# Islanding condition, its causes, mode of origination and detection

Millend Roy

Bachelor of Tehnology in Electrical Engineering, IIT ISM Dhanbad, Dhanbad 826004, Jharkhand

Electrical Engineering, Indian Institute of Technology, New Delhi 110016

## Abstract

With the escalation of energy demands all over and terror of draining conventional fossil fuels, the assimilation of distributed generation networks to centralized traditional networks was introduced. Distributed generation (DG) refers to the production of electricity near the consumption place through renewable energy resources especially wind, solar, tides, biomass, geothermal heat and so on. The heightened perforation of DG, renewable energy application, and the installation of micro-grid concept have changed the architecture of conventional electric power networks. Channelizing the field of study, the different problems and issues faced in the trending DG networks has been an interesting topic for researchers where several faults need to be worked upon. Islanding detection in DG systems is a challenging issue that causes several protection and safety problems. A micro-grid operates in grid connected mode or standalone mode. In the grid connected mode, the main utility network is authoritative for effortless operation in masterminding with the protection and control units, while in standalone mode, the micro-grid operates as a self-reliant and self sufficient power island that is electrically disconnected from the main utility network. Additionally, without stern frequency control, the equity and harmony between load and generation in the islanded circuit will be disrupted, leading to anomalous frequencies and voltages. Hence different anti-islanding detection methods are studied which reports to the control system how to perform in case of any islanding. Here basically an attempt is made to build a real time power system with several trending distributed generation systems being installed at the load end and then to monitor the behaviour of all the electrical parameters in case of any faults or disturbance. Even cases of islanding are considered and next protection of the power system is also considered by inculcating within different relays and circuit breakers wherever required.

Keywords: distributed generation (DG), grid connected mode, islanded mode, fault analysis based on real time, phasor measurement units (PMU)

## Abbreviations

List of Abbreviations
 DER Distributed Energy Resources DG Distributed Generation EMS Enegy Management System GPS Global Positioning System IdMs Islanding Detection Methods IPDN Integrated Power Distribution Network NYISO New York Independent System Operator PCC Point of Common Coupling PDC Phasor Data Concentrator PMU Phasor Measurement Unit RTDS Real Time Digital Power System Simulator SCADA Supervisory Control and Data Acquisition System SE State Estimator SVC Static Variable Compensator UDP User Datagram Protocol

## Background

Power grid architectonics have emerged remarkably since previous decades, from an unidirectional centralized management approach to a rational and decentralised doctrine which allocate autonomous solutions in administering today's amplifying demand complications. The notion of decarbonizing while boosting electrifications have bricked way for DG technologies to knock the existing power grids, affording a substitute power generation that is more handy to consumers. Hence DG is a term that refers to the production of electricity near the consumption place. Solar power generators, wind generators, gas turbines and micro-generators such as fuel cells, micro turbines and so on are all examples of distributed generators. Hence it can be addressed that conventional power distribution systems are passive networks, where electrical energy at the distribution level is invariably outfitted to the customers from power resources which are associated to the bulk transmission scheme.

Conventional centralized power distribution system
Microgrid network consisting of DG sources at load end
Shifting of Technology from Centralized to Decentralized; (Source:- “A Review of Islanding Detection Techniques for Renewable Distributed Generation Systems.”Renewable and Sustainable Energy Reviews, Pergamon, 2 Sept. 2013, www.sciencedirect.com/science/article/pii/S1364032113005650.)

Advantages of integration of DG resources include substantial environmental profits, enlarged adaptability, restraint of transmission and distribution (T&D) capacity promotion, abbreviated T&D line losses, remodeling power quality, providing better voltage support and so on. However diverse problems need to be tackled before the DG units are applied to the networks. These problems include voltage stabilization, frequency ballast, intermittency of the renewable resources and power quality controversies. Such technical challenges are being resolved by professional engineers and researchers using advanced technologies and power economics.[1] ​

Islanding as defined by IEEE standards :- A condition in which a portion of an area of ELECTRIC POWER SYSTEMS (EPS) is energized solely by one or more local EPS through the associated point of common coupling (PCC) while that portion of the area EPS is electrically isolated from the rest of the area EPS.

• 1
• 2

An analogy is a conglomerate of island banded together by a bridge type link like below. The bridges speak for power lines intertwining sections of the grid. If the bridges are fragmented, then each island would become secluded, both tangibly and electrically. In order for the grid to still operate on the island, it must generate enough power to accommodate load requirements. Each island will possess a contrasting frequency. And when the acquaintances are rebuilt, the two connecting islands must synchronize their frequencies before connecting or both islands will fail.

Interlinking of ideas between original island and electrical island; (Source:- Harvepino. “Aerial Footage of Scenic Road Stock Footage Video (100% Royalty-Free) 22185361.” Shutterstock, www.shutterstock.com/video/clip-22185361-aerial-footage-scenic-road-on-tiny-islands.)

An islanding condition can be intentional or unintentional. Intentional islanding is performed due to the obligatory maintenance needed for the main utility, whereas unintentional islanding may occur at any time due to regular faults or other uncertainties in the power system. Therefore, islanding detection is considered as an important task for IPDNs. IEEE 1547 and UL 1741 standards describe DG interconnection, planned and unplanned power islanding, and other important operating considerations. [2]More about intentional and unintentional islanding can be studied from the following journals.​[3]​​[4]

Unintentional islanding is a threat to power system security. All the problems discussed below adds to it.

•   SAFETY CONCERN: - As the grid may still be powered in the event of a power outage, due to electricity supplied by distributed generators. Therefore it may expose utility workers to life critical dangers of shocks and burns who may think that there is no power once the utility power is shut down.
• DAMAGE TO CUSTOMER’S APPLIANCES: - Due to islanding and distributed generation there may be a bi directional flow of electricity. This may cause severe damage to electrical equipments and devices. Some devices which are more sensitive to the voltage fluctuations should be equipped with surge protectors. Now question is what are surge protectors??- Device designed to protect devices from voltage spikes i.e. a transient event, typically lasting 1 to 30 microseconds, which may reach over 1,000 volts. A transient surge protector tries to control the voltage supplied to an electric device by either blocking or shorting current to decrease the voltage under a safe threshold. Blocking is done by utilizing inductors which brings about a sudden change in current. Shorting is achievable by using spark gaps, discharge tubes, zener-type semiconductors, and MOVs (Metal Oxide Varistors), all of which begin to conduct current once a certain voltage threshold is gained, or by capacitors which do not cause a sudden change in voltage. Some surge protectors use multiple elements.
• INVERTER CONTROL MODE SWITCHING: - Several inverters are installed with the distributed generators like solar and wind. The inverters have some control strategy and if the micro grid is islanded from the main grid then the PCC voltage, frequency are going to be changed. The control strategy running inside the inverters of different types of renewable resources takes the voltage and frequency feedback from the point of common coupling (PCC). Now if the frequency and voltage of the PCC change, so input to the controller of the inverter change. Therefore the controller may go to some other mode of operation. Islanding cause problems in proper functioning of the inverters.
• GRID PROTECTION INTERFERENCE: - Different types of relay, reclosures, fuses used in grid for protection may mal-operate during islanded mode of operation.

Hence islanding in power networks is a big challenge for protection engineers. As DG integration increases, the need for unintentional islanding detection will be more significant and challenging. Therefore, many new islanding detection methods (IdMs) have been developed to deal with these problems. Furthermore, the standards for intentional and unintentional islanding provide safe operational strategies and overcome the consequences of DG islanding, if properly implemented. For years, many IdMs have been presented, which reveal the fact that islanding is an open research problem.​[5]

## Statement of the Problems

• The simulation is carried on a software named RSCAD which is a production of REAL TIME DIGITAL based POWER SYSTEM SIMULATOR (RTDS). Real time simulation is a more often than not used tool for studying power system habits in feedback to situations and circumstances. These sort of pragmatic test can unearth potential problems in advance. Remidial measures could then be captured before enforcing the logic in the original system. Hence RTDS is a tool for the design, development and testing of power system protection and control schemes. Since RSCAD is a newly built software, it's working is also under research, hence to access any tool or proper building of the model, proper research requires to be done else it may end up with error prone results.
• Now to study and monitor the real time data obtained from the simulation, PMUs are required. The existency of PMU block in RSCAD helps to collect data and feed to PMU CONNECTION TESTER, where it is checked whether all the PMUs used in the simulation are exporting data or not. Next the data collected is put to the OpenPDC Manager, where all these are converted to an excel file from where they can be assessed.
• Since the modelling is to create a basic power system with all possible faults it may have and even the protection schemes installed within it, thorough study of fault analysis, bus classification, behaviour and working of PMUs, relay connections and linkage to circuit breakers are required as the pre-requisites.
• Studying PMUs and their CYBER SECURITY, how the data can be privatised and kept safe from the malicious programmers are also equally important since use of corrupted and attacked data may cause a huge loss to power economy.

## Objectives of the Work

• An encyclopedic analysis of diversified IdMs regarding their doctrine of operations, advantages and disadvantages, performance assessment, and real time utilizations is bestowed.
• A contrast of varied IdMs in terms of accuracy, computational burden, speed and cost is executed based on the critical reports.
• Manipulating the behaviours of relays and to study the operation so that it can perfectly understand the situation and send proper signals to the circuit breakers.

## Scope of Research

• Since simulation is done in RSCAD, the basics of it are discussed here in this paper as it's a new real time handling software where research is still under progress.
• Now, as the basic skeleton model used is a 9 bus system, ideas regarding buses are compulsory. The basics of bus classification and the different types of buses used in the 9bus system are also studied.
• Fault analysis and studying of different faults that may happen in a power system are taken into account as they serve as the basis of simulation for unintentional islanding.
• Now for analysing the simulation data collected, PMU serves the purpose. Hence why PMU is preferred over SCADA, demerits of SCADA and all other comparisons are studied.
• Next comes the protection of the power system model created, as any unnecessary transients would damage the overpriced instruments used in the model. Hence concepts of relay working logics and circuit breakers whenever faults occur are equally considerable in the field of study.
• Installation of DG resources at the load end of the model would create a realistic view of the modern-day decentralised power system. Therefore logics required for the synchronous working of the DGs with the utility grid have also been studied.
• For synchronous working of both the microgrid and the utility grid, frequency component and peak and rms values of voltage and current should be same. Now due to the varying loads in the load end the frequency changes every second. Hence to have such a situation in our model, import of real load data is made from NYISO where varying load of every 5 mins is uploaded. Now RSCAD has no such provision to accept data from files, it can accept data only through port. Hence UDP is used to send data to the port from where data can be accessed. Hence how frequency varies and code snippet for the UDP connection are also explained to have a clear view.
• Lastly the observations encountered are studied and tried to match the obvious if its possible.

