Você está na página 1de 9

www.ietdl.

org

Published in IET Generation, Transmission & Distribution


Received on 16th November 2009
Revised on 24th March 2010
doi: 10.1049/iet-gtd.2009.0648

ISSN 1751-8687

Protection scheme for transmission lines


based on alienation coefficients
for current signals
M.E. Masoud M.M.A. Mahfouz
Electrical Power and Machines Department, Helwan University, Helwan, Egypt
E-mail: mmahfouz@abcelectric.com.eg

Abstract: In modern digital power system protection systems, statistical coefficients technique is recently used for
fault analysis. An alienation technique is developed for faults detection and discrimination. The proposed
technique is able to accurately identify the condition of phase(s) involved in all ten types of shunt faults that
may occur in extra high-voltage transmission lines under different fault resistances, inception angle and
loading levels. The proposed technique does not need any extra equipment as it depends only on the three
line-currents measurements which are mostly available at the relay location. This technique is able to perform
the fault detection, type and phase selection in about a half-cycle period. Thus, the proposed technique is
well suited for implementation in digital protection schemes. The proposed methodology is applied for a part
of 500 KV Egyptian network. Alternative transient program and MATLAB programs are used to implement the
proposed technique.

1 Introduction knowledge of the experts, which is often quite time-


consuming to be obtained. Fuzzy and fuzzy neural
In modern power system protection, accurate, fast and network-based approaches also require extensive training of
reliable fault detection and selection techniques are an the ANN [6]. Recently, the fault selection techniques
important operational requirement. On the one hand, suitable for a fault recorder have been proposed in [7],
correct information of the type of fault is readily needed for which can identify all ten types of short-circuit faults. In
fault location algorithms [1]. On the other hand, in digital this paper, a fault detection and selection based on
distance protection schemes, for proper operation of the alienation technique is proposed which is able to determine
protective relays, correct determination of the fault type is a accurately, during one cycle of fundamental frequency, all
prerequisite. A significant amount of research work has fault types and phase selection. Also, this technique takes
been directed to address the problem of an accurate fault into consideration the wide variations of operating
selection scheme. Among the various techniques reported conditions such as switching, fault resistance and fault
for fault selection in transmission lines, artificial neural inception angle.
network (ANN) approach is one of the commonly used
techniques. Although the neural network-based approaches
have been quite successful in determining the correct fault
type, the main disadvantage of ANN is that it requires a
2 Fault selection strategy
considerable amount of training effort for good The fault selection algorithm is based on the alienation
performance, especially under a wide variation of operating technique of two half successive cycles with the same
conditions such as, system loading level, fault resistance and polarity. For transmission lines protection, this method
fault inception instance [2 – 5]. Similarly, expert system- needs only three line-current measurements available at the
based approach effectiveness depends largely on the domain relay location.

1236 IET Gener. Transm. Distrib., 2010, Vol. 4, Iss. 11, pp. 1236 – 1244
& The Institution of Engineering and Technology 2010 doi: 10.1049/iet-gtd.2009.0648
www.ietdl.org

2.1 Alienation coefficients calculation 1. A transition is detected if DI . 20%(In), as In is the line


nominal current.
The alienation coefficient (A) is estimated as follows for any
two dependant variables, y1(x) and y2(x) [8] 2. Fault confirmation and faulty phase selection are done
according to the following sequences:
A = (1 − r 2 ) (1)
† Three-phase current alienation coefficients values are
where r is the correlation coefficient and is calculated from calculated. If fault is detected, phase current alienation
the following values are sorted in ascending order and compared.
N
s
Sx=1 y1 (x)y2 (x) − Nsy1y2 The possible fault cases are
r =   (2)
Ns 2 Ns 2
(Sx=1 y1 (x) − Ns (y1 )2 )(Sx=1 y2 (x) − Ns (y2 )2 ) † If the three-phase alienation coefficients are nearly equal
and their values are more than 0.9, then the fault is three-
where Ns is the number of samples per cycle, y1(x) is the phase fault.
sampled measured signal at t0 , y2(x) is the sampled
measured at the next cycle, y1 and y2 are the arithmetic † If the two-phase alienation coefficients are equal and their
means of y1(x) and y2(x), respectively values are nearly 0, while the third phase alienation
coefficient is more than 0.9, the fault is single-phase to
1  ground fault.
N s
y1 = y (x) (3)
Ns x=1 1
† If the two-phase alienation coefficients are equal and their
values are more than 0.9, while the third phase alienation
1 
N s
y2 = y (x) (4) coefficient is nearly 0, the fault is phase – phase to ground
Ns x=1 2 fault.

