Professional Documents
Culture Documents
i i i
= + =
(1)
2) Calculate ' c through equation (2)
=
=
= =
N
1 i
i
N
1 i
i i
N i
x
) ,..., c( c'
(2)
3) Calculate K value, so it fulfills equation (3)
1 k k
x c' x
+
(3)
4) Calculate " c through equation (4)
+ = =
+ = =
+
+
=
N
1 k i
i
A
k
1 i
i A
N
1 k i
i
A
i
k
1 i
i A i
) (x ) (x
) (x x ) (x x
c"
(4)
5) Check, if c' " c
l
= than stop. If not, set " c c'
l
= then return
to step number 2.
The c
r
determination
1) Initialization
i
, through equation (1)
2) Calculate ' c through equation (2)
3) Calculate k value, so it fulfills equation (3)
4) Calculate " c through equation (5)
+ = =
+ = =
+
+
=
N
1 k i
i A
k
1 i
i
A
N
1 k i
i A i
k
1 i
i
A
i
) (x ) (x
) (x x ) (x x
c"
(5)
5) Check, if c' " c
r
= than stop. If not, set " c c'
r
= then
return to step number 2.
Centroid Calculation
The c
l
and c
r
are important in centroid calculation. The
centroid can be obtained through equation (6).
Centroid = (c
l
+c
r
)/2 (6)
Even though centroid calculation is an iteration process, the
number of iteration will not exceeds N, where N equals the
number of sampled values of the primary variable [11].
B. Interval Type2 Fuzzy Logic PSS
The model of IT2FLPSS refers [2] with a type development
on fuzzy logic controller block, and to simplify model
integrator block is also removed. IT2FLPSS model is as shown
in Fig. 8.
Fig. 8 IT2FLPSS Model
Fig. 8 shows that IT2FLPSS consists of blocks, those are
wash-out, gain (Kw, Kp, Ks1), and limiter. Each part is
PROCEEDINGS OF WORLD ACADEMY OF SCIENCE, ENGINEERING AND TECHNOLOGY VOLUME 31 JULY 2008 ISSN 1307-6884
PWASET VOLUME 31 JULY 2008 ISSN 1307-6884 686 2008 WASET.ORG
International Journal of Electrical and Electronics Engineering 3:10 2009
596
explained as follows:
1) Wash Out
This block consists of two wash-out filter with time
constant 10 second.
2) Gain [Kw Kp Ks1]
Gain is needed to normalize the input and output of fuzzy
logic controller.
3) Fuzzy Logic Controller
Fuzzy Logic Controller has a function to produce control
signal as an output appropriate with the input.
4) Limiter [Vsmin Vsmax]
Limiter gives limitation of the PSS output.
In order to set fuzzy logic controller, input membership
function, output membership function, rules, and gain tuning
are definitive. Detail of input membership function, output
membership function, rules, and gain tuning are as follows:
Input Membership Function
Each of input variables are classified by Input membership
function consists of trapezoidal membership function and
triangular membership function, described in Fig. 9.
-6 -4 -2 0 2 4 6
0
0.5
1
Fig. 9 Input Membership Function
The linguistic variable of input membership shown in Fig. 9
can be described as follows:
IBN_UMF =[-5-0.25 -1 -0.2 0.1+0.25 ];
IBN_LMF =[-5+0.25 -1 -0.2 0.1-0.25 ];
IMN_UMF =[-1-0.25 -0.15 0.2+0.25 ];
IMN_LMF =[-1+0.25 -0.15 0.2-0.25 ];
ILN_UMF =[-1-0.25 -0.065 1+0.25 ];
ILN_LMF =[-1+0.25 -0.065 1-0.25 ];
IZ_UMF =[-1-0.25 0 1+0.25 ];
IZ_LMF =[-1+0.25 0 1-0.25 ];
ILP_UMF =[-1-0.25 0.065 1+0.25 ];
ILP_LMF =[-1+0.25 0.065 1-0.25 ];
IMP_UMF =[-0.2-0.25 0.15 1+0.25 ];
IMP_LMF =[-0.2+0.25 0.15 1-0.25 ];
IBP_UMF =[-0.1-0.25 0.2 1 5+0.25 ];
IBP_LMF =[-0.1+0.25 0.2 1 5-0.25 ];
Output Membership Function
Output membership function also consists of triangular
membership function, described in Fig. 10.
-1.5 -1 -0.5 0 0.5 1 1.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Fig. 10 Output Membership Function
The linguistic variable of output membership shown in Fig.