## Power Mismatch Conditions

During islanded mode of operation, main grid is disconnected from the micro grid. Now this does not mean that POWER OUTPUT from the load = POWER GENERATED from the DG sources. This condition is known as power mismatch condition. It totally depend on the mode of operation of the micro grid system i.e. how many loads are switched on or off, how many DGs are in working condition or how many of them are off. [6]

• ∆P (active power mismatch) = P (active load power) – P (DG active power generation) = P (active power taken or given to grid).
• ∆Q (reactive power mismatch)= Q (reactive load power) – Q (DG reactive power generation) = Q (reactive power taken or given to grid).
• 0% ∆P indicates P(load) =P(DG)
• +10% ∆P indicates P (load) > P (DG); the system is loaded and the voltage is going to reduce. Hence under voltage relay is going to experience some tripsing. And if ∆Q is +ve Q (load) > Q (DG), then under frequency relay is going to be actuated.
• -10% ∆P indicates P (load) < P (DG), then over voltage relay is going to be actuated. Similarly if ∆Q is -ve Q (load) < Q (DG), then over frequency relay is going to be actuated.
• 1
• 2
• 3
• 4
• 5
• 6
• 7
Non Detection zone; (Source:- Local and Remote Techniques for Islanding Detection in Distributed Generators - Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/figure/Non-detection-zone-of-OUV-and-OUF-techniques_fig1_221907530 [accessed 11 Jul, 2019])

NON DETECTION ZONE (NDZ) defines for what particular percentage of power mismatch the islanding relay does not detect the islanding condition. All the DERs are to be equipped with under voltage and over voltage relays and under frequency and over frequency relays to detect whether the changed voltage and frequency are within the limits or not. According to IEEE standards OVR=1.11 per unit and UVR= 0.88 per unit; OFR=60.5per unit, UFR=59.3 per unit and the islanding detection should be within 1-2 seconds. [7]

The relation between the thresholds of power mismatch and voltage/frequency limits can be derived using the following equations.​[1]

• $\begin{array}{l}\left\{\left(\frac V{V_{max}}\right)^2-1\right\}\leq\frac{\triangle P}P\leq\left\{\left(\frac V{V_{min}}\right)^2-1\right\}\\\\\end{array}$
• $\left[Q_f\left\{1-\left(\frac f{f_{min}}\right)^2\right\}\right]\leq\frac{\triangle Q}P\leq\left[Q_f\left\{1-\left(\frac f{f_{max}}\right)^2\right\}\right]$

Where Vmax, Vmin, fmax, and fmin are the maximum and minimum voltage/frequency threshold limits of the relays in the DG system; ∆P and ∆Q represent the power mismatches prior to the main grid disconnection; Qf is the load quality factor usually considered to define parallel RLC load.

## Anti-Islanding or Islanding Protection and Detection

The act of preventing islanding from happening is actually what is anti-islanding. It is recommended that all distributed generators shall be equipped with devices that prevent islanding.

Before prevention of islanding, first pre-requisite is how to detect when islanding is happening. There are many ways to detect and resolve islanding:-

Different techniques for detection of islanding
 IdMs (Islanding Detection Methods) Local detection techniques Remote detection Techniques Signal Processing Based Methods Intelligent Based Methods Passive Active Hybrid Power Quality Monitoring Impedance Measurement Voltage unbalance and frequency set point Power Line Carrier Communication Fourier Transform Artificial Neural Network Rate of change of output power Active Frequency Drift Voltage dependency and reactive power shift Signal Produced by Disconnect Wavelet Transform Decision Tree Over/under Voltage protection Sliding Mode Frequency Shift Voltage Fluctuation Injection Supervisory Control and Data Acquisition Technique S-Transform Probabilistic Neural Network Over/under Frequency Protection Sandia Frequency Shift Synchrophasor/PMU based Technique Time-Time transform Support Vector Machine Rate of change of frequency Sandia Voltage shift Auto Correlation Function Fuzzy Logic Impedance Monitoring Kalman Filter Voltage Unbalance Phase Jump Detection Method

## Local Detection Techniques

It is based on the measurement of system parameters especially voltage, frequency etc at the DG site .​[8]​It is classified into:-

## Passive islanding detection

Here the system parameters (including voltage, current, impedance, power and frequency) are monitored at the PCC or DG terminals. These framework reveal symbolic disparity when the main grid is disconnected from the microgrid. The protection relays thus sense these variations and operate to trip the main breaker switch. They have large NDZ where they fail to detect islanding condition and have lesser detection speed [5] [1].

• POWER QUALITY MONITORING/HARMONIC DISTORTION:- Applicable for the inverter based DERs basically solar and wind systems; Inverters work on pulse width modulation techniques to generate low frequency output signals from high frequency pulses and generate higher order harmonics. Therefore in this method the harmonics and ripples in the voltage are measured at the PCC. To understand much more effectively, this journal would help a lot.[9]

Now if ‘d’ is >= certain threshold, then islanding condition is present else not present. One of the major drawbacks is the threshold selection.

• RATE OF CHANGE OF OUTPUT POWER: - The rate of change of output power, $\frac{dp}{dt}$ at the DG side once it is islanded will be higher than that of rate of change of output power before the DG is islanded for the similar percentage of load variation. [10] This method is much more compelling when the DG has unbalanced load inspite of having balanced load. [11] [12]
• OVER/UNDER VOLTAGE PROTECTION and OVER/UNDER FREQUENCY PROTECTION: - Here the accustomed traditional protection relays are established on a distribution feeder to regulate the bizzare circumstances during diversified modes of micro-grid operation.[13] The relays kickoff when the voltage and frequency values overpass a fixed threshold deadline, thus disconnecting the DGs from the main network. Customarily, the utility grid operations will be contingent upon the variations in the active and reactive powers (∆P and ∆Q) antecedent to a detachment from DG sources may crop up.

At the non-zero ∆P, an alteration in the amplitude comes into sight at PCC, which the O/UVP relay identifies and disconnects the DG. Contrarily, at non-zero ∆Q, a load voltage phase shift appears that deflects the inverter current frequency. The O/UFP relay notices this change and disrupts the DG.

• RATE OF CHANGE OF FREQUENCY:- The rate of change of frequency, df/dt, will be very high when the DG is islanded. [15] The rate of change of frequency (ROCOF) can be given by:-$\begin{array}{l}\frac{df}{dt}=\frac{\triangle P\times f}{2\times H\times G}\\\mathrm{Where},\;\triangle\mathrm P\;=\mathrm{power}\;\mathrm{mismatch}\;\mathrm{at}\;\mathrm{the}\;\mathrm{DG}\;\mathrm{side}\;\\\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\mathrm H\;\mathrm{is}\;\mathrm{the}\;\mathrm{moment}\;\mathrm{of}\;\mathrm{inertia}\;\mathrm{for}\;\mathrm{DG}\;\mathrm{system}\\\;\;\;\;\;\;\;\;\;\;\;\;\;\;\mathrm G\;\mathrm{is}\;\mathrm{the}\;\mathrm{rated}\;\mathrm{generation}\;\mathrm{capacity}\;\mathrm{of}\;\mathrm{the}\;\mathrm{DG}\;\mathrm{system}\end{array}$

Large systems have large H and G whereas small systems have small H and G giving larger value for df/dt( ROCOF). [16] Relay audits the voltage waveform and will function if ROCOF is greater than the threshold setting for a convinced span of time. The setting has to be selected in such a manner that the relay will prompt for island condition but not for load variations. This design is highly recommendable when there is enormous mismatch in power but it flops to operate if DG’s capacity matches with its local loads.​[1]

DRAWBACKS:-

• Requires a minimum power mismatch of 15%
• Performance can be affected by the type of loads and generator inertia constant
• IMPEDANCE MONITORING:-

Now at the point of common coupling for grid connected mode of operation; the voltage, current and then impedance= voltage /current are measured. The impedance turns out to be :-

$\left|Z_{normal}\left(j\omega\right)\right|=\frac1{\sqrt{\left({\displaystyle\frac{R_g}{R_g^2+\omega^2L_g^2}}+{\displaystyle\frac1R}\right)^2+\left(\omega C-{\displaystyle\frac1{\omega L}}-{\displaystyle\frac{\omega L_g}{R_g^2+\omega^2L_g^2}}\right)^2}}$

Here the grid is disconnected; hence it is working in islanded mode. Again the voltage, current and then impedance= voltage /current are measured at PCC. The impedance turns out to be :-

$\left|Z_{island}(j\omega)\right|=\frac1{\sqrt{{\displaystyle\frac1{R^2}}+\left(\omega C-{\displaystyle\frac1{\omega L}}\right)^2}}$

Therefore due to change in impedance values, it can be declared whether the microgrid is in islanded mode or grid connected mode. [17]

• VOLTAGE UNBALANCE:- During islanding, DG has to operate so as to continue power supply to the loads in the island. If the variation in loading is huge, then islanding conditions are perfectly identified by surveying several parameters: voltage magnitude, phase displacement, and frequency alteration. Nonetheless, these techniques may not be competent if the variations are less. As the distribution networks mostly include single-phase loads, it is widely possible that the islanding will alter the load equity of DG. Moreover, even though the change in DG loads is small, voltage unbalance will occur due to the change in network condition.​[18]
• PHASE JUMP DETECTION METHOD: - In the PJD method, [19] a phase discrepancy between the voltage and current at DG terminals is monitored for a sudden phase jump. During the islanding condition, an alteration occurs in the phase angle, which is corelated with the pre-set threshold value. The method fails to detect the islanding condition when the DG power and local load are closely matched. Therefore, the PJD method suffers from large NDZ. The NDZ of the PJD method is derived using the following expression as:- ​

$\tan^{-1}\left(\frac{\displaystyle\frac{\triangle Q}P}{1+{\displaystyle\frac{\triangle P}P}}\right)\leq\;\mu_{threshold}$

Where µ threshold denotes the phase-jump threshold. The circuit reactive elements may affect the accuracy of the PJD method; therefore, it is not suitable for industrial applications.​[1]

## Active islanding detection

Active methods directly collaborate with power system transaction by recommending perturbation signals. The extraneous signal injection brings momentous variations in the scheme parameters under the islanding condition and triggers the relays to operate. Here the large NDZ problem is overcome. Drawback: -Injection of high frequency signals produce voltage or current harmonics which affect the performance of such methods.(Total Harmonic Distortion)​[8]

• SIGNAL INJECTION BASED TECHNIQUES: -They inject disturbance signals like low frequency or high frequency signals, pulse signals and negative sequence based signals through inverter controllers. Islanding is detected based on electrical quantities and parameters behaviour at PCC which are measured and estimated due to injected disturbance signals.

