PreprintPDF Available

Short Circuit and Arc Flash Study on a Microgrid Facility

Authors:
Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

Arc flash is one of the main hazards when operating an electrical facility. Without correct Personal Protective Equipment, the operator can be subjected to severe including fatal injuries. By code, facilities are required to properly label their electrical equipment that may be accessed by any operator. While energized, the operator proximity to the equipment can provide the necessary potential for an arc flash accident. The labels are mainly responsible for displaying the equipment short circuit and arc flash levels and the minimum PPE level required to operate it. These electrical hazard aspects become more critical in testbed facilities, usually located inside research centers and universities, where the electrical equipment is more frequently accessed by students and researchers. This paper develops complete modeling of a real microgrid testbed facility to perform short circuit and arc flash studies with the main goal to label the devices accessed by the facility researchers.
Content may be subject to copyright.
XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE
Short Circuit and Arc Flash Study on a Microgrid
Facility
Konrad Erich Kork Schmitt
Global Laboratory for Energy Asset
Management and Manufacturing
Texas Tech University
2500 Broadway, 79409
Lubbock, TX, USA
konradkorkschmitt@ieee.org
Cesar A. Negri
Wind and Science
Engineering Department
Texas Tech University
2500 Broadway, 79409
Lubbock, TX, USA
Cesar.Negri@ttu.edu
Saeed Daneshvardehnavi
Electrical and Computer Engineering
Department
Texas Tech University
2500 Broadway, 79409
Lubbock, TX, USA
Saeed.Daneshvardehnavi@ttu.edu
Stephen Bayne, Ph.D.
Electrical and Computer Engineering Department
Texas Tech University
2500 Broadway, 79409
Lubbock, TX, USA
Stephen.Bayne@ttu.edu
Michael Giesselmann, Ph.D.
Electrical and Computer Engineering Department
Texas Tech University
2500 Broadway, 79409
Lubbock, TX, USA
Michael.Giesselmann@ttu.edu
AbstractArc flash is one of the main hazards when
operating an electrical facility. Without correct Personal
Protective Equipment (PPE) supplemented by proper design of
the protective relays, the operator can be subjected to severe
including fatal injuries. By code, facilities are required to
properly label their electrical equipment that may be accessed
by any operator. While energized, the operator’s proximity to
the equipment can provide the necessary potential for an
incident arc flash accident. The labels are mainly responsible to
display the equipment’s short circuit and arc flash levels, as well
as the minimum PPE level required to operate it. These
electrical hazard aspects become more critical in testbed
facilities, usually located inside research centers and
universities, where the electrical equipment is more easily and
frequently accessed by students and researchers. This paper
develops complete modeling of a real microgrid testbed facility
to perform short circuit and arc flash studies with the main goal
to label the devices accessed by the facility’s researchers.
KeywordsArc Flash, Electrical Facility, Electrical Hazards,
Microgrid, Short Circuit.
I. INTRODUCTION
According to the Occupational Safety and Health
Administration (OSHA), from the United States Department
of Labor, in 2019, 106 work-related cases of electrical
accidents were investigated by OSHA. 59.4% of the accidents
resulted in a fatality [1]. Electrical systems can provide both
direct and indirect risks. The main danger is related to what
the electrical current can do to a human body, such as stoppage
of breathing or burns. Associated indirect dangers are the
damages to the human body that result from a fall, an
explosion, or a fire [2]. An electric shock occurs with the
contact of a human body to any voltage source able to provide
sufficient current flow through the person’s muscles or nerves.
The human body can feel currents from in the milliampere
range and, depending on the current path through the body the
current level can electrocute a person [2]-[3].
An arc flash is an explosion releasing heat, hot gases, and
molten metal caused by a short circuit of energized
conductors. An arc flash can produce by the burning of the
operator’s clothing and skin. Injury to bare skin begins at an
incident energy value of about 1.2 Cal/cm². This amount of
heat energy is equivalent to exposing the skin to a candle
flame for about 1 second [4]-[5]. To avoid any of these
injuries, operators must wear Personal Protective to energize
the systems Equipment (PPE). The PPE is a composite of
different items that provide electrical isolation and safety to
the operator when the person is close to electrical equipment.
There are four different categories of PPE whose protection
levels increase according to the incident energy available in
the electrical equipment [6]. However, even proper PPE will
not guarantee safety. it just assures that second-degree skin
burns will not happen [7].
By considering the standards, facilities need to provide
short circuits, incident energy, PPE level, and other relevant
information for each electrical equipment that may want
examination, adjustment, service, and maintenance. It may
create a potential for arc flash incident occurrence during
energizing.
This information must be organized based on the label and
displayed in a despicable place in front of the equipment,
keeping the operator aware of the potential for injury
associated with the installation. The labels should be installed
on all existing enclosure doors and removable panels that the
operators may access to achieve a maximum level of safety
inside the facility. Once that the facility decides to have an
electrical safety program, it is required to pass through a
sequence of steps and calculations to understand the hazard
level that their energized equipment can provide to the
operators.
The main steps to perform in an arc flash study are:
1) Data collection: collect and organize the equipment
and system’s parameters and information.
2) System’s topologies of operations: define the
different configurations in which the system can operate.
3) Three-phase bolted fault currents: calculate the
short-circuit current and its contribution to arc flash for each
piece of equipment.
4) Arcing current and incident energy: obtain the arcing
current and incident energy through the standard guides based
