FR3152621A1 - Method and device for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) and associated computer program - Google Patents

Abstract

Method and device for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) and associated computer program The invention relates to a method (20) for predicting the electrical consumption of recharging a set of electric vehicle(s) within their place of residence, comprising the following steps: - obtaining (22), for a plurality of daily predetermined times, a prior estimate of the number of electric vehicle(s) present in said set at their place of residence; - predicting (24) the electrical consumption associated with the recharging of said set using a predetermined chart (32) indicating the average instantaneous charging power per vehicle present, as a function of said number of electric vehicle(s) present according to said estimate, said predicted electrical consumption corresponding to the multiplication of said average instantaneous charging power per vehicle present by said number of electric vehicle(s) present at their place of residence. Figure for abstract: Figure 2

Description

Method and device for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) and associated computer program

The present invention relates to a method for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence, said method being implemented by an electronic device for predicting the electrical consumption associated with the recharging of said set of electric vehicle(s).

The invention also relates to a computer program comprising software instructions which, when executed by a computer, implement such a method for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence.

The invention also relates to said electronic device for predicting the electrical consumption associated with the recharging of said set of electric vehicle(s).

The present invention relates to the field of management of electrical networks and electricity markets. In particular, the present invention relates to the prediction of electrical consumption linked to the recharging of electric vehicle(s) (also referred to hereinafter as EV), at home and for a fleet of electric vehicle(s) corresponding for example to a predetermined geographical area such as a region, a country, etc., or to a fleet of electric vehicle(s) managed by a given supplier, or to a fleet of electric vehicle(s) controlled by a consumption flexibility aggregator.

Currently, load aggregation methods exist. However, to predict the electricity consumption related to home charging of a predetermined fleet of electric vehicles, few effective solutions have been developed apart from methods based on machine learning which have the disadvantage of requiring a large amount of training data, complex processing and which are unsuitable for predicting the impact of a control signal equivalent to a request to stop charging.

The aim of the invention is then to propose a method for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence, making it possible to overcome the drawbacks of the state of the art by providing an effective solution, of reduced complexity, for predicting such electrical consumption.

To this end, the subject of the invention is a method for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence, said method being implemented by an electronic device for predicting the electrical consumption associated with the recharging of said set of electric vehicle(s). within their place of residence, said method being implemented by an electronic device for predicting the electricity consumption associated with the recharging of said set of electric vehicle(s), and comprising the following steps:

- obtaining, for a plurality of predetermined daily moments, a preliminary estimate of the number of electric vehicle(s) present making up said set of electric vehicle(s) at their place of residence;

- prediction of the electrical consumption associated with the recharging of said set of electric vehicle(s) using a predetermined chart indicating the average instantaneous charging power per vehicle present at its place of residence at a time t, as a function of said number of electric vehicle(s) present at their place of residence at said time t according to said estimate, said predicted electrical consumption corresponding substantially to the multiplication of said average instantaneous charging power per vehicle present at said time t by said number of electric vehicle(s) present at their place of residence at said time t.

Thus, the present invention advantageously proposes using a relevant chart indicating the average instantaneous charging power per vehicle present at its place of residence at a time t, as a function of said number of electric vehicle(s) present at their place of residence at said time t according to said estimate.

Using such a chart is valuable because it is efficient, precise, and relevant, while also being less complex to use, for those involved in electricity networks and electricity markets. This makes it possible to anticipate network problems, better exploit them, optimize the management of a portfolio of production and consumption assets, and address network issues such as load shedding and even load shedding.

According to other advantageous aspects of the invention, the method for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence comprises one or more of the following characteristics, taken in isolation or according to all technically possible combinations:

- said predetermined abacus corresponds to a closed power trajectory defined by a predetermined direction of time travel of said trajectory and by four points, predetermined by learning, each point being defined by: an abscissa, corresponding to a number of electric vehicle(s) present at their place of residence, and an ordinate corresponding to an average instantaneous charging power per vehicle,

said trajectory substantially comprising a part of a half-ellipse connecting the first three points of said four points, the first point of said four points having a minimum abscissa and corresponding to one end of the horizontal semi-axis of said part of a half-ellipse, the second point of said four points corresponding to the vertex of said half-ellipse,

said half-ellipse part being connected, according to the predetermined direction of time travel, to a broken line formed of two segments including:

- the first vertical segment connects the third and fourth points of said four points, said third and fourth points having the same abscissa corresponding to the number of electric vehicle(s) domiciled at their place of domicile and being representative of the decreasing night-time evolution of the average instantaneous charging power per vehicle present;

- the second segment of said broken line connecting the fourth lower point of said first vertical segment to the first point corresponding to said end of the horizontal half-axis of said half-ellipse part;

- each ordinate of said four points is respectively defined by means of a predetermined parametric equation, for the number of electric vehicle(s) present at their place of domicile corresponding to the abscissa of each of the four points, said predetermined parametric equation comprising at least one constant and having as parameter(s) at least one of the elements belonging to the group comprising at least:

- the average capacity of the battery of said electric vehicle(s) domiciled at their place of domicile;

- the average autonomy factor of said electric vehicle(s) domiciled at their place of domicile;

- the average energy consumption per kilometer of said electric vehicle(s) domiciled at their place of domicile;

- the average behavior coefficient of the drivers of said electric vehicle(s) domiciled at their place of domicile;

- the average distance traveled per day by said electric vehicle(s) domiciled at their place of domicile;

each parameter being an unknown of said predetermined parametric equation, said unknown being associated with a weighting factor whose value and the value of said constant are optimized, for N sets of parameter(s) of distinct values, N being an integer, by minimizing the error between the average instantaneous charging power per vehicle obtained by simulation for said N sets of parameter(s) of distinct values and the result of said predetermined parametric equation associated with said N sets of parameter(s);

- the first point of said four points, corresponding to one end of the horizontal half-axis, has an abscissa corresponding to the minimum number of electric vehicle(s) present at their place of domicile, said minimum number being proportional to the number of electric vehicle(s) domiciled at their place of domicile according to a proportionality factor β, such that 0 < β <1, said proportionality factor β being obtained by statistical learning during said obtaining step;

- the other end of the horizontal semi-axis of said semi-ellipse has as abscissa a maximum fictitious number of electric vehicle(s), said maximum number being proportional to the number of electric vehicle(s) domiciled at their place of domicile according to a proportionality factor α, such that 1 ≤ α, said proportionality factor α being the result of another predetermined parametric equation having as parameter(s) at least one of the elements belonging to the group comprising at least:

