Journal of Modern Power Systems and Clean Energy

ISSN 2196-5625 CN 32-1884/TK

网刊加载中。。。

使用Chrome浏览器效果最佳,继续浏览,你可能不会看到最佳的展示效果,

确定继续浏览么?

复制成功,请在其他浏览器进行阅读

Feasibility of Grid-connected Solar-wind Hybrid System with Electric Vehicle Charging Station  PDF

  • Shakti Singh
  • Prachi Chauhan
  • Nirbhow Jap Singh
Department of Electrical and Instrumentation Engineering, Thapar Institute of Engineering and Technology, Patiala, India

Updated:2021-03-16

DOI:10.35833/MPCE.2019.000081

  • Full Text
  • Figs & Tabs
  • References
  • Authors
  • About
OUTLINE

Abstract

Recently, renewable power generation and electric vehicles (EVs) have been attracting more and more attention in smart grid. This paper presents a grid-connected solar-wind hybrid system to supply the electrical load demand of a small shopping complex located in a university campus in India. Further, an EV charging station is incorporated in the system. Economic analysis is performed for the proposed setup to satisfy the charging demand of EVs as well as the electrical load demand of the shopping complex. The proposed system is designed by considering the cost of the purchased energy, which is sold to the utility grid, while the power exchange is ensured between the utility grid and other components of the system. The sizing of the component is performed to obtain the least levelized cost of electricity (LCOE) while minimizing the loss of power supply probability (LPSP) by using recent optimization techniques. The results demonstrate that the LCOE and LPSP for the proposed system are measured at 0.038 $/kWh and 0.19% with a renewable fraction of 0.87, respectively. It is determined that a cost-effective and reliable system can be designed by the proper management of renewable power generation and load demands. The proposed system may be helpful in reducing the reliance on the over-burdened grid, particularly in developing countries.

I. Introduction

THE availability of limited fossil fuel reserves, climatic effects, and greenhouse gas emissions has compelled the focus toward renewable power generation in the past years. According to the Ministry of New and Renewable Energy, in India, renewable energy generation accounts for approximately 20% of the total power generation. The major contributions are from solar and wind power rather than hydro, thermal, ocean, and biomass energies [

1]. Solar and wind power become prominent technologies to supply electrical load in remote and off-grid locations owing to the recent advancements in power electronic storage devices and the falling prices of the components. Reference [2] stated that solar and wind energies are freely accessible in nature and are more advantageous than other sources for power generation. However, the combination of solar and wind energies increases the complications due to their unpredictable and variable nature, which presents certain technical, environmental, and socio-economic challenges [3]. A hybrid renewable energy system can be designed using mathematical models of different components and optimization techniques [4], [5].

In recent years, extensive research has been conducted to design a solar-wind hybrid system. Reference [

6] developed a solar-wind hybrid system to generate electricity, and concluded that a hybrid system could reduce the fluctuations and complexity of energy. Reference [7] simulated a solar-wind hybrid system to compute the levelized cost of electricity (LCOE) in Tripura, a north-eastern state of India. LCOE was often used to compare power generation technologies. Reference [8] developed a mathematical model of a solar-wind hybrid system to compute the output power and conducted reliability analysis to determine the performance of the system for on-grid and off-grid purposes. Reference [9] determined the size of a solar-wind hybrid system using a genetic algorithm (GA) to minimize the loss of power supply probability (LPSP) and annualized cost of the system (ACS). LPSP signified the maximum allowable limit of power loss to obtain a reliable system. ACS was another useful indicator to compare the costs of different systems because it measured their relative contributions to the total net present cost (NPC) of the system.

Reference [

10] determined the size of a solar-wind hybrid system considering LCOE and LPSP. On the basis of LCOE and LPSP, to determine the optimal configuration, [11] performed economic and technical analyses of a hybrid energy system based on solar power, wind power, and battery. Reference [12] conducted an economic analysis of an electric vehicle (EV) charging station powered by solar energy. Further, it concluded that solar photovoltaic (SPV) panels could satisfy the EV demand with the LCOE of the system measured at 0.097 $/kWh at an interest rate of 6%. Reference [13] optimized the annual cost of investment and maintenance of an EV charging station based on a solar-wind hybrid system using a teaching-learning-based optimization algorithm. Reference [14] conducted an economic analysis of a solar and wind based DC microgrid that supplied power in remote areas and minimized the LCOE of the system. Further, the LPSP and excess electricity were calculated. Reference [15] calculated the NPC of an EV charging station based on solar-wind hybrid power in rural areas and selected the best optimal configuration. Detailed literature on solar-wind hybrid systems can be obtained in [16]-[19].

In recent years, a significant growth of EVs and plug-in hybrid EVs has been observed in the existing power system. The integration of EVs in terms of economic analysis and power management in the renewable energy based system is an interesting and challenging research area. Reference [

20] presented a comprehensive review on the research and advancement of EVs and their interaction with the grid. The feasibility of SPV-wind hybrid energy system was also discussed. The charging and discharging of EVs from a grid or microgrid result in serious problems such as high power loss, voltage deviations, and power quality problems [21].

Reference [

22] emphasized that the evolution of EVs was owing to the increasing pollution from gasoline-operated vehicles and the recent development in the battery, drive motor, and charging methodology of EVs. Reference [23] developed a solar-wind hybrid system for charging EVs with simulation models. Reference [24] presented a comprehensive review based on a solar charging station and discussed different modes of charging EVs. Reference [25] proposed a hybrid solar and wind power charging station, which was implemented in Tangshan, China and developed an LPSP technique to determine the capacities of the components. Reference [26] tracked the maximum power point that can provide the maximum power to the EVs and also interfaced a DC microgrid with bi-directional DC-DC converters. Reference [27] designed an EV charging station to reduce the cost, while considering the emissions and comparing an isolated EV charging station with a grid-connected charging station. However, there were no analysis on the cost of the hybrid system. Reference [28] proposed a charging station at a work place in the Netherlands based on solar and grid-connected stations. Different scenarios for charging EVs were proposed. However, an economic analysis was not included in the work. Reference [29] performed an analysis of the demand response programs and charging of EVs by integrating them in the buses of radial distribution networks. The reliability of radial distribution systems was further elaborated by using a particle swarm optimization (PSO) algorithm with the objective of minimizing the loss of load expectation.

