Proceedings of the 3rd IFAC Conference on

Proceedings
Advances of the 3rd IFAC Conference on Control
in Proportional-Integral-Derivative
Proceedings of the 3rd IFAC Conference on
Proceedings
Advances in of the 3rd IFAC Conference on Control
Proportional-Integral-Derivative
Ghent, Belgium,
Advances May 9-11, 2018 Available
in Proportional-Integral-Derivative online at www.sciencedirect.com
Control
Proceedings
Advances of the 3rd IFAC Conference on Control
in Proportional-Integral-Derivative
Ghent, Belgium, May 9-11, 2018
Ghent, Belgium, May 9-11, 2018
Advances in Proportional-Integral-Derivative
Ghent, Belgium, May 9-11, 2018 Control
Ghent, Belgium, May 9-11, 2018
ScienceDirect
IFAC PapersOnLine 51-4 (2018) 25–30
FOPID
FOPID Controllers
Controllers and
and Their
Their Industrial
Industrial
FOPID Controllers
FOPID Controllers and Their
andofTheir Industrial
Industrial
Applications:
Applications:
FOPID A
A
ControllersSurvey
Surveyand of Recent
Recent
Their Results
Results
Industrial
1
11
Applications:
Applications: A A Survey
Survey of
of Recent
Recent Results
Results 1
Applications: A Survey
Aleksei Tepljakov ∗
Baris of Recent
Baykant Results 1
Alagoz ∗∗
Aleksei Tepljakov ∗ Baris Baykant Alagoz ∗∗
Aleksei
Aleksei Tepljakov
Celaleddin Yeroglu∗∗∗∗
Tepljakov ∗∗Baris Baykant
Emmanuel
Baris Baykant Alagoz
Gonzalez
Alagoz ∗∗
∗∗
∗∗∗
Celaleddin
Celaleddin Yeroglu
Yeroglu Emmanuel
∗∗∗ Emmanuel
∗∗ Gonzalez
Gonzalez
∗∗∗
∗∗∗∗
S. Hassan
Aleksei
Celaleddin HosseinNia
Tepljakov
Yeroglu
∗∗∗∗
Baris Eduard
∗∗∗∗ Baykant
Emmanuel Petlenkov
Alagoz
Gonzalez
∗∗
∗∗∗∗
S.
S. Hassan
Hassan HosseinNia
HosseinNia Eduard
∗∗ ∗∗∗∗ Eduard Petlenkov
∗∗∗∗ Petlenkov ∗∗∗∗
Celaleddin
S. Yeroglu Emmanuel
Hassan HosseinNia EduardGonzalez Petlenkov ∗
∗ Department of Computer Systems, Tallinn University of Technology,
∗
S. Hassan HosseinNia ∗∗∗∗
Eduard Petlenkov ∗
∗ Department of Computer Systems, Tallinn University of Technology,
Department
Tallinn,
∗
Department Estonia of Computer Systems,
(e-mail: {aleksei.tepljakov,
of Computer Tallinn University of Technology,
eduard.petlenkov}@ttu.ee)
Systems, Tallinn eduard.petlenkov}@ttu.ee)
University of Technology,
Tallinn,
Tallinn,
∗∗∗ Department
Estonia
Estonia (e-mail:
(e-mail: {aleksei.tepljakov,
{aleksei.tepljakov, eduard.petlenkov}@ttu.ee)
Department
Tallinn, Estonia of ofComputer
Computer
(e-mail: Engineering,
Systems, Tallinn
{aleksei.tepljakov, Inonu University,
University Malatya,
of Technology,
eduard.petlenkov}@ttu.ee)
∗∗ Department of Computer Engineering, Inonu University, Malatya,
∗∗
Department
Tallinn,
∗∗ Turkey
Estonia
Department of
(e-mail: Computer
of(e-mail:
Computer Engineering,
{baykant.alagoz,
{aleksei.tepljakov,
Engineering, Inonu University,
University, Malatya,
c.yeroglu}@inonu.edu.tr)
eduard.petlenkov}@ttu.ee)
Inonu Malatya,
Turkey
Turkey
∗∗ Schindler
(e-mail:
(e-mail: {baykant.alagoz,
{baykant.alagoz, c.yeroglu}@inonu.edu.tr)
c.yeroglu}@inonu.edu.tr)
∗∗∗
Department
Turkey Elevator
of
(e-mail: ComputerCorporation,
Engineering,
{baykant.alagoz, 1530 Timberwolf
Inonu Dr., Holland,
University,
c.yeroglu}@inonu.edu.tr) Malatya,
∗∗∗
Schindler Elevator Corporation, 1530 Timberwolf Dr., Holland,
∗∗∗ Schindler Elevator Corporation, 1530 Timberwolf Dr., Holland,
∗∗∗
Turkey OH, 43528,
(e-mail: USA (e-mail:
{baykant.alagoz, [email protected])
c.yeroglu}@inonu.edu.tr)
∗∗∗∗Schindler OH,
OH,
Elevator
43528,
43528,
Corporation,
USA
USA (e-mail:
(e-mail:
1530 Timberwolf Dr., Holland,
[email protected])
[email protected])
∗∗∗ Department
∗∗∗∗Schindler OH,Elevator of Precision
43528, and Microsystems
Corporation,
USA (e-mail: 1530 Timberwolf
[email protected]) Engineering,
Dr., Holland, Delft
∗∗∗∗ Department of Precision and Microsystems Engineering, Delft
∗∗∗∗ Department of Precision and Microsystems Engineering,
University
OH, 43528,
Department
University
of Technology,
USA
of Precision
of (e-mail:
Technology,
Delft, The
[email protected])
and Microsystems
Delft, The Engineering, Delft
Netherlands
Netherlands Delft
∗∗∗∗ University
Department
University(e-mail:of Technology,
of Precision
Technology, Delft,
Delft, The
of [email protected])
and Microsystems The Netherlands
Engineering, Delft
Netherlands
(e-mail:
(e-mail: [email protected])
University(e-mail:of [email protected])
Technology, Delft, The Netherlands
[email protected])
(e-mail: [email protected])
Abstract: The interest towards using Fractional-order (FO) PID controllers in the industry is
Abstract:
Abstract: The
The interest
interest towards
towards using
using Fractional-order
Fractional-order (FO)
(FO) PID
PID controllers in the industry is
mainly
Abstract:
mainly
fueled
Theby
fueled by
the fact
interest
the fact
that these
towards
that usingcontrollers
these Fractional-order
controllers
have two
have (FO)
two PID controllers
additional
additional
“tuningin
controllers
“tuning
the
the industry
inknobs”
knobs”
that can
industry
that
is
is
can
mainly
be used
Abstract:
mainly fueled
to adjustby the
Thebyinterest
fueled fact
control
the control
fact that
towards
that lawthese
in
using
these a controllers
way that
Fractional-order
controllers have
would
have twotwo
benefit
(FO) additional
the
PID
additional “tuning
control
controllers
“tuningloop. knobs”
However,
inknobs” that
the industry can
there
thatthere is
can
be
be used
used to
to adjust
adjust the law in a way that would benefit the control loop. However,
are
mainly
be certain
used to points
fueled
adjustby thatcontrol
the are that
fact
control law
rarely
law in
these
in a
a way
addressed that
controllers
way that would
in literature,
have
would benefit
two namely:
benefit the
additional
the control
(1) “tuning
What
control loop.
