Electrical Power and Energy Systems 82 (2016) 579–598

Contents lists available at ScienceDirect

Electrical Power and Energy Systems

journal homepage: www.elsevier.com/locate/ijepes

A new design of dual mode Type-II fuzzy logic load frequency controller
for interconnected power systems with parallel AC–DC tie-lines and
capacitor energy storage unit
N.J. Vinoth Kumar ⇑, M. Mohamed Thameem Ansari 1
Department of Electrical Engineering, Annamalai University, Annamalai Nagar, Tamil Nadu 608002, India

a r t i c l e i n f o a b s t r a c t

Article history: A new design of dual mode Type-II fuzzy logic load frequency controller (DMT-IIFLC) for interconnected
Received 31 October 2014 power systems with parallel AC–DC tie-lines and capacitor energy storage unit (CES) is proposed in this
Received in revised form 19 March 2016 paper. Any optimum controller selected for load frequency control of interconnected power systems
Accepted 29 March 2016
should not only stabilize the power system, but also reduce the system frequency and tie-line power
Available online 26 April 2016
oscillations and settling time of the output responses. In general proportional plus integral controllers
are used for loadfrequency control but it does not eliminate the conflict between the static and dynamic
Keywords:
accuracy. This dispute may be broken up by using the principle of T-IIFLC to utilize expert knowledge and
Load frequency control
PI controllers
being adaptive in nature. The dual-mode concept is incorporated in the proposed controller to improve
AC–DC tie-lines the system performance. A DC link is connected in parallel with AC tie-line to stabilize the frequency of
Dual mode Type-II fuzzy logic controller oscillations. In summation, the CES unit is found to be advantageous for secondary control in the power
Performance Index system and keeps the power quality with the distributed power resources. The CES, which are not aged to
CES the frequent charging and discharging, has been a quick response and outstanding function during
overload conditions. The system performance of the proposed controller have improved significantly
when compared with PI and T-IFLC. It is also found to be less sensitive to the changes in the system
parameters and also robust under different operating modes of the system.
Ó 2016 Elsevier Ltd. All rights reserved.

Introduction It is a well-accepted fact that the classical conventional solution

The interconnected power system encounters a great challenge
in power system design and operation. The load–frequency control
(LFC) problem has gained much importance because of the size and
complexity of modern interconnected power systems. The objec-
tive of LFC is to regulate the output power of the regulating plants
so that the frequency of the power system and tie-line power is
kept within prescribed limits [1–3]. Many control strategies for
LFC, of power systems, have been investigated and proposed by
many researchers over the past several years [4,5]. The application
of control strategy to the LFC problem has found wide acceptance
because of its role in eliminating most of the problems associated
with other multilevel control strategies. The proportional and inte-
gral (PI) controller is most widely applied for the LFC scheme [6].
⇑ Corresponding author. Cell: +91 9994545679.
E-mail addresses: vinothkumarnj@gmail.com (N.J. Vinoth
ansari_aueee@yahoo.co.in (M.M. Thameem Ansari).
1 systems that are too complex or not well defined by mathematical
Cell: +91 9894795210.
of this problem has been one of the first practical applications for
the LFC problem. The advantage of using the PI controller lies in
their simplicity, easy realization, low cost and robust and decen-
tralized nature of the control strategy [7,8]. However, the conven-
tional controller reduces the steady-state to zero but it exhibits
poor dynamic performance and does not reach high performance.
Furthermore, the settling time of the system frequency and tie-
line power deviations are also relatively long. Recently, conven-
tional PI controllers have been replaced by fuzzy logic controllers
(FLC). FLC has received an important role in power systems [9].
An FLC, described completely in terms of Type-I fuzzy sets,
(T-IFS) is called a Type-I fuzzy logic system or Type-I fuzzy logic
controller (T-IFLC). Fuzzy logic is a logical system for formalization
of approximate reasoning, and is used synonymously with fuzzy
set theory, systems introduced by Zadeh and investigated further
by many researchers [10–12]. Since it is able to model human
decision making process and represents vague and uncertain data,
fuzzy set theory is a theory about vagueness and uncertainty.
Kumar), This theory provides a methodology that allows modeling of the

http://dx.doi.org/10.1016/j.ijepes.2016.03.063
0142-0615/Ó 2016 Elsevier Ltd. All rights reserved.
580 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598

Nomenclature

A Type-I fuzzy set Ki integral constant

A~ Type-II fuzzy set DXEi Governor valve position in p.u. MW
T-I FLC Type-I fuzzy logic controller DPTHi deviation in intermediate state of reheat turbine in p.u.
T-II FLC Type-II fuzzy logic controller MW
DMT-II FLC dual-mode Type-II fuzzy logic controller DPGi deviation in thermal turbine output in p.u.MW
FOU footprint of uncertainty DFi frequency deviation of ith area in Hz
FOU upper bound of FOU DPtieij tie-line deviation between ith and jth area
FOU lower bound of FOU DPdi deviation in demand in pu-power
Jx interval # [0, 1]. ei switching limit
U Universe of discourse = [0, 1] Kd,ei, electric governor derivative constant
X input variable (crisp input) Kp,ei, electric governor proportional constant
X Universe of discourse (x domain) Ki,ei electric governor integral constant
v state vector S laplace frequency variable.
l Type-I membership function N number of interconnected areas
lA~ Type-II membership function D incremental change of a variable
lA~ Type-II lower membership function LFC load frequency control
l A~ Type-II upper membership function PI proportional and integral
bi frequency bias constant in p.u. MW/Hz AC alternating current
Ri steady state regulation of governor in Hz/p.u. MW DC direct current
TGi time constant of steam turbine governor mechanism in FLC fuzzy logic controller
seconds ACE area control error
KRi reheat coefficient of the steam turbine ISE integral square error
TRi reheat time constant of the steam turbine in seconds CES capacitor energy storage
TTi time constant of the steam turbine in seconds
KPi gain constant of power system Superscript
TPi time constant of power system Transpose of a matrix
Kdc gain constant of rectifier and converter unit
Tdc time constant of rectifier and converter unit Subscripts
aij area size ratio i, j area indices (i, j = 1, 2, . . . , N)
Tij tie-line power coefficient
Kp proportional constant