A variety of optimization methods such as GA [

9], [30], artificial bee colony (ABC) [31], PSO [32], invasive weed optimization [33], multiplayer harmony search [34], dynamic nonlinear programming [35], biogeography-based optimization algorithm [36], simulated annealing [37], improved harmony search [38] were reported in the literature to solve various problems of engineering design. Among various available optimization algorithms, the PSO algorithm had the least mathematical complexity, whereas the ABC algorithm demonstrated the advantage of having less control parameters [39].

Researchers primarily concentrate on the control and power management of the EVs in the microgrid or grid-connected systems. However, an economic analysis considering the power exchange with the grid is one of the major parameters that must be addressed. The rapid inclusion of EVs presents both challenges and opportunities to the existing power system. The major parameters in designing a renewable energy based system include a smooth power flow between the components, reliability, and ACS of the system. In this paper, a small grid-connected solar-wind hybrid system with EVs is proposed for a small shopping complex located in a university campus in the state of Punjab in India. The primary focus is to formulate a mathematical model of a solar-wind hybrid system incorporating EVs with a grid as the backup. Further, this paper aims to minimize the power exchange with the grid. The optimal configuration of the proposed system while minimizing the ACS is performed using the ABC and PSO algorithms. A detailed comparison of the results with both algorithms is performed. A parameter sensitivity analysis is performed to analyze the impact of the algorithm parameter on the solution quality of both algorithms. Finally, a sensitivity analysis is performed to analyze the impact of the maximum grid sales and the purchase capacities on LCOE.

The remaining paper is organized into five sections. Section Ⅱ explains the mathematical model along with the working methodology of the proposed solar-wind hybrid system. Section Ⅲ considers the objective function and constraints of the system. Section Ⅳ presents a description of the algorithms. Section Ⅴ presents the outcomes and discussions. The conclusion of the work is presented in Section Ⅵ.

II. Mathematical Model

A schematic diagram of the proposed grid-connected solar-wind hybrid system is illustrated in Fig. 1, which is a hybrid AC/DC system. In Fig. 1, PCC stands for the point of common coupling. The DC bus is connected with wind turbines, SPV panels, and a charging station, while the electrical load is connected to the AC bus. A dual converter is proposed to convert AC into DC and vice versa. The power management is more crucial in a hybrid AC/DC system compared to a single AC or DC system. Therefore, a microgrid controller is also proposed to monitor a smooth power flow between various components of the system. The detailed mathematical models of each component in the system are discussed subsequently.

Fig. 1 Schematic diagram of proposed system.

A. Wind Turbines

The power produced by a wind turbine depends on the area through which the wind passes and the velocity of the wind. The power obtained from a wind turbine can be calculated as:

Pwt(t)=0                            V(t)Vcin  or  V(t)VcoffPrwV(t)-VcinVr-Vcin    Vcin<V(t)<VrPrw                         VrV(t)<Vcoff (1)

where V(t), Vcin, Vcoff, Prw, Vr are the wind speeds at the required height, cut-in speed, cut-off speed, rated power of the wind turbine, and rated wind speed, respectively [

40].

Wind speed at different hours of the year is represented in the form of a curve called probability density function (PDF). If a curve is plotted by considering the variable wind speeds, the area under the curve between any two wind speeds is equal to the probability of which the wind is between the two speeds. It can be expressed as:

P(v1vv2)=v1v2fvdv (2)
P(0v)=0fvdv=1 (3)

where fv is the wind speed PDF, i.e., density; and v1 and v2 are any two wind speeds. In wind speed statistics, the most important PDF is Weibull probability function. The Weibull probability function is defined as the starting point for characterizing the statistics of wind speed, and its expression is given as:

fv=kcvck-1exp-vck (4)

where k, c, and v are the shape parameter, scale parameter, and wind speed, respectively.

Wind speed is different over different surfaces of the earth, e.g., over a calm sea and a forest, there are different wind speeds. The variation is based on the height at which the wind speed is measured. The speed of wind at a particular height is calculated as:

V=VoHWTHoα (5)

where V, Vo, and α are the wind speed at a required height HWT, the wind speed at reference height Ho, and the friction coefficient, respectively. The friction coefficient depends on the terrain over which the wind blows. Generally, an approximated value of α is considered to be 1/7.

The total power generated PWT(t) from wind turbines is calculated as:

PWT(t)=NWTPwt(t) (6)

where NWT is the number of wind turbines.

B. SPV Panel

An SPV panel is used to generate the power by harnessing solar energy. Only a small amount of solar radiation falling on the SPV panels is converted into electricity, and the rest is converted to heat. Therefore, the effect of solar radiation along with temperature on output power is considered. The output of an individual SPV panel is given as:

Psol(t)=PmSm1000[1-λ(Tcell-25)] (7)

where Sm is the solar radiation.

The cell temperature and variation of power with respect to the changes in temperature is given by:

Tcell=Tamb+NOCT-200.8Sm (8)

where Tcell, Tamb, and NOCT are the percentage drop in power corresponding to the temperature of the cell at which power is to be calculated, ambient temperature, and cell temperature when ambient temperature is 20 ℃ and solar radiation is 0.8 kW/m2, respectively.

The rated maximum power output Pm is calculated as:

Pm=VmIm (9)

where Vm and Im are the maximum voltage and current ratings, respectively. Further, the total power generated from SPV panels PPV(t) is calculated as:

PPV(t)=NPVPsol(t) (10)

where NPV is the number of SPV panels.

C. EV Charging Station

The charging station is comprised of a dual converter, charging bays, and EVs. To control the power flow, the charging station is connected to a microgrid controller, which helps control the direction of the power flow at a particular time. EVs can be charged according to their state of charge (SOC), which is defined as the ratio of the available capacity to its maximum capacity when a battery is completely charged. Thus, it describes the remaining charging percentage of the battery.

Mathematically, the practical constraints imposed on the charging of EVs are described as:

SOCminSOC(t)SOCmax (11)
Crate(t)Cratemax (12)

where SOCmin, SOCmax, and SOC(t) are the minimum, maximum, and current SOC values for a particular EV at a particular time t, respectively; and Crate(t) and Cratemax are the present charging rate and maximum allowed charging rate of EVs, respectively. At any instant t, the power of the EVs at the charging station is calculated as:

PEV(t)=EmaxSOC(t)100Δt (13)

where Emax is the maximum energy capacity of the vehicle; and Δt is the time interval considered as one hour. The difference in charging requirements of EVs is determined by comparing their SOC with SOCcr, which is the critical SOC. In this paper, a level 2 charger with advantages of easy and user-friendly charging is considered to charge 20 EVs at a charging station.

