An PMU Double Ended Fault Location Scheme

for Aged Power Cables

El Sayed Tag El Din, Senior Member IEEE, Mahmoud Gilany, Mohamed Mamdouh Abdel Aziz,
Member IEEE, Doaa khalil Ibrahim

current method and acoustic method) is used to get the exact

Abstract-- This paper presents an adaptive fault location location. On the other hand, in the on-line methods, fault
scheme for aged power cable using synchronized phasor location is solved by processing of the recorded information
measurements from both end of the cable line. The proposed of fault current and voltage to obtain the unknown parameter
fault location scheme is derived using the two-terminal (distance). Examples of the on-line methods are A/D
synchronized measurements incorporated with distributed line
techniques [4], fault generated high frequency transient
model, modal transformation theory and Discrete Fourier
Transform. The proposed scheme has the ability to solve the method [5, 6], neural networks method [7, 8], and global
problem of cable changing parameters especially the change of positioning systems [9].
the relative permittivity over its age and thus for the operating The performance of Underground cables depends on a
positive, negative, and zero-sequence capacitance changes. number of parameters that may not remain constant over its
Extensive simulation studies are carried out using Alternative age, and the contributions of each added to the performance
Transients Program ATP/EMTP. The simulation studies show in a complex manner. The main problem in the aging cables is
that the proposed scheme provides a high accuracy in fault the change in the 3-phase positive, negative, and zero-
location calculations under various system and fault conditions. sequence capacitance [10]. This change may reach 110% to
The results show that the proposed method responds very well
insensitive to fault type, fault resistance, fault inception angle
200% or more, leads to an error of about 4.65% to 29.2% in
and system configuration. The proposed scheme solves the fault location estimation methods that use the manufacture
problem of aged cables with change of its electric parameters. In parameters of the cables. The adaptive feature of the
addition to, it gives an accurate estimation of the fault resistance. proposed method would represent a solution for this problem.
Index Terms aged cable, distributed line model, fault location, The Alternative Transients Program (ATP), which is
modal transformation, phasor extraction, power cable lines considered to be one of the widely used universal program
protection. system for digital simulation of transient phenomena of
electromagnetic as well as electromechanical nature in
I. INTRODUCTION electric power systems. With this digital program, complex
networks and control systems of arbitrary structure can be
UNDERGROUND cables are an intricate part of a power
system and are depended upon for reliable transmission simulated. ATP has extensive modeling capabilities and
and distribution service. Fault location on underground cables additional important features besides the computation of
can be a time consuming venture that poses a major transients. Simulation studies are carried out using ATP to
inconvenience to customers. Benefits of accurate fault prove that the proposed scheme provides a high accuracy in
location are the fast repair to restore power system, improving fault location in aged cable lines [11].
system availability and performance, reduction of operating In this paper an adaptive fault location scheme is proposed
costs, and saving time and expense of crew searching in bad using distributed line model and modal transformation theory.
weather and tough terrain. It can be used on-line following the operation of digital relays
Variety of fault location techniques are used to achieve the or off-line using the data stored in the digital transient
aim and the above-specified benefits [1]. While numerous recording apparatus. The adaptive scheme by definition,
techniques are now available to the fault locator, some employ "a protective philosophy which permits and seeks to
techniques can be destructive to the cable, some may require make adjustment to various protection functions in order to
instruments not readily stocked, or some may involve make them more attuned to the prevailing power system
detection methods not familiar to the user. Such challenges conditions"[12], the method proposed has no predetermined
led to the need for a comprehensive fault locator reference fixed data except the cable length, it has the ability to adopt
manual. its response according to cable line parameter changes over
In the field of fault location for the underground cables, off- cable age. The method considers the line cable sections to be
line and on-line methods are available. Off-line methods are transposed. No assumptions were made for fault boundary
divided into two main groups : terminal methods and tracer conditions or fault resistance.
