Article

Comparative Study of Different Approaches forIslanding Detection of DistributedGeneration Systems

Ashish Shrestha 1,2,* , Roshan Kattel 3 , Manish Dachhepatic 3, Bijen Mali 4, Rajiv Thapa 5,Ajay Singh 3, Diwakar Bista 1,2, Brijesh Adhikary 1, Antonis Papadakis 6 andRamesh Kumar Maskey 7

1 Department of Electrical and Electronics Engineering, Kathmandu University, Dhulikhel 45200, Nepal2 Centre for Electric Power Engineering, Kathmandu University, Dhulikhel 45200, Nepal3 Khwopa College of Engineering, Tribhuvan University, Bhaktapur 44800, Nepal4 NEA Engineering Company, Kathmandu 44600, Nepal5 Center for Electricity Trade Research and Facilitation, Kathmandu University, Dhulikhel 45200, Nepal6 Department of Electrical Engineering, Frederick University, Nicosia 94014, Cyprus7 Department of Civil and Geomatics Engineering, Kathmandu University, Dhulikhel 45200, Nepal* Correspondence: ashish.shrestha@ku.edu.np; Tel.: +977-11-661-399; Fax: +977-11-661-443

Received: 21 May 2019; Accepted: 22 July 2019; Published: 23 July 2019�����������������

Abstract: The issue of unintentional islanding in grid interconnection still remains a challenge ingrid-connected, Distributed Generation System (DGS). This study discusses the general overview ofpopular islanding detection methods. Because of the various Distributed Generation (DG) types, theirsizes connected to the distribution networks, and, due to the concern associated with out-of-phasereclosing, anti-islanding continues to be an issue, where no clear solution exists. The passive islandingdetection technique is the simplest method to detect the islanding condition which compares theexisting parameters of the system having some threshold values. This study first presents anauto-ground approach, which is based on the application of three-phase, short-circuit to the islandeddistribution system just to reclose and re-energize the system. After that, the data mining-decisiontree algorithm is implemented on a typical distribution system with multiple DGs. The results fromboth of the techniques have been accomplished and verified by determining the Non-Detection Zone(NDZ), which satisfies the IEEE standards of 2 s execution time. From the analysis, it is concludedthat the decision tree approach is effective and highly accurate to detect the islanding state in DGs.These simulations in detail compare the old and new methods, clearly highlighting the progress inthe field of islanding detection.

Keywords: islanding detection; distributed generation system; auto-grounding; data mining

1. Introduction

In the present power system with distributed generators (DGs), there are some issues that arenot resolved yet. One of these issues involves a fast and very accurate detection of islanding [1,2].The number of DGs introduced into electricity distribution systems is increasing day by day, and itis the most challenging concept in modern power system scenarios, since DG islanding can causedegraded proficiency, excellence, reliability and quality of supply [3,4]. The concept of DistributedEnergy Resources (DER) is moving from being a local issue towards a system issue [5]. With theconnection of DGs, the power system becomes more complicated [6]. With the introduction of thereference power transition link, the grid connection process is further ensured to be smooth [7,8].The main difficulties that researchers encounter are not the lack of data, but the efficiency of using

Appl. Syst. Innov. 2019, 2, 25; doi:10.3390/asi2030025 www.mdpi.com/journal/asi

Appl. Syst. Innov. 2019, 2, 25 2 of 19

complex electrical quantities [9]. The issue of unintentional islanding in DGs grid interconnection stillremains a huge challenge in grid-connected systems, since the system may not succeed at activating theprotective devices during the islanded condition [10,11]. This results in the issues related to the powerquality, protection, reverse power flow and system stability [12]. Hence, the condition of islandingmust be detected and cleared as fast as possible as per the recommendation of grid codes [13]. Themaximum time to respond to the islanding detection is recommended to be 2 s by IEEE [14]. Theislanding protection devices being used in the future have to reflect this. Safety measures must betaken for correct operation of these new DGs so as to prevent uncontrolled and undesired powerinjection into the grid and possible damages to equipment and personnel [15]. Thus, a thorough studyis needed in solving this issue. Hence, anti-islanding with fast response time is essential for a DGconnected grid system [10,16]. To overcome the challenges from islanding DGs, researchers proposeda numerous model that deals with the consequences of intentional islanding, and clears it as fast aspossible [11,12,17–29]. Different organizations such as the IEEE, IEA and IEC also set the standardsand were updated regularly to emphasize the importance of islanding detection in a DG connectedgrid, so that a DG connected system will operate smoothly [14,30–32].

In this study, the comparative study of different islanding detection is discussed based onthe literature review. This paper is intended to study the comparative analysis between twotechniques: one traditional auto grounding method, and a relatively new data-mining-decision-treetechnique. This study first introduces the issues of islanding detection techniques and its requirements.Section 2 presents the literature review of different topics associated with the concept of islanding andanti-islanding methods, and their comparisons. In Section 3, applied methodologies and approachesare explained. Section 4 presents the research outcome and their discussion. Finally, in Section 5,conclusions are drawn.

2. Literature Review

2.1. Distributed Generation

Distributed generation (DG) is the system that indicates the energy sources that provide reliableelectricity to the end users. These technologies are site-specific and referred to small renewable energysources such as solar, wind, micro-hydro. It can be at a local level or end-user level. The local level DGis the renewable energy systems that electrify a community, whereas the end-user level includes thetechnologies adopted by the single consumer with similar characteristics of the system. The energystorage system plays a vital complementary role in DG. Although being the decentralization and thesmaller size, DG and the energy storage system are highly preferred for low energy costs with highreliability, and a more efficient, cost-effective and eco-friendly nature [33–40].

Figure 1 presents an example of a DGS containing micro-hydro power plant, solar PV, windturbine, diesel generator as the energy-producing units, the battery bank as a storage device, load,and finally grid as an optional choice. The voltage level of AC and DC bus bars may be the same ordifferent, but in general the DC level is low and controlled through the converter. A circuit breaker(CB) is placed between the grid and the AC bus system. Here, if the CB is open, the energy system,excluding the grid system, can be considered as IES, and, if the status of CB is closed, the system willbe grid-connected. In Figure 1, the system can be operated in either an isolated or grid-connectedmode. The flow of the current in the CB section may be unidirectional or bidirectional, depending onthe nature of the connection and the local policy of energy trading.

Appl. Syst. Innov. 2019, 2, 25 3 of 19

Figure 1. An example of a DGS.

Conventionally, the energy systems are of two types; isolated and grid connected system. Theisolated energy system (IES) is a collection of energy generator, load and the controlling componentsthat are designed to fulfill the energy demand of a community or a system in isolated form. There is noconnection of the national grid line to get power from any reliable service. Generally, IES seems to be apromising option for electrifying remote locations where grid extension is not feasible or economical.The integration of battery banks as a means of energy storage with different renewable energy systemscan enhance the system reliability and its overall performance. Therefore, appropriate choices ofgenerator sizes and the battery bank capacity are critical to the success of such renewable-based,isolated power systems. However, there are numerous challenges in renewable energy based IESsuch as high-cost, poor reliability and power quality, low load factor, periodic nature of renewablesources, problems in maintenance and monitoring activities [41]. Because of the periodic nature anddependency on weather factors, the characteristics of the energy generated from renewable energysources (RES) like solar and wind are unreliable, which can be improved by using sufficient storagedevices or by interconnecting multiple energy resources type, called optimal hybridization [42,43].An optimized hybrid energy system (HES) is quite a promising technique on behalf of cost, powerquality and reliability able to provide the electricity to the area where expansion of the national utility’sgrid system is quite difficult and expensive [44–46]. It can be made more efficient and cost-effectiveby reducing the disadvantages associated with these technologies [41,47]. However, there are otherproblems as well, such as: low load factor, low diversity factor, low reliability, protection, etc. in theisolated energy system. To resolve these problems, the mini-grid concept may be a good option. It isproven that the mini-grid concept contributes to the enhancement of social impact by improving thelocal governance structure via direct participation of the local community, since it is the best option toelectrify the isolated rural communities [45,48]. On the other hand, a grid-connected system is theenergy system containing an energy generator, as well as load components that are interconnectedwith a utility.

