Optimal Battery Energy Storage Size Using Particle ...

International Review of Electrical Engineering (I.R.E.E.), Vol. 10, N. 2 ISSN 1827- 6660 March – April 2015

Optimal Battery Energy Storage Size Using Particle Swarm Optimization for Microgrid System T. Kerdphol, Y. Qudaih, Y. Mitani Abstract – This paper presents the method to evaluate the optimal size of Battery Energy Storage System (BESS) for the microgrid based on load frequency control. The novel optimization and control algorithm considering load shedding as dynamic demand response has been proposed with the combination of the optimal size of BESS in order to control the system frequency of the microgrid. The feasibility of using optimal BESS with load shedding is proposed when the microgrid is disconnected from the main utility grid. Particle Swarm Optimization (PSO) is developed and presented to determine the optimal size, and to achieve the lowest total cost of BESS in the microgrid. Results show that the optimal size of BESS-based PSO with load shedding can achieve higher performance of dynamic stability compared to the optimal size of BESS-based analytic algorithm with load shedding. At the same time, the sizing methods also evaluated the impact of BESS specified costs with different storage technologies for the microgrid system. The total costs of BESS were demonstrated and compared. Copyright © 2015 Praise Worthy Prize S.r.l. - All rights reserved.

Keywords: Battery Energy Storage System, Dynamic Demand Response, Frequency Control, PSO

As the output characteristics of these DGs are quite different from the conventional energy sources, the system should be capable of handling unexpected fluctuation and maintaining system reliability. Moreover, a microgrid can operate in either standalone or grid-connection mode. It is capable to work in parallel with the utility grid, with the capacity to switch to stand-alone operation in case of emergency situation in the grid. When an islanding operation occurs where a DG or a group of DG sources continue to supply the microgrid system that has been separated from the utility grid, the system needs to have a master generator which can provide voltage and frequency support. Generally, a synchronous generator can fulfill this demand. When there is no synchronous generator, Voltage Source Converter (VSC) interfaced batteries can act as the master control. Therefore, storage devices serve as an important aspect in microgrid operations. Presently, battery systems with the capacity of 2 MW have already been utilized to support a microgrid [3], [4]. Battery Energy Storage System (BESS) can be used in various aspects of power system as one key factor for sustainable energy in many countries particularly in Europe, America and Japan. Advantages of BESS include improvement of the system frequency especially when BESS is used for system frequency control in the form of load frequency control (LFC). In the case of small disturbance, BESS is discharging when the system frequency is lower than 50 or 60 Hz. On the other hand, BESS is charging when the system frequency is higher than 50 or 60 Hz. In the case of large disturbance, BESS can enhance the performance

Nomenclature αi Kb Tb Ed0 Eb1 Ebt Eboc IBESS PBESS rc rbt rbs rbp rb1 ∆f

Firing delay angle of converter i Control loop gain Measurement device time constant Maximum DC voltage of the batteries Battery resistance Phase voltage of the battery side Battery open circuit voltage DC current though the battery Active power provided by the batteries Battery overvoltage Terminal voltage of the battery Battery internal resistance Self discharge resistance Overvoltage resistance Frequency deviation

I.

Introduction

Rising electricity demands, decreasing fossil fuel resources associated with conventional generators call for an amendment of alternative electric energy generation method worldwide. Microgrid system, a principle concept integrating distributed generation sources (DGs) is being considered as one of the solutions to this energy concern, and is gaining more attention recently [1]. It can be viewed as a group of distributed generation sources connected to the loads in which the DGs can be fed to the loads alone or be fed to the utility grid [2].