- the average capacity of the battery of said electric vehicle(s) domiciled at their place of domicile;

- the average autonomy factor of said electric vehicle(s) domiciled at their place of domicile;

- the average energy consumption per kilometer of said electric vehicle(s) domiciled at their place of domicile;

- the average behavior coefficient of the drivers of said electric vehicle(s) domiciled at their place of domicile;

- the average distance traveled per day by said electric vehicle(s) domiciled at their place of domicile;

each parameter being an unknown of said other predetermined parametric equation, said unknown being associated with a weighting factor whose value and the value of said constant are optimized, for M sets of parameter(s) of distinct values, M being an integer, by minimizing the error between the average instantaneous charging power per vehicle obtained by simulation for said M sets of parameter(s) of distinct values and the result of said predetermined parametric equation associated with said M sets of parameter(s);

- the second point of said four points corresponding to the apex of said half-ellipse has an abscissa corresponding to the average number of electric vehicle(s) present at their place of domicile, said average number corresponding to the arithmetic mean between said minimum number and said maximum number of electric vehicle(s) which can be simultaneously present at their place of domicile;

- the power instantaneous average load per vehicle present for any point of the half-ellipse part connecting the first, second and third points is suitable for being expressed in the following form, at a time t:

with :

the number of electric vehicle(s) present at said time t; said average number, said minimum number, and E the average instantaneous charging power per vehicle present corresponding to the ordinate of the second point corresponding to the vertex of said half-ellipse and whose abscissa is equal to said average number of electric vehicle(s) present at their place of residence,

said power E being defined by means of a predetermined parametric equation comprising at least one constant and having as parameter(s) at least one of the elements belonging to the group comprising at least:

- the average capacity of the battery of said electric vehicle(s) domiciled at their place of domicile;

- the average autonomy factor of said electric vehicle(s) domiciled at their place of domicile;

- the average energy consumption per kilometer of said electric vehicle(s) domiciled at their place of domicile;

- the average behavior coefficient of the drivers of said electric vehicle(s) domiciled at their place of domicile;

- the average distance traveled per day by said electric vehicle(s) domiciled at their place of domicile;

each parameter being an unknown of said predetermined parametric equation, said unknown being associated with a weighting factor whose value and the value of said constant are optimized, for N sets of parameter(s) of distinct values, N being an integer, by minimizing the error between the average instantaneous charging power per vehicle obtained by simulation for said N sets of parameter(s) of distinct values and the result of said predetermined parametric equation associated with said N sets of parameter(s);

- the method further comprises:

- receiving at least one signal constituting a request to stop said recharging, said signal comprising the time interval during which said stop is required, said interval being defined by two terminals T ₁ and T ₂ , with T ₂ > T ₁ , and

- a first adaptation of said prediction of electrical consumption, the average instantaneous charging power per vehicle present at its place of residence being adapted to each reception of said signal constituting a request to stop said charging to be zero over said time interval;

- the method further comprises, for t > T ₂ , a second adaptation of said prediction of the electrical consumption with respect to the prediction of the electrical consumption which would have been obtained in the absence of a signal equivalent to a request to stop said recharging, the difference between:

- the average instantaneous charging power P _s (t) per vehicle present associated with the prediction of electrical consumption after T ₂ , and

- the average instantaneous charging power P(t) per vehicle present associated with the prediction of the electrical consumption which would have been obtained in the absence of a signal equivalent to a request to stop said charging;

being a positive decreasing function dependent on T₁-T₂, of the energy W which would have been injected into the electric vehicle(s) in the absence of a signal requesting the stopping of said charging such that , and the average number of electric vehicle(s) present between T₁and T₂ ;

- said decreasing positive function is linear such that: P _s (t) – P(t) = T ₂ ) + with

d ₁ < 0.

The invention also relates to a computer program comprising software instructions which, when executed by a computer, implement such a method for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence as defined above.

The invention also relates to a computer program comprising software instructions which, when executed by a computer, implement such a method for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence as defined above.

The invention also relates to an electronic device for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence of said set of electric vehicle(s), characterized in that it comprises:

- an obtaining module, configured to obtain, for a plurality of predetermined daily moments, a preliminary estimate of the number of electric vehicle(s) present making up said set of electric vehicle(s) at their place of domicile;

- a prediction module, configured to predict the electrical consumption associated with the recharging of said set of electric vehicle(s) using a predetermined chart indicating the average instantaneous charging power per vehicle present at their place of residence at a time t, as a function of said number of electric vehicle(s) present at their place of residence at said time t according to said estimate, said predicted electrical consumption corresponding substantially to the multiplication of said average instantaneous charging power per vehicle present at said time t by said number of electric vehicle(s) present at their place of residence at said time t.

The invention will appear more clearly on reading the description which follows, given solely as a non-limiting example, and made with reference to the drawings in which:

- FIG. 1 thereFIG. 1is a diagram illustrating an electronic device for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence according to one embodiment of the invention;

- FIG. 2 there FIG. 2 is a flowchart of a prediction of the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence according to an embodiment of the invention;

- FIG. 3 there FIG. 3 illustrates an example of an abacus used advantageously according to the present invention;

- FIG. 4 FIG. 5 Figures 4 and 5 illustrate respectively an example of representation of the preliminary estimate of the number of electric vehicle(s) over the course of a day and two examples of average charging power of vehicles present at home in the absence and presence of a signal equivalent to a request to stop charging.

In the remainder of the description, the expression “substantially equal to” is understood as a relationship of equality to plus or minus 10%, that is to say with a variation of at most 10%, more preferably as a relationship of equality to plus or minus 5%, that is to say with a variation of at most 5%.

There FIG. 1 is a schematic representation of an embodiment of an electronic device 10 for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence.

More specifically, such an electronic device 10 firstly comprises an obtaining module 12, configured to obtain, for a plurality of predetermined daily instants, a prior estimate of the number of electric vehicle(s) present making up said set of electric vehicle(s) at their place of residence.

For example, we consider a fleet of 10,000 electric vehicles, each electric vehicle being located at an identified location where it can be recharged.

In addition, such an electronic device 10 also comprises a prediction module 14, configured to predict the electrical consumption associated with the recharging of said set of electric vehicle(s) using a predetermined chart indicating the average instantaneous charging power per vehicle present at its place of residence at a time t, as a function of said number of electric vehicle(s) present at their place of residence at said time t according to said estimate, said predicted electrical consumption corresponding substantially to the multiplication of said average instantaneous charging power per vehicle present at said time t by said number of electric vehicle(s) present at their place of residence at said time t.

