IFAC

IFAC Conference
Conference on
on Manufacturing
Manufacturing Modelling,
IFAC Conference
Management and on Manufacturing Modelling,
Control Modelling,
IFAC Conference
Conference
Management
IFAC and on Manufacturing Modelling,
Control
on Manufacturing Modelling,
Management
June 28-30, and Control
2016. Troyes, France Available online at www.sciencedirect.com
Management
June 28-30,
Management and
2016. Control
Troyes,
and Control France
June 28-30, 2016. Troyes, France
June
June 28-30,
28-30, 2016.
2016. Troyes,
Troyes, France
France

ScienceDirect
IFAC-PapersOnLine 49-12 (2016) 047–052
Equipment
Equipment selection and evaluation
Equipment selection
selection and
and evaluation
evaluation
approach
approach for an adaptable assembly line
approach for an adaptable assembly line
for an adaptable assembly line
Angius ∗∗∗ Marcello
Alessio
Alessio
Angius Colledani ∗∗∗ Massimo
Marcello Colledani Manzini ∗∗∗
Massimo Manzini
Alessio
Angius
Alessio
Angius ∗ Marcello Colledani
Marcello ∗
Colledani ∗ Massimo
∗
Massimo Manzini
Manzini ∗
Alessio
Angius Andrea
∗ RattiColledani
Marcello
Andrea Ratti ∗ Marcello Massimo
∗ Marcello
∗ Urgo ∗
Urgo ∗ Manzini ∗
Andrea
Andrea Ratti
Ratti ∗ Marcello Urgo
∗ Marcello Urgo ∗ ∗
Andrea Ratti Marcello Urgo
∗
∗ Department of Mechanical Engineering, Politecnico di Milano, 20156,
∗
Department of Mechanical Engineering, Politecnico di Milano, 20156,
∗ Department
∗
Department of
of
Milan,
Mechanical
Mechanical
Italy,
Engineering,
Engineering,
(E-mails:
Politecnico
Politecnico di
di Milano,
Milano,
alessio.angius@polimi.it,
20156,
20156,
Department of
Milan,
Milan,Mechanical
Italy,
Italy, Engineering,
(E-mails:
(E-mails: Politecnico di Milano,
alessio.angius@polimi.it,
alessio.angius@polimi.it, 20156,
Milan, Italy,
Italy, (E-mails:
(E-mails: alessio.angius@polimi.it,
marcello.colledani@polimi.it,
Milan,
marcello.colledani@polimi.it,alessio.angius@polimi.it,
massimo.manzini@polimi.it,
massimo.manzini@polimi.it,
marcello.colledani@polimi.it,
marcello.colledani@polimi.it, massimo.manzini@polimi.it,
massimo.manzini@polimi.it,
andrea4.ratti@polimi.it, marcello.urgo@polimi.it)
marcello.colledani@polimi.it,
andrea4.ratti@polimi.it, marcello.urgo@polimi.it)
massimo.manzini@polimi.it,
andrea4.ratti@polimi.it,
andrea4.ratti@polimi.it, marcello.urgo@polimi.it)
marcello.urgo@polimi.it)
andrea4.ratti@polimi.it, marcello.urgo@polimi.it)
Abstract: In
Abstract: In anan evolving
evolving manufacturing
manufacturing sector, sector, decisions
decisions about
about production
production systems systems design,
design,
Abstract:
Abstract: In
In an
an evolving
evolving manufacturing
manufacturing sector,
sector, decisions
decisions about
about production
production systems
systems design,
design,
reconfiguration
Abstract:
reconfiguration In anand
and management
evolving
management are
manufacturing
are critical
sector,
critical tasks. In
decisions
tasks. In addition,
about
addition, the
production
the rapid
rapid evolution
systems
evolution and
design,
and
reconfiguration
reconfiguration and
and management
management are
are critical
critical tasks.
tasks. In
In addition,
addition, the
the rapid
rapid evolution
evolution and
and
underlying
reconfiguration uncertainty
underlying uncertainty and
uncertainty of of market
management
of market conditions,
are
market conditions, critical
conditions, make make these
tasks.
make these Indecisions
addition,
these decisions
decisions more morethe important
rapid
more important
important and and
evolution capital
and capital and
capital
underlying
underlying uncertainty of wemarket conditions, makeon these decisions more important and capital
intensive. In
underlying
intensive. Inuncertainty
this paper,
this paper,of we focusconditions,
market
focus our attention
our attentionmake onthese
the decisions
the equipmentmore
equipment selection
important
selection in an
in anand adaptable
capital
adaptable
intensive.
intensive. In
In this
this paper,
paper, we
we focus
focus our
our attention
attention on
on the
the equipment
equipment selection
selection in
in an
an adaptable
adaptable
assembly cell
intensive.
assembly cell
In where
this
where paper,different
we
different easy-changeable
focus our attention
easy-changeable hardware
on
hardwarethe modules can
equipment
modules can be arranged
selection
be arranged
in an around aa
adaptable
around
assembly
assembly cell
cell where different easy-changeable hardware modules can be arranged around aa
skeleton
assembly cell where
skeleton architecture
skeleton where different
architecture
architecture to provideeasy-changeable
provide
different
to provide
to
different
easy-changeableassembly
different assembly
different assembly hardware
hardware modules
technologies
modules
technologies
technologies withcan
with
with can be
be arranged
different
different
different
execution
arranged
execution
execution
around
modes
around
modes
modes a
skeleton
and
skeleton architecture
performances.
architecture
and performances.
performances. In to
this
In to provide
paper
thisprovide different
we present
paper different
we present
present assembly
the
assembly technologies
architectural
technologies
the architectural
architectural with different
framework
with
framework different
of executiondefining
of aaa procedure
procedure
execution
procedure modes
modes
defining
and
and performances. In
In this paper we the framework of defining
the different
and
the
the
different
performances.
different In this
sequencing
sequencing
sequencing this of paper
of
of the we
paper
the
the
tasks
we
tasks
tasks
present
present
in
in
the
the architectural
in relation
relation
relation
to alternative
architectural
to
to
framework
alternative execution
framework
alternative execution
execution
of
of aa procedure
modes, aa defining
procedure
modes,
modes,
possible
aa defining
possible
possible
the
the different
skeleton
different
skeleton sequencing
architecture
sequencing
architecture and
and of
ofa
