protection and control of modern power systems ......hence, the modern power systems become more...

ORIGINAL RESEARCH Open Access

Optimized coordinated control of LFC andSMES to enhance frequency stability of areal multi-source power system consideringhigh renewable energy penetrationGaber Magdy1,2* , G. Shabib2,3, Adel A. Elbaset4 and Yasunori Mitani1

Abstract

With rapidly growing of Renewable Energy Sources (RESs) in renewable power systems, several disturbancesinfluence on the power systems such as; lack of system inertia that results from replacing the synchronousgenerators with RESs and frequency/voltage fluctuations that resulting from the intermittent nature of the RESs.Hence, the modern power systems become more susceptible to the system instability than conventional powersystems. Therefore, in this study, a new application of Superconducting Magnetic Energy Storage (SMES) (i.e.,auxiliary Load Frequency Control (LFC)) has been integrated with the secondary frequency control (i.e., LFC) forfrequency stability enhancement of the Egyptian Power System (EPS) due to high RESs penetration. Where, thecoordinated control strategy is based on the PI controller that is optimally designed by the Particle SwarmOptimization (PSO) algorithm to minimize the frequency deviations of the EPS. The EPS includes bothconventional generation units (i.e., non-reheat, reheat and hydraulic power plants) with inherent nonlinearities,and RESs (i.e., wind and solar energy). System modelling and simulation results are carried out using Matlab/Simulink® software. The simulation results reveal the robustness of the proposed coordinated control strategyto preserve the system stability of the EPS with high penetration of RESs for different contingencies.

Keywords: Renewable energy sources (RESs), Load frequency control (LFC), Egyptian power system (EPS),Superconducting magnetic energy storage (SMES)

1 IntroductionDue to the growing demand on utilizing RenewableEnergy Sources (RESs) as a future solution for en-ergy shortages, many conventional generation unitsare being replaced by the RESs that have several im-pacts on the performance of the renewable powersystems such as lack of system inertia. The RESs ex-change electrical power with the power systemsthrough power electronic devices (i.e., inverters andconverters), which reduce the overall system inertia.

Consequently, the inverter-based RESs will causehigh frequency/voltage fluctuations compared to theconventional generation units [1]. Moreover, the ir-regular nature of the RESs and random load devia-tions can cause severe power generation fluctuations.Therefore, the frequency control becomes more diffi-cult in case of any mismatch between the powergeneration and the load demand, particularly, withhigh-level RESs (e.g., wind and solar energy) pene-tration into the power systems. Hence, the Load

* Correspondence: [email protected] of Electrical Engineering, Kyushu Institute of Technology,Tobata-Ku, Kitakyushu-shi, Fukuoka 804-8550, Japan2Department of Electrical Engineering, Faculty of Energy Engineering, AswanUniversity, Aswan 81528, EgyptFull list of author information is available at the end of the article

Protection and Control ofModern Power Systems

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made.

Magdy et al. Protection and Control of Modern Power Systems (2018) 3:39 https://doi.org/10.1186/s41601-018-0112-2

http://crossmark.crossref.org/dialog/?doi=10.1186/s41601-018-0112-2&domain=pdf

http://orcid.org/0000-0001-6894-695X

mailto:[email protected]

http://creativecommons.org/licenses/by/4.0/

Frequency Control (LFC) is considered one of themost important issues in power systems to maintainthe system frequency and the power variations attheir standard values. Whereas system frequency de-pends on active power and the system voltagegreatly depends on the reactive power. Therefore,the control of power systems can be classified intotwo fundamental issues. a) control of the activepower along with the frequency, b) control of the re-active power along with the voltage regulation [2].To overcome the frequency instability problem, nu-

merous control techniques for power system fre-quency control have been implemented such as fuzzylogic controller [3], artificial neural network (ANN)[4], linear quadratic regulator (LQR) controller [5]and robust controller-based H_infinte [6]. Althoughthe control strategies [3–6] gave a good dynamic re-sponse, they are a dependency on the designer’s ex-perience and need long computational time. On theother hand, real-world LFC is performed based onProportional-Integral-Derivative (PID) or PI control-lers because of it has many merits such as; lowercost, simple structure, robustness performance and asuccessful practical controller that can provideexcellent control performance regardless of theperturbations and variations in the system parameters[7]. However, these controllers suffer from a complicated process of parameters tuning based on trial anderror method. In such a case, the robustness of thesystem is not guaranteed against further perturbationsin the system parameters. Therefore, several optimization algorithms were used to find the optimal pa-rameters of the PI or PID controllers in the LFC loopsuch as particle swarm optimization (PSO) [8],cuckoo optimization algorithm [9], ant colony algo-rithm [10], quasi-oppositional harmony search algo-rithm (QOHSA) [11] etc.According to the previous studies, most research has

focused on the frequency stability analysis in powersystems that are modelled as thermal power plants (i.e.,non-reheat and reheat power plants) or/and hydraulicpower plants depending on the number of areas.However, most of the existing realistic power systemscomprise multi-source dynamic generators (e.g., thermal,hydraulic and gas power plants). Therefore, several typesof power plants should be added to the LFC problem toachieve a realistic study as reported in this research. Fur-thermore, most of the studied power systems are linearand have a simple structure, where it mainly depends onthe conventional generation units. However, severalRESs should be integrated into the power systems toachieve more realistic study for the power systems oftoday. Therefore, recently, a few research and studies onthe integration of several RESs into power systems have

been conducted in [12–15]. However, based on the pre-vious researches, the effects of a high penetration levelof the RESs have not been considered for frequency sta-bility analysis. Hence, several types of RESs with highpenetration levels should be added in the analysis of theLFC issue for achievement more accurate studies fortoday’s power systems.With increasing the utilizing of RESs into the

