TeleWatt: An Innovative Electric Vehicle Charging Infrastructure Over Public Lighting Systems

Transcription

1 2013 International Conference on Connected Vehicles and Expo (ICCVE) TeleWatt: An Innovative Electric Vehicle Charging Infrastructure Over Public Lighting Systems Mario Alvarado Ruiz, Fadi Abi Abdallah,Maurice Gagnaire,Yannick Lascaux Institut Mines TELECOM - TELECOM Paristech - LTCI - UMR 5141 CNRS Networks and Computer Science Department, 23, avenue d Italie, Paris, FRANCE {mario.alvarado,fadi.abi-abdallah,maurice.gagnaire}@telecom-paristech.fr EDELCOM, 7, rue Jean Mermoz Versailles ylascaux@edelcom.fr Abstract The growth of the Electric Vehicle (EV) s market is strongly conditioned by the availability of cost-effective and reliable charging infrastructures. In the current state of the technology, the limited capacity of the batteries limits EV s market to urban moves. The rapid deployment of charging stations in urban areas is a real challenge. In this paper, we introduce an original approach consisting in reusing existing public lighting infrastructures for that purpose. Two main advantages characterize our solution. First, charging stations can be deployed rapidly without costly civil engineering. Second, a fraction of the power non consumed by the lamps at night can be used for the benefit of the charging stations. Meanwhile, such an approach must be achieved while guaranteeing the stability and quality of the lighting system. I. INTRODUCTION In France, more than 2 million of EVs are expected for the year 2020 [1]. Without the proper infrastructure to provide charging services, the growth of the EV s market could be at risk. Charging infrastructures for EVs include a set of elements known as Electric Vehicle Supply Equipments (EVSE) such as connectors, conductors and other equipment associated to the charging stations. Four charging modes offering various charging speeds are today specified in the IEC standard. The strategy to determine the location to install EVSE and the charging modes available in each area (on the basis of the available energy sources) depends on the policies and priorities adopted by each country or region [2]. The most urgent need in all EV markets is to finance charging infrastructures as suggested in [2]. The installation costs of EVSE could be very high. specially when the network must be upgraded or when civil engineering is required. If it is not planned carefully, the charging infrastructure may be over estimated for the need of the market and lead to unused assets. For this reason, in a context of public locations, the deployment of EVSE should not be just maximized, but rather optimized according to the number of potential users, the expected energy demand and the business model that best suits to the service provider. The primary service provided by the Public Lighting System (PLS) is the lighting of the streets. However, during daylight, This work is part of the TELEWATT research project granted by the French National Environmental and Energy Agency - ADEME. The results presented in this paper are the fruit of a collaboration between Institut Mines- Telecoms/Telecom ParisTech and the company Citelum, one of the world leaders in public lighting. the power available at the feeder could be used to provide ancillary services such as EVs charging. The benefits of this approach are the fast and low-cost deployment of the EVSE infrastructure, without needing an upgrade of the PLS in the short term. The feasibility of this approach is demonstrated not only during the daylight but also at night. This paper is organized as follows. In section II, we briefly recall some of the approaches proposed for urban EVs charging. The characteristics and objectives of the TeleWatt project are presented in section III. In Section IV, we demonstrate the feasibility of the TeleWatt s approach through computer simulations. Section V presents the obtained numerical results validated by real measurements. II. RELATED WORKS EV charging configurations can be divided in two categories: dedicated and shared infrastructures. In dedicated infrastructures, EVSE are deployed exclusively for EV charging. Fast charging stations, car parks locations, EVs fleet charging stations or battery switch stations [3] [5] are some examples of dedicated infrastructures. In this context, charging terminals are directly connected to an electrical supply. This kind of architecture implies an important investment in network reinforcements. If a dedicated infrastructure is not judiciously dimensioned, EVSE can be oversized for the actual demand of energy. In shared infrastructures, EV s charging service is provided simultaneously to annex electrical services. The installation of a charging terminal at the customer s premises is a typical example of shared infrastructure. When EV s charging is active, it is considered as a new load added to the domestic electrical network. The resulting requested power must be compatible with the capacity of the electrical network of the considered user. As the level of penetration of EVs on the market increases, shared infrastructures will have to be upgraded. In order to accommodate the additional demand on the network, a smart control of the fluctuant available power is necessary. Different smart charging strategies for EV s charging in residential networks have been proposed and studied in [6] [8]. These strategies are based on the time of departure of the vehicles, the initial state of charge of the batteries and eventually the dynamic pricing of energy. In /13/$ IEEE 741 DOI /ICCVE