The variance between two signals is defined as the alienation † If the three-phase alienation coefficients are not equal and
coefficient; its value is between 0 and 1. their values one phase is more than 0.9, second phase is more
than 0.45, while the third phase alienation coefficient is
To apply the above mathematical method for any phase nearly 0, the fault is phase – phase fault.
current, the two variables are
3. To make sure of distinguishing between double line
y1(x) is the initial instantaneous value of the current at time and double line to ground fault, the zero-sequence current
t0 . (i0) is calculated from the three-phase measured currents as
follows
y2(x) is the instantaneous value of the current at next cycle.

1
2.2 Fault detection and faulty i0 = (ia + ib + ic ) (5)
3
phase selection
To implement our technique, three tasks are starting in
parallel: fault detection, fault confirmation and faulty phase If the value of i0 is nearly 0, the fault is line – line fault.
selection as follows:

Figure 1 Simulated power system

IET Gener. Transm. Distrib., 2010, Vol. 4, Iss. 11, pp. 1236– 1244 1237
doi: 10.1049/iet-gtd.2009.0648 & The Institution of Engineering and Technology 2010
www.ietdl.org

Figure 2 Three-phase current alienation coefficients (Aia , Aib , Aic) for case 1
a Current ia for case 1
b Current ib for case 1
c Current ic for case 1
d Aia for the current ia for case 1
e Aib for the current ib for case 1
f Aic for the current ic for case 1

1238 IET Gener. Transm. Distrib., 2010, Vol. 4, Iss. 11, pp. 1236 – 1244
& The Institution of Engineering and Technology 2010 doi: 10.1049/iet-gtd.2009.0648
www.ietdl.org

If the value of i0 is more than 5% of In , where In is the Table 1 Aix summary (case 1)
line nominal current, the fault is line – line to ground
fault. (A) Pre- (A) At (A) During (A) At fault (A)
fault fault fault clearing Post-
start fault
Aia ¼ 0 0.95 0 0.94 0
3 Case study power system Aib ¼ 0 0.95 0 0.82 0
The proposed technique is applied on the power system
Aic ¼ 0 0.92 0 0.92 0
shown in Fig. 1. The system parameters are obtained from
one-generation unit in El Kureimat power station that
produces 320 MVA [9]. The parameters of the selected Machine 2 (receiving source): Machine 2 has the same
system are as follows: parameters as machine 1 except that the steady-state
voltage phasor angle phase is 08.
Machine 1 (sending source): Rated line voltage is 19 kV, volt–
ampere rating is 320 MVA, frequency is 50 Hz, voltage Main transformers: At each side there is a step-up transformer
phasor angle phase is 208 and number of poles is 2. 340 MVA, 19.57/500 kV (delta/star earthed neutral), its

Figure 3 Three-phase current alienation coefficients (Aia , Aib , Aic) for case 2
a Current ia for case 2
b Current ib for case 2
c Current ic for case 2
d Aia for the current ib for case 2
e Aib for the current ib for case 2
f Aic for the current ic for case 2

IET Gener. Transm. Distrib., 2010, Vol. 4, Iss. 11, pp. 1236– 1244 1239
doi: 10.1049/iet-gtd.2009.0648 & The Institution of Engineering and Technology 2010
www.ietdl.org