loop.are However,
the particular
knobs”
However, that there
can
there
are
are certain
certain points
points that
that are
are rarely
rarely addressed
addressed in
in literature,
literature, namely:
namely: (1)
(1) What are the particular
advantages
be
are used
certain
advantages
(in concrete
to adjust
points
(in the
thatcontrol
concrete
numbers)
are law in
rarely
numbers)
of aFOPID
way that
addressed
of FOPID
controllers
in literature,
controllers
versus
would benefit namely:
versus the (1) What
conventional,
control
What
conventional, loop.are the
the particular
integer-order
areHowever,
integer-order
(IO)
there
particular
(IO)
advantages
PID
are certain
advantages (in
controllers
pointsconcrete
in the
that
(in concrete numbers)
light
are of
rarely
numbers) of FOPID
complexities
addressed
of FOPID controllers
arising
in in the
literature,
controllers versus
namely:
versus conventional,
implementation (1) What
conventional, of integer-order
the
areformer?
the
integer-order (2) (IO)
particular For
(IO)
PID
PID controllers
controllers in
in the
the light
light of
of complexities
complexities arising
arising in
in the
the implementation
implementation of
of the
the former?
former? (2)
(2) For
For
real-time
advantages
PID implementation
(in
controllers concrete
in the light of
numbers)
ofFOPID of controllers,
FOPID
complexities arising approximations
controllers
in the versus are used
conventional,
implementation of that are
integer-order
the former? equivalent
(2) (IO)
For
real-time
real-time implementation
implementation of
of FOPID
FOPID controllers,
controllers, approximations
approximations are
are used
used that
that are
are equivalent
equivalent
to
PID high
real-time order
controllers linear
in the
implementation controllers.
light of of What,
complexities
FOPID then, is
arising
controllers, theinbenefit
the of using
implementation
approximations are FOPID
used of controllers?
the
that former?
are In
(2)
equivalent the
For
to
to high
high order
order linear
linear controllers.
controllers. What,
What, then,
then, is
is the
the benefit
benefit of
of using
using FOPID
FOPID controllers?
controllers? In
In the
the
present
real-time paper,
to high paper, we
order linear attempt
implementation to
of
controllers. address
FOPID these issues
controllers,
What,thesethen,issues
is theby reviewing
approximations
benefit recent
of using are literature
used
FOPID that in the
are
controllers? field
equivalentand
In and
the
present
present paper, we
we attempt
attempt to
to address
address these issues by
by reviewing
reviewing recent
recent literature
literature in
in the
the field
field and
by
to providing
high
present order relevant
linear
paper,relevant
we attempt analysis
controllers. and
to address recommendations.
What, then, is the benefit of using FOPID
these issues by reviewing recent literature in the field and controllers? In the
by
by providing
providing relevant analysis
analysis and
and recommendations.
recommendations.
present
by paper,
providing we attempt
relevant to
analysis address
and these issues
recommendations. by reviewing recent
© 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. literature in the field and
by providing
Keywords: relevant analysis
Fractional-order PID and recommendations.
Control, Industrial applications, Frequency domain analysis
Keywords: Fractional-order PID Control, Industrial applications, Frequency domain analysis
Keywords: Fractional-order PID Control, Industrial
Keywords: Fractional-order PID Control, Industrial applications,
applications, Frequency
Frequency domain domain analysis
analysis
Keywords: Fractional-order PID Control, Industrial
1. INTRODUCTION beenapplications,
carried outFrequency to confirm domain analysis
this (see, e.g., Xue et al.
1. INTRODUCTION
1. INTRODUCTION been
been carried
carried out
out to
to confirm
confirm this
this (see,
(see, e.g.,
e.g., Xue
Xue et
et al.
1. INTRODUCTION (2006);
been
(2006);
Čech
carried
Čech
and Schlegel
out Schlegel
and to confirm (2006); Luo
this Luo
(2006);
and
(see,and Chen
e.g.,
Chen et al.
Xue(2012)).
(2012)).al.
PID controllers are1.widely used for industrial process con- been
INTRODUCTION (2006);
Other
(2006); Čech
FOPID
carried
Čech and
out
and Schlegel
control
to loop(2006);
confirm
Schlegel structures
this
(2006); Luo
(see,
Luo and
have
and Chen
been
e.g.,
Chen Xue (2012)).
reviewed
et
(2012)).al.
PID
PID controllers
controllers are
are widely
widely used
used for
for industrial
industrial process
process con-
con- Other FOPID control loop structures have been reviewed
trol. While
PID controllers more powerful control techniques are readily Other
in, e.g.,
(2006); FOPID control
Tavakoli-Kakhki
Čech and loop
Schlegel structures
and
(2006);HaeriLuo have
(2011);
and been
Chen reviewed
Padula and
(2012)).
trol. While more morearepowerful
widely used
powerful control for industrial
techniquesprocess con- Other
are readily
readily FOPID control loopand structures have been reviewed
trol.
PID While
available,
Whilethe
trol. controllersmorePID controller
arepowerful
widely control
used
control techniques
isforstill popularprocess
industrial
techniques are
duereadily
are its in,
tocon- in, e.g.,
e.g.,
Visioli
Other
in,
Tavakoli-Kakhki
Tavakoli-Kakhki
(2013);
e.g.,FOPID Azarmi loop
control
Tavakoli-Kakhki et al.and Haeri
Haeri
(2015).
structures
and
(2011);
Haeri (2011);
have been
(2011);
Padula
Padula
reviewed
Padula
and
and
and
available,
available, the
the PID
PID controller
controller is
is still
still popular
popular due
due to
to its
its Visioli (2013); Azarmi et al. (2015).
relative
available,simplicity
trol. While themore and
PIDpowerfulapplicability
controllercontrol
is still to a wide
techniques
popular range
are of
duereadily
to in-
its in, Visioli (2013);
e.g.,(2013);
Visioli Azarmi
Tavakoli-Kakhki
Azarmi et al.
et al.and (2015).
Haeri (2011);
(2015). Padula and
relative
relative simplicity and applicability to a wide range of in- However, even though FOPID controllers offer advantages
relative simplicity
dustrial
available,
dustrial
control
the PID
simplicity
control
and
problems
and
problems
applicability
(Åström
controller
applicability
(Åström
and
is still
and
to aaHägglund
wide
wide range
to popular
Hägglund range of
of in-
due(2006)).
to its However,
in-
(2006)). Visioli
However, even
(2013);
even though
Azarmi
though FOPID
et
FOPID al. controllers
(2015).
controllers offer
offer advantages
advantages
dustrial
However, control
it is a problems
commonly (Åström
acknowledged andto aHägglund that(2006)).
fact range only a over IOPID
However, even controllers,
though FOPID the adoption
controllers of the former
offer in the
advantages
relative
dustrial simplicity
However, control
it is and applicability
is aa problems
commonly (Åström
acknowledged and wide
Hägglund
fact that of in-
(2006)).
only aa overover IOPID
IOPID controllers,
controllers, the
the adoption
adoption of
of the
the former
former in
in the
However,
fraction
dustrial
However, of itthe
control
it is commonly
existing
a problems
commonly PI/PIDacknowledged
(Åström controller
acknowledged and fact
based
Hägglund
fact that only
loops
that are
(2006)).
only a
industry
However,
over IOPID
industry
iseven
is
slow
slow
(ChenFOPID
though
controllers,
(Chen
(2006)).
the
(2006)).
