising Matlab Institute of Industrial Engineering and Ergonomics
(IAW), RWTH Aachen University
Jennifer Btzler
Technical Equipment Institute of Industrial Engineering and Ergonomics
(IAW), RWTH Aachen University
4 Industrial robot Kuka KR 30 Jet Marco Faber
Translational axis Institute of Industrial Engineering and Ergonomics
5 rotatorical axes (IAW), RWTH Aachen University
Kuka KRC 2 Sinem Kuz
Industrial robot Kuka KR 16 Institute of Industrial Engineering and Ergonomics
5 6 rotatorical axes (IAW), RWTH Aachen University
Kuka KRC 2 Wolfram Lohse
Tactile 3-Finger-Gripper Schunk SDH-2 3 Laboratory for Machine Tools and Production
6 tactile sensor pads Engineering (WZL), RWTH Aachen University
7 degrees of freedom Simon Mller
Siemens Simotion D to control a conveyor Laboratory for Machine Tools and Production
Festo handling system Engineering (WZL), RWTH Aachen University
Microsoft Kinect Markus Obdenbusch
Computer for image processing: Laboratory for Machine Tools and Production
Quad-Core Intel i7 Engineering (WZL), RWTH Aachen University
8GB RAM
Computer for control unit:
Intel Core 2 Duo
4GB RAM

68

Self-optimising Production Planning
and Control
Introduction
Nowadays, production processes are characterised by
growing customer orientation, greater product variety,
and more complex material flow. Production control
needs to be able to ensure robust production processes
in order to cope with the increasing dynamic within
production. Many software solutions have been deve-
loped in the past to live up to this challenge including
advanced planning and scheduling systems that calculate
an optimal processing order for assignments using specifi-
cally designed algorithms. Currently existing IT-systems
are often criticised for missing responsiveness and poor
prognosis quality. These shortcomings mostly occur due
to bad data quality, and wrong system configurations.
In order to counteract these deficits, a surveillance and
Figure 2: Interactive visualisation
Self-optimising Production Systems 4
2
optimisation system will be developed for an advanced 3
planning and scheduling system in a surrounding of the
demonstrator of the self-optimising production planning
and control (PPC). The surveillance system therefore
reveals discrepancies between the IT-system and reality
and shows optimisation suggestions. A successful start of 4
the project ProSense which is sponsored by the Federal
Ministry of Education and Research (BMBF) was enabled
because of preparatory work of the excellence cluster
concerning this subject.
5
Figure 3: User-interface on a tablet
69
 4 Self-optimising Production Systems

Figure 4: Configuring simulation by specifying individual sequencing

rules for machines

Challenges

Since the beginning of the 1990s, challenges in produc-
tion control have been characterised by short delivery
time and products, which are developed according to
individual customer demands. Consequently, produc-
1 tion life cycles are shortened constantly to live up to the
market demands and to stand up to the international
competition. Concurrently, decreasing lot sizes of single
product variants go hand in hand with an increasing
product variety. Additionally, volatile lot sizes increase the
2 challenges for PPC in such a dynamic environment.
The goal of the PPC is to cope with these dynamics and
to master these complex processes. Typically, a major
issue is the strong deviation between planning results and
3 the actual outcome of the production system. This leads
to a general dissatisfaction in afflicted enterprises and is
accompanied by poor achievements of logistical target.
Therefore, it is paramount to include the complexity of
the order structures and of the material flow in the PPC
to allow real-time adjustments. The production processes
and their planning have to be improved continuously to
react to changed circumstances in a dynamic environ-
ment.
Nowadays, software does not permit this flexible ad-
aption, is configured in erratic intervals and thus often
presents worse control results than with a simple First-In-
First-Out (FIFO)-prioritisation.

Figure 5: Example throughput time distribution Figure 6: Data basis for simulation can be selected freely

70

Figure 7: Extension of the IT structure with cybernetic control loops
Approach
A surveillance system is needed in order to improve the
prognosis results and quality of the control systems. This
surveillance system compares the planning results of the
IT-system with the actual material flow on the shop floor
and identifies potential divergences. These variances can
for instance occur because of an order prioritisation made
by the employees on the shop-floor differing from the
prioritisations made by the IT system. If a divergence is
being identified two reactions are possible: the employee
gets informed and he can change the parameters in the
IT-system or he can adapt the parameters given by the IT-
system and inform fellow employees to follow the given
order.
Figure 8: Potential given in Production Control
Self-optimising Production Systems 4
In addition to the comparison of IT-results and actual ma- 2
terial flow the surveillance system can also reveal produc-
tion control potentials by simulating different scenarios.
Potential measures are visualised for the employee and
an assessment of the improvement in logistical targets is
given. This way, the employees stay in the position of the 3
decision maker while being supported via an user-friendly
and interactive visualisation.
4
Figure 9: PPC-Benchmarking (Throughput Time example)
71
 4 Self-optimising Production Systems

Figure 10: Key indicators of typical production data sets

Technical Challenges

Feedback data from the shop floor is often defective

because of manual feedback, aggregation of resources,
or IT-errors. Thus one of the technical challenges is the
treatment of bad input data quality. Data errors can be
1 corrected by applying algorithms which can be borrowed
from different fields such as market research.

The treatment of mass data is another technical challen-

ge. A midsize enterprise for example can easily produce
2 100.000 data sets per year which can additionally contain
different kinds of information. Thus the calculation and
simulation model is to be developed in order to handle
large amounts of data with low costs, high speed and
robustness. The user interface is implemented in a web
3 surface to guarantee an universal applicability. Feedback
data can be uploaded to the user interface, simulation
can be configured, and results can be reviewed. In all
activities, data privacy is treated with paramount im-
portance.
4

Figure 11: Application by production controller Figure 12: Modelling and analysis of production reality

72
 Self-optimising Production Systems 4

Technical Data Tablets:

3
Up to 500.000 data sets from production Java
Simulation times from 0.1 to 5 min Javascript
HTML
Ajax
Technical Equipment Simtalk 4
Webserver:
Project Team
Hardware:
Intel Core-i7 920 Prof. Gnther Schuh 5
4 processors, 2.67 GHz cycle Laboratory for Machine Tools and Production
24 GB Ram Engineering (WZL), RWTH Aachen University
1 GBit network connection, computation centre Felix Brambring
featuring 440 GBit total width Laboratory for Machine Tools and Production
Software: Engineering (WZL), RWTH Aachen University
Jo Java-Servlets & Java Server Pages under Melanie Luckert
Tomcat 7 Laboratory for Machine Tools and Production
User interface uses JavaScript, Ajax, jQuery, D3, Engineering (WZL), RWTH Aachen University
HTML5 Till Potente
Laboratory for Machine Tools and Production
Simulation computer: Engineering (WZL), RWTH Aachen University
Dr. Christina Reuter
Hardware: Laboratory for Machine Tools and Production
2 x Intel Xeon E5-2687 Engineering (WZL), RWTH Aachen University
16 processors, 3.1 GHz cycle
32 GB Ram
Software:
Siemens Plant Simulation 10.1
Java-program for simulation control

73
 4 Self-optimising Production Systems

Self-optimising Injection Moulding

3
Introduction

Production of injection moulded parts requires robust and
4 consistent operating processes to satisfy high quality stan-
dards. The goal of such processing is to reproduce each
cycle identically and consistently within several process
variables.
5 In particular, process variables such as the melt pressure
and temperatures in the cavity. Conventional control-
ling of injection moulding machine controls is based on
recognised variables, which influence the reproducibility
of moulding processes by additional process disturbances.
During the packing phase for example, temperature or
viscosity fluctuations may affect the reproducibility of the
process. The concept of self-optimising injection moul-
ding therefore, is to compensate for reoccurring process
variations and increase reproducibility.

