Proceedings of the 3rd IFAC Conference on

Proceedings of the 3rd IFAC Conference on

Advances in Proportional-Integral-Derivative
Proceedings of the 3rd IFAC Conference on Control
Advances
Ghent, Belgium, May 9-11, 2018
Available
in Proportional-Integral-Derivative online at www.sciencedirect.com
Control
Proceedings
Advances in of the 3rd IFAC Conference on Control
Proportional-Integral-Derivative
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) 604–609
PID
PID controller
controller design
design for
for load
load frequency
frequency
PID controller
control: Past, design
Present for load frequency
PID controller
control: designand forfuture challenges
load frequency
control: Past, Present and future challenges
Past, Present and future challenges
control: Past, Present ∗and future ∗∗ challenges
Yogesh V. Hote ∗∗ Shivam Jain ∗∗
∗∗
Yogesh V. Hote ∗∗ Shivam Jain ∗∗
Yogesh V. Hote Shivam Jain ∗∗
∗
∗ Yogesh V. HoteRoorkee, Shivam Jain ∗∗ 247667 India
∗ Indian
∗
Indian Institute of Technology, Uttarakhand
Uttarakhand 247667 India
∗ Institute
∗ Indian Institute of(e-mail: of Technology,
Technology, Roorkee,
yhotefee@iitr.ac.in).
Roorkee, Uttarakhand 247667 India
∗
∗∗ Indian Institute (e-mail:
of(e-mail:
Technology, yhotefee@iitr.ac.in).
Roorkee, Uttarakhand 247667 India
∗∗ Indian
∗∗
Indian Institute
Institute of Technology,
of(e-mail:
Technology, Roorkee,
Roorkee, Uttarakhand
yhotefee@iitr.ac.in). Uttarakhand 247667 247667 India India
∗∗
∗∗ Indian Institute(e-mail: of Technology, yhotefee@iitr.ac.in).
shivjain1994@gmail.com)
Roorkee, Uttarakhand 247667 India
∗∗
Indian Institute(e-mail: (e-mail: shivjain1994@gmail.com)
of Technology, Roorkee, Uttarakhand 247667 India
shivjain1994@gmail.com)
(e-mail: shivjain1994@gmail.com)
Abstract: In this paper, a brief literature review for the design of a PID controller for load
Abstract:
frequency control
Abstract: In this
In this (LFC)paper,
paper, in a brief
power
a brief literature
systemsreview
literature for
for the
is presented.
review the design of
of aa PID
The transfer
design PID controller
function
controller for
for load
models for
load
frequency
Abstract:
various
frequency control
In this (LFC)
configurations
control (LFC)
paper, of in
power
in power
a brief
powersystems systems
literature
systemsare is
review presented.
developed.
is for the
presented. For The
design
a
The transfer
of a PID
systematic
transfer function
controller
presentation
function models
forofload
models for
the
for
various configurations
frequency control (LFC) of power
in power systems systemsare developed.
is into
presented. ForThe a systematic
transfer presentation
function models of thefor
review,
various
review, the
the PID
configurations
PID tuning
tuning ofschemes
power
schemes are
systems
are categorised
are developed.
categorised into soft
soft computing
For techniques,
a systematic
computing techniques, robust
presentation
robust control
of the
control
various
schemes,the
review, configurations
fractional
PID tuning order ofschemes
powerPID
based systems
are design are
categorisedand developed.
internal
into softmodel For acontrol
computing systematic(IMC)
techniques, presentation
based of the
PIDcontrol
robust design
schemes,
review,
approaches.
schemes, thefractional
PID
Due tuning
fractional order
diligence based
orderschemes has been
based PID
PID design
are taken
design toand
categorisedand internal
into all
include
internal model
softthecomputing
model control
design
control (IMC)
techniques,
schemes. based
(IMC)Further,
based PID
robust design
thecontrol
PID paper
design
approaches.
schemes,
also enlists
approaches. Due
fractional
various
Due diligence
order
future
diligence has
based been
challenges
has been taken
PID takendesign
that toand
are
to include
internal
still
include all the
model
unresolved
all the design
and
design schemes.
control
can (IMC)
form
schemes. Further,
based
the basis
Further, the
PIDof
the paper
design
future
paper
also enlists various
approaches. Due diligencefuture challenges
has been taken that areinclude
still unresolved and can form the basisthe of future
research
also enlistswork.various future challenges that to all the design
are still unresolved and canschemes.
form Further,
the basis of future paper
research
also enlists
research work.
work.various future challenges that are still unresolved and can form the basis of future
© 2018, IFAC
research
Keywords: work.
PID (International Federation ofrejection,
control, Disturbance AutomaticFractional
Control) Hosting orderby Elsevier
PID, Robust Ltd. control,
All rights reserved.
Keywords:
Mathematical
Keywords: PID
PIDmodelscontrol, Disturbance rejection, Fractional
control, Disturbance rejection, Fractional order PID, Robust control, order PID, Robust control,
Mathematical
Keywords: PIDmodels
Mathematical control, Disturbance rejection, Fractional order PID, Robust control,
models
Mathematical
1. models
INTRODUCTION complex
1. INTRODUCTION complex and and encounter
encounter problems
problems of of robustness
robustness in in an
an un-
un-
1. INTRODUCTION certain environment.
complex and encounter problems of robustness in an un-
1. theINTRODUCTION certain environment.
complexenvironment.
and encounter problems of robustness in an un-
We know that both active and reactive certain
We know that both the active and reactive powerpower require-
require- certain The
The widespread
environment.
widespread use
use ofof PID
PID and and thethe drawbacks
drawbacks of of optimal
optimal
ments
We
mentsknow of
of a
a power
that
powerbothsystem
the active
system are
are never
never steady
and reactive
steady since
power
since the
the con-
require-
con- control
The techniques
widespread usemade
of the and
PID researchers
the feel thatofthere
drawbacks was
optimal
We
mentsknow
sumer of load that
a power both the
and industrial
system areactiveload and
never reactive
are continuously power
steady since the require-
varying.
con- control control
The
a need techniques
widespread
to combine usemade
of PID
the the researchers
and
simplicity the feel
of drawbacks
PID that
controllers there
ofthere was
optimal
with
sumer load and industrial load are continuously varying.
con- acontrol need techniques made the researchers feel that was
ments of
Consequently,
sumer load
Consequently,
a power input
andinput systemsupply,
industrial
supply,
are
load never
i.e.,
aresteamsteadyinput
steam
continuouslysincetothe turbo-
varying. need to
aoptimal combine
techniques
totuning
combine the
the simplicity
made
approaches. It wasof
the researchers
simplicity PID feelcontrollers
of observed
PID thatthat
controllers there with
suchwasa
with
sumer
generatorsload or
Consequently, andwater