Copyright © 2015 Praise Worthy Prize S.r.l. - All rights reserved

277

T. Kerdphol, Y. Qudaih, Y. Mitani

of the system frequency control by integrating BESS with under frequency load shedding scheme, or under frequency generation trip, and over frequency generation trip. With these different functions, BESS can offer a good solution. It can be concluded that BESS is a rapid and flexible element for power system [5]-[7]. Moreover, the purpose of optimal BESS is to smoothen the power in a system with wind or solar energy. In such a system, BESS can play a role by absorbing the surplus power and compensating the power shortage due to the uncertainties of renewable energy. Proceeding [8] derived a dynamic model of BESS implemented to power system; [9] studied the influence of certain size of BESS on frequency regulation in power system stability; [11] applied values analysis of different BESS technologies and applications; [7] combined an incremental model of BESS into LFC of a stand-alone power system with performance observation; [3] implemented BESS to a microgrid system and investigated an operation and control strategy; [6] studied the optimization and dimensioning of BESS based on the state of charge (SOC) which applied to primary frequency control in interconnected networks. Furthermore, the optimization method can be achieved by means of linear programming method, enumerative method, iterative algorithm, genetic algorithm (GA) and particle swarm optimization (PSO) [12]. The advantages of PSO include simplicity, ease of use, high convergence rate and minimal storage requirement. Also, it is less dependent on the set of the initial points compared to other methods which implies that convergence algorithm is robust [13]-[16]. With these reasons, this paper proposes the optimal sizing of BESS using PSO based on the load-frequency control for a microgrid system, and applies the load shedding as dynamic demand response to the microgrid system. In this paper, dynamic demand response based frequency control will turn off noncritical loads in response to frequency deviations in order to maintain the system frequency of the microgrid. One of the past methods [3] was implementing large interconnected power system using small BESS rated power of 2 MW in the comparison of the total volume of the spinning reserve provided by conventional generation of 3000 MW. Therefore, the impacts of BESS on system frequency profile were widely neglected. With the consideration of a microgrid system (e.g., a small power system) the BESS rated power cannot be neglected anymore, and thus the system frequency is more influence by the BESS output power variations [3]. So, the installation of large or inappropriate size of BESS can cause frequency problems, and add the extra cost to the microgrid system. For these reasons, optimal sizing of BESS is an essential factor for a microgrid system [5]-[6]. BESS cost is a function of the storage device power and capacities. These specified costs depend on technology selection. Frequently, those specified costs are difficult to determine, include a lot of uncertainty and develop with research activities [10]. BESS technologies

are presented in this paper are redox-flow batteries. Redox-flow BESS is a relatively new commercially available battery and differs from conventional BESS (e.g., Lead-acid BESS) in such a way that the amount of energy it can store is independent on its power rating [17]. Moreover, redox-flow BESS can be designed for both high power and large energy storage. Due to its new commercialization, recent studies on redox-flow BESS in the microgrid are limited [18], [19]. In the present work, specified costs of BESS are separated in order to compare performances of different BESS technologies in the microgrid. This paper uses the following technologies: Polysulfide-bromine (PB) and Vanadium Redox (VR) BESS technologies. The main purpose of this paper is to improve the frequency control and reduce the total cost of BESS in the microgrid by integrating optimal size of BESS with load shedding [20]-[22]. Load shedding is considered as dynamic demand response in the proposed algorithms [23], [24]. The proposed methodology is based on PSO algorithm in order to define the optimal size of BESS, and to achieve the lowest total cost of BESS for the microgrid. At the same time, as BESS adds the extra cost to the microgrid system, the sizing methods also evaluated the impact of BESS specified costs with different storage technologies for the microgrid system. The total costs of BESS installation were demonstrated and compared.

II.

System Configuration II.1.

Microgrid System

A typical characteristic of a microgrid is that it can be operated either in grid-connected mode or stand-alone mode. Under normal operation, the microgrid is connected to the utility grid. Fig. 1 shows the microgrid system which consists of a 1.2 MW mini-hydro generator, 2 MW hydro generator and 3 MW photovoltaic sources. BESS is connected to the microgrid system at bus 14. The system also consists of group of feeders which could be part of the distribution system.

Fig. 1. Microgrid architecture

The critical load 1 and 4 require a local generation and the non-critical load 2, 3 and 5 are not connected to any local generation. This system is typical to a real system available at the promoting renewable energy in Mae

Copyright © 2015 Praise Worthy Prize S.r.l. - All rights reserved

International Review of Electrical Engineering, Vol. 10, N. 2

278

T. Kerdphol, Y. Qudaih, Y. Mitani

Hong Son project which has initiated and funded by United Nations Development Programme (UNDP) and Global Environment Facility (GEF). This project will operate initially in Mae Hong Son province, which the Ministry of Energy has identified as its target to be the first energy self-sufficient province in Thailand. II.2.