Figure 2: Manual reaction on process disturbances Figure 3: Variables at injection moulding

74

Figure 4: Concept of self-optimising injection moulding
Practical Issues
Machine variables can be referred to as values which
describe the state and action of an injection moulding
machine. Just such a parameter is the hydraulic pressure
of the injection cylinder. Term process variables summari-
se the variables, allowing for a clearer description of the
process at the actual location where the moulding takes
place. The cavity pressure is another example of such
a process variable. Properties of manufactured moul-
ded parts which result from injection moulding can be
described as quality variables, and an example of quality
variables are the part weight and geometry accuracy.
Due to disturbances acting upon the injection moul-
ding process, exclusive control of machine variables are
insufficient to reproduce the process, but cannot ensure
consistant moulding quality. Therefore, the monitoring or
direct control of process variables is desired. Systematic
disturbances such as a fluctuating ambient temperature,
or varying material properties can seriously effect product
quality. This includes the changes in the heat balance of
the injection mould which can occur for example, with a
non-identical repetitive process such as, after changing
machine parameters. In contrast to the machine variables
(such as the hydraulic pressure), process variables (such
Figure 5: pvT-optimised process
Self-optimising Production Systems 4
1
as the cavity pressure) provide detailed information about
processes, during the injection and holding pressure pha-
se. The cavity pressure path correlates with various quality
variables such as the part weight, or moulding accuracy, 2
warpage or shrinkage, as well as morphology and sink
marks.
Approach 3
The concept of a self-optimising system is divided into
model based optimisation (MO-System) and various,
information processing sensor actuator systems (ISA-Sys-
tems). Temperature sensors, like IR-Sensors or thermo- 4
elements are the thermal boundary conditions which
can be measured, and a working point determined by
the ISA-System. A working point is used to calculate the
optimal processing path of the cavity pressure based on
the quality model. This optimal cavity pressure path will 5
describe the internal objective for another ISA System,
which then realises its internal objectives, such as rea-
ching the cavity pressure path independently.
Figure 6: Measuring quality values
75
 4 Self-optimising Production Systems

2 Figure 7: Compensation of thermal fluctuations by self-optimising in injection moulding

During the first phase of the Cluster of Excellence Technical Challenges

research, the concept of a self-optimising injection
3 moulding process was to compensate for temperature
fluctuations and ensure consistent quality was applied,
even with changing boundary conditions or disturbances.
As an example of changing boundary conditions, the fai-
4 lure of two heating zones of a plasticising unit was simu-
lated, by switching off the heating zones and observing
the effect on the part weight. The effect was observed
using a conventional machine control with a fixed holding
pressure path, and a comparison of the parts produced,
5 using the concept of self-optimising injection moulding.
Figure 7 highlights the results of the experiment where a
plate made of polypropylene was used. The effect of the
weight change on the basis of the changing moulding
temperature was significantly lower, when using the self-
optimising concept.
Several initial challenges will be made possible as result of
technological, electronic advances in injection moulding
machines. For one, the robustness and self-adaptation
ability of the cavity pressure control will be tested and
validated. The self-adaptation of the cavity pressure con-
trol will prove essential in later practical feasibility tests
of the self-optimising injection moulding process, and
this will include an autonomous parameterisation of the
controller. Furthermore, the controller architecture will be
selected autonomously from different controller architec-
tures differing in their complexity.

Figure 8: Reduction of process fluctuations Figure 9: Optimising the process model

76
 Self-optimising Production Systems 4

Figure 10: Combined process optimisation

2

As well, the current concept of the self-optimisation 3

injection moulding will be extended by cross-cycle optimi-
sations. The combination of online control and cross-cycle
optimisation will be necessary to compensate for the heat
balance fluctuations, after changing machine parameters.
The compensation of the thermal fluctuations will be ac- 4
complished with the application of the previous concept
(approach) of self-optimising injection moulding machine.
Meanwhile, cross-cycle optimisations will put a special
focus on compensation of material variations, such as the
changing viscosity properties of processed plastic. 5

Figure 11: Additional mould geometries Figure 12: Highly accurate reproduction

77
 4 Self-optimising Production Systems

Technical Data Project Team

3
Physically motivated and self-learning modelling Prof. Christian Hopmann
Algorithms for the implementation of control strategies Institute of Plastics Processing (IKV) at RWTH Aachen
Control cycle of 4 ms University
Prof. Dirk Abel
4 Institute for Automatic Control (IRT), RWTH Aachen
Technical Equipment University
Matthias Reiter
Injection moulding machine: Institute for Automatic Control (IRT), RWTH Aachen
Allrounder 520 A, Arburg GmbH + Co. KG, Loburg University
5 Moulds to produce plane and box shaped moulded Axel Remann
parts Institute of Plastics Processing (IKV) at RWTH Aachen
Pressure and temperature sensors University
Digital and analogue data acquisition
Real-time capable multi-processor system for
implementing control algorithms

78
 Self-optimising Production Systems 4

Image: Trumpf
2

Self-optimising Manufacturing Systems

for Laser Cutting 3

Introduction

Manufacturing in high wage countries has a growing
influence on the local economy. Innovative manufactu-
ring systems are the base of an efficient production of
goods which are expected to fulfil increasing customer
expectations with respect to quality and delivery time.
The market share of laser based manufacturing systems is
facilitated by a continuously increasing level of technolo-
gy. New drives and faster control systems together with 4
enhanced laser sources lead to an overall increase in per-
formance of the systems. Robust manufacturing systems
require this advance to go in line with the stability of the
process itself, considering all the variations in boundary
conditions. 5

Image: Trumpf

Figure 2: Cut head of TruLaser 5030 Figure 3: High quality cut in stainless steel

79
 4 Self-optimising Production Systems

2
Figure 4: Stainless Steel after cutting with CO2 laser at different focus positions

Practical Issues

3 Quality of laser cut products is defined by the clients ap-
plication. If the defined dimensional precision is reached,
properties of the cut face come into scope. One prerequi-
site for a high quality cut is a good command of the laser
beams focus position. In CO2 cutting, this focus position
4 can be determined by mounting a diagnostic measure-
ment device into the machine and measuring the energy
distribution. In the practical application, most industrial
cutting machines provide a calibration routine that allows
the operator to determine the focus position, and transfer
5 its value to the machine. As the focus position is subject
to change over time, the machine is only calibrated when
the calibration routine has just been executed. Once the
focus position is known, other parameters move to the
foreground. The major influences on the process are
associated to technical boundary conditions and material
compositions which altogether determine the absorption
of the laser energy and result in the establishment of
the cut front. Laser cutting is an industrially established
manufacturing process which is widely adopted with an
increasing amount of automation. Future potentials in
technological advance of this manufacturing technology
are seen in innovative, automated assistant systems based
on self-optimising capabilities.

Figure 5: Sample of the focus set up routine Figure 6: Model of the kerf width

80

Figure 7: Simulation of the cut front in laser cutting
Approach
In a multitude of cases, process problems cannot be
solved by simply adding sensors. The focus position for
example cannot be measured directly during cutting as
the relevant part of the kerf is sited in the kerf itself. It is
the width of the kerf that can be measured during pro-
cessing and fortunately, it can well be related to the focus
position. This relationship is embedded into a model
which contains expert knowledge. As such, the measure-
ment value of a surrogate criterion like the kerf width can
be used to determine the focus position. If this approach
is applied to other inaccessible process variables, then
information for the determination of the entire manuf-
acturing systems operating point can be gained. With
this knowledge, the system can be optimised on product
quality.
Contributions to the solution:
Development of numerically evaluable surrogate mo-
dels that describe the laser cutting process
Implementation of decision units based on embedded
process knowledge
Figure 8: Image from the process observation
Self-optimising Production Systems 4
Realisation of an information exchange with product
planning systems on enterprise level 1
Design of interfaces for the interaction between hu-
mans and the self-optimising manufacturing system
Technical Challenges 2
The technological challenge in this task includes the
generation of reliable signals that qualify the state of
the process. The specifics about this challenge lie in the
location of the interaction zone as the quotient of sheet 3
thickness to kerf width is large. Typical values are kerf
widths of 400 m in multi millimetre materials. The task
is to use process signals which describe the process re-
garding the underlying question with sufficient precision.
And as a surrogate criterion, they equally have to enable 4
an inference towards the setting parameters of the
technical system. The required relation is established by a
meta-model which is derived from experiments and physi-
cal descriptions of the process. This meta-model provides
the operating point, which can then be compared to 5
the predictive solution that was computed by the model
based simulation. Decision units can use this to adjust the
Figure 9: System Sketch of the process observation set up
81
 4 Self-optimising Production Systems

2
Figure 10: Concept of Model-Based Self-Optimisation (MBSO)

related setting parameters and guide the system into an and the machine has to provide solutions in cases where
3 optimised operating point. these solutions cannot be fetched from embedded
process knowledge, nor from decision units. Any of these
Achieving this target requires an adaptation of sensors steps adds to an increase in productivity and robustness
and actuators to the high dynamics of the melt based of future manufacturing systems, and aid in utilising
cutting process. The concept of Model-Based Self- unexploited potentials.
4 Optimisation (MBSO) creates the required separation of
time scales with a fast control cycle and an asynchronous
optimisation. One big challenge lies in the implementati-
on of the interaction with the operator. When the manuf-
acturing system is capable of determining its operating
5 point autonomously, communication between humans

Figure 11: Interaction between operator and machine Figure 12: Commercial figures of laser technology