formulation. Therefore, T-IFLC becomes nonlinear and adaptive in
nature, having a robust performance under parameter variations
with the ability to get desired control actions for complex, uncer-
tain, and nonlinear systems without the requirement of their
mathematical models and parameter estimation [13,14]. The gen-
eral framework of fuzzy reasoning handles much of uncertainty; a
Fuzzy system uses T-IFS, which represents uncertainty by numbers
in the range [0, 1]. When something is uncertain, like a measure-
ment, it is difficult to determine its exact value. However, it is
not reasonable to use an accurate membership function for some-
thing uncertain, so in this case we need another type of fuzzy sets,
those that are able to handle these uncertainties, the so-called
Type-II fuzzy logic control (T-IIFLC) [15–17]. Thus, the amount of
uncertainty in a system can be reduced by using T-IIFLC because
it offers better capabilities to handle the linguistic uncertainties
by modeling vagueness and unreliability of information [18].
A T-IIFLC is characterized by a fuzzy membership function, that
is, the membership grade for each element of this set is a fuzzy set
in 0,1 unlike a T-IFS where the membership grade is a crisp number
in the range 0,1 [19–21]. Such sets can be used in situations in
which there is an uncertainty about the membership grades them-
selves, for examples, an uncertainty in the shape of the member-
ship functions or in some of its parameters [21,22]. To achieve
the best system configuration possible, it is important to take into
account the design of the energy storage system.
The energy storage unit is an attractive alternative to augment
demand side management implementation. A quick-acting energy
storage system, in accession to the kinetic energy of the generator
rotors, provides adequate control to mute out the frequency oscil-
lations. The problems such as low discharge rate, increased time
required for power flow reversal and maintenance requirements
have lead to the evolution of a energy storage devices for their
applications in load frequency stabilizers [23,24]. By using energy
storage systems, a low-cost source of electricity can be efficiently
supplied to satisfy the peak requirement [25].
The capacitor energy storage (CES) system is suggested as
energy storage unit for improving the dynamic behavior of LFC
for two area power systems. The CES will, in addition to load lev-
elling, a function conventionally assigned to them, have a wide
range of applications such as power quality maintenance of decen-
tralized power supplies [26–28]. The CES is excellent for short-time
overload output and the response characteristics possessed in par-
ticular. The effect of generation control and the absorption of
power fluctuation needed for power quality maintenance are
required. In this survey, a two area power system with CES unit
is taken to control power flows [28]. In each control area, all the
generators are assumed to form a coherent group.
In the recent days, high-voltage DC transmission system has
witnessed unprecedented development in the power system
because of its performance, economy and environment friendly
nature over the other options. By implementing a DC link in paral-
lel with an AC link, the dynamic performance of the system can be
improved with greater stability margin under small disturbances
in the system [29,30].
An important design concept of the dual mode is used in the
proposed controller because it improves the system performance
and makes it flexible and applicable to actual systems [31,32].
The computer simulation results of the application of the proposed
controller with interconnected power systems proved that the
proposed controller is effective and provides significant improve-
ment in the system performance. Moreover, it has likewise been
discovered that the proposed controller is less sensitive to system
 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598 581

CES-1 CES-2

1 2
P AC
AC link Area-2
Area-1
Inverter Converte Thermal Power System
Thermal Power System
P DC DC link

Change in
Load

Fig. 1. One line diagram of interconnected two-area thermal power systems with parallel AC–DC tie-lines and CES unit.

Fig. 2. Transfer function model of interconnected two area thermal power systems with parallel AC–DC tie-line and CES unit.

parameter variations and is also robust under different operating X

N

conditions [33,34]. n¼ ni þ ðN  1Þ;

i¼1

Statement of the problem

The state variable equation of the minimum realization of the
continuous model of the ‘N’ area interconnected power system is
expressed as [32]
u ¼ ½u1 ; . . . ; uN T ¼ ½DPC1 . . . PCN T ; N  control input vector;
T
d ¼ ½d1 ; . . . ; dN  ¼ ½DPD1 . . . P DN T ; N  disturbance input vector;

X_ ¼ Ax þ Bu þ Cd ð1Þ t ¼ ½t1 ; . . . ; tN T ; N  control output vector;

t ¼ Cx ð2Þ y ¼ ½y1 ; . . . ; yN T ; 2N  measurable output vector;

where A is the system matrix, B is the input distribution matrix, C is

y ¼ Hx ð3Þ the disturbance distribution matrix, C is the control output distribu-
tion matrix, H is the measurable output distribution matrix, x is the
T state vector, u is the control vector, d is the disturbance vector of
x ¼ ½xT1 ; Dpei . . . xTðN1Þ; DpeðN1Þ . . . xTN  n  state vector load changes.
582 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598

where ki1 ¼ ki2 ; kp1 ¼ kp2 is a two dimensional integral and propor-
tional feedback gain vectors.

Modeling of two-area power systems with AC–DC tie-lines and

CES units
Fig. 3. Transfer function model of HVDC link.
Modeling of power system

Output feedback control scheme

The one line diagram and the transfer function model of a two-
area power system under investigation are shown in Figs. 1 and 2
It is known that by incorporating an integral controller the
respectively. The two areas are interconnected through an AC–DC
steady state requirements can be achieved. To introduce an inte-
tie-line. To improve the dynamic performance of the system CES
gral function to the controller the system Eq.(1) is augmented with
unit are also incorporated with this model. The transfer function
a new state variable defined as the integral of the area control error
R model of HVDC link is shown in Fig. 3. The output power of
ACEi ð ti dtÞ, i = 1, 2, 3, . . . , N. The augmented system of the order (N area-1 is given by [29,30],
+ n) can be described as [34],
PGi ¼ PTHi ð8Þ
X_ ¼ Ax þ Bu þ C
d ð4Þ
For small load perturbation in area-i can be written as,
R 
where 
x¼
tdt gN DPGi ¼ DPTHi ð9Þ
x gn
      where i = the area index. PGi = Deviation in thermal turbine output.
0 C ¼ 0 ¼ 0
A¼ ;B and C PTHi = Deviation in intermediate state of reheat turbine.
0 A B C For small load perturbation the tie-line flow from area-1 to
R
As the newly added state variables ð ti dtÞ, i = 1, 2, . . . , N will also area-2 is given by,
be available for feedback in each area, the new measurable output 2pT 12
 can be written as,
vector y DPtieAC ¼ ðDF 1  DF 2 Þ ð10Þ
s
 ¼ Hx
y ð5Þ where DPtieAC = Tie-line power deviation of AC. 2pT12 = Tie-line
power coefficient of area 1 and 2. DF1, DF2 = Frequency deviation
¼
where y TN T
T1 . . . y
½y in area 1 and 2 respectively.
H ¼ ½HT1 . . . HTN ; and For small changes in DC tie-line flow, DPtieDC is given as,

  K dc
0 1...0 DPtieDC ¼ ð DF 1 Þ ð11Þ
H¼ 1 þ sT dc
0 0 . . . Hi
) DPtie12 ¼ DPtieAC þ DPtieDC ð12Þ
The constant matrix Hi ði ¼ 1; 2; . . . ; NÞ is of dimension 2  (N
 is of dimension 2N  (N + n).
+ n). Hence, the matrix H where DPtie12 = Tie-line power deviation between area 1 and 2.
The problem now is to design the decentralized feedback out- DPtieDC = Tie-line power deviation of DC. DPtieAC = Tie-line power
put feedback control law deviation of AC. Kdc = Gain constant of rectifier and converter unit.
Tdc = Time constant of rectifier and converter unit.
T
ui ¼ ki yi i ¼ 1; 2; . . . ; N ð6Þ The area control error (ACE) for the two areas is given by,
To meet the objectives stated in the previous section, the con- ACE1 ¼ b1 DF 1 þ ðDPtieAC þ DP tieDC Þ ð13Þ
trol law, Eq. (6), can be written in terms of ti as,
Z ACE2 ¼ b2 DF 2 þ a12 ðDPtieAC þ DPtieDC Þ ð14Þ
ui ¼ kil ti dt  ki2 ti i ¼ 1; 2; . . . ; N ð7Þ
where a12 is given by,

Fig. 4. System configuration of capacitive energy storage unit.