From the schematic structure of BESS, the output of DC voltage is shown as:

Edo 

The output power of photovoltaic (PV) is uncertain as it is mostly affected by the environmental factors, namely the environmental random changes will inevitably lead to constantly changing of output power of PV [25], [26]. In order to illustrate PV characteristics in operating condition, the influence of solar radiation and atmosphere temperature are designed. The temperature effect is denoted by a temperature coefficient of Tco (1/c°). The efficiency of the inverter is multiplied by the DC output converting DC to AC output as in (1):

PBESS Et

6 6 



rb 1

Ebt

cos

cb1

r

6X rc  co bt 

rbs

E b1

rbp cbp

Eboc

Fig. 3. Equivalent circuit of BESS [7]

The terminal voltage of the equivalent battery can be calculated from:

where nPV is PV modules number, Prate PV is the PV array rated power (W), G is the global insolation on the PV array (W/m2), G0 is the standard amount of insolation rating the capacity of PV modules (W/m2), TA is the ambient temperature, TCO is the temperature coefficient of the maximum power of PV, ηrel is the relative efficiency of the PV modules, ηinv is the efficiency of the inverter. II.3.

Edo

I BESS

PPV  nPV Prate PV  G / G0  1  Tco TA  25 invrel (1)



(2)

where Et is AC voltage between the line to neutral. The equivalent circuit analysis of BESS consists of a converter connected to an equivalent battery as shown in Fig. 3.

Photovoltaic Power Generation



6 6 Et 

Ebt  Edo cos   rc I BESS  

3 6 6 Et  cos 1  cos  2   X co I BESS  

(3)

According to the equivalent circuit of BESS, the expression of DC current flowing into the battery can be expressed as:

Battery Energy Storage System (BESS)

Generally, BESS can be utilized in several implementations such as peak shaving, real power control and load leveling. This paper uses BESS for enhancing the performance of frequency control as BESS can provide active power compensation in a short period of time. Recently, a huge number of electric companies and independent system operators have shown growing interest in BESS due to decreasing cost of batteries. With the fast development of technologies, BESS is expected to be used in several applications including the one proposed in this paper. Considering the optimum BESS as a function of cost and size is modeled in this paper. The structure of BESS consists of power converters, battery cells and control parts [7] which are shown in Fig. 2.

I BESS 

Ebt  Eboc  Eb1 rbt  rbs

(4)

where: Eboc 

Eb1 

rbp

I BESS

(5)

rb1 I BESS 1  STb1

(6)

1  STbp

Tbp  rbp Cbp

(7)

Tb1  rb1Cb1

(8)

From the converter circuit analysis, the active and reactive powers absorbed by BESS are:

Fig. 2. Schematic structure of BESS and interconnection diagram

Copyright © 2015 Praise Worthy Prize S.r.l. - All rights reserved

PBESS 

3 6 Et I BESS  cos 1  cos  2  

(9)

QBESS 

3 6 Et I BESS  sin 1  sin  2  

(10)

International Review of Electrical Engineering, Vol. 10, N. 2

279

T. Kerdphol, Y. Qudaih, Y. Mitani

Only incremental active power is considered in load frequency control. Thus, the P-modulation strategy is introduced in this paper. For P-modulation, α1=- α2= α. Therefore:

where PBESS (MW) and CBESS (MWh) are BESS power and BESS capacity. CP ($/MW) and CW ($/MWh) are the specific power cost of BESS and the specific capacity cost of BESS.

3 6 Et I BESS cos  

(11)

III.2. Operating and Maintenance Cost of BESS

QBESS  0

(12)

The operating and maintenance cost of BESS (CO&M) is shown as:

0 PBESS  I BESS Ed

(13)

CO&M  CMf PBESS  Wannual CMV

PBESS 

Then, the use of BESS in load frequency control is reached by a damping signal ∆Ed :

Ed 

Kb f 1  STb

where CMf ($/MW/year) is the fixed cost of BESS. CMV ($/MWh/year) is the variable operating and maintenance cost of BESS. Wannual (MWh/year) is the energy storage system annual discharge energy. In this paper, the total cost is calculated in two cases depending on BESS technologies. This paper uses the following technologies: Polysulfide-bromine (PB) and Vanadium Redox (VR) BESS technologies which are shown in Table I.

(14)

The ∆f is the beneficial feedback from the power system in order to provide a damping effect. From linearizing (2) to (14), the block diagram of BESS can be shown in Fig. 4.

f

K

Et

bp

1  ST

bp

max PBESS

 PBESS



min BESS

P



Edo cos 0 I BESS

 E co

TABLE I BESS TECHNICAL AND ECONOMICAL DATA [28] Technologies VR BESS PB BESS CP ($/kW) 426 150 CW ($/kWh) 100 65 CMf ($/kW/year) 9 9 CMV ($/kWh/year) 0 0 Life time (years) 15 15 Efficiency (%) 70 65

6 X co  

Ebt

1 Tbs  Tbp





Eb1

I BESS

rb1 1  STb1

(16)



IV.

