FR3102251A1 - Method for optimizing the recharging and / or discharging of batteries for an electric motor vehicle - Google Patents

Abstract

The invention relates to a method of planning an electric recharging of an electric motor vehicle (2) aimed at achieving economical and efficient recharging, comprising steps of pre-processing, optimization, and post-processing in which charging control signals are generated. Figure to be published with the abstract: Fig. 1

Description

Method for optimizing the charging and/or discharging of batteries for an electric motor vehicle

The invention relates to the field of recharging electric accumulators of electric or hybrid motor vehicles.

Lithium batteries are now widely used in electric and hybrid vehicles thanks to their high energy and power density. Nowadays, the charging of electric vehicles is mainly done in a simplistic way. As soon as the vehicle is connected to the electrical network, the battery is charged to the maximum of the power accepted by the battery or supplied by the network, until reaching a full charge. No vehicle load planning is implemented.

With the fluctuation of energy prices over the day, the planning of recharging and/or discharging is an important issue, because it will allow the minimization of a performance criterion, such as the cost of recharging. One of the main effects being the increase in the attractiveness of electric vehicles and the reduction in the overall cost of operating these vehicles.

Extreme outdoor temperatures, 40°C higher and -10°C lower, accelerate the aging and loss of capacity of lithium batteries. Thus, battery charging, taking temperature into account, is also a crucial issue, in order to extend the life of electric vehicle batteries.

There are on-board charging planning algorithms, aimed at determining the optimal battery charging periods.

But, in a context where energy tariffs for recharging electric vehicles can vary to the nearest second, the on-board scheduling algorithms with a constant calculation step turn out to be incapable of carrying out this optimization task because of the large number of decision variables implied by these too many time steps.

Also, there is a general need to optimize the charging of electric motor vehicles in order to obtain economical charging.

Document EP3054552 A1 is known for this purpose, which discloses an apparatus and a method for economical recharging of a motor vehicle. This document discloses in particular a method in which a charge slope is calculated, which represents the speed with which the battery of an electric vehicle is charged, on the basis of information on the actually measured state of charge of the vehicle. This method further calculates a load plan based on the load slope and a load power rate for each time slot. However, this method, which takes into account the charging power of the electrical network, is not optimal from the point of view of the motor vehicle.

Fast charging is also known to negatively impact key considerations for battery life goals and the network. Charging high-performance batteries involves many tweaks to maximize the sometimes conflicting goals of battery life, performance, and maximum availability.

For user convenience, fast chargers have been designed and implemented for public charging stations. These chargers are designed to quickly restore users' access to their electric vehicles. This improved charging speed comes at a potential cost of degraded battery life.

For many electric vehicle users, charging their vehicles as quickly as possible is considered extremely important, and these users choose the fastest charging option whenever possible, even when a slower charging option may be more economical, and/or better for the battery.

Also, there is a need for an economical system and method for recharging a motor vehicle, allowing planning and optimization of the recharging power according to the needs of the user.

For this purpose, a method for planning an electric recharge and/or an electric discharge of an electric motor vehicle is proposed, comprising:

• Une sous-étape d’acquisition d’une pluralité de données de modélisation ;
• Une sous-étape de subdivision au cours de laquelle on subdivise une durée de disponibilité prédéterminée, en une séquence de segments de temps ; la subdivision étant calculée de telle sorte que, pour au moins une donnée de modélisation choisie, ladite donnée de modélisation choisie est constante dans chaque segment de temps ; la subdivision temporelle de ladite donnée de modélisation en un vecteur dont la dimension est égale au nombre de segments subdivisé formant vecteur de données de modélisation,

A pre-processing step comprising:

• A sub-step of acquisition of a plurality of modeling data;
• A subdivision sub-step during which a predetermined availability duration is subdivided into a sequence of time segments; the subdivision being calculated such that, for at least one chosen modeling datum, said chosen modeling datum is constant in each time segment; the temporal subdivision of said modeling data into a vector whose dimension is equal to the number of subdivided segments forming a vector of modeling data,

An optimization step during which an optimal set of decision vectors is determined, by calculating an objective function as a function of said modeling data vectors, said decision vectors comprising decision variables, i.e. i.e. adjustable parameters capable of modifying the cost of said electrical recharge and/or discharge, and each being of the same dimension as said modeling data vectors,

A post-processing step in which charging and/or discharging control signals are generated based on said optimal set of decision vectors and said modeling data.