2.2. Islanding Detection Methods

2.2.1. Passive Method

Passive method is based on continuous monitoring of the local site system parameters like changein voltage, frequency, and harmonic distortion as these parameters vary during the islanding conditioninitiation [49]. These variations are directly influenced by the power mismatch between the supplyand the loads. Due to passive methods of islanding detection being dependent on the changes of theaforementioned parameters, the efficiency is thus dependent on the threshold set [3,50]. Setting a lowerthreshold results in false islanding detection as non-islanding faulty conditions of the power system

Appl. Syst. Innov. 2019, 2, 25 4 of 19

may cause system parameter changes too. The higher setting of the threshold results in the occurrenceof islanding condition being undetected as system parameter changes during islanding will not crossthe set thresholds [51]. Though cheap on implementation, inability of the method to detect islandingduring balanced islanding and large non-detectable zone (NDZ) renders this method unreliable to beused alone [52]. The rate of change of voltage/frequency relies on the fact that power imbalance causestransients in an islanded system and causes the voltage and frequency to vary. The over/under voltageand frequency, however, deal with the thresholds of the voltage and frequency, respectively, to detectan islanded system. The Harmonics Detection Method monitors the total harmonic distortion in thegrid, while the phase monitoring method constantly detects any sudden changes in the output currentof the inverter and its terminal voltage to detect the islanding condition [53]. Explanations to some ofthe passive methods are:

A. Rate of Change of Frequency:

This method uses the equation:

ROCOF =d fdt

=∆P2H

f . (1)

This method’s major drawback is that it cannot detect the islanding condition until the activepower imbalance is higher than 15%. It detects the load and generation mismatch, which in turn affectthe frequency. This change in frequency is then monitored to trip the relay, thereby disconnecting theDG and the power grid [54].

B. Rate of Change of Voltage:

This method involves calculating of rate of change of voltage given by:

ROCOV =dVdt

. (2)

It is based on the fact that the reactive power imbalance causes a transient change and voltagestarts to vary dynamically. Both ROCOF and ROCOV were tested with the conclusion that they candetect islanding condition within 0.125 s [55].

C. Over/Under Voltage and/or Frequency:

This method of detection of the islanding problem is the simplest of the available methods wherethe voltage and/or frequency is continuously measured and when the voltage or frequency is measuredto be outside the predetermined acceptable limits, the connection between the DG and the power gridis disrupted. In the case that the power rating of the DG is less than the power rating of the load, themismatch of the power will be compensated by the grid. However, when the islanding occurs, therewill not be any power supply from the grid, so in this case the voltage and frequency will deviate. Thatis, in this case, when there is islanding effect, the voltage at the Point of Common Coupling (PCC)should increase until active power P (DG) = Real Power of the load P (load). The Non-DetectionZone (NDZ) refers to the zone of within which the islanding condition is not detected by the islandingdetection method.

The frequency should also change such that the reactive power Q (DG) is equal to reactive powerof the load Q (load). This happens due to the fact that the DG inverter will seek a frequency where thecurrent–voltage phase angle of the load equals that of the DG inverter. This is because:

P(load) = (Vˆ2 (PCC))/R, (3)

Q (load) = Vˆ2 (PCC) (1 / (2× pi× f× L) − 2× pi× f×C). (4)

Appl. Syst. Innov. 2019, 2, 25 5 of 19

This deviation can then be effectively measured to find the islanding case. Furthermore, thedetection time of the over/under voltage is found to be 0.03 and 0.05 s and that of the over/underfrequency is found to be 0.02 in both cases [56,57]. The main drawback of the over/under voltage andover/under frequency anti-islanding method is its vast NDZ. This range is given by:

For active power,(V / Vmax)2

− 1 ≤ ∆P / P ≤ (V / Vmin)2 , (5)

For reactive power,

Qf(

1− (f/fmin)2)≤ ∆Q / Q ≤ Qf

(1− (f/fmax)2

). (6)

These two ranges act as the limiting value of the voltage and frequency for the detection for theislanding condition in the net metering.

2.2.2. Active Method

The active method of islanding detection deals with the perturbation to the system directly. Itmeans that an additional external variable, positive feedback or controlled change is introduced, whichinteract with the power system operation providing the significant change in system parameters duringislanded conditions and negligible change during proper operation of the DG in grid. It involvesthe feedback technique or some mechanism to find islanding through parameters change [58]. Thismethod is more accurate and effective than the passive method due to relatively small NDZs, but ithas slow response toward the perturbation. In addition, the power quality and system stability maydegrade due to the external variable introduced in the system [59]. However, this method requires aremarkably higher cost of implementation [49].

A. Active Frequency Drift (AFD):

In this method, a distorted output of the inverter feedback to the utility at a slightly higher orlower frequency at a point of common coupling. AFD frequency (f ’) can be calculated as:

f ′ = f (1/(1−C f )), (7)

where Cf is the chopping fraction. It is the ratio of zero conduction interval to half of the period ofvoltage waveform, which determines the difference between the frequency of inverter’s output currentand the frequency of grid’s voltage.

The added distorted frequency is used as a factor for detection of the islanding condition; however,under normal conditions, the added distorted frequency cannot change the frequency. The weaknessof this method is that it requires a decrease in the output quality.

B. Current Noticing:

The DG inverter appears as a current source to the utility:

i = I sin(w + β). (8)

In this method, a continuous variation is supplied in one of these factors:

∆V = ∆P/2*(R/P)1/2, (9)

where P is active power and R is resistance.At the case of islanding, there is a considerable change in the voltage. This change can be measured

to detect the islanding method. This method cannot be used in the multi-inverter case. This is because,as more variation is imposed on more inverters, the total variation will be less and cannot be used forthe detection islanding condition [51].

Appl. Syst. Innov. 2019, 2, 25 6 of 19

C. Sandia Voltage Shift:

This uses positive feedback to prevent islanding. A positive feedback of voltage and the DGinverter reduces its current and thus its output power. At normal operation, there is no effect when thepower is reduced. However, during the absence of the utility, there is a reduction in the voltage. Thisreduction results in the reduced DG inverter current. These reductions can be calculated to find theislanding condition. The weakness of this method is the inverter’s decreased efficiency and a decreasein the output power quality [50].