Proposed Sizing Strategy and Control

Tbp

IV.1. Objective Function

1  STbp

In this paper, the purpose of objective functions is to improve the frequency control and decrease the total cost of BESS in the microgrid by integrating load shedding with optimal size of BESS. The objective function f1 and f2 have the same weight. To minimize the power and the total cost of BESS, the final objective function are chosen and expressed as:

E boc 0 I BESS

Fig. 4. Linearized BESS model for load frequency control

III. Cost Consideration The total cost for adding the BESS to a microgrid system includes the cost of required inverter. A review of BESS retail prices showed that BESS price is a function of the power and capacity rating. In order to calculate an economical cost of installing BESS, this paper considers the BESS capital, operating and maintenance costs. The parameters of these costs depend on energy purchase costs and chosen BESS technology [27].

Minimize f1 = min (PBESS)

(17)

Minimize f2 = min (CT) = min (CCapital + CO&M )

(18)

under the constraints of: min max PBESS  PBESS  PBESS min max CBESS  CBESS  CBESS

III.1. Capital Cost of BESS

Fmin  F  Fmax

The capital cost of BESS (CCapital) can be expressed as:

CT  CTmax CCapital  CP PBESS  CW CBESS

(15)

Copyright © 2015 Praise Worthy Prize S.r.l. - All rights reserved

International Review of Electrical Engineering, Vol. 10, N. 2

280

T. Kerdphol, Y. Qudaih, Y. Mitani

Step 5) Increase particle jth by 1. Then, check the condition If particle jth +1≤ NP, go back to step 3). Step 6) Adjust Gbest of ith = the best values of Pbest at ith. Step 7) Increase iteration ith by 1. Then, check the condition if iteration ith +1≤ NI, go back to step 2) and update the new position (xi+1) and new velocity (vi+1) for the next iteration. Step 8) If iteration ith +1> NI, the process ends. The overall flow chart of the optimal sizing of BESSbased PSO with load shedding is shown in Fig. 5.

where F is the frequency of the microgrid system (Hz), PBESS is the power capacity of BESS (MW), CBESS is the energy capacity of BESS (MWh), CCapital is the capital cost of BESS ($), CO&M is the operating and maintenance cost of BESS ($), CT is the total cost of BESS ($/15 years). IV.2. Optimal Sizing of BESS-based PSO Algorithm with Load Shedding Particle swarm optimization is an algorithm technique for searching optimal parameters of complicated search spaces. PSO is initiated with a group of random particles to find for optimal parameters by updating generations. In each iteration, each particle is updated by two values. These values are called Pbest and Gbest respectively. Pbest is the best solution acquired by each particle itself in all of the previous generations. Gbest is the best value obtained by any particle in all previous iterations. This value is called the best global solution. Each particle updates its position and velocity by using (19) and (20):

vi 1  vi  c1r1  Pbest  xi   c2 r2  Gbest  xi 

(19)

xi 1  vi 1  xi

(20)

min max PBESS PBESS

where: ith is the iteration number, jth is the particle number, vi is the velocity of a particle at iteration ith, xi is the position of a particle at iteration ith, Pbest is the best solution at iteration ith, Gbest is the best global solution at iteration ith, r1 is the random number one between 0 and 1, r2 is the random number two between 0 and 1, c1, c2 are the learning factors. The learning factors have significant effects on the algorithm convergence rate. Further information for PSO can be found in [12]. The PSO parameters chosen for this proposed algorithm are: • The number of particles in swarm (NP) = 20 • The number of iterations (NI) = 30 • Learning factors are C1= C2 = 1.49445 • Inertia weight (w) = 0.7920 The main steps for the proposed optimal BESS-based PSO algorithm are: Step 1) Initialize the parameters and iteration i = 1 with random position (xi) and velocity (vi). Load shedding is set based on the measured power imported by the grid at the moment of islanding. Step 2) Start the particle j = 1 in the swarm. Step 3) Execute the objective functions for particle jth of iteration ith. Step 4) Find Pbest and Gbest for particle jth of iteration ith. If the fitness value of jth ≤ the best global fitness value, the program sets Pbest = f1 (xj,) and saves the best total cost (Pbest_CT) = f2 (xj.).

