IRJET- ARC Fault and Flash Signal Analysis in DC Distribution Systems

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 03 | Mar 2019 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 3969
ARC FAULT AND FLASH SIGNAL ANALYSIS IN DC DISTRIBUTION
SYSTEMS
Afrin Sheikh1, Piyali K. Saha2
1PG Student, Shri Sai College of Engineering and Technology, Bhadravati
2Assistant Professor, Shri Sai College of Engineering and Technology, Bhadravati
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - In DC electricity distribution systems and dc
microgrids, an oversized variety of electrical connectors and
long wire runs square measure expected. Combined with the
high dc voltage, deterioration of thewire insulationbecauseof
aging or different circumstances like mammal bites and
abrasion because of chaffing with trees, building walls, or
passage throughout installation will cause electrical arcs to
occur. These dc arcs could end in shock hazard, fires, and
system failure or fault within the microgrid.
In this paper, a new approach victimization WT base spectral
energy calibration for arc fault analysis in dc systems is
projected. The method of detection associatearcfaultinvolves
solar PV output voltage analysis and so feature identification.
Simulation models area unit synthesized to review the
theoretical results of the projected methodology and Fourier
analysis remodel analysis on arcing faults. Experimental
information from the dc system of a electrical phenomenon
array is additionally shown to validate the approach.
Key Words: DC arc, arc flash and fault
1. INTRODUCTION
In Photovoltaic (PV) systems, an oversized variety of
electrical connectors and long exposed wire runs are
required to make the series and parallel connection of
individual PV modules into strings and arrays. Combined
with the high DC voltage, deteriorationofthewireinsulation
because of puncture from placental bites and abrasion from
vibration will cause arc faults to occur which can lead to
shock hazard, fires, and system failure in electrical
phenomenon (PV) and different direct-current(DC)systems
like a microgrid [1]. DC electrical arcs in PV systems,
illustrated in Figure one, is series faults - caused by a
separation within the current path like because of a broken
wire or loose connective, or a high-impedance association
like because of corrosion of Associate in Nursing electrical
terminal or connector; or parallel faults – electrical
discharges between conductors of various potential
presumably caused by abrasion of wiring insulation once
mounted to the mounting frame because of thermal
enlargement, vibration, nesting rodents, or failureinsidethe
PV module [2, 3].
The problem of arcing faults exists for small-scale
residential systemsadditionallyaslarge-scaleutilitysystems
and may cause vital threats to human safety. As long as this
drawback exists, the PV business can have vital
considerations concerning liability and alsotheimpactupon
widespread adoption of electrical photovoltaicenergy. Thus,
arc fault detection is very vital for reliable and safe system
operation [4, 5] and may be a necessity for prime
penetration of PV [6].
Even a lot of vital to notice is arc flash, the pre-fault
(before a sustained arc forms) events of sparking and
insulator breakdown. Arc flash might solely last for a brief
length (less than a second), however is Associate in Nursing
early indicator of insidious arc faults. Investigation of arc
flash may be a troublesome drawback as a result of in
contrast to a fastened “hard short” fault during which high
current flows through a metal-to-metal association [7], arc
flash involves short-run current flowingthroughionizing air
or on Associate in Nursing particle path and in PV and
different finite energy systems might not draw sufficiently
high RMS current, or have a high enough I2t energy to trip a
thermal or magnetic electrical fuse, notably underneath low
light-weight (reduced generation) conditions.
Fig-1: Example of locations where arcing may occur in a
PV Array
Because of distributed capacitance of the PV array, noise
from electrical converter shift, and different magnetic
attraction interference, its potential this arcing might go
entirely unobserved [8]. An additional complication to
detection is that arcs aren'tperiodic,andthereforemight not

have simply recognizable amplitudeorfrequencysignatures
for pattern recognition techniques. Spectral analysis
victimization Fourier techniques to decompose the
frequencies of a sustained arc orfastenedfaultneedsa linear
system and a stationary signal [9] and thus isn't capable of
police investigation arc flash.
Fig-2: Example of dc wiring in a ground-mounted PV array
2. PROPOSED METHOD
2.1. Wavelet transform
Wavelet transform not like the FFT, it permits time
localization of various frequency elements of a given signal.
Thanks to the wide selection of signals and issues
encountered in power engineering, there are a unit varied
applications of rippling rework, like fault detection, load
foretelling, and power grid activity. Additionally, data
regarding power disturbance signals is usually a mix of
options that area unit well localized temporally or spatially
like power grid transients. This needs use of versatile
analysis strategies so as to handle signals in terms of their
time-frequency localization that is a wonderful space to use
the special property of wavelets [10].
The rippling analysis procedure relies on a rippling
model operate, known as a "mother rippling" – it provides a
localized signal process technique to decompose the
differential current signal into a series of wavelet elements,
every of that could be a time-domain signal that covers a
particular band [11]. Wavelets area unit notably effective in
approximating functions with discontinuous or sharp
changes like power grid fault signals. With correct selection
of the mother wavelet, wavelet transformation could be a
smart tool for fault detection and have distraction.
There are many sorts of wavelets, like Harr, Daubechies 4,
Daubechies8, Coiflet 3, and Symmlet8, to call a replacement.
One will select among them counting on the actual
application. The wavelet transform incorporates a digitally
implementable counterpart, the discrete wavelet transform
(DWT), almost like the Discrete Fourier transform (DFT)
implementation of the continuous-signal Fourier remodel.
The DWT is outlined as
)()(),(
,,
, ngnskjc
ZkNjZn
kj 

