Fuzzy Sets and Systems 142 (2004) 293 306

www.elsevier.com/locate/fss

A fuzzy multiple linear regression based loss formula in

electric distribution systems
Ying-Yi Honga; , Zuei-Tien Chaoa , Miin-Shen Yangb
a

Department of Electrical Engineering, Chung Yuan Christian University, Chung Li 32023, Taiwan
Department of Applied Mathematics, Chung Yuan Christian University, Chung Li 32023, Taiwan

Received 4 September 2002; received in revised form 22 February 2003; accepted 31 March 2003

Abstract
Estimation of energy loss is an essential task for both operation and planning in electric distribution systems.
The energy losses are considered to be uncertain due to measurements. A new method based on fuzzy-c-number
clustering and fuzzy multiple linear regression analysis is proposed for developing energy loss formulas to
estimate losses in this paper. The proposed method is implemented with three stages. A part of Taipower
distribution system in Taipei is used for illustrating the performance of the proposed method.
c 2003 Elsevier B.V. All rights reserved.


Keywords: Loss formula; Fuzzy multiple linear regression; Fuzzy clustering; Electric distribution systems

1. Introduction
The purposes of energy loss estimation in distribution systems include conductor size design,
substation expansion, rate analysis, loss allocation to customers, energy conversion analysis, and
control strategy development, etc. (see [6]). Therefore, exploration of an energy loss (kW h) formula
is very essential for a utility in the viewpoint of both operation and planning.
There were many existing literatures related to loss estimation for distribution systems (see
[15,710,14]). A sophisticated model for loss estimation was proposed in [14]. An empirical loss
factor equation considering the load composition and various time periods was presented in [7]. An
energy formula further considering constant voltages, power factors and resistors was developed in
[8]. Capacitors and taps were considered as independent (control) variables in an energy loss formula
for the sake of developing proper control strategies in [1]. Unbalanced voltages and currents were
used as independent variables for a loss formula in [4]. A measurement based loss formula was

Corresponding author. Tel.: +886-3-2652300; fax: +886-3-2652399.

E-mail address: yyhong@dec.ee.cycu.edu.tw (Y.-Y. Hong).

c 2003 Elsevier B.V. All rights reserved.

0165-0114/$ - see front matter 
doi:10.1016/S0165-0114(03)00143-X

294

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

proposed for residential wiring system in [5]. A simpliCed loss model for distribution systems was
presented in [3]. Other papers [2,9,10] were also related to loss formulas.
On the other hand, regression is a method for Ctting variables to a function, which may be linear,
nonlinear or multiple. Regression is the determination of coeEcients of the Ctting function. Similarly, fuzzy regression is developed for Ctting fuzzy variables into a fuzzy or real function because
engineers usually use fuzzy linguistic sentences to describe engineering phenomena [15,16,18]. The
fuzzy multiple linear regression method was proposed in [15,16] for forecasting physical quantities.
A new method based on fuzzy clustering incorporated with fuzzy regression to be the cluster-wise
fuzzy regression was presented by Yang and Ko [18]. In the power engineering area, the peak load
was forecasted using fuzzy regression in [13].
It is known that the loss values can be modeled with fuzzy numbers because the system load and
energy source values are uncertain due to measurements [12,13]. In this paper, therefore, a threestage method based on [18] and fuzzy-c-number (FCN) clustering [17] is proposed for obtaining
proper energy loss formulas. The reason for adopting [18] is that this method avoids linear programming used in [15,16]. The FCN algorithm is used Crst to partition the 24-h energy loss pattern
into three clusters covering several segments where each cluster contains several time periods. The
three clusters are designed for the peak-, medium-peak and oI-peak losses. Then each segment is
further clustered into several sub-clusters, which are Ctted by fuzzy multiple linear regression. The
independent variables are the kW h and kVARh loads in this paper.
A realistic distribution system in Taipower system of Taiwan Power Company is studied in this
paper. From the case study results, it is found that the proposed method can eEciently estimate the
system kW h losses for engineers.
2. Characteristics of distribution system losses
The distribution losses are more nonlinear and diversiCed compared with the transmission losses.
A part of the distribution (supply) system for Taipower, including two 161=69 kV and 12 69=11:95
=11=6:9 kV transformers, is illustrated in Fig. 1 for depicting the above phenomenon. This supply
system, Song-Shan area, is a part of the whole Taipei distribution system.
Song-Shan Supply area is a P/S (primary system) fed from six 161=69 kV lines. There are 12
distribution transformers (69=11=11:95=6:9 kV; 25 MVA) connected to four S/S (secondary systems),
namely Nei-Hu, Ming-Shen, Chung-Nun, and Hsin-Ya. Generally, there are two peak loads daily
and therefore the energy loss pattern is a two-peak pattern. Table 1 shows the daily kW h loss range
for diIerent periods in the Song-Shan supply system. As shown in Table 1, a single formula is
inadequate to express the kW h losses in this system. Actually, there are diIerent load characteristics in various hours daily. For example, rotating machine loads are dominated in the peak-load
hours while there are TV and light loads in the oI-peak load hours (at night). There exist some
transition periods between the peak load and oI-peak load periods; therefore, medium-peak load
hours should be considered. Hence, a partition process is required for this loss pattern. As shown
in Table 1, there are seven segments in a day: (1, 2), (2224), (37), (1216), (1921), (811),
and (17,18).
After the three periods (peak-load, medium-peak load, and oI-peak load) are identiCed
(clustered), one should further examine the loss characteristics in each segment. The realistic