 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598 583

Fig. 5. Transfer function model representation of the CES control with capacitor voltage deviation.

Fig. 6. Transfer function model of electric governor.

Pr1
a12 ¼  ð15Þ
Pr2

where Pr1, Pr2 = Power ratings of area-1 and area-2. b1,b2 = Frequency
biased parameter. a12 = Area size ratio.

Description and modeling of capacitive energy storage system Fig. 8. Membership function of interval Type-II fuzzy logic set.

Principle of capacitor energy storage unit

When there is a sudden rise in load demand, the stored energy
is almost immediately put out through the power conversion sys-
tem (PCS) to the grid as pulsed AC. As the governor and other con-
trol mechanisms start working to set the power system to the new
equilibrium condition, the capacitor charges to its initial value of
voltage. The action during sudden releases of load is similar. The
capacitor immediately gets shot down towards its total value, thus
soaking up some portion of the excess energy in the system, and as
the system returns to its steady state, the excess energy absorbed
is released and capacitor voltage reaches its normal value [23–25].
To show the effect of CES on LFC, a small rating CES unit having
maximum capacities of 3.8 MJ is considered. The capacitor voltage
must not be permitted to deviate beyond certain lower and upper
bounds. If the capacitor voltage goes too low during a system upset
and if some other disturbance occurs before the voltage returns to
normal, more energy will be pulled away from the capacitor [25].
This may lead to discontinuous control. To overcome this problem,
the lower limits are placed on the capacitor rating, insulation level
of the capacitor and the rating of the converter bridge.
The normal operating point of the capacitor is set at the maxi-
mum allowable energy absorption which equals the maximum
allowable energy discharge. This makes the CES unit equally effec-
tive in damping the oscillations created by the sudden increase, as
well as decrease, in the load. If Edmax and Edmin denote the maxi-
mum and minimum limits of voltage, respectively, and Edo
denotes the set value of voltage then [26]
ð1=2ÞCE2dmax  ð1=2ÞCE2do ¼ ð1=2ÞCE2do  ð1=2ÞCE2dmin

Fig. 7. Block diagram of a DMT-IIFLC.

584 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598

ðE2dmax þ E2dmin Þ thyristor is provided to accommodate the change of direction of

E2do ¼
2
System configuration of CES unit
The system configuration of CES unit is shown in Fig. 4. The
storage capacitor is connected to the AC grid through a PCS that
includes a rectifier/inverter in 12-pulse configuration. The storage
capacitor C may consist of many discrete capacitance units con-
nected in parallel. The resistance R connected in parallel across
the capacitor represents the dielectric and leakage losses of the
capacitor bank [27].
The capacitor can be loaded to an asset value of potential drop
from the utility grid during normal operation of the two power
schemes. A reversing switch arrangement, using a gate turn-off
ð16Þ the current in the capacitor during charging and discharging, (i.e.,
during rated load period and peak load period). Since the focal
point of the current through the bridge converter cannot change,
the charging modes of switches S2 and S3 are OFF and those of
S1 and S4 are ON. In the discharge mode, switches S2 and S3 are
ON and switches S1 and S4 are OFF.
Modeling of capacitor energy storage unit
The operation of CES unit that is, charging, discharging, the
steady-state mode, and the power modulation during the dynamic
oscillatory period, is controlled by the application of the proper
voltage to the capacitor so that the desired current flows into or
out of the CES. This can be achieved by accomplishing the firing

Fig. 9. Interval Type-II fuzzy reasoning.

OUTPUT
RULE PROCESSING
BASE
CRISP
INPUT DEFUZZIFIER
FUZZIFIER
CRISP
OUTPUT

TYPE-REDUCER

INFERENCE
ENGINE

FUZZY INPUT FUZZY OUTPUT

SETS SETS

Fig. 10. Structure of a Type-II fuzzy logic system.

 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598 585

Fig. 11. Membership function of the interval Type-II fuzzy sets.

angle of the converter bridges [28]. Neglecting the transformer and

Table 1 the convertor losses, the DC voltage would be,
Rule base for DMT-IIFLC membership function.
Edo ¼ 2V do cos a  2Id Rc ð17Þ

where Edo is a DC voltage applied to the capacitor (kV), a the firing

angle (degree), Id the current through the capacitor (kA), Rc the
equivalent commutating resistance (ohm), Vdo the maximum open
circuit bridge voltage of each 6-pulse convertor at a = 0° (kV). The
capacitor is initially billed to its normal voltage, Edo by the PCS. Once
the voltage of the capacitor has reached Edo, it is kept floating at this
voltage by continuing supply from the PCS to compensate for the
dielectric and other leakage losses of the capacitor. The energy
stored at any instant,

Fig. 12. Matlab simulink model of interconnected two area thermal reheat power systems with parallel AC–DC tie-lines and CES unit using DMT-IIFLC.
586 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598

CE2 d different from KCF, the gain constant for frequency deviation as con-
Wc ¼ MJ ð18Þ trol signal. If a signal proportional to ACE ðDF i þ ð1=bi ÞDP ij Þ is used
2
in such a scheme then,
where C = capacitance of CES in Farad.  
The area control error (ACE) can be sensed and used to control K CAi
DIdi ¼ ðDF i þ 1=bi DPij Þ i; j ¼ 1; 2 ð21Þ
the CES current Id. The incremental change in CES current is 1 þ ST dci
expressed as,
  It is necessary to quickly restore the set value of the CES voltage
K CAi after a load disturbance, so that the CES unit is ready to act for the
DIdi ¼ DF i i ¼ 1; 2 ð19Þ
1 þ ST dci next load disturbance. The capacitor voltage deviation can be
sensed and used as a negative feedback signal in the CES control
The ACE of the two areas (i = 1, 2) is defined as
loop to achieve fast restoration of the voltage. Then, the frequency
ACEi ¼ bi DF i þ DPij i; j ¼ 1; 2 ð20Þ deviation as the control signal is, [25–28]

where DFi is the deviation in frequency and DPij the deviation in  

1
tie-line power flow out of area i,j. If ACE is directly used for the DEdi DIdi ð22Þ
C S þ 1=R
control of CES, the gain constant KCA (kA/unit ACE) would be totally

0.02
PI controller
Frequency deviation ΔF 1 in Hz.

0.015
T-IFLC
0.01
DMT-IIFLC
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
PI controller
Frequency deviation ΔF2 in Hz.