D. Power Converter

PV panels and wind turbines feed DC power to EVs at the charging station. However, to satisfy load demand or supply power to the grid or vice versa, converters are required. The size of the converter is chosen according to the maximum grid sales Pgsmax and purchase capacity Pgpmax. The rated power of the inverter Pinv(t) and that of the rectifier Prect(t) are calculated as:

Pinv(t)=Pgsmax/ηinv (14)
Prect(t)=Pgpmaxηrect (15)

where ηinv and ηrect are the efficiency values of the inverter and rectifier, respectively.

III. Operational Strategy and Objective Function

To ensure that the solar-wind hybrid system satisfies the load power demand, as shown in Fig. 1, NPV and NWT must be determined. The objective function is formulated by considering the reliability and LCOE.

A. Reliability

A system is considered reliable if it has sufficient power to satisfy the load demand, which depends on LPSP. Two types of load demands are required to be satisfied: the first is the load demand of EVs PDEV(t) and the second is the AC load demand PAC(t). PDEV(t) at time interval t is calculated as:

PDEV(t)=EmaxSOCcrΔt-PEV(t) (16)

In this paper, it is considered that the vehicles having SOC less than critical SOC are charged before 11:00 a.m.. After 11:00 a.m., the priority is given to all EVs owing to sufficient solar power availability. Further, it is assumed that all EVs arrive at 08:00 a.m. and park throughout the day. The second load demand is PAC(t). These demands are satisfied by two sources, i.e., SPV panels and the wind turbine. The difference between the power generated and the power demand is calculated as:

ΔP(t)=PPV(t)+PWT(t)-PDEV(t)-PAC(t)/ηrect (17)

If the sources are unable to satisfy the load demand, the power is purchased from the grid Pgp(t) to fulfill the requirements. Moreover, if more power is available from the sources after fulfilling the demand, the power is sold to the grid Pgs(t). However, there are limits on purchasing and selling power to the grid, which are defined as the maximum grid purchase capacity Pgpmax(t) and maximum grid selling capacity Pgsmax(t). The power cannot be purchased from the grid or sold to the grid beyond these limits. Therefore, depending on ΔP(t), the cases are formed as follows.

1) If ΔP(t)>0, the total power output obtained from SPV panels and wind turbines is sufficient to charge the EVs and fulfill the requirement of the electric load. Further, the available extra power Pgs(t) is sold to the grid, which is computed as:

Pgs(t)=(PPV(t)+PWT(t)-PDEV(t))ηinv-PAC(t) (18)

2) If ΔP(t)>0 and ΔP(t)>Pgsmax(t), the extra power is supplied to the dump load. Dump load power Pd(t) is calculated as:

Pd(t)=PPV(t)+PWT(t)-PDEV(t)-(PAC(t)+Pgsmax(t))/ηrect (19)

3) If ΔP(t)<0, the power generated from SPV panels and wind turbines is not able to satisfy the power demand of EVs and the electric load. Therefore, the required power is purchased from the grid, which is given as:

Pgp(t)=(PDEV(t)-PWT(t)-PPV(t))/ηrect+PAC(t) (20)

4) If ΔP(t)=0, there is no exchange of power from the grid, and the demand of EVs and electric load is satisfied by the power generated from SPV panels and wind turbines.

When Pgp(t)>Pgpmax(t), both sources and grid are unable to satisfy the load demand. Consequently, power deficiency Pdef(t) occurs, which is calculated as:

Pdef(t)=Pgp(t)-Pgpmax(t) (21)

Pdef(t) must be zero to ensure that the total load demand is served reliably when minimizing the LPSP. Mathematically, LPSP can be calculated as:

LPSP=t=18760Pdef(t)t=18760(PAC(t)+PDEV(t)) (22)

To solve the optimal sizing problem, the LPSP can be maintained within a specific tolerance band. In this paper, the maximum limit of LPSP is considered to be 1%. The energy management algorithm is demonstrated through a simplified flowchart to calculate the power provided to the dump load, power purchased from the grid, and power sold to the grid, as shown in Figs. 2 and 3.

Fig. 2 Flowchart of charging methodology for EVs at charging station.

Fig. 3 Flowchart for operation strategy of proposed system.

B. LCOE

The main objective of this paper is to achieve power exchange between various components of the system and minimize LCOE of the overall proposed system. The decision variables are NPV and NWT required to maintain minimum LPSP and ACS. The ACS includes the costs of installing SPV panels and a wind turbine, costs of energy purchased and sold back to the grid, and costs of the converters.

ACS=F(NPVCPV+NWTCWT+CgpEgp-CgsEgs+Cconv) (23)

where CPV and CWT are the costs of SPV panels and wind turbines, respectively; Cgp and Cgs are the costs of energy purchased and sold to the grid, respectively; Egp and Egs are the amounts of energy purchased and sold to the grid, respectively; and Cconv is the cost of the converters. Further, CPV and CWT are calculated as:

CPV=CacpPV+CarpPV+CmPV+CsvPV (24)
CWT=CacpWT+CarpWT+CmWT+CsvWT (25)

where Cacp is the annual capital cost; Carp is the cost of replacing the components; Cm is the cost of operation and maintrnance; and Csv is the salvage value; and the superscripts PV and WT denote the SPV and wind turbine, respectively.

1) Annual Capital Cost

In the annual capital cost, the installation and purchasing costs of the elements are also included. The annual capital cost of SPV panels and wind turbines are calculated by considering capital recovery factor (CRF) as:

CacpPV=CcpPVCRF(i,n) (26)
CacpWT=CcpWTCRF(i,n) (27)
CRF(i,n)=i(1+i)n(1+i)n-1 (28)

where CcpPV and CcpWT are the initial capital costs of SPV panels and wind turbines, respectively; and n and i are the project life time and the annual interest rate, respectively.

2) Annual Replacement Cost

Annual replacement cost includes the cost of replacing SPV panels and wind turbines if their life time is less than that of the project. The total annual cost of replacing the SPV panels and wind turbine is calculated as:

CarpPV=CrpPVCRF(i,n)1(1+i)y (29)
CarpWT=CrpWTCRF(i,n)1(1+i)y (30)

where CrpPV and CrpWT are the costs of replacing the SPV panel and wind turbine, respectively; and y is the life time of the panel and wind turbine in years.