As an optional variant not shown on the FIG. 1 , the electronic device 10 also comprises a reception module configured to receive at least one signal serving as a request to stop said recharging, said signal comprising the time interval during which said stop is required, said interval being defined by two terminals T ₁ and T ₂ , with T ₂ > T ₁ .

Said signal serving as a request to stop said charging is not an instantaneous signal for the forecasting module, but it is for the charging station. More precisely, such a signal serving as a request to stop said charging is a time series (similar to the time series indicating an estimate of the number of vehicles present at home, it is a signal input to the forecasting module) which indicates the known or potential requests (for example when the user of this forecasting module wants to determine the optimum stop signal for his activity) to stop charging for the coming minutes/hours.

According to this optional variant not shown, the electronic device 10 also comprises an adaptation module configured to implement a first adaptation of said prediction of electrical consumption, the average instantaneous charging power per vehicle present at its place of residence being adapted to each reception of said signal constituting a request to stop said charging to be zero over said time interval.

According to this optional variant not shown, said adaptation module, as an optional addition, is also configured to implement, for t > T ₂ , a second adaptation of said prediction of the electrical consumption with respect to the prediction of the electrical consumption which would have been obtained in the absence of a signal equivalent to a request to stop said recharging, the difference between, on the one hand, the instantaneous average charging power P _s (t) per vehicle present associated with the prediction of the electrical consumption after T ₂ , and on the other hand, the instantaneous average charging power P (t) per vehicle present associated with the prediction of the electrical consumption which would have been obtained in the absence of a signal equivalent to a request to stop said recharging, being a decreasing positive function dependent on T ₁ -T ₂ , of the energy W which would have been injected into the electric vehicle(s) in the absence of a signal equivalent to a request to stop said recharging such that , and the average number of electric vehicle(s) present between T ₁ and T ₂ .

Optionally, said decreasing positive function is linear such that:

P _s (t) – P(t) = T ₂ ) + with d ₁ < 0.

In the example of the FIG. 1 , the electronic device 10 for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence comprises an information processing unit 16 formed for example of a memory 18 and a processor 20 associated with the memory 18.

In the example of the FIG. 1 , the obtaining module 12, the prediction module 14, as well as, as an optional addition not shown, the receiving module and the adaptation module, are each produced in the form of software, or a software brick, executable by the processor 19. The memory 18 of the electronic device for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence is then capable of storing obtaining software, prediction software, as well as, as an optional addition, receiving software and adaptation software. The processor 19 is then capable of executing each of the software among the obtaining software, the prediction software, as well as, as an optional addition, the receiving software and the adaptation software.

In a variant not shown, the obtaining module, the prediction module, as well as, as an optional addition, the receiving module and the adaptation module, are each produced in the form of a programmable logic component, such as an FPGA (from the EnglishField Programmable Gate Array), or an integrated circuit, such as an ASIC (from EnglishApplication Specific Integrated Circuit).

When the electronic device for predicting the electrical consumption associated with the charging of a set of electric vehicle(s) within their place of residence is produced in the form of one or more software programs, that is to say in the form of a computer program, also called a computer program product, it is also capable of being recorded on a medium, not shown, readable by a computer. The computer-readable medium is, for example, a medium capable of storing electronic instructions and of being coupled to a bus of a computer system. For example, the readable medium is an optical disk, a magneto-optical disk, a ROM memory, a RAM memory, any type of non-volatile memory (for example FLASH or NVRAM) or a magnetic card. A computer program comprising software instructions is then stored on the readable medium.

The following is described in relation to the FIG. 2 , an example of operation of the electronic device 10 for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence as illustrated previously in relation to the FIG. 1 .

More specifically, the method 20 for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of domicile of said set of electric vehicle(s), generally comprises a first step 22 of obtaining OBT, for a plurality of daily predetermined times, a prior estimate of the number of electric vehicle(s) present making up said set of electric vehicle(s) at their place of domicile.

In other words, according to said method 20, it is necessary to have access as “input” to an estimate/forecast of the number of electric vehicles EV present at home at each time.

An example of an estimate obtained during this step 22 is notably illustrated by the FIG. 4 described below.

For example, to obtain such an estimate of the number of electric vehicle(s) N(t), for the fleet considered, present at home at time t, a simulation is used beforehand, for example using a means such as a software tool for simulating vehicle movement or the presence of vehicles on a predetermined typology of sites (home, workplace, public places, etc.).

More specifically, the approach, for example, adopted consists of individually simulating a large number of electric vehicles/users, considering possible journeys, for example: from home to work, then from work to school to carry out a "school pick-up"), then from school to home, etc. which are characterized by distances, durations or speeds, start/end times.

Sociological studies of use make it possible to feed these simulations with databases or descriptions in statistical form, an example of which is the document entitled " The mobility of the French: Panorama from the national transport and travel survey 2008 " , accessible in the journal of the General Commission for Sustainable Development CGDD of December 2010.

According to one example among a plurality of possible examples, such a simulation consists of distinguishing working days from non-working days by associating four possible routes with working days against only two with non-working days.

The four journeys associated with working days include in particular a first journey from home going to work then returning from work to home, a second journey from home going to work then to another destination and returning from this other destination to home, a third journey from home going to another destination and returning from this other destination to home, a fourth journey from home going to a first other destination then a second other destination and returning from this second other destination to home.

For example, "other destination" means a place, separate from home and work, for shopping, formalities, leisure activities, private trips, such as school pick-ups or family visits, to accompany or pick up people there or for business trips.

The two routes associated with non-working days (i.e. when it is not necessary to go to work) then only include the third and fourth routes mentioned above.

According to a particular example, for simplicity, for a given electric vehicle, it is assumed that during working days it is always the same route that is taken, and the same for non-working days. The choice of routes, among the two or four possible routes mentioned above, as well as the related characteristics are determined randomly according to predetermined probability laws which are based on data from sociological surveys such as that of the aforementioned document.

These are: the type of journey, the distances travelled for each journey, the departure time in the morning to go to the workplace or the departure time from the workplace for the first and second journeys mentioned above, the departure time from home for days without work for the third and fourth journeys mentioned above, whether the day is a working day or not, the duration during which the electric vehicle EV remains parked for locations of the “other destination” type.