a the
the tasks
procedure
tasks
procedure in
in relation
that,
relation
that, given
given to
to alternative
the selection
alternative
the selection execution
of the
execution
of the modes,
equipment,
modes,
equipment, a possible
provides
possible
provides
skeleton
skeleton architecture
architecture and
and a procedure that, given the selection of the equipment, provides
the evaluation
skeleton
the evaluation
the evaluation of
architecture the
of the
of and a
the associated a procedure
associated
associated performance
procedure
performance
performance
that,
that, to given
to be
given
to the
the selection
compared
be compared
be selection
compared of
of the
against
against
against theequipment,
the
the
the
capability
equipment, provides
requests,
provides
capability requests,
capability requests,
the
the evaluation
expressed
evaluation
expressed in of the
in terms
terms
of theof associated
ofassociated
volumes and
volumes performance
and mix of
performance
mix of the to be
the
to be compared
products.
compared
products. The
Theagainst the capability
modularity
against the
modularity capability requests,
providedrequests,
provided by the
by the
expressed
expressed in terms of volumes and mix of the products. The modularity provided by the
considered in
expressed
considered in terms
architecture
terms
architecture of
of volumes
is also
volumes
is and
also exploited
exploited
and mix
mix of
through
of
through the
the products.
the possibility
products.
the possibilityThe
The ofmodularity
of a fast
modularity
a setupprovided
fast setup of the
provided
of by
the assembly
assembly
by the
the
considered
considered architecture
architecture is
is also exploited through the possibility of a fast setup of the assembly
line, thus
considered
line, thus
line,
allowing
architecture
thus allowing
allowing the is also
also exploited
rearrangement
the rearrangement
the exploitedand
rearrangement through
and
through
and the
the possibility
substitution
substitution
substitution
of the
possibility
of the
of
of
of aa fast
the different fast setup
different
different hardware
setup
hardware
hardware
of
of the
the assembly
modules
assembly
modules
modules
to
to
to
line,
cope
line,
cope thus
with
thus
with allowing
the
allowing
the the
production
the
production rearrangement
of different
rearrangement
of different and
parts.
and
parts. substitution
Both equipment
substitution
Both equipmentof
of the
the different
cost and
different
cost and hardware
performance
hardware
performance modules
are
modules
are taken
taken to
to
cope
cope with
with the the production
production of
of different
different parts.
parts. Both
Both equipment
equipment cost
cost andand performance
performance are
are taken
taken
into consideration
cope
into consideration
with the to identify
production
to identify
of the most
most
different
the promising
parts. Both
promising configurations.
equipment
configurations.cost A testing
and
A testing
performanceof the
of the areapproach
taken
approach
into
into consideration
consideration to
to identify
identify the
the most
most promising
promising configurations.
configurations. A
A testing
testing of
of the
the approach
approach
through
into
through the application
the application
consideration to to a
identify
to a realistic
realistic
the mostcasepromising
case is also
is also provided.
provided.
configurations. A testing of the approach
through
through the application to a realistic case is also provided.
through the the application
application to to aa realistic
realistic case
case isis also
also provided.
provided.
© 2016, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
Keywords: assembly
Keywords: assembly cell, cell, manufacturing,
manufacturing, design, design, analytical
analytical method,
method, optimization
optimization processprocess
Keywords:
Keywords: assembly
assembly cell, cell, manufacturing,
manufacturing, design, design, analytical
analytical method,
method, optimization
optimization processprocess
Keywords: assembly cell, manufacturing, design, analytical method, optimization process
1.
1. INTRODUCTION
INTRODUCTION (ElMaraghy
(ElMaraghy et
et al.
al. (2013)),
(2013)), require
require the
the assembly
assembly resources
resources
1.
1. INTRODUCTION
INTRODUCTION (ElMaraghy
(ElMaraghy et
et al.
al. (2013)),
(2013)), require
require the
the assembly
assembly resources
resources
1. INTRODUCTION to
to bebe
(ElMaraghyendowed
be endowed et
endowed with with
al. technologies
(2013)),
with technologies require designed
the
technologies designed assembly
designed to to
to be be flexible
beresources
flexible
to
to bebeable
endowed with technologies
technologies designed to be be flexible
flexible
The worldwide
The worldwide manufacturing
manufacturing sector sector has has faced
faced a a rapid
rapid andand to and
and able to operate
endowed
to operate
with on more
on more thanthan one product.
designed
one product.
to These
flexible
These
The worldwide manufacturing sector has faced a rapid and and
and able
able to
to operate
operate on
on more
more than
than one
one product.
product. These
The
The worldwide
consistent
worldwide
consistent change
change manufacturing
in the
the last
manufacturing
in last few
fewsector
years.
sector
years. has
has faced
Most
faced
Most ofaaconsumers
of rapid
consumers
rapid and
and technologies
and able
technologies to can be tailored
operate
can be tailored
on moreat process
at process
than level
one (i.e. by
product.
level (i.e. by These
using
These
using
consistent
consistent change
change in the
in have last
the last
last few years.
few years.
years. Most
Most of consumers
of consumers
consumers technologies
technologies can
can be
be tailored
tailored at
at process
process level
level (i.e.
(i.e. by
by using
using
products
consistent
products and goods
change
and goods in the
have incurred
few
incurred into
into a
Most
a shorter
of
shorter life-cycle
life-cycle flexible
technologies
flexible technologies
