Performance improvement of off-grid hybrid renewable energy …€¦ · Performance improvement of oﬀ-grid hybrid renewable energy system using dynamic voltage restorer Wael S.

Alexandria Engineering Journal (2020) xxx, xxx–xxx

HO ST E D BY

Alexandria University

Alexandria Engineering Journal

www.elsevier.com/locate/aejwww.sciencedirect.com

Performance improvement of off-grid hybrid

renewable energy system using dynamic voltage

restorer

* Corresponding author.E-mail address: [email protected] (M.I. Mosaad).

https://doi.org/10.1016/j.aej.2020.03.0371110-0168 � 2020 Faculty of Engineering, Alexandria University. Production and hosting by Elsevier B.V.This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article in press as: W.S. Hassanein et al., Performance improvement of off-grid hybrid renewable energy system using dynamic voltage restorerandria Eng. J. (2020), https://doi.org/10.1016/j.aej.2020.03.037

Wael S. Hassanein a, Marwa M. Ahmed b, M. Osama abed el-Raouf c,

Mohamed G. Ashmawy d, Mohamed I. Mosaad e,*

aDepartment of Industrial Engineering, Faculty of Engineering, King Abdulaziz University, Saudi ArabiabDepartment of Electrical Engineering, Faculty of Engineering, King Abdulaziz University, Saudi ArabiacBuilding Physics and Environmental Research Institute, Housing and Building, National Research Center (hbrc), Cairo, EgyptdDepartment of Electrical Power and Machine Engineering, El-Shorouk Academy, Cairo, EgypteYanbu Industrial College (YIC), Saudi Arabia

Received 28 December 2019; revised 21 March 2020; accepted 28 March 2020

KEYWORDS

Artificial Neural Network;

Dynamic voltage restorer;

Fuel cell;

Hybrid system;

Low voltage ride through;

Photovoltaic;

Wind turbine

Abstract This article proposes an Artificial Neural Network (ANN) controller of Dynamic Volt-

age Restorer (DVR) to improve the performance of a stand-alone hybrid renewable energy system

that is feeding a new community located in Egypt. The hybrid system consists of three renewable

energy sources, namely, solar PV cells, a wind turbines based-permanent magnet synchronous gen-

erator, and fuel cells. These three sources are tied to a common DC link by three boost converters,

one for each source. The common DC link is connected to the AC side via a DC/AC inverter. The

optimal size of the three proposed renewable sources is calculated using the HOMER software

package. The DVR control is attained through regulating the load voltage at different anomalous

working conditions. These conditions are three-phase fault, voltage sag/swell, and unbalanced load-

ing. Two ANNs are utilized to adjust the IGBT pulses of the voltage source inverter (VSI) used to

control DVR by regulating the D-Q axes voltage signals. These D-Q axes components at any load-

ing condition represent the inputs to the two ANNs. The outputs of the two ANNs represent the

IGBT pulses. The input/output data used for training ANNs are obtained by two optimized PI con-

trollers, introduced for regulating the load voltage through DVR-VSI pulses at different abnormal

operating conditions, and accordingly convert the static optimized PI controller into adaptive one

based ANN. The system performance with the proposed ANN-DVR controller is enhanced

through improving the current, voltage, and power waveforms of each generating source. With

compensation of the faulty line voltage, the system retains an uninterrupted operation of the three

renewable sources during fault events and consequently increases the low voltage ride through

(LVRT) capability. Moreover, the total harmonic distortion is reduced.� 2020 Faculty of Engineering, Alexandria University. Production and hosting by Elsevier B.V. This is an

open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

, Alex-

http://creativecommons.org/licenses/by-nc-nd/4.0/

mailto:[email protected]

https://doi.org/10.1016/j.aej.2020.03.037


http://www.sciencedirect.com/science/journal/11100168


http://creativecommons.org/licenses/by-nc-nd/4.0/


Nomenclature

a area of PV

A swept area of rotor in m2

G irradiance (W/m2)ID saturation current of the diode of equivalent pho-

tovoltaic cell circuit

Id & Iq d and q axes currentsIL photocurrent of equivalent photovoltaic cell cir-

cuit

Isc PV cell short-circuit current at 25 �C and 1000W/m2

ISCðT1Þ short circuit current per cell at temp T1

ISCðT2Þ short circuit current per cell at temp T2

k Boltzmann’s constantK0 constant which determined from ISC vs. TKp proportional gain

Ki integral gainNpv number of PV panels

n ideality factor

Ppv power of photovoltaic panelq charge of an electronSr solar radiationSTC Standard Test Conditions

T temperature in �CVg band gap voltagevocðT1Þ open circuit voltage per cell at temperature T1

vocðT2Þ open circuit voltage per cell at temperature T2

voc open circuit voltageVd & Vq d and q axes voltages

Vki velocity of particle i at iteration k

W weighting functionx number of populationsXk

i current position of particle i at iteration k

g PV system efficiency

2 W.S. Hassanein et al.

1. Introduction

Renewable energy plays a major role in feeding electricity to

remote areas. It is the ideal alternative to traditional energy

as it is inexhaustible, and it has proved its importance and

achieved many important privileges as an alternative energy

for the future instead of fossil fuels, which will be exhausted

one day. Many developed and developing countries have

turned to renewable energy to reduce harmful emissions to

the atmosphere, reduce global warming, and save costs.