Thus, the simulation makes it possible to define for each electric vehicle EV its location at each instant of the plurality of predetermined daily instants (i.e. at each simulated time step), and then makes it possible to deduce for each time step the number of electric vehicle(s) EV present at home (i.e. whose location is equal to the “home” value of said electric vehicle EV).

Then, as illustrated by Fig. 2, the method 20 for predicting the electrical consumption according to the present invention comprises a step 24 of predicting PRED the electrical consumption associated with the recharging of said set of electric vehicle(s) using a predetermined chart indicating the average instantaneous charging power per vehicle present at its place of residence at a time t, as a function of said number of electric vehicle(s) present at said place of residence at said time t according to said estimation, said predicted electrical consumption corresponding substantially to the multiplication of said average instantaneous charging power per vehicle present at said time t by said number of electric vehicle(s) present at their place of residence at said time t.

Optionally, as shown in dotted lines on the FIG. 2 , the method 20 for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence further comprises a step 26 of receiving R at least one signal serving as a request to stop said recharging, said signal comprising the time interval during which said stop is required, said interval being defined by two terminals T ₁ and T ₂ , with T ₂ > T ₁ .

According to this option, the reception step 26 is followed by a step 28 of first adaptation A ₁ of said prediction of the electrical consumption, the instantaneous average charging power per vehicle present at its place of residence being adapted to each reception of said signal constituting a request to stop said charging to be zero over said time interval.

In other words, during the time interval between T ₁ and T ₂ , the vehicles are not recharged, which means that the average instantaneous charging power P _s (t) per vehicle is equal to zero during this period, whereas before T ₁ , the average instantaneous charging power P _s (t) per vehicle is equal to the average instantaneous charging power P(t) per vehicle present associated with the prediction of the electrical consumption which would have been obtained in the absence of a signal equivalent to a request to stop said recharging (i.e. without step 26).

As an optional addition to this option, step 28 of first adaptation is further followed for t > T ₂ , by a step 30 of second adaptation A ₂ of said prediction of the electrical consumption with respect to the prediction of the electrical consumption which would have been obtained in the absence of a signal equivalent to a request to stop said recharging, the difference between on the one hand the instantaneous average power P _s (t) of charging per vehicle present associated with the prediction of the electrical consumption after T ₂ , and on the other hand the instantaneous average power P (t) of charging per vehicle present associated with the prediction of the electrical consumption which would have been obtained in the absence of a signal equivalent to a request to stop said recharging, being a decreasing positive function dependent on T ₁ -T ₂ , of the energy W which would have been injected into the electric vehicle(s) in the absence of a signal equivalent to a request to stop said recharging such that:

, and the average number of electric vehicle(s) present between T ₁ and T _{2 .}

In other words, after T ₂ , P _s (t) is calculated based on both P(t) and the energy that should have been injected into the electric vehicles VE during the period T ₁ -T ₂ but which was not due to the control signal acting as a request to stop charging.

According to an optional complementary variant, said decreasing positive function is linear such that: P _s (t) – P(t) = T ₂ ) + with d ₁ < 0 _, with d ₁ and d ₂ which depend on T ₁ -T ₂ , of the energy which would have been injected into the electric vehicles VE if there had not been a control signal equivalent to a request to stop charging and the average number of electric vehicles present between T ₁ and T ₂ .

In other words, we consider that the difference between P _s (t) and P(t) is a positive, decreasing and linear function of time, and that when it reaches zero, we then consider that again P _s (t) = P(t) until receipt of the next control signal equivalent to a request to stop recharging.

Note that in the case where there are two successive periods of control signal, the same approach is applied, considering for example that the first period goes from T ₁ to T ₂ and the second from T ₃ to T ₄ . For the first period from T ₁ to T ₂ , the above-mentioned approach is applied. For the second period from T ₃ to T ₄ , the same approach is applied with one exception, namely that the temporality is considered by means of the difference between on the one hand the average electricity consumption for home charging of all home vehicles in the model over the first period from T ₁ to T 2, and on the other hand the average electricity consumption for home charging of all home vehicles in the model over the second period from T ₃ to T ₄ , in order to find the average electricity consumption for home charging after time T4.

The impact of the optional stages 26 of reception, 28 of first adaptation and 30 of second adaptation will be described in particular subsequently in relation to the FIG. 5 which illustrates two examples of average charging power of vehicles present at home in the absence and presence of a signal equivalent to a request to stop charging.

The implementation of these optional steps makes it possible in particular to consider that at certain times of the day a signal equivalent to a request to stop charging is received and used by the charging system as an order to stop charging. Such a signal corresponds, for example, to a signal representative of a period of high electricity prices (i.e. a peak hour type signal), a signal to reduce consumption (e.g. a reduction of a certain factor, for example expressed as a percentage), or a signal representative of a significant network load which requires an overall reduction in consumption depending on the case at the distribution network or transmission network level, such a signal being suitable for being binary or not with, for example, a request for a gradual reduction in consumption.

There FIG. 3 illustrates an example of an abacus used specifically according to the present invention during step 24 of predicting the FIG. 2 .

More precisely, the predetermined abacus 32 corresponds, according to this example, to a closed power trajectory defined by a predetermined direction of time travel of said trajectory, illustrated by the arrows of the FIG. 3 , and by four points E ₂ , E ₁ , L ₂ and L ₁ , predetermined by learning.

Each point E ₂ , E ₁ , L ₂ and L ₁ is defined by: an abscissa, corresponding to a number N of electric vehicle(s) present at their place of residence, and an ordinate corresponding to an average instantaneous charging power P(kW) per vehicle.

Said trajectory 32 substantially comprises a part of a semi-ellipse connecting the first three points E ₂ , E ₁ , L ₂ of said four points E ₂ , E ₁ , L ₂ and L ₁ .

The first point E ₂ of said four points E ₂ , E ₁ , L ₂ and L ₁ has a minimum abscissa N _min and corresponding to one end of the horizontal half-axis of said part of the half-ellipse.

The second point E ₁ of said four points E ₂ , E ₁ , L ₂ and L ₁ corresponds to the vertex of said half-ellipse.

The said part of the half-ellipse, connecting the points E ₂ , E ₁ , L ₂ , is itself connected, according to the predetermined direction of travel, to a broken line formed of two segments.