industrial
input input
supply, toi.e.,
load aresteam
hydro
i.e., input
input to
continuously
generators turbo-
varying.
tomust
turbo-be optimala need
controller totuning
combine
yields approaches.
the
better It
simplicity
results inwas of
case observed
PID
of that
controllers
parametric such
with
uncer- a
generators
Consequently,
regulated or water
properly,inputelse input
supply,
there toi.e.,
hydro
can steam
be generators
inputintomust
variation turbo-
machine be controller yields better results in case of parametric uncer-a
optimal tuning approaches. It was observed that such
generators or
regulated properly, water elseinput
there to hydro
can generators
be variation must
in machine be optimal tuning
tainties, disturbance approaches.
rejection It was
and observed
alsoofin that such
the case ofuncer- non-a
generators
speed andproperly,
regulated or water
hence a else input
changethere to
incanhydro
frequency
be generators
whichin ismust
variation highly
machine be controller
tainties,
controller
yields better
disturbance
yields
minimumdisturbance better
phase behaviour.
results in
rejection
results and
in
In and
case
casealsoof
this paper,
parametric
in the
parametriccase of non-
uncer-
speed andproperly,
hence highly tainties, rejection also in aathe brief
case review of
of non-
regulated
undesirable
speed and hencein the a change
else
power
apower
changethere incan
system frequency
be variation
operation.
in frequency
which
which in is highly minimum
ismachine
Theoretically, tainties,
various PID
minimum phase
disturbance
based
phase behaviour.
rejection
control
behaviour. In
In and
schemesthis
this paper,
also in athe
is presented.
paper, brief
case
brief review
The of
of main
reviewnon-
of
undesirable
speed
one and hencein the apower
change system operation.
in frequency Theoretically,
which is highly various PID based control schemes is various
presented. The main
one can
can make
undesirable
undesirable makein the
thethechange
change
inpossible.
thechange
power
in
in frequency
system
frequency
system
as
as zero,
operation.
operation.
but,
but, in in prac-
zero,Theoretically,
Theoretically, prac- various minimum
purpose
purposePID of phase
of thisbased
this
behaviour.
paper
paper is
is to
control In this paper,
to summarize
schemes
summarize is various a brief
presented. review
controller
The
controller
of
de-
main
de-
tice,
one
tice, it
can
it is
is not
make
not the
possible. Thus,
in
Thus, there
frequency
there is
is a
aascertain
zero,
certain permissible
but, in prac-
permissible various
signs forof
purpose PIDLFC based
this problem
paper control
is that
to schemes is presented.
have beenvarious
summarize proposed The
in the
controller main
re-
de-
one
limitcan
tice, ismake
for
itfor the the change
variation
not possible. ininfrequency.
Thus, frequency ascertain
zero,deviations
Larger
there is aLarger but, in prac-
permissible in signs purpose forof
cent past.
signs for LFC
LFC
thisproblem
The paper is
paper
problem
that
is organizedhave into
to summarize
that
beenvarious
proposed
beenaaproposed
have into number ofin
controller the de-
re-
sections
insections
the re-
limit
tice, it
frequency is the
not
will variation
possible.
have a in
Thus, frequency.
detrimental there is a
effect certain
on
variation in frequency. Larger deviations in signs deviations
the permissible
consumer in cent past. The paper is organized number of
limit
frequencyfor thewill have aa detrimental effect on cent for LFC
as follows:
past. The problem
section
paper that have
2 discusses
is organized been
theinto aproposed
motivation
number ofinsections
behind theLFC
re-
limit
end
end and
frequency
and
for may
the
willvariation
may lead
have to
lead to
in extensive
an
an
frequency.
detrimental
extensive
Larger
damage
effect
damage on theof
consumer
deviations
of
the the
the
in as
costly
consumer
costly cent
and
as
follows:
past. The
section
follows:
section
3 paper 2 discusses
emphasises
section 2 is organized
discusseson thetheimportance
the
motivation
into a number
motivation
behind
and ofthe
behind
LFC
sections
need
LFC
frequency
equipments will
in have to
the a detrimental effect on the consumer andfollows:
sectionsection
3 emphasises on the importance and the need
end
end
and may
equipments
and may the industry.
in lead
lead industry.
to an
Today,
Today, all
an extensive
extensive all the
the systems
damage
damage
of the
systems
of the
are
are in-
costly
in- and
costly
as
for LFC.
for section
LFC.
Section 4 2presents
3 emphasises
Section 4
discusses
presents
the
onthethethe motivationand
mathematical
importance
mathematical
behind
modelling LFC
the need
modelling
of
of
terconnectedininthe
equipments nature.
industry. Therefore,
Today, power
all the system
systems problem
are in- for and
a powersection
LFC. 3
system.
Sectionemphasises
In
4 section
presents on 5,the
the the importance
overall
mathematical and
transfer the need
function
modelling of
terconnected in nature. Therefore, power system problem is aamodels power system. In section 5,
equipments
is nothing but
terconnected
is nothing
ininthe
but
industry.
anature.
multi-area Today,
systempower
Therefore, allproblem.
thesystem
systemsThus, areitin-
problem is for LFC.are
power Section
system. 4 presents
developed
In for
section thethe
systems
5, the overall
mathematical transfer
consisting
overall transfer function
modelling
of functionof
different
terconnected
highly
is nothing butin aanature.
challenging multi-area
task
multi-area system
Therefore,
to maintain
system problem.
power system
frequency
problem. Thus,
Thus, it is
problem
constant.
it models
atypes
power of
are developed
system.
turbines. In A for
section
brief
systems
5, consisting
the overall
summary of transfer
the
of
various
of
different
function
PID
highly
is nothing challenging
butlarge task
a multi-area to maintain
system problem.frequency constant.
Thus, it is models models
types ofare are developed
turbines. A brieffor systems
summary consisting
of the various different
PID
In
In aa modern
highly challenging
modern large taskscale
scaleto interconnected
maintain
interconnected system,
frequency
system,constant.manual
manual types based control developed
of turbines.schemes isfor
A brief systems
presented
summary inconsisting
section
of the6. of different
6.various
Finally, the
PID
highly
regulation
In challenging
a modern is not task
effective,
largeeffective, to maintain
scale interconnected frequency
leading to thesystem, constant.
need formanual auto- based based
types control
of
future control schemes
turbines.
challenges of ALFC is presented
briefaresummary
presentedin section
of
andthe the Finally,
various the
PID
conclusions