2.2.3. Hybrid Method

With the passive method being cheap and relatively unreliable and the active method beingreliable but with the trade-off in power quality, both of the discussed methods have their own setof advantages and disadvantages. The hybrid method aims to utilize the perks of both methodsdiscussed above while suppressing their disadvantages [60,61]. The passive method is used to monitorthe overall stability and occurrence of islanding, and, in the event of islanding, the active method isemployed to confirm the islanding event [3]. Some techniques include:

A. Technique Based on Voltage Unbalance and High-Frequency Impedance [3,62]:

This technique continuously monitors the voltage unbalance (passive method) as follows:

Voltage Unbalance = (Positive Cycle Voltage)/(Negative Cycle Voltage). (10)

Voltage unbalance can occur in cases like voltage spikes, switching actions and islanding. Thus, toconfirm the occurrence of islanding, positive feedback (active method) is used, where the high-frequencyvoltage is injected into the DG control loop. Furthermore, the voltage and current at the point ofcommon coupling will be measured to estimate the impedance at the injected frequency. In addition, ifnot within a threshold, the islanding event is then detected. The selected of high frequency should notbe equal to any harmonics of any generating the inverter. Furthermore, to make the signal filteringeasier, there must be a spectral separation between the injected frequency and the harmonics of thefundamental frequency. Any frequency meeting these two criteria can be selected. For a 60 Hz system,350 Hz can be selected as it lies between the fifth (300 Hz) and seventh (420 Hz) harmonics.

B. Technique Based on Voltage and Reactive Power Shift [3]:

In this method, the voltage variation of the system is measured over time and a mean is calculated(passive method) and, when this value crosses a threshold, the Adaptive Reactive Power Shift (ARPS)algorithm is used as an active method. This method varies the output reactive power of the DG systemto detect islanding and the results have shown this technique to be effective, while maintaining lessamount of disturbances in the inverter’s grid-connected system operation.

C. Positive Feedback Frequency Shift (PFFS) and Reactive Power Variation (RPV) Method [62]:

This method continuously monitors the system using PFFS and, when the frequency deviates by0.1 Hz, RPV is triggered to detect the islanding. Being capable of synchronizing converters with eachother, this method is thus capable of detecting islanding when multiple converters are paralleled indistributed generation.

2.2.4. Machine Learning Method

This method deals with the studies of characteristics of events from the data set (main data) anduses that knowledge to find the islanding condition in test data. The decision tree approach can makea complex decision-making process into a simpler decision with a solution easier to interpret [63].

A. C4.5 Decision Tree-Based Islanding Detection [64]:

Appl. Syst. Innov. 2019, 2, 25 7 of 19

Based on the C4.5 decision tree model, this method of detection first gathers sampled data obtainedfrom the simulation result. These data are then used to construct a C4.5 classification model. Otherdata are then used to verify the working of the C4.5 decision tree. Each cycle of data processing isthen used for islanding detection, and the result is used to improve the C4.5 decision tree itself—thusresulting in a circular loop of island detection and improvement of the detection scheme.

B. Aggressive Signal Modeling and Support Vector Machine Based Anti-Islanding Method [65]:

System parameters like voltage and current are extracted from the point of common coupling.This is then used as input to the machines and an additional advanced machine learning techniqueis used to determine different scenarios of islanding and non-islanding conditions. Thus, with fewparameters, predictions can be made about the islanding sate of the system. With the accuracy of98.74%, this method was able to detect the islanding condition within 50 ms of the event occurrence.

C. Support Vector Machine (SVM) Classifier Based Machine Learning Approach [66]:

SVM classifiers are fed multiple features extracted from the system parameters such as voltage.The classifiers then determine the islanding condition of the system. Linear, polynomial and GaussianRadial Basis Function kernels are used to train the machine and further islanding and non-islandingdata are used to fine-tune the system. The proposed method is tested with various data with allthe possible combinations of active and reactive power imbalances. The resulting method yieldsexceptional results when compared with some passive methods of islanding detection such as ROCOFand over/under frequency.

D. Reduced NDZ Based Islanding Detection [67]:

With the use of the CART (Classification and Regression Trees) algorithm, the study reduces thenon-detection zones in an islanded condition, thereby overcoming the large NDZ due to reducedpower mismatch in case of multiple feeders network. CART is a decision tree building algorithm thatuses if-then logical conditions to predict accurate conditions of cases. Without prior knowledge ofrelationships between multiple variables and theories governing them, this method can be effectivelyused for data mining and data’s implementation to predict various cases [68]. With the reduction of54% of total NDZ area, higher dependability and improved overall performance are achieved, thoughextensive field experiments are required before any practical device can be realized.

2.2.5. Remote Methods

The remote method of islanding detection is the communication-based method between the utilityand DG. The islanding is known through the status of the utility circuit breaker, and a trip signal issent to the DG unit [69]. With no NDZ regions, no effect on the power quality and fast response time,the remote method of islanding detection provides a highly dependable scheme. However, the use ofextra communication channels and high-cost software like Supervisory Control and Data Acquisition(SCADA) makes this method pricey to implement. The remote islanding detection can be divided intotwo schemes: transfer and trip scheme and power line signaling scheme [70].

Transfer and trip scheme continuously monitor the status of all the reclosers and circuit breakersresponsible for the islanded condition. Thus, a higher requirement of system interaction is essential.The SCADA system can be used for this scheme. Likewise, auto-ground is a low cost, scalable approachfor anti-islanding protection of distributed generation, using a utility owned and operated system,installed just down-line of the utility protection device (substation breaker or inline recloser). Inauto-ground, when the utility breaker opens, the auto-ground opens the substation side device calledSectionalizing Switch (SS) and closes the Auto-grounding Switch (AS) effectively applying a threephase to ground fault. In addition, based on online protection, all DGs connected on their anti-islandingprotection will be forced to disconnect the breaker. However, in the power line signaling scheme, thepower line communication scheme is used to broadcast a continuous signal from the transmission

Appl. Syst. Innov. 2019, 2, 25 8 of 19

system to the distribution feeder using the transmission lines. DGs are provided with signal receiverswhich monitor the continuous signal and, in the case of the islanded condition, the signal disruptionoccurs. Some of the remote-based islanding detection methods are:

A. A Remote Islanding Detection Photovoltaic-Based Distributed Generation Systems and Wind–SolarHybrid Power Plants [49,71]:

Monitoring multiple parameters like changes in the current through circuit breakers, frequencyand voltage at PCC on three sides of the PV system, this study compares them to a pre-definedthreshold. This method uses continuous communication between the control system and on-fielddevices like the circuit breakers. This method is independent from the local loads and inverters andthus is independent from the power mismatch. This results in zero NDZ region. Circuit Breakersare tripped in case of islanding detection. With no NDZ, the islanding detection time is found to be1.65 m-sec for the photovoltaic-based distributed generation systems.

B. Wireless Communication-Based Islanding Detection [72]:

With the introduction of an End Monitor Module (EMM) on the load side, generation side anddistribution side, current, voltage, power, status of generators, and other required parameters aremeasured. These measured values are then communicated with the Local Area Monitor module(LAN) using the ZigBee module. The three EMs are provided with the capability of connecting anddisconnecting loads and sources. This capability of EM is used to control various devices as required atthe time of islanding. The use of wireless communication ensures the modular capability of the schemeand thus does not depend on the restrictions that would have risen when using a more conventionalhard-wired based communication.

A brief comparison of different islanding methods are presented in Table 1 below.

Table 1. Comparison of islanding methods [3,49–51,73,74].

Technique NDZ DetectionTime

PowerQuality Cost Detection

ReliabilityEffect of

Multiple DGs

Passive Broad Short No effect Low Low None

Active Limited Long Degradation Average High Synchronizationissues

Hybrid Limited Long Degradation High High Synchronizationissues

MachineLearning Limited Variable No effect High High None

Remote None Very short No effect High Very high None

3. Method

There are varieties of the methods for the islanding detection of a DGS, each has its own advantagesand disadvantages. In this study, two different approaches: islanding detection by the auto-groundingmethod and the data mining-decision tree algorithm has been presented. The detailed description ofthe approaches are described in Sections 3.1 and 3.2 below.