methods [2,3]. Terminal methods (e.g. the voltage drop ratio,
capacitance method and the bridge technique) are applied to II. FAULT LOCATION FOR A SINGLE PHASE CABLE
the cable from one or two terminals and are usually used as To explain the concept used, the paper starts with the fault
preliminary step of fault location. After finding the location problem of a one single phase cable line and then
approximate location of the fault, tracer methods (e.g. tracing extended to a three phase cable line.
 The voltage and current along a single phase cable are the cable power line is divided into two homogeneous parts
functions of the distance from the end of the line as shown in [14,15]. One is line section SF (with a length of L - D); the
figure1. To represent these voltage and current phasors, we other part is the line section FR (with a length of D). These
consider a very small element (dx) in the line and calculate two line sections still can be regarded as perfect cable power
the difference in voltage and in current between the ends of lines. The voltages at any point on the two sections can be
the element. The rise in voltage over the element in the expressed in terms of the voltages and currents measured at
direction of increasing x is dV. This rise in voltage is also the the healthy end of that section. Moreover, at fault point F the
product of the current in the element flowing opposite to the voltages expressed in terms of these two data sets (VS, IS) and
direction of increasing x and the impedance of the element, or (VR, IR) are identical.
zdxI. Thus dV = z dx I . S IS I+dI I IR R
Similarly, the current flowing out the element is I. The + + +
magnitude and phase of the current I vary with distance along
the line because of the distributed shunt admittance along the VS V+dV V VR
line. Thus dI = y dx V .
_ _ _
The differential voltage and current can be expressed as:
dV sending end dx x
= z I receiving end
dx Fig. 1. A single-phase power line cable.
dI (1)
= y V The voltage on the fault position can be expressed as:
dx
By differentiating equations (1) with respect to dx, the
VR + I R Z C D V R I R Z C D
following are obtained[13] VF = e + e
2 2
d 2V dI 2 VS + I S Z C ( D L ) VS I S Z C ( D L ) (5)
= z = z y V = .V VF = e + e
dx 2 dx 2 2
Equations (5) represents the post-fault voltages at fault
d 2I dV 2 (2) point which are expressed in terms of terminal voltages and
2 = y = z y I = .I
d x dx currents (VS, IS, VR and IR). Solving equation (5) for the fault
Where z is the series impedance of the line in /km, or location, the fault distance D measured from the receiving
z=R+jL , y is the shunt admittance of the line in -1/km, or end is:
y=G+jC. The parameters R, L, G, and C are resistance, (N /M )
D = ln
inductance, conductance and capacitance of the cable per unit 2 (6)
length. The propagation constant is given by: where:-
= ( R + j L ) ( G + j C ) = z y VR I R Z C VS I S Z C + L
By solving equations (2) using the known boundary N = e
2 2
conditions at the receiving end, namely voltage VR and current
VR + I R Z C VS + I S Z C +L
IR , the voltage Vx and current Ix at distance x from the M = + e
receiving end is given by: 2 2
VR + I R Z C +x VR I R Z C x
Vx = e + e III. FAULT LOCATION SCHEME FOR THREE PHASE CABLE
2 2
VR Z C + I R +x VR Z C I R x (3) A. Modal Decomposition
Ix = e e
2 2 In order to implement the method described in section II on
By solving equations (2) using the known boundary three phase systems; the phase domain signals are first
conditions at the sending end, namely voltage Vs and current decomposed into their modal components by means of the
Is, the voltage Vx and current Ix at distance x from the modal transformations matrices. The phase impedance and
receiving end is given by: admittance matrices are used to determine the current
+ ( xL) ( xL)
VS +IS ZC VS IS ZC transformation matrix and the voltage transformation matrix.
Vx = e + e For balanced (equally transposed) multiphase lines, both
2 2
matrices can be easily chosen to one of the different real
VS ZC +IS + (xL) VS ZC IS ( xL) (4) value transformation matrices such as Wedepohl,
Ix = e e Karrenbauer , and Clarke transformation. For untransposed
2 2
multiphase lines, an eigen vector based transformation matrix
where Z C = ( R + j L ) /( G + j C ) = z / y is the which is frequency dependent would be used. In this paper
surge impedance of the cable and L is the cable total length. the well-known de-coupling real Clarke transformation matrix
When a fault occurs at the point F, D km away from below is used with its stationary two-phase components (
receiving end R on a cable power line SR shown in figure 1, and ) and a third component called ground or zero sequence
component. Real Clarke transformation matrix can be used Where m is related to modal data not to phasor data, S for
with any transposed power line. [16]:- sending end data, R for receiving end data, i = 1, 2, and 3 for
+1 +1 +1 , , modal components of signals.

