FR3145902A1 - METHOD FOR MANAGING AN ELECTRIC VEHICLE CHARGING STATION - Google Patents

Abstract

The invention relates to a method for managing an electric vehicle charging station, the charging station comprising a plurality of charging terminals, each charging terminal being capable of being connected to an electric vehicle to deliver electrical power to the electric vehicle over time according to a charging profile, in which each time a new electric vehicle enters the charging station, a corrected charging profile is assigned to the incoming electric vehicle as a combination of a basic charging profile that depends on the charging profiles of the electric vehicles already present in the charging station at the time of entry of the incoming electric vehicle, and a correction profile that depends on characteristics of the day during which the incoming electric vehicle entered the charging station, the correction profile having been determined from a history of activity scenarios of the charging station. Figure for abstract: figure 6

Description

METHOD FOR MANAGING AN ELECTRIC VEHICLE CHARGING STATION FIELD OF THE INVENTION

The invention relates to a method for managing an electric vehicle charging station.

STATE OF THE ART

Electric vehicle users generally have a choice regarding the times and locations of charging their electric vehicles.

In order to limit consumption peaks, users of electric vehicles are encouraged to recharge the batteries of electric vehicles during hours of lowest consumption (or “off-peak hours”) and to avoid recharging the batteries of electric vehicles during hours of highest consumption (or “peak hours”). To this end, it is possible to offer these users two tariffs: a first (lower) tariff for electrical energy consumed during off-peak hours and a second (higher) tariff for electrical energy consumed during peak hours.

However, apart from charging during "off-peak" hours, which is encouraged in some places, in public charging infrastructures, users generally have no reason to make choices that would relieve the electricity distribution network and limit consumption peaks.

Two major challenges for infrastructure and charging service operators seeking to encourage greater flexibility among electric vehicle users are:

1. Dealing with a number of parameters and “constraints” associated with charging that are imperfectly known (or known only through statistical models that necessarily involve errors). Examples of such parameters are the charging needs of electric vehicles (i.e. the electrical energy to be added to the battery), the arrival and departure times of electric vehicles.

2. Adapt information and incentives (charging rates for example) to the flexibility/effort actually made by electric vehicles to move towards good charging (participation in better Quality of Service (QoS) for charging, reduction of local/global constraints linked to the electrical system, etc.).

Typically, an electric vehicle present longer at a charging station (for example from 9:00 a.m. to 6:00 p.m.) than another electric vehicle (for example only during the “lunch break” from 12:00 p.m. to 2:00 p.m.) offers more possibilities to “move” the time of its recharge, therefore more flexibility for the operator of the charging station.

Similarly, a user who agrees to charge their electric vehicle at an empty charging station (even if it means going a few hundred metres further) can also offer flexibility to the operator managing a set of charging stations.

One aim of the invention is to propose a method for managing one or more electric vehicle charging stations, making it possible to limit peaks in electrical energy consumption and/or to offer improved quality of service and/or to reduce the impact of charging on the site of the charging station, while providing an efficient charging service for electric vehicles.

This aim is achieved within the framework of the present invention by means of a method for managing one or more electric vehicle charging stations, the charging station(s) comprising a plurality of charging terminals, each charging terminal being capable of being connected to an electric vehicle to deliver electrical power to the electric vehicle over time according to a charging profile, the method comprising a step of:
a – upon entry of an electric vehicle into the charging station, assigning to the incoming electric vehicle a basic charging profile, the basic charging profile comprising a series of electrical power values to be supplied by one of the charging stations to the incoming electric vehicle, each electrical power value being associated with a time slot, the basic charging profile being determined on the basis of a declared electrical energy requirement value of the incoming electric vehicle, a declared exit time of the incoming electric vehicle, residual electrical energy requirement values of electric vehicles already present in the charging station at the time of entry of the incoming electric vehicle, and declared exit times of electric vehicles already present in the charging station,
b – assign to the incoming electric vehicle a correction profile comprising a series of correction values, each correction value being associated with a time slot, the correction profile being determined on the basis of an entry time of the incoming electric vehicle, the values of residual electrical energy requirements of electric vehicles already present in the charging station at the time of entry of the incoming electric vehicle, the declared exit times of electric vehicles already present in the charging station, and a class of the day during which the incoming electric vehicle entered the charging station,
c - apply to each electric power value of the basic charging profile associated with a given time slot, a corrective value of the correction profile associated with the given time slot, so as to generate a corrected charging profile attributed to the incoming electric vehicle,
d - order a recharge of the incoming electric vehicle via the charging station according to the corrected recharge profile assigned to the incoming electric vehicle, and
e – when a new electric vehicle enters the charging station, repeat steps a to d considering the new electric vehicle as the incoming electric vehicle.

The proposed method makes it possible to take into account not only the declared electrical energy requirement of the incoming electric vehicle to calculate the charging profile assigned to the incoming electric vehicle, but also the declared time of exit of the electric vehicle, charging profiles that have been assigned to electric vehicles already present in the station and a class of the day during which the incoming electric vehicle entered the charging station.

In particular, the correction profile may be determined based on historical data that has been previously recorded, this historical data comprising, on the one hand, actual target charging profiles assigned to vehicles that have been recharged in the charging station and, on the other hand, theoretical charging profiles of these same vehicles as they could have been applied to minimize peaks in electrical energy consumption of the charging station or to improve the quality of charging service for users of electric vehicles or to reduce the impact of charging on the site of the charging station (i.e. limit the electricity bill of the site manager, or limit the maintenance and replacement costs of the equipment of this site – for example power transformers), or a combination of these objectives.

Thus, the application of the correction profile can, for example, make it possible to better smooth out peaks in electrical energy consumption at the charging station, while still meeting the electrical energy needs of users.

Furthermore, the method makes it possible to adapt the charging profile assigned to each new vehicle entering the station based on the events that have occurred previously, during a predefined period of time, thanks to the fact that the correction profile assigned to the incoming electric vehicle depends on the class of the day during which the incoming electric vehicle entered the charging station.

The method can advantageously have the following characteristics:

In one possible embodiment, the method comprises steps of:
f – at the same time as step c, update charging profiles assigned to electric vehicles already present in the charging station, and
g – order a recharge of electric vehicles already present, according to the updated recharge profiles.