3.1. Auto-Grounding Method

In the auto-grounding method, it is assumed that the DG’s line protection has been properlyconfigured in order to detect all faults. In case of inverter based DG, overcurrent protection alonemay not be sufficient and more advanced functions such as overcurrent with voltage restraint maybe required. The auto-ground system consists of three main components: Sectionalizing Switch (SS),Auto-grounding Switch (AS), and the controller, which can be implemented using a variety of re-closeror breaker controls. The SS is required only to mirror the state of the substation breaker, and it is

Appl. Syst. Innov. 2019, 2, 25 9 of 19

preferable to have the SS as a separate device for the simple reason that costs escalate when workwithin the substation is required. Any savings associated with integration of the SS function intothe feeder breaker would be outweighed by the cost of linking it with the AS [65]. In this case, itis not required to interrupt fault currents over any device that can provide automated sectionalizercapabilities. For applications where the auto-ground is paired with an in-line re-closer, the SS is notrequired as its functionality can simply be integrated into the re-closer itself. Figure 2 shows the blockof a system for the implementation of islanding detection via the auto-grounding method, and thefollowing are the three logics/stages of the process:

• Both of the circuit breakers (CB1 and CB2) are in closed conditions. That is the normal condition.Thus, the voltage and current waveforms found at the utilization level are purely sinusoidal.

• CB1 is initially closed and is opened with transition time ‘t’, but CB2 is closed. After switchingoff the CB1, the whole part of the distribution system is isolated from the remaining part of thepower system, which is called the islanding condition.

• CB1 is initially closed and is opened with transition time ‘t’. As another DG is there in the linewhere CB2 is connected, the opening of CB2 is the condition of anti-islanding.

Figure 2. Block diagram of the grid-connected DGS.

3.2. Data Mining-Decision Tree Algorithm

Decision trees are a type of supervised machine learning process where the data is continuouslysplit as per the prescribed parameters. The tree can be explained by two entities, namely decisionnodes and leaves. The leaves are the decisions or the final outcomes, and the decision nodes areplaced where the data is split. It can be divided into two types: Yes/No type and Continuous datatype. It is one of the best independent variable selection algorithms that make a model with logistic(or linear) regressions or with neural networks, which reduce the number of variables by selectingonly the relevant ones: use decision trees. The process is fast as compared to the calculation of simplecorrelations with the target variable, and it also takes into account the interactions between variables.Decision trees are non-parametric means and no specific data distribution is necessary, and easilyhandle continuous and categorical variables.

The main importance of decision trees is that they identify subgroups, in which each terminalor intermediate leaf can be seen as a subgroup/segment of your population. It delivers high-qualitymodels, which are able to squeeze pretty much all information out of the data, especially if you usethem in ensembles. In this study, this concept is used to detect the islanding state of a DGS.

The data from the 13-bus system is extracted into the tree model by using the Classification andRegression Tree Algorithm (CART). The CART algorithm uses the predictive modeling for designingthe decision tree. CART was first developed by Breiman et al. in the 1980s [75], and was first introducedin the Power System by Wehenkel [76,77]. The CART model of the DGS is obtained by recursivelypartitioning the data space and fitting a simple prediction model within each partition, due to whichpartitioning can be represented graphically as a decision tree [78,79]. CART uses the empirical impurityreduction known as Gini Index for splitting the node and selects the split that maximizes the decrease

Appl. Syst. Innov. 2019, 2, 25 10 of 19

in impurity and generates the sequence of sub-trees by growing a large tree and pruning back untilonly one root node is left [80]. The Gini index is given by:

Gini (t) = 1 − Σ [p(j|t)]2, (11)

where p(j|t) is the relative frequency of class j at node t.The decision tree model generated using the CART algorithm is tested in Python software using

tested and validated test data obtained from the auto-grounding method.

4. Result and Discussion

The match between the source and loads lead to the anti-islanding methods being unable to detectthe islanding conditions. The closer the power match between the loads and the sources, the lesserthe changes of the various parameters (like voltage change, frequency change, etc.) occur duringthe islanding event. This means that the anti-islanding methods which depend on monitoring theseparameters will tend not to detect the islanding event. To overcome this, monitoring more than oneparameter is required.

For the developed method, the power mismatch between the source and the loads are 38.41% to−24.39% as shown in Figure 3. That is, the case of any power mismatch between these limits will notbe detected by the proposed methods. This NDZ has to be determined by conducting the load flowanalysis of a sample of 13-bus system as shown in Figure 4.

Figure 3. Changes on load with the limit of NDZ.

Figure 4. Load flow analysis of the 13-bus system.

Appl. Syst. Innov. 2019, 2, 25 11 of 19

4.1. Result of Islanding Detection by the Auto-Grounding Method

Auto-ground, anti-islanding protection is based on applying a three-phase, short-circuit to theislanded distribution system just prior to reclosing or re-energization. It is seen as a compromise interms of cost and performance between the transfer-trip and local passive measurements. The autoground system is installed just down-line of the utility protection device (substation breaker or inlinerecloser). In this configuration, following opening of the utility breaker, the auto-ground opens thesubstation side device, denoted SS and closes the AS effectively applying a three phase to ground fault.All DGs that have not already disconnected based on their anti-islanding protection will be forced todisconnect based on online protection.

For the detection of islanding events, four passive methods are used; over/under frequency,over/under voltage, rate of change of power and rate of change of frequency. The use of passivemethods along with the principle of auto-grounding has been used for the detection of the islandingconditions. Based on the limits defined, the state of the system was monitored for 42 distinct cases.

Among the islanding (1) and non-islanding (0) states with the 42 sequences of events, half were inislanding state, while the other half of the events were not in islanding state as shown in Table 2. Thedeveloped auto-grounding method was implemented to detect these conditions. Furthermore, 20 ofthe total 21 islanded conditions (95.2%) and 19 of the total 21 non-islanded conditions (90.4%) weresuccessfully detected, and listed in Table 3.

Table 2. Forty-two individual events for the islanding detection using the auto-grounding method.

No.Frequency Voltage df/dt dP/dt Islanding States

(Hz) (pu) (Hz/sec) (MW/sec)

1 49.977 0.911 0.748 44.039 12 50.279 0.989 7.309 508.538 13 49.986 0.935 0 1.981 04 49.979 1.05 0.175 16.706 15 49.986 0.969 0 1.15 06 49.986 0.953 0 1.127 07 50.043 0.95 2.364 43.767 18 50.216 1.2 1.967 1974.419 19 50.108 0.987 0.738 1.98 0

10 50.045 1.016 2.294 18.755 111 50.108 0.972 0.738 1.127 012 49.986 0.954 0.738 0.551 013 49.803 0.984 4.868 49.238 114 49.77 1.174 5.684 35.284 115 49.997 0.862 0.007 2.391 016 49.928 1.106 1.72 12.699 117 49.997 0.914 0.007 2.46 018 50.022 0.94 0.619 10.169 019 49.977 0.902 0.233 35.521 120 49.904 1.006 2.052 11.327 121 49.986 0.94 0 0.328 022 49.982 1.038 0.106 6.963 123 49.986 0.977 0 0.282 024 49.986 0.96 0 0.639 025 49.952 0.916 0.857 48.117 126 49.953 1.025 0.837 21.89 127 49.986 0.931 0 2.799 028 49.957 1.059 0.729 15.551 129 49.986 0.964 0 1.653 030 49.986 0.948 0 1.111 031 49.977 0.908 0.233 42.504 132 50.002 1.017 0.39 22.384 133 49.986 0.939 0 2.203 034 50.167 1.002 4.51 16.962 1

Appl. Syst. Innov. 2019, 2, 25 12 of 19

Table 2. Cont.