Fig. 5. Overall flow chart of proposed optimal sizing of BESS-based PSO with load shedding

IV.3. Optimal Sizing of BESS-based Analytic Algorithm with Load Shedding The overall flow chart of the optimal sizing of BESSbased analytic algorithm with load shedding is shown in Fig. 6. The solution algorithm of this method is as follows: Step 1) Read the input data of the microgrid system and get the power values of photovoltaic (PPV), hydro generator (PHV), mini hydro generator

Copyright © 2015 Praise Worthy Prize S.r.l. - All rights reserved

International Review of Electrical Engineering, Vol. 10, N. 2

281

T. Kerdphol, Y. Qudaih, Y. Mitani

(PMini) and the load 1 to 5 (PL). Step 2) Set the desired power of load shedding (PS). Load shedding is based on the measured power imported by the grid at the moment of islanding. Step 3) Calculate the total power of all loads (PLT) after setting the load shedding. Then, the program calculates the differential power (PDP) between generations and loads. Step 4) Set the differential power (PDP) equal to the power of BESS (PB). Step 5) Execute the objective functions. Step 6) Check the condition if the microgrid frequency is within the frequency range set by the user. If the microgrid frequency does not satisfy the condition, the program will increase PBESS by 0.1 MW and the condition is run again. This step will continue until it meets the condition. Step 7) If the microgrid frequency satisfies the frequency condition, the process ends.

V.

Simulation Results

In normal state, all power generated from the microgrid are used to supplying the loads. If the state of charge of BESS is greater than the depth of discharge, the shortage power is supplied by BESS. If BESS could not supply the shortage power, the shortage power is supplied by the utility grid. However, if the generation cannot match the load demand in a short time during a large disturbance (e.g., three phase fault) or islanding, the system frequency will drop drastically. This significant drop will likely cause further extreme consequence such as system collapse. In this case, under frequency load shedding scheme plays an important role in frequency control and BESS can improve the performance of frequency control in load shedding. Whereas the dynamic demand reduction could be considered as load shedding. Non-critical loads allow for demand reduction with relative ease, reducing the demand of industrial processes requires a more complex solution. In fact, dynamic demand response based frequency control will turn on or off non-critical loads in response to frequency deviations in order to maintain the system frequency. In this part, the optimal sizing of BESS with load shedding is accomplished by using DPL script in DIgSILENT PowerFactory, and the performance of the system frequency is discussed in the simulation results. Nevertheless, the proposed methodology and control developed and presented in this paper are not dependent upon the input profile chosen. In the present work, the microgrid system consists of three renewable energy sources namely Mini-hydro generation, hydro generation, and photovoltaic generation with average output power of 1.2 MW, 2 MW and 3 MW respectively. Fig. 7 shows the generated power of photovoltaic system. The load 1 and 4 are the critical load with power of 1.85 MW and 1.9 MW respectively. The load 2, 3 and 5 are the non-critical load with power of 1.7 MW, 1.75 MW and 2.4 MW respectively.

PS

PS PL

PLT  PL  PS

PDP  PLT  PPV  PHY  PMini

3.05

PBESS  PDP 3

PBESS PBESS0.1

PV power output (MW)

PB  0

2.95

2.9

2.85

2.8

F Min  F  F Max

2.75

0

10

20

30

40

50

60

Time (s)

Fig. 7. Power output of Photovoltaic system Fig. 6. Overall flow chart of the optimal sizing of BESS-based analytic algorithm with load shedding

In this paper, an islanding operation is implemented to test the proposed optimization algorithm, and to observe

Copyright © 2015 Praise Worthy Prize S.r.l. - All rights reserved

International Review of Electrical Engineering, Vol. 10, N. 2

282

T. Kerdphol, Y. Qudaih, Y. Mitani

the system performance in the microgrid system. First, the three phase fault has occurred on bus 18 at t = 10.0s. Then, the microgrid system is isolated from the main grid. At 10.5s, the system frequency drop to 59.95 Hz (see Fig. 11) which is the minimum limit of non-critical frequency range in the microdrid system. At that time, the system has applied the load shedding to the noncritical load 5 as dynamic demand response which is the dump load. Normally, the time to clear the fault and restore the system after the islanding operation occurred takes around 1 hour. This paper assumes that BESS can support the microgrid system for about 2 hours after the microgrid system is isolated from the main grid. Fig. 8 displays the convergence rate of the optimal sizing of BESS-based PSO.