0.015
T-IFLC
0.01
DMT-IIFLC
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

x 10
8
Change in tie-line (∆Ptie ) in p.u. MW

6 PI controller
T-IFLC
4 DMT-IIFLC
2

-2

-4

-6

-8
0 5 10 15 20 25 30 35 40 45 50
Time in sec

Fig. 13. Performance analysis of frequency deviations DF1, DF2 and tie-line power deviation DPtie of PI controller, T-IFLC, and DMT-IIFLC for a step load change in area-1
(Power system model + CES + AC–DC tie-line).
 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598 587

 
K CAi to perform low power functions associated with speed-sensing
DIdi ¼ ½K CAi DF i  K v di DEdi  ð23Þ
1 þ ST dci and droop compensation. If an electric governor has been replaced
with a mechanical governor, it would be a practical significance to
If ACE is used as control signal then,
explore the system performance. The transfer function model for
1   the electric governor is shown in Fig. 6. In the figure,
DIdi ¼ K CAi ðDF i þ 1=bi DPij Þ  K v di DEdi i; j ¼ 1; 2 ð24Þ
1 þ ST dci K d;ei ; K p;ei ; K i;ei are the electric governor derivatives, proportional
and integral gains respectively.
where K v di ðkA=kVÞ is the gain corresponding to the DEdi feedback.
The block diagram representation of such a control scheme is
shown in Fig. 5. Design of a dual mode Type-II fuzzy logic controller

Modeling of electric governor system Designing of the dual-mode controller

In the olden days, many of the existing thermal power plants The zero steady state error can be achieved by implementing a
were equipped with mechanical governors. Owing to significant well designed PI controller, but the response of the system
development in scientific research, electrical governors are used becomes slow resulting from high over/under shoot and making

0.02
Frequency deviation ΔF 1 in Hz.

0.015 PI controller
0.01 T-I FLC
DMT-II FLC
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
Frequency deviation ΔF2 in Hz.

0.015 PI controller
0.01 T-I FLC
DMT-II FLC
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

x 10
8
Change in tie-line (∆Ptie ) in p.u. MW

6 PI controller
T-I FLC
4
DMT-II FLC
2

-2

-4

-6

-8
0 5 10 15 20 25 30 35 40 45 50
Time in sec

Fig. 14. Performance analysis of frequency deviations DF1, DF2 and tie-line power deviation DPtie of PI controller, T-IFLC, and DMT-IIFLC for a step load change in area-2
(Power system model + CES + AC–DC tie-line).
588 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598

up time. The proportional controller by itself will improve the Hence, a Type-II membership grade can be any subset in [0, 1],
over-shoot and the speed of the response. Obviously, the comport- the primary membership, and corresponding to each primary
ment of the proportional controller is compelling at transient state membership, there is a secondary membership (which can also
to make the system response faster so as to reduce the overshoot. be in [0, 1]) that defines the possibilities for the primary member-
But incorporating the proportional controller alone fails to bring ship [15,16]. The membership function of a general T-IIFS is
the steady state error to zero [31]. So, the usage of PI controllers denoted by A, ~
simultaneously seems urgent. The PI controller does not get rid Z Z
of the difference between the static and dynamic accuracy. e¼ lA~ ðx; uÞ
A ; J x # ½0; 1 ð26Þ
This conflict may be checked up by applying the principle x2X u2Jx ðx; uÞ
of dual-mode controller [32]. Block diagram of a dual-mode ~ I is defined as,
An interval T-IIFS A
controller is shown in Fig. 7.
R R
Science the principle of dual mode controller can improve the ½ ð1=uÞ
eI ¼
A
x2X u2J X # ½0;1
ð27Þ
system performance a new design of dual mode T-IIFLC is proposed x
in this section. The proposed controller operates in mode ‘‘A” as
long as the significant observed variables to the control actions leA ðx; uÞ; x 2 X; u 2 Jx is a secondary grade and the domain of a sec-
and the system output error are sufficiently large, that is, greater ondary membership function is called the primary membership of
than the switching limit of the controller, otherwise it operators x, where Jx # ½0; 1 8x 2 X.
in mode ‘‘B”. Mode A acts as proportional type T-IIFLC and mode A simple way of representing the Type-II fuzzy membership
B as integral type T-IIFLC. Thus, the control structure of the system function is to use the footprint of uncertainty (FOU), which is a
is changed when switching in each mode of operation. As the pro- two-dimensional representation. Uncertainty in the primary mem-
posed controller is designed based on the switching limit of the berships of a T-IIFS consists of a bounded region that is called the
controller, the performance of the controller gets improved signif- footprint of uncertainty (FOU) [17–19] as shown in Fig. 8. FOU
icantly. The dual-mode Type-II fuzzy logic controller (DMT-IIFLC) is characterizes T-IIFS and is specified as the unification of all primary
an advanced version of T-IFLC [32–34]. memberships:
[
e ¼
FOUð AÞ Jx
Type-II fuzzy logic controller x2X

FOU is associated with the concepts of lower membership func-

The concept of T-IIFLC was introduced by Zadeh in 1975 as an
extension of the concept of an ordinary fuzzy set. Fig. 8 displays
all the definitions adopted for T-IIFS or T-IIFLC. A T-IIFS is charac-
terized by a fuzzy membership function that is, the membership
grade for each element of this set is a fuzzy set in [0, 1], unlike a
T-IFS where the membership grade is a crisp number in [0, 1].
~ is characterized by the membership function:
A T-IIFS A,
e ¼ fððx; uÞ; l~ ðx; uÞÞj 8x 2 X; 8u 2 J # ½0; 1g
A ð25Þ
A x
in which, 0 6 lA~ ðx; uÞ 6 1. In fact J x # ½0; 1 represents the primary
membership of x and lA~ ðx; uÞ.
tion (LMF) and upper membership function (UMF) and models the
uncertainties in the shape and position of the T-IIFS; its uniform
shading denotes interval sets for the secondary membership func-
tions and represents the entire interval T-IIFS lA~ ðx; uÞ; x the UMF
and LMF are in fact two Type-I membership functions that are
~ By shifting the membership func-
bounds for the FOU of a T-IIFS A.
tions of the Type-I fuzzy sets upward and downward by an angle
h1 2 [0, 0.5] that is, (h1 = 0; h1 = 0.5) for lP(DACE) and lN(DACE)
along the membership axes, the boundary membership functions
of the primary memberships of the interval Type-II fuzzy sets are
obtained (i.e., lPL(DACE), lPU(DACE), lNL(DACE), lNU(DACE)).

Table 2
Performance comparison of different control techniques.

