Professional Documents
Culture Documents
org
ISSN 1751-8687
Abstract: Phase portraits are a powerful mathematical model for describing oriented textures. An isotangent-
based approach in a phase portrait is introduced to discriminate between internal and external faults. The
scheme described is new and is a different approach to the problem of relaying ground faults on transmission
lines. The geometrical theory of differential equations is used to drive a symbol set on the basis of the visual
appearance of phase portraits. The ‘phase portrait’ of the instantaneous rate of change of current against
voltage gives valuable information on the transient and stability characteristics of system configuration with
high fault resistance and shunt susceptance of the transmission lines. Hundred percent of the protected zone
is successfully discriminated rather than the external zone. An additional parameter that is considered, which
is sometimes neglected in protection studies, is the shunt susceptance of the transmission lines.
insight into the fault trajectory during the steady state and information about the flow pattern. The critical points of a
transients concerning the effect of high fault resistance and system with differential equations (1) and (2) are the
shunt susceptance. The geometrical theory of differential intersection points of the following equations
equations to describe the transmission line during fault
is obtained. The instantaneous rate of change of current F (x, y) ¼ cx þ dy þ f ¼ 0 (4)
against voltage in a phase portrait is given. Hundred percent G(x, y) ¼ ax þ by þ e ¼ 0 (5)
of the protected zone is successfully discriminated.
The purpose of the phase portrait analysis is to obtain full
2 Phase portraits insight for the fault trajectory during the steady state and
transients concerning the effect of high fault resistance and
Phase portraits provide an analytical tool to study systems of shunt susceptance.
first-order differential equations [23]. The method has
proved to be useful in characterising oriented texture [24,
25]. Let x_ and y_ denote two differentiable functions of 3 Phase portrait classification
time t, related by a system of first-order differential The elements of the characteristic matrix are used to
equations as determine six flow patterns if the characteristic matrix is
non-singular. The two classifications may be added for
x_ ¼ F (x, y) ¼ cx þ dy þ f (1) singular characteristic matrices. The eight classifications
y_ ¼ G(x, y) ¼ ax þ by þ e (2) derived from the characteristic matrix are no-flow, constant
flow, star-node, node, improper node, saddle, centre and
to describe a set of oriented textures comprising saddles, spiral. Some of these shapes are shown in Fig. 1 and can
star-nodes, nodes, improper nodes, centres and spirals be obtained as follows:
(Fig. 1). Equations (1) and (2) can be represented in matrix
notation as Node: the components x(t) and y(t) are exponentials that
either simultaneously converge to, or diverge from, the
X~_ ¼ A
~ X~ þ B
~ (3) fixed point coordinates x0 and y0 .
4 Conventional problems
Fig. 2 shows the tripping quadrilateral characteristic in the
case of high fault resistance. It will be noted that the
resistance of the fault arc takes the fault impedance outside
the relay’s tripping characteristic and, hence, it does not
detect this condition. Alternatively, it is only picked up
either by zone 2 or zone 3 in which case tripping will be
unacceptably delayed.
The infeed effect can be seen from Fig. 3, where there are
Figure 1 Phase portraits for a system of linear first-order other lines and sources feeding current to a fault at F from bus
differential equations R. The relays at bus S are set beyond this fault point to F 0. As
C dv
ic ¼ (8)
n dt
Rearranging (8)
0 1 0 10 1
d(is ic ) R Rf 1 (is ic )
þ
B dt C B L nL nL C B C
@ dv A¼@ A@ A
n
dt 0 v
C
0 R 1 (10)
f
i
B nL r C
þ@B C
A
n
is
C
Figure 5 Phase portrait of is – ic v for different R (R ¼ 0 Figure 7 Phase portrait illustration between internal and
and R ¼ 100) and SLG fault on phase-A external faults and SLG fault on phase-A
which is a straight line for any given w. Since the line Xt (n(1 n)R2 X 0 X 00 ) þ Rt (2n(1 n)RX þ X 000 )
equation is a function of b, every straight line is unique for (14)
~ Rt (n(1 n)R2 X 0 X 00 ) þ Xt (2n(1 n)RX þ X 000 )
a given point set in a given flow pattern. For the matrix A
given in (10), there is a unique straight line with a unique
isotangent, which depends on the system parameters. where Rt ¼ nR þ (1 n)R,
00
Xt ¼ nX þ Xg1 þ (1 n)X000
þ
Xg2 , X 0 ¼ (nX þ Xg1 ), X ¼ ((1 n)X þ Xg2 ) P and XP¼
The paper presents a phase portrait on the basis of the 2n(1 n)RX þ (1 n)RXg1 þ (1 n)RXg2 . The X = R
instantaneous rate of change of current against voltage, that ratio given above becomes totally different if the fault
is, (d(is ic ))=dt dv=dt rather than the instantaneous location is changed along the transmission line, and that
current and voltage (is ic v). The phase portrait gives can be easily recognised using the phase portrait based on
recognisable different isotangents for different cases of the rate of change of the current against voltage.