10 can be described as follows:
OBN_UMF =[-1-0.005 -0.51 -0.439+0.005 ];
OBN_LMF =[-1+0.005 -0.51 -0.439-0.005 ];
OMN_UMF =[-0.515-0.005 -0.439 -0.302+0.005];
OMN_LMF =[-0.515+0.005 -0.439 -0.302-0.005];
OLN_UMF =[-0.439-0.005 -0.302 0+0.005 ];
OLN_LMF =[-0.439+0.005 -0.302 0-0.005 ];
OBZ_UMF =[-0.302-0.005 0 0.302+0.005 ];
OBZ_LMF =[-0.302+0.005 0 0.302-0.005 ];
OLZ_UMF =[-0.0015-0.0005 0 0.0015+0.0005 ];
OLZ_LMF =[-0.0015+0.0005 0 0.0015-0.0005 ];
OLP_UMF =[0-0.005 0.302 0.439+0.005 ];
OLP_LMF =[0+0.005 0.302 0.439-0.005 ];
OMP_UMF =[0.302-0.005 0.439 0.515+0.005 ];
OMP_LMF =[0.302+0.005 0.439 0.515-0.005 ];
OBP_UMF =[0.439-0.005 0.515 1+0.005 ];
OBP_LMF =[0.439+0.005 0.515 1-0.005 ];
Rules
Rules used by IT2FLPSS are identical with T1FLPSS [2].
The Rules are shown in Table III.
TABLE III
FUZZY LOGIC RULES
dw\dPa BP MP LP Z LN MN BN
BN BZ LN MN MN BN BN BN
MN LP BZ LN MN MN BN BN
LN MP LP BZ LN LN MN BN
Z BP MP LP LZ LN MN BN
LP BP MP LP LP BZ LN MN
MP BP BP MP MP LP BZ LN
BP BP BP BP MP MP LP BZ
Gain Tuning
The gains of the proportional and derivative actions of the
IT2FLPSS are given by the following relations:
K
PR
= Ks1
(F{Kp} + F{Kw})
K
DER
= Ks
1
F{K
P
}
With K
PR
, K
DER
, and F{} are proportional gain, derivative
gain, and fuzzy operation, respectively. To obtain the gains
Kw, Kp, and Ks1 a two steps method has been used. These
two steps consist of adjusting Kp and Kw in order to normalize
input and then tuning Ks1 to obtain best result. The value of
Kw, Kp, and Ks1 are 0.55, 50 x Kp, and 3, respectively.
PROCEEDINGS OF WORLD ACADEMY OF SCIENCE, ENGINEERING AND TECHNOLOGY VOLUME 31 JULY 2008 ISSN 1307-6884
PWASET VOLUME 31 JULY 2008 ISSN 1307-6884 687 2008 WASET.ORG
International Journal of Electrical and Electronics Engineering 3:10 2009
597
IV. RESULT AND ANALYSIS
To see the performance of PSS on electrical power system,
fault simulation is demonstrated with connecting a 3 phase-
ground fault block at the middle of transmission line (Fault
Point A) for 8 cycles, in normally cleared condition and
cleared with one line outage condition.
A. Fault Normally Cleared
In this simulation, Kundur Test System is demonstrated
under a fault simulation and then cleared without any line
outage. While fault occurred in electrical power system, it can
deviates power armature that can cause swing on generator,
affect generator voltage output and area power transfer level
until the fault is cleared [1]. When fault is cleared, system will
oscillate to return on its stable point. Electrical power transfer
level from Area1 to Area2, voltage deviation response at G1,
and power armature deviation response at G1 are shown in
Fig. 11 until Fig. 13.