Renewable energy is available continuously and each country

can rely on the most available sources [1]. For example, Arab

countries can use solar energy to generate electricity and water

heating and others as the sun shines most months of the year.

Recently, many developed and developing countries use

renewable energy as unpolluted and renewable alternative, in

place of using contaminated materials [1]. Hybrid systems

are important for several reasons, the most important of which

are: to ensure the continuous feeding of the electrical loads at

various operating conditions, and to optimize the size of the

resources in remote areas [2]. It is preferable to use a hybrid

system, because the total capacity of more generation source

is smaller than the capacity of a system consisting of only

one source, whether solar or wind, as well as for the storage

capacities of batteries. One of the main advantages of using

hybrid systems in remote areas is that securing the electrical

supply from different power sources increases the reliability

and durability of the system [2,3].

A hybrid system (PV-Wind-fuel cell) tied to the grid is uti-lized to feed residential thermal and electrical loads with higheconomic efficiency as indicated in (A. Maleki et al. 2017 a) [4].

Another application for grid-connected PV systems consider-ing economic issues and investigating different technical solu-tions was presented in (O. Ayadi et al. 2018) [5]. Artificial

intelligence techniques have significant role in controlling therenewable energy systems. (A. Maleki et al. 2017) [6], presentedtwo heuristic approaches based-PSO which, are applied to PV-fuel cell-natural gas connected to grid to supply residential

Please cite this article in press as: W.S. Hassanein et al., Performance improvement oandria Eng. J. (2020), https://doi.org/10.1016/j.aej.2020.03.037

loads. (G. Zhang et al. 2018) [7], studied another applicationfor desalination plants that is dependent on the renewableenergy supplies such as wind and solar energies where simu-

lated annealing-chaotic search algorithm were utilized as anoptimization technique. (Shaimaa Barakat et al. 2018) [8], pre-sented a techno economic studies for hybrid renewable energy

systems with different optimization techniques such as opti-mizing a PV/Biomass hybrid system with different batterytechnologies.

There are some problems associated with integrated hybrid

power systems, such as harmonics issue generated from invert-ers that need adding extra equipment with more cost [3,9], inhybrid systems, there are different generation sources, and

any sudden change in the load or in the power generated fromany source can affect the system stability [9]. Load sharingissue in renewable energy hybrid system due to dependence

on renewable generation sources. The power economic dis-patch and calculating cost per unit generation are complicated[10]. Moreover, the short circuit occur in hybrid renewable

energy system sometimes lead to outage of the generationsource from system, especially if the control is not accurate.This short circuit, in case of wind systems may accelerate thegenerator speed to higher values especially without using con-

trollers that lead to disconnecting the generator from the sys-tem or damage the shaft [11].

Some devices were added to overcome these problems for

such hybrid systems [12]. Dynamic Voltage Restorer (DVR)and Static Transfer Switch were used for enhancing the perfor-mance of wind energy system (WES) and the system has the

ability to mitigate the disturbances applied to these systemsas in [13]. A. Prajyusha [14] presented a DVR that is capableof mitigating the power quality issues. Other devices including

Unified Power Quality Conditioner (UPQC) and D-STATCOM were used to enhance the performance of therenewable systems as in [15].

The control of these frameworks is considered as an imper-

ative issue [9,10]. The control process of these devices espe-cially when integrated to the hybrid systems is complicatedas the nonlinearity and complexity. Many control schemes

f off-grid hybrid renewable energy system using dynamic voltage restorer, Alex-


Performance improvement of off-grid hybrid renewable energy system 3

are presented to drive such devices to enhance the performanceof the hybrid systems [16].

PI controller method is introduced in many power system

control applications. PI controller is introduced to regulatethe inverter used to tie Fuel Cell (FC) to the electrical network[17]. PI controller parameters are optimized by many optimiza-

tion techniques from them particle swarm optimization (PSO).This optimization technique for finding the PI controllerparameters is performed off-line and for the purpose of online

control, some adaptations are needed as the optimal controllerparameters and optimal performance are not guaranteed whenchanging the operating conditions. The solution is using adap-tive control techniques for the offline optimal control. Many

adaptive control schemes are introduced such as adaptive PIbased ANN that was introduced to convert the static PI con-trollers into adaptive one as in [18]. Model reference predictive

control of STATCOM for enhancing the WES performancewas presented in [11].

The main contribution of this paper is using adaptive ANN

controller of DVR to enhance the effectiveness of off-gridhybrid system bolstering a new community in Egypt-case studyNew El-Farafra Oasis. The proposed controller (ANN-DVR)

regulates the voltage between DVR and the load to alleviatethe disturbances applied to the system and consequentlyimprove the system performance. A comparison between PItuned by PSO and adaptive ANN for controlling DVR is

summarized.

Fig. 1 (a) Block diagram of the proposed


The paper is organized as follows. Section 2 describes thesystem with all its components. Section 3 represent the model-ing of the three renewable sources while Section 4 describes the

controlled design. Section 5 represents the results and discus-sions. Finally, Section 6 is the conclusions.