During the grid connected mode, I (disturbance) is injected with frequency 30 Hz if the fundamental frequency is 60 Hz. Harmonic currents appear due to non-linear loads, or electronic components of the renewable energy sources. The impedance of the load path is higher than the source path; as a result the 60 Hz current component is going to flow through the load and the rest through the source.

The circuit breaker is open showing islanded mode of operation. As a result the grid current = 0 i.e. all the current flows through the load.

• IMPEDANCE MEASUREMENT: - This technique is dependent on the change in the high frequency impedance of the calculated data accumulated from voltage and current measurements. High frequency impedance turns more significant at the time of islanding condition. Generally used in power systems constituting synchronous DGs, they possess a small NDZ for the systems comprising single inverter based DGs. [20]
• ACTIVE FREQUENCY DRIFT (AFD): - The AFD is based on the injection of a current waveform distortion to the original reference current of the inverter to force a frequency drift in case of islanding operation.
(a) Original reference current and AFD reference current waveforms (b) Original reference current and injected current waveforms; (Source:- Yafaoui, A., Wu, B., & Kouro, S. (2011). Improved active frequency drift anti-islanding detection method for grid connected photovoltaic systems. IEEE Transactions on Power Electronics27(5), 2367-2375.)

By introducing a zero conduction time tz at the end of each half cycle, the phase angle of the fundamental component of current is shifted.

During normal grid connected operation the inverter generally operates with unity power factor and is synchronized to the grid voltage and will conduct at grid frequency. In islanding operation, the perturbation added to the current will generate a permanent drift in the operating frequency towards the local load resonance frequency so as to keep unity power factor. This drift will finally arrive at the frequency boundary limits set for islanding detection. [21]

The dead time tz and the original time period T can be related to define the chopping factor Cf used to perturb the waveform.

$C_f=\frac{2t_z}T$

The AFD reference current waveform can be given by: -

$\begin{array}{l}i_{afd}(t)=\left\{\begin{array}{l}\begin{array}{l}I\sin\left(2\mathrm{πvt}\right)\;\Rightarrow0\leq\omega t<\mathrm\pi-{\mathrm t}_{\mathrm z}\\0\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\Rightarrow\mathrm\pi-{\mathrm t}_{\mathrm z}\leq\mathrm{ωt}<\mathrm\pi\end{array}\\\begin{array}{l}I\sin\left(2\mathrm{πvt}\right)\;\Rightarrow\mathrm\pi\leq\omega t<2\mathrm\pi-{\mathrm t}_{\mathrm z}\\0\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\Rightarrow2\mathrm\pi-{\mathrm t}_{\mathrm z}\leq\mathrm{ωt}<2\mathrm\pi\end{array}\end{array}\right.\\\mathrm{where}\;\mathrm v=\mathrm f\left(\frac1{1-{\mathrm C}_{\mathrm f}}\right)\end{array}$

When this modified waveform is applied to an isolated DG system with RLC load, the frequency of the load will change according to the equation given by:

$arg\left[R^{-1}+(j\omega L)^{-1}+(j\omega C)\right]^{-1}=0.5{\mathrm{πC}}_{\mathrm f}$

• POSITIVE FEEDBACK BASED TECHNIQUES: - Electrical quantities like frequency, phase angle and voltage deviation etc. are given as feedback to inverters. Feedback is designed in such a way that DER operates in unstable region.

1. SLIDING MODE FREQUENCY SHIFT: - The SMFS method uses the phase angle of the inverter output current, which is monitored as a function of terminal voltage frequency. The SMFS method uses positive feedback to change the phase voltage at DG terminals by regulating the frequency deviations, thus identifying the islanding condition. The change in the phase angle of the inverter output current is always relative to the grid output voltage; thus, the frequency of the grid voltage deviates from its nominal value upon the main grid disconnection.​[1]​(http://energyprofessionalsymposium.com/?p=35452)​

2. SANDIA FREQUENCY SHIFT:- The SFS method is basically an extension of the AFD method. The process of applying SFS can be summarized as follows: -

• Inject a current harmonic signal with a limited duration into the Point of Common Coupling (PCC) so as to comply with the maximum THD allowed by interconnection standards.
• The injected current signal distorts the inverter current by presenting a 0A segment.
• The desirable effect of the 0A segment, is that the fundamental component of the inverter current leads the voltage by a small angle ӨAFD, which is frequency dependent and it creates a positive feedback.
• When the grid is disconnected, the frequency of the voltage of the PCC tends to drift, reaching values higher until the frequency is out of the OFP/UFP trip window (Range) and the inverter is disconnected.
• A positive feedback is utilized to prevent islanding.

The NDZ of the SFS highly depends on its design parameters. The design parameters include both chopping factor Cf and feedback gain factor K. If these handlers are not correctly adapted, it may show a failure as outcome or deterioration of the system power quality through injection of large amounts of harmonics. However, SFS may decline to detect islanding considering a fact that the deviations of voltage and frequency are less due to the power balance between DG sources and local loads.[22] The chopping factor is a function of the error in the line frequency and may be computed as follows:

$\begin{array}{l}C_f=C_{f_0}+K\left(f_a-f_{line}\right)\\\\where\;C_{f_0}:-is\;the\;chopping\;factor\;when\;there\;is\;no\;frequency\;error\\\;\;\;\;\;\;\;\;\;\;\;\;K:-\;an\;accelerating\;gain\;that\;does\;not\;change\;direction\\\;\;\;\;\;\;\;\;\;\;\;\;f_a:-is\;the\;line\;frequency\;measured\;at\;PCC\;\\\;\;\;\;\;\;\;\;\;\;\;f_{line}:-\;is\;the\;nominal\;line\;frequency\end{array}$

3. SANDIA VOLTAGE SHIFT:- Sandia Voltage Shift (SVS) applies positive feedback to the amplitude of Va. If there is a decrease in the amplitude of Va (usually it is the RMS value that is measured in practice), the inverter reduces its current output and thus its power output.​[23]

If the utility is connected, there is little or no effect when the power is reduced. When the utility is absent and there is a reduction in Va, there will be a further decrement in the amplitude of Va as governed by the Ohm’s Law response of the (RLC) load impedance to the decreased current. This additional decrement in the amplitude of Va leads to a further reduction in inverter output current, finally leading to reduction in voltage that can be identified by the UVP. It is possible to either have an increment or decrement of the power output of the inverter, leading to an analogous OVP or UVP trip.​[1]​ It is however preferable to respond with a power reduction and a UVP trip as this is less likely to damage load equipment.

## Hybrid islanding detection

System is monitored using passive islanding detection technique, when an island is detected it is further checked by active islanding detection technique. Hence here NDZ problem removed and even the percentage of DGs in the unstable mode is minimized.

• VOLTAGE UNBALANCE AND FREQUENCY SET POINT:- Voltage unbalance =V₂/V₁ where V₂ is negative sequence voltage and V₁ is the positive sequence voltage. Voltage unbalance is the passive technique. Fundamentally here the ratios are taken of certain frameworks (namely voltage) at the terminal of PCC.​[8]​ If the proportion is surpassing certain threshold, then there is islanding. This act as the first indication whether there is islanding or not. Now if it is islanding, then a disturbance signal is injected to the controller of the inverter i.e. the active technique is taken care of; else not required.​[24]​​​
Flowchart showing voltage unbalance and frequency set point; (Source:- 2018, IIT Roorkee July. “Mod04lec18.” YouTube, YouTube, 24 Aug. 2018, www.youtube.com/watch?v=YKcFu4_ONEs&list=PLLy_2iUCG87D59-Bc8Jqfqt43LvPC0KgC&index=37)
• TECHNIQUE BASED ON VOLTAGE AND REACTIVE POWER SHIFT :- Here the rate of change of voltage is the passive technique. Yet the real power shift (active) is enforced to the system if the passive technique cannot admirably discover the islanding condition. [25]

DG generates a level of reactive power flow at the point of common coupling (PCC). This power flow can only be maintained when the grid is linked. Islanding can be detected if the level of reactive power flow is not preserved at the set value. For the synchronous generator based DG, islanding can be disclosed by increasing the internal induced voltage of DG from time to time and observing the change in voltage and reactive power at the terminal where DG is fixed to the distribution system. A large change in the terminal voltage, with the reactive power remaining almost unchanged, indicates islanding.

• VOLTAGE FLUCTUATION INJECTION: - This technique is based on the voltage fluctuation injection, which can be obtained using a high impedance load. The islanding detection correlation factor (CF) is recommended for small-scale DGs, generally less than 1 kW. The two-stage method, which is the passive technique (rate of change of frequency (ROCOF)/rate of change of voltage (ROCOV)) for the protection scheme, and the active technique (CF) as a backup is used to achieve higher effectiveness. This technique utilizes digital signal processing to forecast the ROCOF, ROCOV, and CF of the distribution synchronous generator and to accurately detect islanding and non-islanding disturbances. The NDZ is decreased when the ROCOF is applied with a change in the active technique for islanding detection.