3) Maintenance Cost

The maintenance cost constitutes the labor cost, cleaning cost, and cost of repairing in case of any temporary damage. The maintenance costs of SPV panels and wind turbines are calculated as:

CmPV=NPVCm1PV (31)
CmWT=NWTCm1WT (32)

where Cm1PV and Cm1WT are the maintenance costs of SPV panel and wind turbine, respectively.

4) Salvage Value

The cost of the remaining life for the component after the project ends is known as salvage value. The salvage values of a wind turbine and SPV panel are calculated as:

CsvWT=CrpWTRremy (33)
CsvPV=CrpPVRremy (34)

where CrpWT and CrpPV are the replacement costs of a single wind turbine and SPV panel, respectively; and Rrem is the remaining life.

5) Cost of Exchanging Power

The total amount of electricity purchased Egp and sold back to the grid Egs can be calculated as:

Egp=t=18760Pgp(t) (35)
Egs=t=18760Pgs(t) (36)

The cost of electricity purchased and sold can be calculated as:

Cgp=EgpCgP (37)
Cgs=EgsCgs (38)

where CgP and Cgs are the unit costs of electricity purchased and sold back to grid, respectively.

Moreover, the cost potency of a system can be determined by the objective function LCOE, which is the average cost of energy obtained from the system. It can be calculated as:

LCOE=ACSTtotal (39)

where Ttotal is the total energy served.

The constraints of the objective functions are:

1NPVNPVmax (40)
1NWTNWTmax (41)
PgpPgpmaxPgsPgsmax (42)
0LPSP1% (43)

where NPVmax and NWTmax are the maximum numbers of SPV panels and wind turbines, respectively; and Pgpmax and Pgsmax are the maximum grid purchase and sale capacities, respectively. LPSP is maintained within limits while minimizing the objective function using the exterior penalty method.

IV. Solution Methodology

The proposed hybrid renewable energy system consists of more than one energy source. Therefore, this problem has multiple decision variables resulting in complex optimization problems. This problem requires the identification of energy sources for uninterrupted power supply to the EV charging station and AC load. Hence, the optimization problem incorporates economic objectives. Moreover, it requires the computation of long-term system performance to achieve the best compromise between LPSP and LCOE. LCOE is minimized using ABC and PSO algorithms as they dynamically search for an optimum system configuration while maintaining LPSP within desired limits.

A. ABC Algorithm

This algorithm explores the idea of the ABC algorithm. There are three types of bees in the ABC algorithm: employed, scout, and onlooker bees. In ABC algorithm, there are three steps in each cycle: ① employed bees search for food source or nectar amount and its location is stored in the memory; ② onlooker bees collect the information from employed bees and make the decision for selecting the best food source by doing the quality search, i.e., measuring the nectar amount of food source; ③ scout bees search for new food sources. Half of the population of bees or swarms are employed bees and the rest half are onlooker bees. For each employed bee, there is only one food source. When the food source is abandoned by the employed bees, the new food source location is randomly determined by the scout bees, and it is replaced with the abandoned food source [

39], [40].

The steps for the implementation of ABC algorithm can be summed up as presented in Fig. 4, where NS denotes the dimension of population; NP denotes the colony and population sizes; and ITmax denotes the maximum cycle and iteration.

Fig. 4 Flowchart for proposed ABC algorithm.

B. PSO Algorithm

PSO contains a swarm of particles. Swarm indicates that the population and particles are candidate solutions. By optimizing the objective function, the fitness of each solution is calculated iteratively.

These candidate solutions move around the search space and their movement is directed by the swarm. And the best position in the search space becomes the upgraded position, which directs the swarm. This procedure is repeated until the best solution is obtained. At the ITth iteration, the particle position p is denoted as xpIT, which moves in the search space with velocity vpIT. Therefore, the upgraded position of the particle is calculated as:

xpIT+1=xpIT+vpIT+1 (44)

The upgraded velocity is given by:

vpIT+1=W[vpIT+C1rand(0,1)(ppIT-xpIT)+C2rand(0,1)(gpIT-xpIT)] (45)

where C1 and C2 are the acceleration parameters; rand(0,1) is the uniformly distributed random numbers; ppIT and gpIT are the p-best and g-best, respectively; and W is an inertial weight factor, which is used for managing the capability of the search [

41]. The steps for the implementation of PSO algorithm are summed up in Algorithm 1.

Algorithm 1  : PSO pseudo code

Input: Sm,V(t),  Pgsmax,  Pgpmax, ηrect

Input: PDEV(t),  PAC(t)

Input: NS,  NP,  ITmax

Initialize swarm members xp,  vp,  IT0

 while IT<ITmax do

  p1

  while p<NP do

   while t8760,  IT1 do

    Compute PPV(t),  PWT(t) using (10) and (6)

    tt+1

    Power management as per flow chart shown in Fig. 3

   end while

   f(p)LCOE(p)+rkLPSP(p)

   pp+1

  end while

  XpBXmin(f(XpB),  f(xp)); p=1, 2, ..., NP

  Update swarm velocity and position using (44) and (45)

  ITIT+1

 end while

The system configuration is optimized by employing both the algorithms to determine the optimal configuration that minimizes LCOE. For each system configuration observed in the search process, LPSP is determined. Therefore, for the desired LPSP, the optimal configuration can be selected by obtaining the minimum LCOE while satisfying the maximum LPSP target of 1%.

V. Results and Discussions

Figure 1 shows the sizing of different components for the system, which satisfies the charging requirements of EVs and optimizes the load demand of the shopping complex. The natural resource and electrical data of a small shopping complex located in a university campus is selected. The availability of solar radiation, wind speed, and a histogram representing wind speed frequency at the site is shown in Fig. 5. The selected data show that the site has an average wind speed of 5.9 m/s and an average solar radiation of 5.14 kWh/m2 per day. For this particular geographical location, the data of the entire year are created by using the available resource data on the National Aeronautics and Space Administration website with hybrid optimization model for multiple energy resources software [

42], [43]. The data of electricity demand of the shopping complex for one year are considered in this paper. And the shopping complex is open during the whole week. Figure 6 demonstrates the load demand of the shopping complex throughout the year, considering a 10% day-to-day load variation and 10% time-step-to-time-step variation. The hourly load demand patterns of weekends and weekdays in the summer and winter seasons are illustrated in Fig. 6(a), which demonstrate a peak load of 6.50 kW and 9 kW, respectively. During weekdays, the load demand is comparatively less than that on weekends, i.e., the peak load is 5.60 kW in winters and 8.50 kW in summers. Throughout the year, the average load demand is equal to 110 kWh per day and the overall average is measured at 4 kW. The peak demand is observed to be 12 kW and the load factor is 0.382. The load demand of an entire year is shown in Fig. 6(b) [44].