T = (1 / 3) + 2 1 1

IV. FAULT LOCATION SCHEME FOR THREE PHASE AGED
0 3 3 CABLE
The samples of phase voltage and current are transformed The performance of underground cables depends on a
into ground mode and two aerial modes values. Now we have number of parameters that may not remain constant over a
three decoupled single phase cable power lines. period, and the contributions of each added to the
V
V IA
I performance in a complex manner. The main parameter is the

A
temperature that leads to a change of the 3-phase positive,
V B = T V IB = T I
negative, and zero-sequence capacitance. The change in the
Then ,
VC V

IC

I
relative permittivity r of the cable over its age and thus for
the operating capacitance C, the capacitance current and the
B. Fault location algorithm earth leakage result in significant errors in fault location
The single phase solution is extended to the three phase estimation. Guiding values for Protodur cables may be
line as following:- obtained from figure2 [10], with the basis value at 20 C =1.
(Ni / M i ) For example at 40 C, the ratio of C to C20 is equal to 1.3,
D i = ln leads to an error in fault location techniques which used the
2 i i =1,2,3 (7)
value of C20 of about 12.3%, Another example at 30 C , the
Where i = 1, 2, and 3 are used to represent , and ratio of C to C20 is equal to 1.15 that lead to a change of the
modal components of signals. Each modal is represented with
velocity of propagation of 1 1 .15 = 0.9325 , leads to an
its own propagation constant i and associated surge
error in fault location of 6.75%, Known that
impedance ZCi. Assuming complete transposition, the surge
r
impedance associated with each mode can be shown to be:- C = 2 r / ln
ri
Z C = ( Z S + 2 Z m ) /( Ys 2 Ym ) = Z / Y
Where is the permittivity of free space 8.8510-12 (F/m),
Z = Z = ( Z S Z m ) /( Ys +Ym ) = Z1 / Y1 r is the relative permittivity of the insulation, r is the outer
C C radius of a coaxial conductor (m), ri is the inner radius of a
where ZCi is the i - mode surge impedance , ZS , YS are the coaxial conductor (m).
average sum of all conductor self impedances and
admittances at any frequency. Zm , Ym are the average sum of
all conductor mutual impedances and admittances at any
frequency, Z ,Y are the zero phase sequence impedance and
admittance and Z1 ,Y1 are the positive phase sequence
impedance and admittance [17]. The modal propagation
constants are similarly given in terms of phase sequence and
impedances and admittances by:-
= Z Y , = = Z 1Y 1

Equation (7) is solved using the quantities Ai , Bi , Ci , Ei ,Ni ,
Mi as in equation (8):- Fig. 2 Change in C and r as a function of temperature for Protodur compounds.
Thus any change of change of the 3-phase positive ,
N i = Ai C i
negative , and zero-sequence capacitance of the cable line
leads to a change in phase sequence admittances thus for a
M i
= Ei Bi
change in propagation constant of each mode and its
Ai =
1
2
( V Rmi I Rmi Z Cmi )
associated surge impedance ,therefore an error in fault
location calculation when applying the above equations [10].
The adaptive feature of the proposed method would represent
Bi =
1
2
( V Rmi + I Rmi Z Cmi ) a solution for this problem in the rest of this section.
(8)
+ L A general schematic description of the proposed fault
Ci =
1
2
e
i ( V Smi I Smi Z Cmi )
locator is given in fig.3. A detailed description of each stage
is described in the following sections.
L
Ei =
1
2
e
i ( V Smi + I Smi Z Cmi )
 Data processing V 2 Smi V 2 Rmi
- Gathering samples of ZC i =
voltages and currents from I 2 Smi I 2 Rmi
both sides
1
- Modal transformation
cosh ( Fmi )
i =
- Phasor extraction DFT Lenght (9)
V Smi I Smi + V Rmi I Rmi
Fmi =
Estimation of cable I Smi I Rmi + I Rmi I Smi
changed parameters Where m is related to the modal data not to phasor data.
C. Appropriate Mode Selection according to fault type
Mode selection When applying the theory of fault location problem, a three
fault locations may be obtained for i =1, 2, 3. However, it is
necessary to choose the accurate fault location by selecting
Fault Location according to the appropriate mode , or . Table I showed the
modal phasors appropriate mode for each fault type. As shown, mode is
valid to deal with all type of faults except the fault of B-C.
Thus an appropriate solution is taking fault distance of
fault resistance calculation modal as a true distance for all fault types except for a B-C
fault. In such a case, it is appropriate to take fault distance of
modal. The proposed scheme assumes that the fault type is
input from another diagnostic block. However, using the
Distance voltage and current vectors calculated at the fault in the next
Fig.3 Schematic diagram for the fault locator section, the fault type may also be identified.
A. Data processing TABLE I
The three phase currents and voltages from both sides are SELECTION OF MODE FAULT DISTANCE ACCOEDING TO FAULT TYPE