Figure 3 Continued

primary impedance is 0.0027 + j0.184 V and its secondary 4.1 Three-phase to ground fault (case 1)
impedance is 0.7708 + j61.8 V.
This case studies the performance of the proposed technique
Auxiliary transformers: At each side there is an during the three-phase to ground fault condition. The
auxiliary transformer 32 MVA, 19.57/6.3/6.3 kV (delta/ operating power angle of generator 1 (d1) is 208. Fig. 2
star/star earthed neutral), its primary impedance is shows the simulation results for case 1. Figs. 2a–c present
0.02978 + j0.4894 V and its secondary impedance is the instantaneous values for the three-phase current. In this
0.0039 + j0.0261 V. case, we notice that the three-phase currents during the fault
are higher than the pre-fault currents; their values are nearly
Lines: Total load impedance is 0.0217 + j0.302 V/km with five times the pre-fault current. The alienation coefficients
200 km length for each circuit. Ax are calculated between two successive half-cycles with the
same polarity for the three-phase current signals. The three-
Loads: Each load is represented as impedance of value phase current alienation coefficients Aia , Aib and Aic are
100.0 + j60.0 V. shown in Figs. 2d–f. The values of Aia , Aib and Aic are equal
and close to zero before fault inception and after fault
clearing. At fault start they are equal and greater than 0.95
while they are greater than 0.94 at fault clearing. From the
above results, it is clear that the alienation coefficient values
4 Simulation results at fault initiation are good detector to determine the faulted
A fault F was considered at the middle of one circuit of the phases (ia , ib , ic). Summary of the alienation coefficients for
transmission line assuming that short-circuit is temporary the measured current signals at the different periods are
and not resistive. The developed technique was applied by shown in Table 1.
calculating the alienation coefficients Ax between two
successive half-cycles with the same polarity of current
signals at the sending end where relays would normally be 4.2 Single line – ground fault (case 2)
installed. To implement the present technique, the studied
All parameters are kept as in case 1, except the fault type is
power system configuration was simulated by using ATP
changed from three lines to ground fault (a – b – c – g) to
program [10]. The generated and measured three-phase
line current signals are taken from the transmission line Table 2 Aix summary (case 2)
terminals. Four simulation case studies are done to
discriminate between the faulty phases, by using the (A) Pre- (A) Fault (A) (A) Fault at (A)
proposed method, for fault classification. The current fault at start During clearing Post-
measured signals sampling rate is 50 samples per cycle, fault fault
which means sampling time of 0.4 ms. The total
Aia ¼ 0 0.95 0 0.96 0
simulation time is 0.8 s (i.e. the total number of samples is
2000). The fault inception time is 0.402 s and the fault Aib ¼ 0 0 0 0 0
clearing time is 0.502 s from the beginning of the
Aic ¼ 0 0 0 0 0
simulation time.

1240 IET Gener. Transm. Distrib., 2010, Vol. 4, Iss. 11, pp. 1236 – 1244
& The Institution of Engineering and Technology 2010 doi: 10.1049/iet-gtd.2009.0648
www.ietdl.org

single line to ground fault (a – g). Figs. 3a – c present the current, while both the healthy phase currents are equal
instantaneous values for the three-phase currents. The and nearly 1.35 times the pre-fault value. The three-phase
faulty phase current value is nearly six times the pre-fault current alienation coefficients Aia , Aib and Aic are shown in

Figure 4 Three-phase current alienation coefficients (Aia , Aib , Aic) for case 3
a Current ia for case 3
b Current ib for case 3
c Current ic for case 3
d Aia for the current ia for case 3
e Aib for the current ib for case 3
f Aic for the current ic for case 3

IET Gener. Transm. Distrib., 2010, Vol. 4, Iss. 11, pp. 1236– 1244 1241
doi: 10.1049/iet-gtd.2009.0648 & The Institution of Engineering and Technology 2010
www.ietdl.org

