Method for controlling a collective electrical energy storage system

Abstract

The invention relates to the control of a battery comprising a battery management system (BMS), configured to selectively activate: a discharge of the battery on an electrical network, for frequency support (FCR) thereof; or a discharge of the battery on a device for self-consumption of electrical energy; or a charge of the battery by a device for charging said battery or a charge by the network. The invention is essentially characterized in that the battery management system (BMS) comprises a set of operating parameters, the method comprises steps consisting of: Determining the state of health (SOH) of the battery; Comparing the determined state of health (SOH) with a reference value, and Modifying the value of at least one operating parameter according to the result of the comparison.

Description

Title of the invention: Method for controlling a collective electrical energy storage system

[0001] The present invention relates to the field of electrical energy management.

[0002] Electrical energy is produced by different means. It is transported in particular by a network. An electrical network is a system that must be in balance to function: energy production must be equal to energy consumption. If there is overconsumption, there is a risk of cascading outages; and if there is underproduction then the network must call on a "reserve" and / or load shedding system.

[0003] The present invention is advantageously implemented in the field of automatic power reserve, or frequency stabilization reserve, called FCR (Frequency Containment Reserve in English) or even “primary reserve”, which is a reserve used to adjust the production distributed on a network when consumption or production is disturbed.

[0004] In this area, local electricity producers are asked to declare their FCR capacity to the network manager. If the primary FCR reserve does not allow the nominal frequency value to be restored, then a secondary reserve (aFRR: for automatic frequency restoration reserve) is requested, according to a similar principle.

[0005] The electricity transmission system operator, RTE in France, which is responsible for the public high-voltage electricity transmission system in metropolitan France, sends a signal to all the players in the aFRR secondary reserve. For the FCR primary reserve, the activation signal is directly deduced from the frequency measurement and the calculation of the frequency variation in real time.

[0006] For a day J of supply of FCR, this declaration must be made, for example in France today, at the latest on D-1 (24 hours in advance), in 4-hour slots.

[0007] Beyond the electrical network, electrical energy can be produced locally, for example by renewable energy production devices (REP), in particular photovoltaic (PV) panels. Electrical energy is consumed locally by various consumption devices (homes, industry, etc.).

[0008] In self-consumption (SC), if consumption and production are identical at the same time, then the electricity is produced and consumed locally, otherwise the difference is returned to the network, stored, or is lost.

[0009] In addition to the network, a collective electrical energy storage system, or battery, can be implemented, which makes it possible to supply or store electricity and which is connected to both the power grid, renewable energy production devices and consumption devices. For the purposes of the present invention, the battery is rechargeable. For brevity, it is called a "battery".

[0010] There is therefore a need to optimize battery management, i.e. to control the charging and discharging of such batteries, using a battery management system (BMS).

[0011] The present invention is placed within the framework of participation in the primary FCR reserve of a network, with electricity produced using a renewable energy source (RE) associated with a storage system. By RE, we mean at least any one of the energy sources among: photovoltaic solar, wind, hydroelectric (waterfalls, tides) allowing electricity to be produced.

[0012] The battery management system (BMS) is configured to selectively activate: • a discharge of the battery on the electrical network, for frequency support (FCR) thereof; or • a discharge of the battery on the self-consumption electrical energy device; or • charging the battery by the charging device of said battery or charging by the network.

[0013] The invention makes it possible to calculate and thus regulate the aging of the battery, while optimizing the time slots during which the BMS activates the FCR slots or the self-consumption instructions.

[0014] In this context, the invention relates, according to a first of its objects, to a method for controlling a collective electrical energy storage system, or battery, the battery being connected to:

[0015] an electrical network selectively allowing charging or discharging of said battery;

[0016] a device for charging said battery, separate from the electrical network, and for example using renewable energy;

[0017] a device for discharging said battery, or a device for self-consumption of electrical energy, separate from the electrical network,

[0018] the battery comprising a battery management system (BMS), configured to selectively activate:

[0019] a discharge of the battery on the electrical network, for frequency support (FCR) of the latter; or

[0020] a discharge of the battery on the electrical energy self-consumption device; or

[0021] charging the battery by the charging device of said battery or charging by the network.