82
 Self-optimising Production Systems 4

Technical Data Project Team

3
TRUMPF TruLaser 5030 Prof. Wolfgang Schulz
6 kW CO2-Laser Fraunhofer Institute for Laser Technology (ILT)
Stainless Steel up to 25 mm thick Chair of Non-linear Dynamics of Laser Processing
Maximum gantry speed 300 m/min (NLD), RWTH Aachen University
Torsten Hermanns 4
Chair of Non-linear Dynamics of Laser Processing
Technical Equipment and Tools (NLD), RWTH Aachen University
Thomas Molitor
Coaxial Process Control (CPC) Fraunhofer Institute for Laser Technology (ILT)
Highspeed CMOS Camera with Cameralink Ulrich Thombansen 5
Interface for in process signal acquisition Chair for Laser Technology (LLT),
High power beam coupler RWTH Aachen University
Fiber coupled illumination source
Magnifying optical system for process observation
FPGA based image acquisition and processing
system for synchronous data transfer
Ultrahighspeed CMOS Camera for process
diagnostics
Software
LabView Realtime control system for interfacing
with TruLaser 5030
Cpc recording software for in process recording
of highspeed camera
Cpc analysis and control software, feature
extraction and control signal
MATLAB / Simulink for signal analysis and
processing
Mathematica for process modeling
CPU/GPU computing cluster for process simulation

83
 4 Self-optimising Production Systems

Image: Thilo Vogel

1
Model-based Modules for the
Self-optimisation of Gas Metal Arc
Welding Processes
2
Introduction

In the field of automatic gas metal arc welding processes,
3 clamping devices nowadays are often used to prevent
weld errors by realising an exact positioning of the plates.
Due to a higher individualisation of products, time and
cost related intensive changes of the set-up exist when
a new product is produced. Nevertheless, disturbances
4 to the process can occur leading to process conditions
resulting in an aberrated welding result. To react flexibly
to such altering conditions, the welding process was
enabled to adapt without external intervention by app-
lying intelligent sensor systems combined with process
5 models. For this purpose, process states and their influ-
ence on the welding result are recorded and mathemati-
cally modelled. By utilising these models, new parameter
combinations can be determined if disturbances occur,
so that the varying conditions can be compensated and a
steady welding result can be guaranteed.
Practical Issues
The research activities of the Welding and Joining Insti-
tute (ISF) were to focus on a solution for the above menti-
oned challenge, when using typical applications with thin
sheets with hereby commonly used weld preparations.
As process type, a pulsed gas metal arc welding process
was used which was more applicable to thin sheets, due
to its great reach of efficiency. Furthermore, this process

Figure 2: Gas metal arc welding process Figure 3: Model for the root height

84

Figure 4: Influence of the gap variation on the weld geometry (overlap joint)
type was sufficiently flexible to compensate for changing
boundary conditions.
Approach
In this section of the Cluster of Excellence Integrative
Production Technology for High-Wage Countries, a
cross-process concept for a model-based self-optimisation
of production processes has been developed. The transfer
of this concept to the gas metal arc welding process
includes an automated optimisation to guarantee the
required weld quality on two levels. On the micro level
this is achieved via monitoring and optimisation of the
Figure 5: Optimisation of the weld quality
Self-optimising Production Systems 4
1
process stability with the objective to influence the speci-
fic process parameter of the welding power source. While 3
on the macro level the actual welding result is monitored
and setting parameters are adapted for potential distur-
bances, like variations in the gap width or in the contact
tube distance.
4
Technical Challenges
As well as the development of a global process model
which maps the correlation between setting parameters 5
and welding results adequately, the challenge lies in
providing a suitable sensor system to monitor the process.
Figure 6: Optimisation of the process stability
85
 4 Self-optimising Production Systems
1
Figure 7: Correlation of high speed images and voltage data
2 In the case of process stability on the lower, micro level a
real-time capable sensor module has been realised, which
detects the mass transfer and energy input of pulsed gas
metal arc welding processes. The module uses transient
current and voltage data, and detects therein the point
3 of time of the droplet detachment, because this event
combined with actual electrical capacity has a significant
influence on the homogeneity and the quality of mass
transport. The point in time of the droplet detachment
is indicated within the voltage data by a short increase
4 of the voltage. An algorithm was therefore written using
a special filter method. At a pulse time of about 10 ms,
the evaluation results were available approximately 3
ms before the next pulse started. This short evaluation
time enabled the start of a control procedure in the
5 subsequent pulse, if the droplet had not detached. For
the optimisation on the higher, macro level an optical
Figure 8: User interface of the sensor module
86
sensor system was developed, which offered the possi-
bility of monitoring the formation of the weld pool. For
this, a CMOS-camera had been set up at the welding
torch, which recorded up to 500 frames per second
in full screen mode with a resolution of 1280 x 1024
pixels. The optics had been built according to the special
requirements of gas metal arc welding processes, which
meant that an optical filter was equipped that limited the
spectral range to near infra-red. Furthermore, a remo-
vable protection glass was included to shield the sensor
from welding spatter. And to reduce radiation intensity, a
customised pinhole aperture was added that shielded the
sensor from welding spatter because of its small diame-
ter.
With the help of the newly provided information about
the weld pool in combination with observations from the
Figure 9: Detection of the droplet detachment
 Self-optimising Production Systems 4

1
Figure 10: Correlation of the image analysis with the welding results (macro sections)

material transfer and the energy input, the monitoring 2

of the welding process was guaranteed, which promptly
recognised the influences of disturbances on the welding
result, and indicated a parameter adaption if necessary.
The parameter adaption took place with a model-based
optimisation, targeting an autonomous choice of parame- 3
ters to react to disturbances or changing objectives. A
solution concept for this task will therefore be the usage
of surrogate models for an iterative optimisation. And
for this optimisation, different methods were used which
range from classical gradient based methods to meta- 4
heuristic search techniques. Taking into consideration the
specific requirements for the automated welding produc-
tion, one of the future tasks will be to qualify methods
which are usable for an offline or the time-critical online
application. 5

Figure 11: Application of the camera during welding Figure 12: Arrangement of the camera at the welding torch

87
 4 Self-optimising Production Systems

Technical Data Project Team

3
Calculation speed of the weld pool image analysis: Prof. Uwe Reisgen
20-50 images per second Welding and Joining Institute (ISF), RWTH Aachen
Detection of the droplet detachment within 4 ms University
Guido Buchholz
4 Welding and Joining Institute (ISF), RWTH Aachen
Technical Equipment University
Marion Purrio
Hardware Welding and Joining Institute (ISF), RWTH Aachen
High speed camera system University
5 (100000 frames per second) Konrad Willms
LabVIEW real time system Welding and Joining Institute (ISF), RWTH Aachen
Measurement equipment to acquire transient data University
(2 MHz measuring frequency per channel)

Software
MATLAB
LabVIEW
DIAdem
Modelling with R
Welding process simulation with SimWeld

88

Self-optimisation of the
Weaving Process
Introduction
Weaving is one of the oldest processes for the fabrication
of textiles. The process is characterised by orthogonal
interlacing of warp- and weft-threads. To enable the
interlacing, the warp-threads have to produce the so-
called shed, which is created by raising or lowering of the
weft-threads. After the shed-formation, the weft yarn is
inserted into the open shed. The first looms are dated at
approximately 3000 BC. To produce the required tension
for weaving, stones were used as weights on the warp-
Figure 2: Woven airbag
Self-optimising Production Systems 4
Image: Picanol
3
4
threads of the early looms. In the course of industriali-
sation however, the first mechanically powered looms
have been created. In the 20th century the development
of looms was characterised by an increasing automation
of the weaving process sequence, with an aim of higher 5
efficiency. Due to the cumulative computerisation of the
weaving process, weaving became the most productive
process for producing textile fabrics. On modern looms,
different materials can be processed and a wide variety
Figure 3: Woven 3D-carbon profile
89
 4 Self-optimising Production Systems

Figure 4: Self-optimisation via tablet

2

of woven goods can be produced such as, clothing fabric

like denim, or suit cloth made of cotton, or cashmere-
3 wool to technical goods like parachutes made of polya-
mides or lightweight constructions made of carbon fibres.

Practical Issues
4
Figure 5: Warp tension sensor The modification of the weaving-loom to apply various
materials requires complex adjustments be made to the
machine. For example, yarn guiding elements, weft-inser-
ting-systems or process parameters must all be aligned to
5 the particular material.

Another important characteristic for the judgement of the

weaving process is warp tension (WT). If the warp tension
is too high, the warp threads will break and the weaving
process will stop. If the warp tension is too low, the warp
threads will obstruct the weft insertion and the fabric
cannot be produced. The adjustment of the warp tension
is based on the experience of the weaver. Furthermore,
Figure 6: Grammage sensor the quality surveillance of the fabric does not take place
while the fabric is on the loom.

