Contemporary Engineering Sciences, Vol. 11, 2018, no. 22, 1069 - 1084

HIKARI Ltd, www.m-hikari.com

https://doi.org/10.12988/ces.2018.8237

Study and Analysis of Anti-Islanding Protection for

Grid-Connected Photovoltaic Central of Ghardaïa

Mohamed Redha Rezoug

Departement of Electrical Engineering

Université Kasdi Merbah Ouargla 30000, Algeria

Rachid Chenni

MoDERNa Laboratory Mentouri

University of Constantine1, Constantine 25000, Algeria

Djamel Taibi

Departement of Electrical Engineering

Université Kasdi Merbah Ouargla 30000, Algeria

Copyright © 2018 Mohamed Redha Rezoug, Rachid Chenni and Djamel Taibi. This article is

distributed under the Creative Commons Attribution License, which permits unrestricted use,

distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

This paper aims at providing a simulation study of the nature of the Islanding

phenomenon occurs in the Grid-Tied PV (photovoltaic) System and the different

methods to detect it which permits us to protect the system. We are going to resort

to a number of devices to put into practice these methods including over-current

and under-current (OI/UI) relays, over-voltage and under-voltage (OV/UV) relays

and over-frequency and under-frequency (OF/UF) relays. The protection is

examined in two different circumstances, when the PV system is completely

disconnected from the electric power grid and also during the occurrence of some

various grid faults. These faults can lead to the decoupling of the inverter and

cause of unintentional islanding phenomenon that has dangerous consequences

particularly on equipment and on grid maintenance personnel.

Keywords: PV system, Defaults, Islanding, protection, Grid-Tied PV

1070 Mohamed Redha Rezoug et al.

1 Introduction

The huge consumption of the fossil energy resources in addition to its high cost of

extraction lead for a need to adapt to new modes of energy production and

consumption. Through the implementation of various PV (Photovoltaic) power

plants, we offer a sustainable source of renewable energy for all energy-based

sectors [1].

The integration of new technologies of information and communication to the

actual power grids turns them into Smart Grids capable of responding to different

changes and requirements for the long term. Those grid requirements help in

detecting any harmful defaults protect the equipment and ensure safety of the grid

maintenance personnel [1], [7]. The islanding phenomenon occurs in Grid-Tied

PV systems can be intentional or non-intentional (accidental) and forms a

destructive effect on the whole PV system. In order to face such a phenomenon,

reliable and effective AI (anti-islanding) protection methods should be developed

[5], [12]. This paper provides a simulation study under a Matlab/Simulink

environment of a passive method for islanding detection. This anti-islanding

technique is based on the monitoring of voltage, current and frequency. For our

method to be functional, we compare directly the islanding detection time(s) in

different scenarios of anti-islanding relays such as over-current and under-current

(OI/UI), over-voltage and under-voltage (OV/UV) and over-frequency and under-

frequency (OF/UF). [1] All these scenarios evaluate the performance of AI relays

and then determine which one of the anti-islanding methods is the most effective

and appropriate so as to work on Grid-tied PV system [11].

The work was carried out at the Ghardaïa photovoltaic central in south Algeria at

a latitude of 32°24N and a longitude of 3°48E with an altitude of 566m. The land

is trimmed with a 10-hectare extent.

This area is characterized by a solar irradiation which reaches in summer 900 to

1000W/m² and a Saharan climate whose conditions are very severe given the high

temperatures and sand storms to which the southern regions are subjected.

The central has a rated power of approximately 1100 kWp (kW peak), distributed

as follows (see Figure 1):

Under field 105 KWC monocrystalline silicon fixed structure.

Under field 98.7 KWC in polycrystalline silicon fixed structure.

Under field 105 KWC in monocrystalline structured silicon.

Under field 98.7 KWC in polycrystalline silicon structure motorized.

Under field 100.8 KWC in thin layer (Telluride of cadmium Cd-Te) fixed

structure.

Subfield 100.11 KWC amorphous silicon fixed structure.

Under field 255 KWC in monocrystalline silicon fixed structure.

Under field 258.5 KWC in polycrystalline silicon fixed structure.