collected, sampled, and then fed to a transformation routine Fault type Do D D

as described in section (III-A) to get three decoupled modes AG T T F
, and .
BG or CG T T T
The three modal phasors of currents and voltages are
extracted via Discrete Fourier Transform DFT. ABG , ACG or BCG T T T

B. Estimation of cable changed parameters AB , AC, ABCG F T T

The mathematical model of the cable with its parameters: BC F F T

propagation constant i , its associated surge impedance ZCi Where: - (T) represents the true distance, and (F) false
where i =1, 2, and 3 are used to represent , and modal distance.
components of signals which are derived using one cycle of
the measured voltages and currents signals from both sides D. Fault Location according to modal phasors
before the instant of fault occurrence without using the The calculated quantities in section (IV-B) are used to
manufacturing parameters data of the cable ; by applying the calculate Ai , Bi , Ci , Ei ,Ni and Mi as explained in section (III-
homogenous line equation from both side over the total line B) in equations (8) to calculate Di for i = 2 or 3 (D or D )
length as follows. for modes and according to fault type as mentioned in the
Using equations (3) to get VS and IS as functions of VR , above section with only measuring six currents signals , six
IR, cable surge impedance and propagation constants. Using voltages signals and only known the length of the cable line.
equations (4) to get VR and IR as functions of VS , IS ,cable
surge impedance and propagation constants, where VS, IS, VR, E. Calculation of fault Resistance
and IR are the phasor values of the prefault sending end The proposed fault location algorithm has the ability to
voltage, current and the receiving end voltage and current estimate the fault resistance for various fault types. Using the
respectively calculated using only one cycle before fault. The computed value of Di (fault location distance from the
mentioned four equations are obtained for each modal then be receiving end) , the fault boundary conditions VFi , IFi , where
solved to obtain the propagation constant of each mode and i = 1 , 2 , 3 is used to represent , and modal component
its associated surge impedance as follows: can be obtained as following :-
From the first homogeneous part of the line (length of L-
Di), by applying equations (4) with x = 0, Length =(L Di) to
get VFi and IFiS. From the second homogeneous part of the line
(of length Di), by applying equations (3) with x = Di to get
IFiR. Then we can write: IFi = IFiS + IFiR, where IFiS, IFiR are the when the parameters of the cable are changed over the age.
parts of fault current that are supplied from S and R buses The error in the next tables is calculated as following:-
respectively. Using Clarke transformation to get the voltage Actual fault location Calculated fault
and current vectors at the fault point as follows: % Error = * 100 (11)
Total faulty section length
V The following results show its performance during 40 fault
V I FA I F