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

295

Fig. 1. Song-Shan supply system in Taipei.

measurement data show that the kW h loss characteristics are still very nonlinearly diversiCed. Because the system load and energy source values are uncertain due to measurements, this motivates
one to cluster the losses in each segment and then develop the loss formula with fuzzy multiple linear
regression.

296

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

Table 1
Daily kW h loss pattern for a supply system

OI-peak load
Medium-peak load
Peak load

kW h loss range

Hours

6
613
13

(1, 2), (2224)

(37), (1216), (1921)
(811), (17, 18)

3. The proposed method

There are three stages in the proposed method for developing the proposed kW h loss formula as
follows:
Stage one: This stage includes a clustering process. The FCN algorithm is used to partition the
daily load loss pattern into three periods covering some segments (peak-load, medium-peak load,
and oI-peak load) based on the observations in Section 2.
Stage two: Each segment is further partitioned into several sub-clusters. This clustering will incorporate with the fuzzy multiple linear regression.
Stage three: An optimal sub-cluster number will be identiCed for each segment.
The fuzzy multiple linear regression in the proposed method is extended from the fuzzy linear
regression proposed in [18]. For engineers, an adopted method should be simple, e2cient and feasible
for the problem characteristics. The reasons for using FCN and developing fuzzy multiple linear
regression based on [17,18] are as follows:
(i) Both FCN and the proposed fuzzy multiple linear regression method are based on Lagrangian.
The Crst derivative to Lagrangian may lead to optimum. The proposed method is much simpler,
compared with [15,16] using linear programming.
(ii) Both FCN and the proposed fuzzy multiple linear regression method are based on LR-type fuzzy
numbers. It is feasible to model the kW h loss by the LR-type fuzzy number: the measured
value corresponds to the mean (center) value and the uncertain factors are implemented with
left/right spreads.
(iii) Both FCN and the proposed fuzzy multiple linear regression method are e2cient: the CPU time
is small and the accuracy is acceptable (as described in Section 4).
(iv) The most important factor for considering the proposed fuzzy multiple linear regression method
is that the kW h loss characteristics are still very nonlinearly diversiCed and scattered in
each segment (see Section 4). Further partitioning in each segment is needed. The proposed
fuzzy multiple linear regression method is a cluster-wise approach that obtains the regression
coeEcients and deals with sub-clustering in a single stage. This proposed method is very suitable
for dealing with the loss characteristics in the problem.
3.1. Stage one: partitioning the daily loss pattern
The FCN algorithm is used to partition a set of fuzzy numbers into several clusters [17]. The
FCN algorithm can be summarized as follows:

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

297

According to Yang and Ko [17], we consider a set of LR-type fuzzy numbers LR = {X1 ; X2 ; : : : ; Xn }
([19], also see Appendix A). A FCN clustering objective function on the data set LR is as follows:
J (; W ) =

n

c

im (Xj )d2LR (Xj ; Wi );