2. System description

The proposed system consists of a hybrid PV/WES/FC for bol-stering an off-grid community placed in Egypt, new Farafra

oasis [19,20]. The system consists of PV system tied to the firstDC-DC boost converter; WES is tied to the second DC-DCconverter while FC is tied to the third DC-DC converter.

The three DC-DC converters are connected to a DC link of780 V. Community AC loads are fed through an inverter.Two types of transformers are used, step up transformers are

used to feed off-grid network of the proposed communitywhich stepped up to 11 kV, and step-down transformers areused for supplying the AC loads of the community,11/0.380 kV, 50 Hz. DVR controlled by ANN, is connected

to the system at point of common coupling (PCC) to mitigatethe voltage swell, sag, unbalance and improve the power qual-ity of the system. Any distortion in supply voltage Vs (feeder

F1) may be compensated at the load voltage (feeder F2) byinjecting appropriate voltage from voltage source inverter(VSI) through an injection transformer, Fig. 1. The inverter

hybrid system and (b) DVR component.




output voltage is calculated according to system ratings. Allsystem parameters are given in details in [2].

3. System modeling

3.1. PV modeling

PV equivalent circuit depicted in Fig. 2, contains photo currentIL, and RS & Rsh series and parallel resistances respectively

[21].The modeling of the PV cell can be described as:

I ¼ IL � ID eq vþIRsð Þ

nkT � 1n o

ð1Þ

IL ¼ ILðT1Þð1þ K0 T� T1ð ÞÞ ð2Þ

ILðT1Þ ¼ G� ISC T1;nomð Þ=G nomð Þ ð3Þ

K0 ¼ ðISC T2ð Þ � ISC T1ð ÞÞ=ðT2 � T1Þ ð4Þ

ID ¼ IDðT1Þ � T

T1

� �3n

� e�qvg

nk� 1

T� 1T1

� �ð5Þ

IDðT1Þ ¼ ISCðT1Þ=ðeqvoc T1ð ÞnkT1 � 1Þ ð6Þ

RS ¼ �dv=dIvoc � 1=XV ð7Þ

XV ¼ I0ðT1Þ � q

nkT1

� eqvoc T1ð ÞnkT1 ð8Þ

where k and q are 1.38e�23 and 1.60e�19 respectively.

The PV power at a time t is determined as [21]:

Ppv tð Þ ¼ Sr tð Þ � a� g ð9ÞThe total power of the PV system can be defined as:

PtðtÞ ¼ Npv� PpvðtÞ ð10Þ

3.2. Wind turbine modeling

Kinetic energy from wind is converted to mechanical powerthrough wind turbine. Many types of generators were usedand implemented in WES. Induction generators with their

common types namely self-excited [22] and doubly-fed induc-tion generators [23,24] were used in many WES applications.Switched reluctance generators were also used in WES in both

distribution and generation levels [18,25]. PMSG were used inWES due to their ability of operation at variable wind speedconditions and gearless operation [26]. The WES-based PMSG

was presented in off-grid hybrid renewable system in [26].

Fig. 2 The equivalent circuit of solar cell.


3.3. Fuel cell model

Fuel cell changes chemical energy into electrical energy. Theprinciple of its operation is the opposite of the principle ofthe work of the water electrolysis vessel, which analyzes the

water into its compound’s hydrogen and oxygen, consuminga certain amount of electrical energy. The Solid Oxide FuelCell (SOFC) unit used is depicted in Fig. 3. The fuel cell isfed with hydrogen from the side of the anode, so this electrode

is called a ‘‘fuel electrode”. Electrons that are not passed by theelectrolyte have to flow through an external circuit (e.g., theload) to land on the other pole of the cell, the cathode, which

forms the passage of electricity. At the cathode, the cell is fedwith air to provide the oxygen needed to make it work, thisoccurs in one of the poles which determined by cell type.

The electrochemical reaction within the cell consists of twoparts: the first is at the anode and the second is at the cathode.FC modeling and data are given in [2,27].

3.4. Optimal size of the PV/Wind/FC hybrid system

The Egyptian Code for electrical installations in buildingsstates that” electricity distribution companies in Egypt prepare

schedules to estimate the total electrical loads of all types ofbuildings”. These schedules, give the minimum total electricalcapacity that companies can agree with the owners of the

buildings under which they are fed by electric current. Thecapacity depends on the total area of all the floors of the build-ing and the nature of the activity in the building (residential,

educational, administrative, commercial, etc.. . .) and whetherthe building is located in an urban or rural area and the clas-sification of its location in those areas, etc. In this work, a newcommunity is proposed in new Farafra oases western desert of

Egypt, as a case study [30]. It consists of three loads (residen-tial load with services, light and medium industries loads andagriculture loads) these loads are now existing loads in Egypt,

so these loads are used in a recently found region in a desert ofEgypt as a case study. The new community is planned for20,000 inhabitants where, residential loads with services con-

sumptions represent 11,000 kWh/day, light and medium indus-tries load with services consumptions reflect a 174.8 MWh/day,

Fig. 3 Scheme of solid-oxide fuel cell.




and deferred load includes the agricultural load of thecommunity. The climbed yearly average deferred load is5772 kWh/day and with maximum load of 1443 kW. Agricul-

ture loads (irrigation using pumping water systems) represent5772 kWh/day.