2 [9], the charging of EVs over bus or tree-based distribution infrastructures has been investigated. The voltage sensitivity to additional loads placed at the end of such configurations has been evaluated. The importance of avoiding the connection of EVs to the same phase of the infrastructure has been outlined in order to optimize the number of simultaneous EVs charging. The charging strategies presented in the literature favor vehicles connected close to the Low Voltage (LV) power substations in the distribution grid while ensuring that residential load remains unaffected. Today, PLS can be viewed as a dedicated infrastructure since its only objective is to provide lighting along the streets. In [10], the flexibility of shared electrical distribution grids is discussed. Parallel services may be implemented on such infrastructures thanks to smart metering technologies. In PLS, smart meters are used to regulate the power injected into the lamps in order to save energy. Since power consumption in PLS is easily predictable, the integration of additional services appears as a possibility that we propose to exploit in this paper. A PLS presents similarities with a residential network. Thus, one observes in both cases that the loads connected at the end of bus-based or tree-based topologies will have more significant voltage drops than those connected closer the LV substation. During the day, when the lamps are turned off, the capacity of PLS infrastructures is not exploited. The model of infrastructure proposed by TeleWatt aims to transform dedicated infrastructures like the PLS into shared infrastructures with EVs charging services. To the best of our knowledge, this paper is the first one to consider such an approach not only during the day but also at night. III. TELEWATT OVERVIEW TeleWatt is a National project supported by the ADEME agency, a subsidiary of the French Ministry of Industry. The main objective of this project is to provide public EV charging services through existing PLS infrastructures. In the remaining of the paper, only the case of PLS based on discharge lamps is considered (the great majority of existing configurations). New charging terminals have been designed in order to be attached to existing light poles and connected to the existing LV substations feeding PLS 3-phase buses or trees. In order to communicate with the charging terminals, Citelum [11] (one of the partners in the project) has developed the smart metering technology needed to control in real-time the power injected at the charging terminals and the lamps. A PLS is constituted by a LV substation and a set of conductors that connect multiple light poles, sensors for remote control, and a set of charging terminals. A LV substation corresponds to a 3-single phase transformer with 1 neutral. The power delivered by the transformer is fixed according to the size of the network. It is important to balance the loads connected to the transformer in order to avoid the increment of current on the neutral in order to prevent any service disruption. Each LV substation feeds one or several 3-phase lines. The TeleWatt service is accessible by mean of the users smartphone. Once it has been localized, a TeleWatt client specifies to the TeleWatt server its destination and its current state of charge (SoC). The TeleWatt server sends back to the user a list of available charging terminals close to the destination. Charging terminals are connected to the EV s charger through a standard Mode-3 single-phase AC connection. The Central Control Manager (CCM) of TeleWatt manages the set of charging operations. A. Adapting to changes in the PLS grid When an event (e.g. EV connection/disconnection, switching the lamps on/off, EVSE failure) occurs, the power supplied to the loads connected to the grid must be recalculated with a twofold objective. The former is to guarantee the proper operation of the primary service of the network, which is the lighting service. The latter is to balance the power delivered at each terminal for EVs charging. Our aim is to provide a constant charging rate to the terminals, independently from their position along the transmission line. For this purpose, the CCM simulates in advance the connection of an EV for every charging terminal and determines the feasibility of a new connection. Thus, the CCM dynamically enables or disables charging terminals. Green or red LEDs attached to the charging terminals are activated accordingly to visually display the charger availability to the customer. 