FA
F
V FB = T V F I FB = T I F cases.
, (10)
V FC V F
I FC

I F

A. Influence of the Fault Distance
The fault resistance is calculated according to fault type [18]: To evaluate of influence of the fault location on the
V FA algorithm accuracy different cases of phase-to- ground faults
For A to ground fault : RF = I with 10 ohm of fault resistance at various locations were
FA
considered. The results are presented in table III.
V FB
For B to C to ground fault : R F = I FB + I FC B. Influence of the Fault Type
To analyze the influence of the fault type, ten different
VFB VFC faults, with a fault distance of 10 km from the receiving end,
For B to C fault : R F = I FB I FC 0 inception angle and 1 fault resistance. The results in table
IV show that the accuracy is not affected by fault type.
V. SIMULATED CASE STUDY TABLE III
An underground cable of 30 km for 10 kV applications is FAULT DISTANCE INFLUENCE ON ALGORITHM ACCUARCY
Actual Fault Calculated Percentage
applied on the system shown in figure 4, the cable is Location in km Fault error
represented by its series resistance, series inductance and Location in km
shunt capacitance, all are assumed to be distributed uniformly 2.5 2.5072 0.024
along the line measured in per unit length. The conductance 3.5 3.5093 0.031
5 4.9979 0.007
of the line is neglected. Table II gives nominal values of cable
7.5 7.5030 0.010
parameters. 9.5 9.5081 0.027
11.5 11.5122 0.0407
10 KV , 50 Hz 14 13.9929 0.0237
10 mH 0.06 / km , .019 mh / km , 16 16.0068 0.0227
25MVA source 21.2 21.2138 0.046
2 f / km
25.7 25.6820 0.060
30 km 27.5 27.4829 0.057
Load of
TABLE IV
Fault 16 MW
ESTIMATED LOCATION WITH THE INFLUNCE OF FAULT TYPE
.8 pf
Fault Exact Calculated Percentage
Fig. 4. System under Study. type distance km distance km error %
AG 10 10.023 .0767
TABLE II BG 10 10.0326 .10867
VALUES OF CABLE PARAMETERS CG 10 10.126 .42
Cable Positive and Zero ABG 10 10.0080 .0267
parameter negative sequence value ACG 10 10.0620 .2067
sequence values BCG 10 10.0048 .016
R in ohm/km .06 1.03 AB 10 10.0041 .0137
BC 10 9.9984 .0053
L in mH/km 1.9 2.451 AC 10 10.0624 .208
C in F/km 2 0.764 ABCG 10 10.0443 .147