regulation is not leading to the need for auto- schemes is presented in section 6. Finally, the
In a modern
matic
regulationdigital largeeffective,
control
is not scale
which interconnected
has its toown
leading thesystem,
set
need of problems auto- future
formanual based
are challenges
control of
of LFC
schemes are
are presented
is presented and
and the
in section 6. conclusions
Finally, the
matic
regulation
like
matic digital control
is not
communication
digital whichleading
effective,
control delays.
which has its
has its toown
own set
thesetneed of problems
of problems
for auto- future are drawn
future drawn in
challenges
in
challenges
section
section
of LFC
7.
LFC
7. presented the
are presented and the conclusions
conclusions
like
matic communication
digital controldelays.delays.
which has its own set of problems are drawn in section 7. are drawn in section 7.
like
The communication 2.
like
The
deviations in frequency
communication
deviations in delays. call
frequency
call for the need to design a
for the need to design aa 2. MOTIVATION
MOTIVATION
controller
The
controller which
deviations
which inshould
frequency
should be
be robust
call
robust for and
andthe most
need
most importantly
to design
importantly 2. MOTIVATION
The
simplerdeviations
controller in which
nature.inshould
frequency
It is be known, call for
robust thatandthe
moreneedthan
most to design
90% ofa It has been reported
importantly 2. MOTIVATION
in various publications about the
simpler
controller
industries in nature.
which
still It
should
employ is
PID known,
be robust that
controllers and more
most
owing thanthan 90%
importantly
to their sim-of It has
increasing been reported
number in various
of power publications
grid blackouts observed about
about the
in the
simpler
industries in nature.
still employ It is
PID known, that
controllers more
owing to their90% sim-of It has
increasing been reported
number of in
power various
grid publications
blackouts observed in the
simpler
industries in still
plicity, clear nature.
employ It is
functionality PID known,
and easethat
controllers more
of use.
owing But,than
tomany
their90% con-
sim-of It has years.
recent
increasing been reported
For example,
number in various
of power in
grid 1999,publications was aabout
there observed
blackouts in the
blackout
the
plicity,
industries
trol clear
practitioners
plicity, clear functionality
still employ
pointed
functionality PIDoutand
and ease
controllers
that
ease aof
of use.
owing
PID
use. But, tomany
controller
But, many con-
sim- recent
theirtuned
con- increasing
in
recent years.
Brazil,
years.in For
number example,
August,
For of power
example, 2003, in
grid
in in1999, there
blackouts
Northeast
1999, there was
observedaa blackout
USA-Canada,
was in the
blackout
trol practitioners
plicity, pointed
clear functionality out
and wasthat
ease nota PID controller tuned in Brazil, in August, 2003, in Northeast USA-Canada,
via conventional
trol practitioners approaches
pointed out that aofPID
use. But, many
robust.
controllerTherefore,con- in
tuned recent in years.
andBrazil, in For
2005, example,
in Russia.
August, 2003,Ininfact,
in1999, asthere
Northeastrecentlywas a blackout
as
USA-Canada,in July
via
trol conventional
practitioners
thereconventional approaches
pointed
was a needapproachesfor advancedout was
that not
a
controlPIDrobust.
controller
techniquesTherefore,
tuned
such and and
in in
Brazil,
2012,inIndia 2005,
in in Russia.
August,
suffered In
2003, fact,
in
the Inlargest as
Northeastrecently
everrecently as
USA-Canada,
power as in July
outage in
via
there was aa need for was not robust. Therefore, 2005, in Russia. fact, as in July
via conventional
as sliding
there was mode need for advanced
approaches
control, wascontrol
H-infinity,
advanced not techniques
robust.
Quantitative
control techniquesTherefore,such
such 2012,
feedback and inIndia
the history
2012, India suffered
2005,affecting
in Russia.
suffered the
overInlargest
the 620
largest ever
fact,million
as power
everrecently
people,as
power outage
in July
which
outage in
is
in
as sliding
there was mode
theory(QFT), a need control,
Linear formatrix H-infinity,
advanced inequality
H-infinity, Quantitative
control techniques
(LMI)
Quantitative basedfeedback
such the
tech-
feedback 2012, history
equivalentIndiatoaffecting
suffered
9% of overworld
the
the 620 million
largest ever power
population. people, In which
outage
order is
in
to
as sliding
theory(QFT), mode control,
Linear matrix inequality (LMI) based tech- the the history
equivalent affecting
toaffecting
9% of the over 620
world million
population. people, which
In order to is
as sliding
niques,
theory(QFT), mode
etc. control,
AtLinear
the first H-infinity,
matrix glance, itQuantitative
inequality appeared
(LMI) based thatfeedback
these avoidhistory
tech- equivalent these to issues,
9% of there over
the 620 million
isworld
need for the load
population. people, In which
frequency
order to is
niques,
theory(QFT),
techniques etc. At
are the
Linear
better first
matrix
than glance,
the PID it
inequality appeared
(LMI)
control design that
based these
tech-
but avoid avoid
equivalentthese
be constant. issues,
to 9% there
of
The frequency the is need
world for the
population.
deviation canload frequency
In
directly order
impactto
to
niques,
techniques etc.areAt better
the first than glance,
thethat it appeared
PID control that these
design but be these issues,
constant. The there is need for the load frequency to
niques,
subsequently
techniques
subsequently
etc.areAtit
it
the
was
better
was
first
observed
than
observed
glance,
the it these
PID
that
appeared
control
these
that these
controllers
design
controllers are
but
are
avoid
the
be
the power
constant.
power The frequency
these issues,
system
system
there
frequency
operation.
deviation
is need
operation. for the
The
deviation
The
can
large
can
large
directly
load frequency
deviation
directly
deviation
impactto
can
impact
can
techniques areit better
subsequently than thethat
was observed PIDthese control design but
controllers are the be constant.
power system The frequency
operation. deviation
The large can directly
deviation impact
can
subsequently
2405-8963 ©
Copyright © 2018,
itIFAC
2018 IFAC
was (International
observed that theseofcontrollers
Federation Automatic Control) are604Hosting
the power system
by Elsevier operation.
Ltd. All The large deviation can
rights reserved.
Copyright
Peer review©under
2018 responsibility
IFAC 604Control.
of International Federation of Automatic
Copyright © 2018 IFAC 604
10.1016/j.ifacol.2018.06.162
Copyright © 2018 IFAC 604
 IFAC PID 2018
Ghent, Belgium, May 9-11, 2018 Yogesh V. Hote et al. / IFAC PapersOnLine 51-4 (2018) 604–609 605