can
technologies be like
tailored
like Remote
at
Remote Laser
process
Laser Welding)
level (i.e.
Welding) or/and
by using
or/and
products
products and
and goods
goods have
have incurred
incurred into
into a
a shorter
shorter life-cycle
life-cycle flexible
flexible technologies
technologies like
like Remote
Remote Laser
Laser Welding)
Welding) or/and
or/and
and thus
products
and thus a
and
a shorter
goods
shorter time
have
time to
to market.
incurred
market. In
into
In addition,
a shorter
addition, customers
life-cycle
customers system
flexible
system level
level by
technologies
by using
usinglike the
the aforementioned
Remote Laser
aforementioned Welding)concepts.
or/and
concepts. A
A
and thus aa shorter time
time to to market. In
In addition, customers system system level
level by
by using
using the
the aforementioned concepts. A
and
and
oftenthus
oftenthus
require
a
require shorter
an increasing
shorter
an increasing
time to market.
level of
market.
level In addition, customers
of customization
customization
addition, entailing system
customers
entailing good example
good example
level of this
by
of this
using kind
kindthe of aforementioned
of application is
aforementioned
application concepts.
is represented
represented
concepts. A
by
A
by
often
often require
require an
an increasing
increasing level
level of
of customization
customization entailing
entailing good
good example
example of
of this
this kind
kind of
of application
application is
is represented
represented by
by
the increase
often
the increase
require of
an
of the varietylevel
increasing
the variety of products
of products
of and reduction
customization
and reduction
entailing of good
of assembly
assembly cells
example
cells used
of
usedthisin
in the
kind
the automotive
of application
automotive sector,
is
sector, in which
represented
in which a
a cell
by
cell
the increase of
of the
the variety of
of products and reduction of of assembly
assembly cells
cells used
used in
in the
the automotive
automotive sector,
sector, in
in which a cell
the
the increase
volumes
increase
volumes
volumes
(Wiendahl
of
(Wiendahl
(Wiendahl the variety
et al.
variety
et al.
et of
al. (2007)).products
(2007)).
products
(2007)). In
In order
In
and to
order
and
order reduction
deal
reduction
to
to
with
deal with
deal withof is is
designed
assembly
is designed
designed
and
cells
and
and
used
used
used
used in for
the the
for
for the
the
production
automotive
production
production
of
sector,
of
of in which
different
different
different
a
a cell
which partscell
parts
parts
volumes
these
volumes
these (Wiendahl
market
(Wiendahl
market needs
needs et al.
and
et
and al.
to
to (2007)). In
maintain
(2007)).
maintain In
their
theirorder to deal
deal with
competitiveness,
order to
competitiveness, with is is designed
designed
(S.J.
(S.J. Hu et
Hu and
et and
al. used for
al. (2011)).
(2011)).
used forTothe
Tothe production
exploit
production
exploit these of different
of different parts
these technological
technological parts
fea-
fea-
these market needs and
and toto maintain their
their competitiveness, (S.J. Hu et al. (2011)). To exploit these technological fea-
these marketand
production needs
and manufacturing maintain companies competitiveness,
are asked
asked to to (S.J.tures,Hu an et al.
al. (2011)).
optimized To
To exploit
system design these
needstechnological
to be
be generated. fea-
generated.
these market
production
production
production
needs
and
and
and to
manufacturing
manufacturing
manufacturing
maintain their
companies
companies
companies
competitiveness,
are
are
are asked to
asked to (S.J.tures,
tures,
tures,
For
Huan
an
an
these
et (2011)).
optimized
optimized
optimized
reasons,
system
system
system
design
exploit
design
design
design
and
these
needs
needs
needs
performance
technological
to
to
to be
be
fea-
generated.
generated.
evaluation of
answer these
production
answer theseand challenges with cost-effective
manufacturing
challenges with cost-effective
companies and
and yetasked
are
yet efficient
efficientto tures,
For an
these optimized
reasons, system
design design
and needs
performance to be generated.
evaluation of
answer
answer these
these challenges
challenges with
with cost-effective
cost-effective and
and yet
yet efficient
efficient For
For these
these reasons,
reasons, design
design and
and performance
performance evaluation
evaluation of
of
solutions.
answer
solutions. theseIn
In the last
challenges
the last few
few years,
with
years, several
cost-effective
several concepts
and
concepts yethave been
efficient
have been an
For
an easy
these
easy changeable
reasons,
changeable assembly
design
assembly and cell
cell represent
performance
represent two
two key
evaluation
key tasks
of
tasks
solutions.
solutions. In
In the
the last
last few
few years,
years, several
several concepts
concepts have
have been
been an
an easy
easy changeable
changeable assembly
assembly cell
cell represent
represent two
two key
key tasks
tasks
developed
solutions.
developed Inin in
inthemanufacturing systems
last few years,systems
manufacturing design
several concepts
systems design to to
tohaveaddress for
an
for these
easy
these reconfigurable
changeable
reconfigurable production
assembly cell
production
been for these reconfigurable production systems. This paper
address systems.
represent
systems. two This
key
This paper
tasks
paper
developed manufacturing design address
address for these reconfigurable production
developed
these issues:
developed
these
these
issues:
issues:
in modularity,
in manufacturing
modularity,
manufacturing
modularity,
systems design
reconfigurability
systems
reconfigurability
reconfigurability
design
and
and
toflexibility.
and to flexibility.
address
flexibility.
presents
for these an
presents
presents
an
reconfigurable
an integrated
integrated production systems.
integrated methodology
methodology
methodology to
to
select This
to select
systems.
select
paper
equipment