modern power systems, it becomes much importantto look at methods and techniques to store this en-ergy. Where, there are several Energy Storage Sys-tems (ESSs) such as Superconducting MagneticEnergy Storage (SMES), electric batteries, fuel cells,and others, which have been considered within thedesign of the modern power systems [16]. There-fore, the ESSs can be used for storing the excessenergy from the RESs, as well as, discharging thestored energy to the grid as needed, depending ondemand. Among many ESSs, SMES is most suitedfor improved frequency stability in power systems,due to its outstanding advantages such as fast re-sponse, high efficiency, and long lifetime [17].Therefore, a few research and studies on SMES ap-plications for power systems have been conductedin [18–21]. According to the aforementioned refer-ences, there is no report for SMES system toanalyze the frequency stability for a multi-sourcepower system during high-level RESs penetrationand contingencies. With increasing of penetrationlevel of the RESs into the power systems, it will becaused higher frequency deviations and the LFCmay be failing to maintain the system frequency.Therefore, from the perspective of the LFC, theSMES can be used as a feedback controller in theaim of supporting the frequency control loops (i.e.,primary and secondary frequency controls) for fre-quency stability enhancement of the modern powersystem as reported in this research.Based on the above analysis, this research proposes a

coordinated control strategy between the secondary fre-quency control (i.e., LFC) and SMES unit (i.e., auxiliaryLFC) for frequency stability enhancement of the EPSwith high-level RESs penetration. Therefore, the maincontribution of this work includes the following aspects.(i) this paper presents a real hybrid power system inEgypt that includes both conventional generationsources (i.e., steam, gas and hydraulic power plants) withinherent nonlinearities, and RESs (i.e., wind and solarenergy) for studying the frequency stability analysis ofsuch systems. Where, the conventional generation unitsin the EPS is decomposed into three dynamic subsys-tems; non-reheat, reheat and hydraulic power plants.Moreover, the physical constraints of the governors andturbines such as Generation Rate Constraints (GRCs) of

Magdy et al. Protection and Control of Modern Power Systems (2018) 3:39 of 15

power plants and speed governor deadband (GDB) aretaken into consideration. (ii) this paper proposes a newapplication of SMES unit as a feedback controller in theaim of supporting the frequency control loops (i.e., pri-mary and secondary frequency controls) for frequencystability enhancement of the studied power system.Where the dynamic structure of the studied SMESmodel in the previous researches [2, 18, 20] is too simpleand it is preferable to use a more realistic model. More-over, the inductor current of the studied SMES modelslowly returns to its nominal value after a system strike.However, the induction current of the SMES modelmust be quickly restored to its nominal value after a dis-turbance in the system so that it can respond to the nextload disorder immediately. Therefore, restoring theinductor current to its nominal value can be enhancedby using the inductor current deviation as a negativefeedback signal in the model of SMES control loop asreported in this study. (iii) the proposed coordinatedcontrol strategy of LFC and SMES unit is based on thePI controller that is optimally designed by the PSO algo-rithm to minimize the frequency deviations of the EPS.Moreover, the proposed coordinated can improve andmaintain the frequency stability of the EPS and mainlywhen the RESs are highly penetrated at partial load.According to [22], the renewable power systems withthe proposed coordinated control strategy will providebetter stability and performance for the power systemsof today, and for those of the future that is expected tointegrate more and more RESs; thus, the proposed strat-egy will ensure an avoidance of system instability andsystem collapse.The rest of this paper is structured as follows: Section 2

presents the system configuration including the EPS mod-elling and SMES technology modelling. The state-spacedynamic modelling of the EPS considering RESs andSMES is described in section 3. Section 4 presents thecontrol strategy and problem formulation. Section 5 de-scribes the PSO algorithm. The simulation results and dis-cussion are provided in section 6. Finally, the conclusionis presented in Section 7.

2 System configuration2.1 Modelling of the EPSThe presented power system in this study is a realmulti-source power system in Egypt, which has 180 powerplants. Where, the power plants can be classified into 3categories; a) Non-reheat power plants are represented bygas-turbine power plants and a few numbers of steampower plants. b) Reheat power plants are mainly repre-sented by thermal power plants or combined cycle powerplants. c) Hydraulic power plants such as High Dam inAswan city. Recently, the EPS included several RESs such

as wind and solar energy (e.g., photovoltaics (PV) solarpower, and Concentrated Solar Power (CSP)). These RESscontribute almost 3% of the installed capacity. However,the Egyptian Electricity Holding Company (EEHC) aimsto increase the electric energy from RESs to cover 20% ofthe electric energy demand by the year 2020 [23]. Accord-ing to the last report of the EEHC in 2016, The total gen-eration capacity and peak loads are 38,000MW and29,000MW respectively [23]. Motivated by the aforemen-tioned observation, this paper focuses on modernizationthe EPS via integrating high-level RESs penetration, andSMES technology for facing the future challenges, whichare expected to integrate more and more RESs. There-fore, the RESs includes wind power with a peakpower of 5000 MW and solar irradiation power witha peak power of 2300 MW. The base of the systemfrequency is 50 Hz, while the power base is 38,000MW. The EPS has been simulated and tested in thisresearch to illustrate the proposed coordinated con-trol strategy. The Egyptian grid studied powersystem case 9-machines and 32-bus system [24].The single line diagram of the studied power system(i.e., The EPS) is shown in Fig. 1. Moreover, thesimplified model of the EPS incorporation withRESs and SMES system is shown in Fig. 2. TheNational Energy Control Center (NECC) in Egypthas been advanced a dynamic model of EgyptianLFC in [25]. Therefore, this model will be modifiedusing MATLAB/Simulink by adding high-level RESsintegration and the dynamic contribution of thecontrollable SMES, which represents an auxiliaryLFC alongside with the frequency control loops (i.e.,primary and secondary frequency controls) for fre-quency stability analysis. Figure 3 shows the dy-namic model of the EPS considering RESs with theproposed coordinated control strategy. The NECCin Egypt estimates the system parameters values,which used in the dynamic model of the EPS as in-dicated in Table 1.The important inherent nonlinearities requirements

and the physical constraints enjoined by the system dy-namics of the generation units are taken into consider-ation to obtain an accurate perception for the EPS. Oneof the most important constraints of power plants is therate of generation power change because of the limita-tion of mechanical movements. The physical system dy-namics of power plants is represented by GRC and themaximum/minimum limit of the valve gate (i.e., gov-ernor deadband (GDB)). The GRC limits the generationrate of output power, which is given as 20% pu. MW/mi-nute, and 10% pu MW/minute for non-reheat and re-heat turbines, respectively. However, the actual GRC ofthe hydraulic power plant is about 50% pu MW/minute,which is higher than the generation rate corresponding


to any practical disturbance and hence it will beneglected [22]. On the other side, the GDB restricts thevalve opening/closing. Where the GDB of thenon-reheat and reheat power plants equal ±0.05, whilethe GDB of the hydraulic power plant is ±0.01. In thispaper, the RESs have low-order dynamic models, whichare considered sufficient for frequency stability analysisas reported in [22, 26]. Therefore, the power variations

of RESs; the wind power variation (ΔPWT) and the PVsolar power variation (ΔPPV), and the load power vari-ation (ΔPL) are considered as disturbance signals to theEPS.