internal faults.
Similarly, the input impedance Zin for an external fault at
The X/R ratio of the transmission line is fixed. The Fe can be calculated as follows
(d(is ic ))=dt dv=dt portrait can be considered as a locus
of the instantaneous Z._ The portrait gives more visualisation ZinF e ¼ (R þ jX ) þ jXg1 )==( jXg2 ) (15)
of the instantaneous part of the Z. _ The portrait in such a
case is concerned with the rate of change of the impedance P P
P Equation (15) has the X= R ratio determined as
in thePtransient part, which is highly affected by the X
and R of the transmission line configuration at a point
(i.e. equivalent of X/equivalent of R at the point of fault). R2 þ (X þ Xg1 )(X þ Xg1 þ Xg2 )
(16)
This ratio can be calculated from the input impedance at the RXg2
fault point as will be explained later.
P P
The
P PX = R ratio given in (16) is different from the
In case of external faults, the phase portrait gives an almost X = R ratio given in (14). For extra and ultra high-
fixed isotangent. Thus, the process of discrimination between voltage transmission lines, the installed source capacities
internal and external faults can be easily identified. located at the sending and receiving ends of the TL are
P P almost constant. For an internal fault along the transmission
The isotangent given in (11) depends on the X= R line, the value of the (14) changes according to the fault
ratio (i.e. equivalent of X/equivalent of R at the point of location and, consequently, the isotangent is also changed.
fault) of system configuration. This can be easily calculated
as the input
P impedance
P at the fault point. The isotangent However, for an external fault along the transmission line,
or the X = R ratio is changed during internal faults the source capacity at the far end (Fig. 8) will have very high
based on the fault P location.
P During external faults, the source capacity compared with the transmission line
isotangent or the X = R ratio seen at the fault point is parameters. The rate of change of the impedance in the
almost fixed. Consequently, the isotangent in the phase phase portrait during the transient part depends on the
portrait during internal faults is different from the P P
X = R ratio given in (16), which is totally different
isotangent in case of external faults. from that in (14) and almost with a fixed value.
Consequently, the isotangent in the phase portrait is also
To prove this fact, a simplified configuration given in different and fixed for external faults.
Fig. 8 is considered. For a three-phase transmission line
operating at a certain voltage and having an impedance Figs. 9–11 show the instantaneous rate of change of
connected to the generating station busbar, for a certain current against voltage using a phase portrait for the same
fault at Fi , the input impedance is calculated using fault conditions as shown in Figs. 5 – 7. As shown in these
Thevenin’s as figures, the isotangent for different cases of fault events is
different, and the process of discrimination is easy to be
ZinFi ¼ (n(R þ jX ) þ jXg1 )==((1 n)(R þ jX ) þ jXg2 ) (13) identified.
Figure 10 Phase portrait based on rate of change for Figure 13 Studied configuration system and different
different fault locations events of faults
Table 1 Different cases of fault conditions Fig. 14 shows the phase portrait of the rate of change of the
current against voltage for different cases of fault conditions
Case Fault resistance, V Fault location, km mentioned in (Table 1). The isotangent in the case of an
F1 (SLG) 200 50 external fault is almost fixed. The suggested technique gives
the solutions for the symmetrical and unsymmetrical faults
F2 (SLG) 0 150 with solid and fault resistances. The threshold boundary
F3 (SLG) 100 150 THR for 100% of the transmission line is set. As shown in
the figure, all internal faults rather than the external faults
F4 (SLG) 0 240 have different isotangents (i.e. the isotangent in the case of
F5 (SLG) 0 10 (external) internal faults lies up and down away from the THR
boundary). This indicates that the relay is able to distinguish
F6 (SLG) 50 10 (external) successfully between internal and external faults.