Controller Operating condition Settling time Peak overshoot/undershoot

DF1 in Hz DF2 in Hz DPtie12 in p.u. MW DF1 in Hz DF2 in Hz DPtie12 in p.u. MW
Conventional PI controller With AC tie-line 27.50 26.80 24.25 +0.0018 +0.0031 +0.0039
0.0098 0.015 0.0005
With DC tie-line 25.02 24.02 23.25 +0.003 +0.005 +0.0065
0.0065 0.015 0.0042
With AC–DC tie-line 35.00 34.90 36.25 +0.001 +0.0031 +0.0044
0.0063 0.015 0.0015
AC–DC tie-line with electric governor 34.00 32.25 33.12 +0.001 +0.0011 +0.0027
0.0051 0.012 0.0010
T-IFLC With AC tie-line 17.32 13.25 15.01 +0.005 +0.003 +0.008
0.028 0.023 0.0005
With DC tie-line 17.25 16.95 16.05 +0.000 +0.000 +0.0078
0.027 0.029 0.0000

With AC–DC tie-line 23.32 16.55 15.45 +0.003 +0.005 +0.0003

0.025 0.022 0.0077
AC–DC tie-line with electric governor 10.95 9.80 10.00 +0.001 +0.000 +0.000
0.021 0.023 0.0079
DMT-IIFLC With AC tie-line 15.00 11.25 14.90 +0.000 +0.000 +0.006
0.024 0.020 0.000
With DC tie-line 05.10 04.90 10.00 +0.000 +0.000 +0.0069
0.020 0.019 0.000
With AC–DC tie-line 11.25 10.22 11.23 +0.012 +0.016 +0.012
0.025 0.024 0.0060
AC–DC tie-line with electric governor 4.80 7.23 7.77 +0.001 +0.002 +0.003
0.0230 0.026 0.0007
 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598 589

Where lPL = positive lower of Type-II membership function. The defuzzification is a mapping process from FLC action to a
lPU = positive upper of Type-II membership function. lNL = nega- non-fuzzy (crisp) control action. Defuzzification on a T-IIFLC using
tive lower of Type-II membership function. lNU = negative upper centroid method is shown in Fig. 9. A T-IIFLC is again characterized
of Type-II membership function. by IF-THEN rules, but its antecedent or consequent sets are now of
l~ A~ ðxÞ is the UMF and is associated with the upper bound of Type-II. The T-IIFLC can be used when the circumstances are too
FOUðAÞ ~ It is also denoted as [20,21], uncertain to determine exact membership grades such as when
the training data are corrupted by noise. Similar to a T-IFLC, a T-
~ 8x 2 X
l A~ ðxÞ  FOUðAÞ; ð28Þ IIFLC includes Type-II fuzzifier, a rule base, fuzzy inference engine,
and an output processor, as we can see in Fig. 10. An inference
lA~ ðxÞ is the LMF associated with the lower bound of FOU and is also fuzzy system is a rule-based system that uses fuzzy logic, instead
denoted as of Boolean logic used in data analysis. Its basic structure includes
four components: Fuzzification, inference system, type defuzzifica-
~ 8x 2 X
lA~ ðxÞ  FOUðAÞ; ð29Þ
tor/reductor and knowledge base. These are briefly elaborated
below [20–22].

0.02
PI controller
Frequency deviation ΔF 1 in Hz.

0.015
T-IFLC
0.01 DMT-IIFLC
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
Frequency deviation ΔF2 in Hz.

0.015 PI controller
T-IFLC
0.01
DMT-IIFLC
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

x 10-3
8
Change in tie-line (∆Ptie ) in p.u. MW

6 PI controller
T-IFLC
4
DMT-IIFLC
2

-2

-4

-6

-8
0 5 10 15 20 25 30 35 40 45 50
Time in sec

Fig. 15. Performance analysis of frequency deviations DF1, DF2 and tie-line power deviation DPtie of PI controller, T-IFLC, and DMT-IIFLC for a step load change in area-1
(Power system model + CES + AC tie-line).
590 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598

Fuzzification: Translates inputs (real values) to fuzzy values. Application of the dual mode Type-II fuzzy logic controller for
Inference System: To obtain a fuzzy output, fuzzy reasoning interconnected power systems
mechanism is applied.
Type Defuzzificator/Reductor: To transduce one output to The proposed dual mode T-IIFLC design is applied to intercon-
precise values, defuzzificator is used; the type reductor converts nect two area reheat based thermal power systems. As the system
a T-IIFS into a T-IFS. is exposed to a small change in load during its normal operation,
Knowledge Base: It contains data base which consists of set of the linear model will be sufficient for dynamic [34] representation.
fuzzy rules and a membership functions. The two normalized input
variables DACE and DACE_ are first fuzzified by two interval T-IIFS as
Design of a conventional PI controller and a fuzzy logic controller with
‘‘positive” and ‘‘negative”. It is represented by lP(ACE) and lN(ACE),
output feedback
respectively as shown in Fig 11. The primary memberships are gen-
erated by blurring the trapezoidal T-IFS lP(ACE), lN(ACE), lP(DACE)_
The conventional PI controller with output feedback is designed
_
and lN(DACE). The secondary membership functions of the interval with integral square error (ISE) technique. Gain values of the PI
T-IIFS are all constant. The rule base is given in Table 1. controllers are Kp = 6 and Ki = 0. 61 were obtained according to

0.02
PI controller
Frequency deviation ΔF 1 in Hz.

0.015
T-IFLC
0.01 DMT-IIFLC
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
PI controller
Frequency deviation ΔF2 in Hz.

0.015
T-IFLC
0.01 DMT-IIFLC
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

x 10
8
Change in tie-line (∆Ptie ) in p.u. MW

6
PI controller
4 T-IFLC
DMT-IIFLC
2

-2

-4

-6

-8
0 5 10 15 20 25 30 35 40 45 50
Time in sec

Fig. 16. Performance analysis of frequency deviations DF1, DF2 and tie-line power deviation DPtie of PI controller, T-IFLC, and DMT-IIFLC for a step load change in area-1
(Power system model + CES + DC tie-line).
 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598 591

the system response curve method. The conventional PI controls fuzzifier mode A, the controller will act and if e1 > 0.075 Type-II
and FLC are also designed using the method presented in reference fuzzifier mode B, the controller will act. The input variable of the
[35]. The system output is sampled at the normal sampling rate of proposed controller are DACE (error e) and DACE _ (change of error
2 s and the controller output is also updated at a normal sampling Dce). The membership functions for the input variables scheduled
rate. by only three fuzzy sets with the simple shape membership
functions linguistically labelled as LN, MN and SN for N. SP, MP
Design of a proposed dual mode Type-II fuzzy logic controller using the and LP for P. Z for zero distributed over the intervals a and a that
CES unit with output feedback is, [a = (L2  P2  L2  L2 + P2) and a = (L2  P2L2L2 + P2)] as
shown in Fig. 11. Where LN = large negative, MN = medium
The design of the proposed DMT-IIFLC with an AC–DC tie-line negative, SN = small negative, N = negative, SP = small positive,
and a CES unit with the output feedback scheme is applied in the MP = medium positive, LP = large positive, P = positive and for
two areas interconnected power systems. Since the switching limit Z = zero. The value of a for the fuzzifier mode A is 0.01 (0.01,
value ‘e’ should be greater than the steady state error of the system +0.01) and that of a for the fuzzifier mode B is 0.0004 (0.0004,
output DACE with only proportional T-IIFLC, the value of ‘e’ is +0.0004). The output variable l is characterized by three fuzzy sets
chosen as 0.075 (e1 = 0.075 and e2 = 0.075). If e1 6 0.075 Type-II N, Zo and P+ over the interval a and a. The rule base is shown in