[0022] It is essentially characterized in that:

[0023] the battery management system (BMS) includes a set of parameters of functioning,

[0024] the method comprises steps consisting of:

[0025] Determine the state of health (SOH) of the battery;

[0026] Compare the determined state of health (SOH) to a reference value, and

[0027] Modify the value of at least one operating parameter based on the result of the comparison.

[0028] It can be provided that the set of operating parameters comprises at least one of the parameters among:

[0029] - a value (SOCmin) of minimum state of charge of the battery

[0030] - a value (SOCmax) of maximum state of charge of the battery,

[0031] the modification step comprising at least one of the steps among:

[0032] - increasing the value (SOCmin) of the minimum state of charge of the battery, and

[0033] - the decrease in the value (SOCmax) of the maximum state of charge of the battery.

[0034] It can be predicted that:

[0035] - the step of increasing the value (SOCmin) of the minimum state of charge of the battery includes adding an adjustment value to the value (SOCmin) stored in a memory, the adjustment value being a function of at least one of:

[0036] The determined battery state of health (SOH) value, and

[0037] the value of the difference between the determined state of health (SOH) and the reference value;

[0038] - the step of decreasing the value (SOCmax) of maximum state of charge of the battery comprises removing an adjustment value from the value (SOCmax) stored in a memory, the adjustment value being a function of at least one of:

[0039] • The determined battery state of health (SOH) value, and

[0040] • the value of the difference between the determined state of health (SOH) and the reference value.

[0041] A step can be provided consisting of promoting the discharge of the battery on the electrical network, for frequency support (FCR) of it.

[0042] A step may be provided consisting of minimizing the value of an objective function, the objective function (OF) comprising a relationship between a predetermined set of operating parameters and variables, and corresponding to a predetermined optimization of the charging or discharging of the battery; such that the objective function is determined by the following relationship:

[0043] minFO- mm( / l-ct / 2)

[0044] With f [ Tsj^'PfioutRj'*'lit - P'injetf 'Tr fdt - Rpcz / P Battery ^FCR-) ]

[0046]

[0047]

[0048]

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055]

[0056]

[0057]

[0058]

[0059]

[0060]

[0061]

[0062]

[0063]

[0064]

[0065]

[0066] 0^ / 2 = 0^ 1- Pcons^i*^ \ \ ^ÇQilSOi / / with : PsoutR, i~ ~ Ppri "*■ Pinjcti + P ch / " Poéch / + Pconso, i ' a predetermined coefficient whose value can be dynamic, and which represents, depending on its value, one allocation or another, for example in favor of self-consumption in order to encourage it; Or : T(s,i) is a parameter relating to the purchase (withdrawal) tariff on the network in time slot i; T_(r,i) is a parameter relating to the resale rate (injection) of surplus to the network in time slot i; R_(FCR,i) is a parameter relating to the FCR income on time slot i for 1 MW sold; P_(Battery) is a parameter relating to the maximum power sold for the FCR; P_(conso,i) is a parameter relating to consumption (power) over time slot i; P_(soutR,i) is a variable relating to the power drawn from the network over time slot i; P_(injct,i) is a variable relating to the power injected into the network in time slot i; FCR_i is a binary variable relating to the selection or not of time slot i for FCR; The method comprising a step of modifying the value of the coefficient al as a function of at least one of: The determined battery State of Health (SOH) value, and the value of the gap between the determined state of health (SOH) and the reference value. A step may be provided to determine the aging of the battery, including one of: - calendar aging; - cyclical aging; - the sum of cyclical aging and calendar aging. A step can be provided consisting of limiting the exchanges of electrical energy between the battery and the charging device of said battery or the self-consumption device during a self-consumption slot activated by the battery management system (BMS).