on the system’s characteristics.
5) PPE Category: Selecting the minimum PPE level
required to operate each piece of equipment.
6) Equipment labeling: display all the final information
in a label for each piece of equipment, pasting it to a place of
easy view.
Microgrids are turning to be an indispensable part of
modern electrical systems. They operate in both islanding and
grid-connected modes [8],[9]. The importance of microgrids
has resulted in a significant amount of related research.
Different aspects such as power quality [8]-[10], islanding
detection, integration of renewable resources (wind and
solar), and integration of other assets such as electric vehicles
and electrical machines and generators are reported in [11]-
[14]. Different research approaches focus on simulation and
hardware testing of microgrids. Using a Real-Time Digital
Simulator is a reliable option especially for initial studies that
are used besides the real testbed [15]-[18]. Beyond that, a real
microgrid can assist researchers to do more research. Using
previous data to predict the behavior of MG during and after
faults is an interesting topic for future research. For instance,
[] are two papers in other application which can be applied to
arc flash.
Texas Tech University owns and operates the Global
Laboratory for Energy Asset Management and
Manufacturing (GLEAMM) microgrid testbed facility. The
GLEAMM microgrid is located in Lubbock, TX, USA, and
is a research facility of Texas Tech University that
encourages partnerships between academy and industry,
having daily students and researchers working on its
equipment [19].
This paper explains the complete short circuit and arc
flash study developed for the GLEAMM microgrid. The
remaining part of this paper is subdivided as follows. Section
II shows the facility’s parameters and operation
configurations. Section III develops a complete short circuit
study for each piece of equipment. Section IV provides the
arcing current and incident energy calculations, based on the
standard guide. In section V, all the calculation results and
PPE dimensioning are shown, as well as the required labeling
for each device. At last, section VI concludes the study
accomplished in this paper, emphasizing the importance of an
arch flash study under a real testbed facility and presenting
the future steps for this study.
II. MICROGRID TESTBED TOPOLOGY
The GLEAMM laboratory focuses on advancing new
university innovations and certifying next-generation
industry technologies for protecting, enhancing, and
managing electricity transmission, and distribution systems.
The facility’s main goal is to provide an MW-scale testbed
for advanced studies in Microgrid areas related to grid
modernization, such as energy management, power quality,
control, and operation. Figure 1 shows an overview of the
facility’s diagram and the devices’ connections.
Fig. 1. Microgrid’s One-line Diagram Topology.
The microgrid facility feeds critical and non-critical loads
that are fed by the following assortment of power sources:
1) Utility: The system is mainly connected to the
utility’s 12.47kV feeder.
2) Diesel Generator: It is rated at 569kVA and works as
a backup source in case the grid is not available.
3) Wind Turbines: Three wind turbines are connected to
the microgrid main bus, limited to 400 kVA power
generation.
4) Solar Plant: The PV system is comprised of five
parallel inverters fed by 5 groups of PV panels with a total
peak power rating of 150kW.
5) Dynamic Load Banks: The facility has two resistive
load banks rated in 500kW each and one inductive load bank
rated in 187.5kVAr.
6) Basic Loads Are the loads responsible for the
building’s lighting, air conditioning, and general devices.
7) Outback System: This is comprised of one hybrid
inverter bank and a lead-acid battery of 30kW, being mainly
responsible for the critical loads.
Three power transformers are connected to the main
microgrid bus, the MCC. TR1 is responsible for step down the
utility’s 12.47kV to 480V, the most used voltage level in the
facility. Moreover, the MCC has two 480/208V step-down
transformers, TR2, and TR3. These two devices are
responsible to connect some of the non-critical loads and the
Outback system in the main bus. Table I displays the three
transformers’ parameters.
TABLE I. TRANSFORMERS PARAMETERS
ID
Parameters
S [kVA]
Vp [kV]
Vs [kV]
IZ [%]
X/R
TR1
1000
12.47
0.480
6.4
7
TR2
30
0.48
0.208
3.91
0.76
TR3
45
0.48
0.208
4.07
1.26
All the devices, except for the Outback inverter bank and HA
panel, are located outside the microgrid’s building, being
connected by underground cables to the MCC bus. The cable
connections were modeled o keep the circuit’s fidelity based
on their distance, type, and several sets by phase. Table II
provides the cable parameters for all the main lines connected
in the MCC bus. , the reactance values of the cables were
neglected for the presence modeling.
TABLE II. LINE PARAMETERS
Bus
From
To
Distance [m]
Cable Type
Impedance [Ohm]
Grid
MCC
50
500kcmil
0.00067485
Gen
MCC
25
500kcmil
0.00101227
Wind
MCC
300
500kcmil
0.00404910
Solar
MCC
50
500kcmil
0.00202455
Outback
MCC
10
2AWG
0.00512795
HA
MCC
10
2AWG
0.00512795
L1
MCC
25
500kcmil
0.00101227
L2
MCC
25
500kcmil
0.00101227
Based on the system’s diagram shown in Fig. 1, the
microgrid has two operational topologies. The first topology
is with the Automatic Transfer Switch (ATS) connected to
the grid. Under this configuration, the facility is directly
connected to the utility’s feeder, injecting the excess of power
to the grid or consuming power to feed its loads when needed.
The second possible topology is with the ATS connected to
the diesel generator. In this situation, the microgrid works in
island mode, where the generator is the system’s slack bus
and the remaining inverter-based sources work as grid-
following.
The transition between one topology to another can be
planned or not. During a planned transition, the MCC bus and
its loads are not de-energized. Otherwise, during a non-
planned transition, usually due to a grid fault, the MCC stays
de-energized from the fault until the moment that the
generator is turned on and reaches its steady state. During this
period, the Outback system is responsible to keep the critical
load on, isolating its circuit, and discharging the battery only
to support its critical load.
III. SHORT CIRCUIT STUDY
One of the most critical steps in an arc flash study is to