Table 1. Nomenclature: Power system param-

eters
c Percentage power generated in reheat portion
∆f Incremental frequency deviation (Hz)
Kp Electric system gain
∆Pd Load disturbance (p.u.MW)
R Speed regulation due to governor action (Hz/p.u.MW)
Tg Governor time constant (s)
Tp Electric system time constant (s)
Tr Reheat turbine constant (s)
Tt Turbine time constant (s)
Tw Time constant of hydro turbine (s)
u Input to the power system
Fig. 1. Linear model of a single-area power system
as discussed herein. The nomenclature of different system
damage the equipments, demean the load performance, parameters is given in Table 1. The dynamics of different
affect the various protection schemes in power systems and components of the power system are given as follows:
sometimes lead to a system collapse. Thus, it is important
to keep the system frequency within allowable tolerance Governor
1
range. In order to handle the issues related to frequency Gg (s) = (1)
Tg s + 1
variations, effective robust control strategies are needed
which can improve the performance of the system as well Turbine : A turbine could be classified into non- reheated
as system security. It is found that the most important turbine, reheated turbine and a hydro turbine. Depending
and simple controller to handle these issues is PID type on its type, its transfer function models are given as
controller. Therefore, there is need to explore various PID follows:
and its variants such as two degree freedom PID type Non Reheated turbine
controller, fractional order PID controller, IMC based 1
PID controller, etc. Along with this, there is a need to Gt1 (s) = (2)
Tt s + 1
make PID a robust controller which can handle parametric
uncertainty, disturbance rejection, effect of noise and the Reheated turbine
effect of communication delay. cTr s + 1
Gt2 (s) = (3)
(Tr s + 1)(Tt s + 1)
3. AIMS OF LFC
Hydro turbine
The performance of a complex power system unit deterio- 1 − Tw (s)
Gt3 (s) = (4)
rates due to degradation in power quality caused by load 1 + 0.5Tw (s)
perturbation which results in deviation in tie-line power
interchange and fluctuation in area frequency. Thus, the Load and machine
use of LFC is paramount to withstand the undesirable Kp
Gp (s) = (5)
effects on the system due to load disturbances, parametric Tp s + 1
uncertainties, non-linearities, etc. The transfer function of the overall system without droop
The principle aims of LFC are: characteristics can be computed as follows:
Gwd (s) = Gg (s)Gp (s)Gti (s) (6)
(1) Maintaining a constant frequency (zero steady state
where, i = 1, 2, 3.
error in a minimum time from the onset of distur-
bance) On the other hand, the transfer function of the overall
(2) To withstand load disturbance and counteract it. system with droop characteristics is given as
(3) To minimise unscheduled power flow between neigh- Gg (s)Gp (s)Gti (s)
bouring areas. Gd (s) = G (s)Gp (s)Gti (s)
(7)
(4) To make a power system robust to parametric uncer- 1+ g R
tainties and non-linearities. where, i = 1, 2, 3.
(5) To maintain acceptable time response specifications
within a specified margin of error. 5. TRANSFER FUNCTION MODELS OF SINGLE
AREA SYSTEMS
4. TRANSFER FUNCTION MODELS OF LFC
The overall transfer functions of power system comprising
A power system is normally a large scale system with of non-reheated turbine, reheated turbine and hydro tur-
complex nonlinear dynamics. But, for the purpose of load bine with and without droop characteristics are elucidated
frequency control, it can be approximated by a linear below.
model, linearized about the operating point. The major Non-reheated turbine: The transfer function of a non-
components of a single area power systems (Kundur et al. reheated turbine is a third order model which can be
(1994), Shayeghi et al. (2009)) are alternator, turbine, formulated as follows:
governor and load. Fig. 1 depicts the linear model of a Kp
single area power system consisting of various components G1 (s) = 3 (8)
as + bs2 + cs + d

605
IFAC PID 2018
606
Ghent, Belgium, May 9-11, 2018 Yogesh V. Hote et al. / IFAC PapersOnLine 51-4 (2018) 604–609

where a, b, c, d depend on the parameters of the turbine. a = −RKp Tw , b = RKp , c = 0.5Tg Tp Tw R,