In this
controllers
adoption
In of
this
paper,
theoffer
paper,
we aim
former
we in the
advantages
aim
to
the
to
fraction
fraction of the existing PI/PID controller based loops are industry
explore
over IOPID is
the slow (Chen
reasons
controllers, for (2006)).
this
the by In
adoption this
conductingpaper,
of paper, a
the former we aim
survey to
of
in the
tuned
However,
fractiontoof itthe
is aexisting
ofachieve
the optimal
commonly
existing PI/PID
performance
PI/PID controller
acknowledged
controller based
(O’Dwyer
fact
based loops
that onlyare
(2009)).
loops area industry
explore is slow
the (Chen
reasons for (2006)).
this by Inconducting
this a we aim
survey to
of
tuned
tuned to
to achieve
achieve optimal
optimal performance
performance (O’Dwyer
(O’Dwyer (2009)).
(2009)). explore
recent
industry the
results
is reasons
slowrelated
(Chen for
to this by
advantages
(2006)). Inconducting
of
thisFOPID
paper, a survey
controllers,
we aim of
to
fraction
tuned of the
to achieve existing
optimal PI/PID
performancecontroller based
(O’Dwyer loops
(2009)).are explore
recent the
results reasons
related for
to this by
advantages conducting
of FOPID a survey
controllers,of
In recent years, the emergence of fractional calculus has recent their
explore
recent results
implementation,
the
results related
reasons
related to
and
for
to advantages
industrial
this by
advantages of FOPID
applications.
conducting
of FOPID acontrollers,
Further-
survey
controllers,of
In recent
tuned
In to years,
achieve the emergence
optimal performance of fractional
(O’Dwyer calculus has
(2009)). their implementation, and industrial applications. Further-
In recent
made
maderecent years,
possible
years,the
possible
the
the
emergence
thetransition
emergencefrom
transition
of
of fractional
from
classical calculus
fractional
classical
models and
calculus
models
has
has more,
and
their
recent
their implementation,
we focusrelated
results
implementation, and
on particular
to
and industrial
advantages
advantages
industrial applications.
of FOPIDof using
applications. Further-
FOPID
controllers,
Further-
made possible
controllers the
to those thetransition
described
emergenceby from classical
ofdifferential models
models and
classical equations of more,more, we
we focus on
we focus on particular advantages
on particular advantages of
of using FOPID
of using FOPID
In
maderecent
controllers
controllers
years,
possible the
to those
to those transition
described
described
from
by
by
fractional
differential
differential
calculus
equations
equations
has
andof controllers
of
their
controllers
stemming
more,implementation,
focus
stemming
from
particular
from
frequency
and industrial
advantages
frequency
domain
applications.
domain
analysis
using FOPID
analysis
and
Further-
and
noninteger
made possible
controllers order.
to the
those Thus, fractional-order
transition
described from
by dynamic
classical
differential models models
equations andof controllers
also
more, provide
we
controllers stemming
focus relevant
on
stemming from
particular
from frequency
conclusions.advantages domain
Furthermore,
frequencyFurthermore, of analysis
using
domain analysis we
FOPIDand
also
and
noninteger
noninteger order.
order. Thus,
Thus, fractional-order
fractional-order dynamic
dynamic models
models also provide relevant conclusions. we also
and controllers
controllers
noninteger to were
those
order. introduced.
described
Thus, The
by differential
fractional-order parallel form
equations
dynamic models of
of also
discussprovide
controllers
also the
provide relevant
possible
stemming conclusions.
issues
from
relevantissues related
frequency
conclusions. Furthermore,
to industrialization
domain
Furthermore, we
analysis also
and
we also of
and
and controllers
controllers were
were introduced.
introduced. The
The parallel
parallel form
form of
of discuss the possible related to industrialization of
the fractional-order
noninteger order.were
and fractional-order
controllers PID
Thus, controller
fractional-order
introduced. has been proposed
dynamic models in
of also discuss
FOPID the possible
controllers.
provide issues
relevantissues related
conclusions. to industrialization
Furthermore, we also of
the
the fractional-order PID
PID controller The
controller has been
has
parallel
been form
proposed
proposed in
in
discuss
FOPID the possible
controllers. related to industrialization of
Podlubny
and et
controllers al.
the fractional-order (1997)
were introduced. The parallel
PID controller has been proposed in FOPID form of FOPID
discuss controllers.
the possible
controllers. issues related to industrialization of
Podlubny
Podlubny et
et al. (1997) The rest of the paper is organized as follows. In Section 2,
the
Podlubny et al.
al. (1997)
fractional-order U (s)PID controller−λ
(1997) has been µproposed in FOPID The
The rest
rest
performance
of
of the
controllers.
theand
paper
paper is
is organized
organized
implementation
as
as follows.
follows.
comparison
In
In Section
Section
of IOPID
2,
2,
Podlubny Gcet(s)al. U (s)
= (1997)
U (s) = KP + KI s−λ + KD sµ , (1) The rest of the
performance andpaper is organized ascomparison
implementation follows. In of Section
IOPID 2,
G (s) = E(s)
U (s) = K + K s −λ + KD sµ , (1)
(1) performance
and
The FOPID
rest of theand implementation
controllers
paper is is done
organized comparison
based
as on
follows. recent
In of IOPID
publica-
Section 2,
G c
c (s) =
Gc (s) = E(s) E(s) = K P
P + K I
I s
= KP + KI s + KD s , −λ + KD sµ ,
(1) and performance
and FOPID and implementation
controllers is done comparison
based on recent of IOPID
publica-
U (s) tions. FOPID
In
performance Sectioncontrollers
and 3, recent is
implementation done
research based on
related
comparison recent
to the publica-
prospec-
of IOPID
where (λ, Gc µ)
(s) = E(s) = KP + KI s + KD s , −λ
0. Such a controller has more tuning
µ
(1) and tions.
tions.
FOPID
In
In Sectioncontrollers
3, recent is research
done based on recent
related to publica-
the prospec-
where (λ, µ)
> E(s)
> 0. Such a controller has more tuning tive
and
tions.
tive In Section
industrial
FOPID Section
industrial
use3,
use 3,of
of
recent
controllersFOPID
recent
FOPID is research
controllers
done
research
controllers
related
based is
related
is
to
to the
the prospec-
onreviewed.
recent
reviewed.