This paper
equipment
equipment
these
These
these
These issues:
kind of
issues:
kind modularity,
of modularity,
system paradigms
system reconfigurability
paradigms (ElMaraghy
reconfigurability
(ElMaraghy and
and flexibility.
(2006)) presents
and
have presents
flexibility.
(2006)) have and assess
assessan
an integrated
performance
integrated
performance methodology
of an
methodology
of an to
adaptable
to
adaptable select
select equipment
assembly
equipment
assembly cell
cell
These
These kind
kind of
of system
system paradigms
paradigms (ElMaraghy
(ElMaraghy (2006))
(2006)) have
have and
and assess
assess performance
performance of
of an
an adaptable
adaptable assembly
assembly cell
cell
the
These
the prerogative
kind
prerogative of to
system
to be
be easy and
paradigms
easy and quickly
(ElMaraghy
quickly adaptable
(2006))
adaptable to
to cope
have
cope where
and
where different
assess
different easy
performance
easy changeable
of
changeable an hardware
adaptable
hardware modules
assembly
modules can
can be
cell
be
the
the prerogative
prerogative to
to be
be easy
easy and
and quickly
quickly adaptable
adaptable to
to cope
cope where
where different
different easy
easy changeable
changeable hardware
hardware modules
modules can
can be
be
with
the new
prerogative
with new requirements
new requirements to be
requirements in easyin terms
and
in terms of
quickly
terms of both
of both hardware
adaptable
both hardware to
hardware and and
cope arranged
where
arranged around
different
around a
easy
a skeleton
changeable
skeleton architecture
hardware
architecture
and arranged around a skeleton architecture to provide differ- to
to provide
modules
provide differ-
can be
differ-
with
with new requirements in terms of both hardware and arranged
ent assembly
assemblyaround a skeleton
technologies architecture
with different to to provide
execution differ-
modes
software
with
softwarenew capabilities.
requirements
capabilities. The
Thein impactofof
terms
impact ofboth
thesehardware
these solutionsand
solutions is arranged
is ent around a skeleton
technologies architecture
with different provide
execution differ-
modes
software
software
maximized
software
maximized
capabilities.
capabilities.
when
capabilities.
The
The impact
product,
when product, The
product, impact
process
impact
process of
of
of
and
these
these
andthese
solutions
solutionssys-
production
solutions
production sys-is ent
is
is ent
and
ent
and
assembly
assembly
performances.
assembly
performances.
technologies
technologies
Its
technologies
Its main
main
with
with
aim
with
aim
different
different
is to
different
is to be
be
execution
execution
able to
execution
able to deal
deal
modes
modes
with
modes
with
maximized
maximized when
when product, process
process and
and production
production sys-
sys- and
and performances.
performances. Its
Its main
main aim
aim is
is to
to be
be able
able to
to deal
deal with
with
tem can
maximized
tem can evolve
when
evolve jointly,
product,
jointly, in
in a
a co-evolution
process
co-evolutionand context
production
context (Tolio
sys-
(Tolio high
and
high product
performances.
product variety
variety Its with
main
with low
aim
low volumes
is to
volumes be request.
able
request. to The
deal
The pro-
with
pro-
tem
tem can
can evolve
evolve jointly,
jointly, in
in a
a co-evolution
co-evolution context
context (Tolio
(Tolio high
high product
product variety
variety with
with low
low volumes
volumes request.
request. The
The pro-
pro-
et
tem al.
et al. (2010)).
can evolve
al. (2010)).
(2010)). The The same
jointly,
The same inrapid
same rapid a and
co-evolution
rapid and consistent
and consistent contextchange
consistent change (Tolio
change is is posed
posed approach
approach is
is composed
composed of
of three
is posed approach is composed of three integrated tools. pro-
high product variety with low volumes
three integrated
request.
integrated tools.
The
tools. The
The
et
et al. (2010)). The same rapid and consistent change is posed approach is composed of three integrated tools. The
The
alsoal.affecting
et
also affecting
(2010)). assembly
The
assembly same industry.
rapid
industry. In consistent
and
In fact, the
fact, the shortening
shortening
change is first one
posed
first one addresses
approach
addresses is the definition
the definition
composed of of process
three
of process alternatives
integrated tools.
alternatives for
The
for
also
also affecting
affecting assembly
assembly industry.
industry. In
In fact,
fact, the
the shortening
shortening first
first one
one addresses
addresses the
the definition
definition of
of process
process alternatives
alternatives for
for
of
also
of product
affecting
product life-cycles
assembly
life-cycles and the
industry.
and the increasing
In fact,
increasing set
the
set of product
shortening
of product parts
first
parts to
one
to be
be produced.
addresses
produced. the Process
definition
Process alternatives
of process
alternatives take
take into
alternatives
into consid-
for
consid-
of
of product
product life-cycles
life-cycles and
and the
the increasing
increasing set
set of
of product
product parts
parts to
to be
be produced.
produced. Process
Process alternatives
alternatives take
take into
into consid-
consid-
variants
of product
variants to
to be assembled
life-cycles
be assembled and in
inthethe
the same
increasing
same automated
set
automated of system
product
system eration
parts
eration to both
be
both technology
produced.
technology and
Process
and execution
alternatives
execution modes
take
modes for
into
for a
a given
consid-
given
variants
variants to to be be assembled
assembled in in the
the same
same automated
automated system system eration eration
eration both technology
both technology
technology and execution
and execution
execution modes
modes for for
forina given
a terms
given
set
set of operations.
of operations.
both Moreover,
Moreover,and also requirements
also requirements
modes ina given
variants