2.2 Frequency control based on SMES systemRESs such as wind, solar, waves, and tides are rapidlygrowing into today's power system. Where, the RESs

Fig. 2 A simplified model of the EPS with the proposed coordinated control strategy

Fig. 1 A typical single-line diagram of the Egyptian Power System [24]


exchange power to the power systems through powerelectronic devices that cause to reduce the system inertiaand increase frequency/voltage fluctuations compared tothe conventional generation units. Moreover, the inter-mittent nature of the RESs due to their outputs aredependent on weather will lead to more negative effectson system stability and the LFC may fail to readjust sys-tem frequency. These negative effects can limit highpenetration of the RESs. Therefore, to overcome thisproblem, this paper proposes a new application of SMESsystem that uses as an auxiliary LFC incorporated withthe primary and secondary frequency controls to en-hance the frequency stability of the EPS with high RESspenetration.The SMES technology is one of the most important

energy storage devices, which is considered the most ap-propriate for frequency stability enhancement due to it

has many merits such as fast response, high efficiency,and long lifetime compared to other energy storage de-vices [17]. The SMES system storages the power in themagnetic coil, which made from a superconducting ma-terial with nearly zero loss of energy [19]. Where, theSMES unit comprises a dc superconducting magneticcoil, which is included in helium container, power elec-tronic devices (i.e., inverter/converter) that are used toconnect the dc magnetic coil to the ac power system,and Υ-Δ/Υ-Υ transformer as shown in Fig. 4. The con-trol of the firing angle (α) of the converter provides DCvoltage that appears through an inductor (Ed) to con-tinuously change within a given range of positive andnegative values. The inductor is at first charged to itsrated current Id0 by applying a little positive voltage.Once the current reaches the rated value, it is kept upsteady by diminishing the voltage over the inductor tozero since the coil is superconducting [27]. Neglectingthe losses of the transformer and the converter, the DCvoltage is given by

Ed ¼ 2Vd0 cosα−2IdRC ð1Þ

The charging and discharging processes of the SMESunit can be controlled through variation of the commu-tation angle (α). The power conversion system acts inthe converter mode (i.e., charging mode) When α < 90°but the power conversion system acts in the invertermode (i.e., discharging mode) when α > 90°. In this

Fig. 3 A dynamic model of the EPS considering RESs with the proposed coordinated control strategy

Table 1 System parameters of the studied power system

Parameter Value Parameter Value

D 0.028 R2 2.5

T1 0.4 R3 1.0

T2 0.4 H 5.7096

T3 90 Pn1 0.2529

Td 5 Pn2 0.6107

Th 6 Pn3 0.1364

Tw 1.0 R1 2.5

m 0.5 f 50


study, the SMES system is used as frequency stabilizer(i.e., auxiliary LFC), which can charge and discharge elec-trical power from/to the grid with very short time consid-ering the SMES power limits. Therefore, the DC voltageof the superconducting inductor (Ed) is continuously con-trolled through the input signal of the SMES unit (UPI).Moreover, the current of the superconducting inductormust be quickly restored to its nominal value after a dis-turbance in the system so that it can respond to the nextload disorder immediately. Consequently, the current de-viation of the inductor can be detected and utilized as anegative feedback signal in the SMES control loop so thatthe current recovery can be enhanced to its nominal value[19]. Thus, the gradual change in the voltage applied tothe inductor (ΔEd) and the inductor current deviation(ΔId) can be defined as follows:

ΔEd ¼ 11þ sTc

ΔUPI−K fΔId� � ð2Þ

ΔId ¼ ΔEd

sLð3Þ

The block diagram of the SMES control loop with thenegative feedback of the inductor current deviation is in-cluded in the model of the EPS as shown in Fig. 3.Moreover, the SMES control loop parameters are givenin Appendix [19]. Therefore, in this study, the designedSMES unit gives a good dynamic stability during thetransients even when the system parameters are changedby ±50% of their nominal values. Where the activepower deviation of the SMES unit can be defined as:

ΔPSMES ¼ ΔEd ΔId þ Id0ð Þ ð4ÞThe stored energy E (Joule) in the superconducting

coil and its rated power P (Watt) are described by thefollowing equations:

E ¼ 12LIdo

2 ð5Þ

P ¼ dEdt

¼ LIdodIdodt

ð6Þ

3 State-space dynamic modellingThe studied power system is considered ninth order lin-earized multi-source power system considering RESsand SMES system. The frequency deviation of the stud-ied power system considering the effect of the primarycontrol loop (i.e. governor action), a secondary controlloop (i.e. LFC), and SMES controller (i.e. an auxiliaryLFC) can be obtained as:

Δ f ¼ 12Hsþ D

ðΔPm1 þ ΔPm2 þ ΔPm3 þ ΔPWT þ ΔPPV

� ΔPSMES−ΔPLÞð7Þ

Where

ΔPm1 ¼ Pn1

T1S þ 1� −1

R1� Δ f −ΔPc

� �ð8Þ

ΔPg2 ¼ Pn2

T 2S þ 1� −1

R2� Δ f −ΔPc

� �ð9Þ

ΔPm2 ¼ mþ mThS þ 1

� �� ΔPg2 ð10Þ

ΔPg3 ¼ Pn3TdS þ Pn3

T 3S þ 1� −1

R3� Δ f −ΔPc

� �ð11Þ

ΔPm3 ¼ −TwS þ 10:5 � TwS þ 1

� �� ΔPg3 ð12Þ

ΔPWT ¼ 1TWTS þ 1

� ΔPWindð Þ ð13Þ

ΔPPV ¼ 1TPVS þ 1

� ΔPSolarð Þ ð14Þ

ΔPSMES ¼ ΔEd ΔId þ Id0ð Þ ð15ÞUsing Eqs. (7)–(15) and the dynamic model of the

studied power system with the proposed coordinatedcontrol strategy as shown in Fig. 3, the dynamic equa-tions of the studied hybrid power system can be derivedand written in the state variable form as follows:

_X ¼ AX þ BU þ EW ð16Þ

Y ¼ CX þ DU þ FW ð17Þ

Fig. 4 The schematic diagram of the SMES unit


Hence, the complete state-space equations for the EPSconsidering RESs with the coordinated control strategycan be obtained as in (18).