1) Enabling charging terminals: At any instant, a charging terminal is available for charging if it can start charging any given incoming vehicle. In this case it s signaling green LED is active. In that case,the TeleWatt CCM cannot refuse the service to an incoming EV (apart pricing considerations that are out of the scope of this paper). After the connection of an EV to an available terminal, the charging process begins immediately at the charging rate estimated by the CCM. The charging rate of a terminal is defined as the power provided by the network to the battery through the charger during a period of time. In summary, a charging terminal is available if the following constraints are respected: (i) The available power P av on the installation is sufficient to charge any type of EV at least at its minimum charging rate as defined by the CCM. The power P av is given by: P av (t) =P capacity (t) P consumed (t) (1) where P capacity corresponds to the total power capacity of the considered LV substation. P consumed is the the sum of the power consumed by each lamp and active charger and the total power loss inherent to the feeder cable. We have then: P consumed (t) = n i=0 P i lamps(t) m i=0 P i chargers(t) l i=0 P i losses(t) (2) where n, m, l are the total number of lamps, chargers, and losses inherent to the installation and P lamps, P chargers, and P losses, are the power consumed by the lamp i, the charger i and the power loss along segment i, respectively. (ii) A sudden EV connection or disconnection from a charging terminal causes a drop in the voltage level at all the junction points of the network. Such a drop in the voltage level depends on different factors, which include the distance 742

3 from the transformer to the considered charging terminal, the phase on which the EV is connected and the power delivered to the EV battery. Thus, the voltage level at any point of the network should not exceed the minimum voltage threshold that guarantees the PLS normal operation. The design of the TeleWatt charging terminals is such that if the CCM estimates that at the considered instant, a charging terminal is unavailable (with an active red LED), then the connection by the user of his charger onto this charging point is electrically impossible. B. Update of the Network The CCM defines a series of time steps, in order to obtain the voltage levels and power consumption at all the branching points of the network. The State Of Charge (SoC) of the batteries connected and the power delivered to the charging terminals is periodically updated into the control system. The CCM database is updated each time an EV is connected or disconnected. By doing so, the CCM is able to control power supply for each EV by adjusting the charging rate. C. Controlling power supply for the EVs When an EV is connected to a charging terminal, the charger provides to this EV a constant charging rate accordingly to the type of the considered EV. For that purpose, the Distribution System Operator (DSO) adjusts the charging rate of each charger individually. Slow and standard charges are supported in TeleWatt. The SoC refers to the percentage of remaining energy available in the battery [12]. The time of charging depends on: (i) the size of the battery, (ii) the initial SoC and (iii) the charging rate provided by the terminal. The estimation of the SoC of the battery SoC (t) at any given instant t during the charging process is calculated by ( 3). SoC (t) = SoC (t Δ) +(ChR (t) Time (t (t Δ)) /B size ) (3) Here SoC (t) is the SoC of the battery at instant t, SoC (t (t Δ)) is the SoC of the battery registered at the last update of the network. ChR (t) refers to the charging rate delivered to the battery at time t. Time (t (t Δ)) is the time elapsed from the last update of the network to instant t. B size represents the total capacity of the battery (in kwh). It is assumed that the network can communicate with each EV in order to collect its initial SoC and the size of its battery. D. Typical behavior of a TeleWatt infrastructure over 24 hours The availability and the proportion of power used to charge EVs in TeleWatt depends on the behavior of the PLS. For this reason, it is important to understand PLS operation and determine how the EVs are power-fed. A PLS operation is made of three main stages: 1) Daylight Stage: which takes place at dawn, when all the lamps are turned off. At this stage, the only loads connected to the system are the EVs. 2) Lamp Heating Stage (LHS): which takes place at dusk and all the lamps are turned on. At this stage, the lamps consume generally 40% more of nominal power before the power consumption of the lamps settles down to a nominal value. 