F7 (2LG) 0 50 (external)
F8 (SLG) 200 50 (external) 9 Conclusions
The paper presented a technique for discrimination between
internal and external faults using a phase portrait. The paper
signals are the inputs to the relays, and 250 km is the line
used the geometrical theory of differential equations to drive a
length. The results are described on a phase portrait using
symbol set on the basis of the visual appearance of the phase
the (d(is ic ))=dt dv=dt diagram. The relay is set to
portraits. The ‘phase portrait’ of the instantaneous rate of
protect 100% of the line (250 km); it forms the first zone
change of current against voltage gives valuable information
of the relay. The power system is modelled and different
on the transient and stability characteristics of the system
symmetrical and unsymmetrical faults with solid and fault
configuration with high fault resistance. The paper showed
resistance are simulated using the ATP-EMTP.
that for non-singular, two-dimensional first-order phase
portraits, every isotangent curve is a unique straight line. The
The performance of the proposed technique was evaluated results showed the flow patterns for different internal faults
for different types of internal and external faults, fault with different isotangent lines. In case of external faults, a
location and fault resistance. Results showed faults are unique isotangent line is given. The discrimination between
taken with a fault resistance ranged from 0 to 300 V. internal and external faults can be identified. Furthermore, the
Table 1 shows the different cases of the fault conditions. impact of high fault resistance and remote-end feed is
The THR is adjusted using (16). The parameters of the assessed. The results indicated different events with types of
studied configuration system is given in the Appendix. fault condition, fault resistance and fault location using the
phase trajectory which passes through the phase portrait.
Hundred percent of the protected zone is successfully
discriminated rather than the external zone.
10 References
[1] HOROWITZ S.H., PHADKE A.G.: ‘Power system relaying’
(Research Studies Press, Taunton, Somerset, England, 1992)
[2] SIDHU T.S., GHOTRA D.S., SACHDEV M.S.: ‘An adaptive distance
relay and its performance comparison with a fixed data
window distance relay’, IEEE Trans. Power Deliv., 2002,
17, (3), pp. 691– 697
[3] SIDHU T.S., GHOTRA D.S., SACHDEV M.S.: ‘A fast distance relay
using adaptive data window filters’. Power Engineering
Society Summer Meeting, 2000, 16 – 20 July 2000, vol. 3,
pp. 1407 – 1412
[4] KUDO H., SASAKI H., SEO K., TAKAHASHI M., YOSHIDA K., MAEDA T.:
‘Implementation of a digital distance relay using an
interpolated integral solution of a differential equation’,
IEEE Trans. Power Deliv., 1988, 3, (4), pp. 1475 – 1484
Figure 14 Phase portrait discrimination between internal [5] SACHDEV M.S., SIDHU T.S., GHOTRA D.S.: ‘Implementation of an
and external faults at different fault events given in Table 1 adaptive data window technique in a distance relay’.
Developments in Power System Protection, 2001, 7th Int. [18] LEE J.B., JUNG C.H., KIM I.D., BAEK Y.K.: ‘Protective relay testing
Conf. (IEE), 9 – 12 April 2001, pp. 161– 164 and characteristic analysis for high impedance faults in
transmission lines’. Power Engineering Society Summer
[6] ALTUVE FERRER H.J. , DIAZ VERDUZCO I., VAZQUEZ MARTINEZ E.: Meeting, 1999, 18– 22 July 1999, vol. 2, pp. 1076 – 1081
‘Fourier and Walsh digital filtering algorithms for distance
protection’, IEEE Trans. Power Syst., 1996, 11, (1), [19] CHUL-HWAN K., JEONG-YONG H., AGGARWAL R.K.: ‘An enhanced
pp. 457– 462 zone 3 algorithm of a distance relay using transient
components and state diagram’, IEEE Trans. Power Deliv.,
[7] WAIKAR D.L. , LIEW A.C., ELANGOVAN S.: ‘Design, 2005, 20, (1), pp. 39– 46
implementation and performance evaluation of a new
digital distance relaying algorithm’, IEEE Trans. Power Syst., [20] BO Z.Q., AGGARWAL R.K., JOHNS A.T.: ‘A new approach to
1996, 11, (1), pp. 448– 456 transmission protection using fault generated high
frequency current signals’. Proc. 12th Power Systems
[8] EISSA M.M., MASOUD M.: ‘A novel digital distance relaying Computation Conf., Dresden, Germany, 19 – 23 August 1996
technique for transmission line protection’, IEEE Trans.
