Management of Fluctuating Output Power of Renewable Energy Resources by Optimum Charge/Discharge Scheduling of Electric Vehicles in Smart Parking Lots

The output power of renewable resources fluctuates because of their random nature, which can negatively affect the grid. This problem can be resolved by installing battery energy storage systems (BESS). On the other hand, with the increasing development of electric vehicles (EVs), the scheduling of their presence in the distribution networks has become very important; because, any unplanned presence of EVs in the smart parking lots (SPLs) and their simultaneous charging may affect the performance of the grid, negatively. An optimum scheduling for charge/discharge actions of EVs can resolve this problem, and reduce the need for high cost BESSs. In this paper, the problem of maximizing the operator profit is formulated as a mixed integer linear programming (MILP) problem considering the constraints on the requested charge level of the EVs and the permissible exchanged power. Then, the problem is optimally solved, in which a scheduling scheme is proposed for charge/discharge actions of the EVs in the SPLs in order to control the power fluctuations. Simulation results on the data of a typical parking lot in Tehran demonstrate that the proposed scheduling can reduce the number of required BESSs significantly, which decreases the high expenses of the BESS purchase and installation.


Introduction
In recent years, factors such as environmental issues, reduction of fossil fuel resources, and increasing the energy demand have made the presence of renewable energy sources (RESs) in power systems increasingly important. Energy sources like wind energy, solar energy, fuel cell and micro turbines are included in the RES, and are more environmental friendly [1][2][3].
On the other hand, development of smart grids is very necessary. Growing use of EVs provides various opportunities for future systems, including the prevention of the release of environmental pollutants and the increasing penetration of distributed generations [4,5].
The uncontrolled charging of plug-in hybrid electric vehicles (PHEVs) on the demand side creates a significant load peak over a short period of time and a valley in the network consumption diagram [6]. However, one of the great advantages of PHEVs is the ability to transfer power to the grid. The concept of vehicle-to-grid (V2G) was first used to express the revenue and cost involved in regulating electricity market and the ancillary service market [7]. With the presence of EVs in the power exchange, intermediate commercial entities have emerged between the owners of EVs and network operators, namely EV aggregators and parking lots [7].
These new aggregators create a sophisticated energy storage system (ESS) by keeping large number of EVs next to each other to transmit energy from the grid to vehicle (G2V) and from V2G using new technologies .
Aggregators as a flexible storage resource are an interesting challenging issue in the fields of regulating electricity market and the ancillary service market [8]. The presence of aggregators alongside the RES has a significant impact on the growth of clean energy sources and can also be an appropriate support for the RES [9]. In order to maintain the network healthy as well as reduce the energy costs, the management of EV charging actions has been extensively studied by the researchers [10].
The impact of EVs on the reactive power control and the distribution network is discussed in [11]. While the RESs and vehicles are mainly located in the distribution network, some authors have modeled these sources at the energy transfer levels [12][13][14][15]. Reference [16] demonstrates the cost-saving capability of using hybrid EVs alongside the presence of V2G technology. In [17], in order to maximize profits of the aggregator, the penetration level of EVs in the regulating electricity market is determined.
In order to reduce the peak load in [18], a new charge management strategy for optimal charging of hybrid EVs is presented. The coordination and collaboration between wind energy conversion systems and EVs has been extensively studied in the literature. For example, the authors in [19] evaluate EVs for providing ancillary services based on the network interaction with WECS in the US electricity market. In [20], a two-objective optimization model is used for economical operation and environmental performance of SPLs with respect to the average participation rate of the vehicles in the response and demand program.
Despite the existing researches on the SPL, some cases have not been considered in the studies yet. For example, in [21], in order to reduce the speed of battery aging and extend its lifetime, V2G technique is not used as much as possible. Instead, in order to smooth the output power fluctuations, it stores the excessive power with respect to the upper limit into the EV battery. In some studies, the sampling interval to update the EV charging request is too long .For example, the sampling interval is 30 and 15 minutes in [22] and [23], respectively.
It should be noted that one of the challenges facing the development of wind farms is the high amplitude of their short-term power fluctuations (in terms of several seconds to multiple minutes). Installation of the BESS at the farm side can resolve this problem. However, the high cost of the BESS installation is another challenge. Using the SPLs can reduce the expenses. Although there are a number of works on this issue, some of them have focused on longterm fluctuations [22,23]. Also, the authors in [24] have not considered the profit of the network operator.
In this paper, an optimal scheduling system is proposed for the charge/discharge actions of the EVs in the SPLs, which makes decisions about the exchanged power, and smooths the output power of a RES set. This scheduling scheme is designed so that the requested charging level of the EVs is met at the departure time, and the profit of the network operator is maximized, as well.
The rest of the paper is outlined as follows: In Section II, the formulation of the objective function with problem constraints are presented. In Section III, the input data and the proposed model are implemented in MATLAB environment and the results of the optimized charging scheduling of the vehicles are achieved. Also, the results of integrating the SPLs with the RES are compared in three different scenarios. Finally, Section IV concludes this paper.

Simulation Results
In this section, the impact of using SPLs to smooth the power fluctuations in WPP is discussed. To this end, three scenarios are considered. Fig. 1 shows the power amount that should be exchanged (obtained from Section II) to smooth the power fluctuations.