[0067] It can be expected that the total quantity of energy exchanged Eexchange during the self-consumption slot is greater than A*Cbattery and less than B*Cbattery, with:

[0068] - Battery capacity of the battery,

[0069] - A and B are either two constants whose value is predetermined, or two variables whose value is a function of at least one of:

[0070] The determined battery state of health (SOH) value, and

[0071] the value of the difference between the determined state of health (SOH) and the reference value.

[0072] According to another of its objects, the invention relates to a system for managing the battery (BMS), configured to implement the method according to the invention.

[0073] According to another of its objects, the invention relates to an electric battery comprising a battery management system (BMS) according to claim 9.

[0074] Other characteristics and advantages of the present invention will appear more clearly on reading the following description given by way of illustrative and non-limiting example and made with reference to the appended figures.

[0075] Figures are not to scale. Some details may have been omitted and others enlarged, to facilitate understanding.

[0076] DESCRIPTION OF THE DRAWINGS

[0077] [Fig. 1] illustrates the principle of electrical connections of a battery within the meaning of the present invention. DETAILED DESCRIPTION

[0078] According to the invention, a battery is connected to: • an electrical network selectively allowing charging or discharging of said battery; • a device for charging said battery, separate from the electricity network, and for example using renewable energy; • a device for discharging said battery, or a device for self-consumption of electrical energy, separate from the electrical network.

[0079] The battery comprises a battery management system (BMS), configured to selectively activate: • a discharge of the battery on the electrical network, for frequency support (FCR) of the latter; or • a discharge of the battery on the self-consumption electrical energy device; or • charging the battery by the charging device of said battery or charging by the network.

[0080] The invention makes it possible to optimize which predetermined time ranges should be selected to enable FCR support by an electricity producer. produced in renewable energy. Typically in France, the predetermined time ranges are 4 consecutive and predetermined hours: 0h-4h; 4h-8h; 8h-12h; 12h-16h; 16h-20h and 20h-24h for the primary FCR reserve. For all other markets, these time ranges vary, for example, either 1h or 30 minutes.

[0081] The predetermined time ranges are either stored in a memory or calculated as described later.

[0082] Each electricity producer must declare the day before the predetermined time slot(s) over which it is positioned to support the FCR.

[0083] It is therefore essential to control the destination of the electricity produced in renewable energy, therefore stored in a battery: either in self-consumption or in FCR / aFRR.

[0084] The battery management system (BMS for Battery Management System; or EMS for Energy Management System) allows the battery charge / discharge instructions to be optimized in order to improve its aging.

[0085] The battery management system (BMS) includes a set of operating parameters, described later.

[0086] According to the invention, steps are provided consisting of determining the state of health (SOH) of the battery, which is an indicator of the aging of the battery; then comparing the determined state of health (SOH) with a reference value, and modifying the value of at least one operating parameter according to the result of the comparison.

[0087] Determining the state of health (SOH) of the battery.

[0088] The state of health of the battery can be determined by any known means, for example from document CA3154716. Many other algorithms are accessible to those skilled in the art for determining the SOH of the battery.

[0089] Once the SOH of the battery has been determined, it can be compared to a reference value.

[0090] The reference value comes from a battery aging model which includes the evolution of the SOH as a function of time, i.e. the aging of the battery.

[0091] Battery aging may depend on:

[0092] - Calendar aging;

[0093] - cyclic aging, i.e. the number of charges / discharges of the battery; or

[0094] - of the sum of cyclical aging and calendar aging.

[0095] The value of at least one operating parameter of the battery management system based on the comparison result.

[0096] For example, it may be provided that a parameter is the value (SOCmin) of the minimum state of charge of the battery or the value (SOCmax) of the maximum state of charge of the battery.

[0097] Depending on the result of the comparison, it is then possible to predict the increase in the value (SOCmin) of the minimum state of charge of the battery or the decrease in the value (SOCmax) of the maximum state of charge of the battery. According to Table 1 below, the SOCmin value can for example go from 0.05 to 0.1 and the SOCmax value can for example go from 0.95 to 0.9.