4 Control strategy and problem formulationThe PID Controller has three terms functionality (i.e., P,I and D controllers) covering treatment to both transientand steady-state responses. However, the PID and P con-trollers cannot yield sufficient control performance withthe consideration of nonlinearities and boiler dynamics[28]. To overcome this problem, the PI controller has

been employed for system control. Therefore, the pro-posed coordinated control strategy is based on the PIcontroller in the EPS considering high-level RESs pene-tration and inherent nonlinearities. The PI controller hasbeen validated to be remarkably effective in the regulat-ing of a wide range of processes [7, 28]. However, the PIcontroller suffers from a complicated process of parame-ters tuning-based trial and error method. Therefore, thisresearch uses the PSO algorithm to find the optimumparameters of the PI controller for minimizing the sys-tem frequency deviation. Where, the PSO algorithm has

_X ¼

−D2H

12H

12H

012H

012H

12H

12H

−a1 −1T1

0 0 0 0 0 0 0

−ma2 0 −1Th

b1 0 0 0 0 0

−a2 0 0 −1T 2

0 0 0 0 0

−2b2Dþ 2a3 þ 2b3ð Þ 2b2 2b2 0 2b2−2Tw

� �2Tw

þ 2T3

� �0 0 0

b2D−a3−b3ð Þ −b2 −b2 0 −b2 −1T3

0 0 0

0 0 0 0 0 0 −1

TWT0 0

0 0 0 0 0 0 0 −1

TPV0

1TC

0 0 0 0 0 0 0 −1TC

266666666666666666666666666664

377777777777777777777777777775

Δ fΔPm1

Δpm2Δpg2ΔPm3

ΔPg3

ΔPWT

ΔPPV

ΔEd

26666666666664

37777777777775

þ

0 0 −12H

0 0 00 0 00 0 00 0 −2b20 0 b21

TWT0 0

01

TPV0

0 0 0

266666666666666664

377777777777777775

ΔPWind

ΔPSolar

ΔPL

24

35þ

0 0

−Pn1

T 10

−m�Pn2

T 20

−Pn2

T 20

2�Pn3

T30

−Pn3

T 30

0 00 0

0 −K f

TC

26666666666666666666664

37777777777777777777775

ΔPC

ΔId

� �

Y ¼ 1 0 0 0 0 0 0 0 0½ �

Δ fΔPm1

Δpm2Δpg2ΔPm3

ΔPg3

ΔPWT

ΔPPV

ΔEd

26666666666664

37777777777775

þ 0 0 0½ �ΔPWind

ΔPSolar

ΔPL

24

35þ 0 0½ � ΔPC

ΔId

� �

ð18Þ


many merits such as the ease of use, high convergencerates, minimum storage requirements, and less depend-ing on the set of initial values, implying the robustnesscompared to other methods (i.e., genetic algorithm, arti-ficial neural networks, fuzzy logic, and ant colony) [29].Considering these advantages, this paper uses the PSOalgorithm to tune the PI controller parameters, obtain-ing the optimum PI controller parameters with the ro-bustness of operations. In this study, the integral ofsquared-error (ISE) is used as a fitness function that isThe constants of matrices are:a1 ¼ Pn1

T1R1, a2 ¼ Pn2

T 2R2, a3 ¼ Pn3

T 3R3, a4 ¼ Td

2H , b1 ¼ 2mTh

− mT2,

b2 = a3a4, b3 ¼ P3TdT 3:

the objective function of the proposed optimizationtechnique and can be formulated as follows:

ISE ¼Z tsim

0Δ fð Þ2dt ð19Þ

Subject to bounds of the PI controller parameters asfollows:

½KMinp;i ≤Kp;i≥KMax

p;i �

where (Δf ) is the frequency deviation of the EPS and tsimis the simulation time to execute one run. The PSO al-gorithm is applied in the EPS to obtain the minimumvalue of the objective function (i.e., system frequency de-viation) through getting on the optimal parameters ofthe PI controller.

5 Optimal PI controller design based on PSOalgorithm5.1 Overview of particle swarm optimizationThe PSO is a global optimization algorithm based onevolutionary computation technique. It was presentedby Kennedy and Eberhart in 1995 [30]. The basic op-eration principle of this optimization technique is de-veloped on a swarm of birds flocking. The birds areeither dispersed or go together from one place to an-other for searching their food. Furthermore, one ofthem can discover the place where the food can befound due to the transmitting information to otherbirds at any time while searching the food [31]. Inthe PSO algorithm, instead of using evolutionary op-erators, individuals called particle are used. Therefore,a swarm consist of several particles, each particle rep-resents a potential for the problem. Each particle inthe PSO algorithm flies in the search space accordingto its own flying experience and its companion flyingexperience. Each one of particles is treated as a par-ticle in D-dimension search space. The position ofparticle represented as Xi, the best previous positionof any particle is recorded and called Pbest. Anotherbest value that is tracked by a global version of thePSO (i.e., the overall best value gbest) [32]. The

velocity of particle i is represented by Vi and all parti-cles are updated according to the following equations:

vnþ1id ¼ w:Vn

id þ c1:randðÞ:ðPnid−X

nidÞ

þ c2:randðÞ:ðPngd−X

nidÞ ð20Þ

Fig. 5 Flowchart of PSO-PI controller

Table 2 The control parameters of PSO

Parameter Value

Size of Swarm, S 50

Number of iterations, n 50

Inertia weight factor, w 0.8

Acceleration constant 1, C1 0.12

Acceleration constant 2, C2 2


xnþ1id ¼ xnid þ vnþ1

id ð21Þ

These equations are used to calculate the newvalues of velocity and position of each particle ac-cording to its previous values. Learning factors of theoptimization technique have significant implicationson the algorithm convergence rate. Further informa-tion for the PSO can be found in [29–33].