Fig. 9. Improvement of the system frequency with load shedding

4.5

4

Fitness value

3.5

3

2.5

2

1.5

1

0

5

10

15 Iteration

20

25

Fig. 10. Improvement of the system voltage with load shedding

30

Fig. 8. Convergence rate of the optimal size of BESS-based PSO

The PSO algorithm is converged to its final state approximately after 15 iterations. It is evident that attainment is achieved with Gbest after 15th iterations. The slightly attainment is reached with Gbest after 20th iterations. However, no difference is observed after 20th iterations. The reason for this is that this algorithm is symmetrical around the midpoint. In the case where the system cannot apply load shedding scheme when the islanding operation occurs, the system will collapse as shown in Figs. 9 and 10. Due to the islanding, the dynamic demand reduction could be considered as load shedding which is a method of frequency control protection. In this paper, BESS and load shedding scheme work together at the same time. The performance of the system frequency with optimal size of BESS and load shedding scheme is observed and shown in Fig. 11. From this figure, the noncritical frequency range defines a frequency deviation deadband where no primary frequency regulation is applied. Outside of this deadband, the optimal size of BESS with load shedding is applied. Considering the case of a microgrid system (e.g., a small power system) the BESS rated power cannot be neglected anymore, and thus the grid frequency is more influence to BESS output power variations.

Fig. 11. Frequency deviation of the microgrid system after islanding TABLE II FREQUENCY AND VOLTAGE OF MICROGRID AFTER ISLANDING Power Frequency (Hz.) Voltage (pu.) Case capacity Reference : Reference : (MW) 60.00 0.9890 No BESS 0 41.0258 0.8405 BESS without 2.0000 60.3011 0.9950 optimal size BESS-based 1.3124 60.0012 0.9890 PSO BESS-based 1.4000 60.1905 0.9921 Analytic

Therefore, the installation of large or inappropriate size of BESS can cause frequency problems, and add the

Copyright © 2015 Praise Worthy Prize S.r.l. - All rights reserved

International Review of Electrical Engineering, Vol. 10, N. 2

283

T. Kerdphol, Y. Qudaih, Y. Mitani

extra cost to the microgrid system. Based on Fig. 11, the magnitude of the frequency deviation with the optimal sizing of BESS-based analytic algorithm is higher and takes more time to stabilize than the optimal sizing of BESS-based PSO. However, without optimal size of BESS, the magnitude of the frequency deviation reaches 61 Hz and takes longer to stabilize. In case of no BESS, the system frequency dropped drastically because the power supply cannot meet the load demand. It can be seen in Table II that in case of the optimal sizing of BESS, the performance of system frequency is much better when the optimal sizing of BESS-based PSO is applied simultaneously with load shedding. Besides, the voltage deviation of the microgrid system is shown in Fig. 12. According to this figure, the optimal sizing of BESS-based PSO can still maintain the voltage deviation better than the optimal sizing of BESSbased analytic algorithm. Thus, with the optimal size of BESS-based PSO with load shedding, it is clearly seen that there is a significant improvement of dynamic performances of the microgrid system in terms of peak deviation and settling time. In terms of BESS economical performance, this paper presents a comparative method to evaluate economical performances of BESS installation with different storage technologies in the microgrid. From Table III, the optimal sizing of BESS-based PSO algorithm also gives the lowest total cost of BESS compared to optimal sizing of BESS-based analytic algorithm.

BESS. Thus, the polysulfide-bromine (PB) BESS installation is likely to be more cost-effective than the vanadium redox (VR) BESS installation in the microgrid system for 15 years.

VI.

Conclusion

This paper proposes a control scheme using the optimal sizing of BESS-based PSO with load shedding in order to improve the system frequency and to reduce total cost of BESS in the microgrid system. Load shedding is considered as dynamic demand response in the proposed algorithms. Based on the results, BESS has shown a good performance in load frequency control and can be used for emergency control purposes when the system load shedding is unavoidable. It is obvious that the proposed algorithm provides quicker, smoother and more secure system stability than the optimal sizing of BESS-based analytic algorithm from the emergency situation to the normal equilibrium state (see Fig. 11 and Fig. 12). The optimal sizing method also evaluates the economical performance of BESS with different technologies. The total costs of BESS were demonstrated and compared. It is clear that the polysulfide-bromine BESS installation is likely to be more cost-effective than the vanadium redox-flow BESS installation in the microgrid for 15 years (see Table III). Additionally, the sizing of BESS-based PSO method can help to guarantee the lowest size of BESS with a reasonable and full use of the stand-alone microgrid system, so that the microgrid system can operate at the optimum conditions with optimal size of BESS in terms of investment and reliability requirement of the demand load.