0 200 400 600 800 1000 1200 1400 1600 1800 2000
150
200
250
300
350
400
450
500
550
600
time (1/100 s)
M
e
g
a
W
a
t
t
(
M
W
)
Electrical Power Transfer Level
dw PSS
MB PSS
T1FLPSS
IT2FLPSS
Fig. 11 Electrical Power Transfer Response on Fault Normally
Cleared Simulation
0 200 400 600 800 1000 1200 1400 1600 1800 2000
0.9
0.95
1
1.05
1.1
1.15
time (1/100 s)
p
e
r
u
n
i
t
(
p
u
)
Voltage Output at G1
dw PSS
MB PSS
T1FLPSS
IT2FLPSS
Fig. 12 Voltage Output Response on Fault Normally Cleared
Simulation
0 200 400 600 800 1000 1200 1400 1600 1800 2000
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
time(1/100 s)
p
e
r
u
n
i
t
(
p
u
)
Power Armature at G1
dw PSS
MB PSS
T1FLPSS
IT2FLPSS
Fig. 13 Power Armature Response on Fault Normally Cleared
Simulation
B. Fault Cleared with One Line Outage
In this simulation, Kundur Test System is demonstrated
under a fault simulation and then cleared with opening breaker
on line which fault occurred. Disconnecting one of two tie-line
transmission lines can change the area power transfer level
into single-line power transfer level [1]. System will oscillate
to its new stable point, during that time system parameters will
deviate. Power transfer from Area1 to Area2, voltage deviation
response at G1, and power armature deviation response at G1
are observed and shown in Fig. 14 until Fig. 16.
0 200 400 600 800 1000 1200 1400 1600 1800 2000
150
200
250
300
350
400
450
500
time(1/100 s)
M
e
g
a
W
a
t
t
(
M
W
)
Electrical Power Transfer Level
dw PSS
MultiBand PSS
T1FLPSS
IT2FLPSS
Fig. 14 Electrical Power Transfer Response on Fault Cleared With
One Line Outage Simulation
PROCEEDINGS OF WORLD ACADEMY OF SCIENCE, ENGINEERING AND TECHNOLOGY VOLUME 31 JULY 2008 ISSN 1307-6884
PWASET VOLUME 31 JULY 2008 ISSN 1307-6884 688 2008 WASET.ORG
International Journal of Electrical and Electronics Engineering 3:10 2009
598
0 200 400 600 800 1000 1200 1400 1600 1800 2000
0.9
0.95
1
1.05
1.1
1.15
time(1/100 s)
p
e
r
u
n
i
t
(
p
u
)
Voltage Output at G1
dw PSS
MB PSS
T1FLPSS
IT2FLPSS
Fig. 15 Voltage Output Response on Fault Cleared With One Line
Outage Simulation
0 200 400 600 800 1000 1200 1400 1600 1800 2000
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
time (1/100 s)
p
e
r
u
n
i
t
(
p
u
)
Power Armature at G1
dw PSS
MB PSS
T1FLPSS
IT2FLPSS
Fig. 16 Power Armature Response on Fault Cleared With One Line
Outage Simulation
To evaluate the overshoot value and settling time clearly,
the data of PSS performances are shown on Table IV and
Table V.
TABLE IV
FAULT NORMALLY CLEARED
Power
Transfer Level
Voltage
Output at G1
Armature Power
at G1
PSS
Type
overshoot Settling
time
overshoot Settling
time
overshoot Settling
time
dw PSS
137.27 3.54 0.1112 2.68 0.3322 2.98
MB PSS
158.6083 14.34 0.1360 2.40 0.3321 3.26
T1FLPSS
143.0345 5.88 0.1197 2.56 0.3257 3.14
T2FLPSS
126.3580 5.36 0.1056 2.31 0.3257 2.81
TABLE V
FAULT CLEARED WITH ONE LINE OUTAGE
Power Transfer
Level
Voltage
Output at G1
Armature Power
at G1
PSS
Type
overshoot Settling
time
overshoot Settling
time
overshoot Settling
time
dw PSS
74.7314 >20 0.1222 >20 0.3322 3.61
MB PSS
63.7735 10.42 0.1511 6.53 0.3321 3.20
T1FLPSS
71.2886 10.87 0.1346 9.29 0.3257 2.92
T2FLPSS
68.4138 8.07 0.1246 6.39 0.3257 2.75
V. CONCLUSION
The IT2FLPSS proposed in this paper has been verified by
means of digital simulations of a Kundur Test System in two
fault simulations. From the simulations IT2FLPSS shows
better performance than dw PSS, Multi Band PSS and
T1FLPSS. For power transfer response IT2FLPSS able to
return system to its stable points at 5.36 second with overshoot
value 126.35 MW in fault normally cleared simulation and it
achieves its new stable point at 8.07 second with overshoot
value 68.41 MW in fault cleared with one line outage.
For voltage output at G1 response IT2FLPSS able to return
system to its stable points at 2.31 second with overshoot value
0.1056 pu in fault normally cleared simulation and it achieves
its new stable point at 6.39 second with overshoot value
0.1246 in fault cleared with one line outage.