Fig. 5 Availability. (a) Solar insolation throughout the year. (b) Wind speed throughout the year. (c) Histogram representing wind speed frequency at site.

Fig. 6 Load profile. (a) During winter and summer (weekdays and weekends). (b) Throughout year after 10% variation.

Table Ⅰ lists different costs and specifications associated with the components. The system is proposed for a life time of 20 years. For time value of money, an interest rate of 6% is considered. The life time of SPV panels, wind turbines, and the inverter is considered to be equal to the project life time. Therefore, no replacement is required. The simulation is conducted in MATLAB 2017a, considering the complete data of one year at 1-hour interval to calculate the power exchange and costs of the hybrid system. Moreover, the numbers of SPV panels and wind turbines are optimized by using ABC and PSO algorithms. The different parameters used in optimization technique are listed in Table Ⅱ.

Table I Specification and Cost of System Component
ComponentParameterValue
SPV panel Maximum power Pmax 100 W
Maximum voltage Vmp 18 V
Maximum power current Imp 5.56 A
Open circuit voltage Voc 22.3 V
Short-circuit current Isc 6.1 A
Number of cells 36
Nominal operation cell temperature 45
Capital cost and replacement cost 1084 $/kW
O&M cost 5 $/year
Life time 20 year
Wind turbine Rated power 1 kW
Capital and replacement cost 1098 $/kW
O&M cost 2 $/kW/year 
Cut-out speed Vco 20 m/s
Cut-in speed Vcin 5 m/s
Hub height 50 m
Life time 20 year
Others DC bus voltage Vbus 120 V
Project life N 20 year
Interest rate i 6 %
Converter
Rated power 1 kW
Rectifier and invert efficiencies 90 %
Capital and replacement cost 127 $/kW
O&M cost 1 $/year
Life time 20 year
EV battery specification Battery ampere-hour 210 Ah
Battery type and variant Lithiumion
Number of modules 16
Number of cells 48
Battery energy capacity 5 kWh
Maximum charging rate 0.5 kWh/h
Maximum number of vehicles/bays 20
Table II Parameter of ABC Algorithm
ParameterValue
Colony size 20
Dimension of colony 2
Maximum cycle 100
Employed bees (food number) NP/2
Limit 100
Population size 20
Dimension of population 2
Maximum iterations 100
Inertia weight 0.4-0.9
Weighting factor 2

Table Ⅲ lists the optimized results of the solar-wind hybrid system using both ABC and PSO algorithms. It is inferred from the results that by fixing the grid sales and purchase capacity to 10 kW, the ACS using ABC and PSO algorithms are 2618.3 $/year and 2953.2 $/year, respectively. Moreover, the LCOE using ABC and PSO algorithms are 0.038 and 0.043 $/kWh, respectively. It can be deduced from the table that ABC algorithm provides an acceptable solution as its LCOE is comparatively less. Figure 7 shows the convergence behaviors of both the algorithms, which can be observed in 10 initial iterations.

Fig. 7 Comparison of convergence characteristics of ABC and PSO algorithms.

Table III Optimal Sizing Result for Proposed System
AlgorithmSPV capacity (kW)Wind capacity (kW)Converter capacity (kW)Purchase capacity (kW)Sales capacity (kW)NPC ($)ACSLCOE
ABC 36 20 20 10 10 30021.79 2618.3 0.0380
PSO 41 16 20 10 10 33874.74 2953.2 0.0430

Table Ⅳ lists the energy production and consumption by all components of the proposed system for the configurations using ABC and PSO algorithms. In the case of energy consumption, the EV demand accounts for 25.2% of the total consumed energy, while the shopping complex accounts for 35.7% and grid sales accounts for 39%. The excess electricity is equal to 27976 kWh/year using ABC algorithm and 23488 kWh/year using PSO algorithm, respectively. The excess electricity is greater by using ABC algorithm than PSO algorithm, which is provided to the dump load. Further, LPSP from both algorithms is maintained at 0.19%.

It is inferred from the results that the ABC algorithm provides better results. Therefore, the configuration obtained from the ABC algorithm is selected. Table Ⅴ lists the annualized cost analysis using the ABC algorithm. The total ACS is equal to 2618.3 $/year, which is obtained by adding the annualized capital cost of SPV panels, wind turbines, and inverters while subtracting the cost of annual grid sales. The annualized cost of components is calculated by using CRF. Further, it can be deduced from Table V that the initial costs of 36 kW SPV panels and 20 kW wind turbines is equal to $38880 and $21960, respectively. Figure 8 illustrates the complete monthly power balance throughout the year. In January, there is a deficiency in power generation with increased grid purchase.

Table V Cost Division Achieved by ABC Algorithm
Component

Initial

capital ($)

Capital ($/year)Maintenance cost ($/year)

Total

($/year)

SPV panel 38880 3389.55 144.00 3533.55
Wind turbine 21960 1914.47 40.00 1954.47
Grid purchase 0 0 625.66 625.66
Grid sale 0 0 -3736.80 -3736.80
Converter 2540 221.43 20.00 241.43
Total system 63380 5525.45 -2907.14 2618.31

Fig. 8 Monthly power generation and consumption for whole year.

In certain months such as April, August, and September, better solar and wind power generation is achieved, resulting in more grid sales and less grid purchases.