0.02
Frequency deviation ΔF 1 in Hz.

0.015 PI controller
T-IFLC
0.01
DMT-IIFLC
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
PI controller
Frequency deviation ΔF2 in Hz.

0.015
T-IFLC
0.01
DMT-IIFLC
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

x 10-3
8
Change in tie-line (∆Ptie ) in p.u. MW

6 PI controller
T-IFLC
4 DMT-IIFLC
2

-2

-4

-6

-8
0 5 10 15 20 25 30 35 40 45 50
Time in sec

Fig. 17. Performance analysis of frequency deviations DF1, DF2 and tie-line power deviation DPtie of PI controller, T-IFLC, and DMT-IIFLC for a step load change in area-1
(Power system model with electric governor + CES + AC–DC tie-line).
592 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598

Table 1. The feedback signal is sampled at the normal sampling rate and tie-line power deviation are plotted with respect to time, as
of 2 s and the control output is also updated at normal sample rate. shown in Fig. 14. The nominal parameters are presented in Appen-
dix A. The simulation results show that the proposed controller has
superior performance in settling time and overshoot/undershoot
Simulation results
when compared to conventional PI controller and T-IFLC. The
results are summarized and given in Table 2.
The DMT-IIFLC with a parallel AC–DC tie-line using CES unit
designed in the previous section is implemented in two area inter-
connected reheat based thermal power systems. The Simulink Performance analysis of the proposed controller under different
block diagram of the system with proposed controller is shown operating conditions
in Fig. 12. The performance of this controller is simulated for
0.01 p.u. MW step load change in area 1 and the corresponding fre- To examine the essence of the proposed controller under differ-
quency deviation DF1 in area 1, frequency deviation DF2 in area 2 ent operating conditions the following simulation is also taken
and tie line power deviation DPtie12 are plotted in Fig. 13. For easy away.
comparison, the responses of DF1, DF2 and DPtie12 of the system
with the optimum PI controller designed on the basis of ISE crite- Case (i): Simulation with a AC tie-line only.
rion and T-IFLC are also plotted in the same figure. Similarly, the
performance of the controller is simulated for 0.01 p.u. MW step In Fig. 12 the DC tie-line is removed and simulation is carried
load change in area 2 and the corresponding frequency deviation out for 1% step load disturbance. The results are shown in Fig 15.

4%
0.4
3%
0.3
2%
0.2
Frequency deviation in p.u. Hz

1%
0.1

-0.1

-0.2 DMT-IIFLC
T-IFLC
Disturbance
-0.3

0 10 20 30 40 50
Time in sec

Fig. 18. Performance comparison of frequency deviation in area-1 for a different load disturbances.

DMT-II FLC
3
T-I FLC
Frequency deviation in Hz

2 PI Controller
Disturbance
1

-1

-2

-3

-4
0 5 10 15 20 25 30
Time in sec

Fig. 19. System performance under unknown deterministic power demand.

 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598 593

0.02
0.015

Frequency deviation ΔF 1 in Hz
With CES
0.01 Without CES
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
With CES
0.015
Frequency deviation ΔF2 in Hz

Without CES
0.01
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

x 10-3
8
Change in tie-line (∆Ptie ) in p.u. MW

6 With CES

4 Without CES

-2

-4

-6

-8
0 5 10 15 20 25 30 35 40 45 50
Time in sec

Fig. 20. Performance analysis of DF1, DF2 and tie-lines power deviation of DPtie with DMT-IIFLC for step load change in area-1 for with and without CES.

Case (ii): Simulation with a DC tie-line only. Simulations are carried out for varying load case also. Here the
load varies in step of 1% for every 10 s. (for the first 10 s the distur-
In Fig. 12 the AC tie-line is removed and simulation is bance is 1% and for the next 10 s the disturbance is 2% and so on) as
carried out for 1% step load disturbance. The results are shown in shown in Fig 18. For an easy comparison the disturbance and out-
Fig 16. put responses are plotted in the same figure.
From the above case studies, it is found that the proposed con-
Case (iii): Simulation with electric governor. troller provides a good transient and steady state response even
under different operating conditions. For convenience the results
In Fig. 12 the mechanical governor has been replaced with elec- are summarized in a tabular form and presented in Table 2.
trical governor as shown in Fig. 9. Simulation is carried out for 1%
step load disturbances in area 1 and the results are shown in Performance of the proposed controller under unknown deterministic
Fig. 17. power demand

Case (iv): Simulation with various step load disturbances for Rapidly changing random-load disturbance is a major problem
different time period. in LFC design. To access the performance of the proposed controller
594 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598

0.02
0.015 With CES

Frequency deviation ΔF 1 in Hz
0.01 Without CES
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
0.015
Frequency deviation ΔF2 in Hz

With CES
0.01 Without CES
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

x 10-3
8
Change in tie-line (∆Ptie ) in p.u. MW

6 With CES
Without CES
4

-2

-4

-6

-8
0 5 10 15 20 25 30 35 40 45 50
Time in sec

Fig. 21. Performance analysis of DF1, DF2 and tie-lines power deviation of DPtie with DMT-IIFLC for step load change in area-2 for with and without CES.

Table 3
Performance of interconnected two-area thermal power systems with and without CES unit.