No.Frequency Voltage df/dt dP/dt Islanding States

(Hz) (pu) (Hz/sec) (MW/sec)

35 49.986 0.972 0 0.29 036 49.986 0.957 0 0.596 037 49.952 0.912 0.857 44.299 138 49.953 1.02 0.837 17.891 139 49.986 0.932 0 2.07 040 49.953 1.023 0.837 8.074 141 49.986 0.967 0 1.598 042 49.986 0.951 0 1.165 0

Table 3. Accuracy of the model with the auto grounding method.

Class Cases Correctly Classified

Islanded Condition 21 20 (95.2%)

Non-Islanded Condition 21 19 (90.4%)

Total 42 39 (92.8%)

For instance, the event number 3 (with frequency 49.986 Hz, voltage of 0.935 pu, rate of changeof frequency of 0 Hz/sec and rate of change of power of 1.981 MW/sec) was detected to be in anon-islanding condition, while event number 7 (with frequency 50.043 Hz, voltage of 0.95 pu, rate ofchange of frequency of 2.364 Hz/sec and rate of change of power of 43.767 MW/sec) was detected to bein an islanding condition.

Figure 5 shows the current and voltage waveforms in the event of islanding detection by the autogrounding method. In the event of islanding in 2.5 s, it takes the auto-grounding method 1.705 s forislanding detection. With the auto-grounding scheme grounding the lines, the current increases above500 A at around 4.194 s and thus the line protection scheme disconnects the circuit breaker, therebycreating an anti-islanding scheme.

Figure 5. Islanding detection by the auto-grounding method of voltage and current with respect to time.

Appl. Syst. Innov. 2019, 2, 25 13 of 19

4.2. Result of Islanding Detection by the Data Mining-Decision Tree Algorithm

Similarly, on the other hand, the decision tree is generated using the CART algorithm in thescikit-learn module of the Python programming language [81]. From the total data set, 70% of thedata are used as training data and the remaining 30% are used as testing data. The testing data set aregenerated randomly for each case of islanding state and non-islanding state. The topmost part of thetree is also known as the decision node, which consists of the condition, can be either true or false,following the path of true or false, and it finally ends at a leaf node, while, for a leaf node, the ‘value’consists of two pieces of numeric data. The value on the left-hand side represents the value of thenon-islanding condition and on the right-hand side represents the value of the islanding condition.

The total training samples are 88, out of which 44 samples are for the islanding condition and another44 are for the non-islanding condition. The node is split until the minimum gini for the samples areobtained and all the samples are correctly classified. For instance, the data sample where the value ofactive power is calculated as 1, reactive power as 0.64, rate of change of power (Dp) as −7.89, voltage as0.9857 pu and power angle as 25.25 degrees is taken to explain the workings of the developed algorithm.In accordance with the developed tree, first the algorithm checks the power angle, which is greater than21.75; then, it proceeds to check power (P) which in this case is less than 34.54. The condition is true andit proceeds towards the left side to check if the power angle is greater than 22.316. This is false, whichdirects the algorithm to check Dp/P, which in this case is less than 140.111. Furthermore, it checks if thepower angle is less than or equal to 22.99. In the same manner, the process continues and finally ends atthe leaf node at the bottom left corner, thus determining the condition to be the ‘islanding condition’. Thispath is denoted by the darker shaded decision boxes in Figure 6. The gini index in the box is used todetermine the uncertainty in the result. Here, the gini index is zero, meaning that the result obtained is100% certain. The actual tree diagram of the conducted analysis is shown in Figure 5.

Figure 6. Decision tree for generated data.

Appl. Syst. Innov. 2019, 2, 25 14 of 19

The entire data set was classified into islanding mode and non-islanding mode. The modelsuccessfully classifies all the training set data to the correct classification as shown in Table 4. Thus, theaccuracy of the model with the training set data is 100%. Likewise, for the testing set data, the modelclassified 35 out of 37 events to the correct classification. Thus, the accuracy of the model for testing setis 94.5%. In one of the cases, the occurrence of islanding in 2.5 s led to the detection of islanding within0.162309 s. Simulation of both the methods were conducted using the computer having Intel® Core™i7-7500U @ 2.70 GHz (4 CPU), ~2.9 GHz processor specifications and 8192 MB of RAM.

Table 4. Accuracy of the model.

Classes

Training Set Testing Set

Total Cases CorrectlyClassified Total Cases Correctly

Classified

Islanding 44 44(100%) 18 17 (94.4%)

Non-Islanding 44 44(100%) 19 18(94.7%)

Total 88 88(100%) 37 35 (94.5%)

4.3. Comparison between Auto-Grounding Method and the Data Mining-Decision Tree Algorithm Method

Based on the two simulations, the two islanding detection methods are compared in accordancewith their respective accuracies in Figure 7. While the auto-grounding method is more accuratein detecting islanding conditions with 95.2% accuracy compared to 94.4% accuracy of the datamining-decision tree algorithm method, the data mining-decision tree algorithm method is vastlymore accurate in predicting the non-islanding conditions with the accuracy of 94.7% compared tothe accuracy of auto-grounding method being only 90.4%. Taking both islanding and non-islandingconditions, the auto-grounding method is only 92.8% accurate while the data mining-decision treealgorithm method is 94.5% accurate. Auto-grounding takes a long time to detect the islanding condition(1.705 s), while the data mining-decision tree method is quicker to detect the islanding condition(0.162309 s).

Figure 7. Accuracy comparison between auto-grounding and data-mining-decision-tree algorithm methods.

5. Conclusions

The accurate and precise detection of the islanding condition plays a very significant part in DGS.This paper highlights a relatively new approach of using the auto-grounding method and data miningdecision tree technique to correctly detect the islanding condition of DG. The low-cost auto-groundingmethod was designed and simulated, and it was observed that the accurate anti-islanding can be

Appl. Syst. Innov. 2019, 2, 25 15 of 19

done within 1.706 s. Moreover, the use of the data mining decision tree algorithm based on a machinelearning approach successfully identified the anti-islanding condition with high accuracy and reliabilitywithin 0.162309 s. The comparison between auto-islanding and the machine learning technique showsthat using the data mining decision tree technique presents the fastest operation of the anti-islandingcondition. Furthermore, the data mining decision tree approach was also found to be more reliableand accurate than the auto-grounding method.

An auto-grounding method, which is a type of remote method for islanding detection, and a datamining-decision tree method, which is a type of machine learning methods for islanding detection,were developed, simulated and evaluated. Auto-grounding methods used four passive methods forislanding detection: over/under voltage, over/under frequency, rate of change of power and rate ofchange of frequency. Forty-two individual events were used to simulate the islanding detection usingthe auto-grounding method, out of which 21 events were in islanding state and 21 events were innon-islanding conditions. Similarly, 37 events were used in the data mining-decision tree method,where 18 events were in islanding states and 19 events were in non-islanding states.

The proposed islanding detection methods were simulated in the computer and the followingconclusions were derived:

• The data mining-decision tree method has a higher overall accuracy (94.5%) when compared tothe auto-grounding method (92.8%). Thus, comparing between the two, the data mining-decisiontree method is the more reliable method of islanding detection.

• In comparison between the two discussed methods, the data mining-decision tree method presentsthe fastest and more accurate operation of anti-islanding condition.