3) Night Stage: which takes place after the LHS and lasts until the Day Stage. At this stage, the lamps operate at the nominal power. During the Daylight Stage, all the P av can be used to charge the EVs. In the presented model, it is assumed that the charging rate attributed to an EV remains unchanged, disregarding the charging terminal where the EV is connected. However, the charging rate has been defined according to the type of EV and the size of the battery. Once the LHS starts, due to the increase in the power demanded by the network at the beginning of this stage, the CCM interrupts transitorily all the active EV charging processes. As soon as the LHS ends and the Night Stage starts, the EV charging processes are resumed for the EVs that were connected before LHS. This resume period occurs as long as the power available in the network allows it. Due to available power reduction at that stage, it may not be possible to reactivate simultaneously all the charging procedures after the LHS during the Night Stage. In that case, we propose that the CCM adopts a round-robin procedure for fairness purposes. Once these EVs are fully charged during the night stage, a similar regime as the daylight stage is adopted for new connections, but under a lower available power. At sunrise, the CCM gets back to the Daylight Stage. IV. SIMULATION DATA We have developed a simulator in C# language to emulate the behavior of the CCM applied to a PLS infrastructure. By implementing the technique proposed by [13], it is possible to obtain a graphical representation of any 3-phase network as a topology constituted by several branches and nodes. Without going into the details, specific numerical analysis tools have been used in order to enable computation durations compatible with the real-time constraint of the TeleWatt server. A. Public Lighting Network The test network considered is based on a LV public distribution feeder similar to the French PLS networks operated by Citelum. This network operates at a nominal voltage of 230/400V with a tolerance of ±10%. Different sizes of transformers can be used in the network (32A, 64A and 128A). However 64A transformers are considered. The input voltage V input is of 322V or V rms. The supplying feeder consists of 90 lamps and 9 charging terminals deployed through 2.7km of 3-phase and one neutral copper main cables. It is considered that the distance between the transformer and the first light pole is of 33m. The other light poles are 33m spaced from each other. For modeling purposes, the power consumed by each lamp is set at 145W. It is assumed that only one lamp is connected to a light pole. The light poles are equally distributed through the 3 phases of the transformer. The charging terminals are placed at the nearest light poles 743

4 from the transformer, attaching a single charging terminal by light pole. Specifications for the network model components were supplied by Citelum [11]. In the presented model, the power capacity of the source TPwr capacity is equal to the size of the transformer (T size ) multiplied by the input voltage (Vrms input ): TPwr capacity = T size Vrms input (4) TPwr capacity =64A V =14.572W The entry values for equation ( 2) are n = 90 lamps, m is the total number of chargers connected on the network, 0 m 9 depending on the number of chargers turned on. The losses on the network are considered as the power lost at the conductors (l) that link the different loads to the feeder. B. Stages of the Public Lighting System For modeling purposes, 10 periods of 24-h each are considered. Each period goes from 18:00 to 17:59 of the following day. Every period starts with the LHS at 18:00. It is followed by the Night Stage from 18:30 to 06:00 of the following day. Finally the Daylight Stage takes place from 06:00 until 18:00. Despite the real capabilities of the network operator to regulate the power consumption at each lamp, for the sake of simplicity, it is assumed that from the beginning of the LHS until the end of the Night Stage, all the lamps operate at their nominal power rate. C. Type of Vehicles and Battery Size In 2012, more than 6000 electric vehicles were sold in France including Plug-in Hybrid Electric Vehicle (PHEV) and Battery Electric Vehicles (BEVs) which will be referred simply as EVs. Though the presence in the French market of PHEV is still limited, the model proposed considers both types of vehicles, EVs and PHEV. In order to represent the actual distribution of EVs and PHEV in the model, the total sales of electric vehicles in France in 2012 was taken as reference [14]. The size of the batteries were taken from the models with more presence in the market [15] [17]. The battery size is very homogeneous in the models available for EVs, which is on average of 24kWh. On the contrary for PHEV the sizes are variable. Therefore, two groups of PHEV are considered: PHEV1 with an average battery size of 4.4kWh which represents 62% of the market. The second group, PHEV2 represents the rest of the share of the PHEV market with a battery size of 16.5kWh. For all cases, it is assumed that all the vehicles have Li-ion batteries and they are modeled as constant power loads. 