(1)

j=1 i=1

where
d2LR (Xj ; Wi ) = (mxj mwi )2 + ((mxj lxj ) (mwi lwi ))2
+((mxj + rxj ) (mwi + rwi ))2 ;

i = 1; : : : ; c; j = 1; : : : ; n

(2)

and Xj = (mxj ; xj ; xj )LR ; j = 1; 2; : : : ; n are LR-type fuzzy numbers with mxj , xj and xj representing
the mean value, left spread and right spread, respectively. Wi = (mwi ; wi ; wi )LR , i = 1; 2; : : : ; c, are
the cluster centers.  = (1 ; 2 ; : : : ; c ) is a fuzzy c-partition with ij = i (Xj ),
the membership of Xj
c
in ith cluster. Since the fuzzy c-partition (1 ; 2 ; : : : ; c ) has the restriction
i=1 i (Xj ) = 1 for all
j = 1; : : : ; n, the following Lagrangian should be considered:
 c

c
n





m
2
L(; W; ) =
i (Xj )dLR (Xj ; Wi ) 
i (Xj ) 1 :
(3)
j=1 i=1

i=1

The symbols l and r in Eq. (2) are deCned in Appendix A. The weighting exponent m is larger
than 1. One can take the Crst derivatives of Eq. (3) with respect to all parameters and set them equal
to zero. The necessary conditions for achieving the minimization of the FCN objective function J
are obtained as follows:
n
m
j=1 i (Xj )(3mxj + l(wi xj ) + r(xj wi ))

mwi =
;
(4)
3 nj=1 im (Xj )
n
m
j=1 i (Xj )(mwi mxj + lxj )

;
(5)
 wi =
l nj=1 im (Xj )
n
m
j=1 i (Xj )(mxj mwi + rxj )

(6)
wi =
r nj=1 im (Xj )
and
(d2 (Xj ; Wi ))1=(m1)
ij = i (Xj ) = c LR 2
:
1=(m1)
k=1 (dLR (Xj ; Wk ))

(7)

One may examine the value of ij = i (Xj ) to determine the relation between Xj and the ith cluster.
An example about the kW h loss classiCcation of the Song-Shan supply system is provided in
Appendix B for FCN.
In this paper, the fuzzy numbers LR = {X1 ; X2 ; : : : ; Xn } are the kW h losses. The value of n is
24. The clustering number is 3 for the peak, medium-peak and oI-peak loss levels. Once the daily
losses are clustered, a two-peak loss pattern may be partitioned into several segments. Table 1 shows
seven segments in three clusters.

298

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

3.2. Stage two: fuzzy multiple linear regression

Only one segment is discussed in this section since the same treatment is implemented for all
segments. The purpose of the fuzzy multiple linear regression is to Ct the system energy loads in
various hours to the energy losses in a segment including several hours. In each segment, 1, 2 or 3
loss formulas (sub-clusters) are developed depending the performance of regression.
Yang and Ko [18] had proposed a cluster-wise (simple) fuzzy regression analysis. Let G be the set
of observations (Xj ; Yj ); j = 1; : : : n, where Xj = (mxj ; xj ; xj )LR and Yj = (myj ; yj ; yj )LR are LR-type
fuzzy numbers. Suppose that these observations are heterogeneous and come from c simple linear
regression models,
Yj = a0i + a1i Xj ;

i = 1; : : : ; c; j = 1; : : : ; n;

(8)

where a0i and a1i are unknown real coeEcients. Let ij = i (Xj ) be the membership of Xj in ith
cluster. A cluster-wise fuzzy regression objective function proposed by Yang and Ko [18] is as
follows:
n

c


ijm d2LR (a0i + a1i Xj ; Yj );
(9)
JR (a0i ; a1i ) =
j=1 i=1

where d2LR is the same as deCned in (2) of Section 3.1. However, in this paper the kWh and kVAR h
loads are used to be the independent variables. Thus, the objective function (9) needs to be extended
to the multiple linear regression model with K= 2:
Yj = a0i +

K

aki Xkj ;

i = 1; : : : ; c; j = 1; : : : ; n:

(10)

k=1

Since the loss characteristics in the three clusters (7 segments) are still diversiCed and are found
to be scattered, it is necessary to cluster each segment into sub-clusters. The proposed method incorporates the fuzzy number clustering with the multiple fuzzy regression in a stage. Thus, the proposed
cluster-wise multiple fuzzy regression objective function for K= 2 is then deCned as follows:
JMR (; a0 ; a1 ; a2 )
n

c


ijm d2LR (a0i + a1i Xij + a2i X2j ; Yj )
=
=

j=1 i=1
n



ijm (a0i
j=1

+ a1i mx1j + a2i mx2j myj )2



+ S1i (a0i + a1i mlx1j + a2i mlx2j mlyj )2 + (a0i + a1i mrx1j + a2i mrx2j mryj )2


+ S2i (a0i + a1i mlx1j + a2i mpx2j mlyj )2 + (a0i + a1i mrx1j + a2i mqx2j mryj )2


+ S3i (a0i + a1i mpx1j + a2i mlx2j mlyj )2 + (a0i + a1i mqx1j + a2i mrx2j mryj )2


+ S4i (a0i + a1i mpx1j + a2i mpx2j mlyj )2 + (a0i + a1i mqx1j + a2i mqx2j mryj )2 ;

(11)

where m1 is the weighting exponent as the index of fuzziness; a0i ; a1i and a3i are unknown real
number coeEcients; S1i = 1; S2i = S3i = S4i = 0 if (a1i 0 and a2i 0), S2i = 1; S1i = S3i = S4i = 0 if

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

299

(a1i 0 and a2i 0), S3i = 1; S1i = S2i = S4i = 0 if (a1i 0 and a2i 0), and S4i = 1; S1i = S2i = S3i = 0 if
(a1i
0 and 
a2i 0). Let L(; a0 ; a1 ; a2 ; ) be the Lagrangian with L(; a0 ; a1 ; a2 ; ) = JMR (; a0 ; a1 ; a2 )
+ nj = 1 j ( ci=1 ij 1). Set the Crst derivatives of L with respect to all parameters equal to zero.
Then, the n fuzzy numbers can be partitioned into c clusters and the coeEcients a0i , a1i and a2i
of the fuzzy multiple linear regression can be also obtained. The minimum of JMR can be obtained
through the Crst derivative of L and through 4 sets of solutions for (i) a1i 0 and a2i 0, (ii) a1i 0
and a2i 0, (iii) a1i 0 and a2i 0, and (iv) a1i 0 and a2i 0, respectively. In case of a1i 0 and
a2i 0, the necessary conditions for a minimizer (ij ; a0i ; a1i ; a2i ) of JMR are obtained as follows:
n
m
j=1 ij (mtyj a1i mtx1j a2i mtx2j )

;
(12)
a0i =
3 nj=1 ijm
n
a1i =

j=1

n
a2i =

j=1

ijm (mllrr1j a0i mtx1j a2i mlxrxj )

n
;
m 2
j=1 ij mtx1j

(13)

ijm (mllrr2j a0i mtx2j a1i mlxrxj )

n
;
m m
j=1 ij mtx2j

(14)

(d2LR (a0i + a1i Xj + a2i X2j ; Yj ))1=m1

;
2
1=m1
k=1 (dLR (a0k + a1k X1j + a2k X2j ; Yj ))

ij = c

i = 1; : : : ; c; j = 1; : : : ; n:

(15)

The symbols ms with diIerent subscripts in the above equations are deCned in Appendix C. The
update equations for the cases (ii), (iii) and (iv) can be obtained as similar to Eqs. (12)(15).
For example, suppose that there are 30 energy loss patterns (30 days) available and they are
similar. Hence, there are 305 sets of {X1j ; X2j ; Yj } for developing c loss formulas in the 4th
segment as illustrated in Table 1. We mention that the proposed Stage two is extended from [18].
The diIerences between them can be described as follows:
(i) The proposed fuzzy multiple-linear regression method has three unknowns a0i ; a1i and a2i for
the coeEcients; the method in [18] is related to the fuzzy single linear regression only and the
unknowns are a0i and a1i .
(ii) There are four diIerent conditions for the solutions in the proposed method due to the signs
of a1i and a2i . However, there are only two conditions, positive or negative values for a1i , in
[18].
(iii) Two sets of closed form solutions can be derived in [18]; in other words, the solutions can be
obtained without iterations. On the contrary, the relation among a0i ; a1i and a2i shown in Eqs.
(12)(15) are coupling and iterative steps are required to obtain the solutions. In this paper,
the GaussSeidel method is used.
3.3. Stage three: determination of optimal c
The determination of c value in stage one is trivial. It can be based on the daily loss pattern
and c = 3. However, the determination of optimal c in each segment of individual segment (peak,