Table IV Production and Consumption of Electricity for a Complete Year
EnergyComponentABCPSO
Production(kWh/year)Consumption (%)Production(kWh/year)Consumption (%)

Energy

production

SPV panel (DC) 60015 40.2 68350.0 48.18
Windturbine (DC) 80010 53.7 64008.0 45.12
Gridpurchase (AC) 8938 6.0 9491.6 6.69
Total system 148963 100.0 141849.6 100.00

Energy

consumption

EV load (DC) 28458 25.2 28458.0 25.77
Electric load (AC) 40291 35.7 40291.0 36.49
Grid sale 43962 39.0 41647.0 37.72
Total system 112711 100.0 110396.0 100.00

For better understanding of energy management in the system, two days are selected, i.e., one in winter and the other in summer. Figure 9(a) demonstrates the total power balance of one day in winter. Before 08:00 a.m., there is no charging demand for EVs and no solar and wind power generation. However, to fulfil the demand of the shopping complex, power is purchased from the grid. During the daytime hours from 11:00 a.m. to 02:00 p.m., the maximum EV demand is demonstrated and the power generation and grid sales are also at the maximum value with no power purchased from the grid. In the evening, the power purchased from the grid is at the maximum value as there is no power availability from solar and wind resources. For example, at 11:00 a.m., the electrical demand is 3.086 kW, EV demand is 18.39 kW, and wind and solar power generation values are 12.83 and 23.35 kW, respectively. The total demand is less than the generated power. Thus, the system sells 10 kW power to the grid, which is determined by the limitation in maximum sales capacity. And there is an excess power of 3.26 kW.

Fig. 9 One day power balance in proposed system. (a) Winter. (b) Summer.

Figure 9(b) demonstrates the total power balance of one day in summer. Because of adequate solar radiation and wind speed, there is more power generation in summer than that in winter. From 01:00 a.m. to 06:00 a.m., there is no power generation from solar and wind systems. Therefore, the demand of shopping complex is satisfied by the grid. During the daytime, the maximum solar power generation is achieved, which is sufficient to fulfil EV demand. Moreover, the excess power is sold to the grid. For example, at 01:00 a.m., the shopping complex and EV demands are 6.546 and 15.25 kW, respectively. Owing to adequate solar and wind resources, the generated power are 28.84 and 13.2 kW for solar and wind systems, respectively. The total generation is greater than the demand, therefore, the system have extra power to sell to the grid. The total available power that can be sold to the grid is 18.40 kW. However, owing to the constraint on the maximum grid sales capacity of 10 kW, only 10 kW power is sold to the grid and the remaining 8.40 kW is supplied to the dump load. It is further evident from both the figures that there is no unsatisfied load. All the energy demands are fully satisfied by renewable power generation.

In a grid-connected system, the grid power exchange capacities, i.e., maximum sales and purchase capacities, are two major parameters. To analyze the operation of the proposed system, the effect of grid capacity verses LCOE is evaluated. Firstly, by maintaining the maximum grid purchased capacity at 10 kW, the grid sales capacity varies and LCOE is observed. Figure 10 demonstrates the variation of maximum grid sales and purchase capacities with LCOE.

Fig. 10 Variation of levelized cost of energy with maximum grid sales and purchase capacities.

It can be observed that LCOE decreases significantly as the maximum grid sales increases. Further, by maintaining the maximum grid sales capacity, i.e., 10 kW, the maximum grid purchase capacity varies. Moreover, the results are plotted in Fig. 10. The results clearly depict that with increasing grid purchase capacity, the LCOE of the system decreases.

A. Special Case: Uncertainty in Behavior of EVs

The above case study is proposed considering a time limit in the arrival and departure of EVs at the charging station within office hours. To prove the efficiency of the proposed model, another case study is analyzed and presented in this subsection. A more generalized and flexible, 24×7 charging station is considered. Further, it is assumed that the arrival and departure time of the EVs is uncertain. No charging preference is given to any of the EVs, and it is assumed that the charging occur as per the user requests. The maximum grid sales and purchase capacities are maintained similar to the previous case study.

Table VI lists the results obtained using both algorithms. It is evident from the results that the LCOE demonstrate a marginal increase compared to the previous case. The system sells 11957 kWh/year power to the grid compared to 43962 kWh/year in the previous case, and purchases 19168 kWh/year from the grid compared to 8938 kWh/year in the previous case. The system in the previous case study utilizes more solar power, owing to more power generation during the daytime.

Table VI Optimal Sizing Result for Proposed Hybrid System in a Special Case
AlgorithmSPV capacity (kW)Wind capacity (kW)Converter capacity (kW)Purchase capacity (kW)Sales capacity (kW)NPC ($)ACSLCOE ($/kWh)
ABC 5 10 20 10 10 23893 2083.12 0.0399
PSO 6 9 20 10 10 25885 2256.83 0.0432

Figure 11(a) and (b) demonstrates the power balance for the two days, i.e., one in winter and the other in summer.

Fig. 11 One-day power balance. (a) Winter. (b) Summer.

The power balance shows that the EV load is distributed throughout the day and night compared to the previous case study. The second case is more generic and can be applied to any charging station powered by renewable sources, and the grid can be treated as a backup. LCOE is competitive in both case studies. However, the system will be more grid dependent rather than depending on renewable power sources in the second case.

B. Parameters Considered for Sensitivity Analysis

The performance of metaheuristic algorithms is dependent on the control parameter. Therefore, to verify the effectiveness of the considered algorithms, a comparison of the results of the PSO and ABC algorithms is performed with respect to variations in different control parameters. In both algorithms, the values of the common parameters used such as population size and total evaluation number are chosen to be the same, i.e., 20 and 100, respectively. The other specific parameters considered for sensitivity analysis in the case of PSO are cognitive and social components, which are represented in this paper as C1 and C2, respectively. In the experiments, the cognitive and social components are both set to be 2.0. In the case of the ABC algorithm, excluding the common parameters, only one control parameter is considered as Limit. The aforementioned parameters vary by ±10% of the standard parameter values proposed in the literature [

45]. Table Ⅶ lists the experimental study undertaken for parameter sensitivity analysis of the considered problem. The obtained solution depicts small variations around the mean value. Moreover, the relative deviation from the minimum ACS is less than 1%. A similar trend is observed in both algorithms.