HOMER software is utilized to determine the optimal size

of PV/Wind/FC hybrid system to feed the AC loads of a newcommunity based on some economic criteria [28,29]. The aver-age wind speed at the chosen site is 5.45 m/s and the radiation

is 8 kWh/m2/day.

Fig. 4 (a) Flowchart of DVR operation, (b) Control block

diagram, (c) PI control circuit.


4. Controller design

4.1. Dynamic voltage restorer

As any disturbance is applied to the system, the DVR willintroduce a controlled voltage over an injection transformer

to mitigate that fault. Different disturbances are applied tothe system such as voltage sag, swell, unbalanced operationand three-phase short circuit. The DVR will mitigate these dis-

turbances. The flow chart of DVR process is given in Fig. 4a.The control circuit of DVR along with PI controller aredepicted in Fig. 4b and c. The ABC three-phase coordinatesystem is changed over into dq0 coordinate system as

follows [2]:

Vd ¼ 2

3VaSinwtþ VbSinVc sin wtþ 2p

3

� �� ð11Þ

Vq ¼ 2

3VaCoswtþ VbCos wt� 2p

3

� �þ VcCos wtþ 2p

3

� ��

ð12Þ

V0 ¼ 1

3Vaþ Vbþ Vc½ � ð13Þ

Table 1 DVR DATA.

Description Symbol Specifications

DVR DC voltage Vdc 700 V

DVR capacitor DC link C DC 5000 mF

Fig. 5 Flow Chart of PSO.




The load voltage (VLoad) is sensed and transformed intodq0 coordinates. The difference between the dq and the refer-ence voltages of dq coordinates represent the input to the PI

controllers, the d-axis reference is 1p.u and q-axis referenceis 0, Fig. 5c. Two PI controllers are utilized for d and q errorsignals PId, PIq respectively. The errord represents the input to

the D-axis PI controller while the errorq is for Q-axis PIcontroller.

Fig. 6 ANN controller for d-q coordinates of DVR.

Fig. 7 Nordal grid code.

Table 2 Optimal PI controller’s parameters determined by PSO.

Cases D-axis Q-

Kpd Kid Kp

3-phase fault 9.6 200.03 78

sag voltage

sag 30% 79.2 195.90 85

sag 60% 11.3 210.03 82

sag 90% 9.36 201.63 77

Swell voltage

Swell 30% 17.8 215.1 83

Swell 60% 21.94 217.92 85

Swell 90% 25.84 220.84 90

Unbalance voltage

Unbalance 4.003 154.32 52


The integral of the square of error (ISE) index is utilized tostudy the system effectiveness as:

ISE ¼Z 1

0

e2 tð Þdt ð14Þ

e ¼ errord þ errorq ð15Þ

errord ¼ Vdref � Vd ð16Þ

errorq ¼ Vqref � Vq ð17ÞThe output of the PI controller is converted into ABC then

to PWM to activate the VSI-IGBT. The DVR data are shownin Table 1.

4.2. PSO algorithm for optimal tuning of PI control parameters

There are different optimization techniques for adapting PI

controller. PSO is one of them, where it is developed by search-ing on swarm. The solution of the problem represented by aparticle. Each particle has its own flying which it is adjusted

according to its flying background. In the D- dimensionalspace, the particle treated as a point. The ith particle is denotedas XI = (xi1, xi2. . .). Best value which called (pbest) is calcu-

lated according to minimum fitness value and represented asPI = (PI1, PI2. . .). In PSO technique the velocity of the particleis referred as VI = (VI1, VI2,. . .), the particles are updatedaccording

Vkþ1i ¼ WVk

i þ C1rand1 � pbesti � Xki

� þ C2rand2

� gbest� Xki

� ð18Þ

where Vki : Velocity of particle i at iteration k, W: weighting

function, C: weighting factor, j rand: random number between

0 and 1, Xki : Current position of particle i at iteration k, pbesti:

pbest of particle i, gbest: gbest of the group. The PI controlleroptimized parameters are determined for each test case. The

PSO flowchart for adapting the PI controller parameters isshown in Fig. 5.

axis D-axis Q-axis

q Kiq Vd Vq

.52 191.72 263.3 272.8

.00 124.91 270.4 265.9

.41 296.02 273.7 263.8

.83 190.84 265.3 272.2

.02 266.01 261.8 274

.40 269.03 273.1 265

.83 280.93 268.8 269.8

.43 180.52 263.7 272.6



(a) PV voltage (b) PV current (c) PV power.

(a) Without DVR With DVR

Without DVR With DVR


(b)

(c)

Fig. 8 PV system with and without DVR at three-phase fault.


Please cite this article in press as: W.S. Hassanein et al., Performance improvement of off-grid hybrid renewable energy system using dynamic voltage restorer, Alex-andria Eng. J. (2020), https://doi.org/10.1016/j.aej.2020.03.037


(a) FC voltage (b) FC current (c) Fuel cell power

With DVR Without DVR (a)

(b)

(c)



Fig. 9 FC system with and without DVR at three-phase fault.