Scenario 1: Without the presence of SPLs
To achieve an output power with permissible fluctuations under utility constraints, two-way ESS should be used. As a result, BESS would be a choice. The only drawback of using BESSs is their high prices (especially for highcapacity BESSs). In this scenario, it is assumed that a number of BESSs are used to support the exchanged power shown in Fig. 1 without the presence of SPLs.    (1) is used, in which the price of the EV charging is a random number between 0.15 $ and 0.3 $ per kWh ( [20]) .
Also, the price of discharging the EV battery is three times of that of charging for a given time. The constraint on the amount of exchanged power for charge and discharge actions for each EV is 2775.83 watt per minute. Fig. 3 shows the value of power provided by the SPLs to compensate for the power demanded by the grid. In Fig.  3, the black graph illustrates the power demanded by the grid, and the gray one is the power supplied by the SPLs .
Obviously, the expected power is not provided in many minutes, especially after the time of 16:00. The reason is that, after 16:00, the departure time of the vehicles approaches. However, the SPLs are comparatively successful at providing the requested charge level of the vehicles. Fig. 4 shows the final SOC level of the EVs at the departure time.

Scenario 3: Integration of SPLs with BESS
In order to achieve the first goal, i.e. providing the required exchanged power, and the second one, i.e. providing the requested charge level of the vehicles, the BESS can be used as the supplement of the SPLs. Fig. 5 illustrates the amount of power supplied by the SPLs and the BESS. From Fig. 5, it can be seen that all the required power of the WPP is provided by using the SPLs along with the auxiliary BESS. Also in Fig. 6, the powers of the WPP, SPLs, and BESS are shown separately . With using the BESS, it can be seen that the required power for the smoothing the power fluctuations is fully provided. But the presence of the BESS will have disadvantages. The first disadvantage is the increased cost of the power system. The second disadvantage is the reduction in the average charge level of EVs at the departure time. The constraints related to the BESSs in this scenario are the same as those in the first scenario . The power consumed for the BESS added to the SPLs affects the final SOC of the EVs and reduces the average SOC of the EVs to 74%. Fig. 8 shows the final SOC level of the EVs in scenario 3 .  Table 1 compares the exchanged power values in 8 SPLs at randomly selected minutes. The power values listed in this table are in Watts. Also, a positive power implies that the SPL receives power from the grid, and a negative one implies that the SPL injects power into the grid. Accordingly, it is observed that at a specified time, some of the SPLs receive power from the grid to charge the batteries of the vehicles, and some other deliver the power stored in the batteries of the vehicles to the grid through the discharge actions. This is because some EVs need to be charged for some reasons, such as approaching their departure time or having a low charge level. In addition to providing the required power of the RES, the network operator transfers the power stored in the batteries of the vehicles, whose departure time is not close, to the vehicles in other SPLs. This can also happen for the EVs in the same SPL. Such a power transmission is called V2V power transfer.  By optimally managing the charge/discharge actions of the EVs, the network operator can demand the required power from the SPLs, while delivering the required exchanged power at the same time. According to the decisions made by the operator, the vehicles behave differently at different minutes, so that their charge level reaches to the highest possible level at the departure time, while the required power of the grid is provided, as well. Fig. 10 show several examples of changes in the charge level of the EVs.  Table 2. According to the Table 2, V2G technology is used in the second and third scenarios. In the second scenario, the average charge value of the EVs is sufficient, and thus no BESS is used. However, this scenario has a lower capability to provide the required exchanged power. Also, the number of required BESS in the third scenario is less than the first scenario. Despite its higher expenses because of using the BESS, the third scenario is preferred to the second one, as it can fully provide the exchanged power. In addition, the number of required BESSs is reduced to 10 in the third scenario, compared with 18 BESSs in the first scenario. Assuming that the purchase and installation cost of each Lithium-ion BESS is 200 $/kWh [25], the proposed approach can achieve a reduction of 6.4 million dollars in the BESS expenses, compared with the traditional approach.

Conclusion
In this paper, the impact of using SPLs in the control of RES power fluctuation is investigated. Also, an optimal scheduling scheme for charge/discharge actions of the EVs in the SPLs is proposed, which seeks to reduce the number of BESSs used to supply the required power for the smoothing process, and maximize the profit of the network operator. Considering the constraints on the requested charge level of the EVs and the required exchanged power, this problem is modeled and solved as an MILP problem. In order to investigate the effect of using SPLs and BESSs in the power fluctuation smoothing process, three scenarios including scenario1: 22 unplanned presence of EVs in the intelligent parking lots (IPLs) and their simultaneous charging may affect the performance of the grid, negatively. An optimum scheduling for charge/discharge actions of EVs can resolve this problem, and reduce the need for high cost BESSs. In this paper, the problem of maximizing the operator profit is formulated as a mixed integer linear programming (MILP) problem considering the constraints on the requested charge level of the EVs and the permissible exchanged power. Then, the problem is optimally solved, in which a scheduling scheme is proposed for charge/discharge actions of the EVs in the SPLs in order to control the power fluctuations. Simulation results on the data of a typical parking lot in Tehran demonstrate that the proposed scheduling can reduce the number of required BESSs significantly, which decreases the high expenses of the BESS purchase and installation.