Without droop characteristics: The transfer function for d = R(Tg Tp + Tp Tw + 0.5Tg Tw ),
the system having no droop characteristics is given by (8)
and computed via (6), resulting in the following values of e = R(Tp + Tg + 0.5Tw ) − Kp Tw , f = Kp + R
a, b, c and d.
6. BRIEF REVIEW OF PID BASED CONTROL
a = Tg Tp Tt , b = Tg Tp + Tg Tt + Tp Tt , c = Tg + Tp + Tt , TECHNIQUES
d=1
With droop characteristics: The transfer function for the LFC is basically a disturbance rejection problem wherein
system with droop characteristics is of the form given in the primal objective is the minimisation of frequency
(8) wherein the following values of a, b, c and d are obtained deviations and unwanted tie line interchange. Various
using (7). control schemes for the design of PID controller are given
in literature. Here, we have broadly classified them into
a = Tg Tp Tt , b = Tg Tp + Tg Tt + Tp Tt , c = Tg + Tp + Tt , IMC based control techniques, fractional order (FO) based
Kp
control techniques, soft computing approaches and robust
d=1+ R control schemes.
Reheated turbine: The transfer function of a reheated
turbine is a fourth order model which can be expressed 6.1 IMC based control
as follows:
Internal model control (IMC) is a well laid out mechanism
as + b
G2 (s) = 4 (9) for controller design based on Q parameterization concept
cs + ds + es2 + f s + g
3
and has been developed for integer order (IO) as well as
where a, b, c, d, e and f depend on the parameters of the fractional order (FO), SISO and MIMO continuous time
turbine. and discrete time systems. It provides a good trade-off
between a robust controller and an optimal controller. A
Without droop characteristics: The transfer function for detailed literature review of IMC scheme is presented by
the power system in the absence of droop characteristics is Saxena and Hote (2012). It is primarily used in process
given by (9) with the following values of a, b, c, d, e, f and control industries owing to steady state error minimisation
g. and good disturbance rejection properties. In Liu (2009),
a = Kp cTr , b = Kp , c = Tg Tp Tr Tt , a TDF-IMC design is studied for a single machine infinite
bus system with non reheat, reheat and hydro turbine. A
d = T g T p T t + T p T r Tt + T g T p T r + T g T r T t , unified PID tuning method based on a two degree of free-
e = Tg Tp + Tg Tt + Tg Tr + +Tp Tt + Tp Tr + Tt Tr , dom (TDF) IMC approach and PID approximation pro-
cedure is proposed by Tan (2010) for a single area power
f = Tg + T p + Tr + Tt , g = 1 system comprising of a non-reheated, reheated and hydro
With droop characteristics: The transfer function for the turbine. A reduced order model obtained via Logarithmic
power system with droop characteristics is of the form of approximations is utilized for the design of 2 DOF IMC-
(9) with the values of a, b, c, d, e, f and g given as: PID controller for a single and two area reheat hydrother-
mal power system in Singh et al. (2017) and is observed
a = RKp cTr , b = RKp , c = RTg Tp Tr Tt , to exhibit an excellent dynamic response and robustness
to load disturbance. In Tan et al. (2010), a decentralised
d = R(Tg Tp Tt + Tp Tr Tt + Tg Tp Tr + Tg Tr Tt ),
LFC is tuned for multi area power system with communi-
e = R(Tg Tp + Tg Tt + Tg Tr + +Tp Tt + Tp Tr + Tt Tr ), cation delays via TDF-IMC tuning technique. The studies
discussed above are solely based on the design of integer
f = R(Tg + Tp + Tr + Tt ) + Kp cTr , g = R + Kp and fractional order PID for LFC problem. Apart from
Hydro turbine: The transfer function of a hydro turbine these, various other IMC based control designs for LFC
is a third order model which can be presented as follows: problem are given in Saxena and Hote (2013) and Avvari
et al. (2017).
as + b
G3 (s) = (10)
cs3 + ds2 + es + f 6.2 Fractional order based control schemes
where a, b, c, d, e and f depend on the parameters of the
turbine. A fractional order (PID) can be thought of as a natural
extension of a traditional PID controller and can be
Without droop characteristics: The transfer function for mathematically formulated as follows:
the power system without droop characteristics is given Ki
by (10) with the following values of a, b, c, d, e and f . C(s) = Kp + λ + Kd sµ (11)
s
a = −Kp Tw , b = Kp , c = 0.5Tg Tp Tw , These days, FO systems have garnered immense interests
since a real world system can be better characterized by a
d = Tg Tp + 0.5Tp Tw + 0.5Tg Tw , e = Tp + Tg + 0.5Tw , f = 1
FO differential equation. A FOPID increases the degree of
With droop characteristics: The transfer function for the freedom by 2 over an integer order PID controller. Existing
power system with droop characteristics is of the form studies have confirmed that a best FOPID controller
given by (10) with the following values of a, b, c, d, e and will always surpass the performance of a best integer
f. order PID (IOPID) controller(Chen et al. (2009)). Owing