In one possible embodiment, the corrected charging profile assigned to the incoming electric vehicle is determined based on additional constraints, including for each time slot, a maximum electrical power value that can be provided by the charging station and/or a maximum total electrical power value that can be provided by the charging station during the time slot and/or a value of electrical power consumed by equipment, other than the charging stations, during the time slot.

In one possible embodiment, the base charging profile assigned to the incoming electric vehicle is determined by solving a charging optimization problem having the effect of: minimizing a variance over time of the total electrical power supplied to all the vehicles present in the charging station, or maximizing the quality of charging service of the electric vehicles, or minimizing impact metrics on the charging station, or a combination of these effects.

In one possible embodiment, the method comprises steps of:
h – assigning to the day on which the incoming electric vehicle entered the charging station, a class, from a plurality of predefined classes, by means of a classifier, the days of the same class having common characteristics.

In one possible embodiment, the classifier is configured to assign a class to a day, based on:
- the position of the day in the week, and/or
- information that the day is a public holiday or a working day, and/or
- daily weather data, in particular an average ambient temperature for the day.

In one possible embodiment, the correction profile assigned to the electric vehicle entering step b is generated by a correction profile generation function, the method further comprising steps of:
i – recording an actual scenario of activity of the charging station during a predetermined period of time, the actual scenario of activity of the charging station comprising for each electric vehicle having used the charging station during a predetermined period of time: an arrival time slot, a departure time slot, a value of a declared electrical energy requirement, and a target charging profile comprising a series of electrical power values actually transmitted to a charging terminal of the charging station to control the charging of the electric vehicle,
j – after the predetermined period of time has elapsed, estimating a theoretical scenario of activity of the charging station over the predetermined period of time, the theoretical scenario of activity of the charging station comprising for each electric vehicle having used the charging station during the predetermined period of time: a theoretical charging profile calculated according to the arrival time slots, the departure time slots and the declared electrical energy requirement values of the different electric vehicles having used the charging station during the predetermined period of time, and
k – update parameters of the correction profile generation function based on an error between the actual activity scenario and the theoretical activity scenario.

In one possible embodiment, the predetermined time period has a duration equal to one day.

In one possible embodiment, steps g to i are repeated upon expiration of the predetermined time period, for a subsequent predetermined time period of the same duration.

In one possible embodiment, the method comprises a step of:
j – for each electric vehicle having used the charging station during the predetermined time period, calculate a surplus of electrical energy which would have been delivered to the electric vehicle and/or an anticipated time slot during which the charging of the electric vehicle would have been completed if the theoretical activity scenario estimated in step h had taken place during the predetermined time period.

The invention further relates to a computer program product comprising program code instructions for executing the steps of the method as defined above, when this program is executed by a computer.

The invention also relates to a computer-readable memory storing computer-executable instructions for executing the steps of the method as defined above.

The invention also relates to a module for managing one or more electric vehicle charging stations, the charging station(s) comprising a plurality of charging terminals, each charging terminal being capable of being connected to an electric vehicle and the management module being configured to control each charging terminal so that the charging terminal delivers electrical power to the electric vehicle over time according to a charging profile associated with the electric vehicle, the charging module being configured to:
- upon each entry of an electric vehicle into the charging station, receive data comprising a value of a declared electrical energy requirement of the entering electric vehicle, and a declared departure time slot of the entering electric vehicle,
- assign a charging profile to the incoming electric vehicle, by performing the steps of the process as defined above, and
- once the incoming electric vehicle is connected to one of the charging stations, controlling the charging station so that the charging station delivers electrical power to the incoming electric vehicle according to the charging profile assigned to the incoming electric vehicle.

presentation of drawings

Other features and advantages will emerge from the following description, which is purely illustrative and not limiting, and must be read in conjunction with the attached drawings, among which:

- there schematically represents an electric vehicle charging station,

- there schematically represents steps of a method for managing an electric vehicle research station, in accordance with one embodiment of the invention,

- there schematically represents an example of a charging profile assigned to an electric vehicle,

- there schematically represents steps of a method for updating a correction profile generation function,

- there schematically represents an example of a real scenario of charging station activity and a theoretical scenario of charging station activity,

- there schematically represents an example of implementation of the method for managing the electric vehicle charging station,

- there schematically represents a plurality of charging stations connected to the same electrical energy distribution or transport network.

DETAILED DESCRIPTION OF AN EMBODIMENT

On the , the electric vehicle charging station 1 shown comprises a plurality of charging terminals 2.

Each charging station 2 is connected to a public electrical energy distribution network or to an internal electrical installation network 3. The public electrical distribution network or to an internal electrical installation network 3 is capable of supplying the charging station 2 with electrical energy.

In addition, each charging station 2 is capable of being temporarily connected to an electric vehicle to deliver electrical power to the electric vehicle, over time, according to a predefined charging profile.

Over time, electric vehicles may enter and/or exit charging station 1.

In the example illustrated on the , an electric vehicle 4 (referred to as an “entering electric vehicle”) enters the charging station 1 for recharging.

Several electric vehicles 5 are already present in the charging station. Each of the vehicles already present 5 is connected to a respective charging station 2. The electric vehicles already present 5 are currently being recharged.

An electric vehicle 6 (called an “outgoing electric vehicle”) leaves the charging station 1, after having been recharged via one of the charging stations 2.

The charging station 1 may further include additional electrical equipment 7, connected to the public electrical energy distribution network or to the internal electrical installation network 3.

Furthermore, the charging station 1 comprises a management center 8. The management center 8 comprises a management module 10, configured to control the different charging stations 2, and a database 11, configured to store data from the charging stations 2.

In the example illustrated on the , the management center 8, including the management module 10 as well as the database 11, is part of the charging station 1. However, in another example, the management module 10 and/or the database 11 might not be part of the charging station 1. In particular, the management center 8 and/or the management module 10 and/or the database 11 may be located remotely from the charging station 1. In addition, the management module 10 may be configured to control charging terminals located in several charging stations.

Each charging station 2 is capable of transmitting a charging request signal S1 to the management module 10.

The charging request signal S1 contains data relating to a value of a declared electrical energy requirement of an electric vehicle and a declared output time of the electric vehicle.