[0098] The modification of the value of the parameter can be by an adjustment value, constant and predetermined, for example 0.05 according to the example above.

[0099] It can also be provided that the adjustment value is a function of the value of the state of health (SOH) of the determined battery. In this case, the adjustment value is variable and a function of the value of the SOH. For example, the lower the SOH, the greater the adjustment value.

[0100] Alternatively or in combination, it can also be provided that the adjustment value is a function of the value of the difference between the determined state of health (SOH) and the reference value. In this case, the adjustment value is variable and a function of the relative value of the SOH. For example, the greater the difference between the SOH and the reference value, the greater the adjustment value. BMS operating parameters.

[0101] The system for exchanging electrical energy between the network, the battery, the self-consumption device (therefore battery discharge) and the self-production device (ENR, therefore battery charging) can be modeled according to a system of mathematical equations.

[0102] This system of equations must be solved, in this case by a solver. For example, the solver is the GLPK solver for GNU Linear Programming Kit. It comprises a set of quantities which must satisfy a set of predefined constraints, for example the constraint that the power balance is zero (see below).

[0103] The quantities of the system of equations are a set of variables and a set of parameters, which are some of the operating parameters of the BMS.

[0104] The parameters are values ​​considered constant over a given time range, in this case 30 minutes.

[0105] Table 1 below gives an example of parameters. Table 1 Parameter Description of the PPv i parameter Solar energy production (power) P * conso. i Consumption (power) T ■ * SJ Purchase tariff (withdrawal) from the network on time slot i Rfcrî FCR revenue on the time slot for 1 MW sold n Storage system efficiency, for example value 0.95 Tri Surplus resale tariff (injection) to the network Battery capacity (kWh) P « * Battery Maximum power sold for FCR SOC min Minimum admissible SOC for the battery, for example value 0.05 SOCmax Maximum admissible SOC for the battery, for example value 0.95 eCh FCRj Maximum energy to be stored in the Battery in FCR, for example value 0.7 * CBatt e Déch FCRj Minimum energy to be left in the battery in FCR, for example value 0.3 * CBatt Pcompteur Maximum power of the customer meter

[0106] The value of each parameter is known. Preferably, it is constant over each time step. It can be measured or estimated, for example by a history, in particular from the previous year, possibly coupled with forecasts, in particular meteorological forecasts.

[0107] In addition to the parameters in Table 1, a set of variables is provided.

[0108] Table 2 below gives an example of variables, as well as the domain of each variable.

[0109] In this example there are three binary variables and 5 normalized variables. Table 2 Variable Description of the variable Domain PgoutR i Power drawn from the network in time slot i [0,1] P Chi Power loaded into the battery in time slot i [0,1] ^Déchi Power discharged from the battery in time slot i [0,1] SOC, J State of charge of the battery in time slot i [0,1] Pin jcti Power injected into the network in time slot i [0,1] PCR, If the slot is chosen for FCR on the time slot i 0 or 1 ACchargei If the slot is chosen for charging in Self-consumption on the time slot i 0 or 1 discharged} If the slot is chosen for discharging in Self-consumption on the time slot i 0 or 1