estimate the system’s short circuit levels. The lack of accuracy
and rounding errors in the devices’ parameters implies a
mismatch between theoretical and practical short circuit levels
that the system can provide. Microgrid systems typically have
different energy sources and are known to have higher
dynamics when compared to power systems grids. Generators,
inverters, and intermittent sources increase the system’s
complexity from the transient state point of view [20].
The GLEAMM microgrid has five energy sources, three of
them are inverter-based, namely batteries, solar, and wind
plants. The data for short circuit current levels extracted from
the datasheets for all the devices. The local utility was able to
provide the maximum contribution from the grid in a short
condition on the high and low voltage side of TR1. In the
datasheets of the inverter-based sources, the maximum
current that the power electronics devices can provide under
a short circuit condition is provided. As shown in Table III, a
factor of 1.4 was applied to the inverter-based sources’
maximum current values. This factor works as a safety band
for any non-standard operation of power electronics devices.
TABLE III. SOURCES PARAMETERS
a. Current values are based on the bus nominal voltage level, 208V.
Based on each source’s maximum contribution, the microgrid
is modeled in the PowerWorld software. PowerWorld is a
power systems analysis platform widely used by researchers
and utilities around the world. With this tool, it is possible to
model the system diagram and perform short circuit
simulations for different buses and under different operation
conditions. In the present study, each source was designed and
validated independently. The modeling looked to develop a
source that can provide the theoretical maximum short circuit
current when a three-phase bolted fault was applied in the
source’s bus. Figure 2 shows the GLEAMM microgrid
modeled on PowerWorld.
Fig. 2. Microgrid Modeling on PowerWorld.
The system was modeled and the maximum current from
each source was validated, the short circuit studies were
performed. The study aimed to analyze the most relevant
buses in the GLEAMM microgrid, as the MCC, Gen, Solar,
Wind, OB, HA, L1, and L2. For each of these buses, two
three-phase bolted faults were studied, one under the first
operation topology, microgrid connected to the utility, and
another under the second topology, where the microgrid is
connected to the diesel generator. In total, the short circuit
study collected twenty three-phase symmetrical fault values,
but only the highest value of each bus was used in the arc flash
study. Table IV shows the short-circuit study results.
TABLE IV. SHORT CIRCUIT RESULTS
Source
Parameters
V [V]
I Max [A]
Factor
Isc Max [A]
Utility
480
22131.3
1
22,130
Generator
480
220.6
1
220
Wind Turbines
480
660
1.4
925
Solar
480
181
1.4
250
Outback Battery a
208
169.8
1.4
240
Bus
Topology
Isc at Bus [A]
Tag
Voltage [V]
1ph SC
3ph SC
MCC
480
Grid-Connected
16147
22793
MCC
480
Islanded
1941
1296
Generator
480
Grid-Connected
16140
22783
Generator
480
Islanded
1941
1296
Solar
480
Grid-Connected
16141
22773
Solar
480
Islanded
1941
1296
Wind
480
Grid-Connected
16121
22757
Wind
480
Islanded
1940
1296
OutBack
480
Grid-Connected
16111
22741
OutBack
480
Islanded
1940
1295
OutBacka
208
Grid-Connected
8
52386
OutBacka
208
Islanded
8
2989
HA
480
Grid-Connected
16111
22741
HA
480
Islanded
1940
1295
LA
208
Grid-Connected
8
52365
LA
208
Islanded
8
2989
L1
480
Grid-Connected
16140
22783
L1
480
Islanded
1941
1296
L2
480
Grid-Connected
16140
22783
L2
480
Islanded
1941
1296
a. Current values based on the bus nominal voltage level, 208V.
IV. ARC FLASH STUDY
After the system’s data was collected and the maximum
short circuit current levels were calculated for each circuit’s
bus, the following step is to perform the arc flash calculations.
There are two main standards to guide arc flash studies in the
AC system, the IEEE SA 1584 Guide for Performing Arc-
Flash Hazard Calculation [5] and the NFPA 70E Standard for
Electrical Safety in the Workplace [6]. To analyze DC
systems, Doan's "maximum power method" is suggested, it is
used for linear DC systems, but cannot be used properly for
solar PV systems where we have a nonlinear current-voltage
I-V characteristic [21]. Also, the relevant hazard is considered
on the AC side of the solar inverters. The present study was
based on and followed the guidance provided by the IEEE SA
1584-2018. As a reference, the NFPA 70E equation does
provide results close to the ones obtained in this study.
The main goal of performing an arc flash study is to find
the incident energy and the arc-flash boundary levels. The arc
flash incident energy represents the amount of thermal energy
generated during an electric arc event and the arc flash
boundary means the distance from a prospective arc source at
which the incident energy is calculated to be 5 J/cm² [5].
The IEEE SA 1584-2018 provides a sequence of equations
to find the arc flash parameters, based on short circuit levels
and system characteristics. This study is highly related to the
facility’s setup, mainly with the Electrode Configuration
(E.C.), which can be categorized as:
1) VCB: vertical conductors/electrodes inside a metal
box/enclosure.
2) VCBB: vertical conductors/electrodes terminated in
an insulating barrier inside a metal box/ enclosure.
3) VOA: vertical conductors/electrodes in the open air.
4) HCB: horizontal conductors/electrodes inside a metal
box/enclosure.
5) HOA: horizontal conductors/electrodes in the open
air.
The GLEAMM microgrid is only contained E.C. type VCB.
This category informs the choice of the E.C. coefficients. the
arc duration time (T), the gap distance between conductors
(G), the working distance (D), and the enclosure equivalent
size (ESS) are other important parameters that must be
considered.
The most conservative way to consider the arc duration is to
assume the same arcing current on the faulted bus over the
fault time, which comes from the slowest tripping device
[22]. In this study, the arc duration was considered as 5
cycles. The gap distance was considered as 32 mm and the
working distance as 457.2 mm as respectively suggested by
Tables 8 and 10 from [5]. Using Table 6 from [5] were
obtained an ESS value of 19.999 mm for all of GLEAMM’s
enclosures, which have a height and width of 508 mm. The
ESS parameter enables us to calculate the enclosure size
correction factor (CF) by equation (1).
=++
(1)
where,