300

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

medium-peak, and oI-peak) should be further explored. This paper uses the approach in [11] to
obtain an optimal c. The R2 (c) value deCned in [11] is used to identify the optimal c for each
segment. The optimal c has the property of R2 (c) being the closest to unity. This concept is the
same as that adopted %2 in the traditional regression techniques. The optimal c is the one with
the maximum of R2 (c). R2 (c) is deCned by the weighted mean of the determination coeEcients
R2i ; i = 1; : : : ; c, as follows:
2

R (c) =

c

wi R2i ; i = 1; : : : ; c;

(16)

i=1

where
[Ci (Y; X )]2
;
Ri =
Vi (X )Vi (Y )

n

m
j=1 ij
wi = c n

n
Ci (Y; X ) =

ijm mllrrj


3 nj=1 ijm
j=1

Vi(Y ) =

m 2
j=1 ij mtyj
n
3 j=1 ijm

(17)

j=1

n


ijm mtyj nj=1 ijm mxj
;
2

n
m
3

j=1 ij

j=1

 n

n

ijm

i=1

m
j=1 ij mtyj
n
3 j=1 ijm

(18)

2
:

(19)

It is noted that the criterion R2 (c) for the determination of optimal c could be a good validity index
in the cluster-wise fuzzy regression because of the determination coeEcients R2i being commonly
used as the measure of goodness in the regression model Ctting.
4. Case study results
The Song-Shan regional supply system described in Section 2 is studied in this section. The
weekday measurement data from January 10 to 14, 2000, are used for obtaining the loss formulas
while those from January 17 to 21, 2000, are used for verifying the accuracy of the loss formulas.
The FCN algorithm is used Crst to partition the daily energy loss pattern into three clusters
(peak, medium-peak and oI-peak levels). As shown in Table 1, these three clusters include 7
segments. However, diIerent segments in the same cluster are found to have discriminatory loss
characteristics. More speciCcally, 8th11th (6th segment), and 17th and 18th (7th segment) hours are
in the same cluster (peak loss); however, these two segments have discriminatory loss characteristics
resulting from diIerent customer kW h consumption behavior. Consequently, the loss formulas for
each segment should be developed individually.
The 6th segment covering the 8th11th hours is used to demonstrate the loss fuzziCcation here.
There are 20 sets of data for regression because there are 5 weekdays and 4 operation conditions
(hours) in the 6th segment. The maximum (minimum) value among these 20 energy losses is the
right (left) bound for each measured value. The measured value is considered as the mean value.

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

301

Fig. 2. Energy loss w.r.t. real and reactive loads.

Table 2
Peak-loss formulas for the 6th and 7th segments
Segments

Hours

Peak-loss formulas (kW h)

Optimal c

8, 9
10, 11

Loss = 2:3076 + 0:1763 kW h + 0:0010 kV A R h

Loss = 45:6893 0:2667 kW h + 0:0017 kV A R h

17, 18

Loss = 21:8293 0:0478 kW h + 0:0011 kV A R h

Fig. 2 shows the kW h losses with respect to the kW h and kV A R h loads for the 6th segment.
It can be found the losses are not clustered together in a group: one sub-cluster is for hours 8 and
9 and the other sub-cluster is for hours 10 and 11. Stage two and Stage three, therefore, are used
to develop the kW h loss formula for each segment. The peak loss formulas are shown in Table 2.
There are at least three reasons to indicate that the results shown in Table 2 are feasible to
engineers:
(i) It is expected that the number of the loss formulas is as small as possible and the form of
the formula is as simple as possible to engineers. In the results, there are 16 (4, 9 and 3
for the oI-peak, medium-peak and peak losses, respectively; rather than 24) concise formulas
obtained for the 7 segments. Each segment is veriCed by R2 (c) to develop optimal number of
loss formulas. The engineer can locate a corresponding formula for each hour easily.
(ii) The algorithm is implemented in a PC with Pentium III 450. The CPU time for computing the
coeEcients of a formula is less 5 s on the average in MATLAB 5.3.
(iii) The accuracy of the proposed method is acceptable. The approach for calculating the accuracy
is described as follows:

302

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

Table 3
Errors for the 7 segments
Segments

Min. error

Ave. error

Max. error

1
2
3
4
5
6
7

0.0188
0.0020
0.0059
0.0069
0.0229
0.0203
0.0043

0.0453
0.0138
0.0174
0.0266
0.0290
0.0404
0.0176

0.1116
0.0358
0.0483
0.0658
0.0377
0.0748
0.0336

The realistic data from January 17 to 21, 2000, are served for veriCcation of the accuracy for
all loss formulas. Let (T 2; T 1; T 3)LR and (F2; F1; F3)LR be the true (measured) and estimated fuzzy
energy losses, respectively. The assessment index is deCned as follows:

Error

(T 1 F1)2 + (T 2 F2)2 + (T 3 F3)2

:
(T 1)2 + (T 2)2 + (T 3)2

(20)

Table 3 shows the minimum, average and maximum errors for the 7 segments. It is found that
the maximum error occurred in the 1st segment. The average errors for the rest of the segments
are less than 5%; most of them are close to 2%. The accuracy could be improved by taking more
measured data into account for the fuzzy multiple linear regression.

5. Conclusions
A three-stage method based on the FCN algorithm and the fuzzy multiple linear regression
is proposed for developing energy loss formulas in distribution systems. The FCN is adopted
to partition the daily loss pattern into some segments. The three-stage method is presented because the loss pattern is scattered. The optimal cluster number in a segment can be veriCed by
R2 (c) values. A typical distribution system with realistic operational data in the Song-Shan regional
supply area, Taipei, Taipower is used to validate the performance of the proposed method. We
Cnd that the proposed three-stage method can well estimate the system kW h losses in electric
distribution systems.

Appendix A.
An LR-type fuzzy number can be expressed with (m; ; )LR where m; , and  represent the
mean, left spread and right spread, respectively [19]. If an LR-type fuzzy number is expressed with

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

the triangle membership function T , then

mx

L(x) = 1
for x 6 m;

T (x) =
x

for x m
R(x) = 1

and the variables l and r used in Eq. (2) are deCned as follows [17]:
 1
L1 (x) dx;
l=


r=

1
0

R1 (x) dx;

303

(A.1)

(A.2)
(A.3)

where L1 (x) and R1 (x) are the inverse functions for L(x) and R(x), respectively.
Appendix B.
This appendix provides an example for the solution process of FCN. Let LR = {X1 ; X2 ; : : : ; Xn }
be a set of LR-Type fuzzy numbers. LR includes the fuzziCed kW h loss X1 ; X2 ; : : : ; Xn and n = 24.

0:533 1:693 1:693

5:206 1:454 1:454

6:059 1:411 1:411

6:341 1:373 1:373

6:571 1:331 1:331

9:381 1:368 1:368

10:347 1:467 1:467

16:812 1:604 1:604

19:237 1:931 1:931

16:803 2:184 2:184

15:850 2:221 2:221

9:776 2:376 2:376

LR =
:
8:546 2:291 2:291

12:117 2:272 2:272

12:006 2:290 2:290

12:061 2:313 2:313

15:742 2:381 2:381

15:253 2:599 2:599

8:638 2:621 2:621

8:455 2:571 2:571

7:528 2:519 2:519

4:634 2:377 2:377

1:545 2:145 2:145

1:684 1:936 1:936 LR

304

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

Suppose that C = 3; m = 2; , = 0:001. Let h = 0 and the element values of (0) are given randomly
as follows:

(0)

0:2900
0:2296

0:1977

0:1832

0:1694

0:5262

0:2244

0:3686

0:3617

0:3688

0:3724

0:7742
=
0:0434

0:3881

0:3876
0:3879

0:3729

0:3751

0:0937

0:0363

0:0804

0:2442

0:2836
0:2825

0:2756
0:2080
0:1755
0:1612
0:1479
0:2991
0:7661
0:3966
0:3816
0:3966
0:4055
0:2048
0:0317
0:4935
0:4997
0:4964
0:4066
0:4121
0:0671
0:0270
0:0664
0:2232
0:2678
0:2666