[0110] The relationships between variables and parameters are a set of constraints which are modeled by a system of equations, one embodiment of which is illustrated in Table 3 below. Table 3 No. Descriptive Constraint 1 Balance of wells without ^souiR, i + ^Pv} " in jet} " ^chi + Déch}~^conso, i — 0 2 Only of the load Load}* M + PChj 0 2bis Only of the discharge AC Discharge}? AI + PDech — 3 AC Charge Limit If i multiple of 4 ^0 If not CBattery^ { (1 -SoC^ - ( l-SoCw^.) ) -PChfdt*i] > 0 4 AC Charge Limit If i multiple of 4 CBaiterif( +. - O^FCR (SoC) - Somin ) > Pijprii fictif CBatterie^ f 5 Revente Ener gie des ENR (PV) uniqueme nt SiPPv}-P .>0 M consoi Piniecti~ Ppvi~^~ P . — 0 mjeuf rij consoJ If not Pinjectj = 0 6 Evolution of the S OC If i = first crest of the horizon: SoCi = SoCinitjlll If not: CBattery* SOC( - ( Cg^SOC^ + ' + PCh^dt^ + PFCR» ^dt ) - 0 7 Consumption in priority energy of sPV If PpVj " Pconsoi — 0 PsoutRi ' PDechi ~ 0 If not Psouti ~ Pconsoi + PpVi — 0 8 Sale of FCR over 4h If i+1 not divisible by 4 FCR, -FCRm = =0 9 Return to the initial Soc For the last half hour: CBtUri*{socinitiai-s} , +Pck ^CB„ai*(SOCMtial+e.) 10 Choice of the strategy Chargej "l" ^Cf)éch / 4- FCR- — 1

[0111] In this non-limiting example, the constraints can be described as follows.

[0112] Constraint 1: Balance of active powers on the network.

[0113] Constraints 2 and 2a: sets the value of the binary variable of the charge (discharge) according to the value of the charge (discharge) power. These constraints make it possible to constrain the set powers (charge and discharge) to be zero if the binary variable is zero.

[0114] These equations use the penalty method or "big M method", known for example from the document accessible at the address http: / / staff.univ-batna2.dz / sites / default / files / kalla-salim / files / chap4-bigm-distance-pl.pdf.

[0115] The value M is a value greater than the orders of magnitude of the maximum charging power (Pch j), and in this case fixed at M = 10A6, this makes it possible to have a constraint on the value of AC_charge [0 or 1] according to the value of PCh ,i.

[0116] Constraint 3: This constraint limits the battery charge. If i (hour) is a multiple of 4, then we are at the beginning of a 4-hour slot, we must check if (1-SOC) which corresponds to the space available in the battery, is greater than 30%, to make the FCR. If FCR(i) is zero there is no constraint on the maximum SOC of charge. If not we apply the classic exchange constraint: the SOC must not exceed 1.

[0117] Constraint 4: This constraint represents the discharge limit, according to the same principle as constraint 3. This constraint makes it possible to limit the SOC to zero. In the case of the start of a 4-hour slot, we check with the FCR requirements in addition (min SOC of 30%), we also add a SOCmin to limit deep discharges of the battery in self-consumption.

[0118] Constraint 5: Energy can be injected into the network only if production is greater than consumption. The quantity injected into the network must be only the surplus.

[0119] Constraint 6: The variation of the SOC is the sum of the energy exchanges which take place on the battery. The charged power is divided by the efficiency, we draw more more energy than what is actually stored, converter loss at the load. The discharged power is multiplied by the efficiency, these are the converter losses at discharge. With Nu the efficiency of the conversion system, including for example inverters, DC / DC converters, etc.

[0120] Constraint 7: allows the use of PV / RE energy as a priority when PV production covers consumption, no discharge or withdrawal during these time steps. If production does not cover consumption, we cannot withdraw more than what is necessary for direct consumption. This limits purchase / resale on the network between peak and off-peak hours.

[0121] Constraint 8: The binary variable FCR of i must be the same over the 4-hour slots.

[0122] Constraint 9: To limit the discharge, it is desirable to limit the value of SOC on the last slot, otherwise we can observe a strong discharge. To this end, we can limit the energy exchanges in this case by adding and subtracting a limit e from the initial SOC. This limit e can be equal to 0. In this case it is a value fixed at 0.02, which allows us to relax constraint 9.

[0123] Constraint 10: This constraint limits the number of possible actions in the same slot, the battery can either be charging for self-consumption, or discharging for self-consumption, or in service for FCR.