## Remote Detection Techniques

These are based on communication between utilities and DGs. Although these techniques may have better reliability than local techniques, they are expensive to implement and hence uneconomical .Some of the remote islanding detection techniques are as follows:-

## Power line carrier communication

Here a transmitter (T) is placed near the grid protection switch, and a receiver(R) is installed at the PCC. The transmitter continually broadcasts low energy communication signal (due to low pass filter nature of the power system) through the power line to the receiver to transmit islanded or non-islanded information. When the grid is disconnected, the transmitted signal is cut off because of the substation breaker opening and the signal cannot be received by the DG; which invokes the controller of the inverter to shut down the particular system. No current is going to be supplied from the system to the load. In radial system, only one transmitting generator needed that can continuously relay a message to many DGs in the network.​[26]

The main drawbacks are

i. To connect the transmitter and receiver device to a substation; a high voltage to low voltage coupling transformer is required which costs high.

ii. If the method is applied in the non-radial system, it results in the use of multiple signal generators whose implementation will require three times the cost.

## Signal produced by disconnect

This method is similar to the PLCC, with only difference is the mode of transmission used here specifically microwave link or telephone link. Its strengths are additional supervision and full control of both the grid and DG. The main drawbacks are:-

i. Relative expensiveness

ii.  Significant licensing

iii.  Design complications

## Supervisory control and data acquisition technique

SCADA (central control system) supervises the status of circuit breakers which are present inside the micro grid or smart grid system. Therefore monitors the auxiliary contacts those are liable to check the conditions of islanded operation. Upon islanding, a series of alarms is activated and the corresponding circuit breaker is tripped. [27] Major drawbacks:-

i. Cost of implementation is very expensive since each inverter requires separate instrumentation equipments and sensors.

These communicate with all DGs and offer synchronized operation between micro grid utility grid to avoid any phase difference.

## Synchrophasor measurement unit based technique/ PMU based islanding detection technique

Two PMUs are installed one at the grid side and the other at the DG side. Now the data collected about the voltage and currents signals from the PMUs are transferred to the relay which detects any sort of islanding. A circuit breaker is connected which triggers the decision to be taken by the relay.

Now suppose the major substation is disconnected and the breaker on the left hand side is open. The grid connected voltage phasor consisting of the magnitude as well as the angle is detected by the PMU 1. Similarly the DG connected voltage phasor is detected by the PMU 2. These data are sent to the relay where comparison is done either based on the magnitude or based on the phasor angle or frequency. Disconnection indicates unstabilized changed voltage and frequency which triggers the circuit breaker to open.

## Signal Processing Based Methods

These help to improve the detection performance, decrease detection time and reduce NDZ. These tools help to analyse and extract important features of a measured signal in order to perform efficient power system operations.[28]

## Fourier transform

FT is a frequency domain analysis gizmo used to concentrate the aspects of a signal at the specified frequencies. FT is not capable of considering time domain analysis. DISCRETE FOURIER TRANSFORM(DFT) and FAST FOURIER TRANSFORM(FFT) transform the finite length of the discrete time sequence into discrete frequency sequence. Such tools help to develop efficient and fast IdMs. [29]

DISADVANTAGE:- Reduced spectral estimation and low frequency resolution.

## Wavelet transform

Wavelet transformation is a strong candidate for extracting important features from a distorted voltage, current, or frequency signal. Methods based on WT are associated with STFT (short time fourier transform) and the multi-resolution techniques. [30]

In WT-based IdMs, the wavelet coefficients of the measured signal are compared with the pre-defined threshold value. If these coefficients attain a larger value than the pre-defined threshold value, the islanding condition will be detected. [30] The impacts of mother wavelet selection, threshold settings, and different sampling frequencies are the limitations of such methods. Moreover, WT can only detect the low-frequency band, and therefore, the wavelet packet transform (WPT) is applied to analyse the high-frequency components using the d-q axis of three-phase apparent power.[31] The WPT method is mostly related to the discrete wavelet transform (DWT), except for the fact that WPT gives equal resolution to low- and high-frequency signals. WPT can extract significant features from a measured voltage or current signal using approximation and detail decomposition. [32] [33]

## S-Transform

ST is an extension of the WT concept. It converts a time-domain function into a two-dimensional frequency-domain function.[31]Like other time-domain methods, the ST method is also utilized to extract important features from a measured signal at PCC, thereby detecting the islanding condition. Initially, ST processes the measured voltage or current signals at DG terminals and generates the S-matrix and the equivalent time-frequency contours. Then, the spectral energy content of these contours is calculated as containing the information of frequency and magnitude deviations for islanding detection. To process a signal, the ST method requires more computational memory than the other similar techniques. Moreover, the processing time of such methods is large.

## Time-time transform

The TTT method analyses and transforms one dimensional time domain signal to two dimensional time domain signal by giving time-time distribution on a particular window.

In the TTT method, the low frequency components are concentrated at different places while the high frequency components are concentrated around the localisation point having more energy concentration [34] .

TTT based techniques show good performance in noisy environment.

Kalman Filter is a well-known time frequency based harmonic analysis tool used in power systems to extract and filter harmonic features from the measured voltage and current signals.

## Auto co-relation factor

ACF is a mathematical model used to excerpt the secluded information from a computed power or energy signal. While considering the finite duration sequences, ACF is expressed using the finite summation limits. [28]

This method uses calculated envelopes to extract transient features, for which the variance of samples provides the criteria for islanding detection.

## Intelligent Based Methods

They are analogous to communication or signal processing based approach, except that they do not desire any threshold selection. These artistry are able to sort out the multi-objective problems which the conventional approaches cannot handle.​[1]

First, the input signal in the form of voltage or current measured at the PCC is fed to the system for data training and feature extractions using a training algorithm. This process is performed offline in order to save time and avoid computational burden. The online process is performed using an intelligent classifier model for making the final decision. In general, the intelligent IdMs suffer from the large computational burden.

Algorithm for Intelligent Based Methods; (Source:- Kim, M. S., Haider, R., Cho, G. J., Kim, C. H., Won, C. Y., & Chai, J. S. (2019). Comprehensive Review of Islanding Detection Methods for Distributed Generation Systems. Energies12(5), 837)

## Artificial neural network

The ANN based approaches extract important features from the measuring data, which are used for identifying variations in power system parameters. [35]

The dendrites basically gather the information about the parameters from the PCC. It is actually a computational model of a biological procedure which tries to learn and instruct itself from the past experiences like the biological brain of the neural network.

ANN-based IdMs detect the islanding condition by providing a high-accuracy and suitable operation for the systems using multi-inverters. ANN-based schemes perform the islanding detection efficiently but a large processing time and feature selection with multiple DG configurations still need to be addressed.

Comparison of Biological and logical neural networks; (Source:- Richárd, N. (2018, September 4). The differences between Artificial and Biological Neural Networks. Retrieved May 29, 2019, from Towards Data Science: https://towardsdatascience.com/the-differences-between-artificial-and-biological-neural-networks-a8b46db828b7?gi=d9b833c22055)

## Decision tree

Another classification technique utilized for islanding detection. The DT classifiers are used with wavelet transforms. [36]

Usually the voltage or current signals measured at the DG terminals are fed to WT for feature extraction. The extracted features are then processed using a DT classifier to detect islanding condition. ACCURACY achieved = 98%.

## Probabilistic neural network

PNN is applied in traditional pattern recognition schemes using the artificial neural hardware. PNN contains the following four layers: the input layer, pattern layer, summation layer, and output layer. These layers perform their functions for feature classifications and do not need any learning technique. [37]

## Support vector machine

SVM is a powerful classification tool used for signal and system analysis by constructing a decision boundary (hyper plane) to split the data needed for training by looking at the extremes of the data sets. The SVM classifier, in association with autoregressive modelling, is used to extract the signature features from the measured PCC voltage or current signals. The SVM-based IdMs provide high accuracy and fast detection speed. However, due to the large computational burden regarding the data training and algorithm complexity, SVM-based IdMs are considered impractical for real system implementations. [38]

Support Vectors; (Source:- Fuangkhon, Piyabute & Tanprasert, Thitipong. (2009). An Incremental Learning Algorithm for Supervised Neural Network with Contour Preserving Classification. 740 - 743. 10.1109/ECTICON.2009.5137153.)

## Fuzzy logic

It is actually a superset of Boolean logic (representation of 0 or 1 i.e. either completely true or false; hence discrete). But if we have to represent in continuous manner, Boolean logic stops working. Hence FUZZY LOGIC comes into view which represents the uncertainty (we do not know up to what extent it can go). [39]

To represent the intensity , level or degree FUZZY LOGIC is used.

FL was first introduced using DT transformation, where the combination of fuzzy membership functions and the rule based formulations were used to improve the fuzzy systems. Such methods show an efficient performance when applied in islanding detection algorithms.

Limitations:-  Highly abstract due to several min and max class operations.

Sensitive to noisy data due to repeated generation of rules of membership functions and classifications.

## Concepts

Before going to the simulation process, certain concepts need to be understood that would fill the gap between practical and theoretical concepts.

RTDS is a real time power system simulator, which employs an advanced, and easy to use graphical user interface. The RTDS allows users to accurately develop their models and simulate them efficiently. The software used to architect the power system model in RTDS is RSCAD, which involves a library of power and control system components. RSCAD permits the developer to choose a pictorial portrayal of the power system or control system components from the library in order to design the needed circuit. Once the system has been drawn with the entry of required parameters, the compiler in RTDS generates the low level code that is needed to perform simulation on the RTDS simulator. RTDS is capable of generating the real time signals, which enable the user to simulate the situations, which generally occurs in the power system. [40]

RSCAD is software which permits the user to develop a test case by using the numerous different components present in the RSCAD library. The following steps are required to prepare and run a new simulation case. [40]

Options available on opening RSCAD; (Source:- Kankanala, Padmavathy & Saran, Ankush & N. Schulz, Noel & Srivastava, Anurag. (2008). Designing and Testing Protective Overcurrent Relay using Real Time Digital Simulation.)
• The RSCAD/Draft software module is started.
• A new ‘Project’ and ‘Case’ directory is created in the ‘File Manager’ module.
• Next the new circuit diagram is created for simulation and the circuit is compiled.
• Lastly the simulation case is opened from RSCAD/RunTime.

The RSCAD/Draft software module is used to built the circuit that will be simulated employing the RTDS simulator.