The quality surveillance of a fabric takes place in the so-

called fabric inspection by trained employees, at special
inspection tables after the fabric is removed from the
loom. Therefore, it is not possible to respond to failures,
which occur during the weaving process.

Figure 7: RFID-transponder to store forward process data

90

Figure 8: Concept for the self-optimising weaving loom
Approach
The aim of this project is to design the setup-process of
the loom to such an extent that for a desired product
quality the loom itself will define the necessary process
parameters. For this purpose a meta-model will be gene-
rated and integrated into the loom achieving a premium
worsted fabric with a shortened setup time (30 min) at a
cost reduction of 61.10 to 60.50 per running meter.
Furthermore, there was a reduction in waste material
associated with the setup resulting in the saving of the
expensive materials.
At first, the process has to be modelled with regard to
the warp tension. In addition, a self-optimising-routine
which optimises the weaving parameters of the loom will
have to be realised. An other aspect will be the identifica-
tion of suitable sensors for monitoring the product-quality
using the ITA 9-Step-Tool method and their integration
in the weaving process.
Figure 10: System for self-optimisation
Self-optimising Production Systems 4
Figure 9: Camera system for monitoring the fabric-quality
Technical Challenges
With help of the 9-Step-Tool an X-ray sensor of the 1
BST ProControl GmbH, Freudenberg has been chosen to
monitor the fabric grammage. Following the integration
of the sensor within the weaving process, a control loop
monitoring the fabric grammage was realised with help
of a smith-predictor. For additional system monitoring 2
of the fabric-quality, an optical camera system has been
developed and integrated into the loom.
With the use of a thread sensor which has been integra-
ted into the loom, the warp tension could be measured 3
online. The obtained data has been used for model-
building by quadratic regression. By the dint of the
Gauss-Markov theorem, and additional quality criteria, an
optimised set of parameters could be calculated. In this
manner a self-optimising-routine was generated, which 4
was implemented with the help of products (Soft SPS) by
the company iba AG, Frth into the weaving loom.
Trials in the technical centre of ITA show that using this
routine can reduce the warp thread tension up to 13 %. 5
The optimising-routine has been field tested at the wea-
ving plant Weyermann Technical Textiles GmbH & Co.
KG, Wegberg. An air jet loom of the company Picanol
91
 4 Self-optimising Production Systems

1
Figure 11: Measured warp thread tension with or without self-optimising routine

2 nv, Ieper, Belgium was connected to the routine. Due to impact on the warp thread tension. There is still however,
the new settings after completing the optimising-routine, the question whether a regression model can provide the
the loom runs 100 rpm faster. Furthermore, the loom optimum results for the described additional input- and
produced two shifts without process-related stops and output-parameters. Hence alternative models have to be
the fabric quality fulfilled the requirements. analysed. Likewise the control of the self-optimisation by
3 so-called Smart Devices like Tablets or Mobiles will be of
On-going work will aim to embed further parameters interest as well as the storage and forwarding of process
as e.g. the energy consumption, or the position of the data using RFID technology.
looms components into the optimisation-routine. And for
that purpose, both appropriate sensors and additional ac-
4 tuators were needed to be integrated into the loom. One
solution to monitor the air consumption was the installa-
tion of a flow sensor between the compressed air supply,
and the loom. Present-day measuring techniques enable
the wireless transfer of measured data at sufficient speed.
5 By integration of stepper motors e.g. the position of the
warp stop motion can also be adjusted automatically. The
position of the warp stop motion therefore has a strong Figure 12: Project partners

92
 Self-optimising Production Systems 4

Technical Equipment Technical Data

3
Weaving looms: OmniPlus 800 air-jet loom by Picanol nv, Ieper,
Picanol nv, Ieper, Belgium Belgium
OmniPlus 800 air-jet loom Working width 1.90 m
Gamma rapier loom Quadruple weft insertion
Lindauer DORNIER GmbH, Lindau Max. speed: 1400 weft insertions per minute 4
Air-jet loom Actively powered back-rest bar
Rapier loom Electric warp let-off and fabric take off
Multiaxial loom Reluctance motor as main drive
MAGEBA Textilmaschinen GmbH & Co. KG,
Bernkastel-Kues Grammage sensor by BST ProControl GmbH, 5
SL shuttle loom Wenden
SAMT multi-needle-loom Acceleration voltage below 5 kV
Jakob Mller AG, Frick, Schweiz Can be operated in Germany without permission
Carbon-needle-loom Measuring range from 50 to 1000 g/m2
Needle-loom Solution of 0,1 g/m2
Accuracy of 0,3 g/m2
Sensors for:
Movement and acceleration of looms components
Thread-load in warp and weft Project Team
Online Inspection System for fabric appeareance
Grammage Prof. Thomas Gries
Institut fr Textiltechnik (ITA), RWTH Aachen University
Measurement data acquisition and processing / Adjunct. Prof. Yves-Simon Gloy
Soft SPS der Firma iba AG, Frth Institut fr Textiltechnik (ITA), RWTH Aachen University
ibaPDA-V6 Maximilian Kemper
ibaPADU-S-IT-16 Institut fr Textiltechnik (ITA), RWTH Aachen University
ibaLOGIC-V4 Marco Saggiomo
Institut fr Textiltechnik (ITA), RWTH Aachen University

93
 4 Self-optimising Production Systems

1
Self-optimised Metal Cutting Processes

2 Introduction

Cutting processes are one of the key technologies used in
many industries to manufacture products for the global
market. The processes include, inter alia, five axis milling
3 and gun drilling operations which are used in industries
with automotive, aeronautical or medical technology. For
production in high-wage earning countries like Germany,
it is of key importance that these cutting processes be
controllable. This becomes increasingly difficult as increa-
4 sing demand for quality, higher degrees of individualisa-
tion, and shorter product life cycles require more flexible
cutting processes which can be operated safely at their
technological limit. Furthermore, the economic and eco-
logical aspects will be more and more in focus. Reasons
5
for this include the increasingly competitive pressure from
manufacturers in low-wage earning countries, the global
shortage of resources, or legal requirements. The result is
in an urgent need to continue the progress of controlla-
bility in cutting processes while taking the global scope of
action into account.
Practical Issues
Due to the manufacturing task, the cutting process has to
meet different demands, which include not only pro-
duct specific quality aspects, but also economic targets.

Figure 2: 5-axis milling Figure 3: Gun drilling

94

Figure 4: Overall procedure
In order to adjust the cutting process to these respec-
tive requirements, the process is nowadays designed
in advance. For this purpose, extensive resources and
technological knowledge are required which depend
upon the complexity of the manufacturing task. Further-
more, cutting processes have to deal with a multitude of
variations and disturbances, which can only be partially
considered during the design phase. So that, during
production the demanded process stability is not always
maintained resulting in a quality below pre-defined to-
lerances. This may lead to scrap or re-machining opera-
tions which creates immense additional costs and has a
negative, long-term effect on the manufacturer. Nowa-
days, feedback from process disturbances which affords
Figure 5: BLISK (Blade Integrated Disk)
Self-optimising Production Systems 4
1
2
appropriate countermeasures, is mainly dependent upon
the experience of the operator. The ability to respond by
the manufacturing company is often limited.
3
Approach
To increase controllability, the idea of self-optimisation is
therefore transferred to metal cutting processes. For de- 4
monstration purposes, monitoring and control strategies
are developed for the 5-axis milling process and the gun
drilling process. Process stability is ensured by strate-
gies for detecting disturbances, and their subsequent
compensation by closed control loops. Furthermore, 5
Figure 6: Injection body (Image: Bosch)
95
 4 Self-optimising Production Systems
1
Figure 7: Connectivity of virtual and real world for information exchange
2
planning aspects are integrated into the control loops to
achieve an optimum concerning external target values of
3 the production plant. An important requirement is the
development of new or adaption of existing technological
models that are used as a basis for decision-making in
self-optimisation. For the implementation, sensor systems
are developed, empirical and simulative investigations
4 performed, process information extracted and transferred
to knowledge in machine-readable models. Further tasks
are the cross-linking of single components of the overall
system to ensure a consistent data exchange. Concerning
this latter, a special task is the linkage between CAM and
5 physical process monitoring that is established by a positi-
on related monitoring and simulation strategy.
Figure 8: Communication structure
96
Technical Challenges
The challenge of introducing self-optimisation to metal
cutting processes is caused by the complexity of the
processes. The abstraction of process coherences and the
reproduction in technical models, are today probably the
most challenging tasks. For modelling the 5-axis milling
process and the gun drilling process, transfer functions
are used and implemented to black box models. The ex-
tension by physical descriptions of phenomena and cau-
se-effect relationships transfers the models to the class of
grey box models. For the development of suitable models,
the interconnection of virtual and real world is necessa-
ry. This includes the joining of experimental results and
simulations. For example, the position-correlated forces
of the 5-axis milling process are measured and connected
Figure 9: Simulation of engagement conditions