Power Deliv., 2001, 16, (3), pp. 380 – 384 [21] JOHNS A.T., AGARAWAL A.P.: ‘New approach to power line
protection based upon the detection of fault induced
[9] ZHNAG N., MLADEN K. : ‘Transmission line boundary high frequency signals’, Proc. Inst. Elect. Eng. Gen. Trans.
protection using wavelet transform and neurasl network’, Distrib., 1990, 137, (4), pp. 307– 313
IEEE Trans. Power Deliv., 2007, 22, (2), pp. 859 – 869
[22] PATHIRANA V., MCLAREN P.G.: ‘A hybrid algorithm for high
[10] VAZQUEZ E., CASTRUITA J., CHACON O., CONDE A. : ‘A new speed transmission line protection’, IEEE Trans. Power
approach traveling wave distance protection-part I: Deliv., 2005, 20, (4), pp. 2422 – 2428
algorithm’, IEEE Trans. Power Deliv., 2007, 22, (2),
pp. 795– 800 [23] WYLIE C.R., BARRET L.C.: ‘Advanced engineering
mathematics’ (McGraw-Hill, 1995, 6th edn.)
[11] WAIKAR D.L., ELANGOVAN S., LIEW A.C.: ‘Fault impedance
estimation algorithm for digital distance relaying’, IEEE [24] RAO A.R., JAIN R.C.: ‘A taxonomy for texture description
Trans. Power Deliv., 1994, 9, (3), pp. 1375 – 1383 and identification’ (Springer-Verlag, 1990)
[12] DJURI M., RADOJEVI Z., TERZIJA V.: ‘Time domain solution of [25] CHO K.H., KIM J.R., BAKE S., CHOI H.S., CHOO S.M.: ‘Inferring
fault distance estimation and arcing faults detection on bimolecular regulatory networks from phase portraits of time
overhead lines’, IEEE Trans. Power Deliv., 1999, 14, (1), series expression profiles’, FEBS Lett., 2006, 580, pp. 3511–3518
pp. 60– 67
[26] ROUSSEUW P.J.: ‘Leat median of squares regression’,
[13] TERZIJA V., RADOJEVIC Z.: ‘Numerical algorithm for adaptive J. Am. Stat. Assoc., 1984, 79, (388), pp. 871– 880
autoreclosure and protection of medium voltage overhead
lines’, IEEE Trans. Power Deliv., 2004, 19, (2), pp. 554– 559 [27] RAO R., SCHUNK B.: ‘Computing oriented texture fields’.
Proc. Conf., Computer Vision and Pattern Recognition, 1989
[14] XIA Y.O., LI K.K., DAVID A.K.: ‘Adaptive relay setting for stand-
alone digital distance protection’, IEEE Trans. Power Deliv.,
1994, 9, (1), pp. 480 – 491 11 Appendix
[15] EREZZAGHI M.E., CROSSLEY P.A.: ‘The effect of high resistance The parameters of the three-phase 500 kV transmission line
faults on a distance relay’. Power Engineering Society R0 ¼ 0:1896 V=km
General Meeting, 2003, 13– 17 July 2003, vol. 4
R1 ¼ 0:018 V=km
[16] LI Y., DE-SHU C., XIANG-GEN Y., ZHANG Z.: ‘The uniform formula L0 ¼ 3:456 mH=km
of single phase earth – fault distance relay with L1 ¼ 0:936 mH=km
compensation’. Transmission and Distribution Conference C0 ¼ 0:00828 mF=km
and Exposition, 2003, IEEE PES, 7 – 12 September 2003,
C1 ¼ 0:01134 mF=km
vol. 1, pp. 245– 249
The source capacity rating is
[17] SIDHU T.D., BALTAZAR D.S., PALOMINO R.M., SACHDEV M.S.: ‘A new
approach for calculating zone-2 setting of distance relays GVA ¼ 10 (sending end)
and its use in an adaptive protection system’, IEEE Trans. GVA ¼ 6 (receiving end)
Power Deliv., 2004, 19, (1), pp. 70– 77