Auto-grounding being the older and the data mining-decision tree method being the comparativelynewer method of islanding detection, these simulations highlight increased accuracy and speed ofthe newer islanding detection method when compared to the older one. The power grid system’scomplexity is growing on a yearly basis. The introduction of DGs and increased power capacity of thepower grid has led to the requirement of a faster protection system. Two seconds is not a fast-enoughislanding detection speed in this scenario. As such, micro-grids and smart-grids are relying more onnewer and faster islanding detection methods.

In successful simulations of the data mining-decision tree method, the research has shown, withpositive results, a newer form of islanding detection method with short detection speed and highaccuracy. This method promises to satisfy the meticulous islanding detection criteria of the futuregrids comprised of micro-grids and smart grids. To further the findings of this paper, the followingexplorations can be conducted:

• The detections of islanding conditions were done in 1.705 s and 0.162 s by auto-grounding anddata mining-decision tree methods. The speed of islanding condition detections can be improvedupon by simulating these two methods in a computer with better specifications.

• A larger amount of training data can be fed to a data mining-decision tree method to furtherincrease the accuracy of this method.

• Expanding on the presented method by making it compatible with the grid code. This mayinclude and not be limited to introducing fault ride through requirements.

Author Contributions: All authors contributed in the paper writing and correcting task. A.S. (Ashish Shrestha),R.K. and M.D. conceived the idea and focused on algorithm development. B.M. and R.T. assisted A.S. (AshishShrestha) in the experimentation work and whole result analysis. A.S. (Ajay Singh) performed the resultvisualization task. D.B. and B.A. contributed on the project management activities, and result validation. A.P. andR.K.M. supervised the whole project, to provide an efficient way to execute this study and to provide critical review.

Funding: This study was supported by Energize Nepal Project Office at Kathmandu University, Nepal under thefunding of the Royal Norwegian Embassy in Nepal (Project ID: ENEP-CEPE-18-01).

Acknowledgments: This study is supported by the EnergizeNepal Project Office at Kathmandu University,Nepal under the funding of the Royal Norwegian Embassy in Nepal, and conducted at the Center for Electric

Appl. Syst. Innov. 2019, 2, 25 16 of 19

Power Engineering, Kathmandu University and Khwopa College of Engineering, Tribhuvan University (ProjectID: ENEP-CEPE-18-01). The authors would like to acknowledge the Department of Electrical and ElectronicsEngineering, Kathmandu University, the Center for Electric Power Engineering, Kathmandu University, andthe Department of Electrical Engineering, Khwopa College of Engineering, Tribhuvan University for their kindsupport during the whole period of this study.

Conflicts of Interest: The authors declare no conflict of interest.

References

1. Khamis, A.; Shareef, H.; Bizkevelci, E.; Khatib, T. A review of islanding detection techniques for renewabledistributed generation systems. Renew. Sustain. Energy Rev. 2013, 28, 483–493. [CrossRef]

2. Laghari, J.; Mokhlis, H.; Karimi, M.; Bakar, A.; Mohamad, H. An islanding detection strategy for distributionnetwork connected with hybrid dg resources. Renew. Sustain. Energy Rev. 2015, 45, 662–676. [CrossRef]

3. Mahat, P.; Chen, Z.; Bak-Jensen, B. Review of islanding detection methods for distributed generation. InProceedings of the Third International Conference on Deregulation and Restructuring and Power Technologies(DRPT 2008), Nanjing, China, 6–9 April 2008; pp. 2743–2748.

4. Greacen, C.; Engel, R.; Quetchenbach, T. A Guidebook on Grid Interconnection and Islanded Operation of Mini-GridPower Systems up to 200 kw; Lawrence Berkeley National Lab.(LBNL): Berkeley, CA, USA, 2013.

5. Hirsch, A.; Parag, Y.; Guerrero, J. Microgrids: A review of technologies, key drivers, and outstanding issues.Renew. Sustain. Energy Rev. 2018, 90, 402–411. [CrossRef]

6. Xue, F.; Xu, Y.; Zhu, H.; Lu, S.; Huang, T.; Zhang, J. Structural evaluation for distribution networks withdistributed generation based on complex network. Complexity 2017, 2017, 7539089. [CrossRef]

7. Yingram, M.; Premrudeepreechacharn, S. Over/undervoltage and undervoltage shift of hybrid islandingdetection method of distributed generation. Sci. World J. 2015, 2015, 654942. [CrossRef] [PubMed]

8. Dong, Y.; Bing, C.; Qun, L.; Jianhua, L. Control strategy of low voltage microgrid based on equivalent referencecurrent control. In Proceedings of the 2014 China International Conference on Electricity Distribution (CICED),Shenzhen, China, 23–26 September 2014; pp. 70–76.

9. Sorrell, S. Reducing energy demand: A review of issues, challenges and approaches. Renew. Sustain. EnergyRev. 2015, 47, 74–82. [CrossRef]

10. Eltawil, M.A.; Zhao, Z. Grid-connected photovoltaic power systems: Technical and potential problems—areview. Renew. Sustain. Energy Rev. 2010, 14, 112–129. [CrossRef]

11. Bukhari, S.B.A.; Zaman, M.S.U.; Haider, R.; Oh, Y.-S.; Kim, C.-H. A protection scheme for microgrid withmultiple distributed generations using superimposed reactive energy. Int. J. Electr. Power Energy Syst. 2017,92, 156–166. [CrossRef]

12. Chen, X.; Li, Y.; Crossley, P. A novel hybrid islanding detection method for grid-connected microgrids withmultiple inverter-based distributed generators based on adaptive reactive power disturbance and passivecriteria. IEEE Trans. Power Electron. 2018, 34, 9342–9356. [CrossRef]

13. Pedrino, E.C.; Yamada, T.; Lunardi, T.R.; de Melo Vieira, J.C., Jr. Islanding detection of distributed generationby using multi-gene genetic programming based classifier. Appl. Soft Comput. 2019, 74, 206–215. [CrossRef]

14. Photovoltaics, D.; Storage, E. IEEE standard for interconnection and interoperability of distributed energyresources with associated electric power systems interfaces. IEEE Std. 2018, 1547–2018. [CrossRef]

15. Kim, M.-S.; Haider, R.; Cho, G.-J.; Kim, C.-H.; Won, C.-Y.; Chai, J.-S. Comprehensive review of islandingdetection methods for distributed generation systems. Energies 2019, 12, 837. [CrossRef]

16. Bignucolo, F.; Cerretti, A.; Coppo, M.; Savio, A.; Turri, R. Impact of distributed generation grid coderequirements on islanding detection in LV networks. Energies 2017, 10, 156. [CrossRef]

17. Khamis, A.; Xu, Y.; Dong, Z.Y.; Zhang, R. Faster detection of microgrid islanding events using an adaptiveensemble classifier. IEEE Trans. Smart Grid 2016, 9, 1889–1899. [CrossRef]

18. Bukhari, S.B.A.; Haider, R.; Zaman, M.S.U.; Oh, Y.-S.; Cho, G.-J.; Kim, C.-H. An interval type-2 fuzzy logicbased strategy for microgrid protection. Int. J. Electr. Power Energy Syst. 2018, 98, 209–218. [CrossRef]

19. Pouryekta, A.; Ramachandaramurthy, V.; Padmanaban, S.; Blaabjerg, F.; Guerrero, J. Boundary detectionand enhancement strategy for power system bus bar stabilization—investigation under fault conditions forislanding operation. Energies 2018, 11, 889. [CrossRef]