 This
to
research
be assembled
was supported
in by
thethe
same
EU
automated
project
system
RobustPlaNet, set
set of
of operations.
operations. Moreover,
Moreover, also
also requirements
requirements in terms
terms

This
This research was supported by the EU project RobustPlaNet,
research was supported by the EU project RobustPlaNet, of volumes
set
of volumes
of and product
operations.
and product
Moreover,mix are
mix are
also considered.
requirements
considered. Thein
The in terms
second
terms
second
FP7-2013-NMP-ICT-FOF,
 This
This research
research was
FP7-2013-NMP-ICT-FOF, was supported Grantby
Grant
supported 609087
the
the EU
609087 and project
and by Smart
by SmartRobustPlaNet,
Manufactur-
Manufactur- of volumes and product mix are considered. The second
FP7-2013-NMP-ICT-FOF, Grantby 609087 EU
and project
by SmartRobustPlaNet,
Manufactur- of
one
of
one volumes
considers
volumes
considers and
andthe
theproduct
selection
product
selection mix
mixof of are
are considered.
hardware
considered.
hardware modules
modules The second
The able
second
able to
to
ing
ing 2020
2020 of
of the
FP7-2013-NMP-ICT-FOF,
the
FP7-2013-NMP-ICT-FOF, Cluster
Cluster Tecnologico
Grant
Tecnologico
Grant Nazionale
609087 and
Nazionale
609087 and Fabbrica
by Smart
Fabbrica
by Smart
ing 2020 of the Cluster Tecnologico Nazionale Fabbrica Intelligente.
Intelligente.
Manufactur-
Intelligente.
Manufactur- one
one considers
considers the
the selection
selection of
of hardware
hardware modules
modules able
able to
to
ing 2020 of the Cluster Tecnologico Nazionale Fabbrica Intelligente.
ing 2020 of the Cluster Tecnologico Nazionale Fabbrica Intelligente.
one considers the selection of hardware modules able to
Copyright 2016 IFAC
2405-8963 © 2016,
Copyright IFAC 47 Hosting by Elsevier Ltd. All rights reserved.
IFAC (International Federation of Automatic Control)
Copyright © 2016
2016 IFAC 47
47
Copyright
Peer review©
Copyright 2016
©under IFAC
2016 responsibility
IFAC 47
of International Federation of Automatic
47 Control.
10.1016/j.ifacol.2016.07.548
IFAC MIM 2016
48
June 28-30, 2016. Troyes, France Alessio Angius et al. / IFAC-PapersOnLine 49-12 (2016) 047–052
provide requested technologies and execution modes, and
arrange them within a given system architecture. Finally,
the third one is a performance evaluation based analysis:
it consists in an analytical method to estimate main KPI s
(i.e. throughput and WIP) for given configuration and
execution modes selected. Given the results of performance
evaluation, the configuration tool is able to select best
set of pieces of equipment and the associated arrange-
ment able to manage the production of requested parts
and volumes at minimal cost. A testing of the developed
approach to a realistic case study in automotive industry
is also reported.
The structure of the paper is the following: Section 2
provides a review of the available approach, considering
both the configuration and performance evaluation area.
Section 3 describes the approach in terms of the function
and characteristics of the different tools. Section 4 reports
results of the test on an industrial use case, while Section
5 reports the conclusions and foreseen evolution of the
presented approach.
2. STATE OF ART
Regarding the presented methodology, it is possible to
classify the related works in two groups of interest: produc-
tion cell configuration problem, and production cell perfor-
mance evaluation. The first group contains all the contri-
butions about how it is possible to design a manufacturing
system in order to choose the minimum-cost configuration.
The second group instead, contains all the contributions
and advancements in the field of performance evaluation
of production cells.
Regarding first point, contributions taken into account
concern the connection between configuration and opti-
mization problem: first works addressing this issue are
presented in Wesolowski (1973) and Rosenblatt (1986). In
these papers, authors consider the arrangement of physical
facilities within a production system by optimizing a cost
function. Their most important contribution is on prob-
lem formalization: they highlight the need of connection
between layout design, production process and market
demand. In cell manufacturing design problem, major
differential cost is the handling system movement and
routing, as exposed in Massoud (1999). Other examples of
cost minimization are represented by worker’s assignment
(Aalaei et al. (2010)) or machine reliability and mainte-
nance (Das et al. (2007)). Regarding the design of flexible
lines it is possible to include in this investigation Tolio
and Urgo (2013). In this paper we propose an innovative
framework, because we address the resource needing for a
production process by list several ways in which a certain
assembly activity can be run using modular devices.
Regarding the second group of interest, there are two main
families of performance evaluation techniques, namely sim-
ulation and analytical methods. Simulation techniques
usually involves precise estimation of systems KPI s at the
expense of high computation times, model development
and validation. Analytical usually involves fast computa-
tion times at the expense of high error on KPI s estimation.
We will focus on the second family of techniques by nar-
rowing the analysis on analytical techniques for modeling
production systems with general distributed failure and re-
48
pair times, multiple-product assembly and non-negligible
transportation time. Several researchers have developed
approximate analytical models for manufacturing systems
with assembly operations, in particular Levantesi et al.
(2003) analyzed assembly production systems with multi-
ple failure modes machines. However, this model assumed
homogeneous production lines. In Colledani et al. (2005)
and Colledani et al. (2008) authors analyzed systems hav-
ing automated machines with multiple failure modes, with-
out incorporate assembly operations. Currently, the case of
multiple products can only be treated by approximating
the multiple product flows as an aggregated flow of an
averaged and fictitious equivalent single product. However,
this approach leads to accurate results only if processing
times and reliability parameters for two or more products
are the same. Otherwise, the accuracy of this approach is
very poor if the results are compared to DES. In literature
there are also several contributions about the modeling
and analysis of production lines with general distributed
failure and repair events. In Tan and Gershwin (2009),
authors present the modeling of a two-machine line with
general distributed failure and repair events; in Tolio and
Ratti (2013) a two-machine line with general Markovian
machines and generalized thresholds is faced, while in
Colledani and Gershwin (2011) authors present a decom-
position methodology to analyze long production lines
composed by Markovian machines. As advancement from
aforementioned contributions, in Colledani et al. (2015)
authors present an efficient technique to evaluate multi-
product assembly manufacturing systems, with general
distributed failure and repair events. In this last contri-
bution, transportation time is not explicitly considered. In
the context of the presented paper, analytical method pro-
posed involves the modeling of general distributed failure
and repair events, multi-product assembly and transporta-
tion time.
3. SOLUTION APPROACH
The presented approach is composed of three different
modules that perform a loop iteration, as depicted in
Figure 1. These three modules are described below:
(1) Demand and processes requirements: it addresses en-
vironment representation through volumes and pro-
duction mix requested by the market and, moreover,
it addresses assembly process representation of each
product using different execution modalities, that are
different ways to execute the same set of operations
using different equipment pieces.
(2) Equipment selection tool : it provides the selection
of hardware modules needed to execute a given set
of operations in a given execution modality. It also
arranges these pieces of equipment onto the assembly
line architecture and calculates the investment cost
of selected equipment. After the evaluation of the
performance, it also selects best combination of pieces
of equipment to match the production requests at a
minimal cost.
(3) Performance evaluation tool : it automatically gener-
ates a dynamic model of system behavior to evaluate
the performance in an analytic way. Performances
are estimated under the assumption that the cell can
manage the production of a single product at a time,
 IFAC MIM 2016
June 28-30, 2016. Troyes, France Alessio Angius et al. / IFAC-PapersOnLine 49-12 (2016) 047–052 49