The management module 10 is configured to, based on the data contained in the charging request signal S1, generate a charging profile assigned to the electric vehicle, and to generate a charging control signal S2 intended for the charging terminal 2 having emitted the charging request signal S1. The charging control signal S2 contains data representative of the charging profile assigned to the electric vehicle.

Charging station 2 is configured to charge the electric vehicle according to the charging profile transmitted via the charging control signal S2.

Furthermore, the management module is configured to record in the database 11 data relating to the entry time slots, the exit time slots, the electrical energy requirements of the vehicles, as well as data representative of the charging profiles calculated by the management module for the different electric vehicles.

In practice, the management module 10 comprises one or more processor(s). The management module 10 comprises or is coupled to a memory in which is recorded a management program for the charging station 1 and, where applicable, data manipulated by the program (input data of the program, or output data generated by the program).

There schematically represents steps of a method 100 for managing the charging station 1, implemented by the management module 10 to generate a charging profile.

According to a first step 101, when an electric vehicle 4 enters the charging station 1, the user of the electric vehicle 4 declares a value of an electrical energy requirement (i.e. a quantity of energy with which he wishes to supply the battery of the electric vehicle) and an exit time (i.e. the date and time at which he wishes the charging to be completed).

This declaration can be made, for example, by the user, via an interface of a charging station or other equipment (for example a user terminal, such as a “smartphone” or multifunction mobile), for example by manually entering the value of the electrical energy requirement and the output time.

The values of the declared electrical energy requirement of the incoming electric vehicle and the declared exit time of the electric vehicle are transmitted to the management module 10, for example via the charging request signal S1.

According to a second step 102, the management module 10 assigns to the incoming electric vehicle 4 a charging terminal 2 from among the available charging terminals of the charging station 1. This is done on the basis of a simple rule set in advance by the manager of the station(s); for example (i) in descending order of maximum power of the charging terminals (the electric vehicles arriving first benefiting from a higher maximum charging power); (i.bis) possibly by taking into account the maximum charging power of the electric vehicles so as not to assign them a terminal with a higher maximum power than theirs (if known); (ii) in the case of several stations by applying rule (i) or (i.bis), possibly combined with directing the electric vehicles towards the stations with the greatest number of charging terminals at the time of their arrival.

According to a third step 103, the management module 10 assigns a basic charging profile to the incoming electric vehicle 4.

There schematically represents, in the form of black rectangles, an example of a basic recharge profile calculated by the management module 10.

The basic charging profile comprises a series of electrical power values to be provided by the charging station that has been assigned to the incoming electric vehicle, each electrical power value being associated with a time slot, the basic charging profile being determined based on a value of the declared electrical energy requirement of the incoming electric vehicle 4, the declared exit time of the incoming electric vehicle 4 and values of residual electrical energy requirements of electric vehicles 5 already present in the charging station 1 at the time of entry of the incoming electric vehicle 4, and declared exit times of the electric vehicles 5 already present in the charging station 1.

According to a first example, the basic charging profile can be calculated by the management module 10 by an algorithm for allocating electrical power by “water-filling”. The algorithm for allocating electrical power by water-filling makes it possible to reduce the variance over time of the total electrical power supplied to all the vehicles present in the charging station.

In addition, the basic charging profile can also be calculated based on additional constraints, including for each time slot, a maximum total electrical power value that can be provided by the charging station during the time slot and/or a value of electrical power consumed by equipment (for example the additional electrical equipment 7), other than the charging stations, during the time slot.

According to a second example, the basic charging profile can be calculated by the management module 10 by solving a charging optimization problem seeking to maximize the quality of service for charging electric vehicles. This makes it possible to provide an “improved” quality of service to electric vehicles already present, in terms of the quantity of energy recharged – via a potential surplus of energy compared to the value of the declared electrical energy requirement – and the end of charging – via an early end compared to the value of the declared starting time.

According to a third example, the basic charging profile can be calculated by the management module 10 by solving a charging optimization problem seeking to minimize impact metrics on the electric vehicle charging station, and the site to which it is attached: (i) economic impact via the electricity bill; (ii) physical impact, for example accelerated aging of transformers – resulting in maintenance and equipment replacement costs.

According to a fourth example, the basic charging profile can be calculated by the management module 10 by solving a charging optimization problem seeking to optimize a “combined” - or “aggregated” - metric of the metrics of the three previous examples.

According to a fourth step 104, the management module 10 assigns a correction profile to the incoming electric vehicle 4.

There schematically represents, in the form of dotted rectangles, an example of a correction profile calculated by the management module 10 for the incoming vehicle 4.

The correction profile comprises a series of correction values, each correction value being associated with a time slot. The correction profile is determined based on an entry time of the incoming electric vehicle 4, the values of residual electrical energy requirements of electric vehicles 5 already present in the charging station 1 at the time of entry of the incoming electric vehicle 4, the declared exit times of the electric vehicles 5 already present in the charging station 1, and a class of the day during which the incoming electric vehicle 4 entered the charging station 1.

The correction profile assigned to the incoming electric vehicle 4 can be generated via a correction profile generation function. The correction profile generation function determines the correction profile based on a class to which the day of entry of the incoming vehicle 4 belongs, an entry time of the incoming vehicle 4 and the charging profiles of the vehicles 5 already present in the charging station 1.

The class to which the day of entry of the vehicle belongs can be predefined in advance based on characteristics of that day, for example based on the position of the day in the week and/or information that the day is a public holiday or a working day and/or weather data of the day. The weather data of the day includes an average ambient temperature of the day.

Weather data influences “non-flexible” electrical energy consumption (i.e. electrical energy consumed by equipment, such as additional electrical equipment 7, other than charging stations 2) and the charging needs of electric vehicles, via higher electrical energy consumption “on the road” in the event of extreme temperatures.

Thus, the class of the current day can be chosen from a plurality of predefined classes. The number of predefined classes can be between 5 and 15 classes, for example 10 classes.

The number of “weather categories” (hot/intermediate/cold) is chosen to have enough observed data from the charging station in each category (crossed with the other characteristics of the days), so that the statistical learning to generate the correction profile generation function is “representative”.