In Sec-
publica-
prospec-
In Sec-
where
freedom and thus Such
also a a
wider controller
region of has more
parameters tuning
that tive industrial use of FOPID controllers is reviewed. In
In Sec-
(λ, µ) > 0. tion
tions. 4, we
In attempt
Section to cover the issue of comparing FOPID
where (λ,
freedom andµ) thus
> 0.also Such
a a controller
wider region of has more tuning
parameters that tive
tion industrial
4, we use3,ofto
attempt recent
FOPID
cover research
controllers
the issue related
of is to the prospec-
reviewed.
comparing Sec-
FOPID
freedom
stabilize and
the thus
µ)plant also
under a wider
a control, region of
and offers parameters
improvements that tion 4,
where
freedom
stabilize and
(λ,
the thus
> 0.also
plant Such
under a controller
wider
control, region
and ofhas
offers more tuning
parameters
improvements that tive tion 4, we
controller
controller we attempt
industrial use ofto
implementations
attempt
implementations
cover
cover the
toFOPID to
the
to
issue
issue of
general
controllers
general
comparing
of high
is
comparing
high order
FOPID
order integer-
reviewed. In Sec-
FOPID
integer-
stabilize
in controlthe
theloopplant under
undera control,
thusrobustness. and
and offers
Corresponding improvements
studies havethat controller
order implementations
4,controllers. In
to Section to general of high order integer-
freedom
stabilize
in controland plant
loop also
robustness. wider
control, region
Corresponding of parameters
offers improvements
studies have tion
controller
order we attempt
implementations
controllers. In cover the
Section to5,
5,
heuristic
issue
general
heuristichigh tuning
ordermethods
comparing
tuning FOPID
integer-
methods
in control
stabilize
in control loop
theloopplant robustness.
under
robustness. Corresponding
control, and
Corresponding offers studies
improvements
studies have
have order
for FOPID
controllercontrollers. In
controllers
implementations Section
are to 5, heuristic
reviewed.
general Then,
high tuning
in
order methods
Section
integer- 6
1 This study is based upon works from COST Action CA15225, a order
for FOPIDcontrollers. In
controllers Section
are 5, heuristic
reviewed. Then, tuning
in methods
Section 6
1 This study is based upon works from COST Action CA15225, a
in control
network
1 loop
supported
This study is basedrobustness.
by COST Corresponding
(European
upon works from Cooperation
COST Action studies have
in CA15225,
Science and a for
the
order
for FOPID
question
FOPID controllers
controllers.of are
industrialization
In Section
controllers reviewed.of
5, heuristic
are reviewed. Then,
FOPID
Then, in
tuning Section
controllers
methods
in Section 6
is
6
1 This study
network is based
supported upon works
by COST (Europeanfrom Cooperation
COST Action in CA15225,
Science and a the
the question
question of
of industrialization
industrialization of
of FOPID
FOPID controllers
controllers is
is
Technology).
network supported by COST (European Cooperation in Science and addressed,
for
the FOPID
question and of some
controllers related
are
industrialization patents
reviewed.of are
FOPID reviewed.
Then, in In
Section
controllers Sec- 6
is
1 This study is based upon works from COST Action CA15225, a
network supported
Technology). by COST (European Cooperation in Science and addressed,
addressed, and
and some
some related
related patents
patents are
are reviewed.
reviewed. In
In Sec-
Sec-
Technology).
network supported by COST (European Cooperation in Science and
Technology). the question of industrialization of
addressed, and some related patents are reviewed. In Sec- FOPID controllers is
Technology).
Copyright
2405-8963 ©© 2018,
2018 IFAC
IFAC (International Federation of Automatic Control)
addressed,
25 Hosting and some related patents are reviewed. In Sec-
by Elsevier Ltd. All rights reserved.
Copyright © 2018 IFAC 25
Copyright
Peer review©under
2018 responsibility
IFAC 25 Control.
of International Federation of Automatic
Copyright © 2018 IFAC 25
10.1016/j.ifacol.2018.06.014
Copyright © 2018 IFAC 25
IFAC PID 2018
26
Ghent, Belgium, May 9-11, 2018 Aleksei Tepljakov et al. / IFAC PapersOnLine 51-4 (2018) 25–30

tion 7 general items for discussion are provided. Finally, in the cases that these specifications are simultaneously
in Section 8 conclusions are drawn. achievable by an FOPD controller ...” thus showcasing the
tuning flexibility of FOPID controllers compared to IOPID
2. PERFORMANCE AND IMPLEMENTATION controllers.
COMPARISON OF IOPID AND FOPID Let us also consider related critical research. One interest-
CONTROLLERS ing critical work is that of Padula and Visioli (2016) which
uses the term fragility with respect to FOPID controllers
Based on the feedback from industrial partners (Vansovits designed for FOPDT plants. In the paper, the authors
et al. (2012)), the main issue of adopting FOPID con- assume that the parameters of the controller are subject to
trollers in the industry can be formulated as follows: “Can variation and thus devise measures to study this scenario.
the complexity of implementation of FOPID controllers One practical example is when the parameters of a tuned
outweigh the benefits of additional tuning flexibility?”. FOPID controller are changed manually. In this case, hav-
Indeed, in the most basic case, to implement a conven- ing an idea about the fragility of the FOPID controller
tional PID controller digitally one just implements the would be useful. There is also a critical paper published
corresponding control law in software in a Russian journal (Zhmud and Zavorin (2013)) the title
n of which can be translated to English as “On the Inad-
u(k) = Kp e(k) + Ki e(k) + Kd (e(k) − e(k − 1)) , (2) visability of using Fractional-order PID Controllers”. The
j=0 main conclusion of the paper is that by applying a proper
optimization procedure it is possible to obtain IOPID con-
whereas for a FOPID controller one usually has to use trollers that are superior to FOPID controllers obtained
approximations which are often more complicated and re- in Bettou and Charef (2009). However, the premise itself
quire considerable computational resources. Still, modern is based on insufficient evidence, so the conclusion, taken
embedded software solutions have been found to easily more generally, is questionable.
handle the additional implementation complexity (Tepl- The advantages of FOPID controllers can be seen most
jakov (2017)). Thus, in what follows, we focus on the clearly by invoking frequency domain analysis and control
benefits of FOPID controllers with respect to achievable design methods. Seminal works on the subject include
performance improvement. Monje et al. (2008, 2010); Sabatier et al. (2015). Results on
In general, to make a valid comparison between the per- H∞ based design methods were reported by Padula et al.
formance of IOPID and FOPID controllers one can turn (2013). A metric for measuring reference-to-disturbance
to global optimization based methods for tuning both ratio (RDR) was proposed in Alagoz et al. (2015) which
controllers because that way the best possible controller can be used to compare the performance of IOPID and
gains and orders are assumed to be obtained (Beschi et al. FOPID controllers. Robust design for parametric varia-
(2015)). We expand on this issue in Section 5. In Merrikh- tions of control systems can be achieved also through the
Bayat and Jamshidi (2013), IOPID and FOPID controllers assessment of maximum sensitivity properties thereof. A
are designed for the control of a nonlinear boost converter related study is conducted in Ranganayakulu et al. (2016).
using an artificial bee colony algorithm. Based on sim- A clear theoretical demonstration of the advantages of
ulation results, the authors conclude that the “proposed fractional-order control was provided by Vinagre et al.