5.2 Implementation of a PSO-PI controllerThis study uses the PSO algorithm to tune the PIcontroller gains (Kp and Ki) in the model of the EPSconsidering high penetration of RESs. Where, eachparticle in the search space introduces a probable so-lution for the PI gains, which are a 2-dimensionalproblem. The performance of the probable solutionpoint is determined by the fitness function as seen inEq. (19). Moreover, the size of the swarm determinesthe requirements of global optimization and computa-tion time. Therefore, the steps of the PSO algorithmfor optimum PI controller in the EPS are illustrated

in Fig. 5. The performance of the PSO algorithm insearching the PI controller parameters of the coordi-nated control strategy in the EPS has been validatedby using the characteristics of the PSO as given inTable 2. Where these optimal characteristics are con-sidered enough after many trials. Therefore, the opti-mal parameters of the PI controller-based PSOalgorithm under the system operation condition with-out RESs are {Kp = 7.48710, and Ki = 1.171355}, whichproduce the optimal control signal to the coordinatedcontrol strategy for frequency stability enhancementof the EPS when high RESs penetration. Where, theparameters of the PI controller lie in the range [0,10], which are considered the most common valuesfor LFC in the industry [34].

6 Simulation results and discussionsIn this study, the coordinated control strategy be-tween the secondary frequency control loop (i.e.,LFC) and the SMES system (i.e., auxiliary LFC) isproposed for frequency stability enhancement of theEPS mainly when the RESs are highly penetrated.

Fig. 7 Conventional generation and SMES power responsesof the EPS for scenario 1

Fig. 6 The frequency deviation of the EPS for scenario 1

Table 3 The performance specification of the EPS for scenario 1

Scenario 1 With SMES-Basedoptimal PI controller

With SMES Without SMES

MUS (pu) 1.911 × 10−3 8.401 × 10− 3 1.523 × 10− 2

MOS (pu) 0.0 6.440 × 10−5 1.420 × 10− 3

TS (s) 17.297 26.703 27.421

Fig. 8 The frequency deviation of the EPS for scenario 2: (a)+ 50% system parameters variations, (b) − 50% systemparameters variations


Where, the proposed coordinated control strategy isbased on the PI controller, which is optimallydesigned by the PSO algorithm to obtain the mini-mum value of the EPS frequency deviations. More-over, the performance of the proposed coordinatedcontrol strategy is compared with both; the optimalLFC with/without SMES system under high-levelRESs penetration and system parameters variations(i.e., system uncertainties). The simulation results ofthe studied power system are carried out usingMATLAB/Simulink® software to validate the effectiveness of the proposed coordinated control strategy.The code of the PSO as an m-file is interfaced withthe model of the EPS to execute the optimizationprocess. The EPS frequency stability with the pro-posed coordinated control strategy is investigatedunder different operating conditions through the fol-lowing scenarios:

6.1 System performance evaluation without the RESsThe model of the studied power system (i.e., the EPS)without the RESs is considered as the test system tovalidate the effectiveness of the proposed coordinatedcontrol strategy for system frequency stability. Theproposed coordinated control strategy using the de-signed PI controller-based PSO algorithm is tested bya sudden load change, which is represented as a stepload perturbation (SLP) of 10% pu at time t = 200 s.Therefore, the sudden change in the load demand im-pacts on the EPS frequency stability can be obviouslyseen from these test scenarios:

Scenario 1: in this scenario, the studied powersystem (i.e., the EPS) is assumed to have thedefault parameters with 100% of default systeminertia as indicated in Table 1. Figure 6 shows theEPS frequency deviations with the studied threecontrol strategies. From Fig. 6, It is clear that theSMES controller can improve the frequencyresponse and gives a better damped than the EPSwithout SMES. Compared to the EPS with/withoutthe SMES controller, the proposed coordinatedcontrol strategy-based the optimal PI controller canprovide a smooth and secure frequency perform-ance. Therefore, the frequency response of the EPSis improved by using the proposed coordinatedcontrol strategy-based the optimal PI controller.Fig. 7 shows that the proposed coordinated controlstrategy success for decreasing the required power fromthe conventional generation units (i.e., non-reheat, reheatand hydraulic power plants) during the sudden loadchange at timet = 200 s. Hence, the SMES power is greatly dischargedby the proposed control strategy. The performancespecifications; maximum overshoot (MOS), maximumundershoot (MUS), and maximum settling time (TS) of theEPS for this scenario have been compared in Table 3.Scenario 2: in this scenario, the dynamicperformance of the EPS with the proposedcoordinated control strategy is investigated undersystem parameters variations (i.e., systemuncertainties). The variable parameters are T1, T2,T3, Th, Td, Tw, m, R1, R2, R3, H, and D, which are


Scenario2

With SMES-based optimal PI controller With SMES Without SMES

MUS (pu) MOS (pu) TS (s) MUS (pu) MOS (pu) TS (s) MUS (pu) MOS (pu) TS (s)

+ 50% 1.857 × 10−3 0.0 15.848 7.548 × 10− 3 6.150 × 10−5 23.424 1.266 × 10− 2 2.209 × 10− 3 26.723

-50% 1.982 × 10− 3 0.0 17.935 9.692 × 10− 3 1.249 × 10− 4 38.820 2.237 × 10− 2 2.042 × 10− 3 41.019

Fig. 9 Conventional generation and SMES power responses of the EPS for scenario 2: (a) + 50% system parameters variations, (b)− 50% system parameters variations


changed by ±50% of their nominal value. Figure 8shows the frequency deviations of the studiedpower system with the proposed three controlschemes under these conditions. It can be clearfrom these results the proposed coordinated controlstrategy can effectively regulate the systemfrequency and guarantee robust performance againstsystem uncertainties. Furthermore, the settling timehas lower values using the proposed controlstrategy than that by using other control strategies.Hence, the designed PI controller for the proposedcoordinated control strategy is a robust controllerwhere it does not need to re-tuning its parameters

to deal with system uncertainties. The transientspecification of the EPS like MOS, MUS, and TS isindicated in Table 4. Figure 9 shows that theconventional generation units significantly generatedthe needed power in cases of with/without SMEScontroller, while the proposed coordinated controlstrategy could significantly reduce the neededpower from the conventional generation unitsduring the sudden load change at time t = 200 s.Moreover, the SMES power is greatly delivered bythe proposed control strategy-based the optimal PIcontroller.