C. Influence of the Fault Resistance

Extensive simulations were carried out using the To estimate the fault resistance influence, different fault
Alternative Transients Program. The voltage and current types were considered with fault distance of 12 km from the
signals from both sides are determined, at different fault sending terminal, the fault resistance varied from 10 to 50
locations ranging from 2 to 28 km away from the receiving and the results are shown in table V. Table VI represents the
end. Different fault types and different fault inception angles ability of the locator to estimate the fault resistance under
are also studied. The fault impedance varied from 0 to 100 different fault types at different fault locations.
ohm.
D. Influence of the inception angle
VI. SIMULATION RESULTS AND ALGORITHM EVALUATION. In the analysis of the inception angle influence, cases
The performance of the scheme was tested to check corresponding to double phase to ground, with fault distance
whether the protection scheme is suitable for determine an at 23 km from the sending terminal with inception angle
accurate fault location in different fault conditions especially varied from 0 to 90O. The results are shown in table VII.
E. Influence of the cable parameters changes with .9 4.9989 .0036
temperature and age. 1.2 4.9901 .033
1.8 4.9982 .006
The evaluation of influence of the cable parameters change 2.5 4.9844 .052
in the algorithm accuracy is examined by changing the 3 4.9853 .049
positive and zero sequences capacitances of the cable to
TABLE IX
simulate any change on cable temperature as discussed above ESTIMATED PARAMETERS IN COMARISON WITH ACTUAL MODIFIED PARAMETERS
on section (IV). Parameter C/ C20
In this analysis, cases corresponding to single phase to 1.2 3 5
ground faults were considered, with fault distance at 5 km Actual Zc 10.6085 6.7094 5.1971
from the receiving terminal, varying the system capacitance Estimated Zc 10.6117 6.7105 5.1985
Actual Zc 10.6085 6.7094 5.1971
from .9 to 3 times of the nominal capacitance at 20O (C20). Estimated Zc 10.5998 6.7021 5.1986
The results presented in table VIII show that the method Actual c 0.0080 0.0126 0.0163
accuracy is practically the same irrespective to any Estimated c 0.0080 0.0127 0.0162
parameters changes for aging power cables. And that is the
most significant advantage of the algorithm proposed. VII. CONCLUSION
Moreover, the ability of the proposed scheme to detect the
In this paper a fault location scheme is proposed for aged
change of the parameters with age and estimate the modified
power cables using synchronized phasor measurements from
parameters is shown in table IX.
both end of the cable. The proposed scheme estimates the
TABLE V
changed parameters of aged cables. Using the post fault
ESTIMATED LOCATION WITH THE INFLUNCE OF FAULT RESISTANCE phasor measurements; it calculates the fault location using the
Fault Fault Calculated Percentage two aerial modals of the cable line.
type resistance distance km error % The proposed scheme showed also its ability for adaptation
ohm
by accurate estimation of cable parameters using only one
Single 10 11.9972 .0093
phase to 20 12.0087 .029 cycle of double ended voltages and currents before fault and
ground 50 12.0127 .0423 the cable length. In other words, the algorithm results are not
sensitive to the variations in the parameters values of the
Double 10 12.0234 .078
cable power lines over its age. So this method can be applied
phase to 20 12.0384 .128
ground 50 12.0047 .0156 on any cable with double ended measurements of voltages
and currents without any predetermined cable data except the
Three 10 12.0427 .1423 cable length. The simulation results show that the proposed
phase to 20 11.9998 .00067
method responds very well insensitive to fault type, fault
ground 50 12.0427 .1423
distance, fault resistance, fault inception angle and system
configuration. Accurate results for all fault conditions were
TABLE VI evaluated, as the maximum percentage error in the simulated
ESTIMATED FAULT RESISTANE
fault cases is .235 % of the 30 km cable (about 70 meters).
Fault RF Fault at 10 km Fault at 20km
type Moreover, the proposed method introduced a good method
Estimated % Error Estimated % Error
for estimating the value of fault resistance for all fault types
RF RF through resistance.
SLG 10 10.1516 1.516 11.1017 11.017
20 19.8907 0.546 20.8418 4.2090 VIII. REFERENCES
100 98.3750 1.625 99.4753 0.5247
[1] M. Kezunovic, B. Perunicic, Fault Location, Wiley Encyclopedia
DLG 10 9.9378 0.622 10.8912 9120
of Electrical and Electronics Terminology, Vol. 7, pp. 276-285, John
30 29.0624 3.125 30.0719 0.2397
Wiley, 1999.
LL 15 15.0001 0.01 15.0019 0.0190 [2] C. M. Wiggins, D. E. Thomas, T. M. Salas, F. S. Nickel, and H. W.
Ng, A novel concept for URD cable fault location, IEEE Trans.
Power Delivery, vol. 9,no. 1, pp. 591597, January 1995.
TABLE VII [3] S. Potivejkul, P. Kerdonfag, S. Jamnian and V. KinnaresN, Design
ESTIMATED LOCATION WITH THE INFLUNCE OF THE INCEPTION ANGLE
of a low voltage cable fault detector, Power Engineering Society
Inception angle Calculated Percentage Winter Meeting, 2000. IEEE , pp. 724-729 Vol. 1, 2000.
degree distance km error [4] C. M. Wiggins, D. E. Thomas, T. M. Salas, F. S. Nickel, and H. W.
0 23.0706 .235 Ng, On-line fault location system for 66kv underground cables with