undergoing a setup to move to new part production. It is possible to exploit this modality because of 7-
Without taking into account batch size and setup axes flexibility and versatility, indeed it is sufficient to
times, the tool perform an average evaluation of sys- mount a particular tool (e.g. glue gun for an adhesive
tem capability. joining) on the ending part of the robot in order to
use it as a production machine.
(4) Part is handled by 7-axes robot and the machine works
on it. In this situation the 7-axes robot moves the
part near the module (containing the machine) and
handles it during the processing.
(5) Part goes outside the line because needed machine is
placed in another dedicated cell. This is the simpler
situation, in which the needed technology is not
available and thus the part has to be moved outside
the cell.

Fig. 1. Integrated tool architecture.

3.1 Demand and process requirements

The first module is a twofold objective formalization

tool. On one hand, it addresses environment description
through a product mix and production volumes. Indeed,
for each product i ∈ I it is possible to obtain its requested
volume volumei from the market.On the other hand, this
module addresses assembly process representation through
the identification of different execution modalities to run
each operation. An execution modality represents a com-
mon practice in industrial operation, that is how to run
a certain process with a given set of equipment and ma-
chines. It is possible to identify different modalities for
each operation: these modalities diverge from one another
for different sub-operation and elements required. Execu-
tion modalities identified are also useful for configuration
generator (depicted in Figure 1 as an input of Equipment
selection tool ) that considers them for arranging requested
equipments within the cell. Modular devices, containing
machines or a collection of tools or fixtures, are also con-
sidered during the generation. These devices have the main
characteristic to be easy changeable and thus, by changing
a module containing (for example) spot welding devices
with one accommodating a tabletop hemming, an easy cell
reconfiguration is achieved. It is clear how, with modular Fig. 2. Execution modalities description: for each one an exemplar
devices and a easy cell reconfiguration, the opportunity configuration and task sequencing is given.
to run the same operation in different ways has to be
taken into account during cell design. The main execution As already mentioned, there are two main differences
modalities, represented in Figure 2, are: from an execution modality to another: set of equipments
(machines, tools, fixture) required and the set of sub-
(1) Part is worked inside the module. This is the case of a operations, and thus task sequence. It is possible to see in
machine that has a working cube. The part has to be Figure 2 these two characteristics: on the left side of the
transported inside the working cube from the input picture element arrangements for each execution modality
station (or previous module) by 7-axes robot with a are depicted, instead, on the right side different exemplar
handling tool. Once the part is inside the module, it task sequences are enlisted. These two characteristics af-
can be worked. fect cell configuration and its performance evaluation be-
(2) Part is blocked in the fixture and machine works it. cause considering different task sequences we have different
In this case the machine has an anthropological shape production processes and thus different performances.
that allows it to work on a part in the nearby of it. In
this situation it is sufficient for the machine to have 3.2 Equipment selection tool
the part available near it blocked in a fixture. Again,
a 7-axes robot handling robot has to move the part The Equipment selection tool has the aim of define a
within the cell, through the chosen fixture. candidate set of hardware modules together with their
(3) Part is blocked in the fixture and 7-axes works on it. In arrangement within the assembly line architecture. This
this case the 7-axes robot has the main manufacturing candidate configuration (conf in formulation) is passed
duty: work the part while it is blocked in the fixture. to the performance evaluation tool (P E in formulation)

49
IFAC MIM 2016
50
June 28-30, 2016. Troyes, France Alessio Angius et al. / IFAC-PapersOnLine 49-12 (2016) 047–052
to assess the associated performance. Grounding on this,
the Equipment selection tool aims at defining the pieces
of equipment that must be acquired together with their
configuration of different part types to produce. After the
loop with the third step tool, the aim of Equipment selec-
tion tool becomes the satisfaction of production requests
at a minimal cost. It means that the second iteration of
presented tool (in Figure 1) evaluates the feasibility and
optimality of each cell configuration generated in the pre-
vious iteration. The cost function taken into consideration
is:
min (T otalCost) = P urchCost(conf ) + OpCost · T tot
subject to
conf = generator (ExM odalitiesik ) ∀i, ∀k
T Haveragei = P E (conf, ExM odalitiesik ) ∀i, ∀k
I
T tot =
volumei
i
T Haveragei
T tot ≤ T available
i = 1···I
k = 1···K
Hence, equation (1) is able to choose the minimum-cost
arrangement configuration conf considering the purchas-
ing cost P urchCost(conf ) for each equipment to be ac-
commodated in the cell, and also the operational cost
OpCost · T tot (e.g. electric energy or human operators
cost) needed to run the assembly line. This second cost
type considers the time spent to produce all the products
i ∈ I requested from the market, as described in (4).
The throughput of a single product is affected by selected
execution modality and equipment configuration. These
parameters are handled from the Performance evaluation
tool as depicted in (3) with notation P E. The time given
in output from P E tool represents the average throughput
that selected system is able to perform in producing the
given product i. In order to investigate all possible layout
solutions, the performance evaluation has to consider all
feasible execution modalities for each operation k ∈ K
and each product i ∈ I (highlighted with configuration
generator in (2)). As a final constraint, total production
time has to comply with the considered time horizon, like
in (5).
3.3 Performance evaluation tool
In this section we provide the mathematical description of
the system as composition of N operational stations (also
called modules) and a 7-axes robot (only 7-axes from now-
on). The state of the system is fully described by a vector
s = |c1 , c2 , . . . , cN , r| where:
• ci , 1 ≤ i ≤ N , describes the state of ith module by
assuming the values: Operative (O), Starved (S) and
Blocked (B).
• r describes the state of 7-axes by taking values
from the interval [0, 1, . . . , N, N + 1] where: 0 is a
bookmark that corresponds to the case in which 7-
axes is idle and the values i ∈ [1, . . . , N + 1], are the
operative states in which 7-axes is working a part that
will be placed in module i at the end of the service.
Note that, in order to provide a compact notation,
N + 1 is a fictitious module that corresponds the
outside of the system. For this reason, module N +1 is
always starved and unaffected by changings of state.
In order to describe the system dynamics, following con-
siderations have been made.
• Parts are moved inside the system only by 7-axes.
• Raw material is always available at the first module
and the loading, that requires negligible time, does
not require 7-axes. Therefore, the system will never
reach a state in which m1 = S.
(1) • No storage is possible between modules. Hence, the
module is blocked after end of service until 7-axes
takes the part.
(2) • The 7-axes moves parts only in modules that do not
(3) contains parts already. Therefore, 7-axes can be idle
even if there is a blocked module; furthermore, this
(4) also means that a module i, 2 ≤ i ≤ N , becomes
starved once the 7-axes removes a part from it.
• The 7-axes is always able to unload a part that have
(5)
been processed inside the last module of the system.
• 7-axes takes negligible time to remove a part from
a module. Therefore, a machine gets empty immedi-
ately after 7-axes remove the part.
• All the machines of the system work asynchronously
among each others.
The considerations written above lead to a set of events
that fully describes the dynamics of the system. These
events are conveniently depicted by logical expressions
that describe what is verified immediately before and
immediately after the occurrence of an event. We provide
the description of an event in terms of its pre-conditions
and post-conditions: i) < pre − conditions > is the logical
expression describing under which conditions an even can
occur; ii) < post − conditions > describes how the state of
the system will change if such event occurs. We denote an
event with < pre − condtions > → < post − condtions >.
In sake of brevity, each logical expression will indicate only
those variables that are directly involved in the event.
Given a generic state s, the events describing the system
dynamics are the following:
(1) ci = O ∧ ci+1 = S ∧ r = I → ci = S ∧ r = i + 1, 2 ≤ i ≤ N ;
(2) ci = O ∧ (ci+1 = S ∨ r = I) → ci = B, 2 ≤ i ≤ N ;
(3) c1 = O ∧ c2 = S ∧ r = I → c1 = O ∧ r = 2;
(4) c1 = O ∧ (c2 = S ∨ r = I) → c1 = B;
(5) r = k ∧ k > 1 ∧ ∃i > 1 : (ci = B ∧ ci+1 = S) → ck = O ∧
ci = S ∧ r = i;
(6) r = k ∧ k > 1 ∧ c1 = B ∧ c2 = S → ck = O ∧ ci = O ∧ r = 2;
(7) r = k ∧ k > 1∧  ∃i : (ci = B ∧ ci+1 = S) → ck = O ∧ r = I.
Events (1) and (2) model the end of a service at ith
module. In particular, event (1) corresponds to the case in
which 7-axes is idle and can remove the part from module
i immediately. On the contrary, event (2) corresponds to
the case in which 7-axes is already processing a part.
Therefore, in (1) ci will be blocked whereas in (2) it will be
starved. Events (3)-(4) follow the same path but refer to
the first module that, when not blocked, returns operative
instead of starved. Finally, events (5)-(7) correspond to
the end of service of 7-axes. In detail, event (5) and
(6) describe the case in which 7-axes finishes to move
a part, then, takes another part from another module i.
The difference between these two events is that module
50
 IFAC MIM 2016
June 28-30, 2016. Troyes, France Alessio Angius et al. / IFAC-PapersOnLine 49-12 (2016) 047–052 51