FOPID controller can improve the startup response of (2003) by considering Bode’s ideal transfer function. The
the boost converter by using less on-off switching actions system under study is composed of a fractional controller
compared to the optimal PID controller” and stress the in the form C(s) = sγ and double integrator plant function
practical importance of this result. This essentially means in the form of G(s) = A/s2 . The plant function is used to
an improvement in the control law. Moreover, better rejec- model a fundamental system that represents single-degree-
tion of disturbances and better output voltage regulation of-freedom translational and rotational motion in the field
are also cited as advantages. Further, authors of the con- of robotics. The corresponding closed-loop system is in the
ference paper de Castro et al. (2016) applied both IOPID form of Bode’s ideal transfer function GCL (s) = A/(s2−γ +
and FOPID controllers to a liquid level control problem A). The main advantages of this closed loop system are:
while tuning both using genetic programming. The results
show that “... the PIλ Dµ has performed slightly better for (1) Gain margin is infinite. This property provides the
the response signal ...”. However, one remarkable result is advantage of the system being insensitive to gain
that a more desirable control law is obtained in this case changes;
as well by measuring the variance of u(t). For IOPID it (2) The phase margin is constant at ϕm = π(1 − (2 −
measures 3.44, while for FOPID the value obtained is 1.39. r)/2). The variations of the gain result in the change
Furthermore, the same conclusion regarding the reduction of the crossover frequency ωc , but the phase margin
of the control effort is reached in Marinangeli et al. (2018). of the system is unchanged.
This result can be considered very important in cases
The desired robustness motivated the use of fractional-
where reducing energy due to control effort matters, like
order controllers in classical control systems to enhance
process systems and precision positioning systems where
their performance (Oustaloup (1995); Chen et al. (2009);
the generation of heat reduces precision.
Pommier-Budinger et al. (2008); Banos et al. (2008)).
In Badri and Tavazoei (2015), the design of a FOPD A comparitive introduction of CRONE and TID type
motion controller was investigated. One important item in controllers was provided in Xue and Chen (2002). Chen et
the conclusions reads “... the inefficiency of PID controllers al. demonstrated that several fractional-order controllers
for simultaneously ensuring specifications [was shown] ...

26
 IFAC PID 2018
Ghent, Belgium, May 9-11, 2018 Aleksei Tepljakov et al. / IFAC PapersOnLine 51-4 (2018) 25–30 27

can be tuned to meet the criterion of robustness to gain In Lino et al. (2017), FOPI controllers were designed
variations which is given by for industrial electrical drives. The papers deals with a
  particular design method. To cite the authors directly, “It
d arg (C(jω)P (jω))
= 0. (3) is remarkable that the fractional pre-filters almost cancel
dω ω=ωc the oscillations. The improvement is even more relevant
This property is known as iso-damping and has been in the case of speed control. The overshoot is greatly
frequently cited as one of the most attractive features reduced and the settling and rise times are also reduced
of FO control (Monje et al. (2010)) and forms in part with respect to the PI-controlled system.”
the basis of the so called fractal robustness (Oustaloup Finally, in Krijnen et al. (2017), a precision positioning
(1995)). It was further investigated in, e.g., Pommier- system based of FO control is designed and analyzed. Two
Budinger et al. (2008); Feliu-Batlle et al. (2009); Sabatier cascaded control loops with decoupled SISO controllers are
et al. (2015); Beschi et al. (2015); Azarmi et al. (2016); implemented for the moving mass controlled on a mass-
Bettayeb et al. (2016). spring system which can be modeled as a fourth order
system. Applying fractional order derivative compared to
3. INDUSTRIAL APPLICATIONS OF FOPID integer order one, the bandwidths are extended by 14.6%
CONTROL and 62%, for the inner and outer loop respectively. A
closed-loop positioning bandwidth of the wafer of 60Hz is
With respect to industrial use of FOPID controllers, a achieved, resulting in a positioning error of 104nm, which
survey paper is immediately found (Efe (2011)). The main is limited by sensor noise and pressure disturbances.
content of the paper comprises the description of various
classical control techniques (PID control, sliding mode
control, backstepping, MRAC) extended to make use of
fractional-order calculus. There is a single example of 4. FOPID CONTROLLERS VERSUS HIGH ORDER
a control problem provided where a control system is INTEGER ORDER CONTROLLERS
developed for a cement mill.
For real-life implementation of FOPID controllers high
In Ghasemi et al. (2014), a fractional controller is designed order IO transfer function approximations are generally
for a wind turbine generator. Here, based on simulations, used (Monje et al. (2010)). This brings about the question:
the authors claim the following: (1) “... while the fractional “If integer-order approximations are used anyway, why
order PI controller ... properly tracks the input command, not just use high order integer-order controllers instead of
the simple integer order PI controller is not capable to cope FOPID approximations?” Although this is a very impor-
with the nonlinearity due to backlash phenomenon.” and tant issue from the modeling standpoint, surprisingly few
(2) “... the fractional order control system accurately tracks publications deal with this matter. A few (almost) relevant
the reference input [under plant parameter variation] ... papers are reviewed below.
however, the integer order control system becomes unsta-
ble [under the same conditions]” once again highlighting In Merrikh-Bayat (2012), the author proposes some rules
advantages of FOPID controllers. for selecting the parameters for approximating FOPID
controllers using the Oustaloup recursive filter method.
In Jiangbo and Junzheng (2015), a FOPI controller is This can potentially help to reduce the order of the
developed for an electro-hydraulic system with a partic- resulting controller. However, this does not solve the
ular emphasis on the energy saving aspect. To validate “FOPID versus high order IO controller” issue that is the
the performance of the obtained control system, a PI topic of this section.
controller is designed and experiments with both PI and
FOPI controllers are conducted (including variants thereof In one instance, the authors of Mansouri et al. (2010)
where an orifice compensation (OC) system is added). suggest to use FO models for “compressing” high order
The authors claim that the FOPI+OC method “has the integer-order models. The proposed method has certain
minimum tracking error, and common PI controller has limitations (e.g., it works only for real transfer function
the biggest one.” poles). In Barbé et al. (2013) this approach is extended.
Here, the authors use fractional-order models for creat-
The authors of Azarmi et al. (2015) design FOPID con- ing compact models obtained by (1) identifying a high
trollers for twin rotor systems. Here, the authors evidently order integer-order model; (2) converting the model to
show that “... disturbance rejection by applying the [set- a fractional-order one; (3) optimizing the fractional-order
point weighted FOPID] controller is always better done model. Both of these contributions propose the basic idea
in comparison with the designed [setpoint weighted PID], that FO models can be used for compact description of
[filtered FOPI] and [filtered PI] controllers ...”. high order systems but offer little with respect to the FO
In Hassan and Zolotas (2017) the impact of FO methods control discussion.
on tilt control of rail vehicles is discussed. Here, both the Clearly, more relevant research devoted to this issue is
design of FOPID controllers, and the reduction of corre- required. As a preliminary conclusion, we consider the
sponding approximations are investigated. The results of following:
FOPID control are compared to that achieved with con-
ventional PID controllers. To cite the authors, “[Relevant (1) FOPID controllers are not generally speaking equiv-
figure] illustrates the immense benefit of fractional order alent to high order integer-order controllers, and we
based control on improving tilt following (with full order also assume that an ideal implementation of a FOPID
control).” controller is available.

27
IFAC PID 2018
28
Ghent, Belgium, May 9-11, 2018 Aleksei Tepljakov et al. / IFAC PapersOnLine 51-4 (2018) 25–30

(2) We focus on the fact that a FOPID controller is an industrial applications (Li et al. (2016); Aghababa (2015);
extension to an IOPID controller and treat its two Wang et al. (2017); Li et al. (2017a)).
additional parameters as “tuning knobs”.