[0124] For each time step i, in this case 30 minutes, the solver solves all the equations of the system and thus determines the values ​​of all the variables that meet all the constraints, in this case P_(soutR,i); P_(Ch,i); P_(Dch,i); SOC_(,i); P_(injct,i); FCR_i; AC_(charge,i) and AC_(decharge,i). We can therefore obtain a plurality of possible values ​​per variable.

[0125] According to the invention, the problem to be solved is of the MILP (Mixed Interger Linear Programming) type. Optimization makes it possible to determine the ranges, or slots, most conducive to self-consumption. This optimization makes it possible to choose the best slots, i.e. those which optimize the balance between FCR and self-consumption while promoting self-production.

[0126] For each time step i, each variable has only one possible value. These values ​​are stored in a memory.

[0127] Thus, over a predetermined time range, we obtain a set of variable values. In this case, for a time step i of 30 minutes over a range of 4 hours, we obtain a set of 8 values ​​per variable, or 48 values ​​per variable over a range of 24 hours. Objective function

[0128] To optimize system management, it is appropriate to select, for each variable, a unique value from the set of calculated values.

[0129] To select a unique value per variable from the set of values ​​per variable, we plan to minimize an objective function.

[0130] The objective function corresponds to the desired optimization of the collective electrical energy storage system.

[0131] The objective function therefore depends on what we wish to optimize within the system.

[0132] Typically, the objective function aims to find the set of variables that allows to maximize or minimize a function defined as a linear combination (MILP type problem) of parameters and variables, for example to maximize the power sold for the FCR, to maximize solar production, etc. and typically, to optimize at least one of: the power withdrawn from the network, the power charged into the battery, the power discharged from the battery, the state of charge of the battery, and the power injected into the network.

[0133] The objective function can be defined as a kind of constraint, but which is resolved by the solver after the resolution of the system of mathematical equations.

[0134] The objective function is a mathematical function comprising a relationship between some of the parameters and / or some of the variables described above.

[0135] In this case, in a non-limiting example, the objective function FO is determined as follows:

[0136] minFO- 1(1111( / 1^^2)

[0137] With JI “ [ Tsf PsoutRj? ~ Pinjctj " P-FCRj * P Battery *'* FCÿ ] a / f2^a^EnergyACN 1- .......— Pcwi*dt

[0140] with: * PsoutR, i “PPVJ + Pinjctj + Pchj - Pl)échj 3- Pconso, i ' • has a predetermined coefficient whose value can be dynamic, and which represents, depending on its value, one allocation or another, for example in favor of self-consumption in order to encourage it.

[0141] In this case, the objective function is calculated over N predetermined time steps of 30 minutes (N = 48). All data, variables, instructions and parameters are over a 30-minute step.

[0142] The optimization determines the optimal slots, in this case 4 hours, over a determined time horizon, in this case 48 hours. That is to say that the objective function is calculated over a time horizon of 48 hours, and not over the 30-minute time step because this is too short for fine management of the battery charge / discharge.

[0143] We choose a two-day optimization because it is difficult to store energy over long periods. Indeed, the FCR requires a SOC of 50% because the frequency variations are not predictable. Energy storage can be done mainly in the same self-consumption slot.

[0144] Thanks to the present invention it is thus possible for example to control the management of energy in a battery, so that it can serve both as an energy source for self-consumption and as an energy source for support of the network frequency (FCR).

[0145] According to the invention, this energy management can be hierarchical: if a 4-hour slot is determined to operate the battery in support of the FCR, then this slot is blocked and the battery is only operated for this purpose. If a 4-hour slot is not determined to operate the battery in support of the FCR, then the battery can be controlled for local charging or discharging (self-consumption), depending on its state of charge, the energy demand and its recharging capacity via renewable energy sources.

[0146] For example, the battery can be in FCR support over a time slot from 8 a.m. to 12 p.m., then in the following slot: charging over a time slot from 12 p.m. to 2 p.m., then discharging from 2 p.m. to 3 p.m., then charging from 3 p.m. to 4 p.m., then in the following slot again in FCR support from 4 p.m. to 8 p.m., etc.