1) Connected Vehicles: In order to obtain an approximation of reality, 500 vehicles arriving during a period of 10 days have been simulated. Three vehicle distributions were tested in the model: (a) Only EVs connected. (b) A mix of EVs and PHEV: Constituted by 409 EVs, 64 PHEV1 and 27 PHEV2. (c) Only PHEV: A total of 310 PHEV1 and 190 PHEV2 were considered. Type of Vehicle Average Battery Size (kwh) Charging Rate (kw) EV / 3.3 PHEV PHEV / 3.3 TABLE I SIZE OF BATTERY AND CHARGING RATE ACCORDING TO THE TYPE OF VEHICLE [15] [17] Arrival Group μ σ (in hours) Number of arrivals AM (μ 1,σ 1 ) 09h % MD (μ 2,σ 2 ) 14h % PM (μ 3,σ 3 ) 19h % TABLE II DISTRIBUTION OF THE ARRIVAL TIMES OF THE VEHICLES DURING THREE CHARGING PERIODS 2) State of Charge and Charging Rates: It is assumed an initial SoC for all the vehicles when arriving to the charger of 20%. To avoid an aging phenomenon on the batteries, we interrupt the charging operation once their SoC reaches the 80%. For modeling purposes the charger delivers a constant power to the vehicle. The model only considers two types of charges: (i) Slow charge set at 1.7kW and (ii) Standard Charge set at 3.3kW. EVs and PHEV2 can be charged at both speeds of charge depending on the considered policies. However, it is considered that the group PHEV1 can only be charged at a slow charging rate. A summary of the size of the batteries and the charging rate is presented in Table I. 3) Aggregation of a Fleet of Vehicles: Once the number of vehicles arriving to the charging terminals has been established, and the charging rate for each type of vehicle defined, the time of arrival was selected. In order to distribute the arrival times among the 500 vehicles, the arrivals were divided into three groups. The first two groups represent 30% of the fleet each. They consider a time of arrival to the charging terminal around 09h00 (Group AM) and 14h00 (Group MD); while the third group (PM), with the 60% of remaining vehicles, considers a time of arrival around 19h30 to the charging terminal. The instants of arrival for all the vehicles are randomly and normally distributed during the three periods previously defined as it is presented in Table II. D. Selection of a Charging Terminal by the EV In order to reduce the effects on the voltage sensitivity by connecting the loads at the extreme of the network, all the charging terminals are placed at the nearest possible position from the LV substation. However, in order to analyze the impact of a new connection caused by its position in the line, three scenarios are considered for selecting the order of connection to a charging terminal by the user at its arrival. (i) Closest Charger First (CCF): This scenario considers that the user is always directed towards the nearest charging terminal available. (ii) Farthest Charger First (FCF): This scenario considers that the user always selects the farthest charging terminal available. (iii) Random Charger Selection (RCS): This scenario gives absolute freedom to the user for selecting any charging terminal available, disregarding its position along the line. 744

5 E. New Connection Test and Charging Constraints In the presented model, the charger of the terminal is simulated as a branch on a given topology. This branch represents a variable resistor connected between the neutral cable and one of the three phases. The value of the resistor is calculated in order to obtain the charging rate requested by the charger. In order to provide a constant power to the battery, the value of the resistor must be variable. If a fixed value of the resistor was chosen, the power delivered to the charger would variate through the network, decreasing at the extreme of the line and not delivering a constant power as requested. The value of all the resistors has to be balanced until finding the value that delivers the power requested at the new connection, while respecting the charging rates of the vehicles previously connected. In order to test the feasibility of a new connection at any available charger, the addition of a new load is simulated for each one of the three phases of the source transformer. The connection is tested at the worst possible charger available, which in the model is considered to be the farthest charger that does not have a vehicle connected. Once the charger has been selected, the value of the resistor that delivers the power requested at the charger is calculated. This operation has to be performed for all the vehicles connected until the resistors are balanced and the requested charging rate is