Operating conditions Settling time in sec. Peak overshoot/undershoot

DF1 in Hz DF2 in Hz DPtie12 in p.u. MW DF1 in Hz DF2 in Hz DPtie12 in p.u. MW
With CES for area-1 10.00 11.00 11.00 +0.0031 +0.003 +0.005
0.021 0.022 0.006
Without CES for area-1 13.00 15.00 14.00 +0.005 +0.008 +0.002
0.0026 0.027 0.065
With CES for area-2 18.00 17.31 14.10 +0.004 +0.005 +0.004
0.023 0.02 0.0008
Without CES for area-2 20.00 24.5 30.00 +0.013 +0.006 +0.0051
0.02 0.023 0.0025
 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598 595

under unknown deterministic power demand, DPdi was taken to be The performance of this controller is simulated for 0.01 p.u. MW
slowly time-varying and given by [36], step load change in area 1 and the corresponding frequency devi-
 ation DF1 in area 1, frequency deviation DF2 in area 2, and tie-
0:01 sinðpt=20Þ 0 < t 6 10
DP di ¼ ð30Þ line power deviation DPtie12 are superimposed in Fig. 20. Similarly,
0:01 t > 10
the controller is simulated for 0.01 p.u. MW step load change in
where sinp = Trigonometric operator. DPdi = Deviation in power area 2 and the corresponding frequency deviation DF1 in area 1,
demand in p.u.-power. frequency deviation DF2 in area 2 and tie-line power deviation
The simulation results of such a case with PI controller, T-IFLC, DPtie12 are plotted in Fig. 21. The results are summarized in Table 3.
and DMT-IIFLC are superimposed as shown in Fig. 19 which clearly Figs. 20 and 21 and Table 3, show the effectiveness of the CES unit.
indicates that the proposed controller is robust than the other two System performance is better with CES unit than without. The
controllers under unknown deterministic power demand. results are summarized and given in Table 3.

Effectiveness of the CES unit Performance analysis of the proposed controller under parameter
variation
To study the effectiveness of the energy storage device, the CES
unit simulations are given in two phases. The performance of the proposed controller has been analyzed
under parameter variation. The parameters Tg1, Tg2, Tt1, Tt2, Tr1
Phase (i): Power system model + controller. and Tr2 are varied by ±20% from the nominal value one at a time,
Phase (ii): Power system model + CES + controller. and simulations are carried out. The simulation results are shown

0.02
Frequency deviation ΔF1 in Hz

0.015 with nominal parameters

with +20% variation in Tg1
0.01
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
Frequency deviation ΔF1 in Hz

with nominal parameters

0.015
with +20% variation in Tt1
0.01
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
Frequency deviation ΔF1 in Hz

with nominal parameters

0.015
with +20% variation in Tr1
0.01
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

Fig. 22. Frequency deviations under parameter variations.

596 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598

0.02
with nominal parameters

Frequency deviation ΔF2 in Hz

0.015
with -20% variation in Tg2
0.01
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
Frequency deviation ΔF2 in Hz

0.015 with nominal parameters

0.01 with -20% variation in Tt2
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

0.02
Frequency deviation ΔF2 in Hz

0.015 with nominal parameters

with -20% variation in Tr2
0.01
0.005
0
-0.005
-0.01
-0.015
-0.02
-0.025
-0.03
0 5 10 15 20 25 30 35 40 45 50
Time in sec

Fig. 22 (continued)

Fig. 23. Block diagram of a realistic power system.

 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598 597

0.01

PI controller
0.005 T-I FLC
DMT-II FLC

Frequency deviation in Hz 0

-0.005

-0.01

-0.015
0 10 20 30 40 50 60
Time in sec

Fig. 24. Performance analysis of frequency deviations for realistic power system of PI controller, T-IFLC, and DMT-IIFLC for 0.01 p.u. MW step load changes.

in Fig. 22. It is clear that, even though the parameters of governor, Appendix A
turbine, and reheat unit are varied by +20% or 20% with the
proposed controller, the variations are marginal. This proves that A.1. Data for the interconnected two-area thermal reheat power
the proposed DMT-IIFLC controller is robust for parameter system
variations.
Pr1 = Pr2 = 1000 MW; KR1 = KR2 = 0.55; R1 = R2 = 2.4 Hz/p.u. MW;
TG1 = TG2 = 0.08 s; TR1 = TR2 = 10.2 s; a12 = 1; DPd1 = 0.01 p.u. MW;
Performance analysis of the proposed controller for realistic power
Tt1 = Tt2 = 0.3 s; KP1 = KP2 = 120 Hz/p.u. MW; TP1 = TP2 = 20 s;
system
b1 = b2 = 0.425 p.u. MW/Hz; 2pT12 = 0.545 p.u. MW/Hz.
To assess the performance of the proposed controller for
practical system, a realistic power system [37,38] is considered. A.2. Data for realistic power system
This system is a combination of multi source generators in a
control area with their corresponding participation factor is more Prt = 2000 MW, PL = 1840 MW; f = 60 Hz; H = 5 MW; D =
realistic for the study of LFC. The simulink block diagram of the (rPL/rf)  (1/Prt) p.u MW/Hz = 0.0153 p.u MW/Hz; Kps = 1/D Hz/p.
system consist of linearized model of governor, reheat-turbines, u MW; Tps = (2H/fD)s; TSG = 0.08 s; RTH = RHY = RG = R = 2.4 Hz/p.
hydro turbines, gas turbines are taken to study the power system u MW; KR = 0.3; TR = 10 s; TW = 1.0 s; TRS = 5 s; TRH = 28.75 s;
as shown in Fig. 23. Simulation is carried out and the frequency aTH = 0.543478; aHY = 0.326084; ag = 0.130438; TGH = 0.2 s;
deviation is plotted as a function of time as shown in Fig. 24. XG = 0.6 s; YG = 1.0 s; Cg = 1; bg = 0.055; TF = 0.23 s; TCR = 0.001 s;
The simulation result show that the proposed DMT-IIFLC has TCD = 0.2 s; DPD = 0.01 p.u.
superior performance when compared to conventional PI con-
troller and T-IFLC. A.3. Data for CES unit

C = 1.0 F; R = 100 X; Edmax = 2.76 kV; Edmin = 0.6 kV; Edo = 2.0 kV;
Conclusions Tdc = 0.05 s; KACE = 70 kA/unit MW; Kvd = 0.20 kA/kV.