6.2 System performance evaluation with the RESsThe EPS considering high RESs penetration asshown in Fig. 3 is considered as the test system toconfirm the robustness and effectiveness of the pro-posed coordinated control strategy. The proposedcoordinated control strategy of LFC and SMESsystem (i.e., auxiliary LFC) using the designed PIcontroller-based PSO algorithm is tested byimplementation; high fluctuated wind power at timet = 1000 s, low fluctuated solar irradiation power attime t = 0 s and a sudden load change with 10% puat time t = 200 s. Figure 10 shows the power vari-ation of wind and solar power generations. There-fore, the EPS frequency stability with the proposed

Fig. 10 Power variation of wind and solar power generations

Fig. 11 The frequency deviation of the EPS for scenario 3


Scenario 3 With SMES-Basedoptimal PI controller

With SMES Without SMES

MUS (pu) 1.933 × 10− 3 8.358 × 10− 3 1.508 × 10− 2

MOS (pu) 1.908 × 10− 3 8.005 × 10− 3 1.469 × 10− 2

TS (s) 19.432 26.288 27.001

Fig. 12 Conventional generation and SMES power responsesof the EPS for scenario 3


coordinated control strategy is investigated underdifferent operating conditions through the followingscenarios.

Scenario 3: in this scenario, the EPS consideringhigh-level RESs penetration as seen in Fig. 3 isassumed to have the default parameters with 100%of default system inertia as indicated in Table 1.Figure 11 shows that the EPS frequency responseis affected by the sudden load change and theRESs fluctuations. From Fig. 11, it can be notedthat the system response using the proposed

coordinated control strategy is faster, has a lowersteady-state error and better damped than otherscontrol strategies. In addition, the numericalresults of the transient specification (i.e., MOS,MUS, and TS) for the three control strategiesunder these conditions are within acceptableranges as indicated in Table 5. Hence, there is noneed to redesign the designed optimal PIcontroller. On the other hand, in cases of the EPSwith/without SMES controller, the conventionalgeneration units significantly generated the neededpower during the sudden load change at timet = 200 s, while the proposed coordinated controlstrategy could significantly reduce the neededpower from the conventional generation units asshown in Fig. 12. Moreover, the SMES power isgreatly charged/discharged by the proposedcoordinated control strategy from/to the EPS,according to the EPS needing.Scenario 4: the main target of this scenario is toinvestigate the performance of the EPS with theproposed coordinated control strategy under variationin system parameters (i.e., system uncertainties). Thedefault system parameters as indicated in Table 1 arechanged by ±50% of their nominal values. Figures 13and 14 show the frequency deviations of the EPS withthe three control strategies under these conditions.From these figures, it can be concluded that theproposed coordinated control strategy could addressthe system uncertainties and frequency deviation isquickly driven to zero. Where the proposed controlstrategy gives a little transient compared with the othercontrol strategies. Furthermore, the numerical resultsof the transient specification (i.e., MOS, MUS, and TS)for the three proposed control schemes under variationin system parameters are very close to that of thenominal value of system parameters and within theacceptable ranges of the system frequency, according tothe European network of transmission system operatorsfor electricity codes [35], as indicated in Table 6.Although, the PI controller is very sensitive to thesystem uncertainty and non-linearity, which may berepresented by the main demerit of this controller insome industrial applications [12]. It does not need re-tuning its parameters when the EPS considering highRESs penetration subjected to system uncertainties.This demonstrates the robustness and superiority ofthe proposed coordinated control strategy-based theoptimal PI controller in regulation the system fre-quency in case of system parameters variations as wellas high penetration level of the RESs. Figures 15 and 16show the responses of the conventional generationsources and SMES power for the EPS under thesecondition of system uncertainties (i.e., ±50% of system

Fig. 13 The frequency deviation of the EPS for scenario 4,(+ 50% system parameters variations)

Fig. 14 The frequency deviation of the EPS for scenario 4,(− 50% system parameters variations)


parameters) and high-level RESs penetration. Fromthese figures, it is obvious that the conventional gener-ation units largely generated the required power whenapplied a sudden load change at t = 200 s in cases of theEPS with/without SMES, while the proposed coordi-nated control strategy could largely reduce the requiredpower from the conventional generation units. Thus,the SMES power is fastly discharged by the proposedcoordinated control strategy when the connection ofload change, while the SMES power is greatly chargedby the proposed coordinated control strategy when theconnection of solar irradiation and wind power at timet = 0 s, and t = 1000 s, respectively.

7 ConclusionUtilizing Renewable Energy Sources (RESs) isattracting great attention in today's power systemto face the challenges of future energy shortages.However, the irregular nature of RESs and randomload deviations cause large frequency/voltage fluctu-ations as well as reducing the system inertia thatresults from replacing the synchronous generatorswith RESs. Where these effects resulting fromutilizing the RESs can limit their penetration. Inorder to benefit from a maximum capacity of theRESs, this paper proposes a new application ofSuperconducting Magnetic Energy Storage (SMES)

system based on an optimal PI controller that isoptimally designed by the Particle SwarmOptimization (PSO) algorithm to enhance the fre-quency stability of the Egyptian Power System(EPS) considering high RESs penetration. Further-more, the proposed SMES-based the optimal PIcontroller is coordinated with the secondary fre-quency control loop (i.e., Load Frequency Control(LFC)) for improvement and preservation the fre-quency stability of the EPS considering high RESspenetration. The conventional generation units inthe EPS is decomposed into three dynamic subsys-tems; non-reheat, reheat and hydraulic power plantsconsidering inherent nonlinearities (i.e., governordeadband and generation rate constraints of thepower plants). To prove the effectiveness of theproposed coordinated control strategy, the EPS con-sidering high-level RESs penetration has been testedusing Matlab/SIMULINK® software. The simulationsresults proved that the proposed coordinated con-trol strategy has achieved an effective performancefor maintaining the EPS frequency stability. Hence,the proposed coordinated control strategy betweenthe LFC and SMES system (i.e., auxiliary LFC)using the optimal PI controller-based the PSO algo-rithm will ensure an avoidance of power system in-stability and system collapse owing to high-levelRESs integration.