Table VII Analysis of Parameter Sensitivity
AlgorithmParameterScopeStepACS
Mean ($/year)

Minimum

($/year)

Maximum

($/year)

Standard deviation

($/year)

Variation of ACS (%)
PSO C1 0-4 0.2 2957.72 2953.2 2963.0 3.259 0.1101
C2 0-4 0.2 2957.62 2953.2 2967.8 3.325 0.1124
ABC Limit 0-200 20.0 2629.39 2618.3 2639.0 5.975 0.2272

VI. Conclusion

Hybrid renewable energy systems have the potential to provide electricity to grid and off-grid locations economically and reliably. The efficiency of the system is enhanced if the renewable energy systems are integrated with the grid. In this paper, a detailed mathematical model and operation strategy are presented to deduce the component sizing of solar-wind hybrid system that incorporates an EV charging station. The optimal configuration consists of solar capacity of 36 kW and wind capacity of 20 kW along with grid sales and purchase capacities of 10 kW. The EV demand of 28 MWh/year and the shopping arcade demand of 40 MWh/year are completely managed by the renewable energy sources and the grid. Excess electricity of 27 MWh/year is given to the dump load. The energy sold to the grid is 43 MWh/year, which is much higher than the energy purchased from the grid, i.e., 8 MWh/year. LCOEs obtained from ABC and PSO algorithms are measured to be 0.038 and 0.043 $/kWh, respectively. LCOEs using both algorithms are highly competitive with the cost of energy purchased from the grid, while maintaining the LPSP to be 0.19%.

Moreover, a more generic case study is considered by relaxing the time constraints, and the system is satisfactorily economical. However, the grid dependency increases in this case. Further, a sensitivity impact analysis on the maximum grid sales and purchase capacities on the LCOE is performed. In a grid-connected system, the maximum grid sales and purchase capacities demonstrate a significant impact on LCOE and ACS of the system. Finally, a parameter sensitivity analysis is performed to analyze the impact of the algorithm parameters on the solution quality of both algorithms. The obtained solution shows minor variations around the mean value, and the relative deviation from the minimum ACS is less than 1%. The proposed system will be helpful in promoting renewable energy sources in smart grid and can reduce the dependence of a small community on the grid.

References

1

Ministry of New and Renewable Energy Annual Report. (2019, Dec.) [Online]. Available: https://mnre.gov.in/knowledge-center/publication [百度学术

2

M. Nehrir, C. Wang, K. Strunz et al., “A review of hybrid renewable/alternative energy systems for electric power generation: configurations, control, and applications,” IEEE Transactions on Sustainable Energy, vol. 2, no. 4, pp. 392-403, May 2011. [百度学术

3

X. Chen, M. B. Mcelroy, Q. Wu et al., “Transition towards higher penetration of renewables: an overview of interlinked technical, environmental and socio-economic challenges,” Journal of Modern Power Systems and Clean Energy, vol. 7, no. 1, pp. 1-8, Aug. 2019. [百度学术

4

W. Zhou, C. Lou, Z. Li et al., “Current status of research on optimum sizing of stand-alone hybrid solar-wind power generation systems,” Applied Energy, vol. 87, no. 2, pp. 380-389, Feb. 2010. [百度学术

5

A. Kanase-Patil, R. Saini, and M. Sharma, “Integrated renewable energy systems for off grid rural electrification of remote area,” Renewable Energy, vol. 35, no. 6, pp. 1342-1349, Jun. 2010. [百度学术

6

V. J. Babrekar, S. D. Bandawar, and A. R. Behade, “Review paper on hybrid solar-wind power generator,” International Journal of Computer Applications, vol. 165, no. 5, pp. 36-40, May 2017. [百度学术

7

S. Bhattacharjee and S. Acharya, “PV-wind hybrid power option for a low wind topography,” Energy Conversion and Management, vol. 89, pp. 942-954, Jan. 2015. [百度学术

8

G. Tina, S. Gagliano, and S. Raiti, “Hybrid solar/wind power system probabilistic modelling for long-term performance assessment,” Solar Energy, vol. 80, no. 5, pp. 578-588, May 2006. [百度学术

9

H. Yang, W. Zhou, L. Lu et al., “Optimal sizing method for stand-alone hybrid solar-wind system with LPSP technology by using genetic algorithm,” Solar Energy, vol. 82, no. 4, pp. 354-367, Apr. 2008. [百度学术

10

A. Kaabeche, M. Belhamel, and R. Ibtiouen, “Sizing optimization of grid-independent hybrid photovoltaic/wind power generation system,” Energy, vol. 36, no. 2, pp. 1214-1222, Feb. 2011. [百度学术

11

S. Diaf, M. Belhamel, M. Haddadi et al., “Technical and economic assessment of hybrid photovoltaic/wind system with battery storage in corsica island,” Energy Policy, vol. 36, no. 2, pp. 743-754, Feb. 2008. [百度学术

12

B. Ye, J. Jiang, L. Miao et al., “Feasibility study of a solar-powered electric vehicle charging station model,” Energies, vol. 8, no. 11, pp. 13265-13283, Nov. 2015. [百度学术

13

A. Koochaki, M. Divandari, E. Amiri et al., “Optimal design of solar-wind hybrid system using teaching-learning based optimization applied in charging station for electric vehicles,” in Proceedings of IEEE Transportation Electrification Conference and Expo, Long beach, USA, Jun. 2018, pp. 1-6. [百度学术

14

R. Kaur, V. Krishnasamy, and N. K. Kandasamy, “Optimal sizing ofwind-PV-based DC microgrid for telecom power supply in remote areas,” IET Renewable Power Generation, vol. 12, no. 7, pp. 859-866, May 2018. [百度学术

15

M. Nizam and F. R. Wicaksono, “Design and optimization of solar,wind, and distributed energy resource hybrid power plant for electric vehicle charging station in rural area,” in Proceedings of 5th International Conference on Electric Vehicular Technology, Surakarta, Indonesia, Oct. 2018, pp. 41-45. [百度学术

16

M. Habib, S. Said, M. El-Hadidy et al., “Optimization procedure of a hybrid photovoltaic wind energy system,” Energy, vol. 24, no. 11, pp. 919-929, Nov. 1999. [百度学术

17

J. Li, W. Wei, and J. Xiang, “A simple sizing algorithm for stand-alone [百度学术

PV/wind/battery hybrid microgrids,” Energies, vol. 5, no. 12, pp. 5307- [百度学术

5323, Dec. 2012. [百度学术

18

H. Yang, L. Lu, and W. Zhou, “A novel optimization sizing model for hybrid solar-wind power generation system,” Solar Energy, vol. 81, no. 1, pp. 76-84, Jan. 2007. [百度学术

19

G. Bekele and G. Boneya, “Design of a photovoltaic-wind hybrid power generation system for ethiopian remote area,” Energy Procedia, vol. 14, pp. 1760-1765, 2012. [百度学术