606
 IFAC PID 2018
Ghent, Belgium, May 9-11, 2018 Yogesh V. Hote et al. / IFAC PapersOnLine 51-4 (2018) 604–609 607

to its advantages, FOPID has been extensively used in fuzzy PID controller. Dhillon et al. (2015) have employed a
LFC design. Sondhi and Hote (2014) designed a FOPID combination of fuzzy based inferences and PSO algorithm
controller via stability boundary locus (SBL) for three to efficiently tune a PID controller for a five area LFC
types of turbine, i.e., non-reheated, reheated and hydro model, and Jagatheesan et al. (2017) have used FFA to
turbines. An exhaustive comparative analysis validated optimize the parameters of a PID controller and compared
the robustness and load disturbance rejection capability the results with GA and PSO. In Ahuja et al. (2014), PSO
of the proposed controller. In Ahuja et al. (2014), the algorithm is used to obtain a robust FOPID controller for
parameters of a FOPID controller are tuned via PSO a single area non reheated type power system. In Zamani
meta-heurisic search technique, whereas in Topno and et al. (2016), the authors have incorporated the advantages
Chanana (2016), Differential evolution (DE) has been of Gases Brownian Motion Optimization (GBMO) and
utilized for obtaining the controller parameters for a two fractional order to design a FOPID controller by taking
area hydrothermal power system. A FOPID controller is consideration of the saturation limits of the governor,
designed for an interval model of the single area LFC via while in Sahu et al. (2013), differential evolution (DE) is
Kharitonov’s theorem in Sondhi and Hote (2016) and is employed to design a 2- degree of freedom PID controller
found to be robust to a larger range of system parameter for a realistic power system which incorporates the effects
variations as compared to the existing controllers. It also of physical constraints such as time delay, generation rate
gives a choice of range of parametric variations desired by constraint (GRC) and governor dead band (GDB). Fur-
the control practitioner. In Alomoush (2010), the authors ther, the superiority of the proposed approach is demon-
have investigated LFC and automatic generation control strated by comparison of the results with Craziness based
(AGC) for interconnected and isolated power system via PSO (CPSO) for an interconnected two area thermal
FO controllers which results in an improvement in the system and the authors have used an modified objective
stability and time response. Conflicting design objectives function which is derived by using weighted integral time
are contemplated by Pan and Das (2015) for a multi absolute Error (ITAE), damping ratio of dominant eigen-
objective based LFC to tune the gains of FOPID controller values, settling times of frequency and peak overshoots. A
in an interconnected power system. Chathoth et al. (2015) two degree of freedom based proportional integral double
have implemented FOPID as a supplementary controller derivative (PIDD) controller is presented by Debbarma
for the enhanced performance of an interconnected multi et al. (2015) for a three unequal area thermal system and is
area deregulated power system and it is found to best found to be robust for wide change in the position of step
amongst integer order hybrid fuzzy PI controller, GA load perturbation (SLP). A new optimization technique
tuned PI controller and GA tuned PID controller. A called as quasi-oppositional grey wolf optimization algo-
hybrid linear FO fuzzy PID (FOFPID) controller is applied rithm (QOGWO) is used for the first time to solve a LFC
to Egyptian power system in Ghany et al. (2016) and problem in Guha et al. (2016) and the results are compared
the results are compared with integer order PID tuned with other intelligent methods like fuzzy logic, artificial
via bacterial foraging algorithm (BFA) and a FOPID neural network (ANN) and adaptive neuro-fuzzy interface
tuned via genetic algorithm. The FOFPID is found to system (ANFIS). Further, sensitivity analysis is performed
exhibit better performance in comparison to the existing to investigate the robustness in an uncertain environment.
techniques in the literature. The FO controllers have thus Apart from the aforementioned techniques, several other
been combined with existing tuning techniques and have soft computing based techniques are elucidated for LFC
introduced an exciting field for further research in the field problem in Yesil et al. (2014), Abdelaziz and Ali (2016),
of LFC. Abdelaziz and Ali (2015) and El-Hameed and El-Fergany
(2016).
6.3 Soft computing

Soft computing techniques belong to the Computational 6.4 Robust control schemes
Intelligence (CI) family, which comprises of a vast number
of innovative heuristic search techniques inspired from Robustness can be simply defined as the ability of the
broadly four branches like mathematics, biology, physics system to withstand changes in parameters, uncertain
and chemistry. A total of 134 such techniques are en- environment, measurement noise, load disturbances, etc.
listed in Bo Xing (2014). Some of the widely known It is the capacity of system to perform optimally, provide
and popularly used are Genetic algorithm (GA), Particle that the uncertain parameters are found within a typi-
swarm optimization (PSO), Big Bang Big crunch algo- cally compact set. Thus, trade-off between robustness and
rithm (BBBC), Firefly algorithm (FFA), Base optimiza- performance is one of the key issues of controller design.
tion algorithm, etc. They exploit a combination of ran- Various robust techniques include linear quadratic Gaus-
domness, tolerance to imprecision and inductive reasoning sian (LQG) control, quantitative feedback theory (QFT),
to circumvent the problems observed in the traditional H-infinity, Lyapunov based control, etc. In Tanaka et al.
optimization algorithms. Owing to their numerous advan- (2016), a control method is proposed based on H-infinity
tages, researchers have also applied them in the tuning control theory for a power system with distributed gen-
of parameters of PID in LFC problem. In Kumar et al. eration (DG) and validated via power transmission simu-
(2017), BBBC algorithm has been utilized to tune the lator. The proposed technique results in the suppression
parameters of a fractional order PID controller owing to of maximum frequency-deviation parameter and there is
its more flexible nature as compared to a traditional PID a marked reduction in overall frequency deviation also.
controller, whereas in Yesil (2014), same algorithm is used Davidson and Ushakumari (2016a) suggest a decentralized
to tune the scaling factor and footprint of uncertainty PID control via H-infinity approach based loop shaping
(FOU) for the membership functions of an interval type-2 method for a two area deregulated power system. The