Scenario4

With SMES-based optimal PI controller With SMES Without SMES

MUS (pu) MOS (pu) TS (s) MUS (pu) MOS (pu) TS (s) MUS (pu) MOS (pu) TS (s)

+ 50% 1.850 × 10− 3 2.57 × 10− 5 17.869 7.56 × 10− 3 7.36 × 10− 3 24.512 1.269 × 10− 2 1.225 × 10− 2 27.018

-50% 2.015 × 10− 3 2.024 × 10− 3 18.559 9.681 × 10− 3 9.417 × 10− 3 26.142 2.188 × 10− 2 2.104 × 10− 2 27.119

Fig. 15 Conventional generation and SMES power responsesof the EPS for scenario 4, (+ 50% system parametersvariations)

Fig. 16 Conventional generation and SMES power responsesof the EPS for scenario 4, (− 50% system parametersvariations)


AcknowledgementsNot applicable.

FundingThis paper was funded by the Cultural Affairs and Missions Sector of theEgyptian Ministry of Higher Education.

Availability of data and materialsData sharing not applicable to this article as no datasets were generated oranalyzed during the current study.

Authors’ contributionsGM as a corresponding author, contributed significantly to analysis,manuscript preparation and manuscript submission. GS, AAE and YM assupervisors helped to perform the study analysis with constructivediscussions, professional advice and revised the manuscript. All authors readand approved the final manuscript.

Competing interestsThe authors declare that they have no competing interests.

Author details1Department of Electrical Engineering, Kyushu Institute of Technology,Tobata-Ku, Kitakyushu-shi, Fukuoka 804-8550, Japan. 2Department ofElectrical Engineering, Faculty of Energy Engineering, Aswan University,Aswan 81528, Egypt. 3Higher Institute of Engineering and technology, KingMarriott, Alexandria 23713, Egypt. 4Department of Electrical Engineering,Faculty of Engineering, Minia University, Minia 61517, Egypt.

Received: 10 August 2018 Accepted: 26 November 2018

References1. Rakhshani, E., Remon, D., Cantarella, A., et al. (2016). Analysis of derivative

control based virtual inertia in multi-area high-voltage direct currentinterconnected power systems. IET Gener Trans Distrib, 10(6), 1458–1469.

2. Magdy, G., Mohamed, E. A., Shabib, G., et al. (2018). SMES based a new PIDcontroller for frequency stability of a real hybrid power system considering highwind power penetration. IET Renewable Power Generation, 12(11), 1304–1313.

3. Cam, E., Gorel, G., & Mamur, H. (2017). Use of the genetic algorithm-basedfuzzy logic controller for load-frequency control in a two areainterconnected power system. Appl Sci, 7(3), 1–22.

4. Nasiruddin, I., Sharma, G., Niazi, K. R., et al. (2017). Non-linear recurrent ANN-based LFC design considering the new structures of Qmatrix. IETGeneration Trans Distribut, 11(11), 2862–2870.

5. Shahalami, S. H., & Farsi, D. (2018). Analysis of load frequency control in arestructured multi-area power system with the Kalman filter and the LQRcontroller. Int J Electron Commun, 86(1), 25–46.

6. Sun, Y., Wang, Y., Wei, Z., et al. (2018). Robust H□ load frequency control ofmulti-area power system with time delay: A sliding mode control approach.IEEE/CAA J Auto Sin, 5(2), 1–8.

7. Shabib, G., Abd-Elhamed, E. H., & Magdy, G. (2018). Robust digital redesignof continuous PID controller for power system using plant-input-mapping.WSEAS Trans Power Syst, 13(1), 31–39.

8. Illias, H., Fikri, A., Zahari, M., et al. (2016). Optimisation of PID controller forload frequency controller in two-area power system using evolutionaryparticle swarm optimisation. Int J Elect Syst, 12(2), 315–324.

1 AppendixTable 8 SMES data [19] SMES control loop parameters are; L = 3H, Tc = 0.03, Ido = 20KA, Kf = 0.001

Nomenclature

Δf The frequency deviation of the EPS (Hz) xnid The current position of particle i at iteration n

D System damping coefficient of the EPS (pu MW/Hz) vnid The velocity of particle i at iteration n

H Equivalent inertia constant (pu sec) c1, c2 Acceleration constant

T1 Valve time constant of the non-reheat plant (sec) n Number of iterations

T2 Steam valve time constant of reheat plant (sec) Pnid pbest of particle i at iteration n

T3 Water valve time constant hydro plant (sec) Pngd gbest of particle i at iteration n

Td Dashpot time constant of hydro plant speed governor (sec) ρ Air density (Kg/m3)

Th The time constant of reheat thermal plant (sec) AT The rotor swept area (m2)

Tw Water starting time in hydro intake (sec) VW The rated wind speed (m/s)

ΔPL Load variation (MW pu) CP The power coefficient of the rotor blades

m The fraction of turbine power (intermediate pressure section) PW The output power of the wind turbine (W)

R1 Governor speed regulation non-reheat plant (Hz/pu MW) β The pitch angle

R2 Governor speed regulation reheat plant (Hz/pu MW) rT The rotor radius

R3 Governor speed regulation hydro plant (Hz/pu MW) λ The tip-speed ratio (TSR)

Pn1 Nominal rated Power output for the non-reheat plant (MW pu) Ed The dc voltage applied to the inductor (KV)

Pn2 Nominal rated Power output for reheat plant (MW pu) Vdo The maximum bridge circuit voltage (KV)

Pn3 Nominal rated Power output for the hydro plant (MW pu) α The firing angle (degree)

ΔPc1 Regulating the system frequency of non-reheat plant (Hz) Id The current through the inductor (KA)

ΔPc2 Regulating the system frequency of reheat plant (Hz) Rc The equivalent commutation resistance (KΩ)

ΔPc3 Regulating the system frequency of hydro plant (Hz) Kf The feedback gain of the inductor current deviation

TWT Wind turbine time constant (sec) Tc The converter time delay

TPV Solar system time constant (sec) L The coil inductance (H)

w Inertia weight factor ΔPSMES The active power deviation of SMES unit

rand () A random number between 0 and 1


9. Gheisarnejad, M. (2018). An effective hybrid harmony search and cuckoooptimization algorithm based fuzzy PID controller for load frequencycontrol. Appl Soft Comput, 65(1), 121–138.