Table 3 Aix summary (case 3) Figs. 3d – f. Their values are equal and close to zero before
fault inception and after fault clearing. At fault starts, Aib
(A) Pre- (A) At (A) During (A) At fault (A) and Aic are equal to zero, whereas Aia is greater than 0.95.
fault fault fault clearing Post- At fault clearing, Aib and Aic are equal to zero, whereas Aia
start fault is greater than 0.96. Therefore the alienation coefficient
Aia ¼ 0 0.97 0 0.98 0 value at fault initiation is good indicator to determine the
faulted phase A. Summary of the alienation coefficients for
Aib ¼ 0 0.95 0 0.80 0 the measured signals at different periods are shown in
Aic ¼ 0 0 0 0 0 Table 2.

Figure 5 Three-phase current alienation coefficients (Aia , Aib , Aic) for case 4
a Current ia for case 4
b Current ib for case 4
c Current ic for case 4
d Aia for the current ia for case 4
e Aib for the current ib for case 4
f Aic for the current ic for case 4

1242 IET Gener. Transm. Distrib., 2010, Vol. 4, Iss. 11, pp. 1236 – 1244
& The Institution of Engineering and Technology 2010 doi: 10.1049/iet-gtd.2009.0648
www.ietdl.org

4.3 Double line –ground fault (case 3) Table 4 Aix summary (case 4)

This case studies the effect of double line to ground fault (A) Pre- (A) At (A) During (A) At fault (A)
condition on the performance of the proposed algorithm. fault fault fault clearing Post-
Therefore all the parameters are kept as in case 1, except start fault
that the fault type is changed to double line to ground
Aia ¼ 0 0.57 0 0.48 0
fault (a – b – g). Figs. 4a –f show the simulation results for
the three-phase currents as the two faulty phases currents Aib ¼ 0 0.93 0 0.98 0
are nearly ten times the pre-fault values while the healthy
Aic ¼ 0 0 0 0 0
phase current is nearly 1.35 times its pre-fault current
value. The values of Aia , Aib and Aic are equal and close
to zero before fault inception and after fault clearing. At
fault starts, Aia and Aib are greater than the value of 0.95,
whereas Aic is still zero. At fault clearing, Aia and Aib are Summary of the alienation coefficients for the measured
greater than 0.9 while Aic is still zero. From the above signals at different periods is shown in Table 4. These
results, it is clear that the alienation coefficient value at results show that the alienation coefficient value at fault
fault initiation can define the double line to ground fault. initiation is good detector to determine the faulted phases
Summary of the alienation coefficients for the measured A and B.
signals at the different periods is shown in Table 3.
These show that the alienation coefficient value at fault 4.5 Discrimination between isolated
initiation is good detector to determine the faulted phases
A and B.
and ground double line faults
The results for both the cases, double line to ground and
double line faults, shows that the value of the alienation
4.4 Double line fault (case 4) coefficients appears for both the faulty phases either in
This case studies the effect of double line fault condition isolated or ground line fault. Although the alienation for
on the performance of the proposed algorithm. Therefore the faulty phases in case of line – line to ground fault has
all the parameters are kept as in case 1, except that the the same value, whereas in case of isolated line – line fault,
fault type is changed to double line fault (a– b). their values are different. The proposed technique has to
Figs. 5a– f show the simulation results for the three-phase distinguish clearly between both the faults. The zero-
currents as the two faulty phases currents are nearly ten sequence current calculation from the measured three-phase
times the pre-fault values, whereas the healthy phase currents is an ideal tool to identify whether the line – line
current is nearly 1.35 its pre-fault value. The values of fault is grounded or isolated without adding any extra
Aia , Aib and Aic are equal and close to zero before fault measuring equipment. Figs. 6a and b illustrate the zero-
inception and after fault clearing. At fault start, Aia , Aib sequence currents in both line – line and line – line to
and Aic have different values with values of 0.57, 0.93 ground, respectively. The zero sequence is so high in case
and 0, respectively. While at fault clearing, Aia , Aib and of grounded fault while its value is nearly zero for isolated
Aic have values of 0.48, 0.98 and 0, respectively. fault.