The first segment of said broken line, according to the predetermined direction of travel of said trajectory, illustrated by the arrows of the FIG. 3 , is vertical and connects the third and fourth points L ₂ and L ₁ of said four points E ₂ , E ₁ , L ₂ and L _{1 .}

The third and fourth points L ₂ and L ₁ have the same abscissa N _ev corresponding to the number of electric vehicle(s) domiciled at their place of domicile and being representative of the decreasing nocturnal evolution of the average instantaneous charging power per vehicle present (i.e. as they are recharged and consequently no longer require charging compared to the time of their return home).

The number N _ev of electric vehicle(s) domiciled at their place of domicile corresponds to the size of the electric vehicle fleet considered and is known as an input to said process.

The second segment of said broken line, according to the predetermined direction of travel of said trajectory, illustrated by the arrows of the FIG. 3 , connecting the fourth lower point L ₁ of said first vertical segment [L ₂ L ₁ ] to the first point E ₂ corresponding to said minimum abscissa end of the horizontal semi-axis of said half-ellipse part connecting the points E ₂ , E ₁ , L ₂ .

The part, called main, corresponds to the diurnal trajectory connecting the points L ₁ to E ₂ to E ₁ to L ₂ , and is used during the day temporally by following the direction arrows shown on the FIG. 3 The vertical part connecting points L ₂ to L ₁ along the vertical line corresponds to night use between returning and leaving home, and by hypothesis this segment is traveled by default linearly with time.

According to an optional addition, each ordinate F ₁ , E, F ₃ and F ₂ of said four points E ₂ , E ₁ , L ₂ and L ₁ respectively is defined by means of a predetermined parametric equation, for the number of electric vehicle(s) present at their place of domicile corresponding to the abscissa N _min , N _mid , N _ev of each of the four points, said predetermined parametric equation comprising at least one constant and having as parameter(s) at least one of the elements belonging to the group comprising at least:

- the average battery capacity B _avg (kWh) of said electric vehicle(s) domiciled at their place of domicile, corresponding to the average of the battery capacities of the fleet of electric vehicles VE considered, for example between 20 kWh and 70 kWh;

- the average autonomy factor R _avg of said electric vehicle(s) domiciled at their place of domicile, the value of which is for example between 1 and 2;

- the average energy consumption per kilometer D _avg (kWh/km) of said electric vehicle(s) domiciled at their place of domicile, corresponding to the average mileage consumption of the electric vehicles VE in the fleet of electric vehicles considered, and which may possibly be “weather dependent”, for example between 0.12 and 0.20 kWh/km;

- the average behavior coefficient U _avg of the drivers of said electric vehicle(s) domiciled at their place of domicile, for example between 0.8 and 1.5;

- the average distance travelled per day K _avg (km) of said electric vehicle(s) domiciled at their place of domicile, corresponding to the average daily distance travelled by the electric vehicles VE of the fleet of electric vehicles considered, for example between 10 km and 50 km.

It should be noted that the average autonomy factor R _avg and the average behavior coefficient U _avg are notably defined in the document by F. Gonzalez et al. entitled “ Plug-in behavior of electric vehicles users: Insights from a large-scale trial and impacts for grid integration studies ”, eTransportation, vol. 10, p. 100131, Nov. 2021.

More precisely, according to these definitions, the average autonomy factor R_avg corresponds to a number greater than one and indicates the average margin that users take with regard to the urgent need to recharge the electric vehicle EV given the trips planned for the next day. Indeed, drivers do not systematically plug their electric vehicle EV into their terminal when returning home. If a driver considers that the battery is not sufficiently charged for the next day's trips, then in this case he plugs in the electric vehicle EV.

According to these definitions, the average behavior coefficient U _avg then characterizes whether or not a driver plugs in their EV, because in the case where the autonomy is sufficient for the next day's journeys, the driver is not obliged to plug in their EV to its terminal.

Each parameter B _avg , R _avg , D _avg , U _avg or K _avg is characteristic of both the electric vehicle fleet and the behavior of their driver.

Each parameter B _avg , R _avg , D _avg , U _avg or K _avg is treated according to the present invention as an unknown of said predetermined parametric equation, said unknown being associated with a weighting factor whose value and the value of said constant are optimized, for N sets of parameter(s) of distinct values, N being an integer, by minimizing the error between the average instantaneous charging power per vehicle obtained by simulation for said N sets of parameter(s) of distinct values and the result of said predetermined parametric equation associated with said N sets of parameter(s).

More precisely, according to this optional addition, the ordinate F ₁ of the point E ₂ , corresponding to one end of the horizontal half-axis of said part of the half-ellipse, is for example obtained by applying the following first-degree parametric and optimized equation:

F₁= -0.0077 -0.00003 B_avg+ 0.0585 D_avg + 0.0008 U_avg + 0.0009 R_avg+0.0006 K_avg

after optimization using N sets of parameter(s) B _avg , R _avg , D _avg , U _avg or K _avg of distinct values (i.e. at least one parameter value being distinct from one set to another) with for example B _avg values between 20 kWh and 70 kWh, R _avg values between 1 and 2, D _avg values between 0.12 and 0.20 kWh/km, U _avg values between 0.8 and 1.5 and K _avg values between 10 km and 50 km.

Similarly, according to this optional complement, the ordinate E of the point E ₁ corresponding to the vertex of said half-ellipse, is for example obtained by applying the following first degree parametric and optimized equation:

E = -0.6375 -0.0053 B _avg +4.1829 D _avg +0.0132 U _avg +0.1685 R _avg +0.0172 K _avg

after optimization also using the N sets of parameter(s) B _avg , R _avg , D _avg , U _avg or K _avg of distinct values (i.e. at least one parameter value being distinct from one set to another) with for example values of B _avg between 20 kWh and 70 kWh, values of R _avg between 1 and 2, values of D _avg between 0.12 and 0.20 kWh/km, values of U _avg between 0.8 and 1.5 and values of K _avg between 10 km and 50 km.

Note that some parameter sets are likely to provide a negative resulting E value, which is relatively normal because such sets are not realistic. For example, if the value of B _avg is 70 kWh, which roughly corresponds to a fleet that would currently be composed of Tesla electric vehicles, the value of K _avg is 10 km, which corresponds to a low daily distance, and the value of D _avg is 0.12, which corresponds to low consumption, the "average" battery then provides an "average" autonomy of almost two months of use of the electric vehicle, so that combined with U _avg values of 0.8 and R _avg of 1, representative of the fact that drivers of the electric vehicle fleet considered have a strong tendency to wait until the last moment before recharging their electric vehicle, the value of E corresponding to this unrealistic combination of parameters would be such that E = -0.155 kW.