Figure 10: Control loop fort the gun drilling
to engagement conditions of the simulation. By this, first
optimisations can be done in the CAM system. On the
other hand the information about engagement condi-
tions can be used for the continuous identification of
important parameters for the control loop and the more
detailed processing of monitoring data.
Another task is the development of robust process
control strategies. Process control is used to compensate
disturbances to ensure a safer process. To achieve this,
control strategies have to be implemented that adapt
Figure 11: Process without axial force control loop
Self-optimising Production Systems 4
their control behaviour depending upon the actual pro-
cess state. Furthermore, the developed process models
must be integrated into the control concept. Therefore, 3
interfaces have to be established that enable a smooth
communication within the control system.
4
Figure 12: Process with axial force control loop
97
 4 Self-optimising Production Systems

Technical Data Project Team

3
Data acquisition with up to 2.5 MHz Prof. Fritz Klocke
Computer-based modelling of 5-axis milling Laboratory for Machine Tools and Production
Simulation of engagement conditions for any kind of Engineering (WZL), RWTH Aachen University
milling tool Fraunhofer Institute for Production Technology (IPT)
4 Realtime calculations of process values by efficient Oliver Adams
simulation models Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Thomas Auerbach
Technical Equipment Laboratory for Machine Tools and Production
5 Engineering (WZL), RWTH Aachen University
Hardware: Lothar Glasmacher
Machine tools: Fraunhofer Institute for Production Technology (IPT)
MAZAK Variaxis 630 5X II T Gunnar Keitzel
TBT ML 200 Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Sensors: Meysam Minoufekr
Kistler Dynamometer 9257B Fraunhofer Institute for Production Technology (IPT)
Microphone Dr. Draen Veselovac
Accelerometer Laboratory for Machine Tools and Production
Eddy current sensors Engineering (WZL), RWTH Aachen University
Temperature sensor
Chip detector (self-development)

Data acquisition and processing

FPGA
NI CRio 9004
Labview 2011

98

Self-optimisation of the Radial Braiding
Process
Introduction
Thomas Walford, a Briton from Manchester, claimed
the first patent for a braiding machine in 1748. On this
machine, as well as on the subsequent machines, it was
possible to manufacture two-dimensional structures, such
as laces and straps.
According to DIN 60000 braids are described as fabrics
with closed selvedges. Braids consist of at least two
thread systems that intersect alternately and have a de-
fined angle at the selvedges, the so-called braiding angle.
Figure 2: Principle of Braiding
Self-optimising Production Systems 4
3
4
At least two thread systems are needed to manufacture a
braided structure.
With braids it is possible to produce 2-dimensional as well
as 3-dimensional structures. The most commonly known 5
methods for braiding are flat braiding, packing braiding,
rotary braiding and overbraiding. A special method of
braiding is the radial braiding. The bobbins face radially
towards the centre of the machine. Threads from the
bobbins pass to the centre of the ring body, where the
Figure 3: Triaxial Braiding Structure
99
 4 Self-optimising Production Systems
1
Figure 4: Sinusoidal Path of the Bobbins
2 Practical Issues
braiding ring collects them. Due to the radial arrange-
ment of the bobbins, thread deflections are minimised.
This ensures lower thread damage during braiding and
3 thus enables braiding of delicate materials. The braided
structure is created in a radial braiding process by in-
tersecting of two thread systems on two paths by moving
in opposite directions (clockwise and anticlockwise).
Consequently, the threads of the one thread system
4 cross the threads of the other one. Braided structures can
be divided into biaxial and triaxial braids. Biaxial braids
consist only of the two thread systems which form the
braid. Triaxial braids are created by adding 0-threads, the
so-called stationary threads, to the biaxial structure.
5 mple only half-fitted with stationary threads, the friction
Figure 5: Braiding Bobbin
100
A consistent quality over the entire overbraided core,
through the length as well as diameter, is desired. For
example the deviation of the braiding angle from the
local target value and the damage of the fibres are
example quality criteria. The braiding angle varies with
the geometry of the braiding core and the take-off speed
of the core moved by a robot through the braiding ring.
The tension of each thread is another factor which affects
the braiding angle. Furthermore, the thread placement is
affected by the friction of the filaments with each other.
The more stationary threads (0-threads) and braiding
threads are used, the greater the friction of the threads
with each other. If the radial braiding machine is for exa-
of each thread changes consequently at an angle of 180
degree. In addition, the combination of materials has an
influence on the friction. The braiding angle is influenced
by varying the speed of the core that is pulled through
the braiding ring. Furthermore, the complexity of the core
affects the braiding angle.
Figure 6: Dependence of Thread Friction

Figure 7: Concept to Regulate the Thread Tension
Currently, the thread tension is controlled mechanically
via a spring. Thereby, the thread passes from the coil
over the guiding plate and then through an eyelet. The
eyelet moves upwards against a spring when the thread
is pulled. When the maximum suspension travel of the
spring is reached, the bobbin releases more material.
Thus, the thread tension is about the tension that results
by the spring during the pull process of the thread from
the bobbin. However, the spring can only compensate
the difference in take-off speed from the bobbin, but not
the fitting of stationary threads and other factors, such as
unequal thread tension. The unequal tension can result in
damages of the thread.
In case of unequal yarn tension is the possibility of the
thread to be damaged by too much tension. In case of
low thread tension, there is a risk that a braid may deflect
by overbraiding of the core.
Figure 8: Principle of Developed System
Self-optimising Production Systems 4
1
2
Approach
The approach here is to monitor and control the braiding
angle and thread tension by the interaction of a sensor 3
system in a bobbin developed at ITA and an Active Carrier
system designed by ITA for regulation of the thread
tension.
The Active Carrier System includes an electric motor, a 4
tension sensor at the pedestal of each of the 144 bobbins
of the braiding machine and a data processing unit. The
tension sensor measures the thread tension, which is
than processed and the electric motor releases the thread
or holds it back. In this way, the thread tension of each 5
bobbin can be individually detected and controlled.
Figure 9: Measuring System
101
 4 Self-optimising Production Systems
Figure 10: Exemplary Illustration of Tension Fluctuations in a Braiding Thread
2
Technical Challenges
3 a fast realisation of a desired component configuration is
A standard for quality evaluation of braids needs to be
developed. This measurement basis is required for bench-
marking of measurements and results.
4 Another challenge is the online process monitoring. A
concept for an active default of bobbin tension is already
available. However, the implementation is taking some
time due to its complexity.
5 Furthermore, the exact interrelations between machine,
material parameters and resulting parameters of other
components are still unknown. ITA has a large amount of
experimental data available. A solution containing the lin-
Figure 11: Simulated Braiding Cell
102
king of all the parameters and thus methodically defining
still missing. Particularly, the automated determination of
a trajectory of the robot, which guides the braiding core
through the braiding ring is relevant. A module based
on a knowledge database and the component geometry
will be adding great value. The module should write and
simulate a program for the robot and dictates necessary
machine configuration. The knowledge database is inten-
ded to learn continuously. Furthermore, there is a lack of
solutions for monitoring the winding process as well as
the integration in the monitoring and braiding process.
Furthermore, a solution for detection and influences to
changes in thread tension between the bobbins and
the braiding ring, which is dependent on the stationary
threads and the material pairing has to be found.
Figure 12: Braided Hub of Motorcycle Wheel Rim
 Self-optimising Production Systems 4

Technical Equipment Project Team

3
Radial Braiding Machine Herzog RF1/144-100 Prof. Thomas Gries
Industrial Robot Kuka KR 150 for the core handling Institut fr Textiltechnik (ITA), RWTH Aachen University
Kuka Sim Pro Viktor Reimer
Thread Tension Sensor (self-development) Institut fr Textiltechnik (ITA), RWTH Aachen University
Bobbin Winding Machine Britta Kuckhoff 4
3D-Rotary Braiding Machine 12x12 of company Herzog Institut fr Textiltechnik (ITA), RWTH Aachen University
3D-Rotary Braiding Machine 5x5 of company Herzog Michael Lengersdorf
3D-Hexagonal Braiding Machine Institut fr Textiltechnik (ITA), RWTH Aachen University
APODIUS Vision System (AVS) for quality management Jens Schfer
Minitab 16 Statistical Software Institut fr Textiltechnik (ITA), RWTH Aachen University 5

103
 4 Self-optimising Production Systems

Self-optimising Assembly of Large

3
Components

Introduction
4
An increasing number of industrial companies are Approaches for self-optimisation are currently being re-
confronted with the challenge of automating complex searched for the application in the assembly of large com-
assembly processes, in a flexible manner, in order to ponents (e.g. planes, wagons, or wind turbines). Large
compete on international markets. With ever increasing components are posing special challenges for production
5 complexity of a process however, the complexity of engineering, as they tend to deviate significantly from a
production equipment and control systems increases as theoretically planned geometry (because of deformation
well. And as not all production states can be planned in due to gravity, manufacturing tolerances, or fluctuations
advance, a reliable control of these processes cannot be in temperature).
guaranteed. A current research approach for mastering
this complexity is the use of self-optimising systems,
which aside from controlling can also fulfill planning
tasks, thus ensuring control over the complex system.