obtained. At every step, the minimum voltage drop threshold and the total power consumption are verified. If the constraints are respected all the chargers connected to the same phase are set as available and the next phase is evaluated. If the connection simulated to the charger selected does not respect the network constraints, the red LED of the charging terminal is activated and the next charger on the same phase is selected. In this way, it is possible to guarantee that if a vehicle selects a charging terminal that is displayed as available, the battery will be charged at least at the minimal requested charging rate. F. Charging Rate Regulation at the Charger When the vehicle is connected to the charging terminal, the charging process should not stop until the SoC of the battery reaches its 80%. The only allowed interruption takes place during the LHS, which will be followed by a reconnection of the EV previously connected according to the power available at the network. The present work has considered two different policies for regulating the charging rate delivered by the charger. (i) Fixed Charging Rate: This policy considers a standard charging rate for the EVs and PHEV1, and a slow charging rate for PHEV2. The charging rates cannot be regulated. Therefore, in order to display a charging terminal as available, the CCM must guarantee that any type of EV can be connected. By consequence, the standard charging rate is taken as a reference in order to decide if a terminal is available or not. (ii) Flexible Charging Rate: In this policy it is assumed that EVs and PHEV1 can be charged at a slow rate when the power available in the network is not enough to support a standard charging rate. In case that the power available P av at the installation is SlowChargingRate P av StandardChargingRate, the EVs and PHEV1 will charge at slow rate, until the power in the network allows to increase the power and charge at the standard charging rate. Once the LHS is over, if the total power available in the network is not enough to resume simultaneously the charge of the vehicles that were previously connected, the vehicles will be charged by using the well known circular Round-robin algorithm. Time slots of 20 minutes are defined. During this time the number of vehicles that can be charged simultaneously will resume the charging process. At the end of the slot, one of the vehicles attributed to the slot of time is disconnected and the next vehicle on hold is connected. This process continues until the power available at the network is enough to resume the charge of the remaining vehicles simultaneously or until the end of the Night Stage. At this time, all the vehicles that have not reached a SoC of 80% are reconnected at once. V. RESULTS AND DISCUSSION A. Total Number of Accepted Vehicles One of our major interrogations at the beginning of this work was the possible impact on the number of accepted vehicles according to the position of the selected charging terminal. This was considered due to the voltage sensitivity which is affected by the distance from the transformer. However as it is presented on table III, the total number of vehicles accepted was not affected by the position of the charger, or the order of connection in the different scenarios considered (CCF, FCF or RCS). The acceptance rate depends more on the type of vehicle connected and the hour of arrival to the charger. However, this parameter can be improved by adopting smart charging strategies. Smart charging strategies could accept vehicles for charging and postpone the time when the charging process begins. In the model presented, a small improvement can be observed on the table III. In this case, just by accepting a vehicle to start charging at a slow charging rate when the power available is not enough to provide a standard charging rate. Thus, the network was capable of treating more vehicles than when the charging rate is not flexible. At the same time, the adopted strategy did not reduce the time of charging of the vehicles previously connected. B. Voltage Drop It was observed that despite the order of selection of the charging terminals, and its position on the network, the minimum voltage level registered never exceeded the network constraints. The most important voltage drop occurred when a vehicle was placed following the FCF scenario, with a mix of EVs, PHEV1, and PHEV2. In this case, the minimal voltage value registered was V. Due to the initial assumptions, that placed all the charging terminals at the nearest light poles from the transformer, it was not possible to truly observe an impact on the drop of voltage. 745