Alternatively, a parametric equation of another degree may be used. For example, the ordinate E of the point E ₁ corresponding to the vertex of said half-ellipse is obtained by applying the following optimized second-degree parametric equation:

E = -0.5821 -0.0081 B _avg +2.7157 D _avg -0.0351 U _avg +0.4888 R _avg +0.0117 K _avg +0.00003 B _avg ² + 4.8245*D _avg ² +0.01988*U _avg ² - 0.1067*R _avg ² + 0.00007*K _avg ² , and this after optimization also using the N sets of parameter(s) B _avg , R _avg , D _avg , U _avg or K _avg of distinct values (i.e. at least one parameter value being distinct from one set to another) with for example values of B _avg between 20 kWh and 70 kWh, values of R _avg between 1 and 2, D _avg values between 0.12 and 0.20 kWh/km, U _avg values between 0.8 and 1.5 and K _avg values between 10 km and 50 km.

Similar to what was previously indicated, according to this optional addition, the ordinate F ₂ of the point L ₁ is for example obtained by applying the following first degree parametric and optimized equation:

F ₂ = -0.1928 +0.0015 B _avg +0.697 D _avg -0.015 U _avg +0.0012 R _avg +0.0025 K _avg

after optimization using the N sets of parameter(s) B _avg , R _avg , D _avg , U _avg or K _avg of distinct values (i.e. at least one parameter value being distinct from one set to another) with for example values of B _avg between 20 kWh and 70 kWh, values of R _avg between 1 and 2, values of D _avg between 0.12 and 0.20 kWh/km, values of U _avg between 0.8 and 1.5 and values of K _avg between 10 km and 50 km.

The ordinate F ₃ of the point L ₂ can be obtained in a similar manner or more simply by considering that the point L ₂ is the point of intersection between the semi-ellipse passing through the points E ₂ and E ₁ and the vertical of abscissa N _ev corresponding to the number of electric vehicle(s) domiciled at their place of domicile, which amounts to the size of the fleet of electric vehicles considered, known as an input to said method.

As an optional addition, the first point E ₂ of said four points E ₂ , E ₁ , L ₂ and L ₁ , this first point E ₂ corresponding to one end of the horizontal half-axis, has an abscissa N _min corresponding to the minimum number of electric vehicle(s) present at their place of domicile, said minimum number being proportional to the number of electric vehicle(s) domiciled at their place of domicile according to a proportionality factor β, such that 0 < β <1, said proportionality factor β being obtained by statistical learning during said obtaining step 22. For example, N _min = β N _ev = 0.1638 N _ev .

According to a particular variant, β is directly obtained from said estimate presented in the form of a time series from which such a minimal number N _min can be directly extracted.

As an optional addition, on the FIG. 3 , the other end E ₃ of the horizontal half-axis of said half-ellipse is also represented although not belonging to the abacus used as such. This other end E ₃ has as abscissa a maximum fictitious number N _max of electric vehicle(s), said maximum number being proportional to the number N _ev of electric vehicle(s) domiciled at their place of domicile according to a proportionality factor α, such that 1 ≤ α. More precisely, said proportionality factor α is the result of another predetermined parametric equation also having as parameter(s) at least one of the elements belonging to the group comprising at least:

- the average capacity of the battery B _avg (kWh) of said electric vehicle(s) domiciled at their place of domicile, for example between 20 kWh and 70 kWh;

- the average autonomy factor R _avg of said electric vehicle(s) domiciled at their place of domicile, for example between 1 and 2;

- the average energy consumption per kilometer D _avg (kWh/km) of said electric vehicle(s) domiciled at their place of domicile, for example between 0.12 and 0.20 kWh/km;

- the average behavior coefficient U _avg of the drivers of said electric vehicle(s) domiciled at their place of domicile, for example between 0.8 and 1.5;

- the average distance traveled per day K _avg (km) of said electric vehicle(s) domiciled at their place of domicile, for example between 10 km and 50 km.

Each parameter B _avg , R _avg , D _avg , U _avg or K _avg is characteristic of both the electric vehicle fleet and the behavior of their driver.

Each parameter B _avg , R _avg , D _avg , U _avg or K _avg is treated according to the present invention as an unknown of said other predetermined parametric equation, said unknown being associated with a weighting factor whose value and the value of said constant are optimized, for M sets of parameter(s) of distinct values, M being an integer, by minimizing the error between the average instantaneous charging power per vehicle obtained by simulation for said M sets of parameter(s) of distinct values and the result of said predetermined parametric equation associated with said M sets of parameter(s).

More precisely, according to this optional addition, the proportionality factor α defining the abscissa N _max of the point E ₃ , namely a maximum fictitious number of electric vehicles, corresponding to one end of the horizontal half-axis of said part of the half-ellipse, is for example obtained by applying the following first-degree parametric and optimized equation:

=0.9533 +0.00065 B _avg +0.1714 D _avg -0.009 U _avg +0.0009 R _avg +0.0007 K _{avg .}

As an optional addition, the second point E ₁ of said four points E ₂ , E ₁ , L ₂ and L ₁ , said second point E ₁ corresponding to the vertex of said half-ellipse has an abscissa N _mid corresponding to the average number of electric vehicle(s) present at their place of residence, said average number N _mid corresponding to the arithmetic mean between said minimum number and said maximum number of electric vehicle(s) that can be simultaneously present at their place of residence. In other words:

As an optional extra, the power instantaneous average load per vehicle present (i.e. the ordinate) for any point of the part of the semi-ellipse connecting the first, second and third points E ₂ , E ₁ and L ₂ is suitable for being expressed in the following form, at a time t:

with :

the number of electric vehicle(s) present at said time t; said average number, said minimum number, and E the average instantaneous charging power per vehicle present corresponding to the ordinate of the second point E ₁ corresponding to the vertex of said half-ellipse and whose abscissa is equal to said average number of electric vehicle(s) present at their place of residence, said power E being defined as described previously by means of a predetermined parametric equation comprising at least one constant and having as parameter(s) at least one of the elements belonging to the group comprising at least B _avg , R _avg , D _avg , U _avg or K _avg as described previously.

Such an equation also makes it possible to determine the ordinate F ₃ of the point L ₂ which is the last point of the part of the half-ellipse taken into account within the abacus used advantageously according to the present invention.