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
In Draft window; (Source:- Kankanala, Padmavathy & Saran, Ankush & N. Schulz, Noel & Srivastava, Anurag. (2008). Designing and Testing Protective Overcurrent Relay using Real Time Digital Simulation.)

After the successful compilation of the system, the user simulates the system by using the RSCAD/ RunTime software module.

In Runtime window; (Source:- Kankanala, Padmavathy & Saran, Ankush & N. Schulz, Noel & Srivastava, Anurag. (2008). Designing and Testing Protective Overcurrent Relay using Real Time Digital Simulation.)

In RunTime mode, the user is able to add switches, buttons, meters and sliders for fault application, can plot graphs for voltage, current, power, fault, and frequency etc, can increase or decrease different physical quantities. After completion of the selection of plots in the RunTime, the user is able to simulate the system by pressing a button “Start Simulation” which is available in the RSCAD/RunTime module.

## Bus classification and concept of load flow

A bus is a node at which one or many loads and generators are connected. In a power system, each node is co-related with 4 quantities namely magnitude of voltage, phase angle of voltage, active or true power and reactive power in load flow problem two out of these four quantities are specified and remaining two are require dto be determined through the solution of equations. Depending on the quantities that have been specified, the buses are classified into 3 categories. Buses are classified according to which two out of the four variables are specified.

• Slack /Swing/Reference bus:- For the slack bus, it is assumed that the voltage magnitude$\left|V\right|$ and voltage phase ∂ are known, whereas real and reactive powers Pg and Qg are obtained from load flow equations. So one bus has to be made reference i.e. it has to be taken as $1\angle0^\circ$Another way to define it is :- slack bus is a generator bus i.e. connected to generator which is taken as reference.$P_{slack}+P_{generator}=P_{demand}+P_{loss}$
• Generator/ Voltage controlled/PV bus:- Here the voltage magnitude corresponding to the generator voltage and real power Pg corresponds to its rating are specified. It is required to find out the reactive power generation Qg and phase angle of the bus voltage. It is also called voltage controlled bus where we use compensator compensation techniques like static VAR compensator (SVC).
• Load/PQ bus:- No generator is connected to the bus. At this bus the real and reactive power are specified. It is desired to find out the voltage magnitude and phase angle through load flow solutions. It is required to specify only Pd and Qd at such bus as at a load bus voltage can be allowed to vary within the permissible values.
• 1
• 2
• 3
9 Bus System Line Diagram; (Source:- Use of ANFIS Control Approach for SSSC based Damping Controllers Applied in a Two-area Power System - Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/figure/Single-line-diagram-of-a-3-machine-9-bus-system-All-impedances-are-in-per-unit-on-100MVA_fig3_276427706 [accessed 11 Jul, 2019])

The 9 bus system is shown above along with all the buses and their classification to have a clearer view of the classes.

The active and reactive power flow equations of a power system are non-linear with unknown bus voltages and phase angles. The load flow feature of RSCAD DRAFT shown in fig 14, uses the Fast-Decoupled Newton-Raphson method for the power flow calculations which provides a solution to this set of non-linear equations. Prior to running the power flow, a bus label component (where bus type must be specified whether slack bus, PV bus or PQ bus) is required at every bus in the power system.

Bus nodes; (Source:- Taken from oursimulationat IIT Delhi)

From the power flow solution, the voltages and angles at all the buses can be determined. Once the power flow is completed, the RSCAD simulation should be compiled. The compile uses the power flow results to initialize the sources, generators, dynamic loads, exciters and governors which therefore creates a quasi-steady state starting point for the simulation. If the power flow does not converge it means that either the set of nonlinear equations itself does not have a solution or the equations could be ill conditioned and the numerical method used to solve the nonlinear equations is inappropriate. When the algorithm takes a large number of iterations it may indicate that the steady state solution is close to voltage instability.

## Fault analysis

Electrical systems ocassionally experience several types of faults. These faults are hazardous to the safety of both equipment and people. Faults in a 3 phase system can be unsymmetrical:-single line to ground, double line to ground, line to line or three phase symmetrical. Power system operation during any of these faults can be analysed using sequence components discovered by Charles L. Fortescue.

• Symmetrical Faults :- Before the fault, voltage and current were 120º phase displaced and balanced i.e. magnitudes of all the phases were same. Even after the fault, voltage and current magnitudes and phases remain balanced. Therefore this type of faults are known as symmetrical faults. Now faults can be either short circuit (L-L-L) or ground (L-L-L-G) faults.
• Unsymmetrical Faults :- Before the fault, the system is in balanced condition, but after sudden faults, they become unbalanced. Specifically line to ground (L-G) , line to line (L-L) and double line to ground (L-L-G) faults are common.

Fortescue Theorem states that if there are n unsymmetrical phasors, then this phasors can be resolved into n-1 symmetrical phasors and 1 co-phasal quantity. Since here dicussions are related to current and voltage vectors, therefore there are only 3 unsymmetrical vectors which can be divided into 2 symmetrical vectors and 1 co-vector.

Now since faults can be permanent or temporary i.e. transient for a short period, hence for the simulation purpose both the faults have been considered.

• 1
• 2
• 3
• 4
Logic for Fault control both temporary and permanent used in the simulation; (Source:- Taken fromour simulationat IIT Delhi)

## Phasor measurement units (PMUs)

In a power system operational paradigm, focussing on the transmission levels; data are gathered through the sensors like the remote terminal units which get data from CTs and PTs. Traditionally these analog values or signals are given to the SCADA technology; which communicates data and sends it to the control centers. In the control centers, Energy Management System (EMS) is present; and all the data that come in; get into different algorithm that run as engines in EMS. Next EMS tells the user that whether the power system is operating as they wanted to operate, or whether any problem has occured and what are the control actions to solve the problem.

• Next important factor is how fast the data is being transferred from SCADA to the EMS. Typically the output data rate of SCADA has been once in 4-6 seconds i.e. data fed to the algo get the data every 4-6 seconds.

If some disturbance takes place, and if one has to study those disturbances or has to study how the system behaves after the disturbance (post disturbance analysis)then if the output rate is just once in every 4-6 seconds then probably that is not enough. Because the power system tend to change very quickly; it's a dynamic system, so if we are not able to capture the values of power system in a good percentage, then we loose our data. Therefore we are not able to view the exact behaviour of the power systemand if the exact behaviour is not extracted, then there might be a chance of ending up with wrong control actions.

As a result, data is required at a much faster rate that helps to stagnate and apply all the numerical values of data. From here to eliminate this diadvantage emerges a newr technology named synchrophasor.

• In SCADA, we are trying to gather data from different parts in power system and even though the data might have some timing, but the problem is sending data to a control center and some other substation located quite far away and is getting to send the data as well to the same control center, because of the geographical distance the substation which is closer will get to send the data much quicker than the other substation. Hence local clocks are used to stand the data coming from the local substations; so that we get to know what time those data belong to. So time synchronized wide area system data is required to have an accurate view of the entire power system.
• SCADA tech. gives magnitudes of different electrical quantities like voltages, currents, frequency and so on. Networks one has to analyse are AC which has both magnitude as well as angle phasors. Therefore looking into just magnitude leads to loosing of information and ending up monitoring the system wrongly.
PMU block in RSCAD; (Source:- Taken from oursimulationatIIT Delhi)

So to solve all the above mentioned problems PMUs act as life saviours. PMUs are devices that produce synchro-phasors, frequency and rate of frequency estimates from voltage or current signals received from PTs and CTs and a time synchronized signal which we can get from GPS satellites through GPS clock.

## Power system protection

Two main equipments serves the purpose:- relays and circuit breakers. The main operation of the protection system is to isolate the healthy section from the faulty section of the power system in shortest time possible causing the minimum disturbance. If the main protection fails to operate, there should be a backup protection for which proper relay co-ordination is necessary.

Relay Operation; (Source:- Sona Electronics Team. (2015, May 18). Retrieved June 28, 2019, from http://electronicsteam.blogspot.com/2015/05/4protective-relays.htmlB)

Working Of Relays:-When fault occurs in X, information regarding the abnormal magnitudes of currents and voltages are sent to the relay through CT and PT. Therefore fault sensed and the trip circuit is actuated by the relay. Hence current flows in the trip circuit and as the trip coil of circuit breaker (CB) is energized; CB gets actuated and it operates for the opening of the system. Therefore the main functions are to sound an alarm, close the trip circuit of a circuit breaker and disconnect the abnormally operating part from the healthy part of the circuit. Certain Terminologies play defining role in the working of a relay.​[28]

• Pick-up level of actuating signal - After crossing the threshold value (actuating quantity may be current or voltage), relay initiates to operate. As the value increases the electromagnetic effect of the relay coil is heightened and after exceeding a certain level (pick-up), the moving mechanism of the relay just begins to move.
• Reset Level - Value of current or voltage less than which a relay will open its contacts and come in normal position.
• Operating time of relay - Time which elapses between the instant when actuating quantity exceeds the pickup value to the instant when the relay contacts close.

Requirement of auto-reclosers in circuit breakers:-Now depending on the fault type whether permanent or temporary, relay logic needs to be built. Since in temporary, the logic should be such that it uses auto reclosing technique after the fault removal but for permanent, the logic should have facility for manual reclosing since here the fault has to be removed manually by the utility workers. [41]

Since around 80% faults are transient, a lot of unnecessary downtime will result if the protection scheme simply isolates the section of a network where a fault occurs and then halts for someone to go out into the field, check the status of the fault and reset the circuit breaker. A much more effective strategy is to isolate the affected part of the network and then, after a short delay, re-energise it to determine whether the fault has cleared. If it has, the network can operate normally but if it hasn’t, the affected area can once again be isolated. This isolation followed by re-energisation process is exactly what an automatic circuit recloser (ACR) does. In principle, it’s a simple device. It comprises of a circuit breaker with protection relays, a mechanism that will oblige it to be closed after a trip automatically, a controller that provides the auto-reclose service.[42]

Protection Schemes used in relays for the simulation:-In this article main focus was to protect the transmission and distribution lines from faults, hence certain type of relays comes into picture which helps us to attain so. Distance relay with directional features specifically MHO relay are mainly used in the transmission lines with overcurrent directional relays and differential relays in the distribution lines. RSCAD provides the facility to use their inbuilt models of distance protection components.