Please cite this article in press as: W.S. Hassanein et al., Performance improvement of off-grid hybrid renewable energy system using dynamic voltage restorer, Alex-andria Eng. J. (2020), https://doi.org/10.1016/j.aej.2020.03.037


Fig. 10 WES with and without DVR at three-phase fault.


4.3. Adaptive control techniques (ANN controller)

As discussed in previous section, the difficulty and more run-ning time are considered as a disadvantage in using offline con-trol techniques, so that adaptive control plays an important

role to increase the system effectiveness with more rapidresponse. In the present application, two ANNs are used,


one for D-axis and the other for the Q-axis. Each ANN hasthree layers namely input layer representing the D-axis andQ-axis values of the load voltage change for first and second

ANN respectively, a hidden layer including 20 neurons andan output layer representing the D-axis and Q-axis voltagesfor the first and second ANN respectively with supervised

training via a back-propagation as depicted in Fig. 6. The out-put voltage signals of the two ANN in the D-Q are trans-formed into (Va, Vb, Vc) to drive PWM of voltage source

inverter of DVR. The ANN training is obtained by eightabnormal operating conditions counting three phase fault,voltage sag/swell and unbalanced operation. At each conditionthe load voltage is sensed and transformed to D-Q frame that

represent the input to the two ANNs. Tuning of the two PIcontroller proposed by PSO is introduced at each time (eighttimes) for optimal control of DVR to increase the hybrid sys-

tem execution and the D-Q frame of the voltage control signalsare used as the output of the two ANNs. The D-Q frame-controlled voltage signals are converted into Vabc components

to drive the PWM of IGBT used in the VSI of DVR system. Inthis work ISE performance function has been used to representthe error between the target and input values. In this paper, the

number of epochs required for training D coordinate ANN is500 and the best validation performance is 0.51137 at epoch500 and the number of epochs required for training Q coordi-nate ANN is 450 and the best validation performance is

0.40037 at epoch 400.

4.4. Assessment of integrating DVR into hybrid system throughLVRT

The DVR is integrated into the hybrid PV/wind/fuel cell tosupport the dynamic performance of the proposed system dur-

ing abnormal operating conditions. Through using LVRTcapability, the hybrid system especially for wind generators,the system will be able to remain in service during abnormal

operating conditions. Nordal grid code [31] is used in this arti-cle, as an example, as shown in Fig. 7.

5. Results and discussions

The proposed system bolstering a new community in Egypt,given in Fig. 1 is introduced for testing the validity of the pro-posed ANN-DVR controller over PI controller optimized by

PSO considering sag, swell, unbalance voltages and three-phase short circuit.

5.1. Homer results

Homer software [28] was used for determining the optimal rat-ing of each generating source introduced in this paper. With

eight different scenarios for all possible combinations of threesources, the results show that, the optimal combination ofthree renewable sources are 30.687, 13.5 and 46 MW for PV,

WES and FC systems respectively, generic 1 kWh lead-acidbattery (117,089 strings) and system converter (32.911 MW)with a dispatch strategy of load. Moreover, the added to netdisplay cost is $ 265 M, the capital price is $ 193 M and the

cost of electricity is $ 0.293/kWh.




5.2. DVR control by PSO-PI controllers

In this part, the DVR integrated into the hybrid renewableenergy system is regulated by two PI controllers optimizedby PSO to keep to the Nordal grid code and to improve the

overall system performance. Eight test cases were presentedcounting three-phase fault, sag/swell with three levels of volt-age for each, and unbalanced voltage. These different casesare simulated for obtaining the PI control parameters required

for control DVR and the corresponding D/Q voltage signalsrequired for training the ANNs. The PSO optimized PI con-troller parameters for all the above-mentioned cases are given

in Table 2. Through these cases, the regulated voltage andeffects of DVR on load voltages are studied. Also, the effectson rotor wind speed before and after using a DVR are illus-

trated. Moreover, the role of connecting DVR to the systemto keep the normal operation of PV/Wind/Fuel cell without

(a) Output load voltage

WW

WiWiLV

(a)

(b)

Fig. 11 Output voltage with and w


disconnection during voltage dips due to several fault eventsare taken place at F1.

5.3. Test case 1: three-phase fault at F1

In this case three-phase fault is applied at feeder F1 with abetween 0.5 and 0.75 s. This fault affects the output of the gen-

eration sources. The voltage of PV decreases, the currentincreases and as a result the power increases system as shownin Fig. 8.

With this fault event, a decrease is taken place in the FCvoltage. While increase in the FC current to 150% that will dis-connect FC from the system, Fig. 9.

Due to this fault, wind system current increases as shown inFig. 10b. Through this fault, the electrical torque throbs incor-porates more motions with low-frequency esteem and this maylead to failure of the system due to large vibrations on the

, (b) Voltage at PCC

ithout DVR ith DVR PI-PSO

thout DVR th DVR PI-PSO RT

ithout DVR at three-phase fault.