As results of random search nature of heuristic optimiza-
It is also apparent that managing (e.g., tuning) all of the tion methods, the objectivity of assessments based on a
parameters of a high order integer-order controller is more single test is somewhat questionable because of unrepeata-
difficult than in case of a FOPID controller. bility and possibility of being stuck at a local minimum.
To make findings of these studies more solid, a statistical
5. HEURISTIC OPTIMIZATION METHODS IN THE evaluation of results is necessary. Hence, in comparison
TUNING OF FOPID CONTROLLERS of controller, consideration of an average value of perfor-
mance indices and revealing standard deviation of results
for repeated tests is important in publications to increase
Heuristic optimization techniques are commonly used for objectivity and consistency of findings of heuristic opti-
optimization of high complexity problems and perform mization algorithms.
set-trail search to find out an optimal solution. This
property makes it a straightforward solution even for Consequently, heuristic optimization algorithms are low
online tuning of parameters of real systems. Since industry computational complexity tools for multi-objective and
looks for shortcut techniques to fix the problems, heuristic constrained optimization, that is, it is relatively easy
optimization techniques can provide a low-cost solution to implement these algorithms. As a future perspective,
for FOPID tuning of real time control systems. Control heuristic optimization can be utilized in the realization
literature has extensively demonstrated utilization and of plug&play type FOPID controller devices implemented
advantage of heuristic optimization techniques (Tepljakov on low-cost programmable control cards for industrial
et al. (2012); Ateş and Yeroglu (2016)). Many recent works applications.
have addressed several heuristic optimization methods for
FOPID controller tuning in an offline (simulation based) 6. INDUSTRIALIZATION OF FOPID CONTROLLERS
and online (experimental based) manners. Some studies,
presenting with its application, are as follows. Industrialization of controllers is much of a challenge nowa-
days in a competitive world where companies tend to
Zamani et al. (2017) used a multi-objective cuckoo search invest resources in order to bring down the cost of goods
optimization approach for FOPID tuning to adjust the and improve the products’ benefits to the customers. As
contact force of piezoelectric friction dampers for semi- it was demonstrated, FOPID controllers have technical
active control of base-isolated structures during far- advantages over its integer-order counterparts, but the cost
field and near-field earthquake excitations. In Li et al. of producing such controllers and the cost-benefit the end
(2017b), a gravitational search algorithm combined with user would get are still something that should be investi-
the Cauchy and Gaussian mutation, named as CGGSA, gated. Where a FOPID controller shows better technical
is proposed and used to optimize the FOPID controller performance in a twin rotor system in Azarmi et al. (2015),
parameters to control a pump turbine governing system it does not necessarily mean that the cost of creating
(PTGS). From the perspective of multi-objective optimiza- such FOPID controllers for commercial or industrial use is
tion, Zeng et al. (2015) proposed a novel FOPID controller similar to ordinary PID controllers and its families. In such
design method based on an improved multi-objective ex- a case, it is suggested to look how to industrialized FOPID
ternal optimization (MOEO) algorithm for an automatic controllers considering financial factors. This is an open
voltage regulator (AVR) system. The effectiveness of the question that needs to be addressed in future research.
proposed MOEO-FOPID algorithm on AVR system was
also demonstrated. In Xu et al. (2016), gravitational search Patents can be seen as important factors in the industri-
algorithm was improved for FOPID design to control alization process. To conclude this section, we investigate
pumped storage hydro unit. The results are compared with patents related to fractional-order control and implemen-
PSO (particle swarm optimization), GSA (gravitational tation thereof.
search algorithm) and BCGSA (bacterial-foraging chemo- The TID controller patent (Lurie (1994)) describes a tilt-
taxis gravitational search algorithm). In Haji and Monje integral-derivative controller. It is similar to a PID con-
(2017), further enhancement of the PSO algorithm’s rate troller, but the proportional gain is replaced with the tilt
of convergence and the minimization of the fitness function component that has a transfer function of s−1/n . The
was achieved based on dynamic control parameters selec-
advantage of this controller was explained in Chen et al.
tion. The method was applied to fractional order fuzzy-
(2009) as “The resulting transfer function of the entire
PID control of a combined cycle power plant. The results
compensator more closely approximates an optimal loop
are compared with several different evolutionary algo-
transfer function, thereby achieving improved feedback
rithms. In Yeroğlu and Ateş (2014), an online auto-tuning
control performance. Further, as compared to conven-
of FOPID controllers on a real time running experimen-
tional PID compensators, the TID compensator allows for
tal test system (Twin Rotor MIMO System-TRMS) was
simpler tuning, better disturbance rejection, and smaller
performed using Stochastic Multi-parameter Divergence
effects of plant parameter variations on the closed-loop
Optimization (SMDO) algorithm. In Alagoz et al. (2013),
response”.
SMDO algorithm was used for tuning FOPID parameters
to track the response of the Bode’s ideal control loop. The Apart from that, only a few US-based patents can be
results are used to control DC Rotor of an experimental found. In Abbisso et al. (2004), an invention implementing
test system (TRMS). Many other heuristic optimization noninteger (i.e., fractional) systems is described and is
algorithms were proposed for FOPID tuning for different based on artificial neural networks. In Chen (2009), a tun-

28
 IFAC PID 2018
Ghent, Belgium, May 9-11, 2018 Aleksei Tepljakov et al. / IFAC PapersOnLine 51-4 (2018) 25–30 29

ing method for fractional controllers is proposed. Finally, Åström, K. and Hägglund, T. (2006). Advanced PID control. The
in Almadhoun et al. (2016) a fractional-order capacitor Instrumentation, Systems, and Automation Society (ISA).
is described. This latter patent could be very important Ateş, A. and Yeroglu, C. (2016). Optimal fractional order PID design
in the sense of achieving an ideal implementation of FO via tabu search based algorithm. ISA Transactions, 60, 109–118.
doi:10.1016/j.isatra.2015.11.015.
operators, and, as a direct consequence, also of FOPID-
Azarmi, R., Tavakoli-Kakhki, M., Sedigh, A.K., and Fatehi, A.
type controllers. (2015). Analytical design of fractional order PID controllers
based on the fractional set-point weighted structure: Case study
7. DISCUSSION in twin rotor helicopter. Mechatronics, 31, 222–233. doi:
10.1016/j.mechatronics.2015.08.008.
Due to the lack of coherent research results related to Azarmi, R., Tavakoli-Kakhki, M., Sedigh, A.K., and Fatehi, A.
comparing high order IO approximations of FOPID con- (2016). Robust fractional order PI controller tuning based on
trollers and high order IO controllers, this item is left for bode’s ideal transfer function. IFAC-PapersOnLine, 49(9), 158–
163. doi:10.1016/j.ifacol.2016.07.519.
discussion. In addition, any feedback related to industrial
Badri, V. and Tavazoei, M.S. (2015). Achievable performance
adoption of FOPID controllers and corresponding issues is region for a fractional-order proportional and derivative motion
welcome. controller. IEEE Transactions on Industrial Electronics, 62(11),
7171–7180. doi:10.1109/TIE.2015.2448691.