20. Mahat, P.; Chen, Z.; Bak-Jensen, B. A hybrid islanding detection technique using average rate of voltagechange and real power shift. IEEE Trans. Power Deliv. 2009, 24, 764–771. [CrossRef]

Appl. Syst. Innov. 2019, 2, 25 17 of 19

21. Zeineldin, H.H.; Kirtley, J.L. A simple technique for islanding detection with negligible nondetection zone.IEEE Trans. Power Deliv. 2009, 24, 779–786. [CrossRef]

22. Do, H.T.; Zhang, X.; Nguyen, N.V.; Li, S.S.; Chu, T.T.-T. Passive-islanding detection method using the waveletpacket transform in grid-connected photovoltaic systems. IEEE Trans. Power Electron. 2015, 31, 6955–6967.[CrossRef]

23. Reigosa, D.; Briz, F.; Charro, C.B.; García, P.; Guerrero, J.M. Active islanding detection using high-frequencysignal injection. IEEE Trans. Ind. Appl. 2012, 48, 1588–1597. [CrossRef]

24. Far, H.G.; Rodolakis, A.J.; Joos, G. Synchronous distributed generation islanding protection using intelligentrelays. IEEE Trans. Smart Grid 2012, 3, 1695–1703. [CrossRef]

25. Tran, T.; Nguyen, D.; Fujita, G. Islanding detection method based on injecting perturbation signal and rate ofchange of output power in dc grid-connected photovoltaic system. Energies 2018, 11, 1313. [CrossRef]

26. Bei, T.-Z. Accurate active islanding detection method for grid-tied inverters in distributed generation.IET Renew. Power Gener. 2017, 11, 1633–1639. [CrossRef]

27. Khodaparastan, M.; Vahedi, H.; Khazaeli, F.; Oraee, H. A novel hybrid islanding detection method forinverter-based DGs using SFS and ROCOF. IEEE Trans. Power Deliv. 2015, 32, 2162–2170. [CrossRef]

28. Akhlaghi, S.; Meshginkelk, H.; Akhlaghi, A.; Ghadimi, A. A novel hybrid islanding detection method forinverter-based distributed generation based on frequency drift. Aust. J. Electr. Electron. Eng. 2014, 11,161–174. [CrossRef]

29. Pouryekta, A.; Ramachandaramurthy, V.K.; Mithulananthan, N.; Arulampalam, A. Islanding detection andenhancement of microgrid performance. IEEE Syst. J. 2017, 12, 3131–3141. [CrossRef]

30. International Energy Agency; Photovoltaic Power Systems. Evaluation of Islanding Detection Methods forPhotovoltaic Utilityinteractive Power Systems; Report IEA PVPS T5-09; International Energy Agency: Paris,France, 2002.

31. IEC 62116:2014 Utility-Interconnected Photovoltaic Inverters-Test Procedure of Islanding Prevention Measures;International Electrotechnical Commission: Geneva, Switzerland, 2014.

32. Ghanbari, T.; Farjah, E.; Naseri, F. Power quality improvement of radial feeders using an efficient method.Electr. Power Syst. Res. 2018, 163, 140–153. [CrossRef]

33. Jenkins, N. Distributed Generation; The Institution of Engineering and Technology: Stevenage, UK, 2010.34. Pepermans, G.; Driesen, J.; Haeseldonckx, D.; Belmans, R.; D’haeseleer, W. Distributed generation: Definition,

benefits and issues. Energy Policy 2005, 33, 787–798. [CrossRef]35. Shrestha, A. Planning, Design In addition, Optimization Of Distribution System For Affected Area Of Upper

Karnali Hydropower Project. Master’s Thesis, Kathmandu University, Kathmandu, Nepal, 2017.36. Mohajeryami, S.; Jennings, R.; Alkhbbaz, G. Integration of energy storage and distributed generation (dg) in

distribution systems: Economic analysis and development perspective. Munich Personal RePEc Archive 2015.37. Hasan, N.S.; Hassan, M.Y.; Majid, M.S.; Rahman, H.A. Review of storage schemes for wind energy systems.

Renew. Sustain. Energy Rev. 2013, 21, 237–247. [CrossRef]38. Lopes, J.P.; Hatziargyriou, N.; Mutale, J.; Djapic, P.; Jenkins, N. Integrating distributed generation into

electric power systems: A review of drivers, challenges and opportunities. Electr. Power Syst. Res. 2007, 77,1189–1203. [CrossRef]

39. Qiu, J.; Xu, Z.; Zheng, Y.; Wang, D.; Dong, Z.Y. Distributed generation and energy storage system planningfor a distribution system operator. IET Renew. Power Gener. 2018, 12, 1345–1353. [CrossRef]

40. Zhang, X.; Chen, H.; Xu, Y.; Li, W.; He, F.; Guo, H.; Huang, Y. Distributed generation with energy storagesystems: A case study. Appl. Energy 2017, 204, 1251–1263. [CrossRef]

41. Shrestha, P.; Shrestha, A.; Adhikary, B. Comparative Analysis of Grid Integration on Distributed EnergySystem. In Proceedings of the 5th International Conference on Developments in Renewable EnergyTechnology, Kathmandu, Nepal, 29–31 March 2018.

42. Bhandari, B.; Ahn, S.-H.; Ahn, T.-B. Optimization of hybrid renewable energy power system for remoteinstallations: Case studies for mountain and island. Int. J. Precis. Eng. Manuf. 2016, 17, 815–822. [CrossRef]

43. Sah, S.; Shrestha, A.; Papadakis, A. Cost-Effective and Reliable Energy System for Kathmandu UniversityComplex. In Proceedings of the 11th International Conference on Deregulated Engineering Market Issues,Nicosia, Cyprus, 20–21 September 2018.

Appl. Syst. Innov. 2019, 2, 25 18 of 19

44. Shrestha, A.; Singh, A.; Khanal, K.; Maskey, R. Potentiality of off-grid hybrid systems for sustainable powersupply at Kathmandu University campus. In Proceedings of the 2016 IEEE 6th International Conference onPower Systems (ICPS), New Delhi, India, 4–6 March 2016; pp. 1–6.

45. Bhandari, V.; Dahal, R.; Adhikary, B.; Maskey, R.K.; Shrestha, N. MiniGrid: Design and development ofmini-grid for efficient use of distributed hydropower systems-experience from the first year of the project.Rentech Symp. Comendium 2012, 1, 4.

46. Bhandari, B.; Lee, K.-T.; Lee, G.-Y.; Cho, Y.-M.; Ahn, S.-H. Optimization of hybrid renewable energy powersystems: A review. Int. J. Precis. Eng. Manuf. Green Technol. 2015, 2, 99–112. [CrossRef]

47. Shrestha, A.; Rana, L.B.; Singh, A.; Phuyal, S.; Ghimire, A.; Giri, R.; Kattel, R.; Karki, K.; Jha, S.K. Assessmentof electricity excess in an isolated hybrid energy system: A case study of a dangiwada village in rural nepal.Energy Procedia 2019, 160, 76–83. [CrossRef]

48. Bhattacharyya, S.C.; Palit, D. Mini-Grids for Rural Electrification of Developing Countries: Analysis and CaseStudies from South Asia; Springer: Berlin, Germany, 2014.

49. Kunte, R.S.; Gao, W. Comparison and review of islanding detection techniques for distributed energyresources. In Proceedings of the 40th North American Power Symposium, Calgary, AB, Canada, 28–30September 2008; pp. 1–8.