1 always returns operative immediately after the part is 4. APPLICATION

removed from it, whereas every other module gets starved
because it must wait for a new part. Event (7) describes
the ending of 7-axes service when system does not contain
any other part ready to be taken; in this situation, 7-
axes gets idle. As further constraint, we assume that 7-
axes serves always the module that is the closest to the
end of the line amongst those blocked. Therefore, every
state s is reachable by one event of type (5)-(6) at the
most. At least in principle, different policies are possible
but we leave this investigation as future work. We assume
that every change of state occurs according to Markovian
distributions in such a way that the underlying stochastic
process is a Continuous Time Markov Chain (CTMC ).
We assume phase-type distributions (PH ) to model the
processing times of modules and 7-axes. Let us remind that
a random variable T is PH distributed if its cumulative
distribution function (cdf ) corresponds to the time till
absorption of a CTMC, given a pre-fixed initial distribu-
tion. The structure of PH distributions allows the fitting
of general distributions by matching their higher moments
(see for example (Horvath and Telek, 2007)). Furthermore,
PH -type distribution allows us to model the system by
using a component-based approach in which the behavior
of every module and the 7-axes is a CTMC that describes,
possibly complex, dynamics such as model failures, repairs,
degradations etc.. This lead to an infinitesimal generator,
denoted by Q, that is composed of blocks, referred as
Qs,s , that describe the motion of the process between
two states of the system. Let us denote the infinitesimal
generator, the initial vector and the firing vector of the
PH -type describing the dynamics of the ith module with
Λi , αi and fi respectively. Similarly let Λa , αa and fa
be the same quantities for 7-axes robot. Diagonal blocks
Qs,s represent the sojourn distribution time in state s;
thus, they correspond to the Kronecker sum of PH -type
distributions that are enabled in s. In formula

Qs,s =
∀i∈A(s)
where A(s) is the set of PH -type that are active in state
s. Out diagonal blocks Qs,s are not null only if they
correspond to events that are able to move the state of the
system. This allows us to define the blocks in reference to
the events described above. Thus, we have:
The viability of the proposed approach has been also
faced in a realistic industrial case in automotive sector.
In particular, we are focusing on the assembly of spare
parts for the aftermarket, that have small production vol-
umes. In order to face the production of these parts in a
gainful way, an assembly cell containing modular devices
is used. In this way it will be not needed an assembly cell
for each product type but only a flexible assembly cell
able to manage different assembly processes. Presented
application case foresees the presence of three different
products to be assembled with a requested volume of 300
parts for each product 1, 2 and 3. In addition, for each
product it is foreseen the same assembly process composed
by an input activity and three generic operations A, B
and C, with a product transportation after each activity,
executed by the 7-axes robot. Each of these operations
could be executed with execution modality number 1 or
2, following the classification exposed in Section 3.1: oper-
ation times, for which an Erlang distribution parameter
4 is assumed, are reported in Table 1. Purchasing and
operational costs (in Euros/h) are described in Table
2. The total available time is 160 hours (a month, one
shift a day, 8 hours/shif t). Grounding on these data,
three different configurations are generated and evaluated.
The first solution foresees execution modality number 1
for operations A and C, and number 2 for operation
B. The second one foresees execution modality number 1
for operation A, and number 2 for operations B and C.
Instead, the last one foresees execution modality number 2
for operation A, and number 1 for operations B and C. As
already explained, each configuration differs from others
for the set of equipment chosen and arranged within the
cell.
Table 1. Process times for each product and for each
execution modality.
Product A B C
ex mod 1 18 210 27
1
ex mod 2 22 185 43
ex mod 1 27 177 23
2
ex mod 2 31 152 25

(1) h1 (s) ⊗ · · · ⊗ fi ⊗ · · · ⊗ hN (s) ⊗ αa ex mod 1 21 221 26

3
(2) h1 (s) ⊗ · · · ⊗ fi ⊗ · · · ⊗ hN (s) ⊗ ha (s) ex mod 2 25 191 34
(3) f1 ⊗ α1 ⊗ h2 (s) ⊗ · · · ⊗ hN (s) ⊗ αa
(4) f1 ⊗ α1 ⊗ h2 (s) ⊗ · · · ⊗ hN (s) ⊗ ha (s) Table 2. Application case miscellaneous costs.
(5) h1 (s) ⊗ h2 (s) ⊗ · · · ⊗ αi ⊗ · · · ⊗ hN (s) ⊗ fa αa
(6) α1 ⊗ h2 (s) ⊗ · · · ⊗ hN (s) ⊗ fa αa Miscellaneous costs ex mod 1 ex mod 2
(7) h1 (s) ⊗ h2 (s) ⊗ · · · ⊗ hN (s) ⊗ fa αa
A equipments 25 000 28 000
B equipments 50 000 45 000
where C equipments 45 000 48 000
 
I i ci = O Ir r = O 7-axes robot 75 000
hi (s) = ha (s) =
1 otherwise 1 otherwise input station 45 000
output station 15 000
and Ix is an identity matrix having the same dimension- operational cost 50
ality of matrix Λx . Let us give an insight of how these
formulas work by considering case (3) where the first
module finishes the service, wakes up 7-axes and starts Results in term of throughput (in parts/second), total
the processing of another part: i) the Kronecker product time (in hours) and total cost (in Euros) for each solution
between the first two terms describes the end and the are exposed in Table 3. As it is possible to see solution
restart of the service at the first module; ii) matrices h number 3 exceeds time constraint (time needed is greater
preserve the aging of the PH -type active in state s; finally, than 160 hours) and the minimum-cost one is the number
the last term models the starting of 7-axes service. 1, which representation is reported in Figure 3.

