High Impedance Fault Detection using FeaturePattern Based Relaying A. M. Sharaf
Guosheng Wang
and
11. SAMPLE STUDY SYSTEM Abstract-- The paper presents a novel time domain High Impedance Fault (HE) detection scheme based on low frequency
x distance to fault (km)
(Third,Fifth harmonic Excursion Patterns and Phast+Portraits). The proposed relaying scheme utilizes a recognition method.
U%
pattedshape
138/25KV
I
,
h
0-4(km) A
Keywords-- Detection, Feature-Vector, HIF Arc-Faults, Pattern, Phase Portrait, Relaying
I.
INTRODUCTION
H
igh Impedance Arc-Type Faults (HF) on Radial Electrical Distributiofltilization networks are characterized by an intermittent Arc-type nature and lowlevel of the fault currents. The detection of low-level groundcurrents using any conventional over-current or ground fault type relays is both difficult and sometimes inaccurate. [ 1-71. The Arc-nature of all (HIF) faults, being intermittent in nature, characterized by very low level fault current is also rich in low harmonic content and high frequency noise spectra that identify and fingerprint any existence of Arc hazardous HIF faults. High Impedance Arc-Type faults are safety hazard to Humans, Live stocks and Electric Utility Personnel. If left undetected for hours, weeks, and possibly months can prove fatal to humans and animals at typical current level of 50 milliampere or above. The need to find a simple, fast and accurate detection scheme using only time domain substation voltage and current signals (few cycles) at the feeder is the motivation for this novel PattedShape Based Recognition and Identification approach. Other techniques by the first author [8-91 encompass Neutral Residual harmonics, excursion harmonic-vector, frequency-based nonlinear transformations, Artificial Neural Network and this new concept of the Harmonic Pattems/Phase Portraits (HF'P-Concept) introduced in this paper.
Relay
f
Fig1. Single Line Diagram of the sample study system
Fig2. Per Phase Equivalent Radial Circuit Model with HIF Arc type fault (Rfnonlinear)
The radial distribution study system is shown in Fig1. Fig2 depicts the equivalent per phase circuit model used to study the arc-fault condition at different locations (x) along the distributionhtilization feeder. The System, Fault, Relay Data and Model Parameters are given in the Appendix.
~~
A. M. Sharaf is with the Department of Electrical and Computer Engineering, University of New Brunswick, Fredericton, NJ3, Canada (e-mail:
[email protected]). G. Wang is a graduate student in the Department of Electrical and Computer Engineering, University of New Brunswick, Fredericton, Nl3, Canada (e-mail:
[email protected]).
0-7803-8110-6/03/$17.00 02003 IEEE
Localized Load
222
In.
SAMPLE MATLAB/SMULINK RESULTS
PI vs.time
VI
0.5
Fig3. MATLAB/SIMLJLINKUnified Block functional d e l
0
The Unified Sample Radial Distribution System was subjected to an Arc type fault at different locations (x=O to full feeder length) measured from the substation bus, and the MATLAB/SIMULINK simulation model is shown in Figure 3. Only voltage and current signals (vl, il) at the feeder terminals were utilized as the detection signals. Figure 4 shows the Relay PattedShape Based Phase Portrait detection scheme, based on (1-2) cycle FlT spectrum. The Feature Vector X
vs I1
OO
I
[email protected]
0.1
2 ' I0
0.05
0 U.1
FigS. System with No Fault (Units in per unit)
is based on 31d, 5" harmonic levels.
vl
XF
p l vs.time
(v3f, i3f)--3rdharmonic
-+
VI vs il
10 0
il
-1 0 0.1
vi5yyj3
-0.5
4 b . o P -
vspii5
XF- Fault Feature Vector a=v3fri3f/p3f; b=v5PiSf/pSf; ep3Wplf; c=y3f=i3/v3; d=y5f=iS/v5; f=p5f/plf Fig4. Pattern Based Phase-Portraits (*) Detection Scheme with FlT (3d,5*)
i
Figures 5-8 depict the dynamic time response, Pattern and ShapesPhase-Portrait Signals for all 3d, 5" harmonic based phase-portraits in addition to the real time (volt-Ampere) and instantaneous power (pl) variations. These strange figures illustrate the distinct key patterns, shapes and phase-portrait features associated with the four possible systemlfeeder conditions. 1. No-fault (Fig4) 2. Bolted-fault (FigSA, B) for x=0.31, x=0.71 locations 3. Linear fault (Fig6A, B) for x=0.31, x=0.71 locations 4. Nonlinear fault (Arc-model) (Fig7A, B) for x=0.31, x=0.71 locations, where l=full feeder length.
U l-hdl.l 'oooom m 2ooo OO
sea#
10000
,o
%l.mt
0.1
5000u 00
0.05
0 90- - 4 A0.05 4A
0.1
0.1
Fig6A. System with Bolted Fault (x=0.31) (Units in per unit)
223
n l vs.time
VI vs i l
DI vs. time
VI vs il
2 0
-2 0
"bW3
.