(a) Sag conditions (b) swell conditions

Without DVR DVR with static PI DVR with retuned PI


(a)

(b)

Fig. 12 RMS of the load voltage with and without DVR.

(a) Voltage at PCC, (b) RMS output load voltage


Without DVR DVR with static PI DVR with retuned PI LVRT

(a)

(b)

Fig. 13 Phase-a voltage with and without DVR.

Without DVR DVR with static PI DVR with retuned PI DVR with adapted ANN LVRT

Fig. 14 Voltage at PCC without and with DVR based PI-PSO,

based ANN.


mechanical part of the wind system. The generator speed isincreased to high values (20%) without using DVR that willlead to disconnecting the generator from the system as the

speed surpasses greatest permissible cutoff of 17% [11]. Thewind turbine rotor speed is depicted with and without DVRin Fig. 10d.

The output voltage when using controlled DVR shows bet-

ter performance than without using DVR, Fig. 11a during thisthree-phase fault. DVR has the capacity to regulate the voltageat PCC with less harmonics during this fault conditions using

PSO-PI controller and consents to the Nordal grid code pro-posed, Fig. 11b.

5.4. Test case 2: Voltage sag/swell conditions

In this case, 30% voltage sagging condions were considered toassess the effictivness of the tunned PI controller for DVR to

amelorate the stsyem effectivness. Moroever, a comaprisonbetween staic tunned PI parameters obtained for the three-phase fault case (test case1) and the retuned PI control param-etes obtained for the 30% voltage sagging condition is intro-

duced as given in Table 2. RMS load voltage enhancement isshown in Fig. 12a. Although the load voltage levels areimproved using DVR with both static and retuned PI con-

trollers, the retuned PI controllers give better performancedue to the PSO is tunned at the 30% sagging condition. Con-sequently and in order to get the best (optimal) performance at


each abnormal condition, PI controller parameters should beretuned at each condition that can not be achieved as this tun-ning process is performed off-line rather than long time taken

for each tuning. The adaptive control is the key solution forthe retunned PI control paramerets hence, adaptive ANN con-trol will be used. Eight test cases summerized in Table 2 are




used for training the ANN. On the other hand, with 30% swellconditions, the RMS output load voltage enhancement at swellvoltage is shown in Fig. 12b. With DVR and the PI controllo-

ers are not retuned for this condition (vaules of case study1),the load voltage at this condition reaches to 430 V. With theretuned PI controller parameters, the value is 380 V. This

ensure the need for adaptive control rather than optimizedPI controller.

(a) Voltage at F1, (b)(d) Rotor wind turb

WWW

(a)

(d)

Fig. 15 System performance during three-phase


5.5. Test case 3: Unbalanced voltage condition

A single line to ground fault at phase-a, is applied to the sys-tem in this case. Without using DVR, the PCC voltage reaches0.37p.u with more oscillations, but with DVR and static PI

controller, the PCC voltage reaches 0.58p.u while with retunedPI controller parameters, the value is 0.68p.u., Fig. 13a. Theload voltage with and without DVR at unbalance voltage is

Load (c) Injected voltages ine speed

ithout DVR ith DVR PI ith DVR ANN

(b)

(c)

short-circuit with DVR based adapted ANN.




shown in Fig. 13b. Although voltage enhancement at PCC andRMS load voltage, this tunning process is performed off-linerather than long time taken for this tuning, so that adaptive

ANN controller will be used.

5.6. ANN adaptive control

Now ANN adaptive control technique is applied to the hybridsystem. As mentioned before, two ANN networks are used, thefirst for D-axis and the second for Q-axis. The input to ANN is

the change of load VL in D-Q coordinates and the output isupdated three phase voltages (Vd, Vq, V0) control parametersused to drive PWM of voltage source inverter of DVR. Differ-

ent test cases are used to test the effectiveness of PI controller.Eight test cases summerized in Table 2 are used for training theANN with obtaining an online tuning for control parameters.With three-phase fault and without using DVR, the PCC volt-

age raches 0.39p.u, Fig. 14. With DVR and the PI controlloersare not retuned for this condition (vaules of case study1), theminimu PCC volage at the sagging interval is 0.58p.u while

with retuned PI controller parameters, the minimum value is0.78p.u, while adaptive ANN controller, the minimum valuereaches to 0.9 p.u. the voltage levels at PCC are improved

using DVR with adaptive ANN controller more than both sta-tic and retuned PI controller, with adaptive ANN controllers,there is no need for retuning PI parameters.

Fig. 16 THD with DVR based


DVR based ANN could regulate the voltage at PCC withdiminished harmonics through faulty conditions as comparedto the PI-PSO as shown in Fig. 14. Fig. 15a to c show the volt-

age at F1, load and injected voltage waveforms during three-phase short circuit with ANN controller, which has a veryrapid response with smoothing waveforms than PI controller.

With DVR based PI-controller, the generator speed isincreased to values (10%), but with using ANN adaptive con-troller where, the generator rotor speed reaches to values (8%),

Fig. 16d. A comparison of Wind rotor speed without using aDVR, with using PI-PSO and with using adapted ANN isshown in Fig. 16.