This paper presents a design of DMT-IIFLC for interconnected

two area thermal reheat power systems with parallel AC–DC tie-
line using CES system. The proposed controller is designed by tak-
ing advantage of dual mode concept and T-IIFLC. Simulation results
are carried out using MATLAB 2009b version demonstrated that
the designed controller capable to guarantee the robust stability
and robust performance such as precise reference frequency track-
ing and disturbance attenuation under a wide range of parameter
uncertainty and area load conditions. The simulation results indi-
cate that the proposed DMT-IIFLC can guarantee the stability of
the overall system for large parametric uncertainty and achieves
good performance even for a less sensitive to change in the param-
eter of the system.
Acknowledgement
The authors wish to thank the authorities of Annamalai Univer-
sity, Annamalai Nagar, Tamil Nadu, India for the facilities provided
to prepare this paper.
A.4. Data for DC link
Kdc = 1.0; Tdc = 0.5 s.
A.5. Data for electric governor
Kp = 1.0; Ki = 5.0; Kd = 4.0; f = 50 Hz.
References
[1] Kothari ML, Nanda J, Kothari DP, Das D. Discrete-mode automatic generation
control of a two-area reheat thermal system with new area control error. IEEE
Trans Power Syst 1989;4:730–8.
[2] Ramar K, Velusami S. Design of decentralized load-frequency controllers using
pole placement technique. Electr Mach Power Syst 1989;16:193–207.
[3] Barisal AK. Comparative performance analysis of teaching learning based
optimization for automatic load frequency control of multi-source power
systems. Int J Electr Power Energy Syst 2015;66:67–77.
[4] Chaturvedi DK, Satsangi PS, Kalra PK. Load frequency control; a generalized
neural network approach. Int J Electr Power Energy Syst 1999;21:405–15.
Elsevier Sci.
598 N.J. Vinoth Kumar, M.M. Thameem Ansari / Electrical Power and Energy Systems 82 (2016) 579–598
[5] Pandey Shashi Kant, Mohanty Soumya R, Kishor Nand. A literature survey on
load-frequency control for conventional and distribution generation power
systems. Renew Sust Energy Rev 2013;25:318–34.
[6] Shabani Hamed, Vahidi Behrooz, Ebrahimpou Majid. A roboust PID controller
based on imperialist competitive algorithm for load-frequency control of
power systems. ISA Trans 2013;52:88–95.
[7] Farahani M, Ganjefar S, Alizadeh M. PID controller adjustment using chaoit
optimisation algorithm for multi-area load frequency control. IET Control
Theory Appl 2012;6:1984–92.
[8] Fereidouni Alireza, Masoum Mohammad AS, Moghbel Moayed. A new adaptive
configuration of PID type fuzzy logic controller. ISA Trans 2015;56:222–40.
[9] Feng Gang. A survey on analysis and design of model-based fuzzy control
systems. IEEE Trans Fuzzy Syst 2006;14:676–97.
[10] Talaq J, Al-Basri F. Adaptive fuzzy gain scheduling for load frequency control.
IEEE Trans Power Syst 1999;14:145–50.
[11] Zuniga KV, Castilla I, Aguilar RM. Using fuzzy logic to model the behaviour of
residential electrical utility customers. Appl Energy 2014;115:384–93.
[12] Yang L, Entchev E. Performance prediction of a hybrid microgeneration system
using adaptive Neuro-Fuzzy inference system (ANFIS) technique. Appl Energy
2014;134:197–201.
[13] Zadeh LA. Fuzzy logic. Computer 1988;1:83–93.
[14] Zadeh LA. Knowledge representation in fuzzy logic. IEEE Trans Knowl Data Eng
1989;1:89–100.
[15] Melin Patricia, Castillo Oscar. A review on the applications of type-2 fuzzy logic
in classification and pattern recognition. Exp Syst Appl 2013;40:5413–23.
[16] Mendel JM, Mouzouris GC. Type-2 fuzzy logic systems. IEEE Trans Fuzzy Syst
1999;7:643–58.
[17] Castillo Oscar, Melin Patricia. A review on interval type-2 fuzzy logic
applications in intelligent control. Inf Sci 2014;279:615–31.
[18] Karnik NN, Mendel JM. An introduction to Type-2 fuzzy logic systems. Univ of
Southern Calif, Los Angeles, CA; June 1998b.
[19] Mendel JM, John RIB. Type-2 fuzzy sets made simple. IEEE Trans Fuzzy Syst
2002;10:117–27.
[20] Castillo Oscar, Melin Patricia. Type-2 fuzzy logic: theory and applications,
studies in fuzziness and soft computing, vol. 223. Springer; 2008.
[21] Sudha KR, Vijaya Santhi R. Load frequency control of an Interconnected Reheat
Thermal System using type-2 fuzzy system including SMES units. Electr Power
Energy Syst 2012;43:1383–92.
[22] Sudha KR, Vijaya Santhi R. Robust decentralized load frequency control of
interconnected power system with Generation Rate Constraint using Type-2
fuzzy approach. Int J Electr Power Energy Syst 2011;33:699–707.
[23] Tripathy SC. Improved load frequency-control with capacitor energy storage.
Energy Convers Manage 1997;38:551–62.
[24] Tripathy SC, Juengst KP. Sampled data automatic generation control with
superconducting magnetic energy storage in power systems. IEEE Trans
Energy Convers 1997;12:187–92.
[25] Tripathy SC, Balasubramanian R, Chandramohanan Nair PS. Small rating
capacitive energy storage for dynamic performance improvement of automatic
generation control. IEEE Proc-C 1991;138.
[26] Mukherjee V, Ghoshal SP. Application of capacitive energy storage for
transient performance improvement of power system. Electr Power Syst Res
2009;79:282–94.
[27] Tripathy SC, Mishra IP. Dynamic performance of wind-diesel power system
with capacitor energy storage. Energy Convers Manage 1996;37:1787–98.
[28] Banerjee S, Chatterjee JK, Tripathy SC. Application of magnetic energy
storage unit as load frequency stabilizer. IEEE Trans Energy Convers
1990;5:46–51.
[29] Ganapathy S, Velusami S. MOEA based design of decentralized controllers for
LFC of interconnected power systems with nonlinearities, AC–DC parallel tie-
line and SMES units. Energy Convers Manage 2010;51:873–80.
[30] Ibraheem, Nizamudin, Bhatti TS. AGC of two area power system
interconnected by AC/DC links with diverse sources in each area. Int J Electr
Power Energy Syst 2014;55:297–304.
[31] Thameem Ansari MMd, Velusami S. Dual mode linguistic hedge fuzzy logic
controller for an isolated wind-diesel hybrid power system with
superconducting magnetic energy storage unit. Energy Convers Manage
2010;51:169–81.
[32] Velusami S, Chidambaram IA. Decentralized biased dual mode controllers for
load frequency control of interconnected power systems considering GDB and
GRC non-linearities. Energy Convers Manage 2007;48:1691–702.
[33] Sathya MR, Ansari MT. Load frequency control using bat inspired algorithm
based dual mode gain scheduling of PI controllers for interconnected power
system. Int J Electr Power Energy Syst 2015;64:365–74.
[34] Velusami S, Chidambaram IA. Design of decentralized biased dual mode
controllers for load-frequency control of interconnected power systems. Electr
Power Comput Syst 2006;34:1057–75. Taylor & Francis.
_
[35] Cam Ertugrul, Kocaarslan Ilhan. Load frequency control in two area power
system using fuzzy logic controller. Energy Convers Manage 2005;46:233–43.
[36] Miniesy SM, Bohn EV. Optimum load frequency continuous control with
unknown deterministic power demand. IEEE Trans Power Appar Syst 1972;
PAS-91(5):1910–5.
[37] Bevrani Hassan. Robust power system frequency control. New York: Springer;
2009. p. 15–31.
[38] Singh Parmar KP, Majhi S, Kothari DP. Load frequency control of a realistic
power system with multi-source power generation. J Electr Power Energy Syst
2012;42:426–33.