enclosure size correction factor;

equivalent enclosure size (mm);
coefficients for a typical VCB E.C.,
Table 7 of [5].
The arc flash study is categorized by systems with voltage
levels less or equal to 600V and greater than 600V. The
microgrid’s main voltage level is 480V, so the study followed
the first option. Before getting the average root mean square
(RMS) arcing current for our 480V system, we must obtain
the arcing current at 600V open circuit voltage through
equation (2).
_=10

+
+
+
+
+
+
(2)
where,
_
average RMS arcing current at

=
0.6  (kA);

bolted fault current three-phase fault
(kA);
the gap distance between electrodes
(mm);
coefficients for VCB E.C., Table 1 of
[5].
Based on the arcing current at 600V open circuit voltage
we can obtain the final arcing current for our 480V system
with equation (3).
 =
1
0.6
1
_
0.6
0.6
(3)
where,

final rms arcing current at

=
0.48  (kA);

open-circuit voltage

= 0.48 
(kV).
The arcing current is one of the characteristic parameters
that represent the system’s hazards. The IEEE SA 1584-2018
provides different equations according to the system voltage
level to find the incident energy. As the microgrid is a 480V
system, again it is used the equation for any open circuit
voltage less or equal to 600V. Based on it, we can obtain the
incident energy through equation (4).
=
12.552
50
10
 _









(4)
where,

incident energy for

0.6 
(J/cm²);
arc duration (ms);
working distance (mm);
coefficients for VCB E.C., Table 3 of
[5].
The incident energy level represents the amount of
thermal energy that that bus can provide during a short circuit
event. This value determines the minimum PPE category
required to work in that device when energized. Finally, with
the parameters collected and calculated, it is possible to
obtain the arc-flash boundary through equation (5). Similarly,
to the incident energy equation, the arc flash boundary has
different equations for different voltage ranges. In the
microgrid’s case, with a 480 level, it must use the equation
for any open circuit voltage less or equal to 600V.