8. CONCLUSIONS Banos, A., Cervera, J., Lanusse, P., and Sabatier, J. (2008). Bode
optimal loop shaping with CRONE compensators. In MELECON
Based on the evidence reviewed in the present paper, the 2008 - The 14th IEEE Mediterranean Electrotechnical Confer-
following conclusions can be drawn: ence. IEEE. doi:10.1109/melcon.2008.4618409.
Barbé, K., Rodriguez, O.J.O., Moer, W.V., and Lauwers, L. (2013).
• FOPID controllers offer clear advantages over IOPID Fractional models for modeling complex linear systems under poor
controllers as seen from both simulations and experi- frequency resolution measurements. Digital Signal Processing,
ments with real-life objects, the comparison is solidly 23(4), 1084–1093. doi:10.1016/j.dsp.2013.01.009.
Beschi, M., Padula, F., and Visioli, A. (2015). The generalised
based on global optimization based tuning for both
isodamping approach for robust fractional PID controllers de-
types of controllers. sign. International Journal of Control, 90(6), 1157–1164. doi:
• One significant advantage of FOPID controllers when 10.1080/00207179.2015.1099076.
applied to industrial problems is the potential re- Bettayeb, M., Mansouri, R., Al-Saggaf, U., and Mehedi, I.M. (2016).
duction of the control effort which also results in Smith predictor based fractional-order-filter PID controllers de-
reduction of wasted energy. sign for long time delay systems. Asian Journal of Control, 19(2),
• The most common design method for fractional-order 587–598. doi:10.1002/asjc.1385.
controllers is based on frequency domain analysis. Bettou, K. and Charef, A. (2009). Control quality enhancement using
The properties of Bode’s ideal transfer function and fractional PIλ dµ controller. International Journal of Systems
Science, 40(8), 875–888. doi:10.1080/00207720902974546.
the iso-damping property are essential characteristics
Chen, Y.Q. (2006). Ubiquitous fractional order controls? IFAC
that define “fractal” robustness and showcase the Proceedings Volumes, 39(11), 481–492. doi:10.3182/20060719-3-
advantage of fractional-order controls. pt-4902.00081.
• Heuristic FOPID tuning methods can be seen as at- Chen, Y.Q. (2009). Tuning methods for fractional-order controllers.
tractive due to the relative simplicity of implementing Chen, Y., Petráš, I., and Xue, D. (2009). Fractional order control
the corresponding algorithms, but do suffer from is- - a tutorial. In Proc. ACC ’09. American Control Conference,
sues related to the assessment of results. 1397–1411. doi:10.1109/ACC.2009.5160719.
• An ideal, “basic building block” implementation of FO de Castro, F.A., Bernardes, N.D., de S. L. Cuadros, M.A., and
operators should be sought to compensate for the high de Almeida, G.M. (2016). Comparison of fractional and integer
PID controllers tuned by genetic algorithm. In 2016 12th IEEE In-
order approximation issue. This implementation must
ternational Conference on Industry Applications (INDUSCON).
be cost effective to facilitate industrial adoption of FO IEEE. doi:10.1109/INDUSCON.2016.7874592.
controllers. Efe, M.O. (2011). Fractional order systems in industrial
These items form the basis for future research. automation—a survey. IEEE Transactions on Industrial Infor-
matics, 7(4), 582–591. doi:10.1109/TII.2011.2166775.
Feliu-Batlle, V., Pérez, R.R., García, F.C., and Rodriguez, L.S.
REFERENCES (2009). Smith predictor based robust fractional order con-
trol: Application to water distribution in a main irrigation
Abbisso, S., Caponetto, R., Diamante, O., Porto, D., Di Cola, E., canal pool. Journal of Process Control, 19(3), 506–519. doi:
and Fortuna, L. (2004). Non-integer order dynamic systems. 10.1016/j.jprocont.2008.05.004.
Aghababa, M.P. (2015). Optimal design of fractional-order PID Ghasemi, S., Tabesh, A., and Askari-Marnani, J. (2014). Application
controller for five bar linkage robot using a new particle swarm of fractional calculus theory to robust controller design for wind
optimization algorithm. Soft Computing, 20(10), 4055–4067. doi: turbine generators. IEEE Transactions on Energy Conversion,
10.1007/s00500-015-1741-2. 29(3), 780–787. doi:10.1109/TEC.2014.2321792.
Alagoz, B.B., Ates, A., and Yeroglu, C. (2013). Auto-tuning of PID Haji, V.H. and Monje, C.A. (2017). Fractional order fuzzy-
controller according to fractional-order reference model approxi- PID control of a combined cycle power plant using particle
mation for DC rotor control. Mechatronics, 23(7), 789–797. doi: swarm optimization algorithm with an improved dynamic pa-
10.1016/j.mechatronics.2013.05.001. rameters selection. Applied Soft Computing, 58, 256–264. doi:
Alagoz, B.B., Deniz, F.N., Keles, C., and Tan, N. (2015). Disturbance 10.1016/j.asoc.2017.04.033.
rejection performance analyses of closed loop control systems by Hassan, F. and Zolotas, A. (2017). Impact of fractional order
reference to disturbance ratio. ISA Transactions, 55, 63–71. doi: methods on optimized tilt control for rail vehicles. Fractional
10.1016/j.isatra.2014.09.013. Calculus and Applied Analysis, 20(3). doi:10.1515/fca-2017-0039.
Almadhoun, M., Elshurafa, A., Salama, K., and Alshareef, H. (2016).
Fractional order capacitor.

29
IFAC PID 2018
30
Ghent, Belgium, May 9-11, 2018 Aleksei Tepljakov et al. / IFAC PapersOnLine 51-4 (2018) 25–30

Jiangbo, Z. and Junzheng, W. (2015). The fractional order PI control Podlubny, I., Dorčák, L., and Kostial, I. (1997). On fractional deriva-
for an energy saving electro-hydraulic system. Transactions of tives, fractional-order dynamic systems and PIλ Dµ -controllers. In
the Institute of Measurement and Control, 39(4), 505–519. doi: Proc. 36th IEEE Conf. Decision and Control, volume 5, 4985–
10.1177/0142331215610184. 4990. doi:10.1109/CDC.1997.649841.
Krijnen, M.E., van Ostayen, R.A., and HosseinNia, H. (2017). Pommier-Budinger, V., Janat, Y., Nelson-Gruel, D., Lanusse, P.,
The application of fractional order control for an air-based and Oustaloup, A. (2008). Fractional robust control with iso-
contactless actuation system. ISA Transactions. doi: damping property. In 2008 American Control Conference. IEEE.
10.1016/j.isatra.2017.04.014. In press. doi:10.1109/acc.2008.4587279.
Li, C., Mao, Y., Zhou, J., Zhang, N., and An, X. (2017a). Design of Ranganayakulu, R., Babu, G.U.B., Rao, A.S., and Patle, D.S. (2016).
a fuzzy-PID controller for a nonlinear hydraulic turbine governing A comparative study of fractional order PIλ /PIλ dµ tuning rules
system by using a novel gravitational search algorithm based on for stable first order plus time delay processes. Resource-Efficient
cauchy mutation and mass weighting. Applied Soft Computing, Technologies, 2, S136–S152. doi:10.1016/j.reffit.2016.11.009.
52, 290–305. doi:10.1016/j.asoc.2016.10.035. Sabatier, J., Lanusse, P., Melchior, P., and Oustaloup, A. (2015).