These classes can be assigned in advance to each day of the calendar, or alternatively, these classes can be assigned by a classifier.

According to a fifth step 105, the management module 10 generates a corrected charging profile from the basic charging profile and the correction profile assigned to the incoming vehicle 4.

As schematically illustrated in the , the corrected charging profile can be obtained by adding the basic charging profile and the correction profile assigned to the incoming vehicle 4.

Thus, for each time slot, the management module 10 applies to each electrical power value of the basic charging profile associated with the given time slot, the corrective value of the correction profile associated with the given time slot.

Alternatively, the corrected recharge profile can be obtained as follows:

For each time slot, the management module adds each electrical power value of the basic charging profile associated with the given time slot with the corrective value of the correction profile associated with the given time slot, and:
- if the sum of the electrical power value of the basic charging profile and the corrective value of the correction profile is less than or equal to a maximum power value, then the value of the corrected charging profile is considered equal to this sum, and
- if the sum of the electrical power value of the basic charging profile and the corrective value of the correction profile is greater than the maximum power value, then the value of the corrected charging profile is considered equal to the maximum power value.

In other words, for each time slot, the corrected charging profile value is equal to the minimum of the sum of the electrical power value of the basic charging profile and the corrective value of the correction profile, and the maximum power value.

The maximum power value can be the maximum power value that the charging station assigned to the electric vehicle can deliver. This maximum power value is set for each charging station.

This avoids generating a corrected charging profile including power values higher than the maximum power value that can be delivered by the charging station.

According to a sixth step 106, once the incoming electric vehicle 4 is connected to the charging station 2 assigned to it, the management module 10 controls the charging station to recharge the battery of the incoming electric vehicle 4 according to the corrected charging profile assigned to the incoming electric vehicle 4.

For this purpose, the control module 10 transmits to the charging terminal 2 the signal S2 containing data representative of the corrected charging profile.

According to a seventh step 107, upon entry of a new electric vehicle into the charging station 1, steps 101 to 106 are repeated by considering the new electric vehicle as the incoming electric vehicle.

According to a possible implementation of the method 100, during step 104, the management module 10 can also update charging profiles assigned to the electric vehicles 5 already present in the charging station, then during step 106, order a recharge of the electric vehicles already present, according to the updated charging profiles.

There schematically represents steps of a method 110 for generating the correction profile generation function by learning from training data.

The training data comprises a set of charging station activity scenarios, each charging station activity scenario having been recorded for a predetermined period of time. The duration of the predetermined period of time may be one day. For example, the predetermined period of time may start at 00:00 and end at 23:59.

The method 110 may comprise the following steps:

According to a first step 111, the management module 10 records a real activity scenario of the charging station during a predetermined period of time.

The actual activity scenario of the charging station 1 includes for each electric vehicle having used the charging station during the predetermined time period: an arrival time slot, a departure time slot, a value of a declared electrical energy requirement, and a target charging profile including a series of electrical power values as it was actually transmitted to a charging terminal 2 of the charging station to recharge an electric vehicle connected to the charging terminal 2.

The data from this real activity scenario can be recorded by the management module 10, as the day progresses, in the database 11.

According to a second step 112, after the predetermined period of time has elapsed, the management module 10 estimates a theoretical scenario of activity of the charging station over the predetermined period of time.

The theoretical activity scenario of the charging station is the scenario that would have been obtained if the management module 10 had been aware in advance of the events occurring during the predetermined time period, that is to say: the arrival time slots, the departure time slots and the values of the declared electrical energy requirements of the different electric vehicles having used the charging station during the time period.

The theoretical activity scenario of charging station 1 includes, for each electric vehicle having used the charging station during the predetermined period of time, a theoretical charging profile calculated according to the arrival time slots, the departure time slots and the declared electrical energy requirement values of the different electric vehicles having used the charging station during the period of time.

The theoretical charging profiles can be calculated by the management module 10 using an electrical power allocation algorithm by “water-filling” (“water-filling algorithm” in English).

In addition, the theoretical charging profile can also be calculated based on additional constraints, including for each time slot, a maximum total electrical power value that can be provided by the charging station during the time slot and/or a value of electrical power consumed by equipment (for example the additional electrical equipment 7), other than the charging stations 2, during the time slot considered.

There schematically represents an example of a real activity scenario of charging station 1 and a theoretical activity scenario of charging station 1. More precisely, the represents the total power consumed by charging station 1 during each time slot.

According to a third step 113, the management module 10 updates the parameters of the correction profile generation function, based on an error between the actual activity scenario and the theoretical activity scenario.

Steps 111 to 113 are repeated at each new time period. For example, steps 111 to 113 are repeated every day, with the data recorded over the course of a day.

Furthermore, the method 110 may comprise an additional step according to which, for each electric vehicle having used the charging station during the predetermined period of time, the management module 10 calculates a surplus of electrical energy which would have been delivered to each electric vehicle and/or an anticipated time slot during which the charging of the electric vehicle would have been completed if the theoretical activity scenario estimated in step 112 had taken place during the predetermined period of time.

There schematically represents an example of implementation of the method for managing the electric vehicle charging station.

As illustrated in the , the management method comprises on the one hand a first “intra-daily” method, which is implemented during the course of a day, and on the other hand a second method of “training the correction profile generation function”, which is implemented once a day, at the end of the day, from data collected during the past day.

The first “intra-daily” process takes place as follows:

At each entry of an electric vehicle i into the charging station at time ti, the management module receives the data associated with the electric vehicle i, necessary to calculate the charging profile assigned to the electric vehicle i. The data received include in particular a value of a declared electrical energy requirement of the electric vehicle i, and a declared exit time of the electric vehicle i. In addition, the management module already has the following data: values of residual electrical energy requirements of electric vehicles already present in the charging station at the time of entry of the electric vehicle i, and declared exit times of the electric vehicles already present in the charging station, as well as data relating to possible additional constraints.

These additional constraints may include, for each time slot, a maximum total electrical power value that can be provided by the charging station during the time slot and/or a value of electrical power consumed by equipment (for example additional electrical equipment 7), other than the charging stations, during the time slot.

Then the management module calculates a charging profile assigned to the electric vehicle i (step Pi2). This charging profile takes into account the correction profile determined by the correction profile generation function.