=
10
 _










(5)
where,

arc-flash boundary for

0.6 
(mm);
coefficients for VCB E.C., Table 3 of
[5].
These steps must be followed for each of the system’s
devices, applying the five equations, and obtaining values
that represent the hazards in each bus. Section V compiles
and displays all the short circuit and arc flash study results,
looking to organize this information in labels that must be
paste into each device as a warning of possible hazards.
V. RESULTS AND EQUIPMENT LABELING
Based on the short circuit and arc flash results we can
understand the hazards present in each device that the operator
can be in touch with. Moreover, with these results we can rate
the minimum PPE category required to work on these devices,
avoiding possible injuries. The PPE has four categories that
depend on the incident energy level [6].
1) Category 1: minimum arc rating of clothing 4
cal/cm².
2) Category 2: minimum arc rating of clothing 8
cal/cm².
3) Category 3: minimum arc rating of clothing 25
cal/cm².
4) Category 4: minimum arc rating of clothing 40
cal/cm².
In the GLEAMM’s microgrid facility arc flash study, all
the buses presented incident energy smaller than 4 cal/cm²,
assuming then category 1. As the first PPE level, category 1
requires that the operator use an arc-rated long sleeve shirt
and pants, as well as a face shield. According to the device’s
location and access conditions, the operation can also be
required to wear an arc-rated jacket, rainwear, parka, and
hardhat liner. In addition to these items, the person is also
encouraged to wear heavy-duty leather gloves, hardhat, eye
protection, hearing protection, and leather footwear.
Table V provides an overview of the performed study
results, displaying the bolted three-phase fault current, the
arcing current, incident energy, arc-flash boundary and the
PPE category for each bus analyzed.
TABLE V. SHORT CIRCUIT AND ARC FLASH RESULTS
Bus 
[V] 
[kA] 
[kA] 
[cal/cm²] 
[cm]
PPE
Hazard
Category
MCC
480 22.79 16.81 3.6 91 1
Generator
480 22.78 16.85 3.6 91 1
Solar
480 22.77 16.85 3.6 91 1
Wind
480 22.76 16.82 3.6 91 1
Outback
480 22.74 18.82 3.6 91 1
Outbacka
208 52.38 16.12 3.8 94 1
HA
480 22.74 16.82 3.6 91 1
LAa
208 52.36 16.12 3.8 94 1
L1
480 22.78 16.85 3.6 91 1
L2
480 22.78 16.85 3.6 91 1
a. Current values based on the bus nominal voltage level, 208V.
With all studies completed, the last step is to label all the
devices to easily inform the users about the hazards present
inside that encloser. In the facility, we used the “Arc Flash
and Shock Hazard” warning label. Figure 3 shows the
warning label developed for the GLEAMM’s microgrid MCC
bus.
Fig. 3. MCC Bus Labeling.
This label displays the system nominal voltage, the arc-
flash boundary, incident energy, working distance, PPE
category, and the required and additional PPE clothes for
work if the respective device is energized. Moreover, with the
NFPA 70E, we have a limited and restricted approach
distance that is based on the system’s nominal voltage. All
this information is neatly on the label and must be pasted in
an easy view place.
VI. FINAL CONSIDERATIONS
The present paper proposed to develop a complete short
circuit and arc flash study for a microgrid facility, the
GLEAMM installation. As a research center, the GLEAMM
microgrid daily has students and researchers on its visiting
the facility that may be close and even operating some of the
de-energized and energized devices. The microgrid presented
a considerable short circuit level in all of its buses, 22.75 kA
on average with five different power sources. The arc flash
study looked to measure the hazards present in the facility,
obtaining the incident energy and the arc-flash boundary level
for each of the system’s buses. The purpose of calculating
these parameters is to inform the operators of the hazards that
they can face inside the facility. The importance of this study
is emphasized when considered that incident energy of 1.2
cal/cm² is enough to injury a bare skin and the microgrid’s
average incident energy was 3.63 cal/cm². Therefore, even
that the GLEAMM microgrid has the function to be a real
testbed, open for new researchers and looking to improve the
students’ practical skills, all persons, when working with the
facility, must be aware of its hazards and understand that
without the right PPE category that place has the potential to
provide them different types of injuries. The presented study
will take continuity analyzing the relation between the
different generation levels that each resource has with the
short-circuit and incident energy levels. GLEAMM has a
historical database with generation levels measured in the
range of seconds. The next research will collect the data and
with a probabilistic analysis will check which are the most
common combination between solar and wind generation and
their arc-flash levels. As well as with the statistical analysis,