References [1]

[2]

[3]

Fig. 12. Voltage deviation of the microgrid system after islanding TABLE III TOTAL COST OF BESS INSTALLATION FOR MICROGRID Case Power Energy Total cost of Total cost of capacity capacity VR BESS PB BESS (MW) (MWh) ($/15 years) ($/15 years) No BESS 0 0 0 0 BESS without 2.000 4.000 1,522,000 830,000 optimal size BESS-based 1.3124 2.6248 998,736 544,646 PSO BESS-based 1.400 2.800 1,065,400 581,000 Analytic

[4]

[5]

[6]

[7]

It is seen that in the microgrid system with polysulfide-bromine (PB) BESS, the total cost is lower than the microgrid system with vanadium redox (VR)

[8]

Copyright © 2015 Praise Worthy Prize S.r.l. - All rights reserved

H. Karimi, H. Nikkhajoei, and R. Iravani, Control of an electronically coupled distributed resource unit subsequent to an islanding event, IEEE Trans. Power Delivery, vol. 23, no. 1, Jan. 2008, pp. 293-501. R. Wai and W. Wang, Grid-connected photovoltaic generation system, IEEE. Trans. Circuits System, vol. 55, no. 3, Apr. 2008, pp. 953-964. F. Katiraei, M. Iravani, and P. Lehn, “Micro-grid autonomous operation during and subsequent to islanding process,” IEEE Trans Power Delivery, vol. 20, no. 1, Jan. 2005, pp. 248-257. Barone, G., Brusco, G., Burgio, A., Menniti, D., Pinnarelli, A., Sorrentino, N., A power management and control strategy with grid-ancillary services for a microgrid based on DC bus, (2014) International Review of Electrical Engineering (IREE), 9 (4), pp. 792-802. P. Mercier, R. Cherkaoui, and A. Oudalov, Optimizing a battery energy storage system for frequency control application in an isolated power system, IEEE Trans. Power System, vol. 24, Aug. 2009, pp. 1469-1477. A. Oudalov, D. Chartouni, and C. Ohler, Optimizing a battery energy storage system for primary frequency control, IEEE Trans. Power System, vol. 22, Aug. 2007, pp. 1259-1266. S. K. Aditya and D. Das, Application of battery energy storage system to load frequency control of an islanded power system, International Journal of Energy Research, vol. 23, 1999, pp. 247258. C.F. Lu, C. C. Liu, and C. J. Wu, Dynamic modelling of battery

International Review of Electrical Engineering, Vol. 10, N. 2

284

T. Kerdphol, Y. Qudaih, Y. Mitani

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20] [21]

[22]

[23]

[24]

[25] [26]

[27]

[28]