In addition, the data associated with the electric vehicle i are recorded in the database (step Q1). These data constitute the real activity scenario l1 of the charging station.

The second process of “training the correction profile generation function” is as follows:

At the end of the day, the management module calculates a theoretical scenario of l2 activity of the charging station (step Q2) from the data associated with the electric vehicles that have been recorded in the database during the day that has just passed.

In addition, the management module calculates an error between the theoretical activity scenario l2 and the real activity scenario l1 of the charging station (step Q3), and updates parameters of the correction profile generation function based on the error l2 – l1 thus obtained.

The correction profile generation function thus updated is then used to calculate correction profiles used the following day, by the first “intra-daily” process to generate corrected recharge profiles.

The method of managing the charging station has been described in relation to a single charging station. However, as illustrated in the , it would be possible to apply the same process to a plurality of charging stations connected to the same electrical energy distribution or transport network.

In addition, each charging station may include one or more fast charging terminals and one or more slow charging terminals. In this case, the characteristics of each charging terminal (fast/slow) are taken into account by the management module to assign a charging profile to an incoming electric vehicle.

Example 1: Algorithm for allocating electrical power by Water-Filling

The management module 10 determines the charging profile of the electric vehicles 4 connected to the charging stations 2 during the time period . This period of time is represented by a set containing a number of time slots: , each time slot having a duration . In this example, the set can represent a day, the duration of a time slot can be two hours, so the number of time slots is 12.

Electric vehicles are divided into different classes . Each class groups together electric vehicles with an arrival time slot and a departure time slot municipalities. More specifically, electric vehicles of the class arrive at the charging station during the time slot and leave the charging station during the departure time slot , and can therefore only be recharged during the time slots . For example, electric vehicles of a class in which can only be recharged during the time slot . The set of time slot pairs is noted . The total electrical energy requirement aggregated across all electric vehicles in the class is noted .

For each class , the operator wishes to determine the power at each time slot aggregated across all vehicles in the class , so that the need for electrical energy is satisfied within the departure time slot , that is to say that . For all , .

Determination of load profiles in the case of a theoretical activity scenario (second method of “training the correction profile generation function”)

When implementing the second method of "training the correction profile generation function", the management module 10 knows all the arrival ranges and departure and the declared electrical energy requirements before the start of the time period .

The aggregated load profile selected by the management module 10 (vector noted ) is the one that minimizes a cost function (whose minimum value is noted ), which also depends on electricity consumption of the charging station, other than consumption due to the charging of electric vehicles and called "non-flexible consumption" (typically consumption of electrical equipment 7 of the charging station or of buildings located on the site of the charging station, other than charging stations 2). This non-flexible consumption may also include electrical energy generation equipment, such as photovoltaic panels. Thus, the term can be negative, meaning that there is more electrical energy produced locally in the charging station than electrical energy consumed by electrical equipment 7 other than the charging stations 2.

The cost function depends on the cost of the total power load during a time slot given, represented by a function increasing and convex. The function may represent the economic mechanism of the charging station or the congestion effects of the local electricity distribution network (losses, impact on the voltage plan, aging of electrical equipment).

The management module 10 calculates the optimal charging profiles by solving the problem following :
such as , (

Determination of load profiles in the case of a real activity scenario (first “intra-daily” process)

When implementing the first “intra-daily” method, the management module 10 does not know in advance all the arrival and departure time slots, nor the electrical energy requirements of the electric vehicles that will enter the charging station 1. The management module 10 has access to the following data: the arrival and departure time slots of an incoming vehicle 4 and a declared value of the electrical energy requirement of the incoming electric vehicle 4, only at the time when the electric vehicle 4 enters the charging station 1 and communicates this information to the management module. Thus, for the entire time period , the management module 10 waits for the arrival of a new electric vehicle in the charging station 1 to update the charging profiles of the electric vehicles present in the charging station.

1. Le module de gestion met à jour les quantités d’énergie restant à charger à partir de cette plage horaire d’arrivée pour chaque instant de départ . Pour chaque , est obtenue à partir du besoin en énergie électrique agrégé sur les véhicules électriques qui sont arrivés au cours de la plage horaire et qui repartiront au cours de la plage horaire , et des besoins en énergie électrique restants pour charger les véhicules électriques qui sont arrivés plus tôt (et qui repartiront également au cours de la plage horaire ).
   A noter que si aucun véhicule électrique n’arrive avant la plage horaire , la quantité est simplement égale à . L’ensemble ordonné (croissant) de plages horaires de départ où est noté . Vu de la plage horaire , il correspond à toutes les plages horaires (futures) de départs pour lesquelles un besoin en énergie électrique non-nul doit être satisfait.
2. Le module de gestion 10 exécute l’Algorithme 1 ci-dessous pour calculer la valeur optimale et des profils de recharge agrégés par classe ( , ), solutions du problème suivant. Ce problème correspond aux prévisions de recharge « infra-journalier » des besoins en énergie électrique résiduels, par classe, restant à charger à l’instant d’arrivée :
   
   tel que , (
   A noter que est la puissance de charge programmée pour la plage horaire et agrégée sur tous les véhicules électriques qui sont arrivés à la station de recharge 1 pendant la plage horaire ou avant et quittent la station de recharge 1 au cours de la plage horaire : . La puissance de charge de ces véhicules électriques pendant la plage horaire peut être mise à jour plus tard.

Each time an electric vehicle arrives during a time slot , the management module 10 proceeds as follows:

1. The management module updates the energy quantities remaining to be loaded from this arrival time slot for each moment of departure . For each , is obtained from the electrical energy requirement aggregated on electric vehicles that arrived during the time slot and who will leave during the time slot , and the remaining electrical energy requirements for charging electric vehicles that arrived earlier (and will also leave during the time slot ).
   Please note that if no electric vehicle arrives before the time slot , the quantity is simply equal to . The ordered (ascending) set of departure time slots Or is noted . Seen from the time slot , it corresponds to all (future) departure time slots for which a non-zero electrical energy requirement must be satisfied.
2. Management module 10 executes Algorithm 1 below to calculate the optimal value and aggregated charging profiles by class ( , ), solutions to the problem next. This problem corresponds to the “intra-daily” recharge forecasts of residual electrical energy needs, by class, remaining to be loaded at time of arrival :
   
   such as , (
   Please note that is the scheduled charging power for the time slot and aggregated across all electric vehicles that arrived at charging station 1 during the time slot or before and leave charging station 1 during the time slot : . The charging power of these electric vehicles during the time slot may be updated later.