607
IFAC PID 2018
608
Ghent, Belgium, May 9-11, 2018 Yogesh V. Hote et al. / IFAC PapersOnLine 51-4 (2018) 604–609

proposed approach is advantageous as compared to the Alomoush, M.I. (2010). Load frequency control and au-
conventional H-infinity controller as it simplifies the trial tomatic generation control using fractional-order con-
and error procedure to obtain various weighting functions. trollers. Electrical Engineering, 91(7), 357–368. doi:
In Hanwate et al. (2018), an adaptive scheme for LFC 10.1007/s00202-009-0145-7.
is reported for different cases of single and multi-area Avvari, R.K., Kotturi, R., and Ganta, V.A.R. (2017).
power systems to achieve an improved performance. A new A bip2arametric variation (Tt-Tg) degrades the per-
robust PID design based on QFT and convex-concave opti- formance of TDF-IMC scheme for load frequency
mization is developed in Mercader et al. (2017). The design control. In 2017 Innovations in Power and Ad-
problem is articulated as a convex-concave optimization vanced Computing Technologies (i-PACT), 1–5. doi:
problem and it extends the MIGO method to include 10.1109/IPACT.2017.8244887.
process uncertainty characterized by a set of plants. In Bo Xing, W.J.G. (2014). Innovative Computational In-
Satpati et al. (2008), PSO algorithm is employed in a telligence: A Rough Guide to 134 Clever Algorithms,
reduced order robust controller design for a non-minimum volume 62. Springer International Publishing, 1 edition.
phase hydro power plant to ease the computational effort doi:10.1007/978-3-319-03404-1.
and automate the loop shaping stage of QFT design. Chathoth, I., Ramdas, S.K., and Krishnan, S.T. (2015).
Various other approaches based on QFT are elucidated Fractional-order proportional-integral-derivative-based
in Khodabakhshian et al. (2006), Khodabakhshian and automatic generation control in deregulated power sys-
Golbon (2005) and Davidson and Ushakumari (2016b). tems. Electric Power Components and Systems, 43(17),
Yazdizadeh et al. (2012) presented a decentralized ro- 1931–1945. doi:10.1080/15325008.2015.1072256.
bust optimal MISO PID controller based on characteristic Chen, Y., Petras, I., and Xue, D. (2009). Fractional order
eigen values and Lyapunov method. Further, the proposed control - a tutorial. In 2009 American Control Confer-
scheme is validated on a practical control area comprising ence, 1397–1411. doi:10.1109/ACC.2009.5160719.
of Karoon3 and Dez hydro power plants in Khozestan Davidson, R.A. and Ushakumari, S. (2016a). H-infinity
which is a province is southwestern part of Iran. Apart loop-shaping controller for load frequency control of
from these, a few other robust control techniques are an uncertain deregulated power system. In 2016 In-
elucidated in Horie et al. (2017) and Khooban et al. (2016). ternational Conference on Electrical, Electronics, and
Optimization Techniques (ICEEOT), 2185–2191. doi:
7. CONCLUSION AND FUTURE CHALLENGES 10.1109/ICEEOT.2016.7755079.
Davidson, R.A. and Ushakumari, S. (2016b). H-infinity
A brief literature review of the recent PID tuning tech- loop-shaping controller for load frequency control of a
niques for load frequency control is done in this paper. The deregulated power system. Procedia Technology, 25,
authors have attempted to include most of the advances in 775–784. doi:10.1016/j.protcy.2016.08.172.
the PID design, but the presence of omissions is inevitable, Debbarma, S., Nath, A., Sarma, U., and Saikia, L.C.
considering a extremely large number of papers that are (2015). Cuckoo search algorithm based two degree of
published in this area each year. The primary challenge freedom controller for multi-area thermal system. In
in LFC is to develop a highly robust PID controller that 2015 International Conference on Energy, Power and
maintains frequency deviation strictly in specified lim- Environment: Towards Sustainable Growth (ICEPE), 1–
its even in the presence of nonlinearities like Governor 6. doi:10.1109/EPETSG.2015.7510113.
deadband (GDB) and Generation Rate constraint (GRC), Dhillon, S.S., Lather, J.S., and Marwaha, S. (2015). Multi
physical constraints and uncertain environment and vali- area load frequency control using particle swarm opti-
date it on an appropriate hardware setup. Further, there is mization and fuzzy rules. Procedia Computer Science,
a need of proper investigation on the effect of communica- 57, 460 – 472. doi:10.1016/j.procs.2015.07.363. 3rd In-
tion delay for transmitting the control signal from control ternational Conference on Recent Trends in Computing
center to remote terminal unit and ways and means to 2015 (ICRTC-2015).
minimise it. Thus, there is an ample scope of research that El-Hameed, M.A. and El-Fergany, A.A. (2016). Water cy-
the future publications can focus on to further robustify cle algorithm-based load frequency controller for inter-
the power system. connected power systems comprising non-linearity. IET
Generation, Transmission Distribution, 10(15), 3950–
REFERENCES 3961. doi:10.1049/iet-gtd.2016.0699.
Ghany, M.A.A., Ghany, A.M.A., Bensenouci, A., and
Abdelaziz, A.Y. and Ali, E.S. (2016). Load frequency Bensenouci, M.A. (2016). Fractional-order fuzzy
controller design via artificial cuckoo search algorithm. PID for the Egyptian load frequency control. In
Electric Power Components and Systems, 44(1), 90–98. 2016 Eighteenth International Middle East Power
doi:10.1080/15325008.2015.1090502. Systems Conference (MEPCON), 954–961. doi:
Abdelaziz, A. and Ali, E. (2015). Cuckoo search algorithm 10.1109/MEPCON.2016.7837012.
based load frequency controller design for nonlinear Guha, D., Roy, P.K., and Banerjee, S. (2016). Load fre-
interconnected power system. International Journal of quency control of large scale power system using quasi-
Electrical Power & Energy Systems, 73, 632–643. doi: oppositional grey wolf optimization algorithm. Engi-
10.1016/j.ijepes.2015.05.050. neering Science and Technology, an International Jour-
Ahuja, A., Narayan, S., and Kumar, J. (2014). Robust nal, 19(4), 1693 – 1713. doi:10.1016/j.jestch.2016.07.004.
FOPID controller for load frequency control using par- Hanwate, S., Hote, Y.V., and Saxena, S. (2018). Adap-
ticle swarm optimization. In 2014 6th IEEE Power tive policy for load frequency control. IEEE Trans-
India International Conference (PIICON), 1–6. doi: actions on Power Systems, 33(1), 1142–1144. doi:
10.1109/POWERI.2014.7117663.