Study and analysis of anti-islanding protection 1071

Figure 1: The different types of PV technologies in the PV central in Ghardaïa

2 The Problematic of Islanding

In an electric power grid and in the presence of decentralized energy production

(DEP), particularly of photovoltaic installations, it appears a phenomenon called

“Islanding” [4], [3]. It happens when a sub-grid having one or more DEPs is

disconnected from the main grid, these DEPs continue supplying local loads.

Islanding can be intentional or accidental. Indeed, during a maintenance operation

on the power grid, the disconnection of the grid may lead to the islanding of the

generator. Since the loss of the grid is voluntary, the islanding is known and can

be put off-voltage by the operating personnel. Unintentional islanding linked to an

accidental disconnection of the grid is of great interest. This situation highlights

the dangers of maintaining a voltage in the islanded grid and can generate risks on

[7], [11]:

Electrical equipment during high drifts of voltage or frequency.

The generators when the protections are reset (false coupling)

People near equipment or during maintenance operations.

It is therefore essential to detect any islanding situation and to reduce the running

time of the system. This situation must be detected in order to:

To avoid feeding a fault or leaving a faulty system running.

1072 Mohamed Redha Rezoug et al.

To avoid feeding the islanding with an abnormal voltage or frequency.

To enable automatic reclosing systems to operate.

Also, the development of anti-islanding protection which is sensitive and reliable

is very important to encourage the integration of the distributed power generation

system (DPGS) in the electrical grid and avoid its untimely launching [3] - [6].

The design of PV inverters will be influenced by the requirements of the power

grid, including the anti-islanding (AI) requirement which is considered the most

technically difficult.

3 Method of Detecting Islanding

There are several methods of detecting islanding (or detection of loss of the main

grid). These methods can be divided into three categories: passive methods, active

methods, and methods of using communications between the main grid and the

PV inverter [12], [9].

3.1 Passive methods

They are based on the monitoring of the parameters of voltage and frequency or

their characteristics “harmonic, speed of change” etc. These methods require a

definition of the thresholds. If the preset threshold is exceeded, the inverter is

therefore disconnected. They are simple and easy to install with low currents and

do not need any additional materials. They do not cause disturbances to the grid or

inverters and have a rapid detection time. However, they are a great disadvantage

regarding to the definition of a threshold [1], [10], [2].

3.2 Active Methods

They are based on the injection at the output of the inverter (or the grid) of small

disturbances that can deviate a magnitude and thus detect more quickly the

islanding. However, the fact that harmonic currents are injected at the connection

node can cause variations in voltage, power or resistance of the grid and thus

degradation in the quality of the energy supplied. They are inefficient in the case

of several inverters in parallel (possible unjustified disconnection).

3.3 Methods Using Means of Communication

They are based on communication between the PV plant and the grid. They are

very fast allowing the non-degradation of the quality of the energy supplied. They

are therefore, very effective but their major disadvantage is their high cost and

that they are difficult to implant and need communication infrastructures as well.

4 Simulation of Anti-Islanding Protection Systems

In our work, we present a simulation study on a Matlab / Simulink environment of

a passive method of islanding detection which is based on the monitoring of the

Study and analysis of anti-islanding protection 1073

parameters related to the voltage on the DC side and the point of connection to the

PCC (Point of Common Coupling) grid, Current and frequency [7], [11], [4]. This

function is provided by the implementation of anti-islanding AI relays in the PV

system such as over-voltage and under-voltage (OV / UV), over-current and

under-current (OI / UI) relays and over-Frequency and under-frequency (OF /

UF). The simulation model of the PV system studied here is shown in Figure2.

Figure 2: Simulation model of the Grid-Tied PV System

Two cases are studied:

1st case: islanding is produced when the circuit breaker CB1 is open.

2nd case: islanding is produced when various grid faults are produced at a

distance of 8 Km away of the Point of Common Coupling PCC.

4.1 Passive Methods

An islanding state is simulated when the three-phase circuit breaker CB1 is

opened at time (t) = 0.3s and is closed at time (t) = 0.45s. In this case, we study

three scenarios depending on the connected load to the PV system.

Scenario 1: Local load is higher than local generated power.