51
IFAC MIM 2016
52
June 28-30, 2016. Troyes, France Alessio Angius et al. / IFAC-PapersOnLine 49-12 (2016) 047–052
Table 3. Application case results.
Results sol 1 sol 2 sol 3
TH prod 1 0.001252 0.001245 0.000971
TH prod 2 0.001876 0.001866 0.001352
TH prod 3 0.001155 0.001151 0.000840
Total time 152.60 153.33 205.57
Total cost 257 630 260 666 268 278
Fig. 3. Exemplar cell describing solution 1.
5. CONCLUSION
In this article we propose an analytical approach able to
select the minimum-cost configuration of an assembly cell.
We present also a set of execution modalities, useful con-
cept for formalize different way to run a process; formaliza-
tion of these modalities represents a connection between
human designer expertize and an analytical method. They
give also an additional degree of freedom in the production
system design field. In addition, presented method is used
only for the configuration of a cell by choosing the set of
equipments that can ensure minimum cost, but it possible
to extend it for a reconfiguration aim: in this case, it is
needed to consider several time buckets by describing the
environment evolution in terms of volumes and product
mix, and also by describing the reconfiguration behavior
of the cell. Another developing way could be represented
by the introduction of a risk index in order to take into
account performance variance. At least, we consider only
a configuration generator for a cell layout that assume
a spatial allocation with the 7-axes robot between two
rows of machines (like in Figure 3); it could be possible
to change the generator in order to face another kind of
layout paradigm and thus use the entire instrument as a
flexible tool, on a different configuration problem.
REFERENCES
Aalaei, A., Paydar, M.M., Solimanpur, M., and Mahdavi,
I. (2010). Designing a mathematical model for dynamic
cellular manufacturing systems considering production
planning and worker assignment. Computers and Math-
ematics with Applications, 60(4), 1014–1025.
Colledani, M., Gandola, F., Matta, A., and Tolio, T.
(2008). Performance evaluation of linear and non-
linear multi-product, multi-stage lines with unreliable
machines and finite homogeneous buffers. IIE Transac-
tion, 40, 612–626.
Colledani, M. and Gershwin, S. (2011). A decomposition
method for approximate evaluation of continuous flow
52
multistage lines with general markovian machines. An-
nals of Operations Research.
Colledani, M., Matta, A., and Tolio, T. (2005). Perfor-
mance evaluation of production lines with finite buffer
capacity producing two different products. OR Spec-
trum, 27, 243–263.
Colledani, M., Ratti, A., and Senanayake, C. (2015). An
approximate analytical method to evaluate the perfor-
mance of multi-product assembly manufacturing sys-
tems. Procedia CIRP, 33, 358–364.
Das, K., Lashkari, R.S., and Sengupta, S. (2007). Reliabil-
ity consideration in the design and analysis of cellular
manufacturing systems. International Journal of Pro-
duction Economics, 105(1), 243–262.
ElMaraghy, H., Schuh, G., ElMaraghy, W., Piller, F.,
Schonslben, P., Tseng, M., and Bernard, A. (2013).
Product variety management. CIRP Annals - Manu-
facturing Technology, 62(2), 629–652.
ElMaraghy, H.A. (2006). Flexible and reconfigurable man-
ufacturing systems paradigms. International Journal of
Flexible Manufacturing Systems, 17, 261–276.
Horvath, A. and Telek, M. (2007). Matching more than
three moments with acyclic phase type distributions.
Stoch. Models, 23, 167–194.
Levantesi, R., Matta, A., and Tolio, T. (2003). Perfor-
mance evaluation of continuous production lines with
machines having different processing times and multiple
failure modes. Performance evaluation, 51, 247–268.
Massoud, B.L. (1999). Layout designs in cellular manufac-
turing. European Journal of Operational Research, 112,
258–271.
Rosenblatt, M.J. (1986). The dynamics of plant layout.
Management Science, 32(1), 1–12.
S.J. Hu, S.J., Ko, J., Weyand, L., ElMaraghy, H.A., Lien,
T.K., Koren, Y., Bley, H., Chryssolouris, G., Nasr, N.,
and Shpitalni, M. (2011). Assembly system design
and operations for product variety. CIRP Annals -
Manufacturing Technology, 60, 715–733.
Tan, B. and Gershwin, S. (2009). Analysis of a general
markovian two-stage continuous-flow production system
with a finite buffer. International Journal of Production
Economics.
Tolio, T., Ceglarek, D., ElMaraghy, H.A., Fischer, A.,
J., H.S., Laperrire, L., Newman, S.T., and Váncza,
J. (2010). Co-evolution of products, processes and
production systems. CIRP Annals - Manufacturing
Technology, 59, 672–693.
Tolio, T. and Ratti, A. (2013). Performance evaluation
of two-machines line with generalized thresholds. In
Proceedings of the 9 th International Conference on
Stochastic Models of Manufacturing and Service Opera-
tions.
Tolio, T. and Urgo, M. (2013). Design of flexible trans-
fer lines: A case-based reconfiguration cost assessment.
Journal of Manufacturing Systems, 32(2), 325–334.
Wesolowski, G.O. (1973). Dynamic facility location. Man-
agement Science, 19(11), 1–9.
Wiendahl, H.P., ElMaraghy, H.A., Nyhuis, P., Z ä h,
M.F., Wiendahl, H.H., Duffie, N., and Brieke, M. (2007).
Changeable manufacturing - classification, design and
operation. Annals of the CIRP, 56(2), 783–809.