5.7. Total harmonic distortions

The Total Harmonic Distortion (THD) is measured by usingFFT analysis. THD is calculated as, [17]

THD ¼ 1

V1

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiX1n¼2;3;��:

Vn2

vuut ð19Þ

where V1: is the main harmonic magnitude in RMS, nth har-monics magnitude. Fig. 16a and b show the total harmonic dis-tortions with DVR based PI-PSO (4.31%) and ANN (0.86%).

The simulation is carried using FFT analysis in MATLABsoftware.

(a) PI-PSO and (b) ANN.




6. Conclusions

The application of a DVR based ANN connected to hybridPV/Wind/FC supplying the electrical power to a new off-grid

community proposed in Egypt is investigated in this paper.ANN controller of DVR has the ability to overcome somefault events include, voltage sag/swell, unbalanced operation

and three-phase faults at PCC. The proposed controller suc-ceeded at keeping the three renewable sources (PV/Wind/fuelcell) in continues operation during these fault events withoutdisconnection from the system. The main idea behind this pro-

posed controller is to regulate the voltage between DVR andconsequently improve the voltage profile at PCC. A compar-ison between this proposed scheme and PI controller tuned

by PSO was introduced. The proposed controller showed abetter performance rather than static and retuned PI con-troller. Moreover, the proposed ANN controller of DVR

showed has a rapid response with minimum total harmonicdistortion level.

Declaration of Competing Interest

The authors declare that they have no known competing

financial interests or personal relationships that could haveappeared to influence the work reported in this paper.

Acknowledgment

This project was funded by the Deanship of Scientific Research(DSR), king Abdulaziz University, Jeddah, under grant

No. (DF-202-135-1441). The authors, therefore, gratefullyacknowledge DSR technical and financial support.

References

[1] M.O.A. El-Raouf, M.I. Mosaad, M.A. Al-Ahmar, F.M. El

Bendary, MPPT of hybrid solar-wind-grid power generation

system, Int. J. Ind. Electron. Driv. 2 (4) (2015) 234–241.

[2] Mohamed I. Mosaad, M. Osama Abed El-Raouf, Mahmoud A.

Al-Ahmar, Fahmy M. El Bendary, Optimal PI Controller of

DVR to Enhance the Performance of Hybrid Power System

Feeding a Remote Area in Egypt, Sustain. Cities Soc. 47(May)

(2019) 101469–101483.

[3] M. Osama abed el-Raouf, Mohamed I. Mosaad, Adel

Mallawany, Mahmoud A. Al-Ahmar and Fahmy M. El

Bendary, Optimal Control of DVR to Enhance The Power

Quality of Pv/Wind/Fuel Cell Hybrid System Feeding a New

Community, in: 25th International Conference on Electricity

Distribution CIRED, Madrid, 3–6 June 2019, pp. 599–603.

[4] A. Maleki, M. Rosen, F. Pourfayaz, Optimal Operation of a

Grid-Connected Hybrid Renewable Energy System for

Residential Applications, Sustainability 9 (8) (2017) 1314.

[5] O. Ayadi, R. Al-Assad, J. Al Asfar, Techno-economic

assessment of a grid connected photovoltaic system for the

University of Jordan, Sustain. Cities Soc. 39(May) (2018) 93–98.

[6] A. Maleki, M.G. Khajeh, M.A. Rosen, Two heuristic

approaches for the optimization of grid-connected hybrid

solar–hydrogen systems to supply residential thermal and

electrical loads, Sustain. Cities Soc. 34 (Oct.) (2017) 278–292.

[7] G. Zhang, B. Wu, A. Maleki, W. Zhang, Simulated annealing-

chaotic search algorithm-based optimization of reverse osmosis

hybrid desalination system driven by wind and solar energies,

Sol. Energy 173 (Oct.) (2018) 964–975.


[8] Shaimaa Barakat, M.M. Samy, Magdy B. Eteiba and Wael

Ismael Whaba, Optimization of an Off-Grid PV/Biomass

Hybrid System with Different Battery Technologies, Sustain.

Cities Soc. J. 40(July) (2018) 713–727.

[9] A.S. Anees, Grid integration of renewable energy sources:

Challenges, issues and possible solutions, in: 2012 IEEE 5th

India International Conference on Power Electronics (IICPE),

Delhi, 2012, pp. 1–6.

[10] S. Anand, B.G. Fernandes, J. Guerrero, Distributed Control to

Ensure Proportional Load Sharing and Improve Voltage

Regulation in Low-Voltage DC Micro grids, IEEE Trans.

Power Electron. 28 (4) (2013) 1900–1913.

[11] Mohamed I. Mosaad, Model reference Adaptive Control of

STATCOM for Grid-Integration of Wind Energy Systems, IET

Electr. Power Applic. J. 12 (5) (2018) 605–613.

[12] X. Wang, F. Blaabjerg, W. Wu, Modeling, and analysis of

harmonic stability in an AC power-electronics-based power

system, IEEE Trans. Power Electron. 29 (12) (2014) 6421–6432.

[13] SenerAgalar, Yusuf AlperKaplan, Power quality improvement

using STS and DVR in wind energy system, Renew. Energy 118

(April) (2018) 1031–1040.