Where C(j,k) is the corresponding wavelet coefficient, n
is the sample number, s(n) is the signal to be analyzed and
gj,k(n) is the discrete scaling function (also called the father
wavelet), which for dyadic-orthonormal wavelet transform
is defined by
)2(2)( 2/
, kngng jj
kj  
The auxiliary functionto thisisthewaveletfunction(also
called the mother wavelet). With this initial setting, there
exists an elegant algorithm, the multi-resolution signal
decomposition (MSD) technique, which can decompose a
signal into levels with different time and frequency
resolution. At each level j, approximation and detail signals
Aj (represented by linear combinations of father waveletsat
jth level) and Dj (represented by linear combinations of
mother wavelets at jth level) can be created.
The words "approximation" and "detail" are justified by
the fact that Aj-1 is an approximation of A1, taking into
account the “low frequency" of Aj, whereas the detail Dj
corresponds to the "high frequency" correction.
As shown in Figure 3, for a reference level J, there are two
categories of details: 1) those details associated withindices
j ≥J correspond to the scales 2-j/2 ≤ 2-J/2, which are the fine
details; and 2) the other details correspond to j < J and are
the coarse details, which define an approximation of the
signals which signify that s is the sum of itsapproximation AJ
improved by the fine details [20].


jJ
jJ DAs
Fig-3: Wavelet decomposition tree
2.2. Cassie Arc Model
Circuit-breaker’s performance in power system is
analyzed by representing the circuit-breaker characteristics
by a operate of electrical parameters like current/voltage,
and mixing with, although difficult, facility circuit. For such
functions, supposed “Black-box modeling” is applied, in

which, despite of actual circuit-breaker hard-ware like
contact form, pressure, range of snapping point, etc., a
function of electrical parameters is introduced. Within the
chapter, as being popularly used equations, Mayr arc model
and cassie arc model are concerned.
In Cassie arc modeling, the assumptions are:
-Heat loss depends on the arc flow (convection loss).
-Heat loss, hold on heat, and electrical conductance
phenomenon area unit proportional to the cross section
space. Then, because the result, the subsequent is obtained.






 1
11
2
2
OE
E
dt
dG
G 
…………………….(1)
Where, E = Arc voltage,
E0 = Constant,
θ= Arc time constant,
G = Arc conductance.
The above assumptions correspond to relatively high
current of arc, such as higher than several hundred A, so
Cassie arc model is applicable to higher current of arc.
Introducing to EMTP-TACS, following rewriting is useful.
)(2
OO GGGG 
sO
O
E
I
G


1
1
2
2
For steady state, i.e. d/dt = 0, arc voltage E equal constant
E0. Therefore, as the equation is to be applied to relatively
long time interval of high current region, to introduce just
appropriate E0 value is important. As an example, so called
zero skipping current breaking near a synchronous
generator is taken up.
3. PROPOSED APPROCH IMPLEMETATION
Figure 4 shows the block diagram of proposed approach
arc fault and arc flash analysis of solar PV dc grid system
using wavelet based spectral energy calibration. In these
system cassie Arc model are utilized for development of DC
arc at different location of solar grid system. After DC arc
simulation in solar grid system then arc voltage was
measures and send to the wavelet multiresolution analysis
matlab block. Then we get detail and approaximate signals
after successful wavelet multiresolution analysis then send
that signal for spectral energy calibration. In future that
spectral energy value for different arc fault condition will be
utilized for classification of fault zone in DC grid system.
Fig-4: Block diagram of proposed approach
3.1. MATLAB Simulation model
MATLAB Simulation model was enforced in MATLAB
2015 software system version. Simulink model of the PV
array dc system with one 120 Hz double-frequency line
ripple (ac voltage 2), a 2 KHz switch ripple (ac voltage 1),
and series arcing (Cassie arc model details square measure
listed in Table I). The arc model at first behaves as a
electrical conductance with the worth 1e4 Siemens till the
arc “switches on,” then it's ruled by equation (1). This
simulates the separation of the electrodes that initiates the
arc.
Fig-5: MATLAB simulation model of proposed approach
Table-1: Parameters used with the cassie arc model
tau 1.2e-6 Sec
Uc 100 Volt
g(0) 1e4 Sec
Contact separationstarttime for
fault zone location
0.2 Sec