Scenario 2: Local load is equal to the local generated power.

Scenario 3: Local load is lower than local generated power.

The implementation scheme of the various AI relays: OVdc, UVdc, OC, UC, OV,

UV, OF, UF are shown in the following Figures.

1074 Mohamed Redha Rezoug et al.

Figure 3: Simulation model of over/under Vdc relay

Figure 4: Simulation model of over/under Frequency relay

Figure 5: Simulation model of over/under Current relay

Study and analysis of anti-islanding protection 1075

Figure 6: Simulation model of over/under Voltage relay

4.2 Simulation result

In Figures (7, 8, 9, 10) there is a representation the variations of the voltage Vdc,

the efficient value of the voltage and current and the variation of frequency for the

three scenarios during islanding.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8200

250

300

350

400

450

500

550

600

650

700

Time (S)

Vdc voltage during islanding of PV system (Volts)

Reference Vdc voltage

Local load greater than local generation

Local load matches with local generation

Figure 7: The voltage Vdc during islanding for the different local loads

1076 Mohamed Redha Rezoug et al.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80

0.5

1

1.5

2

2.5

3x 10

4

Time (S)

RMS voltage during islanding at point PCC-B2 (Volts)

Local load greater than local generation

Local load matches w ith local generation

Local load less than local generation

Figure 8: The effective value of voltage during islanding for the different local

loads

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80

5

10

15

20

25

30

Time (S)

RMS current during islanding at point PCC-B2 (A)

Local load greater than local generation

Local load matches with local generation

Local load less than local generation

Figure 9: The effective value of current at the PCC point during islanding for

different values of the local load

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.849.5

50

50.5

51

51.5

52

Time (S)

Frequencey during islanding at point PCC-B2 (Hz)

Local load greater than local generation

Local load matches with local generation

Local load less than local generation

Figure 10: The variation of frequency during islanding for the different local loads

Study and analysis of anti-islanding protection 1077

Figures (11, 12, 13) represent the measured voltage of reference Vdc, the current

with reference Id, measured Id and Iq, the compound voltage Vab-Vsc (V) at the

output of the inverter, instant voltage and current at the point of PCC and the

power at point PCC during islanding for the case of a higher local load than the

local generated power (scenario 1).

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8480

500

520

540

560

580

600

620

640

660

Time (S)

Vdc reference voltage

Vdc measure voltage

Figure 11: Voltage Vdc reference and measured Vdc during islanding (load is

higher than local generated power)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

Time (S)

Id reference (pu)

Id measure (pu)

Iq measure (pu)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8-1000

-500

0

500

1000

Time (S)

Vab- VSC (Volts)

Figure 12: The compound voltage Vab-Vsc (V) during islanding

1078 Mohamed Redha Rezoug et al.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8-4

-2

0

2

4x 10

4

Time (S)

Voltage at PCC-B2 (Volts)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8-40

-20

0

20

40

Time (S)

Current in PCC-B2 (A)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8-500

0

500

1000

Time (S)

Power at PCC-B2 (Kw)

Figure 13: Voltage of phase Va, line current and power at the point PCC during

islanding

A graphic representation of the detection time of different anti-islanding AI relays

for test case 1 (the circuit breaker in open position) is shown in Figure 14.

When comparing the performance of AI relays by detection time, we notice that

[11]:

The OC relay has a very longer detection time in the two cases of equal and

higher load than the local generated power. On the contrary, in the case the load is

lower than the local generated power, the detection time is short.

The UV relay is the fastest in detecting the islanding with respect to the other

AI relays and it has the same detection time for the three load levels.

UVdc took a long detection time in cases of higher and lower loads and failed

to detect in the case of load equal to the local generated power.

UC, OV, UF, OVdc completely failed to detect islanding in the three load-

level cases.

Study and analysis of anti-islanding protection 1079

Figure 14: Anti-islanding relays detection time in the case the circuit breaker is in

open position

4.3 Test case 2 faults that occurs in the power grid

Different grid faults (single-phase-to-ground, phase-to-phase, phase-to-phase-to-

ground, three-phase and three-phase-to-ground faults) lasting for 150 ms were

simulated at 8 Km away from the PCC, Figure 15.