Algorithm 1 : (executed now) ) Resolution of (

1. Available data: departure times and electrical energy needs
2. For each starting moment chronologically ordered to execute
3. Calculation of the optimal charging profile of electric vehicles leaving the charging station at the moment using the solution of Water-Filling (see definition 1):
4. Calculation of the minimum cost for the electricity distribution network manager
Output: Charging profiles and cost

Definition 1 :

If is increasing and convex and is ordered increasing, the solution vector of the following recharge problem during the time period :
, such as ( )
East :
, (2)
Or And are such that , with For And And .

This is a standard solution called Water-Filling (“WF”). In this solution, the total load (electrical energy used to charge electric vehicles added to the so-called “non-flexible” electrical energy consumed by other equipment than the charging stations) at each time slot used for charging operations is the same (it has the same “water level”), while the total load is higher in the unused time slots. Equation (2) shows that a function increasing and convex leads to the same optimized recharge profile , which smooths as much as possible the total power profile (consisting of the terms of the recharging of electric vehicles and the terms of non-flexible consumption). It should also be noted that unlike the problem of "training the correction profile generation function" for which an optimization is necessary, the solution here takes an explicit form; it is obtained "immediately".

Algorithm 1 first solves the problem of recharge planning (solution which is written ) introduced in Definition 1 for outgoing electric vehicles that leave the charging station during a first departure time slot , depending on the aggregate electrical energy requirement of the class (a, and the non-flexible vector . Then this standard optimization problem is resolved for outgoing vehicles that leave the charging station during a second departure time slot , depending on and a non-flexible fictitious vector which includes the charging profile of electric vehicles that will have left the charging station earlier (starting with ), and so on in the order of the times of departure from the charging station.

Example 2: Determination of the charging power profiles of electric vehicles by an optimization problem maximizing the quality of service of charging for users of electric vehicles

Temporal and class modeling of electric vehicles by arrival and departure time slots is identical to that described in Example 1. Two resolution times of the recharge optimization problem are also distinguished – identically to Example 1.

Determination of load profiles in the case of a theoretical activity scenario (second method of “training the correction profile generation function”)

The modeling used is identical to that of Example 1, with certain constraints and the optimization objective, which translate in the present example the maximization of the quality of recharging of users of electric vehicles (and not the smoothing of the total consumption of the recharging station under the constraint of satisfaction of the recharging need in Example 1). In this Example 2, the management module 10 calculates the optimal recharging profiles by solving the problem following :
such as , (

and with (i) (importance placed on the “surplus energy” metric, via weighting), (also to encourage early departure), and; (ii) the maximum charging power of the electric vehicle class , taken equal to the sum of the maximum powers of the charging stations which have been allocated to the electric vehicles of this class: (iii) And are parameters which come to normalize the corresponding metric (respectively “surplus energy” and “early departure”), so that for example when , both metrics have equal weight (the values of And can be defined relative to the optimal value of executed the day before).

This problem being a “Linear Program” (LP), and of relatively small size (for hours, as mentioned for example in Example 1, only 12 variables and 25 constraints for each class of electric vehicles ), it can be solved by a known Linear Program solver.

Determination of load profiles in the case of a real activity scenario (first “intra-daily” process)

For the implementation of the first “intra-daily” method, the conditions are the same as for Example 1, namely that the management module 10 does not know in advance all the arrival and departure time slots, nor the electrical energy requirements of the electric vehicles that will enter the charging station 1. The notations used are also identical to those of Example 1. The method to be applied is then the same as in Example 1, replacing the resolution of the optimization problem by that of the optimization problem following :

such as , (

with the maximum charging power of electric vehicles present during the time slot in the charging station and leaving during the time slot , taken equal to the sum of the maximum powers of the charging stations which have been allocated to these electric vehicles.

Just like , this optimization problem can be solved by a known Linear Programming solver. Note that unlike Example 1, an algorithm (like Algorithm 1) is not used here to solve this problem; the solution is done "frontally" (in one go), and not sequentially starting time after starting time.

Example 3: Determination of the charging power profiles of electric vehicles by an optimization problem minimizing the impact of charging on the metrics of the electric vehicle charging station, and on those of the charging site

• lors de l’ « entrainement de la fonction de génération de profils de correction », à remplacer le problème d’optimisation par le problème d’optimisation :

The flow in this Example 3 is identical to that of Example 2, replacing charging service quality metrics (for electric vehicle users) with impact metrics on the charging station and the charging site. As an example, two metrics will be minimized here: (i) the electricity bill of the charging station manager and; (ii) the aging of the (power) transformers of the charging station site. This leads to:

• when "training the correction profile generation function", to replace the optimization problem by the optimization problem :

such as , (

• en « infra-journalier », à remplacer le problème d’optimisation par le problème d’optimisation suivant :

with (i) (importance placed on the “electricity bill” metric, via weighting), (also, to minimize the impact on the electrical equipment of the electric vehicle charging site; in this example the transformer), ; (ii) the price of electricity paid by the operator of the electric vehicle charging station, typically set by the relationship with an electricity supplier; (iii) (respectively ) a function giving an estimate of the temperature of the “hot spot” (or “HS”) according to the total electricity consumption on the site of the charging station (respectively the contribution of the temperature of the transformer over the time slot t to its cumulative aging) ¹ ; (iv) the maximum charging power of the electric vehicle class defined as in Example 2. By taking linear formulations for the functions of (iii), this optimization problem can be solved with known Linear Programming solvers;

• in “intra-daily”, to replace the optimization problem by the optimization problem following :

such as , (

1 The function can be fixed by linearizing a standard function from the literature (cf. Jauregui-Rivera, L., & Tylavsky, DJ (2008). Acceptability of four transformer top-oil thermal models—Part I: Defining metrics. IEEE Transactions on Power Delivery , 23(2), 860-865 for example); the one expressing aging can be taken up in Beaude, O., Lasaulce, S., Hennebel, M., & Mohand-Kaci, I. (2016). Reducing the impact of EV charging operations on the distribution network. IEEE Transactions on Smart Grid , 7 (6), 2666-2679.