a short circuit and incident energy daily profile will be
computed, showing which is the safer and most reasonable
time to make maintenance based on the short circuit
capability of the facility along its historical operation day.
ACKNOWLEDGMENT
The authors would like to acknowledge the Texas Tech
University, the Reese Technology Center, and the Global
Laboratory for Energy Asset Management and Manufacturing
(GLEAMM) for all the structural and financial support to
develop the present research.
REFERENCES
[1] “Fatality Inspection Data: Work-related fatalities for cases inspected
by Federal or State OSHA,” Occupational Safety and Health
Administration. Accessed on: Jul. 30, 2020. [Online]. Available:
https://www.osha.gov/fatalities
[2] C. F. Dalziel, “Dangerous electric currents,” AIEE Trans. Power App.
Syst., vol. 65, pp. 579585, 1946.
[3] T. Bernstein, “Electrocutions and fires involving 120/240 V
appliances,” IEEE Trans. Industry App., vol. 1A-19, no. 2, pp. 155-
159, 1983.
[4] D. R. Doan and R. Sweigart, "A Summary of Arc Flash Energy
Calculation," IEEE Transactions on Industry Applications, vol. 39, no.
4, pp. 1200-1204, 2003.
[5] IEEE-SA Standards Board, "IEEE 1584-2018 - IEEE Guide for
Performing Arc-Flash Hazard Calculations," IEEE, Piscataway, NJ,
USA, 2018.
[6] The standard for Electrical Safety Requirements for Employee
Workplaces, NFPA 70E, National Fire Protection Association, 2015.
[7] B. Stacho, J. Veleba, and J. Dudek, "Arc Flash Risk Assessment -
Overview of scope and different approaches in the US and the EU,"
20th International Scientific Conference on Electric Power
Engineering (EPE), Kouty nad Desnou, Czech Republic, 2019.
[8] Ghaedi, Amir, Saeed Daneshvar Dehnavi, and Hadi Fotoohabadi.
"Probabilistic scheduling of smart electric grids considering plug-in
hybrid electric vehicles." Journal of Intelligent & Fuzzy Systems 31.3
(2016): 1329-1340.
[9] 1Mortazavi, Seyed Mohammad Bagher, et al. "Optimal planning and
management of hybrid vehicles in smart grid." Ciência e Natura 37
(2015): 253-263.
[10] S. Dinkhah, C. A. Negri, M. He, and S. B. Bayne, "V2G for Reliable
Microgrid Operations: Voltage/Frequency Regulation with Virtual
Inertia Emulation," 2019 IEEE Transportation Electrification
Conference and Expo (ITEC), 2019, pp. 1-6, doi:
10.1109/ITEC.2019.8790615.
[11] J. A. Mueller, M. Rasheduzzaman and J. W. Kimball, "A Model
Modification Process for Grid-Connected Inverters Used in Islanded
Microgrids," in IEEE Transactions on Energy Conversion, vol. 31, no.
1, pp. 240-250, March 2016.
[12] S. Nikolovski, D. Mlakić, and H. R. Baghaee, "Arc Flash Incident
Energy Simulation in PV Power Plant Connected to Distribution
Network," 2018 International Conference on Smart Systems and
Technologies (SST), Osijek, 2018.
[13] L. N. Canha, P. R. Pereira, R. Milbradt, A. da Rosa Abaide, K. E. Kork
Schmitt and M. de Abreu Antunes, "Intelligent voltage regulator to
distributed voltage control in smart grids," 2017 52nd International
Universities Power Engineering Conference (UPEC), 2017, pp. 1-6,
doi: 10.1109/UPEC.2017.8231977.
[14] L. N. Canha, P. R. Pereira, R. Milbradt, A. da Rosa Abaide, K. E. Kork
Schmitt and M. de Abreu Antunes, "Intelligent voltage regulator to
distributed voltage control in smart grids," 2017 52nd International
Universities Power Engineering Conference (UPEC), 2017, pp. 1-6,
doi: 10.1109/UPEC.2017.8231977.
[15] Dehnavi, S. D., Negri, C., Bayne, S., & Giesselmann, M. (2021).
Dynamic Voltage Restorer (DVR) with a novel robust control strategy.
ISA transactions.
[16] Negri, Cesar A., Saeed Daneshvar Dehnavi, and Michael G.
Giesselmann. "Universal Model of Pulsed Alternators Using LTspice."
IEEE Transactions on Plasma Science 48.10 (2020): 3435-3442.
[17] A. Majd and R. Luo, "An Improved Arc Flash Energy Calculation
Method and Its Application," IEEE Transactions on Industry
Applications, vol. 53, no. 5, pp. 5062-5067, 2017.
[18] K. E. K. Schmitt, L. N. Canha, M. d. A. Antunes and P. R. Pereira, "A
smart local voltage regulator methodology for dynamic integration
between volt-var control and distributed energy resources," 2018 IEEE
PES Transmission & Distribution Conference and Exhibition - Latin
America (T&D-LA), 2018, pp. 1-5, doi: 10.1109/TDC-
LA.2018.8511728.
[19] Ghaseminejad, Ali, and Venkatesh Uddameri. "Physics-inspired
integrated space-time artificial neural networks for regional
groundwater flow modeling." Hydrology and Earth System Sciences
24.12 (2020): 5759-5779.
[20] Shirvani, Abootaleb, and Dimitri Volchenkov. "A Regulated Market
Under Sanctions. On Tail Dependence Between Oil, Gold, and Tehran
Stock Exchange Index." Journal of Vibration testing and System
Dynamic 3.3 (2019): 297-311.
[21] Shirvani, Abootaleb. "Stock returns and roughness extreme variations:
A new model for monitoring 2008 market crash and 2015 flash crash."
Applied Economics and Finance 7.3 (2020): 78-95.