By this method, it is possible to determine in a relatively simple and fast way an economical and efficient planning of the electric recharging of an electric motor vehicle.

Advantageously and in a non-limiting way, a chosen modeling data item comprises, for the predetermined availability duration, the representative data:

maximum charging power

the maximum possible discharge power;

charging energy tariffs; And

discharge fees.

Thus, the modeling data are relatively simple to process and few in number.

Advantageously and in a non-limiting manner, the optimization step is also a function of a plurality of functions of technical constraints and of use, the optimization step comprising a sub-step for verifying the conformity of said optimal set of decision vectors with said constraints. Thus, it can be ensured that the planned planning meets the functional constraints of the motor vehicle.

Advantageously and in a non-limiting manner, said optimal set of decision vectors is determined as being the set of decision vectors minimizing said objective function. Thus, the step of determining said optimal set of decision vectors involves a relatively simple and inexpensive optimization in computation time.

Advantageously and in a non-limiting manner, said objective function corresponds to the equation: In which : a recharge energy cost value an energy discharge cost value
c1 and c2 are predetermined weighting coefficients,
X is said set of decision vectors,
T the size of said decision vector,
and t is the index corresponding to a time segment for the corresponding values of said set of decision vectors.

This function makes it possible in particular to make the modeling rapid and the planning obtained particularly reliable and economical.

The invention also relates to a device for planning an electric recharge and/or an electric discharge of an electric motor vehicle comprising:

• Acquérir une pluralité de données de modélisation ;
• Subdiviser une durée de disponibilité prédéterminée, en une séquence de segments de temps ; la subdivision étant calculée de telle sorte que, pour au moins une donnée de modélisation choisie, ladite donnée de modélisation choisie est constante dans chaque segment de temps ; la subdivision temporelle de ladite donnée de modélisation en un vecteur dont la dimension est égale au nombre de segments subdivisés formant un vecteur de données de modélisation,

Pre-treatment means suitable for:

• Acquire a plurality of modeling data;
• Subdivide a predetermined uptime into a sequence of time segments; the subdivision being calculated such that, for at least one chosen modeling datum, said chosen modeling datum is constant in each time segment; the temporal subdivision of said modeling data into a vector whose dimension is equal to the number of subdivided segments forming a vector of modeling data,

Optimization means capable of determining an optimal set of decision vectors, by calculating an objective function as a function of said data vectors, said decision vectors comprising decision variables, that is to say parameters adjustable capable of modifying the cost of said electrical recharge and/or discharge, and each being of the same dimension as said modeling data vectors,

Post-processing means capable of generating recharging and/or discharging control signals as a function of said optimal set of decision vectors and of said modeling data.

The invention also relates to an electric motor vehicle comprising at least one electric storage battery, and a planning device as described above.

The invention also relates to an electrical assembly comprising an electric motor vehicle as described above and a charging and/or discharging terminal.

Other features and advantages of the invention will become apparent on reading the description given below of a particular embodiment of the invention, given by way of indication but not limitation, with reference to the appended drawings in which:

is a schematic view of an electrical assembly according to one embodiment of the invention;

is a schematic flowchart of a planning method according to one embodiment of the invention;

is an example of the result of the implementation of the subdivision sub-step of the method of FIG. 2;

is a flowchart of the sub-steps of the optimization step of the method according to FIG. 2.

According to a first embodiment of the invention, with reference to FIGS. 1 to 4, an electrical assembly 1 comprises an electric motor vehicle 2 and a charging terminal 3 able to recharge the electric accumulators 21 of the motor vehicle 2 or to recover the energy accumulated in the motor vehicle, what is called in the present invention the discharge.

In order to allow the motor vehicle to be recharged, the latter comprises means of connection to the charging terminal comprising in particular a charge controller 22.

This charge controller cooperates with a recharging planning device 23 implementing a recharging optimization and planning process 30 also called more simply planning process 30.