Figure 15: Simulation model of PV system during islanding created by defaults

The simulation model for sub-fields 7 and 8 is shown in Figure 16. Two circuit

breakers CB3 and CB4 are implemented at the output of the inverter.

1080 Mohamed Redha Rezoug et al.

Figure 16: Simulation model of sub-fields 7 and 8 with the implementation of

controlled circuit-breakers

Figure 17 shows the comparison of detection times of the AI relays for test case 2

if different defaults occur in the grid. We notice that:

The UV relay has the best performance as it is noted in the study of the first

case. It is the fastest.

The UC relay also set a short detection time of islanding for the 4 faults:

phase-to-phase, phase-to-phase-to-ground, three-phase and three-phase-to-ground

and it were late for the single-phase-to-ground fault.

The UF relay was very late in detecting all faults.

The OC, OV, OF, OVdc and UVdc relays completely failed to detect

islanding for all types of defaults.

Study and analysis of anti-islanding protection 1081

Figure 17: Anti-islanding relays detection time in the case of different grid faults

To compare the detection times of the AI relays in both cases: the case of the

circuit-breaker in open position with the three scenarios of variation of the local

load and the case of the occurrence of different types of faults at point 8Km, the

simulation results are shown in the Table 1.

It gives the results of the theoretical simulation of the AI detection methods

considering: the state of the relays and the time (s) for detecting insularity by relay

protection. The results of the electrical grid monitoring are presented for both

cases: in case of simulation 1, when the three-phase circuit breaker CB1 is in the

open position for scenarios where the power generated locally by the photovoltaic

central is higher, less and roughly balanced with the power of the consumers

connected to Local level; And in case of simulation 2, for scenarios where

different types of defects are simulated in the utility public grid.

The minimum current relay (OC), does not meet the conditions for detecting an

abnormal operating situation in any of these cases. The maximum frequency relay

(OF), is only activated in case 1, when the power of locally connected consumers

is less than or equal to the power generated locally. The maximum voltage relay

(OV), is activated only in case 2. Right after the isolation conditions and three-

phase faults in the mains supply, the voltage drops, the current increases and the

frequency changes.

* Status of the relay;

** Detection time (s) of anti-islanding condition.

1082 Mohamed Redha Rezoug et al.

Table 1: Detection time of anti-islanding relays

Test

case scenario

Type of anti-islanding detection method

OC UC OV UC OF UF OVdc UVdc

* ** * ** * ** * ** * ** * ** * ** * **

1.

CB1

open

Higher load

than

production

1

0.0

2

0 - 0 - 1

0.0

1

1

0.1

0

0 - 0 - 1

0.1

6

Equal load to

the

production

1

0.1

6

0 - 0 - 1

0.0

1

1

0.0

7

0 - 0 - 0 -

Lower load

than

production

1

0.1

8

0 - 0 - 1

0.1

3

1

0.0

6

0 - 0 - 1

0.2

0

2.

Fault

in the

grid

Single phase

to ground 0 - 1

0.0

7

0 - 1 0

.01

0 - 1

0.1

7

0 - 0 -

Isolated

biphase 0 - 1

0.0

3

0 - 1

0.0

1

0 - 1

0.4

5

0 - 0 -

Biphased to

ground 0 - 1

0.0

3

0 - 1

0.0

1

0 - 1

0.3

2

0 - 0 -

isolated

three-phase 0 - 1

0.0

3

0 - 1

0.0

1

0 - 1

0.2

5

0 - 0 -

Three-phase

to ground 0 - 1

0.0

3

0 - 1

0.0

1

0 - 1

0.3

8

0 - 0 -

5 Conclusion

This paper presents the results of a simulation study concerning the islanding

phenomenon occurs in a photovoltaic PV system. Due to the harmful effects of

islanding on the equipment and on the maintenance personnel as well, we study

the different methods for its detection. Three methods are analyzed (passive,

active and communication between grid and PV inverter) in an attempt to

elaborate a passive protection method consisted of monitoring the parameters of

voltage, current and frequency of DC and AC sides in the PV system by

implementing different anti-islanding relays at fixed points on the PV installation.