50. Merino Fernández, J.; Mendoza-Araya, P.; Venkataramanan, G.; Baysal, M. Islanding detection in microgridsusing harmonic signature. IEEE Trans. Power Deliv. 2015, 30, 2102–2109. [CrossRef]

51. Najy, W.K.; Zeineldin, H.; Alaboudy, A.H.K.; Woon, W.L. A bayesian passive islanding detection method forinverter-based distributed generation using esprit. IEEE Trans. Power Deliv. 2011, 26, 2687–2696. [CrossRef]

52. Wang, F.; Xu, Z. Evaluation of inverter resident passive islanding detection method for grid connected PVsystem under constant power mode. In Proceedings of the Chinese Control and Decision Conference, Guilin,China, 17–19 June 2009; pp. 3086–3089.

53. Velasco, D.; Trujillo, C.; Garcerá, G.; Figueres, E. Review of anti-islanding techniques in distributed generators.Renew. Sustain. Energy Rev. 2010, 14, 1608–1614. [CrossRef]

54. Chang, W.-Y.; Yang, H.-T. An integrated passive islanding detection method for distributed generators.WSEAS Trans. Syst. 2009, 8, 1207–1216.

55. Merlin, V.L.; Santos, R.C.; Grilo, A.P.; Vieira, J.C.M.; Coury, D.V.; Oleskovicz, M. A new artificial networkbased method for islanding detection of distributed generators. Int. J. Electr. Power Energy Syst. 2016, 75,139–151. [CrossRef]

56. Hatata, A.; Abd-Raboh, E.-H.; Bishoy, E.S. A review of anti-islanding protection methods for renewabledistributed generation systems. J. Electr. Eng. 2016, 16.

57. Skocil, T.; Gomis-Bellmunt, O.; Montesinos-Miracle, D.; Galceran-Arellano, S.; Rull-Duran, J. Passive andactive methods of islanding for PV systems. In Proceedings of the 13th European Conference on PowerElectronics and Applications, Barcelona, Spain, 8–10 September 2009; pp. 1–10.

58. Abokhalil, A.G.; Awan, A.B.; Al-Qawasmi, A.R. Comparative study of passive and active islanding detectionmethods for PV grid-connected systems. Sustainability 2018, 10, 1798. [CrossRef]

59. Abd-Elkader, A.G.; Saleh, S.M.; Eiteba, M.M. A passive islanding detection strategy for multi-distributedgenerations. Int. J. Electr. Power Energy Syst. 2018, 99, 146–155. [CrossRef]

60. Bower, W.; Ropp, M. Evaluation of islanding detection methods for utility-interactive inverters in photovoltaicsystems. Sandia Rep. SAND 2002, 3591, 2002.

61. Khamis, A.; Shareef, H.; Mohamed, A.; Bizkevelci, E. Islanding detection in a distributed generationintegrated power system using phase space technique and probabilistic neural network. Neurocomputing2015, 148, 587–599. [CrossRef]

62. Mohiti, M.; Mahmoodzadeh, Z.; Vakilian, M. A hybrid micro grid islanding detection method. In Proceedingsof the 13th International Conference on Environment and Electrical Engineering (EEEIC), Wroclaw, Poland,1–3 November 2013; pp. 342–347.

63. Yin, J.; Chang, L.; Diduch, C. A new hybrid anti-islanding algorithm in grid connected three-phase invertersystem. In Proceedings of the 37th IEEE in Power Electronics Specialists Conference, Jeju, Korea, 18–22 June2006; pp. 1–7.

64. Mostafaei, S.; Beheshti, S.; Wu, B.; Zhang, D.; Wang, H.; Xing, H. A novel hybrid active anti-islandingmethod for multi-converter fed distributed generation systems. In Proceedings of the 2013 IEEE InternationalConference on in Industrial Technology (ICIT), Cape Town, South Africa, 25–28 February 2013; pp. 654–660.

Appl. Syst. Innov. 2019, 2, 25 19 of 19

65. Matic Cuka, B. Machine Learning Approach to Islanding Detection for Inverter-based Distributed 596Generation. Ph.D. Thesis, Texas A & M University, College Station, TX, USA, 2014.

66. Lin, X.; Dong, X.; Lu, Y.; Wei, Z.; Liu, Y. Application of intelligent algorithm in island detection of distributedgeneration. Aust. J. Electr. Electron. Eng. 2010, 7, 203–210. [CrossRef]

67. Kong, X.; Xu, X.; Yan, Z.; Chen, S.; Yang, H.; Han, D. Deep learning hybrid method for islanding detection indistributed generation. Appl. Energy 2018, 210, 776–785. [CrossRef]

68. Alam, M.R.; Muttaqi, K.M.; Bouzerdoum, A. Evaluating the effectiveness of a machine learning approachbased on response time and reliability for islanding detection of distributed generation. IET Renew. PowerGener. 2017, 11, 1392–1400. [CrossRef]

69. Cui, Q.; El-Arroudi, K.; Joós, G. Islanding detection of hybrid distributed generation under reducednon-detection zone. IEEE Trans. Smart Grid 2017, 9, 5027–5037. [CrossRef]

70. STATISTICA Features Overview. Available online: http://www.statsoft.com/Products/STATISTICA-Features(accessed on 23 July 2019).

71. Pujhari, T. Islanding Detection in Distributed Generation. Master’s Thesis, National Institute of Technology,Rourkela, India, 2009.

72. Bayrak, G. A remote islanding detection and control strategy for photovoltaic-based distributed generationsystems. Energy Convers. Manag. 2015, 96, 228–241. [CrossRef]

73. Bayrak, G.; Kabalci, E. Implementation of a new remote islanding detection method for wind–solar hybridpower plants. Renew. Sustain. Energy Rev. 2016, 58, 1–15. [CrossRef]

74. Lee, K.-J.; Kim, K.-M.; Moon, C.; Ahn, H.-S.; Jeong, G.-M. A new type of remote power monitoring systembasedon a wireless sensor network used in an anti-islanding method applied to a smart-grid. In Proceedingsof the International Conference on Multimedia, Computer Graphics, and Broadcasting, Jeju Island, Korea,8–10 December 2011; pp. 358–367.

75. Sun, R.; Wu, Z.; Centeno, V.A. Power system islanding detection & identification using topology approachand decision tree. In Proceedings of the 2011 IEEE in Power and Energy Society General Meeting, Detroit,MI, USA, 24–28 July 2011; pp. 1–6.

76. Abbey, C.; Brissette, Y.; Venne, P. An autoground system for anti-islanding protection of distributed generation.IEEE Trans. Power Syst. 2013, 29, 873–880. [CrossRef]

77. Breiman, L.; Friedman, J.; Olshen, R.A.; Stone, C.J. Classification and Decision Trees; Wadsworth InternationalGroup: Belmont, CA, USA, 1984.

78. Wehenkel, L.; Van Cutsem, T.; Ribbens-Pavella, M. An artificial intelligence framework for online transientstability assessment of power systems. IEEE Trans. Power Syst. 1989, 4, 789–800. [CrossRef]

79. Wehenkel, L.; Pavella, M. Decision trees and transient stability of electric power systems. Automatica 1991,27, 115–134. [CrossRef]

80. Loh, W.Y. Classification and regression trees. Wiley Interdiscip. Rev. Data Min. Knowl. Discov. 2011, 1, 14–23.[CrossRef]

81. Breiman, L. Classification and Regression Trees; Routledge: Abingdon, UK, 2017.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open accessarticle distributed under the terms and conditions of the Creative Commons Attribution(CC BY) license (http://creativecommons.org/licenses/by/4.0/).