Figure 2: Influences of the component geometry Figure 3: Deformed large component

104

Figure 4: Flexible aircraft production by robot-based assembly systems
Practical Issues
In order to compensate for deviations in assembly
processes, conventionally fixed steel jigs are used,
which exactly map the geometry of the component and
compensate the deformations. Since the precision in the
manufacturing of these jigs directly affects the dimen-
sional accuracy of the components, they have to be
produced within a tenth millimetre tolerance or have to
be measured during commissioning. Not only has the pre-
cision of these tools led to enormous costs, but also their
size. As they are custom designed to fit the geometry of
one component, multiple of the large, expensive jigs have
to be procured to cover a variety of components.
A more flexible technology for the positioning of compo-
nents is the use of robots. As the large components have
to be supported at several grasping points, the produc-
tion of task-specific handling units can be an economic
Figure 5: Metrology assisted assembly
Self-optimising Production Systems 4
2
alternative to industrial robots. By a simple reprogram-
ming of the robot motions, various components can be
moved and positioned within the workspace. The low 3
accuracy of a robotic system, in comparison with a fixed
tool, can be compensated by the integration of exter-
nal measurement systems to measure and re-adjust the
positional deviation of the manipulating unit. The usage
of cooperating handling units combined with metrology, 4
instead of purely fixed jigs, leads to a higher installation
complexity and a larger planning effort for operating this
system.
5
Figure 6: Steps of self-optimising
105
 4 Self-optimising Production Systems
Figure 7: Structural design of a airplane section
1
Approach
For the execution of such complex production processes,
the use of self-optimisation is currently being re-
2 searched. Compared to state of the art control systems,
self-optimising assembly systems get assigned more
degrees of freedom for planning and control. By that, the
system will be able to make decisions autonomously,
change its behaviour flexibly, and reach the assembly
3 goal self-optimising, within predefined limits. There-
fore, the system needs to be able to perceive its environ-
ment with suitable sensors, and possess sufficient degrees
of freedom for adapting the process to the current situa-
tion. In order to fulfil the requirements of the production
4 the desired level of changeability.
Figure 8: Requirements for changeability
106
process self-optimising, a changeable assembly system is
required. This system has to offer at least the mechanical
degrees of freedoms that are required to adept itself to
the task. Additional mechanical degrees of freedoms for
a self-optimising assembly process are achieved by a re-
configurable handling technology. A promising approach
is the manipulation principle PARAGRIP, which integrates
the handling object, in this case an aircraft shell, as a part
in the kinematic chain of the created structure. By an au-
tonomous selection of suitable contact points, the system
can be reconfigured without mechanical modifications.
This creates additional degrees of freedoms to achieve

Figure 9: Vision of Local shaping
Technical Challenges
In order to transfer planning tasks to a machine control,
mathematical and physical models have to be developed
which describe the states and correlations between com-
ponent, process and assembly system. While kinematic
and dynamic models for the description of a handling
system mostly exist, there is a fundamental need for
research on the description of the components behavi-
our under the influence of external forces. The goal is
to create models, based on measurement data and to
derive robot motion based on the individual component,
in order to compensate for its deformation. The new
manipulation technology has to be designed and imple-
mented based on the requirements of a changing process
Figure 10: Component geometry measurement
Self-optimising Production Systems 4
2
3
with focus on flexibility and changeability. Therefore, a
consistently modular assembly system is required, which
can be used to react quickly and simply to altering requi-
rements. In this respect, the selection of the structure and
its dimensions pose a major challenge. Furthermore, the 4
additional degrees of freedoms lead to a much higher
complexity of the manipulation system. For still being able
to find the optimal operating point, the required planning
algorithms for the systems configuration and motion
planning need to be deduced first. Thus, a changeable 5
and efficient operation can be guaranteed.
Figure 11: Universal manipulation system
107
 4 Self-optimising Production Systems

3 Technical Data Project Team

Component dimensions: 1,7 m x 2,1 m Prof. Burkhard Corves

Component weight: 23 kg (material: CFK) Department of Mechanism Theory and Dynamics
Required accuracy of component positioning: < 1 mm of Machines (IGM), RWTH Aachen University
4 Prof. Robert Schmitt
Technical Equipment Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Robot: Felix Bertelsmeier
Kuka-robot (KR16 and KR 60) Laboratory for Machine Tools and Production
5 Handling principle demonstrator Paragrip Engineering (WZL), RWTH Aachen University
Tim Detert
Sensors: Department of Mechanism Theory and Dynamics
Indoor-GPS of Machines (IGM), RWTH Aachen University
Light section laser Dr. Mathias Hsing
Force-torque-sensor Department of Mechanism Theory and Dynamics
of Machines (IGM), RWTH Aachen University
Software: Markus Janen
Labview Laboratory for Machine Tools and Production
Matlab Engineering (WZL), RWTH Aachen University
Easy-Rob Stefan Kurtenbach
Department of Mechanism Theory and Dynamics
of Machines (IGM), RWTH Aachen University
Christoph Storm
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Dr. Walter Kimmelmann
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University

108

Self-optimising Assembly of
Laser Systems
Introduction
Photonic technologies play an important role in industri-
al production with increasing relevance. Short product
life cycles require improved and novel solutions for the
production of laser systems which are capable of meeting
the demands regarding mass customisation.
Practical Issues
Customer-specific requirements and relatively low pro-
duction volumes between one and a thousand units per
year are characteristic of the production of micro-optical
laser systems today. Versatility leads to a high degree of
manual processes for assembly and thus leads to rising
product costs. Therefore manual processes claim up to
80 % of the added value of such a laser. In a few cases
automated solutions have been installed. Rigid auto-
mation concepts can only be economic for high-volume
production. Rigid automation is based on the dominating
Figure 2: Compact micromanipulator
Self-optimising Production Systems 4
1
2
imperative programming paradigm which requires the
manual planning and storing of all decision points that
can possibly occur during assembly and respective pro-
cess steps of handling, alignment and bonding. Product
variance and even scaling the production usually leads 3
to massive re-programming efforts for the stored com-
plex decision points and can also leads to a modification
of the used hardware. The demonstrator for self-optimi-
sing assembly of laser systems validates methods which
intend to increase flexibility of assembly systems and to 4
allow economic production of low-volume production.
5
Figure 3: Marking laser in planar design
109
 4 Self-optimising Production Systems
Figure 4: Process chain of laser assembly
1
Approach
For decreasing assembly costs, manual assembly shall
be replaced by flexibly automated processes. The efforts
2 for planning and commissioning of processes on the
assembly platform need to be reduced. An increase of the
degree of autonomy reduces the need of implementing
decision points in an imperative manner. Self-optimisation
is applied in order to achieve the goals. A self-optimising
3 production system is characterised by the ability of
detecting tolerance chains, finding optimisation patterns
and compensation strategies which allow an automatic
reaction to changes in external influences and adapt own
behaviour to new situations. Adaptation capabilities inclu-
4 de the change of behaviour (structure or logic of internal
state machine). Due to this flexible use of the production
system a reduction of setup times occurring to the chan-
ge from one product to another increase the production
flexibility.
5
Figure 5: Assembly of the MicroSlab laser system
110
Technical Challenges
High flexibility is only possible if production tools provide
standardised interfaces with regard to hardware and
software interfaces, which allow fast re-configuration,
and easy change of the control program. The task is to
provide interfaces which serve as a common standard,
while not limiting future applications, and at the same
time are easy to implement. Flexible production processes
need to be specified in a flexible manner so that optimi-
sation potential can be realised. Especially the demanding
field of laser assembly requires the use of simulation
models during production in order to optimise the optical
function of the system. Simulation-based approaches
need to be capable of compensating for work piece to-
lerances and limited machine accuracy. A core challenge
is the mapping of measured data into simulation models.
Robots with large work space as well as ultra-precise
micromanipulators need to be integrated in the control
Figure 6: Table-top version of the micromanipulator