Two cases are tested, the first when circuit-breaker is in open position and the

second by creation of various defaults at a point away from the Point of Common

Coupling PCC.

Study and analysis of anti-islanding protection 1083

By comparing the anti-islanding relays detection time, the simulation results allow

us to determine the most efficient relays in each case. These results are of great

importance for the creation of AI protection devices for Grid-Tied PV Systems.

Acknowledgements: Any collaboration with central GHARDAÏA

Nomenclature

AI: Anti-islanding

CB: Circuit breaker

OC: Maximum current relay

OF: Maximum frequency relay

OV: Maximum voltage relay

PV: Photovoltaic

UC: Minimum current relay

UF: Minimal frequency relay

UV: Minimal voltage relay

VSC: Voltage Source Converter

References

[1] A. Khamis, H. Shareef, E. Bizkevelci and T. Khatib, A review of islanding

detection techniques for renewable distributed generation systems,

Renewable and Sustainable Energy Reviews, 28 (2013), 483 - 493.

https://doi.org/10.1016/j.rser.2013.08.025

[2] A.V. Soumya and J. Belwin Edward, Review of Islanding Detection

Techniques for Distributed Energy Sources, National Conference on

Emerging Trends in Engineering and Technology (NCETET’16), Vol. 3,

(2016), 135 - 138.

[3] B. Yu, M. Matsui and G. Yu, A review of the current methods for

photovoltaic power system, Solar Energy, 84 (2010), 745 - 754.

https://doi.org/10.1016/j.solener.2010.01.018

[4] C. Li, C. Cao, Y. Cao, Y. Kuang, L. Zeng and B. Fang, A review of islanding

detection methods for microgrid, Renewable and Sustainable Energy

Reviews, 35 (2014), 211 - 220. https://doi.org/10.1016/j.rser.2014.04.026

[5] D. Rekioua and E. Matagne, Optimization of Photovoltaic Power Systems,

Modelization, Simulation and Control, Springer, London, 2012.

https://doi.org/10.1007/978-1-4471-2403-0

1084 Mohamed Redha Rezoug et al.

[6] I.V. Banu, M. Istrate, D. Machidon and R. Pantelimon, A study on anti-

islanding detection algorithms for grid-tied photovoltaic systems, 2014

International Conference on Optimization of Electrical and Electronic

Equipment (OPTIM), Bran, Romania, (2014).

https://doi.org/10.1109/optim.2014.6850940

[7] I.V. Banu and M. Istrate, Study on three-phase photovoltaic systems under

grid fault, 2014 International Conference and Exposition on Electrical and

Power Engineering (EPE), Bran, Romania, (2014).

https://doi.org/10.1109/icepe.2014.6970086

[8] J. Dvarioniene, J. Šinkūnienė, I. Gurauskienė and Ž. Stasiškienė, Analysis of

integration of solar collector systems into district heat supply networks,

Environmental Engineering and Management Journal (EEMJ), 12 (2013),

2041 - 2050.

[9] M. Ciobotaru, V. Agelidis and R. Teodorescu, Accurate and less-disturbing

active anti-islanding method based on PLL for gridconnected PV Inverters,

IEEE Power Electronics Specialists Conference (PESC), (2008), 4569 -

4576. https://doi.org/10.1109/pesc.2008.4592685

[10] S. Raza, H. Mokhlis, H. Arof, J.A. Laghari and L. Wang, Application of

signal processing techniques for islanding detection of distributed generation

in distribution network: A review, Energy Conversion and Management, 96

(2015), 613 - 624. https://doi.org/10.1016/j.enconman.2015.03.029

[11] S. Shrivastava, S. Jain, R. Nema and V. Chaurasia, Two level islanding

detection method for distributed generators in distribution networks,

International Journal of Electrical Power & Energy Systems, 87 (2017), 222

- 231. https://doi.org/10.1016/j.ijepes.2016.10.009

[12] W. Y. Teoh and C. W. Tan, An overview of the methods of detection of

islands in photovoltaic systems, International Journal of Electrical and

Computer Engineering, 5 (2011), 1336 - 1344.

Received: February 24, 2018; Published: April 23, 2018