608
 IFAC PID 2018
Ghent, Belgium, May 9-11, 2018 Yogesh V. Hote et al. / IFAC PapersOnLine 51-4 (2018) 604–609 609

10.1109/TPWRS.2017.2755468. Saxena, S. and Hote, Y.V. (2013). Load frequency

Horie, S., Yukita, K., Mathumura, T., and Goto, Y. (2017). control in power systems via internal model control
Load frequency control using H infinity control in case of scheme and model-order reduction. IEEE Trans-
considering PV. In 2017 20th International Conference actions on Power Systems, 28(3), 2749–2757. doi:
on Electrical Machines and Systems (ICEMS), 1–6. doi: 10.1109/TPWRS.2013.2245349.
10.1109/ICEMS.2017.8056445. Saxena, S. and Hote, Y.V. (2012). Advances in in-
Jagatheesan, K., Anand, B., Samanta, S., Dey, N., Ashour, ternal model control technique: A review and future
A.S., and Balas, V.E. (2017). Design of a proportional- prospects. IETE Technical Review, 29(6), 461–472. doi:
integral-derivative controller for an automatic genera- 10.4103/0256-4602.105001.
tion control of multi-area power thermal systems using Shayeghi, H., Shayanfar, H., and Jalili, A. (2009). Load
firefly algorithm. IEEE/CAA Journal of Automatica frequency control strategies: A state-of-the-art survey
Sinica, PP(99), 1–14. doi:10.1109/JAS.2017.7510436. for the researcher. Energy Conversion and Management,
Khodabakhshian, A. and Golbon, N. (2005). Design 50(2), 344–353. doi:10.1016/j.enconman.2008.09.014.
of a new load frequency PID controller using QFT. Singh, J., Chattterjee, K., and Vishwakarma, C. (2017).
In Proceedings of the 2005 IEEE International Sym- Two degree of freedom internal model control-
posium on, Mediterrean Conference on Control and PID design for LFC of power systems via loga-
Automation Intelligent Control, 2005., 970–975. doi: rithmic approximations. ISA Transactions. doi:
10.1109/.2005.1467145. doi.org/10.1016/j.isatra.2017.12.002.
Khodabakhshian, A., Rahimi, H., and Golbon, N. (2006). Sondhi, S. and Hote, Y.V. (2014). Fractional order
QFT design for load frequency control of non-minimum PID controller for load frequency control. Energy
phase hydro power plant. In 2006 IEEE Confer- Conversion and Management, 85, 343 – 353. doi:
ence on Computer Aided Control System Design, 2006 10.1016/j.enconman.2014.05.091.
IEEE International Conference on Control Applica- Sondhi, S. and Hote, Y.V. (2016). Fractional order
tions, 2006 IEEE International Symposium on Intel- PID controller for perturbed load frequency control
ligent Control, 1380–1385. doi:10.1109/CACSD-CCA- using Kharitonovs theorem. International Journal of
ISIC.2006.4776843. Electrical Power & Energy Systems, 78, 884 – 896. doi:
Khooban, M.H., Niknam, T., Blaabjerg, F., Davari, P., and 10.1016/j.ijepes.2015.11.103.
Dragicevic, T. (2016). A robust adaptive load frequency Tan, W. (2010). Unified tuning of PID load fre-
control for micro-grids. ISA Transactions, 65, 220 – 229. quency controller for power systems via IMC. IEEE
doi:10.1016/j.isatra.2016.07.002. Transactions on Power Systems, 25(1), 341–350. doi:
Kumar, N., Tyagi, B., and Kumar, V. (2017). Application 10.1109/TPWRS.2009.2036463.
of fractional order PID controller for AGC under deregu- Tan, W., Yu, M., and Zhang, H. (2010). Decentral-
lated environment. International Journal of Automation ized LFC tuning for power systems with communi-
and Computing. doi:10.1007/s11633-016-1036-9. cation delays. In IEEE ICCA 2010, 209–214. doi:
Kundur, P., Balu, N.J., and Lauby, M.G. (1994). Power 10.1109/ICCA.2010.5524287.
System Stability and Control. McGraw-Hill. Tanaka, Y., Yukita, K., and Goto, Y. (2016). Experiment
Liu, J. (2009). A novel PID tuning method for load of load frequency control using H-infinity control. In
frequency control of power systems. In 2009 3rd In- 2016 19th International Conference on Electrical Ma-
ternational Conference on Anti-counterfeiting, Security, chines and Systems (ICEMS), 1–4.
and Identification in Communication, 437–442. doi: Topno, P.N. and Chanana, S. (2016). Fractional order
10.1109/ICASID.2009.5276973. PID control for LFC problem of a hydro-thermal power
Mercader, P., strm, K.J., Baos, A., and Hgglund, T. system. In 2016 11th International Conference on
(2017). Robust PID design based on QFT and Industrial and Information systems (ICIIS), 1–6.
convex-concave optimization. IEEE Transactions on Yazdizadeh, A., Ramezani, M.H., and Hamedrahmat, E.
Control Systems Technology, 25(2), 441–452. doi: (2012). Decentralized load frequency control using a
10.1109/TCST.2016.2562581. new robust optimal MISO PID controller. International
Pan, I. and Das, S. (2015). Fractional-order load-frequency Journal of Electrical Power & Energy Systems, 35(1),
control of interconnected power systems using chaotic 57 – 65. doi:10.1016/j.ijepes.2011.09.007.
multi-objective optimization. Applied Soft Computing, Yesil, E., Savran, A.I., and Guzay, C. (2014). Load
29, 328 – 344. doi:10.1016/j.asoc.2014.12.032. frequency controller design using new big bang -
Sahu, R.K., Panda, S., and Rout, U.K. (2013). DE big crunch 2 algorithm. In 2014 IEEE Interna-
optimized parallel 2-DOF PID controller for load fre- tional Symposium on Innovations in Intelligent Sys-
quency control of power system with governor dead- tems and Applications (INISTA) Proceedings, 7–13. doi:
band nonlinearity. International Journal of Electri- 10.1109/INISTA.2014.6873590.
cal Power & Energy Systems, 49, 19 – 33. doi: Yesil, E. (2014). Interval type-2 fuzzy PID load fre-
10.1016/j.ijepes.2012.12.009. quency controller using big bang-big crunch optimiza-
Satpati, B., Bandyopadhyay, I., Das, G., and Koley, C. tion. Applied Soft Computing, 15, 100–112. doi:
(2008). Robust controller design for load frequency 10.1016/j.asoc.2013.10.031.
control of non-minimum phase hydro power plant using Zamani, A., Barakati, S.M., and Yousofi-Darmian, S.
PSO enabled automated quantitative feedback theory. (2016). Design of a fractional order PID controller
In 2008 Annual IEEE India Conference, volume 2, 349– using GBMO algorithm for load-frequency control with
354. doi:10.1109/INDCON.2008.4768748. governor saturation consideration. ISA Transactions,
64, 56 – 66. doi:10.1016/j.isatra.2016.04.021.

609