30 23.0239 .08 fast O/E and A/D technique IEEE Trans. Power Delivery, vol.
45 23.0355 .1183 9,no. 1, pp. 579-584, January. 1994.
60 23.0582 .194 [5] Z.Q Bo, R.K Aggarwal and A.T Johns, "A very accurate fault
90 23.0706 .235 location and protection scheme for power cable using fault generated
high frequency voltage transients", Proceedings of the 8th Electo-
TABLE VIII tecnical Conference, Industrial Applications in power systems
ESTIMATED LOCATION WITH THE INFLUNCE OF THE CABLE PARAMETERS ,MELECON 96, Bari Italy, pp.777-780 , May 1996 .
CHANGES [6] E.M. Tag Eldin, M.I. Gilany, M.M. Abdel Aziz and D.K. Ibrahim,
C/ C20 Calculated Percentage A Wavelet-Based Fault Location technique for Aged Power Cable
distance km error Lines, accepted for presentation in the IEE International Conference
 On Reliability Of Transmission and Distribution Networks, London Ph. D. in High voltage Engineering from Cairo University in 1996 and 2000,
February 2005. respectively. During 1993- 2000 he was a demonstrator and research assistant in
[7] M. R, Haghifam, S. Hosseini, A.A. Yaghti, Fault location in Cairo University. In October 2000 he became an assistant professor with Cairo
underground power cables using artificial neural networks, University, while continuing his research.
Proceedings of the Second IASTED International Conference on Dr Tag Eldin is currently contributing in a UNDP/GEF Project for Energy
Power and Energy Systems (EUROPES), Crete, Greece, pp. 634- Efficiency Improvement and Greenhouse Gas Reduction and in a World Bank
639, June 2002. Project in Higher Education Development in Egypt. His research interests
[8] M. T. Glinkowski and N. C. Wang, ANNs pinpoint underground include electric fields near power lines, risk analysis, harmonics, power quality,
distribution faults, IEEE Computers applications in power, pp. 31 distributed generation, digital protection and digital signal processing.
34, October 1995.
[9] K. Amano, H. Kon and Y. Kumada, Fault Location System for Mahmoud Gilany was born in Egypt in 1964.He graduated from the Cairo
Power Cable Using Global Positioning System - FLAG -, Fujikura University, Faculty of Engineering, Electrical Power and Machines Department
Technical Review No. 31 , January 2002. in 1986 and received the M. Sc. from Cairo University in 1989 and received the
[10] H. Lother , Power cables and their applications ,Published by PhD. in Electrical Eng., from Calgary University, Alberta, Canada in the field of
Siemens Aktiengesellschaft, Berlin and Mucher 1979. Microprocessor Based Relays 1993. During 1986-1989 he was a demonstrator
[11] Alternative Transient Program, Rule Book, 1987. in Cairo University. In 1993 became a teaching position as an assistant
[12] CIGRE Working Group 34.02,"Adapive protective and Control"- professor in Cairo University. From May 1999 till now he is an Associated
CIGRE final report, August 1995. Professor in Cairo University .Research interests include power system
[13] H. Saadat, Power system analysis, WCB/McGraw-Hill Book protection, Digital relaying, Area of generator.
Company, 1999.
[14] J. A. Jiang , J. Z. Yang, Y. H. Lin, C. W. Liu, and J. C. Ma, An Mohamed Mamdouh Abdel Aziz (M80) received the B.Sc. (Hons.),
adaptive PMU based fault detection/location technique for M.Sc., and Ph.D. degrees in electrical power and machines, Cairo University,
transmission Part I: Theory and algorithms, IEEE Trans. Power Giza, Egypt, in 1970, 1972, and 1975, respectively. Currently, he is a Professor
Delivery, Vol. 15, pp.486493, April 2000. in the Department of Electrical Power and Machines at Cairo University. He
[15] J. A. Jiang , C. W. Liu, and C. S .Chen , A Novel Adaptive PMU- was an Instructor in the Department of Electrical Power and Machines, Cairo
Based Transmission-Line RelayDesign and EMTP Simulation University, from 1970 to 1972. He was also a Teaching Assistant in the
Results , IEEE Trans. Power Delivery, vol. 17, no. 4, pp. 930937, Department of Electrical Power and Machines at Cairo University from 1972 to
October 2002. 1975. Dr Abdel Aziz is the author or coauthor of many referred journal and
[16] L. S. Martine, J. F. Martins, V. Fernao Pires and C.M. Alegria, The conference papers. His areas of research include cables, contact resistance,
application of neural networks and Clarke -Concordia transformation harmonics, power quality, photovoltaic systems, and wind energy systems.
in fault location in distribution power systems, IEEE/PES T&D
Transmission and Distribution 2002 Asia Pacific, pp. 2091-2095, Doaa khalil Ibrahim was born in Egypt in Dec 1973. She graduated from
October 2002. the Cairo University, Faculty of Engineering, Electrical Power and Machines
[17] A. T. Johns and S. K. Salman , Digital protection for power Department in 1996 and received the M. Sc. in digital protection from Cairo
systems , Peter Peregrinus Ltd 1995 . University in 2001. From 1996 till now she was a demonstrator and research
[18] Jr. William D. Stevenson, Elements of Power system analysis, assistant in Cairo University. Research interests include utilization and
McGraw-Hill Book Company, third edition 1975. generation of electric power and digital protection.

BIOGRAPHIES
El Sayed Tag El Din (Senior Member, IEEE, Member IEE) was born in
Egypt. He is graduated from the Cairo University, Faculty of Engineering,
Electrical Power and Machines Department in 1993 and received the M. Sc. and