Figure 6 Zero-sequence current


a Zero sequence L –L (isolated) fault
b Zero sequence L – L ground fault

IET Gener. Transm. Distrib., 2010, Vol. 4, Iss. 11, pp. 1236– 1244 1243
doi: 10.1049/iet-gtd.2009.0648 & The Institution of Engineering and Technology 2010
www.ietdl.org

5 Conclusions network’, IEEE Trans. Power Deliv., 1999, 14,


pp. 1250 – 1256
In this paper, an alienation technique of transmission line for
fault identification and faulty phase selection has been [3] LIN W.-M., YANG C.-D., LIN J.H.: ‘A fault classification method
proposed. The main achievements of this work are as follows: by RBF neural network with OLS learning procedure’, IEEE
Trans. Power Deliv., 2001, 16, pp. 473 – 477
1. Three-phase current measurements are sufficient to
implement this technique. [4] FERNANDEZ A.L.O., GHONAIM N.K.I.: ‘A novel approach using a
FIRANN for fault detection and direction estimation for high
2. It is accurate to identify all ten types of short-circuit faults. voltage transmission lines’, IEEE Trans. Power Deliv., 2002,
17, pp. 894 – 901
3. It is efficient to distinguish between line – line isolated and
grounded faulty without needing any extra equipments. [5] POELTL A., FROHICH K.: ‘Two new methods for fast fault
type detection by means of parameters fitting and
4. Fast and simple method, as the time taken by this method artificial neural networks’, IEEE Trans. Power Deliv., 1999,
is about 20 ms (for a 50-Hz system). 14, pp. 1269 – 1275

5. The reliability of the proposed method is quite high. [6] YANG H.T., CHANG W.Y., HUANG C.L.: ‘On line fault diagnosis of
power substation using connectionist expert system’, IEEE
6. It is quite effective over a wide range of a pre-fault power Trans. Power Syst., 1995, 10, pp. 323– 331
level, fault resistance and fault inception angle.
[7] ADU T.: ‘An accurate fault classification technique for
7. The effects of DC components and harmonics are power system monitoring devices’, IEEE Trans. Power
eliminated with estimation of alienation coefficients. Deliv., 2002, 17, pp. 684 – 690

8. The technique does not use the power system element [8] HAUSCHILD W., MOSCH W.: ‘Statistical techniques for high
data as it needs only measuring the three-phase line voltage engineering’ (Hand book, English edition
currents. published by Peter Pere Grinus Ltd., London, UK, 1992),
Ch. 2, pp. 78– 79
6 References
[9] Generator Electrical Equipment: Instruction Manual
[1] SAHA M.M. , IZYKOWSKI J., ROSOLOWSKI E., ET AL .: ‘A new for Generator Electrical Equipment, Upper Egypt
accurate fault location algorithm for series compensated Electricity Production Company El kureimat Q 750 MW
lines’, IEEE Trans. Power Deliv., 1999, 14, pp. 789 – 797 Combined Cycle Project, Steam Turbine Generator &
Auxiliaries, Hitachi, Ltd., Tokyo, Japan
[2] AGGRAWAL R.K. , XUAN Q.Y. , DUNN R.W., BENNETT A. : ‘A
novel fault classification technique for double-circuit line [10] ATP: Version 3.5 for Windows 9x/NT/2000/XP – Users’
based on a combined unsupervised/supervised neural Manual – Preliminary Release No. 1.1 – October 2002

1244 IET Gener. Transm. Distrib., 2010, Vol. 4, Iss. 11, pp. 1236 – 1244
& The Institution of Engineering and Technology 2010 doi: 10.1049/iet-gtd.2009.0648

Você também pode gostar