Furthermore, in view of the above, before returning to the moment which corresponds on the abscissa to the number minimum number of electric vehicle(s) EV present in their home (i.e. temporally according to the direction of travel of the abacus 32 on the segment going from point L ₁ to point E ₂ ), the average home charging power P(t) is defined using two linear functions associated respectively with the two segments [L ₂ L ₁ ] then [L ₁ E ₂ ] forming the broken line including the points L ₂ with coordinates ( , F ₃ ), L ₁ of coordinates ( , F ₂ ) and E ₂ with coordinates ( , F ₁ ) _.

The trajectory along the bisected line L ₂ -L ₁ -E ₂ corresponds roughly to the period extending from the end of the evening, when almost all electric vehicles (VE) have returned home, to the time of the morning when there are the fewest vehicles present.

On the first segment [L ₂ L ₁ ] corresponding to the fact that N(t) = N _ev , the average home charging power P(t) is then defined using the following linear function:

with the period during which N(t) is equal to N _ev , which corresponds in particular to the night, when we consider that there is no longer any movement of vehicles.

On the second segment ]L ₁ E ₂ ] corresponding to the fact that N(t) < N _ev the average home charging power P(t) is then defined using the following linear function:

with And .

Thus, for each part of the abacus 32, in particular the half-ellipse part from E ₂ to E ₁ then L ₂ and the broken line part from L ₂ to L ₁ then to E ₂ , the total power for home charging can be predicted at each time step by multiplying the average home charging power P(t) by the number of N(t) electric vehicles EV present at home at each time step.

There FIG. 4 illustrates an example of estimate 33 obtained during step 22 of the FIG. 2 . More precisely, curve 33 represents on the ordinate 34 the number of electric vehicles present at their respective homes over a day whose hours are represented on the abscissa 36. According to this estimate 33, a minimum number N _min of electric vehicles, less than a thousand electric vehicles, is notably estimated around 1 p.m., while the number N _ev of electric vehicles in the fleet considered is of the order of 5,000 electric vehicles as estimated to be present at their homes between midnight and 5 a.m.

There FIG. 5 allows for the respective illustration of two operating modes of the method according to the present invention, with firstly view 38 which represents the average recharging power 40 of the vehicles present at home, in the absence of a signal equivalent to a request to stop said recharging (i.e. without implementation of the optional steps 26, 28 and 30 of the FIG. 2 ).

View 42 illustrates the average recharging power 44 of the vehicles present at home, in the presence of a signal equivalent to a request to stop said recharging (i.e. with implementation of the optional steps 26, 28 and 30 of the FIG. 2 ), said signal corresponding for example to a peak hour type signal from 4 p.m. to 6 p.m. then from 7 p.m. to 9 p.m. Such a view 42 makes it possible to illustrate the impact of such a control signal acting as a request to stop said recharging, in particular the “rebound” or “deferral” effect of consumption between 6 p.m. and 7 p.m., the prediction of which is made possible thanks to the present invention.

Those skilled in the art will understand that the invention is not limited to the embodiments described, nor to the particular examples of the description, the embodiments and variants mentioned above being suitable for being combined with each other to generate new embodiments of the invention.

The present invention thus makes it possible to effectively predict the electricity consumption associated with the recharging of a fleet of electric vehicles, which is useful for distribution system operator (DSO) or transport system operator (TSO) managers who can, based on this consumption, predict the network load and balance production versus consumption.

Such a prediction is also useful for flexibility aggregators, able to predict the impact of a control signal equivalent to a request to stop said charging on the consumption profile of a fleet of EV electric vehicles.

Moreover, such a prediction is useful for suppliers to anticipate their needs for electrical energy to purchase on the markets, or for traders on the electricity market to anticipate changes in electricity prices.

In other words, the prediction of the load chart specifically used according to the present invention, and optionally of the impact on the latter of a control signal equivalent to a request to stop said charging, has value for the players in the electricity networks and electricity markets who can thus anticipate problems on the networks, better exploit the networks, or even optimize the management of a portfolio of production or flexibility assets on the electricity networks.

Claims (12)