In the existing embedded planning processes in the state of the art, the calculation step is fixed. In other words, the calculation step Δt is predetermined and fixed at the start before starting the process.

In the present invention, the planning method 30 comprises a variable calculation step, in other words a variable Δt, this Δt is chosen in an optimized manner and can vary from one calculation step to another.

To this end, the planning method 30 comprises a first pre-processing step 31 implementing a first sub-step for acquiring data useful for optimization and planning, also called modeling data.

These modeling data include in particular data acquired from the battery management device 21 of the motor vehicle 2 as well as from the charging station 3.

In particular, the battery management device 21 acquires the measurement of the initial temperature of the batteries, the outside temperature, the initial state of charge, better known by the English name of State of Charge (SoC), as well as the capacity drums.

We also acquire from the charging station 3 the values (with their reference used in the description):

- The maximum charging power available over 24h: PG2V_max depending on time;

- The maximum possible discharge power ( Grid support ) over 24 hours: PV2G_max depending on time;

- Charging energy tariffs over 24 hours: price_G2V depending on time;

- 24-hour discharge remuneration: price_V2G as a function of time.

The charging station 3 sends the data over a maximum availability period of 24 hours, this step restricts the time base to the actual period of availability of the electric vehicle on the charging station, i.e. between the time arrival and departure of the vehicle. However, these durations are given on a non-limiting basis and can be freely adapted by those skilled in the art.

In particular, the duration of availability may differ to be greater or less than 24 hours.

The pre-processing step then implements a second subdivision sub-step 312 during which the charging time interval is subdivided 3121 into a sequence of time segments 3122 for which the maximum charging power PG2V_max, the power of maximum discharge PV2G_max, the charging energy tariff prix_G2V and the discharge remuneration prix_V2G are constant, as represented in figure 3'.

This subdivision step 312 thus calculates the size T of the decision vectors, corresponding to the number of segments 3122 determined.

In the example of Figure 3', which is not limiting and simply aims to illustrate the method 30, there are thus 10 segments 3122, forming modeling data vectors such as these:

PG2V_max = [10, 10 10,10, 1, 3, 6, 6, 6, 6];

PV2G_max = [5, 2 1.10, 10, 2, 2, 3, 3, 6]

Price_G2V = [0.1, 0.1, 0.1, 0.2, 0.2, 0.3, 0.3, 0.3, 0.04, 0.04]

V2G Price = [0.1, 0.1, 0.1, 0.25, 0.25, 0.3, 0.3, 0.3, 0.15, 0.15]

The number of time slots serving as the basis for the calculations is therefore here T = 10.

Also, the preprocessing step 31 makes it possible to prepare the implementation of the following steps by acquiring the necessary data and parameters.

In this step, the upper bounds of the decision variables and the optimization parameters are defined.

A subdivision into time segments 3122 is used in which all modeling data has a constant value. In our case, the following four modeling data must be constant, namely: PG2V_max, PV2G_max, price_G2V and price_G2G.

However, it is possible to extend this preprocessing step to apply it to a problem with more variables and parameters and to a segmentation comprising a larger number of parameters.

At the end of this preprocessing step 31, the time base is constructed, the values Δt_max, PG2V_max, PV2G_max are determined, and the parameters Prix_G2V, Prix_V2G are then known. Δt_max is a vector of size T comprising the T values of the time intervals of the subdivision 3122. Indeed the following step will determine on each of these intervals an optimized value of the charging or discharging time interval whose value will be less than Δt_max _t . In this request, the index t corresponds to the index of a value corresponding to a time slot t inside a vector of size T, t varying from 1 to T.

Once the preprocessing step 31 is completed, an optimization step 32 is implemented.

The optimization step 32 comprises the modeling of the electric vehicle charging/discharging problem as a function of the useful data acquired previously and as a function of the modeling data vectors calculated previously.

Modeling means the combination of input parameters and decision variables, obtained during the preprocessing step, into an objective function 42 as well as a set of imposed constraints 43.

The objective here is to optimize 32 the output of this model, in other words the optimal value obtained by the objective function 42.