[22] S. D. Dehnavi, C.A. Negri, K. E. Kork Schmitt. M. G. Giesselmann1,
S. Bayne, A Case Study for Connecting Bidirectional PEV Station for
Reactive Power Support to the GLEAMM Prototype Microgrid (PV
and Storage), Greentech 2021.
[23] Dehnavi S.D., Shahparasti M., Simab M., Mortazavi S.M Employing
interface compensators to enhance the power quality in hybrid AC/DC
microgrids (2015), 10.5902/2179460X20796.
[24] Dehnavi S.D., Shayani E. Compensation of voltage disturbances in
hybrid AC/DC microgrids using series converter Ciencia Nat, 2179-
460X, 37 (2015), pp. 349-356, 10.5902/2179460X20794.
[25] B. S. Martins, K. E. Kork Schmitt, L. N. Canha, M. de Abreu Antunes,
P. R. da Silva Pereira, and S. Cristina Nunes Araujo, "Real-Time
Hardware-In-the-Loop Testbed Applied to Voltage Control in
Distribution Smart Grids," 2019 IEEE PES Innovative Smart Grid
Technologies Conference - Latin America (ISGT Latin America),
2019, pp. 1-6, doi: 10.1109/ISGT-LA.2019.8894977.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
An integrated space–time artificial neural network (ANN) model inspired by the governing groundwater flow equation was developed to test whether a single ANN is capable of modeling regional groundwater flow systems. Model-independent entropy measures and random forest (RF)-based feature selection procedures were used to identify suitable inputs for ANNs. L2 regularization, five-fold cross-validation, and an adaptive stochastic gradient descent (ADAM) algorithm led to a parsimonious ANN model for a 30 691 km2 agriculturally intensive area in the Ogallala Aquifer of Texas. The model testing at 38 independent wells during the 1956–2008 calibration period showed no overfitting issues and highlighted the model's ability to capture both the observed spatial dependence and temporal variability. The forecasting period (2009–2015) was marked by extreme climate variability in the region and served to evaluate the extrapolation capabilities of the model. While ANN models are universal interpolators, the model was able to capture the general trends and provide groundwater level estimates that were better than using historical means. Model sensitivity analysis indicated that pumping was the most sensitive process. Incorporation of spatial variability was more critical than capturing temporal persistence. The use of the standardized precipitation–evapotranspiration index (SPEI) as a surrogate for pumping was generally adequate but was unable to capture the heterogeneous groundwater extraction preferences of farmers under extreme climate conditions.
Article
Full-text available
We use Student’s t-copula to study the extreme variations in the bivariate kinematic time series of log–return and log–roughness of the S&P 500 index during two market crashes, the financial crisis in 2008 and the flash crash on Monday August 24, 2015. The stable and small values of the tail dependence index observed for some months preceding the market crash of 2008 indicate that the joint distribution of daily return and roughness was close to a normal one. The volatility of the tail and degree of freedom indices as determined by Student’s t-copula falls down substantially after the stock market crash of 2008. The number of degrees of freedom in the empirically observed distributions falls while the tail coefficient of the copula increases, indicating the long memory effect of the market crash of 2008. A significant change in the tail and degree of freedom indices associated with the intraday price of S&P 500 index is observed before, during, and after the flash crash on August 24, 2015. The long memory effect of the stock market flash crash of August 2015 is indicated by the number of degrees of freedom in the empirically observed distributions fall while the tail coefficient of the joint distribution increases after the flash crash. The small and stable value of degrees of freedom preceding the flash crash provides evidence that the joint distribution for intraday data of return and roughness is heavy-tailed. Time-varying long-range dependence in mean and volatility as well as the Chow and Bai-Perron tests indicate non-stability of the stock market in this period.
Article