[14] A Prajyusha, N Srinivas Rao, B. Naresh, Development of ultra-

capacitor-based DVR for power quality improvement, Int. J.

Innov. Technol. Res. 6(1) (2018).

[15] Rakeshwri Pal, Dr. Sushma Gupta, Recapitulation of Various

Topologies and Control Strategies Implicated in Dynamic

Voltage Restorer (DVR) for Power Quality Improvement, Int.

J. Appl. Eng. Res. 13(5) (2018) 3057–3073 (ISSN 0973-4562).

[16] W.J. Praslin, J. Belwin Edward, A Review on Impacts of Power

Quality, Control and Optimization Strategies of Integration of

Renewable Energy Based Microgrid Operation, I.J. Intell. Syst.

Appl. 3 (2018) 67–81.

[17] Mohamed I. Mosaad, H.S. Ramadan, Power Quality

Enhancement of Grid Connected Fuel Cell Using

Evolutionary, Comput. Tech. Int. J. Hydrogen 43 (25) (2018)

11568–11582.

[18] Mohamed I. Mosaad, Nagy I. Elkalashy, Mohamed G.

Ashmawy, Integrating Adaptive Control of Renewable

Distributed Switched Reluctance Generation and Feeder

Protection Coordination, Electric Power Syst. Res. J. 154

(January) (2018) 452–462.

[19] S. Abouelfadl, El-Lithy KH Impact Assessment of Global

warming on Egypt, Int. J. Eng. Advtechnol. 3 (2014) 352–360.

[20] S. Abouelfadl, K.H. Ouda, A. Atia, N. Al-Amir, The Egyptian

Gardens’ City- A New City in the Western Desert (EGCWD) -A

primary master plan, International Conference on Energy and

Sustainability, NED University of Engineering and Technology,

Pakistan, 2013.

[21] Mohamed I. Mosaad ,M. Osama abed el-Raouf, Mahmoud A.

Al-Ahmar, Fahd A. Banakher, Maximum Power Point Tracking

of PV system Based Cuckoo Search Algorithm; review and

comparison, Energy Proc. 162(April) (2019) 117–126.

[22] Mohamed I. Mosaad, Fawzan Salem, Adaptive Voltage

Regulation of Self Excited Induction Generator Using FACTS

Controllers, Int. J. Ind. Electron. Driv. 1 (4) (2014) 219–226.

[23] Mohammed I. Mosaad, Ahmed Abu-Siada, Mohamed

Elnaggar, Application of Superconductors to Improve the

Performance of DFIG-based WECS, IEEE Access J. 7(1)

(2019) 103760–103769.

[24] Mohamed I. Mosaad, Ahmed Alenany, A. Abu-Siada,

Enhancing the Performance of Wind Energy Conversion

Systems using Unified Power Flow Controller, IET Gen.

Trans. Distrib. J. (2020), https://doi.org/10.1049/iet-

gtd.2019.1112.

[25] Mohamed I. Mosaad, Direct Power Control of Switched

Reluctance Generator based- Wind Energy Conversion

Systems Using Optimized Fractional Order PI Controller, IET

Electric Power Appl. https://doi.org/10.1049/iet-epa.2019.0194.


http://refhub.elsevier.com/S1110-0168(20)30139-3/h0005
















































https://doi.org/10.1049/iet-gtd.2019.1112

https://doi.org/10.1049/iet-gtd.2019.1112

https://doi.org/10.1049/iet-epa.2019.0194



[26] M. Osama abed el-Raouf, Mohamed I. Mosaad, Adel

Mallawany, Mahmoud A. Al-Ahmar and Fahmy M. El

Bendary, MPPT of PV-Wind-Fuel cell of off-grid Hybrid

System for a New Community, in: Twentieth International

Middle East Power Systems Conference (MEPCON), 2018, pp.

480–487.

[27] M.M. Samy, S. Barakat, H.S. Ramadan, A Flower Pollination

Optimization Algorithm for an Off-Grid PV-Fuel Cell Hybrid

Renewable System, Int. J. Hydrogen Energy (IJHE) 44 (4) (2019)

2141–2152, https://doi.org/10.1016/j.ijhydene.2018.05.127.

[28] HOMER (Hybrid Optimization of Multiple Energy Resources).

Available at: https://www.homerenergy.com/homer-pro.html.


[29] M.M. Samy, N. Alshammari, J. Asumadu, Optimal Economic

Analysis Study for Renewable Energy Systems to Electrify

Remote Region in Kingdom of Saudi Arabia, in: The Twentieth

International Middle East Power Systems Conference

(MEPCON), Cario University, Egypt, December 18–20, 2018,

pp. 1040–1045.

[30] The Egyptian Code for the foundations of design and conditions

for the implementation of electrical connections and

installations in buildings, vol. 1, 2, 3, 1994.

[31] Nordel connection code wind turbines, 15 January 2007.


https://doi.org/10.1016/j.ijhydene.2018.05.127

https://www.homerenergy.com/homer-pro.html


Performance improvement of off-grid hybrid renewable energy …€¦ · Performance improvement of oﬀ-grid hybrid renewable energy system using dynamic voltage restorer Wael S.

Documents