In a first part of the optimization step 32, a set of decision vectors X ₀ is initialized 41 with X ₀ =[PG2V_max,0,0,0]. This set of optimal decision vectors X is indeed a set of 4 vectors of size T, that is [ , the vector PG2V comprising T charging power values, the vector PV2G comprising T discharge power values, the vector comprising T values of charging time intervals, the vector comprising T discharge time interval values, which optimize the objective function described below,

Then we try to evaluate the following objective function 42:

Where X =[ is a set or matrix of 4 vectors of size T, and :X _t = [ is a vector of size 4,

: Charging power

: The discharge power

charging and discharging calculation steps

Charging energy cost

Energy discharge cost

c1 and c2 are empirical weighting coefficients determined by those skilled in the art.

T is here the size of the decision vectors, i.e. T=10 in the example of figure 3.

This optimization problem is for example formulated using the Fmincon solver from Matlab®, and the resolution is carried out with the SQP (Sequential Quadratic Programming) algorithm.

This modeling necessarily having to meet predefined technical and usage constraints, a constraint verification step 43 is implemented, which are in this embodiment, and without limitation, the following:

- Adoption of the following convention (charge/discharge difference) : ;

- Duration limitation for each segment 3122: ; And duration of each segment for t varying from 1 to T

- Minimum charger operating time: = min of durations or 0.1h.

- Maximum power: ;

: maximum charging power for each segment

: maximum discharge power for each segment

- Minimum charger power: Or Or

- Charging or discharging: i.e. Or

- Power limitation by the battery: Here the "Powermap" function a function that allows to estimate the amount of power accepted by the battery according to the temperature of the battery and the SOC. For example at 5°C and a SOC of 10% the battery can accept 15kW and at -10°C and the same SOC of 10% the battery can only accept 5kW. It is indicated in more detail in the Renault patent EP2878034. The values of this function depend on the type of battery. Of course, as a variant, a map can be used.

- Limitation of the state of charge (SOC):

;

- Minimum state of charge, determined according to the usual use of the motor vehicle, for example according to the work-home distance for daily use: = from 0 to 1

- Respect for the need for mobility:

The need for mobility here is the customer's demand in terms of energy distance. This need can be expressed according to a percentage of SOC that the customer wants his battery to be recharged (for example: 20% of the SOC, to go from 40% to 60% at the end of the charge) or according to an amount of energy to charge (for example: 10kWh to go from 30kWh to 40kWh at the end of the charge.)

= participation authorized by the customer

The customer authorized participation is the amount of energy that the customer authorizes to be used to participate in the V2G (vehicle-to-grid) market. For example: A customer with a 52kWh battery can allow 5kWh of their battery to be used for V2G.

Thus, we can obtain an optimization of the planning by solving the optimization problem mentioned above in the form of the minimization of the objective function.

If the constraint verification step makes it possible to determine that the objective function evaluated satisfies them, it is then checked whether the objective function 44 is minimal.

- If the objective function is minimum 45, we return 46 the last value , which is the global minimum.

- If not 47, another value of X 48 is chosen then the process is repeated from the step of evaluating the objective function 42, on the basis of this other value of X.

Then the planning method 30 implements a post-processing step 33 comprising the generation of control signals (recharging/discharging power), the estimation of the evolution of the temperature of the battery and of the state of load (SoC) over the entire period of availability of the electric vehicle.

The signals are generated by applying the following equation:

With :

MCp, R, and Rthext thermal constants dependent on battery chemistry

T the battery temperature

Text room temperature

I the battery current

From this power, we deduce the energy absorbed by the battery:

And the state of charge value:

It should be noted that in reality an estimation of SOC and of temperature is also made before evaluating the constraint of power limitation by the battery.
This implicit calculation is not put forward, because it is an intermediate calculation just to have the SOCt+1 and Tt+1, the most important being the verification of the power limitation constraint which reflects the effect of the battery temperature. To conclude, the temperature is estimated twice, instantly after each draw of potential solution and during a final estimation of the temperature which gives the global evolution of the battery temperature for the optimal solution X. In the modeling of the problem we has however assumed that the energy absorbed is equal to the energy requirement, for simplification of calculation.