With the advancement of technology, electric equipment and loads have become more sensitive to problems related to power quality, such as voltage sag, swell, imbalances, and harmonics. To detect faults and to protect sensitive loads from these voltage distortions, a Dynamic Voltage Restorer (DVR) series compensator is among the best available cost-effective solutions. One of the main goals of the DVR is to achieve a control structure that is robust, stable, and can handle properly the disturbances (e.g., grid voltage issues, load current, and fluctuations at the DC link voltage) and model uncertainties (e.g., inverters and filter parameters). In this work, a novel framework control strategy based on Uncertainty and Disturbance Estimator (UDE) is proposed to improve the response of the DVR to properly compensate the load voltage under a variety of power quality issues, particularly the ones associated with the grid voltage disturbances. Additionally, the stability of the proposed control system is analyzed and validated using the Lyapunov stability theory. The advantages of the new control system are robustness, simplified design, good harmonic rejection, low tracking error, fast response, and sinusoidal reference tracking without the need for voltage transformations or specific frequency tunning (e.g., abc-dq0 and Proportional-Resonant). This research uses the MATLAB/Simulink software to validate the effectiveness of the proposed scheme under a diverse set of conditions with no control limitations. Moreover, the designed controller is tested under real conditions using Hardware-In-the-Loop (HIL) validation with OPAL-RT real-time simulator coupled with a TI Launchpad microcontroller. The results demonstrate a good performance of the proposed control strategy for a quick transient response and a great harmonic rejection when subject to grid voltage distortions.
Article
The high power and energy density of synchronous machines turn them into reliable sources of energy for pulsed power applications. In this article, two mathematical models using both actual and normalized (per-unit) parameters have been developed and simulated in LTspice, which is a powerful and a free tool to simulate electrical/electronic circuits. In this article, an efficient solver based on the flux and current equations is presented. The model has been validated against published results with both actual and per-unit parameters. It has been observed that for models with normalized parameters, the simulation time is significantly reduced. The validated model of the machine has been used to show the effects of nonlinear loads on the voltages and currents, in particular the reaction currents in the damper winding inside of the generator. In addition, the model has been used to study the effect of the RC time constant of the load on the peak power provided by a synchronous machine used to store energy for pulsed power applications.
Conference Paper
This paper studies the changes which has been occurring by the connection of distributed energy resources (DER) in the electrical distribution system. The big number of houses acquiring the photovoltaic generation system connected on-grid, increases the challenge of the distribution system company in maintain the service entrance voltage within service range. One problem of the on-grid system is that the network presents a big voltage fluctuation by the unpredictable resultant load profile in each residence that depends the instantaneous house power generation and consumption. While the big injection increases, the low injection of energy decreases the nominal house voltage, and without some kind of control, this voltage fluctuation is able to impact not only the own voltage profile, but also the neighbors’ voltage profile. For this reason, this paper proposes a smart local voltage control that uses an energy storage system (ESS) and Volt-VAr control in each house, triggering them when it is more conventional, maintain the nominal house voltage within the service range. Different from others studies, this local control is developed looking for small changes that when added, can impact in a positive way the entire system, improving the voltage profile and not overloading the voltage regulators and capacitor banks along the network. All control was developed, tested and analyzed on Matlab/Simulink, where was used a modified version of the IEEE 34-bus test feeder as case study.
Conference Paper
This paper presents a method for computation of arc flash incident energy in mixed alternating current (AC)/direct current (DC) systems with DC units such as strings, panel cabinets, inverters in Photovoltaic (PV) power plant, as well as on AC side behind the inverters connected to the distribution network. It also gives measures to reduce the risk and consequences of electrical arc flash that may occur due to live work or maintenance on DC and AC part of the PV power plant. By exploiting an advanced smart protection software for coordination of protective devices, this paper shows how we can reduce the risk of these harmful consequences for the personnel working in such power plants. An example case of a PV power plant with rated power of 300 kW is modelled with with its all components and existing protection settings. Then, the power of the electric arc flash will be evaluated using the suggested analytical-simulation-based method according to IEEE 1584 standard, followed by selection of proper protection devices settings for minimum arc power and risk for the staff.