2
-1 v ~ o.I-~+-----, 4
~ t p i s
5
~
looou U :U 0.05
'0
0.1
'0
0.05
0.1
Fig6B. System with Bolted Fault (xd.71) (Units in per unit) p l vs.time
00
0.05
0.1
0
0.05
0.1
Fig7B. System with Linear Fault (x=0.71) (Units in per unit) p l vs.time
VI vsil
VI vsil
1
5
o.20 -5 0
v a w 3
0.1 o,.f5 -1 O
W
5
"
100
'0
0.05
0.1
0
0.05
'0 0.1
Fig7A. System with Linear Fault (xd.31) (Units in per unit)
0.05
0.1
0
0.05
0.1
FigSA. System with NL Arc Fault (x=0.31) (Units in per unit)
224
tsl vs.time
A. M. Sharaf, L. A. Snider and K. Debnath, “Residual Third Harmonic Detection of High Impedance Fault in Distribution System Using perception Neural Networks”, Proceeding of the ISEDEM 93, Singapore, October 1993. [9] A. M. Sharaf, L. A. Snider and K.Debnath, “Harmonic based Detection of High Impedance Faults in Distribution Networks Using Neural Networks”, Proceeding of the IASTED conference 1993, Pittsburg, PA. [lo] A. M. Sharaf, R. M. EI-Sharkawy, H. Jalaat and M. A. Badr, “Fault Detection on Radial and Meshed Transmission Systems. [8]
VI vs il
VI.
APPENDIX
(a)AC System
Vu = 2 5 KV,V, =V,sinm, w=377 radlsec, R, =0.7 52, Ls=7 mH (b)Fault Model (HIF)
I ) Bolted HIF fault
R, = L , = O 2 ) Linear HIF fault R, = 30
Cl, L, = 3mH
3) Nonlinear (Arc-type)HIF fault
loou 00
0.05
0.1O
0 .
L, = 3 m H , R f =R,,(l+a(-)‘f
0.05 : U 0.1
Fig8B. System with NL Arc Fault (x=0.71) (Units in per unit)
lf 0
WhereRfo =20,if, = 7 0 , a=0.6, p=2
(c) Structure Detection Rules (Table I ) for Relaying/Isolation Rules
IV. CONCLUSION
TABLE I
The paper presents a novel low-frequency (3d, 5”) harmonic component-pattedshape based relay for High Impedance arc type Fault Detection. The relay is based on time domain (patterdshape) identification of the phaseportraits of harmonic voltage, current, power and v/i ratios, at only (3rd and 5”) harmonics. The proposed scheme is fast and accurate as it relies on time domain measurements and the specific shape of harmonic-ripple (31d,5”) associated with any linear or nonlinear (Arc-type) faults. The Proposed Scheme can be extended to other relaying schemes by utilizing time domain patterdshape recognition technology.
EXCURSION-VECTOR (3RD, 5TH)HARMONIC PATERN BASED RELAY
No fault
Ellipse shaped
HI “Application Guide on Protection of Complex Transmission Network Configurations”, CIGRE,SC34-WG04, May 1991. P I A. T. Johns, “Correspondence on Microprocessor Based Algorithm for
r41
r51 r61 r71
High-Resistance Earth-fault Distance Protection”, Pro. Inst. Elect. Eng., Part C, 132 (1985) A. Wiszniewski, “Accurate Fault Impedance Locating Algorithm”, Pro. Inst. Elect. Eng., Part C, 130, (1983) 311-314. T. Takagi, Y. Yamakoshi, “A New Algorithm of an Accurate Fault Location for EHVlLTHV Transmission Lines”, Part-I, Fourier transform method, IEEE trans., PAS-100 (1981) 1316-1322. Q. Yang and I. Momson, “Microprocessor-based Algorithm for High resistance Earth-fault Distance Protection”, Proc. Inst. Elect. Eng., Part C, 130, (1983) 306-310. B. Jeyansurya and W. SmoIinski, “Design and Testing of a Microprocessor-based Distance Relay”, IEEE Trans., PAS-I03 (1984 1104-1110. C. Roth and J. Femando, “A Microcomputer-based Interactive Transmission Line Simulator”, IEEE Trans., E-26 (1983) 151-156.
225
Mdti
IQm
Two
crossed mltual
bak)
crmsed lmps
openloop
boP5
Bolted (HIP)
circle shaped Reduced
size
V. REFERENCES
[31
B)
[HIP)
Skewed shaped
Flawer shaped ”loop
(Two
Multi
hacks)
adjacela
TWOdouble cosentric Retraced h s d
small hooks open loop Callapsed (collapsed) xhaced
b P
Crmsed Imp
Thngle shaped
(d) Transmission line
RLine= 0.25 S21 km ,LLine= 0.99472 mH I km , C,,, =0.01117p.FIkm
VII. BIOGRAPHIES Adel M. Sharaf (M76, SM’83) obtained his B.Sc degree in Electrical Engineering from Cairo University in 1971. He completed an M.Sc degree in Electrical Engineering in 1976 and the PH.D degree in 1979 from University of Manitoba, Canada. He was employed by Manitoba Hydro as special studies Engineer, responsible for engineering and economic feasibility studies in Electrical Distribution System Planning and Expansion. He authored and co-authored over 360 scholarly technical journals, conference papers, and engineering reports. Dr. Sharaf holds a number of US and intemational Patents (Pending) in electric energy and environmental devices. He is the president & Technical Director of both Sharaf Energy System Inc. & Intelligent Environmental Energy Systems Inc., Fredericton, Canada.
Guosheng Wang received the B.Sc in Electrical Engineering from North China Electric Power University in 1997. He was employed by Hanfeng Power as an Electrical Engineer, responsible for the Power Electronics and Control System feasibility. Currently he is an M.Sc Candidate in Electrical & computer Engineering of University of New Brunswick, Fredericton, NB, Canada. His research interests include power quality, stability and power electronics.
226