Example 4: Determination of the charging power profiles of electric vehicles by an optimization problem combining the optimization of several types of metrics – here for example the quality of service of the charging (for users) and the impact of the charging on the charging station, and on its site (for the operator of the station and/or the site)

The procedure for this Example 4 is identical to that of Example 2 or Example 3, replacing in the “training of the correction profile generation function” process (respectively in the “intra-daily” process) the problem Or ( by an optimization problem combining ( And ( (respectively the problem ( Or ( by an optimization problem combining And ( )).

• concernant le « sens » de l’optimisation (maximisation ou minimisation) : (i) si les deux problèmes « à combiner » ont le même sens, garder ce sens pour le problème « combiné » ; (ii) si les deux problèmes ont un sens distinct (par exemple ( et ( ), choisir arbitrairement l’un des deux sens (par exemple la maximisation) ;
• concernant les variables d’optimisation, prendre l’union des variables des deux problèmes d’optimisation à combiner (en supprimant celles redondantes). Par exemple, comme ici et ( ont les mêmes variables d’optimisation, cela donne :

To apply this Example 4, it remains to combine two optimization problems:

• regarding the "direction" of optimization (maximization or minimization): (i) if the two problems "to be combined" have the same direction, keep this direction for the "combined"problem; (ii) if the two problems have a distinct direction (for example ( And ( ), arbitrarily choose one of the two directions (e.g. maximization);
• regarding the optimization variables, take the union of the variables of the two optimization problems to be combined (removing the redundant ones). For example, as here And ( have the same optimization variables, this gives:

• regarding the “combined” objective function, sum the objective functions of the two problems to be combined, with weightings, and taking the opposite of the objective for a problem whose direction has been reversed. For example, here this objective function is written:

• concernant l’ensemble de contraintes à considérer dans le problème combiné, prendre l’union des ensembles de contraintes des deux problèmes d’optimisation (en supprimant celles redondantes). Par exemple, comme ici les contraintes des problèmes et ( sont identiques, cet ensemble serait :

with (importance placed on the quality of charging service – on the side of electric vehicle users, cf. , (similarly, for charging station and/or site metrics, see ( ), . Note that the opposite is taken regarding the contribution of ( to this “combined” objective function, because the meaning of this optimization problem is minimization (and here the “combined” problem operates a maximization);

• regarding the set of constraints to be considered in the combined problem, take the union of the sets of constraints of the two optimization problems (removing the redundant ones). For example, as here the constraints of the problems And ( are identical, this set would be:

,

The problem is obtained in the same way, by combining ( And . Note that these combined problems directly "inherit" membership in the "Linear Programming"category; they can thus be solved with known solvers.

Obtaining the correction profile generation function:

The correction profile generation function (load) is obtained by a statistical method as follows:

This generation function is chosen from a parameterized family, for example linear in its arguments. As a reminder, a correction profile is specific to the type of day , at the time slot at which the correction is applied (e.g. at each time step where one or more electric vehicles arrive at the charging station), at the departure time step of the electric vehicle considered, as well as the quantities that remains to be recharged (vector indexed on the departure time slots ). And for example, the correction profile can be given by:

• Initialement, tant qu’aucun scénario de réalisation des données de recharge sur une journée n’est disponible, un choix arbitraire est fait pour cette fonction de génération de profils de correction. Par exemple, pour des paramètres donnés, si aucun n’a été précédemment calculé, alors on pose , c’est-à-dire un vecteur nul – de la taille du profil à générer
• Ensuite, à mesure que des scénarios lors des journées de recharge sur la ou les stations de recharge se réalisent, on procède de la manière suivante :
  • Le procédé d’ « d’entrainement de la fonction de génération de profils de correction » permet de calculer les profils de correction (optimaux) qu’il aurait fallu appliquer « théoriquement », i.e. si toutes les données avaient été connues à l’avance. Ces corrections ( ) sont définies comme la différence entre le profil de recharge « théorique » calculé à la fin de la journée ( ), et le profil de recharge qui aurait été effectivement appliqué ce jour-là si aucune correction n’avait été utilisée. Autrement dit, ce profil de recharge effectif étant composé des profils « online » ( ) recalculés au cours de la journée, corrigés par les différentes corrections , on a :

with the , all correction profiles applied on previous days and corresponding to the same parameters , and to any values for the vector . is the set of correction profiles available for a day j and an arrival time slot a . The weightings used for this calculation of barycenter – of vectors – being the inverse of the distance (for example in norm 2) between And .

• Initially, as long as no scenario for realizing recharge data over a day is available, an arbitrary choice is made for this correction profile generation function. For example, for parameters given, if none has not been previously calculated, then we set , i.e. a zero vector – of the size of the profile to be generated
• Then, as scenarios during the charging days at the charging station(s) occur, we proceed as follows:
  • The "training of the correction profile generation function" method allows to calculate the (optimal) correction profiles that should have been applied "theoretically", i.e. if all the data had been known in advance. These corrections ( ) are defined as the difference between the “theoretical” recharge profile calculated at the end of the day ( ), and the recharge profile that would have actually been applied that day if no correction had been used. In other words, this effective recharge profile being composed of the “online” profiles ( ) recalculated during the day, corrected by the various corrections , we have:

.

• A chaque réalisation des arguments de la fonction génération de profils de correction mesurée lors d’une journée de recharge, il est donc possible d’associer ce profil de correction théorique. Ceci vient enrichir une base de couples (réalisation des arguments de la fonction de génération de profils de correction, profil de correction théorique). Autrement dit, pour tous les paramètres parcourus pendant ce jour , on pose :

Please note that is thus dependent on the arguments of the correction profile generation function ; even if not explicitly noted here.

• For each realization of the arguments of the correction profile generation function measured during a recharge day, it is therefore possible to associate this theoretical correction profile. This enriches a base of pairs (realization of the arguments of the correction profile generation function, theoretical correction profile). In other words, for all the parameters traveled during this day , we pose:

• In the particular case where a (with exactly the same parameters ) has already been defined before, we can for example ask:

.