[0147] According to the invention, it can be provided that the step consisting of modifying the value of at least one operating parameter of the BMS comprises modifying the value of the coefficient al of the objective function.

[0148] It can be provided that the value of the coefficient al is determined as a function of the value of the state of health (SOH) of the battery determined.

[0149] It can also be provided that the value of the coefficient al is determined as a function of the value of the difference between the determined state of health (SOH) and the reference value. Charging or discharging instructions outside FCR.

[0150] Once the slots have been selected for FCR support, battery management can be provided for the other slots, called self-consumption (AC) slots, typically in order to maximize self-production and self-consumption of electrical energy.

[0151] This is a second optimization loop, downstream of the previous optimization, and which makes it possible to calculate the charge and discharge instructions as close as possible to real time.

[0152] Indeed, in this case, it is possible to charge the battery using the charging device (ENR) or to discharge the battery in the discharging device, depending on the power available by the charging device (ENR) and the power requested by the discharging device.

[0153] In this case, the battery management system selectively emits: - either a battery charging instruction by the charging device; - either a battery discharge instruction in the discharge device; - either an inhibition instruction, which amounts to not issuing any instruction charging or discharging the battery.

[0154] Advantageously, the 4-hour slots can be subdivided into a set of sub-slots, each sub-slot being, for example, of a time range between 1 a.m. and 4 a.m., and each sub-slot can be assigned a respective instruction.

[0155] Thus, a sub-slot can be of the same duration as a (FCR) slot, i.e. 4 hours. However, as the climatic conditions can vary over such a 4-hour slot, it may be clever to subdivide the 4-hour slots into a set of sub-slots and thus optimize the management of the energy produced by photovoltaic panels for example and consumed for self-consumption.

[0156] In this sense, a sub-slot is an optimization period.

[0157] In this case, each sub-slot is 1 hour each, the instructions being issued according to values ​​determined according to a 30-minute step preceding the sub-slot. Note that as two successive 4-hour slots can be self-consumption slots, we can have a succession of sub-slots which spread over a time span of 8 hours.

[0158] According to the present invention, each instruction allows the charging or discharging of the battery depending on: - electrical production by charging device, - electrical consumption by discharging device, and - the state of charge (SOC) of the battery,

[0159] according to an optimization described below, and similar to the objective function seen previously for the selection of slots in support of FCR.

[0160] In this case, the electrical production is determined by estimation or measurement, the electrical consumption is determined by estimation or measurement and the SOC is measured.

[0161] Typically, for a given sub-slot: - if the determined electrical production is greater than or equal to the determined electrical consumption, then the battery management system (BMS) can issue a battery discharge instruction in the discharge device and / or a battery charge instruction, for example a surplus charge instruction; - if the determined electrical production is lower than the determined electrical consumption, then the battery management system (BMS) issues a setpoint which depends on the value of the battery SOC: if the battery SOC value is greater than a predetermined threshold value, then the battery management system (BMS) can issue a battery discharge instruction; if the battery SOC value is lower than the said threshold value predetermined, then the battery management system (BMS) can issue an inhibition instruction.

[0162] As the sub-slots have a low hourly amplitude (Ih), it is estimated that the values ​​are constant and reliable in each sub-slot.

[0163] Similar to the determination of time slots to enable FCR support, the aim here is to determine an instruction for each sub-slot, using a second objective function. In this case, the aim is to maximize the second objective function in the following manner: [°1641 mavf lf a- ( 1 ( l mdX ( 1-( £ JJ Econso + ( 1- Epv JJ Epv

[0165] With

[0166] Esout = Energy withdrawn (consumed) from the network over the optimization period (kWh);

[0167] Econso = Total energy consumed over the optimization period (kWh);

[0168] EPV = Total energy produced over the optimization period (kWh) by the charging device;

[0169] Einjc = Total energy injected into the network (resold) over the optimization period (kWh); and

[0170] E^^fdt

[0171] with N the number of time steps of the given self-consumption slot and the index x representing the type of energy: sout, conso, PV or injet respectively for withdrawn, consumed, photovoltaic and injected.