1. Distance Protection:- It depends upon the distance of feeding point to the fault. The time of operation of such a protection is a function of the ratio of current and voltage i.e. admittance or voltage and current i.e. impedance. Since MHO relay used; is an admittance relay, therefore discussions related to it are more specific in this report. MHO relay is a high speed relay and is also known as angle impedance relay. In this relay administering torque is purchased by the volt ampere element and the restraining torque is refined due to the voltage element. Therefore a MHO relay is a voltage restrained directional relay. The equation of the torque can be given as :-

T = K1VI cos(θ-τ) - K2V2 - K3

where K3 is the control spring effect and θ and τ are defined as positive when I lags behind V. At balance point:-

0 = K1VI cos(θ-τ) - K2V2 (considering spring effect as 0)

K1VI cos(θ-τ) = K2V2

Therefore, Z = K1/K2cos(θ-τ) which reflects an equation of a circle passing through origin whose diameter is K1/K2 =ZR (ohmic setting).

RSCAD view of MHO relay component or Distance function block; (Source:- Dean. (2013). 8relay)
Operating characteristic of MHO relay; (Source:- Shanaya. (n.d.). Engineering Notes. Retrieved July 3, 2019, from http://www.engineeringenotes.com/electrical-engineering/distance-relays/distance-relays-and-its-classification-devices-electrical-engineering/32872)
MHO relay

In Fig 23 (a) the inputs to the component would be the voltages and currents of different phases detected by the secondary of CTs and PTs connected in the main transmission lines. Whereas the output will be in the form of WORD that needs to be converted to BITs to be accessed by other logics.

Output Signals for the Distance Function Block; (Source:- Dean.(2013). 8relay)
 Signal Description Format 21-start Distance Element Bits: Integer (word) 1: Zone1 start signal 2: Zone2 start signal 3: Zone3 start signal 4: Zone2 time delay operate signal 5: Zone3 time delay operate signal 1P/3PT Tripping Element Bits: Integer (word) 1: Trip A phase 2: Trip B phase 3: Trip C phase 4: Trip 3 phase RECL Reclosing Element Bits: Integer (word) 1: Reclose Signal 2: Reset Signal 3: Cycling Signal 4: Lockout Signal 68-B or 68-T Out of Step Element Integer

Therefore the logic used here is such that it compares the admittance or even impedance with the threshold zone impedance already set inside the relay. If the impedance comes out to be less than the set impedance then it can be concluded that fault has occured in that zone due to increase in current flow. [43]

Summary of impedance formula according to type of faults
 EVENT TYPE IMPEDANCE FORMULA AG VA/(IA + 3K0I0) BG VB/(IB + 3K0I0) CG VC/(IC + 3K0I0) AB and ABG VAB/IAB = (VA-VB)/(IA-IB) BC and BCG VBC/IBC = (VB-VC)/(IB-IC) CA and CAG VCA/ICA = (VC-VA)/(IC-IA) ABC and ABCG VA/IA  or  VB/IB or VC/IC

2. Zonal and Directional Protection with Backup relay:- In the figure shown below, the circles represented by black lines show the zones of relay A attached to bus, B1. Simimlarly the circles with blues lines are zones of relay B on bus B2. Now zone 1 covers 80% of the transmission line on which the relay is fixed. Zone 2 covers around 120% line in the forward direction and zone 3 covers around -100% of the transmission line. Now building up of 3 zones is basically because of the action as a backup relay. Suppose a relay fails to work due to mechanical defect of moving parts of the main relay, failure of DC power supply to the relay, failure of tripping pulse to the breaker from relay, or because of failure of current or voltage to the relay from CT or PT circuits etc, then the relay acting as backup comes into view to protect the zone that the faulty relay should have done. Hence the operating times of zone 1 is instantaneous while the starting time of zone 2 is after 20 cycles of start of the fault, whereas zone 3 starts after 60 cycles of start of the fault.

• 1
• 2
• 3
Illustration on Zonal And directional protection
• If F1 occurs:- then the primary relays should work. But suppose if they fail, then relay A may detect as it falls in zone 3 after 60 cycles of fault operation. It would stop the flow of current from right side. The left side flow current will be stopped by backup relays present on the other side.
• If F2 occurs:- Besides the primary relays, it can be detected by relay A as well as relay B as it falls under zone 3 and zone 2 respectively. Although the operation of relay B will be faster since it will operate only after 20 cycles of fault starting, relay A will detect if even relay B has certain failures that too after 60 cycles of operation. But both the relays stop the current flow from right by opening the corresponding breaker. The left side flow current will be stopped by backup relays present on the other side.
• If F3 occurs:- Relay A would detect instantaneously since its in zone 1, while relay B will detect after 20 cycles of Fault operation. Therefore relay A stops the current flow from left side and the relay B stops the current flow from the right end.
• If F4 occurs:- Relay A and B would detect almost at the same time instantaneously since its in the zone 1 of both the relays. Hence both the currents from left and right ends are terminated.
• If F5 occurs:- Relay B would detect instantaneously since its in zone 1, while relay A will detect after 20 cycles of Fault operation. Therefore relay A stops the current flow from left side and the relay B stops the current flow from the right end.
• If F6 occurs:- Besides the primary relays, it can be detected by relay A as well as relay B as it falls under zone 2 and zone 3 respectively. Although the operation of relay A will be faster since it will operate only after 20 cycles of fault starting, relay B will detect if even relay A has certain failures that too after 60 cycles of operation. But both the relays stop the current flow from left by opening the corresponding breaker. The right end flow of current will be stopped by backup relays present on the other side.
• If F7 occurs:-the primary relays should work. But suppose if they fail, then relay B may detect as it falls in zone 3 after 60 cycles of fault operation. It would stop the flow of current from left side. The right side flow current will be stopped by backup relays present on the other side.

The directional features of power flow can be justified by whether the relay has voltage and current in the leading or lagging phase. Suppose if originally current is leading voltage by suppose δ phase angle then after fault current reverses and becomes out of phase, hence ultimately lagging voltage. The power flow equations thus changes sign showing the directional characteristics.

3. Over-current Protection:- Similar to distance protection, here too a comparison is made between the rms current flowing through the transmission lines and the setting threshold values. If the values move on to a higher pitch than the threshold then over protectional relays help to trip the corresponding circuit breakers by sending the trip signal. The logic used has been shown below.

• 1
• 2
Logic for overcurrent protection; (Source:- Taken from oursimulationat IIT Delhi)

## Implementation of DG resources

DG resources used in the simulation are PV models, Li-ion batteries for storage of excess power and diesel generators. Hence studies regarding all these sources are elaborated here.

• PV model :- The PV controls are based on the decoupled dq current control startegy with internal startups and protection logic blocks. To know more about dq current control of grid tied voltage, studying the following resource would be plentiful. The PV sources are interfaced to the AC grid using dynamic average model of their DC/AC voltage source converter.
• 1
• 2
• 3
PV logic used in the simulation; (Source:- Taken from oursimulationat IIT Delhi)

The Photovoltaic module can be constructed in 2 ways : as a 4 parameters model or as a 5 parameters model. The difference between them is in a shunt resistance (Rsh). [44]

4 parameters scheme
5 parameters scheme
PV Schemes; (Source:- Metelev, I. S., Isakov, R. G., Ferenetz, A. V., Gilmanshin, I. R., Gilmanshina, S. I., & Galeeva, A. I. (2018, September). Using of functional capability of RTDS hardware and software complex in the educational process of electrotechnical cpeciality master’s training in ecomony global digitalization conditions. In IOP Conference Series: Materials Science and Engineering (Vol. 412, No. 1, p. 012053). IOP Publishing.)

The circuit with shunt resistance is more precise, but it requires more complex calculations. Therefore the 4 parameters scheme is considered and the equation for current- voltage dependence will be as follows:-

1)

2)

3)

4) $\alpha=\alpha_{ref}(\frac T{T_{ref}})$

5) ${V}_{T}=\frac{kT}{q}$

where, Iph = photocurrent generated by a single solar cell, A

I0 = Diode inverse saturation current, A

Nc = No. of solar cells connected in series in a solar module

α = Diode ideality factor

VT = diode thermal voltage, V

G = solar insolation, W/m2

ki = temperature coefficient at short-circuit current

T = temperature, К

Eg = photovoltaic element energy gap, eV

αref, Iphref = calculated value

Gref and Tref = insulation and temperature parameters at which solar module parameters were measured (typically Tref = 250ºC, Gref = 1000 W/m3)

q = electron charge, 1.602∙10-19 C

k = Boltzmann’s constant, 1.38∙10-23 J/K

The passport parameters of solar panels are:
Voc− open-circuit voltage, V;
Vm- voltage at maximum power, V;
Isc– short-circuit current, A;
Im - current at maximum power, А;
ki– temperature coefficient at short-circuit current;
kv– temperature coefficient at open-circuit voltage

PV cell to array; (Source:- University of Central Florida. (n.d.). The Florida Solar Energy Center (FSEC). Retrieved June 21, 2019, from http://www.fsec.ucf.edu/en/consumer/solar_electricity/basics/cells_modules_arrays.htm)
Overlapped bypass diodes; (Source:-Ekpenyong, E. E., & Anyasi, F. (2013). Effect of shading on photovoltaic cell. IEEE, 01-06)
No-overlapped bypass diodes; (Source:- Ekpenyong, E. E., & Anyasi, F. (2013). Effect of shading on photovoltaic cell. IEEE, 01-06.)
Bypass Diodes
• Li-ion Batteries:- The working principle of a grid-tied PV system is synchronizing the installation with the electric utility voltage, frequency, and phase, which basically turns the PV system into a part of the grid. If at any given moment the system output exceeds the owner’s demand by such an amount that the surplus production is automatically exported to the electric utility grid and metered even after redirecting to a battery bank by a charge controller.
Li-ion logic; (Source taken from our simulationat IIT Delhi)

Here Li-ion batteries act as chargeable batteries i.e. when in discharging mode it generates power that feeds the load but when in charging mode it takes in power from the surplus generation. The rate of charging and discharging are controllable by users and hence act as a great source in times of requirement. Similar to PV arrays they also produce DC power which is converted to AC through DC/AC converter.