6 Fixed Charging Rate Flexible Charging Rate EVs Only PHEVs Only EVs / PHEVs Mix CCF FCF RCS CCF FCF RCS CCF FCF RCS Total EVs Accepted Total PHEV1 Accepted Total PHEV2 Accepted Total Energy Demanded(kWh) 1, ,269 Total Power Consumed (kw) 2,776 2,787 2,784 2,401 2,407 2,405 2,764 2,775 2,774 Total Power Lost (kw) Total EVs Accepted Total PHEV1 Accepted Total PHEV2 Accepted Total Energy Demanded(kWh) 1, ,322 Total Power Consumed(kW) 2,834 2,845 2,842 2,462 2,468 2,468 2,818 2,828 2,826 Total Power Lost (kw) TABLE III TOTAL CONSUMED POWER AND TOTAL LOST POWER ACCORDING TO THE STRATEGY FOR SELECTING A PARTICULAR CHARGER IN THE NETWORK AND THE TYPE OF CONNECTED VEHICLE C. Total Power Consumed and Total Power Loss The total power loss registered in the network was obtained by calculating the power at all the segments which represent a transmission line. Despite the minimum impact in the voltage drop on the network due to the selection of the charging terminal, the power consumed in the network presents a different result as it is shown on the table III. When the loads are added at the farthest charging stations (FCF scenario), it is possible to observe how the power consumed and by consequence the power lost on the network increases. It is observed that the total power lost in the network is less significant when the charging rate is fixed. However, this is related to the total amount of energy demanded. As the flexible charging rate policy allows more connections of vehicles, there is more power delivered to the batteries and therefore more loss of power registered. VI. CONCLUSIONS AND PERSPECTIVES In the very next years, EV s market expansion will be conditioned by two factors: the advances in batteries technologies and the availability of low cost charging infrastructures in urban areas. We have shown through the TeleWatt project that PLS are a promising solution for non-dedicated charging systems. The obtained numerical results have been validated by on-site measurements. If the number of usable charging stations per PLS remains modest, going by a rule of thumb, we can estimate that around 10,000 charging stations could be deployed in one year at low cost over the city of Paris. This represents a much wider and faster deployment than dedicated infrastructures. Our coming studies will consist in considering more flexible charging modes in order to increase the number of accepted EVs. We shall also design pricing strategies as incentive for the drivers to shift their charging periods in order to exploit at the best the day-by-day predictable solar and wind energy sources. REFERENCES [1] A. Foley, I. J. Winning, and B. Gallachoir, State-of-the-art in electric vehicle charging infrastructure, in Vehicle Power and Propulsion Conference (VPPC), 2010 IEEE, 2010, pp [2] Global EV Outlook, Understanding the Electric Vehicle Landscape to 2020, [Online]. Available: [3] V. Zdornov and Y. Birk, Battery and energy management in fleets of switchable battery evs, in Innovative Smart Grid Technologies (ISGT Europe), nd IEEE PES International Conference and Exhibition on, 2011, pp [4] Identification of control and management strategies for {LV} unbalanced microgrids with plugged-in electric vehicles, Electric Power Systems Research, vol. 80, no. 8, pp , [5] J. Mossberg and M. Kuhlich, Adaptability and flexibility of a charging station for electric vehicles, [Online]. Available: [6] K. Mets, T. Verschueren, F. De Turck, and C. Develder, Evaluation of multiple design options for smart charging algorithms, in Communications Workshops (ICC), 2011 IEEE International Conference on, 2011, pp [7] J. Lopes, F. Soares, and P. Almeida, Integration of electric vehicles in the electric power system, Proceedings of the IEEE, vol. 99, no. 1, pp , [8] K. Clement, E. Haesen, and J. Driesen, Coordinated charging of multiple plug-in hybrid electric vehicles in residential distribution grids, in Power Systems Conference and Exposition, PSCE 09. IEEE/PES, 2009, pp [9] P. Richardson, D. Flynn, and A. Keane, Optimal charging of electric vehicles in low-voltage distribution systems, Power Systems, IEEE Transactions on, vol. 27, no. 1, pp , [10] S. Karnouskos, D. Ilic, and P. Silva, Using flexible energy infrastructures for demand response in a smart grid city, in Innovative Smart Grid Technologies (ISGT Europe), rd IEEE PES International Conference and Exhibition on, 2012, pp [11] CITELUM expert in remote lighting management. [Online]. Available: [12] N. Saker, M. Petit, and J.-C. Vannier, Electric vehicles charging scenarios associated to direct load control programs (dlc), in North American Power Symposium (NAPS), 2011, 2011, pp [13] T. S. Oepomo, A step-by-step method for z-loop construction using graph theory and topology for power system studies, in Computational Technologies in Electrical and Electronics Engineering, SIBIR- CON IEEE Region 8 International Conference on, [14] Report of sales of electric vehicles in france in [Online]. Available: [15] Chevrolet volt models and specs. [Online]. Available: [16] Nissan leaf specifications sheet. [Online]. Available: httop:// [17] Toyota prius plug-in hybrid specifications sheet. [Online]. Available: 746