For armature power at G1 response IT2FLPSS able to
return system to its stable points at 2.81 second with overshoot
value 0.3257 pu in fault normally cleared simulation and it
achieves its new stable point at 2.75 second with overshoot
value 0.3257 in fault cleared with one line outage.
REFERENCES
[1] Imam Robandi, Modern Power System Design (Desain System
Tenaga Modern, in Bahasa Indonesia), Andi Offset Publisher: 2006.
[2] M. Dobrescu, I. Kamwa, A New Fuzzy Logic Power System Stabilizer
Performances, IEEE, 2004.
[3] P. Kundur, Power System Stability and Control, McGraw-Hill,1993.
[4] Timothy J. Ross, Fuzzy Logic with engineering Applications,
McGraw-Hill, 1997.
[5] ...., Fuzzy Logic Toolbox for Use with MATLAB, The Mathworks,
2002.
[6] QilianLiangand, Jerry M. Mendel, Interval Type-2 Fuzzy Logic System
Theory and Design, IEEE, October, 2000.
[7] Jerry M. Mendel, Robert I.Bob John, Type-2 Fuzzy Sets Made
Simple, IEEE, April, 2002.
[8] Juan R. Castro, Oscar Castillo, Interval Type-2 Fuzzy Logic for
Intelligent Control Applications, IEEE, 2007.
[9] . , MATLAB help version 7.1.0.246 (R14) Service Pack 3, 2005.
[10] . , Fuzzy Logic Toolbox for use with MATLAB users guide,
version 2, 1998.
[11] Jerry M.Mendel, Feilong Liu, Super-Exponential Convergence of the
KarnikMendel Algorithms for Computing the Centroid of an Interval
Type-2 Fuzzy Set, IEEE, April 2007.
PROCEEDINGS OF WORLD ACADEMY OF SCIENCE, ENGINEERING AND TECHNOLOGY VOLUME 31 JULY 2008 ISSN 1307-6884
PWASET VOLUME 31 JULY 2008 ISSN 1307-6884 689 2008 WASET.ORG
International Journal of Electrical and Electronics Engineering 3:10 2009
599
LIST OF ABREVIATION
Abbreviation Contraction
IBN_UMF Input Big Negative Upper Membership Function
IBN_LMF Input Big Negative Lower Membership Function
IMN_UMF Input Medium Negative Upper Membership Function
IMN_LMF Input Medium Negative Lower Membership Function
ILN_UMF Input Low Negative Upper Membership Function
ILN_LMF Input Low Negative Lower Membership Function
IZ_UMF Input Zero Upper Membership Function
IZ_LMF Input Zero Lower Membership Function
ILP_UMF Input Low Positive Upper Membership Function
ILP_LMF Input Low Positive Lower Membership Function
IMP_UMF Input Medium Positive Upper Membership Function
IMP_LMF Input Medium Positive Lower Membership Function
IBP_UMF Input Big Positive Upper Membership Function
IBP_LMF Input Big Positive Lower Membership Function
OBN_UMF Output Big Negative Upper Membership Function
OBN_LMF Output Big Negative Lower Membership Function
OMN_UMF Output Medium Negative Upper Membership Function
OMN_LMF Output Medium Negative Lower Membership Function
OLN_UMF Output Low Negative Upper Membership Function
OLN_LMF Output Low Negative Lower Membership Function
OBZ_UMF Output Big Zero Upper Membership Function
OBZ_LMF Output Big Zero Lower Membership Function
OLZ_UMF Output Low Zero Upper Membership Function
OLZ_LMF Output Low Zero Lower Membership Function
OLP_UMF Output Low Positive Upper Membership Function
OLP_LMF Output Low Positive Lower Membership Function
OMP_UMF Output Medium Positive Upper Membership Function
OMP_LMF Output Medium Positive Lower Membership Function
OBP_UMF Output Big Positive Upper Membership Function
OBP_LMF Output Big Positive Lower Membership Function
BN Big Negative
MN Medium Negative
LN Low Negative
Z Zero
LP Low Positive
MP Medium Positive
BP Big Positive
BZ Big Zero
LZ Low Zero
PROCEEDINGS OF WORLD ACADEMY OF SCIENCE, ENGINEERING AND TECHNOLOGY VOLUME 31 JULY 2008 ISSN 1307-6884
PWASET VOLUME 31 JULY 2008 ISSN 1307-6884 690 2008 WASET.ORG
International Journal of Electrical and Electronics Engineering 3:10 2009
600