2. INTRODUCTION:
• Design and selection of proper protection schemes are very much
essential for control and operation of power systems.
• It helps in better power reliability, less damage to power equipment
and safety of operational personnel.
• Penetration of the RES in power systems is steadily increasing.
• The derivative-based protection scheme has limitation in case of
faults in the islanding mode of operation, different microgrid
topologies, varying distributed generations (DGs) penetration, and
against the measurement noise.
3. WHAT IS LVDC SYSTEM:
• The low-voltage dc (LVDC) microgrid possesses numerous benefits and their penetration
in the power system has increased rapidly in recent years.
• The microgrids are classified into three categories: ac, dc, and hybrid ac/dc microgrid.
• LVDC system have the following benefits:
It can fulfill the electricity requirement in remote locations.
Has higher power transfer capability,
Better power quality, improved reliability,
Lesser maintenance and operation costs
Higher efficiency due to lower losses and lesser number of electronic converter.
4. TRADITIONAL PROTECTION
SCHEME:
OCRs
Overcurrent relays are most
widely used protection system
for distribution lines.
OCRs considered as the
primary protection technique
in dc microgrids.
Shabani and Mazlumi
proposed a communication-
assisted OCR for primary and
backup protection of LVDC
microgrids.
OCR is limited for high-
resistance faults.
Differential Protection
scheme
It is communication-based
differential current protection
for primary protection of a
medium voltage dc microgrid.
This improves the reliability
of the protection system
But it need additional cost
and maintenance
These scheme are effective
only in case of microgrids
integrated with PV farm only.
Cyberattack Detection and
prevention
It is for differential relays in
dc microgrids.
It proposed a multifunction
monitor which provides load
awareness to the relaying
system for assisting in the
detection of challenging fault
cases including high
resistance faults.
5. NONTRADITIONAL PROTECTION
SCHEME:
Overcurrent based
discrete wavelet
transform-assisted
protection scheme
For LVDC microgrids , this
scheme detects low and
high resistance faults using
overcurrent and wavelet
features respectively.
The inclusion of passive
elements increases the cost
of the relaying system.
Wavelet-based Protection
scheme
This scheme utilitzes the
rate of change of fault
current.
The performance of
wavelet-based schemes get
affected severely by the
decomposition levels and
selection of mother
wavelets.
AI and ML techniques
ML bases techniques can
overcome the challenges
faces by traditional and
signal processing based
relaying schemes including
high resistance fault and
faults with varying DG
penetration.
It has fast computation and
decision making abilities.
ANN technique
•Artificial neural network
and wavelet transform is
used for the classification
of faults.
•Computation time and
the burden for this
relaying scheme are high.
•Wavelet and support
vector mechine is used
for fault detection and
classification in the LVDC
microgrid.
6. WHY ENHANCED PROTECTION IS NEEDED:
• The utilization of LVDC microgrids on a larger scale is restricted due
to the absence of fast and reliable protection schemes and the lack of
standardization.
• The rate of rising of fault current is significant in dc microgrids,
which can damage the equipment and impose safety hazards for
operating persons.
• The dc microgrids do not have natural zero-crossing points for the
opening of circuit breakers. The impedance of cables is low, which
results in high fault currents in the events of a short circuit
7. WHY ENHANCED PROTECTION IS NEEDED:
• As it is found out that existing traditional, signal processing and ML-
based protection systems for LVDC microgrids have several
limitations.
• The existing techniques either fail to operate or operate with delayed
response time for faults in the islanding mode of operation, faults
with different topologies of the microgrid, and high-resistance faults.
• Furthermore, the existing schemes fail to distinguish the fault
condition from external faults, critical no-faults, and transient
conditions.
• Volatile nature of the DGs and microgrid with different types of DGs
has not been carried out in most of the protection schemes.
8. ML BASED RELAYING TECHNIQUE
• Current-assisted ML approach for LVDC microgrid protection
that utilizes the decision tree (DT) technique for the detection
of fault and the K-nearest neighbor (KNN) algorithm for fault
classification.
• This detect Pole to Ground and Pole-Pole faults with an
extensive disparity in fault resistance, fault location, and
operating conditions.
• The performance of the relaying technique remains unaffected
by the variable and intermittent generation of DGs.
9. FAULT ANALYSIS OF TWO BUS
SYSTEM
• From the mathematically
analysis the fault current is
depend on prefault dc-link
voltage and line parameters.
10. FAULT ANALYSIS
• The differential current and its first derivative are utilized for
the detection of the fault in the proposed scheme. The
mathematical expression of the first derivative of differential
current (FDODC)
• The second derivative of the differential current (SDODC)
• The combination of differential current, FDODC, and SDODC,
the classification accuracy becomes highly promising.
11. DECISION TREE AND K-NEAREST NEIGHBOR
TECHNIQUE:
• DT technique is used to identify the fault conditions. While KNN
method is utilized for classification purposes.
• DT has been trained with the features differential current and
FDODC.
• The KNN algorithm will check the KNN’s class; thus, the class with a
higher number of data points out of all k neighbor points will be
assigned as the class of unseen data.
• Both DT and KNN methods require offline training to predict the fault
condition and further fault classification online.
12. FLOWCHART OF ENHANCED DIFFERENTIAL
PROTECTION SCHEME
• KNN method is employed for the
classification of faults. Both the
models are first trained and then
tested for different conditions.
• The current and its first and second
derivatives are used to classify the
type of faults. At the time of trip
initiation, the corresponding data
are sent to the trained KNN model
for the classification of faults.
13. FEATURE EXTRACTION
• The critical no-fault conditions are the scenarios when the grid
is subjected to certain transients, which creates sharp variation
in voltage current and frequency patterns, similar to the fault
cases.
• The protection scheme should not detect the transients and
external fault cases as fault conditions.
14. MAIN CHALLENGES WITH INTEGRATION OF
RENEWABLE ENERGY SOURCES:
• False tripping or sympathetic tripping:
Integration of large scale RESs in distribution systems results in the
bidirectional flow of the fault current on most of the feeders/lines. A
non-directional over current relay may fail to provide the desired
protection for these networks during infeed from the RES.
• Islanding Problems:
if the fault current level sensed by relay is sufficient to trip it, then it
will lead to an islanding operation of the RES with its local connected
load.
15. MAIN CHALLENGES WITH RES
• Loss of Coordination:
The false/sympathetic/blinding operation of the relays from
downstream to upstream feeders leads to the sequentially false
operation of the relays. This type of false tripping of the relay in a
cascade manner is known as loss of coordination.
• Auto Recloser Problems:
, when the fault is partially cleared from the recloser bus, it still feeds
the RES. The injected RES fault current may cause energization of the
arc across the recloser and may convert a temporary fault into a
permanent fault.
20. FUTURE WORK:
• The testing of the proposed scheme in a practical LVDC
microgrid for high-resistance faults (above 20 ohm).
• The implementation on an FPGA board for developing the relay
module for commercial applications are being considered as
one of the future research works