• Diesel Generator:-Diesel Generator produces electricity but runs on diesel fuel. It can also be used as and when required. Logic behind the building of Diesel Generator is shown below.
Logic behind usage of Diesel Generator; (Source:- Taken from oursimulationat IIT Delhi)

Since all the simulation is done on real time, hence it would be better if the load is varied and how synchronization is done by the power generators are observed. Synchronization involves voltage and current magnitudes and phasors as well as frequency.

Generator takes mechanical energy as input and gives electrical energy as output which is equal to load requirement. Now if suddenly load is increased by 10%, at this instant electrical energy required is more than the mechanical energy provided. So, stored rotational kinetic energy in rotor of generator is converted to supply excess electrical energy.

Ns=120*f/P where Ns is the speed of the rotor, f is the frequency and P is the no. of poles of the Machine.

Now suppose electrical generation exceeds load requirement which indicates that Ns is more and so frequency generated is more which is to be brought down by slowing down the rotor speed and maintained at reference. Similarly the opposite happens when electrical generation is less than the load requirements. In this way frequency synchronization is done.

Now Load data is taken from an official NYISO which has an excel sheet that contains varied load after every 5 mins in 11 different regions.(https://www.nyiso.com/load-data) This excel data has to be fed in the RSCAD model where simulation is to be done. Now data can only be fed through port. So using UDP protocol and writting code in Python data are sent to the required ports from where they are accessed. Here is the following code snippet.

import socketimport timeimport openpyxlstr2=""str1=""wb=openpyxl.load_workbook("9BUS_RAW LOAD DATA FEED from NYISO.xlsx")ws=wb["Mysheet"]maxrow=ws.max_rowmaxcol=ws.max_columnhost="10.194.11.44"port=5454s= socket.socket(socket.AF_INET,socket.SOCK_DGRAM)#means we are using IPv4 and UDP protocol)initial = time.time()for i in range (2,maxrow+1):    for j in range(2,maxcol+1):        cell_obj = ws.cell(row=i, column=j)        str2=str2 +" "+str(cell_obj.value)    data = bytes(str2,'utf-8')    s.sendto(data, (host, port))    print("sent:",data)    print( time.time())    time.sleep(300)    str2=str1wb.save("9BUS_RAW LOAD DATA FEED from NYISO.xlsx")
Server side or sending side code snippet
import sockethost="10.194.11.44"port=5454s= socket.socket(socket.AF_INET,socket.SOCK_DGRAM)#means we are using IPv4 and UDP protocol)s.bind((host,port))while 1:    print(s.recv(40))
Client side or receiving side code snippet

## Method

• First of all the model is built i.e. connecting utility grid as well as microgrid from the draft window in RSCAD and the transmission lines data are fed from theT_Line window by specifying the length and positive and zero sequence impedances as required.
• Next the faults are made to occur in the transmission lines depending on what type of fault user needs to watch in the runtime window of the simulation. It may be a permanent fault provided by a switch or a temporary fault provided by a push button.
• Next signals from CTs and PTs are sent for sampling the voltage and current signals and then put into DFTs to extract samples of 60 Hz component.
• PMUs detect these data wherefrom the extracted data are sent to relays where decisions are made what step next to be taken and send them accordingly to the circuit breakers.
• When the Circuit breakers open from the utility grid, the microgrid becomes islanded as during fault ;frequency , current and voltage magnitudes and phasors are not syncronized. Due to lack of synchronization circuit breakers receive the trip signal from relays and get disconnected.
• The time period for disconnection may be a short period if temporary fault occurs. If permanent faults occur then first the circuit breakers trip then again reconnect after certain time interval hoping that the fault has cleared. When they find that the fault has not cleared they open once again and it continues for a total of 4 times until they quite surely understand that the fault occured is permanent and locks the breaker in the open position for the rest period of time. Workers from maintenance have to come and clear the fault and close the breakers to ensure continuity of power flow.
• Now during islanding a small checking is done which turns out to be quite helpful at the end. The load requirement and the power generated from the DG resources are compared. If load requirement is less than the power generated quickly a communicaton signal is sent to the inverter to shut the DG since this may burn the other home electrical appliances and even may trace back to the faulty transmission line location where the maintenance workers are working and may threaten their lives. If the load requirement is more than the DG, then do the necessity so as to synchronize and turn other DGs on if available else continue the DG to operate in low voltage operation until power from grid returns.

The 9 bus model used in the simulation is shown below:-

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
Final 9 bus power system model with DG sources installed; (Source:- Taken from oursimulation at IIT Delhi)

## RESULTS AND DISCUSSION

Runtime Model; (Source:-Taken from oursimulationat IIT Delhi) .

Besides creating the model in the Draft window, the above given figure provides a better outlook of a part of the model (preferably the load end side) in the Runtime window. The grid bus shown represents the model before the distribution lines including the transmission lines and the synchronous generators. The grid switch is kept for manually tripping the circuit breaker or more technically performing intentional islanding. Unintentional islanding occurs when the fault button is pressed and the spark sign turns red. Automatically after some time the circuit breaker opens (turn red) hence causing unintentional islanding. The auto opening of the circuit breaker is possible because of addition of relay working to the model. If the fault is temporary, then auto recloser of the circuit breaker also takes place. If the fault is permanent, then one has to manually turn on the circuit breaker for continual flow of power from generator end to the load end. Even provisions are made for the battery to charge and discharge as and when required. The switches infront of the DERs provide the ability to use DERs according to our requirements. Through the sliders one can adjust the active and reactive power flow into the load.

Normal Condition when both Microgrid and utility grid are working; (Source:- Taken from our simulationat IIT Delhi)
• 1
• 2
• 3
• 4
• 5
• 6
Signals during Islanded mode of operation; (Source:- Taken from oursimulationat IIT Delhi)
After fault clearing, signals trying to synchronize with the main utility grid; (Source:- Taken from oursimulationat IIT Delhi)
Runtime results window

a) First of all the runtime model is made to run. Initially the DERs are connected and the load is supplied by both the micro grid and the utility grid. In Fig34a) the battery is set to discharge (Pref = +0.5) and the excess power is sent to the grid. Receiving power by any source is provided with a minus sign, while sending power is administered with a positive sign. Next the state of charge (SOC) of the battery is observed. If it falls below 50%, then PREF value is set to -0.5 for pushing it to charging mode. But overall there would be no such disturbance shown in the signals as long as synchronized operation takes place.

b) Next cause a fault to happen by pressing the push button switch. (The change in colour of the spark sign denotes it) It will come into observation that the Square representing CB will also turn red representing the diconnection of the main utility grid. This can be noticed from Fig34b) which shows the current in the main circuit breaker as well as the grid turning to zero. At the same time the diesel generator present tries to compensate and feed the load with the sufficient power that the main grid was providing. Hence an increase in the overall current signal of the circuit breaker is shown.

c) Now as a temporary fault was forced to happen, it gets cleared after sometime; and the reclose signal from the relay is sent to the circuit breaker. The circuit breaker recloses and checks whether the fault persists or not. As it happens to see the fault to be recleared; it tries to synchronize the signals with the other generators operating and helps in continual flow power to the load. In Fig34c), how the signals try to resynchronize is shown, after which it returns to the state shown by Fig 34a).

## CONCLUSION AND RECOMMENDATIONS

Based on the review shown above, clearly each IdM has certain advantages and disadvantages. Passive methods are simple and can detect the islanding condition using conventional protection methodologies. However these methods fail to detect the islanding condition when the DG and the load power are balanced, therefore suffering from large NDZ. The active methods provide a fast detection speed and small NDZ, but their impact on power quality may deteriorate the performance of the power system. Remote methods have fast detection speed, high reliability and the ability to perform well with multiple system configurations. However the computational efficiency, implementation cost and malfunctioning due to the failure of communication links rae the main limitations associated with remote IdMs.

The signal processing IdMs are becoming more reliable and efficient due to the application of signal processing, pattern recognition and artificial intelligence tools. The intelligent methods, due to the presence of various training and testing procedures, suffer from a large computational burden which makes them less favourable in comparison with other signal processing IdMs.

Therefore based on the above discussions, it can be observed that the signal processing methods are preferred on the basis of simplicity, cost, low computational burden, accuracy and real-time industrial applications.

Coming to the future scope of the research, which the main aim is; behind studying all such results and observations, have a much broader aspect. There are two ways in which a relay can get disturbed signals. One being the situation when really fault has occured and the other when cyber criminals feed in wrong data to the CTs and PTs thus manipulating the original data and providing an erroneous data to the relay. Now anticipating the relay operation, if it acts just according to the CT and PT adjoint to the relay and send a trip signal to the circuit breaker then the outcome obtained may not be a completely gratifying one, hence implementing a wrong decision which may thwart the smooth run of a particular EPS. Hence the main reason behind studying the PMU data received is to form a trend that could be fed to the relay depending on which it would make decisions. This is because if an invader is able to manipulate a single PMU data, then hopefully the other PMUs acting when a fault has occured would provide with correct result. Now getting out the trend that the PMUs follow when a fault occurs is the main area of research since cyber protection comes along with the need to make the grid smarter through different communication channels.

## ACKNOWLEDGEMENTS

I would like to express my gratitude to my honourable esteemed guide, Prof. Swades De, Professor of Department of Electrical Engineering in IIT, New Delhi for his guidance and constant support. His perspective of how to approach in my work has inspired me to carry on through out the internship. I am glad to work with him.

I am very thankful to my mentor Mr. Akash Kumar Mandal, PhD student at IIT Delhi, for his endless help that he has provided during my internship. His vast experience in power system and linking the subject with communication touch has been an inexhaustible source of information. I am most grateful to him for making me understand all the brand new things that I overcome in such simple process by giving eternal practical examples. I would like to thank him for all the mindful and thought stimulating discussions we had, which prompted us to build logics much easier and cooler than before without dislodging any of the main characteristics. His thoughtfulness is a gift which I will always like to treasure. This dissertation stands as a testament to his unconditional encouragement.