20

F. Mwasilu, J. J. Justo, E.-K. Kim et al., “Electric vehicles and smart grid interaction: a review on vehicle to grid and renewable energy sources integration,” Renewable and Sustainable Energy Reviews, vol. 34, pp. 501-516, Jun. 2014. [百度学术

21

S. Singh, S. Jagota, and M. Singh, “Energy management and voltage stabilization in an islanded microgrid through an electric vehicle charging station,” Sustainable Cities and Society, vol. 41, pp. 679-694, Jun. 2018. [百度学术

22

H. C. Yu and C. G. Lu, “Recent development of electric vehicles,” Applied Mechanics and Materials, vol. 490, pp. 968-971, Jan. 2014. [百度学术

23

C. Chellaswamy, V. Nagaraju, and R. Muthammal, “Solar and wind energy based charging station for electric vehicles,” International Journal of Advanced Research in Electrical Electronics and Instrumentation Engineering, vol. 7, no. 1, pp. 313-324, Jan. 2018. [百度学术

24

S. Khan, A. Ahmad, F. Ahmad et al., “A comprehensive review on solar powered electric vehicle charging system,” Smart Science, vol. 6,no. 1, pp. 54-79, Dec. 2017. [百度学术

25

H. Li, H. Liu, A. Ji et al., “Design of a hybrid solar-wind powered charging station for electric vehicles,” in Proceedings of International Conference on Materials for Renewable Energy and Environment, Chengdu, China, Aug. 2014, pp. 977-981. [百度学术

26

M. Nashed and I. Edwar, “Wind/PV hybrid of DC electric vehicle charging station with bi-directional converter,” in Proceedings of World Engineering Conference and Convention, Kyoto, Japan, Dec. 2015, pp. 1-6. [百度学术

27

O. Hafez and K. Bhattacharya, “Optimal design of electric vehicle charging stations considering various energy resources,” Renewable Energy, vol. 107, pp. 576-589, Jul. 2017. [百度学术

28

G. C. Mouli, P. Bauer, and M. Zeman, “System design for a solar powered electric vehicle charging station for workplaces,” Applied Energy, vol. 168, pp. 434-443, Apr. 2016. [百度学术

29

O. Sadeghian, M. Nazari-Heris, M. Abapour et al., “Improving reliability of distribution networks using plug-in electric vehicles and demand response,” Journal of Modern Power Systems and Clean Energy, vol. 7, no. 5, pp. 1189-1199, May 2019. [百度学术

30

A. E. S. A. Nafeh, “Optimal economical sizing of a PV-wind hybrid energy system using genetic algorithm,” International Journal of Green Energy, vol. 8, no. 1, pp. 25-43, Feb. 2011. [百度学术

31

M. R. Javadi, A. Jalilvand, R. Noroozian et al., “Optimal design and economic assessment of battery based stand-alone wind/PV generating system using ABC,” in Proceedings of the 3rd Conference on Thermal Power Plants, Tehran, Iran, Jan. 2011, pp. 1-7. [百度学术

32

L. Wang and C. Singh, “Multicriteria design of hybrid power generation systems based on a modified particle swarm optimization algorithm,” IEEE Transactions on Energy Conversion, vol. 24, no. 1, pp. 163-172, Jan. 2009. [百度学术

33

M. Kharrich, M. Akherraz, and Y. Sayouti, “Optimal sizing and cost of a microgrid based in PV, Wind and BESS for a school of engineering,” in Proceedings of International Conference on Wireless Technologies, Embedded and Intelligent Systems, Fez, Morocco, Apr. 2017, pp. 1-5. [百度学术

34

M. Nazari-Heris, B. Mohammadi-Ivatloo, S. Asadi et al., “Large-scale combined heat and power economic dispatch using a novel multi-player harmony search method,” Applied Thermal Engineering, vol. 154, pp. 493-504, May 2019. [百度学术

35

O. Hoseynpour, B. Mohammadi-Ivatloo, M. Nazari-Heris et al., “Application of dynamic non-linear programming technique to nonconvex short-term hydrothermal scheduling problem,” Energies, vol. 10, no. 9, p. 1440, Sept. 2017. [百度学术

36

R. Kumar, R. Gupta, and A. K. Bansal, “Economic analysis and power management of a stand-alone wind/photovoltaic hybrid energy system using biogeography based optimization algorithm,” Swarm and Evolutionary Computation, vol. 8, pp. 33-43, Feb. 2013. [百度学术

37

M. Nazari-Heris, A. F. Babaei, B. Mohammadi-Ivatloo et al., “Improved harmony search algorithm for the solution of non-linear nonconvex short-term hydrothermal scheduling,” Energy, vol. 151, pp. 226-237, May 2018. [百度学术

38

O. Ekren and B. Y. Ekren, “Size optimization of a PV/wind hybrid energy conversion system with battery storage using simulated annealing,” Applied Energy, vol. 87, no. 2, pp. 592-598, Feb. 2010. [百度学术

39

D. Karaboga and B. Akay, “A comparative study of artificial bee colony algorithm,” Applied Mathematics and Computation, vol. 214, no. 1, pp. 108-132, Aug. 2009. [百度学术

40

D. Karaboga and B. Basturk, “A powerful and efficient algorithm for numerical function optimization: artificial bee colony algorithm,” Journal of Global Optimization, vol. 39, no. 3, pp. 459-471, Nov. 2007. [百度学术

41

J. Kennedy and R. Eberhart, “Particle swarm optimization,” in Proceedings of ICNN’95-International Conference on Neural Networks, Perth, Australia, Mar. 1995, pp. 1942-1948. [百度学术

42

Surface meteorology and solar energy. (2019, Dec.). [Online]. http://eosweb.larc.nasa.gov/sse/RETScreen, NASA [百度学术

43

HOMER. (2019, Dec.). [Online]. https://www.homerenergy.com/products [百度学术

44

S. Singh and S. C. Kaushik, “Optimal sizing of grid integrated hybrid PV-biomass energy system using artificial bee colony algorithm,” IET Renewable Power Generation, vol. 10, no. 5, pp. 642-650, Apr. 2016. [百度学术

45

N. J. Singh, J. S. Dhillon, and D. P. Kothari, “Synergic predatorprey optimization for economic thermal power dispatch problem,” Applied Soft Computing, vol. 43, pp. 298-311, Jun. 2016. [百度学术