energy storage system and application to power system stability,”Proc. Inst. Elect. Eng., Gen., Transm., Distrib., vol. 142, no. 4, Jul. 1995, pp. 429-435. D. Kottick, M. Blau, and. D. Edelstein, Battery energy storage for frequency regulation in an island power system, IEEE Trans. Energy Convers., vol. 8, no. 3, Sep. 1993, pp. 455-458. Nguyen, M.Y., Yoon, Y.T., Battery modeling for economic operation of energy storage systems in electricity markets, (2013) International Review of Electrical Engineering (IREE), 8 (3), pp. 1000-1007. A. Ouadalov, D. Chartouni, C. Ohler, and G. Linhofer, Value analysis of battery energy storage application in power systems, Proc. 2nd IEEE PES Power System Conf. Expo., GA, 2006, pp. 2206-2211. J. Kennedy and R. Eberhart, Particle swarm optimization, in Proc IEEE Int. Conf. Neural Networks, Perth, Australia, vol. 4, 1995, pp. 1942-1948. A. A. Esmin, G. Lambert-Torres, and A. C. Zambroni de Souza, A hybrid particle swarm optimization applied to loss power minimization, IEEE Trans. Power Systems, vol. 20, no. 2, May. 2005, pp. 859-866. Z. L. Gaing, A particle swarm optimization approach for optimum design of PID controller in AVR system, IEEE Trans. Energy Conversion, vol. 19, no. 2, Jun. 2004, pp. 284-391. Z. L. Gaing, Particle swarm optimization to solving the economic dispatch considering the generator constraints, IEEE Trans. Power Systems, vol. 18, no. 3, Aug. 2003, pp. 1187-1195. Dhayalini, K., Sathiyamoorthy, S., Christober Asir Rajan, C., Particle swarm optimization technique for the coordination of optimal wind and thermal generation dispatch, (2013) International Review of Electrical Engineering (IREE), 8 (6), pp. 1843-1849. T. A. Nguyen, X. Qiu, J. D. Guggenberger, M. L. Crow, and A. C. Elmore, Performance characterization for photovoltaicvanadium redox battery microgrid systems, IEEE Trans. Sustainable Energy, Mar. 2014, pp. 1-10. J. Guggenberger, A. C. Elmore, J. Tichenor, and M. L. Crow, Performance prediction of a vanadium redox battery for use in portable, scable microgrid, IEEE Trans. Smart Grid, vol. 3, Dec. 2012, pp. 2109-2116. Darabi, A., Hosseina, M., Gholami, H.R., Vanadium redox flow battery control in flexible microgrids, (2013) International Review of Electrical Engineering (IREE), 8 (4), pp. 1341-1348. Y. Halevi and D. Kottick, Optimization of load shedding system, IEEE Trans. Energy Conversion, vol. 8, Jun. 1993, pp. 207-213. P. M. Anderson and M. Mirheydar, “A low-order frequency response model,” IEEE Trans. Power System, vol. 5, Aug. 1990, pp. 720-729. P. M. Anderson and M. Mirheydar, An adaptive method for setting underfrequency load shedding relays , IEEE Trans. Power System, vol. 7, May 1992, pp. 647-655. Cai, G.W., Yang, D.Y., Jiang, Y.T., Synchrophasor associated adaptive control strategy for under frequency protection and load shedding in smart grid, (2013) International Review of Electrical Engineering (IREE), 8 (1), pp. 227-235. A. M. Garcia, F. Bouffard, and D. S. Kirschen, Decentralized demand-side contribution to primary frequency control, IEEE Trans. Power Systems, vol. 26, no.1, Feb. 2011, pp. 411-419. F. Lasnier and T. G. Ang, Photovoltaic engineering handbook, Institute of physics publishing Ltd., 1990. Mohamed, R.G., Ibrahim, D.K., Youssef, H.K., Rakha, H.H., Optimal sizing and economic analysis of different configurations of photovoltaic systems, (2014) International Review of Electrical Engineering (IREE), 9 (1), pp. 146-156. C. Lo and K. Kwan, Economic dispatch and optimal sizing of battery energy storage systems in utility load-leveling operations, IEEE Trans. Energy Conversion, vol. 14, no.3, 1999, pp. 824829. F. Chacra, P. Bastard, G. Fleury, and R. Clavreul, Impact of energy storage costs on economical performance in a distribution substation, IEEE Trans. Power Systems, vol. 20, no. 2, pp. 684691.

Authors’ information Thongchart Kerdphol received his B.Eng., M. Eng from Kasetsart University, Bangkok, Thailand in 2010 and 2012. Presently, he is pursuing in D. Eng. degree at Kyushu Institute of Technology. He is currently a research assistant at the Department of Electrical Engineering and Electronics, Kyushu Institute of Technology (KIT), Japan. His research interests are in the area of the application of artificial intelligence in power systems, power system stability, Smart Grid, renewable energy, power system dynamics and controls. He is a member of IEEE Yaser Soilman Qudaih graduated from the University of Engineering and Technology, Lahore, Pakistan in 1996 as an electrical engineer. He Completed his M.Sc. and PhD from Kumamoto University, Japan in Electrical Engineering. He is currently an assistant professor at the Department of Electrical Engineering and Electronics, Kyushu Institute of Technology (KIT), Japan. His area of interest including power system is renewable energy and Smart Grid Applications. He is a member of IEEE. Yasunori Mitani received B.Sc., M.Sc. and D. Eng. degrees in Electrical Engineering from Osaka University, Japan in 1981, 1983 and 1986, respectively. He was a Visiting Research Associate at the University of California, Berkeley, from 1994 to 1995. He is currently a professor at the Department of Electrical Engineering and Electronics, Kyushu Institute of Technology (KIT), Japan. At present he is the Head of Environmental Management Center of KIT. His research interests are in the areas of analysis and control of power systems. He is a member of the Institute of Electrical Engineers of Japan and IEEE.

Copyright © 2015 Praise Worthy Prize S.r.l. - All rights reserved

International Review of Electrical Engineering, Vol. 10, N. 2

285