Instead of uniformly weighting all past and current achievements (with coefficients ½ in the previous equation), another example of application would be to apply a rule proportional to the number of achievements (weighting 1 for the current one, and number of past achievements for those achieved so far).

Claims (13)

Method for managing one or more charging stations (1) for electric vehicles, the charging station(s) (1) comprising a plurality of charging terminals (2), each charging terminal (2) being capable of being connected to an electric vehicle (5) to deliver electrical power to the electric vehicle (5) over time according to a charging profile, the method comprising a step of:
a – upon entry of an electric vehicle (4) into the charging station (1), assigning to the incoming electric vehicle (4) a basic charging profile, the basic charging profile comprising a series of electrical power values to be supplied by one of the charging stations (2) to the incoming electric vehicle (4), each electrical power value being associated with a time slot, the basic charging profile being determined as a function of a declared electrical energy requirement value of the incoming electric vehicle, a declared exit time of the incoming electric vehicle (4), residual electrical energy requirement values of electric vehicles (5) already present in the charging station (1) at the time of entry of the incoming electric vehicle (4), and declared exit times of the electric vehicles (5) already present in the charging station,
b – assigning to the incoming electric vehicle (4) a correction profile comprising a series of correction values, each correction value being associated with a time slot, the correction profile being determined as a function of an entry time of the incoming electric vehicle (4), the values of residual electrical energy requirements of electric vehicles (5) already present in the charging station at the time of entry of the incoming electric vehicle (4), the declared exit times of the electric vehicles (5) already present in the charging station (1), and a class of the day during which the incoming electric vehicle (4) entered the charging station (1),
c - applying to each electric power value of the basic charging profile associated with a given time slot, a corrective value of the correction profile associated with the given time slot, so as to generate a corrected charging profile attributed to the incoming electric vehicle (4),
d - order a recharge of the incoming electric vehicle (4) by the charging station (2) according to the corrected recharge profile assigned to the incoming electric vehicle, and
e – when a new electric vehicle enters the charging station, repeat steps a to d considering the new electric vehicle as the incoming electric vehicle. The method of claim 1, comprising the steps of:
f – at the same time as step c, update charging profiles assigned to electric vehicles (5) already present in the charging station (1), and
g – order a recharge of electric vehicles (5) already present, according to the updated recharge profiles. Method according to one of claims 1 and 2, in which the corrected charging profile assigned to the incoming electric vehicle (4) is determined according to additional constraints, including for each time slot, a maximum electrical power value that can be provided by the charging station (2) and/or a maximum total electrical power value that can be provided by the charging station during the time slot and/or a value of electrical power consumed by equipment (7), other than the charging stations (2), during the time slot. Method according to one of claims 1 to 3, in which the basic charging profile assigned to the incoming electric vehicle (4) is determined by solving a charging optimization problem having the effect of: minimizing a variance over time of the total electrical power supplied to all the vehicles present in the charging station, or maximizing the quality of charging service of the electric vehicles, or minimizing impact metrics on the charging station, or a combination of these effects. Method according to one of claims 1 to 4, comprising steps of:
h – assigning to the day on which the incoming electric vehicle (4) entered the charging station, a class, from a plurality of predefined classes, by means of a classifier, the days of the same class having common characteristics. The method of claim 5, wherein the classifier is configured to assign a class to a day, based on:
- the position of the day in the week, and/or
- information that the day is a public holiday or a working day, and/or
- daily weather data, in particular an average ambient temperature for the day. Method according to one of claims 1 to 6, in which the correction profile assigned to the incoming electric vehicle (4) in step b is generated by a correction profile generation function, the method further comprising steps of:
i – recording an actual activity scenario of the charging station (1) for a predetermined period of time, the actual activity scenario of the charging station (1) comprising for each electric vehicle having used the charging station for a predetermined period of time: an arrival time slot, a departure time slot, a value of a declared electrical energy requirement, and a target charging profile comprising a series of electrical power values actually transmitted to a charging terminal (2) of the charging station to control the charging of the electric vehicle,
j – after the predetermined period of time has elapsed, estimating a theoretical scenario of activity of the charging station over the predetermined period of time, the theoretical scenario of activity of the charging station comprising for each electric vehicle having used the charging station (1) during the predetermined period of time: a theoretical charging profile calculated according to the arrival time slots, the departure time slots and the declared electrical energy requirement values of the different electric vehicles having used the charging station during the predetermined period of time, and
k – update parameters of the correction profile generation function based on an error between the actual activity scenario and the theoretical activity scenario. The method of claim 7, wherein the predetermined time period has a duration equal to one day. Method according to one of claims 7 and 8, in which steps g to i are repeated at the end of the predetermined time period, for a following predetermined time period, of the same duration. Method according to one of claims 7 to 9, comprising a step of:
j – for each electric vehicle having used the charging station during the predetermined time period, calculate a surplus of electrical energy which would have been delivered to the electric vehicle and/or an anticipated time slot during which the charging of the electric vehicle would have been completed if the theoretical activity scenario estimated in step h had taken place during the predetermined time period. Computer program product comprising program code instructions for executing the steps of the method according to one of claims 1 to 10, when this program is executed by a computer. Computer readable memory storing computer executable instructions for carrying out the steps of the method according to one of claims 1 to 10. Management module (10) of one or more electric vehicle charging stations, the charging station(s) (1) comprising a plurality of charging terminals (2), each charging terminal (2) being capable of being connected to an electric vehicle (4) and the management module (10) being configured to control each charging terminal (2) so that the charging terminal delivers to the electric vehicle (4) an electrical power over time according to a charging profile associated with the electric vehicle (4), the charging module (10) being configured to:
- upon each entry of an electric vehicle (5) into the charging station (1), receive data comprising a value of a declared electrical energy requirement of the entering electric vehicle, and a declared departure time slot of the entering electric vehicle,
- assigning a charging profile to the incoming electric vehicle, by executing the steps of the method according to one of claims 1 to 10, and
- once the incoming electric vehicle is connected to one of the charging stations, control the charging station so that the charging station delivers electrical power to the incoming electric vehicle according to the charging profile assigned to the incoming electric vehicle.