3.2. Wavelet and Filter Banks
Multiresolution signal analysis exploitationDWTwill be
enforced by filter banks wherever a moving ridge and a
scaling perform are related to a high-pass and a low-pass
filter, severally. As shown in Figure 3, on every level of
decomposition, the signal is split into a lowerfrequency part
and the next frequency part. With II moving ridge filters
(WT), solely the low-frequency half is additional rotten. As
compared, binary wavelet transform filters (wavelet
packets), that split each low- and high-frequency parts on
every level, result in rotten signals with associate degree
equal information measure.
3.3. Selection of mother wavelet
The criteria for selecting the mother wavelet adopted in
this paper are summarized in [12] and [13] are as follows.
1) The wavelet function should have a sufficient number of
vanishing moments to represent the salient features of the
disturbances.
2) The wavelet should provide sharp cutoff frequencies to
reduce the amount of leakage energy into the adjacent
resolution levels.
3) The wavelet basis should be orthonormal.
4) For applications where the information lasts for a very
short instant, wavelets with fewer numbers of coefficients
are better choices; on the other hand, for signal signature
spread over a longer period of time, wavelets with larger
numbers of coefficients tend to show smoother results.
There are several well-known families of orthogonal
wavelets. An incomplete list includes Harr, Meyer family,
Daubechies family, Coiflet family, and Symmlet family [39].
Daubechies wavelets are chosen in this paper due to their
outstanding performance in detecting waveform
discontinuities [13], [14].
4. MATLAB SIMULATION RESULTS
MATLAB Simulation results are classified in three
sections, first section shows results for arc parameters,
second section shows the results regarding wavelet
multiresolution analysis for different arc fault locations,and
third section shows result of wavelet spectral energy
calibration for different fault condition.
4.1. Arc parameters results
Figure 6 to 8 shows the source voltage, harmonicsvoltage
in source supply, arc voltage and load voltage for normal
condition, when arc flash at zone 1 and when arc flash at
zone 2 respectively. In these result it is clear that during
normal condition voltage across arc model is nothing but
normal supply voltage. When fault takes place at zone 1 and
zone 2 that time arc voltage increases their magnitude as
shown in figure 7 and 8 respectively for zone 1 and zone 2
respectively.
Fig-6: Result of DC source voltage, Harmonics voltage, Arc
voltage and load voltage during normal condition
voltage and load voltage during arc takes place in zone 1
voltage and load voltage during arc takes place in zone 2
4.2. Wavelet Multiresolution analysis results
Figure 9 to 11 shows the waveletmultiresolutionanalysis
of load voltage during normal condition, whenarcflashtakes
place at zone 1 and when arc flash takes place at zone 2
respectively. In which detail signals D1, D2, D3, D4 and A4
signals are available for different frequency band ranges. In
which it is observed that during normal condition high

frequency components are very less as compared with arc
flash conditions at zone 1 and zone 2.
Fig-9: Wavelet multiresolution analysis results of details
D1, D2, D3, D4 and Approximation A4 signal during
normal condition
D1, D2, D3, D4 and Approximation A4 signal during arc
flash takes place at zone 1
D1, D2, D3, D4 and Approximation A4 signal during arc
flash takes place at zone 2
4.3. Wavelet spectral energy calibration results
Figure 12 to 14 shows the wavelet spectral energy of
wavelet multiresolution analysis of details and
approximation signals during normal conditionandarcflash
conditions. In which it observed that during normal
condition energy of signals are normal around 0.04 to 0.1
range but during arcing fault conditiontheseismorethat 0.1
ranging.
Fig-12: Wavelet spectral energy of Detail D1, D2, D3, D4
and A4 signals during normal condition
and A4 signals during arc fault takes place at zone 1 at 0.4
second simulation time
and A4 signals during arc fault takes place at zone 1 at 0.8
second simulation time
5. CONCLUSION
This paper has proposed a new approach for arc
analysis in dc microgrid systems based on WT. The
fundamental feasibility of applying WT has been presented.
The presence of switching harmonics and ambient electrical
noise can mask the arc signal, making detection of an arc