Figure 7: Schematic illustration of the challenge regarding tolerances
architecture. Calibration plays an important role in order
to link the real world with the world of simulation mo-
dels. Tolerances can be exploited in order to compensate
for abberations by dynamically changing the order of as-
sembly and applying tolerance matching techniques. This
requires the characterisation of individual parts and their
influence on the optical quality. The process of enginee-
ring self-optimising control requires novel methodologies
in order to minimise costs for software development.
Figure 8: Optimised housing of the laser crystal
Self-optimising Production Systems 4
1
Demonstration Process
The demonstrator will realise industrial assembly pro-
cesses by applying new approaches. One focus is the
assembly of collimation optics for diode laser systems, 2
which are used to parallelise diverging laser beams, as
they occur in diode laser systems. Such elements require
the highest precision to six degrees of freedom covered
by a micromanipulator developed in the first phase of the
project. 3
The complexity of tasks requires the use of model-based
and self-optimising control. The integration of sensor
technology and intelligent evaluation algorithms are able
to map reality to models during assembly and to compen- 4
sate for possible deviations dynamically. Active alignment
is carried out operating the laser source based on online
sensor data. Novel concepts regarding the analysis of
Figure 9: MicroRod Laser-System
111
 4 Self-optimising Production Systems

2
Figure 10: Assembly cell at the Chair for Technologies of Optical Systems at the Fraunhofer ILT

3 measurement data used for beam shape characterisation

are applied.

The demonstrator consists of an assembly cell at

Fraunhofer IPT, as well as a robot cell for measurement
4 development at the Chair for Technology of Optical Sys-
tems, which are continually enhanced by new concepts.
Through the distributed nature of the demonstrator, the
transferability of software architecture is demonstrated.

Figure 11: Sensor-guided pick-up process

Figure 12: Lasercomponent to adjust the outcoupling mirror

112
 Self-optimising Production Systems 4

Technical Data Project Team 3

Sub-m precision Prof. Christian Brecher
Algorithms for mapping sensor data and simulation Fraunhofer-Institute for Production Technology (IPT)
models Prof. Peter Loosen
Laser marking system with 2 4 W output power Chair for Technology of Optical Systems (TOS), RWTH 4
Aachen University
Martin Holters
Technical Equipment Chair for Technology of Optical Systems (TOS), RWTH
Aachen University
Microassembly station MicRohCell compact Sebastian Haag 5
Micromanipulators Fraunhofer-Institute for Production Technology (IPT)
Laser sources Marcel Prochnau
Camera-based tools for beam analysis Chair for Technology of Optical Systems (TOS), RWTH
Simulation software for optical systems Aachen University

113
114

Collaboration Productivity
The stronger integration of information and communica-
tion technologies into production enables a new type of
productivity growth by collaboration. This so-called colla-
boration productivity in integrative, interdisciplinary teams
is in the focus of the Cross Sectional Processes. Here
three main perspectives of collaboration are considered:
results, employees and structure.
The research results of the interdisciplinary project teams
aim at increasing productivity both in the production and
in the product development process, especially concer-
ning individualised products. The aim is to integrate the
contributions of the different research projects by a holi-
stic approach to a new theory of production. This theory
is supposed to increase the predictability of the behaviour
of socio-technical production systems and therefore sup-
ports the decision-maker.
The basis of collaboration productivity are integrative,
interdisciplinary teams that are working together on
technologies and models of the integrative production
Collaboration Productivity 5
3
technology. Focus of Scientific Cooperation Engineering
is to support and analyse collaboration in and across
the teams. On the one hand innovation and knowledge
management in interdisciplinary research teams are analy- 4
sed, on the other hand the performance and productivity
of those teams are measured.
The Research area Technology Transfer aims to
strengthen the collaboration structures within the Cluster 5
of Excellence by sustainable networking. With the Scien-
tific Cooperation Portal networking structures can be te-
sted internally. In the long-term the aim is to simplify the
search of technologies for external partners. The network
PROTECA (Production Technology Aachen) transfers the
results and competencies into the local industry.
115
 5 Collaboration Productivity
Theory of Production
3
Introduction
The classical, neoclassical, and decision-orientated Theo-
4 ries of Production represent three historical milestones of
hitherto scientific analysis and description of production
systems. They originate from the core of business and
economic research and investigate operational processes
under economical influences. The classical and neoclassi-
5 cal Production Theories set their focus on the description
via formula of input factors of a production process (raw
materials, intermediates) and their output factors (finis-
hed products) in the form of production functions. Cost
functions were used to evaluate the cost effectiveness
of the described transformational processes. Production
resources in the form of material, maintenance resources,
and human workforce were therefore the fundamental
model parameters of Production Theory. The decision-
orientated Theory of Production represents an alternative
approach, whose main focus is not set in the combination
of production factors, but in the way operational relevant
decisions of processes are passed on. The human factor,
which represents the leading instance in a production
system becomes increasingly the centre of attention.
116
Practical Issues
To cope with the claim of a holistic theory, a multiplicity
of engineering disciplines as well as social and socio-
technical point of views necessary for the description and
design of production systems, had to be taken into consi-
deration in the resulting construct. The claim of feasibility
of the theory at the same time requires an adequate
reduction in the covered complexity. The percentage of
knowledge workers respectively, analytical and arranging
functions, in production systems increased continuously
thereby stressing the importance of the human factor.
Approach
To address the above mentioned claims and build up the
hitherto findings of operational Theory of Production,
effort was put into a Theory of Production that integra-
ted deterministic, as well as cybernetic models, into a
superordinate organisation frame. Models and theories
from different research areas of the CoE were linked,
thereby contributing to a technology-orientated Theory of
Production.

Scientific Cooperation Engineering
Introduction
Scientific Cooperation Engineering (SCE) aims at the
interdisciplinary cross-linkage of scientific processes
within the Cluster of Excellence (CoE). Furthermore, the
evaluation of the performance concerning the
organisational development within these research
processes is addressed.
Challenges
Interdisciplinary cooperation takes on greater significance,
especially in science: Central research questions are more
than ever located at or between the boundaries of
different cooperating disciplines. The growing dynamics
and complexity emerging from interdisciplinary
cooperation processes require new approaches to the
cross-linkage of scientists and knowledge repositories.
Focusing on a successful scientific cross-linkage
interdisciplinary cooperation does not only have to be
shaped but also to be transformed into sustainable and
robust structures. Therefore, the central research question
of SCE can be posed as follows:
Which actions are necessary to support the transfer of
highly complex, dynamic and interdisciplinary research
cooperation of the CoE into sustainable and robust
structures?
Collaboration Productivity 5
2
Approach
Since 2006, a consortium consisting of 3
the Institute of Information Management in
Mechanical Engineering & Center for Learning and
Knowledge Management & Associated Institute of
Management Cybernetics of RWTH Aachen University 4
(IMA/ZLW & IfU RWTH Aachen University,
Coordination),
the Human-Computer Interaction Center (HCIC),
the Department of Ferrous Metallurgy (IEHK) and
the Department Nonlinear Dynamics of Laser 5
Processing (NLD), RWTH Aachen University
has addressed the challenge of interdisciplinary coope-
ration within the CoE. On the basis of an actor-based
approach, various means have been developed and
implemented in order to enable a cross-linkage between
the researchers within the CoE. Those means comprise
further trainings and colloquia for the clusters employees
on the one hand. On the other hand strategic workshops
as well as the identification and advancement of success
factors within cooperation of interdisciplinary teams are
conducted. Models and proceedings regarding perfor-
mance measurement as well as the support of result and
knowledge dissemination (e.g. via a virtual approach:
the Scientific Cooperation Portal), are also part of the
SCE means. All of these means can be structured into
four fields of action: Knowledge & Cooperation Enginee-
ring, Interdisciplinary Innovation Management, Diversity
Management and Performance Management. In combi-
nation with a continuous evaluation, a system-oriented
management model for the CoE is formed.
117
 5 Collaboration Productivity
1 Scientific Cooperation Portal
Introduction
2 All main social challenges (energy, mobility, production,
health etc.) need appropriate solution strategies, which
often require an integration of methods and knowledge
from different entities. One important condition of sci-
entific innovation capability is based on the possibility to
3 connect existing knowledge with new knowledge.
A central aspect for this connection is a broad exchange
between great research associations. Complex and highly
specific knowledge as well as a high employee turnover
require a systematic analysis and the storage of know-
4 ledge. The collected knowledge of an organisation exists
in the memory of its employees. The central issue of
cross-divisional processes in knowledge-intensive orga-
nisations is to save and utilise the collected knowledge
for all actors. The research of these organisations highly
5 contributes to manage knowledge strategically and thus
supports the innovation capability.
Figure 2: Technology Profiles
118
Technical Challenges
Interdisciplinary knowledge exchange has to deal with
problems, such as missing criteria, guidelines and
communication strategies, as well as different specific
ratings. Hence respective scientific disciplines do not only
operate with different terminologies, but are also based
on diverse methodological understanding and cogni-
tive models. The resulting differences are often neither
consciously realised nor reflected, and up against this
background can hardly communicable.
Approach
To support the employees of the Excellence Cluster Inte-
grative Production Technology for High-Wage Countries
in their daily work the Scientific Cooperation Portal (SCP)
was initiated. The portal offers access to employee pro-
files, which show the respecting interest of research and
Figure 3: Overview discipline-specific definitions