0:4344
0:5624

0:6268

0:6556

0:6827

0:1747

0:0095

0:2347

0:2567

0:2346

0:2221

0:0210
;
0:9249

0:1184

0:1127
0:1156

0:2205

0:2127

0:8392

0:9367

0:8532

0:5326

0:4486
0:4509

w(0) = [2:201

1:729

1:957];

w(0) = [1:891

2:169

2:060]:

(1) (1)
Then (m(1)
w ; w ; w ) will be:

m(1)
w = [11:147

11:790

7:322];

w(1) = [2:180

1:837

2:078];

w(1) = [1:974

2:131

2:147]

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

305

and

(1)

0:2328
0:1064

0:0596

0:0435

0:0312

0:4497

0:7373

0:3738

0:3743

0:3739

0:3699

0:5808
=
0:1647

0:0832

0:0526
0:0701

0:3692

0:3649

0:1948

0:1407

0:0051

0:1322

0:2176
0:2152

0:2039
0:0845
0:0456
0:0328
0:0232
0:2270
0:2105
0:4910
0:4514
0:4911
0:5157
0:2443
0:1000
0:9131
0:9455
0:9271
0:5188
0:5351
0:1166
0:0869
0:0035
0:1069
0:1880
0:1855

0:5633
0:8091

0:8948

0:9237

0:9456

0:3233

0:0522

0:1352

0:1744

0:1350

0:1145

0:1749
;
0:7353

0:0037

0:0019
0:0029

0:1120

0:1000

0:6886

0:7724

0:9915

0:7609

0:5945
0:5993

When h = 32;

(32)

0:0573
0:3159

0:5643

0:6448

0:7062

0:9874

0:9134

0:0053

0:0738

0:0044

0:0038

0:9686
=
0:9935

0:5867

0:6132
0:6001

0:0063

0:0261

0:9943

0:9885

0:9035

0:1784

0:0207
0:0163

0:0162
0:0356
0:0445
0:0438
0:0419
0:0062
0:0545
0:9930
0:8977
0:9943
0:9951
0:0172
0:0024
0:3468
0:3204
0:3334
0:9920
0:9673
0:0022
0:0041
0:0232
0:0242
0:0052
0:0040

0:9265
0:6485

0:3912

0:3113

0:2519

0:0064

0:0321

0:0017

0:0285

0:0014

0:0011

0:0142

0:0041

0:0666

0:0664
0:0665

0:0017

0:0066

0:0035

0:0074

0:0734

0:7975

0:9741
0:9797

and
|(31) (32) | = 0:0007 ,; the algorithm stops:
It can be found that X3 X7 , X12 X16 and X19 X21 have strong relation in the 1st cluster (column)
due to their larger membership values compared with the values of the other columns. Moreover,
the membership values of X8 X11 and X17 ; X18 are close to unity in the second column. Finally,
X1 ; X2 ; X22 X24 are found to be in the third cluster because they have membership values of 0.9265,
0.6485, 0.7975, 0.9741 and 0.9797, respectively. The result is summarized in Table 1.

Appendix C.
mx1 lx1 = mlx1 ; mx1 + lx1 = mpx1 ;
mx1 + rx1 + mrx1 ; mx1 rx1 = mqx1 ;
mx1 + mlx1 + mrx1 = mtx1 ; mx1 + mpx1 + mqx1 = mox1 ;
m2x1 + m2lx1 + m2rx1 = m2tx1 ; m2x1 + m2px1 + m2qx1 = m2ox1 ;