We have also not taken into account the energy consumed by the accessories because we do not know the exact quantity of this energy and we do not have a global battery model with all the units around it.

However, these simplifications can alternatively be processed in order to refine the model.

However, in this way the process is relatively simple and quick. It suffices to increase the actual energy requirement by a factor of 1.1 to 1.25, in order to compensate for the energy consumed by the accessories.

With the addition of this factor, it is thus possible to reach, or even exceed, the final state of charge desired by the customer.

The power profile of the optimized vehicle is thus obtained from the prices and the available powers received at the input.

According to an alternative implementation of the invention, the planning is carried out according to the energy, by taking the product of the power times the duration for G2V and V2G as the decision variable, which reduces the problem to two variables decision for each time segment. As the need for mobility is an energy need, modeling in energy is possible and gives results comparable to the first embodiment. However, it is then necessary to plan the energy distribution on each segment to have the control power to send to the charger.

Claims (8)

Method for planning (30) an electric recharge and/or an electric discharge of an electric motor vehicle (2) characterized in that it comprises:

• A pre-processing step (31) comprising:
  • A sub-step of acquiring a plurality of modeling data;
  • A subdivision sub-step (312) during which a predetermined availability duration is subdivided (3121) into a sequence of time segments (3122); the subdivision being calculated such that, for at least one chosen modeling datum, said chosen modeling datum is constant in each time segment (3122); the temporal subdivision of said modeling data into a vector whose dimension is equal to the number of subdivided segments forming a vector of modeling data,
• An optimization step (32) during which an optimal set of decision vectors is determined, by calculating an objective function as a function of said modeling data vectors, said decision vectors comprising decision variables c' that is to say adjustable parameters capable of modifying the cost of said electrical recharge and/or discharge, and each being of the same dimension as said modeling data vectors,
• A post-processing step (33) during which charging and/or discharging control signals are generated as a function of said optimal set of decision vectors and of said modeling data.

Planning method (30) according to Claim 1, characterized in that the said at least one chosen modeling datum comprises, for the predetermined duration of availability, the data representative of:

• of the maximum charging power (PG2V_max)
• the maximum possible discharge power (PV2G_max);
• charging energy tariffs (prix_G2V); And
• discharge remuneration (price_V2G).

Planning method (30) according to claim 1 or 2, characterized in that the optimization step (32) is also a function of a plurality of functions of technical and usage constraints, the optimization step comprising a verification sub-step (43) of the compliance of said optimal set of decision vectors with said constraints. Planning method (30) according to any one of Claims 1 to 3, characterized in that the said optimal set of decision vectors is determined (44) as being the set of decision vectors minimizing the said objective function. Planning method (30) according to any one of Claims 1 to 4, characterized in that the said objective function corresponds to the equation:

in which :
J ₁ a recharge energy cost value;
J ₂ an energy discharge cost value;
c1 and c2 are predetermined weighting coefficients;
X said set of decision vectors;
T the size of said decision vector; And
t the index corresponding to a time segment for the corresponding values of said set of decision vectors. Scheduling device (23) of an electric recharge and/or an electric discharge of an electric motor vehicle (2) characterized in that it comprises:

• Pre-processing means (31) capable of:
  • Acquire from a plurality of modeling data;
  • Subdivide (3121) a predetermined uptime into a sequence of time segments (3122); the subdivision being calculated such that, for at least one chosen modeling datum, said chosen modeling datum is constant in each time segment (3122); the temporal subdivision of said modeling data into a vector whose dimension is equal to the number of subdivided segments forming a vector of modeling data,
• Optimization means (32) capable of determining an optimal set of decision vectors, by calculating an objective function as a function of said modeling data vectors, said decision vectors comprising decision variables i.e. i.e. adjustable parameters capable of modifying the cost of said electrical recharge and/or discharge, and each being of the same dimension as said modeling data vectors,
• Post-processing means (33) capable of generating recharging and/or discharging control signals as a function of said optimal set of decision vectors and of said modeling data.

Electric motor vehicle (2) comprising at least one electric storage battery (21) and a planning device (23) according to claim 6. Electrical assembly (1) comprising an electric motor vehicle (2) according to claim 7 and a charging and/or discharging terminal (3).