Figure 4: Homepage of the Scientific Cooperation Portal
competences, as well as contact information. Furthermo-
re, the SCP represents a virtual meeting point to collect
and represent central appointments, documents and
results. Moreover, the SCP also offers a range of appli-
cations to further assist the interdisciplinary cooperation
and to tailor the development of measures to the support
of the research alliance.
One application serves the recording and visualisation of
the terminologies used within the cluster. It should assist
in revealing and overcoming the diversity of the termi-
nologies, and resulting barriers of the discipline-specific
cooperation. A further functionality is represented by the
technology platform, which depicts the key characteri-
stics of cluster-specific technologies, and the responsible
Figure 5: Communication tool FlowChart
Collaboration Productivity 5
2
people in charge. The application FlowChart supports
project management in the development, visualisation
and communication of project schedules. To visualise the 3
structures of the interdisciplinary cooperation SCP also
shows various network relationships, which are for exa-
mple based on publication dates. They can be viewed as
a general overview of the cluster, or from the individual
perspective of one single researcher. 4
5
Figure 6: Analysis of the interdisciplinary cooperation
119
 5 Collaboration Productivity
3 Project Team Scientific Cooperation
Engineering and Portal
Susanne Aghassi
Fraunhofer Institute for Production Technology (IPT
4 Prof. Wolfgang Bleck
Department of Ferrous Metallurgy (IEHK), RWTH
Aachen University
Andr Brkling
Fraunhofer Institute for Production Technology (IPT)
5 Dr. Andr Calero-Valdez
Human Computer Interaction Center (HCIC), RWTH
Aachen University
Dr. Kirsten Dahmen
Department of Ferrous Metallurgy (IEHK), RWTH
Aachen University
Toni Drescher
Fraunhofer Institute for Production Technology (IPT)
Ulrich Jansen
Department Nonlinear Dynamics of Laser Processing
(NLD), RWTH Aachen University
Prof. Sabina Jeschke
Institute of Information Management in Mechanical En-
gineering (IMA) & Center for Learning and Knowledge
Management (ZLW) & Assoc. Institute for Management
Cybernetics e.V. (IFU), RWTH Aachen University
Dr. Claudia Joo
Institute of Information Management in Mechanical En-
gineering (IMA) & Center for Learning and Knowledge
Management (ZLW) & Assoc. Institute for Management
Cybernetics e.V. (IFU), RWTH Aachen University
120
Dr. Ulrich Prahl
Department of Ferrous Metallurgy (IEHK), RWTH Aa-
chen University
Prof. Anja Richert
Institute of Information Management in Mechanical En-
gineering (IMA) & Center for Learning and Knowledge
Management (ZLW) & Assoc. Institute for Management
Cybernetics e.V. (IfU), RWTH Aachen University
Anne Kathrin Schaar
Human Computer Interaction Center (HCIC), RWTH
Aachen University
Thomas Thiele
Institute of Information Management in Mechanical En-
gineering (IMA) & Center for Learning and Knowledge
Management (ZLW) & Assoc. Institute for Management
Cybernetics e.V. (IfU), RWTH Aachen University
Markus Wellensiek
Fraunhofer Institute for Production Technology (IPT
Prof. Martina Ziefle
Human Computer Interaction Center (HCIC), RWTH
Aachen University
Prof. Wolfgang Schulz
Department Nonlinear Dynamics of Laser Processing
(NLD), RWTH Aachen University

Project Team Theory of Production
David Bendig
Innovation and Entrepreneurship (WIN),
RWTH Aachen University
Arne Bohl
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Christian Dlle
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Niklas Friederichsen
Innovation and Entrepreneurship (WIN),
RWTH Aachen University
Annika Hauptvogel
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Collaboration Productivity 5
3
Michael Keller
Innovation and Entrepreneurship (WIN),
RWTH Aachen University
Till Potente 4
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Jan-Philipp Prote
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University 5
Marius Rosenberg
Innovation and Entrepreneurship (WIN),
RWTH Aachen University
Anja R. Weber
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Cathrin Wesch-Potente
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
Bartholomus Wolff
Laboratory for Machine Tools and Production
Engineering (WZL), RWTH Aachen University
121
 5 Collaboration Productivity

1
Technology Transfer

2 Introduction Approach

The increasing complexity of technology and thus, the
related considerable development effort required for the
next generation of technologies, in accordance with shor-
3 ter amortisation periods, will often exceed the resources
of an organisation. Hence companies will be forced into
a development cooperation. The related search for exter-
nal, available technologies and suitable cooperation part-
ners will become a challenge for most companies in the
4 near future. So that, special importance will be attached
to the technology transfer from research into industry.
Various efforts from the government, funding sponsors
and universities has therefore been made to promote and
encourage inter-organisational technology transfer, from
5 science into economy, and the implementation of excel-
lent research results into products for the market.
In conclusion a community-orientated approach should
be adopted, in which a technology transfer portal for
the excellence-cluster based on modern social software
solutions be developed. Initially, development will be
for cluster-internal use only, but the portal should create
transparency around the technologies being developed
within the cluster, as well as the people behind it. As
part of the Scientific Cooperation portal, the portal
should be made accessible to external parties after a
successful test phase within parts of the cluster. This will
then enable and encourage the exchange between indus-
trial organisations and the possible cooperation partners.

Challenges

Conditions for a successful technology transfer initially

will be to find transfer partners. But experts agree that
the main barrier is the information deficit, or information
imbalance between (technology-) demand and supply.
To overcome this problem, it will have to be transparent,
which technologies are available, and who is interested in
them. The efficient access on technology-related informa-
tion and the direct link of the information with the corre-
sponding person of contact can however, counteract an
information deficit and support the technology transfer in
its early stages (e.g. initial phase).

122

PROTECA Integrative Production
Technology Aachen
PROTECA is a network consisting of scientists and
industry representatives and builds a bridge between
research and practice. It transfers applicable solutions into
industry, in order to strengthen the competitiveness of
that producing industry in Europe. The aim of PROTECA is
also to push forward efficient, close cooperation between
science and industry. The focus will not only be on global
players, but also small and medium enterprises.
In total, PROTECA bundles production technology
expertise from more than 25 institutes inside the Cluster
of Excellence, and transfers this expertise to industry.
Together, with Industrie 4.0 which is driven by the
increasing integration of information and communication
technology into the production, new scientific results
from different research areas in Individualised Produc-
tion, Virtual Production Systems, Integrated Technologies
and Self-optimising Production Systems will be directly
transferred into application, with the key focus being,
to leverage the full potential of both production and
engineering.
The exchange between science and local industry will
be supported and actively shaped by cooperation with
regional associations, like the Department of Industrial
Collaboration Productivity 5
3
Development of the City of Aachen, the Stdteregion
Aachen, the Industry-Dialogue of the Region of Aachen 4
and the Chamber of Commerce and Industry Aachen.
Networking events, panel discussions and workshops are
being organised, as well as public events benefiting from
the mutual cooperation and assistance.
5
Driven by the Cluster ProduktionNRW, a joint project
founded by the state of NRW and the Verband Deutscher
Maschinen- und Anlagenbau (VDMA), there will be an
exchange within the state of (NRW) Nordrhein-Westfalia.
Public appearances such as an exhibition at the
Hannover Messe, will strengthen the public presence of
the network, make its expertise highly visible, and help in
acquiring new partners.
A strong collaboration, networking and clustering at all
levels will be the basis of ensuring the level of entrepre-
neurial innovative capacity, prosperity and social stability
for High-Wage Countries over the long term.
123
Cluster of Excellence
Integrative Production Technology
for High-Wage Countries
c/o Werkzeugmaschinenlabor WZL
RWTH Aachen University
Steinbachstrae 19
D-52074 Aachen

Phone: +49 (0) 241 80-25322

Fax: +49 (0) 241 80-22293
Email: xcluster@wzl.rwth-aachen.de
Website: www.production-research.de