Method (20) for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence, said method being implemented by an electronic device for predicting the electrical consumption associated with the recharging of said set of electric vehicle(s), and comprising the following steps:
- obtaining (22), for a plurality of daily predetermined times, a prior estimate of the number of electric vehicle(s) present making up said set of electric vehicle(s) at their place of residence;
- prediction (24) of the electrical consumption associated with the recharging of said set of electric vehicle(s) using a predetermined chart (32) indicating the average instantaneous charging power per vehicle present at its place of residence at a time t, as a function of said number of electric vehicle(s) present at their place of residence at said time t according to said estimate, said predicted electrical consumption corresponding substantially to the multiplication of said average instantaneous charging power per vehicle present at said time t by said number of electric vehicle(s) present at their place of residence at said time t. Method (20) according to claim 1, in which said predetermined abacus (32) corresponds to a closed power trajectory defined by a predetermined direction of time travel of said trajectory and by four points (E ₂ , E ₁ , L ₂ , L ₁ ), predetermined by learning, each point being defined by: an abscissa, corresponding to a number of electric vehicle(s) present at their place of residence, and an ordinate corresponding to an average instantaneous charging power per vehicle,
said trajectory (32) substantially comprising a part of a half-ellipse connecting the first three points of said four points (E ₂ , E ₁ , L ₂ , L ₁ ), the first point (E ₂ ) of said four points, having a minimum abscissa and corresponding to one end of the horizontal semi-axis of said part of a half-ellipse, the second point (E ₁ ) of said four points (E ₂ , E ₁ , L ₂ , L ₁ ) corresponding to the vertex of said half-ellipse,
said half-ellipse part being connected, according to the predetermined direction of time travel, to a broken line formed of two segments including:
- the first vertical segment connects the third (L ₂ ) and fourth (L ₁ ) points of said four points, said third and fourth points (L ₂ , L ₁ ) having the same abscissa corresponding to the number of electric vehicle(s) domiciled at their place of domicile and being representative of the decreasing nocturnal evolution of the average instantaneous charging power per vehicle present;
- the second segment of said broken line connecting the fourth lower point of said first vertical segment to the first point corresponding to said end of the horizontal half-axis of said half-ellipse part. Method (20) according to claim 2, in which each ordinate of said four points is respectively defined by means of a predetermined parametric equation, for the number of electric vehicle(s) present at their place of domicile corresponding to the abscissa of each of the four points, said predetermined parametric equation comprising at least one constant and having as parameter(s) at least one of the elements belonging to the group comprising at least:
- the average capacity of the battery of said electric vehicle(s) domiciled at their place of domicile;
- the average autonomy factor of said electric vehicle(s) domiciled at their place of domicile;
- the average energy consumption per kilometer of said electric vehicle(s) domiciled at their place of domicile;
- the average behavior coefficient of the drivers of said electric vehicle(s) domiciled at their place of domicile;
- the average distance traveled per day by said electric vehicle(s) domiciled at their place of domicile;
each parameter being an unknown of said predetermined parametric equation, said unknown being associated with a weighting factor whose value and the value of said constant are optimized, for N sets of parameter(s) of distinct values, N being an integer, by minimizing the error between the average instantaneous charging power per vehicle obtained by simulation for said N sets of parameter(s) of distinct values and the result of said predetermined parametric equation associated with said N sets of parameter(s). Method (20) according to claim 2 or 3, in which the first point (E ₂ ) of said four points, corresponding to one end of the horizontal half-axis, has an abscissa corresponding to the minimum number of electric vehicle(s) present at their place of domicile, said minimum number being proportional to the number of electric vehicle(s) domiciled at their place of domicile according to a proportionality factor β, such that 0 < β <1, said proportionality factor β being obtained by statistical learning during said obtaining step. Method (20) according to any one of claims 2 to 4, in which the other end (E ₃ ) of the horizontal semi-axis of said semi-ellipse has as abscissa a maximum fictitious number of electric vehicle(s), said maximum number being proportional to the number of electric vehicle(s) domiciled at their place of domicile according to a proportionality factor α, such that 1 ≤ α, said proportionality factor α being the result of another predetermined parametric equation having as parameter(s) at least one of the elements belonging to the group comprising at least:
- the average capacity of the battery of said electric vehicle(s) domiciled at their place of domicile;
- the average autonomy factor of said electric vehicle(s) domiciled at their place of domicile;
- the average energy consumption per kilometer of said electric vehicle(s) domiciled at their place of domicile;
- the average behavior coefficient of the drivers of said electric vehicle(s) domiciled at their place of domicile;
- the average distance traveled per day by said electric vehicle(s) domiciled at their place of domicile;
each parameter being an unknown of said other predetermined parametric equation, said unknown being associated with a weighting factor whose value and the value of said constant are optimized, for M sets of parameter(s) of distinct values, M being an integer, by minimizing the error between the average instantaneous charging power per vehicle obtained by simulation for said M sets of parameter(s) of distinct values and the result of said predetermined parametric equation associated with said M sets of parameter(s). Method (20) according to claims 4 and 5, in which the second point (E ₁ ) of said four points corresponding to the vertex of said half-ellipse has an abscissa corresponding to the average number of electric vehicle(s) present at their place of residence, said average number corresponding to the arithmetic mean between said minimum number and said maximum number of electric vehicle(s) which can be simultaneously present at their place of residence. The method (20) of claim 6, wherein the power instantaneous average load per vehicle present for any point of the half-ellipse part connecting the first, second and third points is suitable for being expressed in the following form, at a time t:

with :
the number of electric vehicle(s) present at said time t; said average number, said minimum number, and E the average instantaneous charging power per vehicle present corresponding to the ordinate of the second point corresponding to the vertex of said half-ellipse and whose abscissa is equal to said average number of electric vehicle(s) present at their place of residence,
said power E being defined by means of a predetermined parametric equation comprising at least one constant and having as parameter(s) at least one of the elements belonging to the group comprising at least:
- the average capacity of the battery of said electric vehicle(s) domiciled at their place of domicile;
- the average autonomy factor of said electric vehicle(s) domiciled at their place of domicile;
- the average energy consumption per kilometer of said electric vehicle(s) domiciled at their place of domicile;
- the average behavior coefficient of the drivers of said electric vehicle(s) domiciled at their place of domicile;
- the average distance traveled per day by said electric vehicle(s) domiciled at their place of domicile;
each parameter being an unknown of said predetermined parametric equation, said unknown being associated with a weighting factor whose value and the value of said constant are optimized, for N sets of parameter(s) of distinct values, N being an integer, by minimizing the error between the average instantaneous charging power per vehicle obtained by simulation for said N sets of parameter(s) of distinct values and the result of said predetermined parametric equation associated with said N sets of parameter(s). A method (20) according to any preceding claim further comprising:
- receiving (26) at least one signal serving as a request to stop said recharging, said signal comprising the time interval during which said stop is required, said interval being defined by two terminals T ₁ and T ₂ , with T ₂ > T ₁ , and
- a first adaptation (28) of said prediction of electrical consumption, the average instantaneous charging power per vehicle present at its place of residence being adapted to each reception of said signal constituting a request to stop said charging to be zero over said time interval. Method (20) according to claim 8, further comprising, for t > T ₂ , a second adaptation (30) of said prediction of the electrical consumption with respect to the prediction of the electrical consumption which would have been obtained in the absence of a signal equivalent to a request to stop said recharging, the difference between:
- the average instantaneous charging power P _s (t) per vehicle present associated with the prediction of electrical consumption after T ₂ , and
- the average instantaneous charging power P(t) per vehicle present associated with the prediction of the electrical consumption which would have been obtained in the absence of a signal equivalent to a request to stop said charging;
being a decreasing positive function dependent on T ₁ -T ₂ , of the energy W which would have been injected into the electric vehicle(s) in the absence of a signal equivalent to a request to stop said recharging such that , and the average number of electric vehicle(s) present between T ₁ and T _{2 .} Method (20) according to claim 9, wherein said decreasing positive function is linear such that: P _s (t) – P(t) = T ₂ ) + with d ₁ < 0 _. Computer program comprising software instructions which, when executed by a computer, implement a method for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of domicile of said set of electric vehicle(s) according to any one of the preceding claims. Device (10) for predicting the electrical consumption associated with the recharging of a set of electric vehicle(s) within their place of residence of said set of electric vehicle(s) characterized in that it comprises:
- an obtaining module (12), configured to obtain, for a plurality of daily predetermined times, a prior estimate of the number of electric vehicle(s) present making up said set of electric vehicle(s) at their place of residence;
- a prediction module (14), configured to predict the electrical consumption associated with the recharging of said set of electric vehicle(s) using a predetermined chart indicating the average instantaneous charging power per vehicle present at their place of residence at a time t, as a function of said number of electric vehicle(s) present at their place of residence at said time t according to said estimate, said predicted electrical consumption corresponding substantially to the multiplication of said average instantaneous charging power per vehicle present at said time t by said number of electric vehicle(s) present at their place of residence at said time t.