Li, C., Zhang, N., Lai, X., Zhou, J., and Xu, Y. (2017b). Design Fractional Order Differentiation and Robust Control Design.
of a fractional-order PID controller for a pumped storage unit Springer Netherlands. doi:10.1007/978-94-017-9807-5.
using a gravitational search algorithm based on the cauchy and Tavakoli-Kakhki, M. and Haeri, M. (2011). Fractional order
gaussian mutation. Information Sciences, 396, 162–181. doi: model reduction approach based on retention of the domi-
10.1016/j.ins.2017.02.026. nant dynamics: Application in IMC based tuning of FOPI and
Li, X., Wang, Y., Li, N., Han, M., Tang, Y., and Liu, F. (2016). FOPID controllers. ISA Transactions, 50(3), 432–442. doi:
Optimal fractional order PID controller design for automatic 10.1016/j.isatra.2011.02.002.
voltage regulator system based on reference model using particle Tepljakov, A. (2017). Fractional-order modeling and control of
swarm optimization. International Journal of Machine Learning dynamic systems. Springer-Verlag GmbH. doi:10.1007/978-3-319-
and Cybernetics, 8(5), 1595–1605. doi:10.1007/s13042-016-0530-2. 52950-9.
Lino, P., Maione, G., Stasi, S., Padula, F., and Visioli, A. (2017). Tepljakov, A., Petlenkov, E., and Belikov, J. (2012). A flexible
Synthesis of fractional-order PI controllers and fractional-order MATLAB tool for optimal fractional-order PID controller design
filters for industrial electrical drives. IEEE/CAA Journal of subject to specifications. In Proceedings of the 31st Chinese
Automatica Sinica, 4(1), 58–69. doi:10.1109/jas.2017.7510325. Control Conference, 4698–4703.
Luo, Y. and Chen, Y. (2012). Stabilizing and robust fractional order Vansovits, V., Petlenkov, E., Vassiljeva, K., and Guljajev, A. (2012).
PI controller synthesis for first order plus time delay systems. Identification of industrial water boiler for model predictive con-
Automatica, 48(9), 2159–2167. trol of district heat plant. In Proceedings of the 13th biennial
Lurie, B.J. (1994). Three-parameter tunable tilt-integral-derivative Baltic Electronics Conference, 314–318.
(TID) controller. Čech, M. and Schlegel, M. (2006). The fractional-order PID con-
Mansouri, R., Bettayeb, M., and Djennoune, S. (2010). Approxima- troller outperforms the classical one. In Process control 2006,
tion of high order integer systems by fractional order reduced- 1–6. Pardubice Technical University.
parameters models. Mathematical and Computer Modelling, Vinagre, B.M., Chen, Y.Q., and Petráš, I. (2003). Two direct
51(1), 53–62. doi:10.1016/j.mcm.2009.07.018. Tustin discretization methods for fractional-order differentia-
Marinangeli, L., Alijani, F., and HosseinNia, S.H. (2018). Fractional- tor/integrator. Journal of the Franklin Institute, 340(5), 349–362.
order positive position feedback compensator for active vibration Wang, H., Zeng, G., Dai, Y., Bi, D., Sun, J., and Xie, X. (2017). De-
control of a smart composite plate. Journal of Sound and Vibra- sign of a fractional order frequency PID controller for an islanded
tion, 412(Supplement C), 1 – 16. doi:10.1016/j.jsv.2017.09.009. In microgrid: A multi-objective extremal optimization method. En-
press. ergies, 10(10), 1502. doi:10.3390/en10101502.
Merrikh-Bayat, F. and Jamshidi, A. (2013). Comparing the Per- Xu, Y., Zhou, J., Xue, X., Fu, W., Zhu, W., and Li, C. (2016).
formance of Optimal PID and Optimal Fractional-Order PID An adaptively fast fuzzy fractional order PID control for pumped
Controllers Applied to the Nonlinear Boost Converter. ArXiv storage hydro unit using improved gravitational search algo-
e-prints. rithm. Energy Conversion and Management, 111, 67–78. doi:
Merrikh-Bayat, F. (2012). Rules for selecting the parameters of 10.1016/j.enconman.2015.12.049.
Oustaloup recursive approximation for the simulation of linear Xue, D. and Chen, Y. (2002). A comparative introduction of four
feedback systems containing PIλ dµ controller. Communications fractional order controllers. In Proceedings of the 4th World
in Nonlinear Science and Numerical Simulation, 17(4), 1852– Congress on Intelligent Control and Automation, 3228–3235.
1861. doi:10.1016/j.cnsns.2011.08.042. Xue, D., Zhao, C., and Chen, Y. (2006). Fractional order PID control
Monje, C.A., Vinagre, B.M., Feliu, V., and Chen, Y.Q. (2008). of a DC-motor with elastic shaft: a case study. In 2006 American
Tuning and auto-tuning of fractional order controllers for industry Control Conference, 3182–3187. doi:10.1109/ACC.2006.1657207.
applications. Control Engineering Practice, 16(7), 798–812. Yeroğlu, C. and Ateş, A. (2014). A stochastic multi-parameters
Monje, C.A., Chen, Y.Q., Vinagre, B.M., Xue, D., and Feliu, V. divergence method for online auto-tuning of fractional order PID
(2010). Fractional-order Systems and Controls: Fundamentals controllers. Journal of the Franklin Institute, 351(5), 2411–2429.
and Applications. Advances in Industrial Control. Springer Verlag. doi:10.1016/j.jfranklin.2013.12.006.
O’Dwyer, A. (2009). Handbook of PI and PID Controller Tunning Zamani, A.A., Tavakoli, S., and Etedali, S. (2017). Fractional
Rules. Imperial College Press, 3 edition. order PID control design for semi-active control of smart base-
Oustaloup, A. (1995). La dérivation non entière: Théorie, synthèse isolated structures: A multi-objective cuckoo search approach.
et applications. Hermes Science Publications. ISA Transactions, 67, 222–232. doi:10.1016/j.isatra.2017.01.012.
Padula, F., Vilanova, R., and Visioli, A. (2013). H∞ optimization- Zeng, G.Q., Chen, J., Dai, Y.X., Li, L.M., Zheng, C.W., and
based fractional-order PID controllers design. International Jour- Chen, M.R. (2015). Design of fractional order PID controller
nal of Robust and Nonlinear Control, 24(17), 3009–3026. doi: for automatic regulator voltage system based on multi-objective
10.1002/rnc.3041. extremal optimization. Neurocomputing, 160, 173–184. doi:
Padula, F. and Visioli, A. (2013). Set-point weight tuning rules for 10.1016/j.neucom.2015.02.051.
fractional-order PID controllers. Asian Journal of Control, 15(3), Zhmud, V. and Zavorin, A. (2013). O necelesoobraznosti primenenija
678–690. doi:10.1002/asjc.634. drobno-stepennyh PID-reguljatorov. Avtomatika i programmnaja
Padula, F. and Visioli, A. (2016). On the fragility of fractional-order inzhenerija, 4(2), 7–21.
PID controllers for FOPDT processes. ISA Transactions, 60, 228–
243. doi:10.1016/j.isatra.2015.11.010.

30