difficult. Fourier analysis is usually not able to discover
transient signals and abrupt changes like sudden arc faults
and arc flashes. If the duration of the arc flash lasts fora very
short period of time in comparison with the sampling
window of FFT, it is likely that the arc flashes will not be
observable.
However, WT is extraordinarily effective with detecting
the exact instant the signal changes. The results suggestthat
the WT approach is not just capable of analyzing arc fault in
dc systems but that it also provides a more readily
detectable signal and better performance than the FFT
method.
REFERENCES
[1] J. Johnson, “Overview of arc-faults and detection
challenges,” U.S. Dept.Energy,Sandia Nat.Lab.,Albuquerque,
NM, USA, Tech. Rep. 284-9586, Feb. 2011.
[2] M. Rabla, E. Tisserand, P. Schweitzer, and J. Lezama, “Arc
fault analysis and localisation by cross-correlation in 270 V
DC,” in Proc. 59th IEEE Holm Conf. Elect. Contacts (HOLM),
Newport, RI, USA, 2013, pp. 1–6.
[3] A. Lazkano, J. Ruiz, E. Aramendi, and L. A. Leturiondo,
“Evaluation of a new proposal for an arcing fault detection
method based on wavelet packet analysis,” Eur. Trans.Elect.
Power, vol. 14, no. 3, pp. 161–174, May/Jun. 2004.
[4] N. I. Elkalashy, M. Lehtonen, H. A. Darwish, M. A.
Izzularab, and A. M. I. Taalab, “Modeling and experimental
verification of high impedance arcing fault in medium
voltage networks,” IEEE Trans. Dielect. Elect. Insul., vol. 14,
no. 2, pp. 375–383, Apr. 2007.
[5] S. Harb, M. Kedia, Z. Haiyu, and R. S. Balog,“Microinverter
and string inverter grid-connected photovoltaic system—A
comprehensive study,” in Proc. IEEE 39th Photovolt. Spec.
Conf. (PVSC), Tampa, FL, USA, 2013, pp. 2885–2890.
[6] G. D. Gregory and G. W. Scott, “The arc-fault circuit
interrupter, an emerging product,” in Proc. IEEE Ind.
Commer. Power Syst. Tech. Conf., Edmonton, AB, Canada,
1998, pp. 48–55.
[7] C. C. Grant, Fire Fighter Safety and Emergency Response
for Solar Power Systems, Fire Protect. Res. Found., Quincy,
MA, USA, 2010.
[8] G. Yunmei, W. Li, W. Zhuoqi, and J. Binfeng, “Wavelet
packet analysis applied in detection of low-voltage DC arc
fault,” in Proc. 4th IEEE Conf. Ind. Electron. Appl. (ICIEA),
Xi’an, China, 2009, pp. 4013–4016.
[9] F. M. Uriarte et al., “A DC arc model for series faultsinlow
voltage microgrids,” IEEE Trans. Smart Grid, vol. 3, no. 4, pp.
2063–2070, Dec. 2012.
[10] D. C. T. Wai and X. Yibin, “A novel technique for high
impedance fault identification,” IEEE Trans. Power Del., vol.
13, no. 3, pp. 738–744, Jul. 1998.
[11] C. E. Restrepo, “Arc fault detection and discrimination
methods,” in Proc. 53rd IEEE HOLM Conf. Elect. Contacts,
Pittsburgh, PA, USA, 2007, pp. 115–122.
[12] W. Li, A. Monti, and F. Ponci, “Fault detection and
classification in medium voltage DC shipboard power
systems with wavelets and artificial neural networks,” IEEE
Trans. Instrum. Meas., vol. 63, no. 11, pp. 2651–2665, Nov.
2014.
[13] C. Parameswariah and M. Cox, “Frequency
characteristics of wavelets,” IEEE Trans. Power Del., vol. 17,
no. 3, pp. 800–804, Jul. 2002.
[14] S. Mallat, A Wavelet Tour of Signal Processing (Wavelet
Analysis & its Applications), 2nd ed. San Diego, CA, USA:
Academic Press, 1999.
[15] L. Zhang and P. Bao, “Edge detection by scale
multiplication in wavelet domain,” Pattern Recognit. Lett.,
vol. 23, no. 14, pp. 1771–1784, 2002.
[16] Wang, Zhan, and Robert S. Balog. "Arc fault and flash
detection in photovoltaic systems using wavelet transform
and support vector machines." 2016 IEEE 43rdPhotovoltaic
Specialists Conference (PVSC). IEEE, 2016.
[17] Lu, Shibo, B. T. Phung, and Daming Zhang. "A
comprehensive review on DC arc faults and their diagnosis
methods in photovoltaic systems." Renewable and
Sustainable Energy Reviews 89 (2018): 88-98.
[18] Yuventi, Jumie. "DC electric arc-flash hazard-risk
evaluations for photovoltaic systems." IEEETransactionson
Power Delivery 29.1 (2014): 161-167.

IRJET- ARC Fault and Flash Signal Analysis in DC Distribution Systems

More Related Content

What's hot (17)

Similar to IRJET- ARC Fault and Flash Signal Analysis in DC Distribution Systems (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- ARC Fault and Flash Signal Analysis in DC Distribution Systems