I cannot end without thanking Indian Academy of Sciences, Indian National Science Academy and The National Academy of Sciences, India for selecting me after completion of second year in B-tech to this summer research internship and providing me with such wonderful and great experiences and oppurtunities.

#### Notes

• Only around 7% of faults are permanent. The vast majority of faults – around 80% – are transient, caused by events like lightning strikes and arcing, will disappear in less than a second. The remaining 13% of faults are semi-permanent and are typically caused by animals or branches bridging the power lines. Even these faults usually burn away and clear in a relatively short time.

• Now in transmission lines there should be 3 pole tripping while in distribution lines single pole tripping will be more helpful as single pole tripping helps in continuation of power supply to the other phases while 3 pole tripping helps in better synchronization once the cicuit breakers are reconnected.

• A typical silicon solar cell generates a volatge about 0.7V. To achieve the necessary current and voltage levels, solar cells are usually connected in parallel or in series and creates a solar module, which generates higher voltage of about 16-36V. This modules when connected in series create a PV panel. Now PV panels when grouped together make a PV array.

• [45]Shading provides a massive crunch on the administration of solar photovoltaic panels. It is clear that the creamy solution is to dodge shading altogether, though practically this is impossible due to circumstances like cloud, rain etc.. Even if a minute sector of the solar panel is in shade, the conduct of the total photovoltaic panel will consequently curtail since solar panels actually subsist a number of photovoltaic cells which are wired in sync into a series circuit. This channelizes to the fact that when the power harvest of a single cell is irresistibly reduced, the power output for the total model in series is scaled down to the current level crossing the weakest cell. Therefore, a little amount of shading can drastically decrease the performance of an entire solar photovoltaic panels system. These PV modules are composed of photovoltaic cells (PV cells) serially or parallely composed with diodes indulged in different configurations. Furthermore, the modules have diodes that permit the current flows through a different path, when enough cells are shaded or damaged. There are two typical configurations of bypass diodes: overlapped and no-overlapped. It should be noted that the analysis in modules with overlapped diodes is a more complex one, because there may be different paths for current flow.

#### References

• Kim, M. S., Haider, R., Cho, G. J., Kim, C. H., Won, C. Y., & Chai, J. S. (2019). Comprehensive Review of Islanding Detection Methods for Distributed Generation Systems. Energies12(5), 837.

• Menon, D., & Antony, A. (2016, March). Islanding detection technique of distribution generation system. In 2016 International Conference on Circuit, Power and Computing Technologies (ICCPCT) (pp. 1-5). IEEE.

• Krishnan, G., & Gaonkar, D. N. (2012, December). Intentional islanding operations of distributed generation systems with a load shedding algorithm. In 2012 IEEE International Conference on Power Electronics, Drives and Energy Systems (PEDES) (pp. 1-6). IEEE.

• Noor, F., Arumugam, R., & Vaziri, M. Y. (2005, October). Unintentional islanding and comparison of prevention techniques. In Proceedings of the 37th Annual North American Power Symposium, 2005. (pp. 90-96). IEEE.

• Mahat, P., Chen, Z., & Bak-Jensen, B. (2008, April). Review of islanding detection methods for distributed generation. In 2008 third international conference on electric utility deregulation and restructuring and power technologies (pp. 2743-2748). IEEE.

• Maki, K., Kulmala, A., Repo, S., & Jarventausta, P. (2007, July). Problems related to islanding protection of distributed generation in distribution network. In 2007 IEEE Lausanne Power Tech (pp. 467-472). IEEE.

• Ugale Bhagyashri, and N.J. Phadkule. “Islanding Detection Based on Reactive Power Control.” International Journal of Engineering Science and Computing, vol. 7, no. 5, May 2017, pp. 11683–11687.

• Trujillo, C., Velasco, D., Figueres, E., & Garcera, G. (2010). Local and remote techniques for islanding detection in distributed generators. In Distributed Generation. IntechOpen.

• Phipps, J. K., Nelson, J. P., & Sen, P. K. (1994). Power quality and harmonic distortion on distribution systems. IEEE transactions on industry applications30(2), 476-484.

• Redfern, M. A., Usta, O., & Fielding, G. (1993). Protection against loss of utility grid supply for a dispersed storage and generation unit. IEEE transactions on power delivery8(3), 948-954.

• Redfern, M. A., Barrett, J., & Usta, O. (1995). A new microprocessor based islanding protection algorithm for dispersed storage and generation units. IEEE Transactions on Power Delivery10(3), 1249-1254.

• Rostami, A., Jalilian, A., Hagh, M. T., Muttaqi, K. M., & Olamaei, J. (2017, October). Islanding detection of distributed generation based on rate of change of exciter voltage and mechanical power with circuit breaker switching strategy. In 2017 IEEE Industry Applications Society Annual Meeting (pp. 1-8). IEEE.

•  Li, C., Cao, C., Cao, Y., Kuang, Y., Zeng, L., & Fang, B. (2014). A review of islanding detection methods for microgrid. Renewable and Sustainable Energy Reviews35, 211-220.

• Pai, F. S., & Huang, S. J. (2001). A detection algorithm for islanding-prevention of dispersed consumer-owned storage and generating units. IEEE Transactions on Energy Conversion16(4), 346-351.

• Karimi, H., Karimi-Ghartemani, M., & Iravani, M. R. (2004). Estimation of frequency and its rate of change for applications in power systems. IEEE Transactions on Power Delivery19(2), 472-480.

• Budoya, D. E., & Baptista, F. G. (2018). A comparative study of impedance measurement techniques for structural health monitoring applications. IEEE Transactions on Instrumentation and Measurement67(4), 912-924.

• Von Jouanne, A., & Banerjee, B. (2001). Assessment of voltage unbalance. IEEE transactions on power delivery16(4), 782-790.

• Llaria, A., Curea, O., Jiménez, J., & Camblong, H. (2011). Survey on microgrids: unplanned islanding and related inverter control techniques. Renewable energy36(8), 2052-2061.

• O'kane, P., & Fox, B. (1997). Loss of mains detection for embedded generation by system impedance monitoring.

• Yafaoui, A., Wu, B., & Kouro, S. (2011). Improved active frequency drift anti-islanding detection method for grid connected photovoltaic systems. IEEE Transactions on Power Electronics27(5), 2367-2375.

• Hatata, A. Y., Abd-Raboh, E. H., & Sedhom, B. E. (2018). Proposed Sandia frequency shift for anti-islanding detection method based on artificial immune system. Alexandria engineering journal57(1), 235-245.

• Velasco, D., Trujillo, C. L., Garcerá, G., & Figueres, E. (2010). Review of anti-islanding techniques in distributed generators. Renewable and sustainable energy reviews14(6), 1608-1614.

• De Mango, F., Liserre, M., Dell'Aquila, A., & Pigazo, A. (2006, August). Overview of anti-islanding algorithms for PV systems. Part I: Passive methods. In 2006 12th International Power Electronics and Motion Control Conference (pp. 1878-1883). IEEE.

• Yin, J., Chang, L., & Diduch, C. (2006, June). A new hybrid anti-islanding algorithm in grid connected three-phase inverter system. In 2006 37th IEEE Power Electronics Specialists Conference (pp. 1-7). IEEE.

• Wang, W., Kliber, J., Zhang, G., Xu, W., Howell, B., & Palladino, T. (2007). A power line signaling based scheme for anti-islanding protection of distributed generators—Part II: Field test results. IEEE Transactions on Power Delivery22(3), 1767-1772.

• Bayrak, G., & Kabalci, E. (2016). Implementation of a new remote islanding detection method for wind–solar hybrid power plants. Renewable and Sustainable Energy Reviews58, 1-15.

• Paithankar, Yeshwant G., and S.R Bhide. Fundamentals of Power System Protection. PHI Learning Private Limited, 2014.

• Pei, S. C., Tseng, C. C., & Yeh, M. H. (1999). A new discrete fractional Fourier transform based on constrained eigendecomposition of DFT matrix by Lagrange multiplier method. IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing46(9), 1240-1245.

• Gilles, J. (2013). Empirical wavelet transform. IEEE transactions on signal processing61(16), 3999-4010.

• Bahrani, B., Karimi, A., Rey, B., & Rufer, A. (2012). Decoupled dq-current control of grid-tied voltage source converters using nonparametric models. IEEE Transactions on Industrial Electronics60(4), 1356-1366.

•  Barros, J., & Diego, R. I. (2005). Application of the wavelet-packet transform to the estimation of harmonic groups in current and voltage waveforms. IEEE Transactions on Power Delivery21(1), 533-535.

• Thompson, Stan. “Auto-Recloser-Safety and Minimising Downtime.” Transmission and Distribution; Wilson Transformer; Australasian Power Technologies, 2018, pp. 12–15.

• Samantaray, S. R., Samui, A., & Babu, B. C. (2011). Time-frequency transform-based islanding detection in distributed generation. IET renewable power generation5(6), 431-438.

• Tay, A. L., Zurada, J. M., Wong, L. P., & Xu, J. (2007). The hierarchical fast learning artificial neural network (HieFLANN)—an autonomous platform for hierarchical neural network construction. IEEE Transactions on Neural Networks18(6), 1645-1657.

• Swain, P. H., & Hauska, H. (1977). The decision tree classifier: Design and potential. IEEE Transactions on Geoscience Electronics15(3), 142-147.

• Kowalski, P. A., & Kusy, M. (2017). Sensitivity analysis for probabilistic neural network structure reduction. IEEE transactions on neural networks and learning systems29(5), 1919-1932.

• Huang, K., Jiang, H., & Zhang, X. Y. (2017). Field support vector machines. IEEE Transactions on Emerging Topics in Computational Intelligence1(6), 454-463.

• Yang, J. M., Cheng, K. W. E., Wu, J., Dong, P., & Wang, B. (2004, November). The study of the energy management system based-on fuzzy control for distributed hybrid wind-solar power system. In Proceedings. 2004 First International Conference on Power Electronics Systems and Applications, 2004. (pp. 113-117). IEEE.

• Saran, A., Kankanala, P., Srivastava, A. K., & Schulz, N. N. (2008, June). Designing and testing protective overcurrent relay using real time digital simulation. In Edinburgh, Scotland: Grand Challenges in Modeling and Simulation, Summer Simulation Conference.