306

Y.-Y. Hong et al. / Fuzzy Sets and Systems 142 (2004) 293 306

mx2 lx2 = mlx2 ; mx2 + lx2 = mpx2 ;

mx2 + rx2 = mrx2 ; mx2 rx2 = mqx2 ;
mx2 + mlx2 + mrx2 = mtx2 ; mx2 + mpx2 + mqx2 = mox2 ;
m2x2 + m2lx2 + m2rx2 = m2tx2 ; m2x2 + m2px2 + m2qx2 = m2ox2 ;
mx1 my + mlx1 mly + mrx1 mry = mllrr1 ;
mx1 my + mpx1 mly + mqx1 mry = mplqr1 ;
mx2 my + mlx2 mly + mrx2 mry = mllrr2 ;
mx2 my + mpx2 mly + mqx2 mry = mplqr2 ;
mx1 mx2 + mlx1 mlx2 + mrx1 + mrx2 = mlxrx ;
mx1 mx2 + mpx1 mlx2 + mqx1 + mrx2 = mplqr ;
mx1 mx2 + mpx1 mpx2 + mqx1 mqx2 = mpxqx :
References
[1] R. Baldick, F.F. Wu, Approximating formulas for the distribution system: the loss function and voltage dependence,
IEEE Trans. Power Systems 6 (1) (1991) 252259.
[2] P. Caramia, P. Verde, D. Menniti, Evaluation of the losses in unbalanced three-phase power systems, in: Proc.
Electrotechnical Conf., 1996, MELECON96, 8th Mediterranean, Vol. 2, 1996, pp. 715 720.
[3] C.S. Chen, J.C. Hwang, M.Y. Cho, Y.W. Chen, Development of simpliCed loss models for distribution system
analysis, IEEE Trans. Power Delivery 9 (3) (1994) 15451551.
[4] H.D. Chiang, J.C. Wang, K.N. Miu, Explicit loss formula, voltage formula and current Tow formula for large-scale
unbalanced distribution systems, IEEE Trans. Power Systems 12 (3) (1997) 10611067.
[5] W.C. Chu, B.K. Chen, P.H. Yu, The measurement and economic evaluation of the residential wiring system losses,
IEEE Trans. Power Systems 9 (3) (1994) 16791683.
[6] T. Gonen, Electric Power Distribution System Engineering, McGraw-Hill, New York, 1986.
[7] M.W. Gustafson, Demand, energy, and marginal electric system losses, IEEE Trans. Power Apparatus and Systems
PAS-102 (9) (1983) 31893195.
[8] M.W. Gustafson, Approximating formulas the system losses equation, IEEE Trans. Power Systems 4 (3) (1989)
850855.
[9] IEEE PES Transformers Committee, IEEE loss evaluation guide for power transformers and reactors, IEEE Std
C57.120-1991.
[10] Insulated Conductors Committee, Task Group 7-39, Cost of Losses, Loss evaluation for underground transmission
and distribution system cable systems, IEEE Trans. Power Delivery 5(4) (1990) 16521659.
[11] K. Jajuga, Linear fuzzy regression, Fuzzy Sets and Systems 20 (1986) 343353.
[12] P.A. Mastorocostas, J.B. Theocharis, A.G. Bakirtzis, Fuzzy modeling for short-term load forecasting using the
orthogonal least-square method, IEEE Trans. Power Systems 14 (1) (1999) 2936.
[13] J. Nazarko, W. Zalewski, The fuzzy regression approach to peak load estimation in power distribution systems, IEEE
Trans. Power Syst. 14 (3) (1999) 809814.
[14] D.I.H. Sun, S. Abe, R.R. Shoults, M.S. Chen, P. Eichenberger, D. Farris, Calculation of energy losses in a distribution
system, IEEE Trans. Power Apparatus Syst. PAS-99 (4) (1980) 13471356.
[15] H. Tanaka, H. Ischibuch, Possibilistic regression analysis based on linear programming, in: J. Kacprzyk, M. Fedrizzi
(Eds.), Fuzzy Regression Analysis, Physica-Verlag, Heidelberg, 1992, pp. 4760.
[16] H. Tanaka, S. Uejima, K. Asai, Linear regression analysis with fuzzy model, IEEE Trans. Systems Man Cybernet.
12 (6) (1982) 903906.
[17] M.S. Yang, C.H. Ko, On a class of fuzzy c-number clustering procedures for fuzzy data, Fuzzy Sets and Systems
84 (1996) 4960.
[18] M.S. Yang, C.H. Ko, On cluster-wise fuzzy regression analysis, IEEE Trans. Systems Man Cybernet. Part B:
Cybernet. 27 (1) (1997) 113.
[19] H.J. Zimmerman, Fuzzy Set Theory and its Application, Kluwer Academic Publishers, Dordrecht, 1991.