10. El-Saddy, G., Ibrahim, E. A., & Donkol, A. A. (2017). Ant Colony PID Controllers forNonlinear Load Frequency Control System. Int J Power Eng Energy, 8(1), 719–724.

11. Shankar, G., & Mukherjee, V. (2016). Load frequency control of anautonomous hybrid power system by quasi-oppositional harmony searchalgorithm. Elect Power Energy Syst, 78(1), 715–734.

12. Hasanien, H. M. (2017). Whale optimization algorithm for obtaining theoptimal PID controller parameters in interconnected modern power systemincluding renewable sources. IET Gener Transm Distrib, 12(3), 607–616.

13. Ma, M., Liu, X., & Zhang, C. (2017). LFC for multi-area interconnected powersystem concerning wind turbines based on DMPC. IET Generation TransDistrib, 11(10), 2689–2686.

14. Kouba, N. E., Menaa, M., Hasni, M., et al. (2016). LFC enhancementconcerning large wind power integration using new optimised PIDcontroller and RFBs. IET Generation Trans Distrib, 10(16), 4065–4077.

15. El-Fergany, A. A., & El-Hameed, M. A. (2017). Efficient frequency controllersfor autonomous two-area hybrid microgrid system using social-spideroptimiser. IET Generation Trans Distrib, 11(3), 637–648.

16. Marzband, M., Ghazimirsaeid, S. S., Uppal, H., et al. (2017). A real-timeevaluation of energy management systems for smart hybrid homemicrogrids. Electr Power Syst Res, 143(1), 624–633.

17. Amrouche, S. O., Rekiouab, D., Rekiouab, T., et al. (2016). Overview ofenergy storage in renewable energy systems. Int J Hydrog Energy,41(45), 20914–20927.

18. Zargar, M. Y., Mufti, M. U., & Lone, S. A. (2017). Adaptive predictive control ofa small capacity SMES unit for improved frequency control of a wind-dieselpower system. IET Renewable Power Generation, 11(1), 1832–1840.

19. Khosraviani, M., Jahanshahi, M., Farahani, M., et al. (2018). Load-frequencycontrol using multi-objective genetic algorithm and hybrid sliding modecontrol-based SMES. Int J Fuzzy Syst, 20(1), 280–294.

20. Magdy, G., Shabib, G., Elbaset, A. A., et al. (2018). A robust control strategyfor mitigating renewable energy fluctuations in a real hybrid Powe systemcombined with SMES. AIP Conf Proc, 1968(1), 1–12.

21. Padhan, S., Sahu, R. K., & Panda, S. (2014). Automatic generation control withthyristor-controlled series compensator including superconductingmagnetic energy storage units. Ain Shams Eng J, 5(1), 759–774.

22. Bevrani, H. (2014). 'Robust power system frequency Control’ (2nd ed.). NewYork: springer.

23. Egyptian Electricity Holding Company, Annual Report, available at: http://www.eehc.gov.eg/eehcportal/YearlyReports.aspx, 2016.

24. Mobarak, Y. A. (2015). Voltage collapse prediction for Egyptianinterconnected electrical grid EIEG. Int J Elec Eng Inform, 7(1), 79–88.

25. Magdy, G., Bakeer, A., Shabib, G., et al. (2017). Decentralized Model PredictiveControl Strategy of a Realistic Multi Power System Automatic GenerationControl’. IEEE Conf., 16th international Middle East power systems conference(MEPCON’19) (pp. 190–196). Egypt: Menoufia University.

26. Bevrani, H., Habibi, F., Babahajyani, P., Watanabe, M., & Mitani, Y. (2012).‘Intelligent frequency control in an AC microgrid: Online PSO-based fuzzytuning approach’, IEEE trans. Smart Grid, 3(4), 1935–1944.

27. Abraham, R. J., Das, D., & Patra, A. (2007). Automatic generation control of aninterconnected hydrothermal power system considering superconductingmagnetic energy storage. Electrical Power Energy Syst, 29(1), 571–579.

28. Bhanu, P., Bhushan, C., Sujatha, K., et al. (2011). Load Frequency Controller with PIController Considering non-Linearities and Boiler Dynamics (pp. 290–294). India:2nd international conference on sustainable energy and intelligent system(SEISCON).

29. Kerdphol, T., Qudaih, Y., & Mitani, Y. (2016). Optimum battery energy storagesystem using PSO considering dynamic demand for microgrids. ElectricalPower and Energy Syst, 83(1), 58–66.

30. Kennedy, J., & Eberhart, R. C. (1995). ‘Particle swarm optimization’, proc. IEEE IntConf Neural Net Perth, Aust IEEE Service Center Piscataway, 4(1), 1942–1948.

31. Ang, K. H., Chong, G., & Li, Y. (2005). PID control system analysis, design, andtechnology. IEEE Trans on Control Syst Technol, 13(4), 559–576.

32. Gaing, Z. L. (2004). A particle swarm optimization approach for optimumdesign of PID controller in AVR system. IEEE Trans on Energy Conversion,19(2), 384–391.

33. Valarmathi, K., Devaraj, D., & Radhakrishnan, T. K. (2005). Particle SwarmOptimization based PI Controller Tuning for Fermentation Process (pp. 1–6).

Vienna: IEEE Conf., International conference on computational intelligencefor modelling, Control and Automation.

34. Abd-Elazim, S. M., & Ali, E. S. (2016). Load frequency controller design viaBAT algorithm for nonlinear interconnected power system. Int J Electr PowerEnergy Syst, 77(1), 166–177.

35. European Network of Transmission System Operators for Electricity (ENTSO-E). (2013). Supporting Document for the Network Code on Load-FrequencyControl and Reserves’, chapter 4 (pp. 17–36).


http://www.eehc.gov.eg/eehcportal/YearlyReports.aspx

http://www.eehc.gov.eg/eehcportal/YearlyReports.aspx

protection and control of modern power systems ......hence, the modern power systems become more...

Documents