[0172] For example the time step is 30 minutes.

[0173] It is possible to plan to simulate the entire operation of the battery, and to calculate the energy exchanges requested by the FCR and the aFRR based on frequency signal histories (FCR) or the N signal given by RTE (aFRR).

[0174] Advantageously, the variables are the same as those in table 2 above.

[0175] On the other hand, the parameters of the following table 4 are provided: Table 4 Parameter Description of the parameter Ppv, i Solar energy production P * conso. i Consumption T • ■* SX Purchase tariff (withdrawal) on the network on the slot i RpCRi FCR revenue on the niche for 1 MW sold n Storage system efficiency c Battery Battery capacity in kWh P 1 Battery Max power sold for FCR 8OCmin Minimum admissible SOC for the battery SOCmax Maximum admissible SOC for the battery upper terminal Upper Soc terminal for constraint 9 Lower SoCforne fa Lower Soc terminal for constraint 9 Pmeter Max power of the electric meter

[0176] And we predict the constraints of the following table 5: Tableau 5 N° Descriptif Contrainte 1 Bilan des puissa nce P sont i PPvi ” Pinjcti " ^chi 3“ PDéchj~Pconso, i 2 Uniquement de 1 a charge Charge} ' 3" Pchi — 2bis Uniquement de 1 a décharge AC Déchargeî' AI + P Dech — 3 Limite de Charg e en C Batterie^ ( 1 " SoCi ) ' (1 " J ' Pch. 0 4 Limite de Décha rge en AC fïacîr [ Cb^^SoQ-SoC^)- >0 5 Revente Energi e des PV unique ment Si PPvi-P . >0 - consoj Piniecti- PPvi+ P . <0 Si non Pinject} = 0 6 Evolution du S OC Si i = prime créneau de l'horizon : SoCt = SoCinitùlj Si non : Cb^SOC^Cj^SOC^ „ +Pch^dt*n [-0 7 Conso en priori té énergie des PV Si PP\j "Pconsoj — 0 PsoutRj " PDechj “ Si non Psmitj-Pconsoj Ppvj — 0 8 Choix de la stratégie AC Charge! + ^-'Déchi — 1 9 Limite d'échang e d'énergie P exchange U “b Pch,il^'S Si SoC^ > SoC^^ supérieur VJ Battrie — l^exclumge — v Baitrie Si SoC^ SûCborne lower Batterie — P exchange ^02 '^Batterie Si non - D <f     <0 v batterie ~ exchange —        batterie

[0177] In this non-limiting example, the constraints can be described as follows.

[0178] Constraint 1: Balance of active power on the network.

[0179] Constraints 2 and 2a: sets the value of the binary variable of the charge (discharge) according to the value of the charge (discharge) power, allows the setpoint powers (charge and discharge) to be forced to be zero if the binary variable is zero.

[0180] Constraint 3: This constraint limits the battery charge. If i (hour) is a multiple of 4, we are at the beginning of a 4-hour slot. We must check whether (1-SOC), which corresponds to the space available in the battery, is greater than 30%, to perform the FCR. If FCR(i) is zero, there is no constraint on the maximum SOC charge. If not, we apply the classic exchange constraint: the SOC must not exceed 1.

[0181] Constraint 4: It represents the discharge limit, same principle as constraint 3. This constraint makes it possible to limit the soc to zero. In the case of the start of a 4-hour slot, we check with the FCR requirements in addition (min SOC of 30%), we also add a SOCmin to limit the deep exchange of the battery in self-consumption.

[0182] Constraint 5: We can inject into the network only if production is greater than consumption. The quantity injected into the network must only be the surplus.

[0183] Constraint 6: The variation of the SOC is the sum of the exchanges which take place on the battery. The charged power is divided by the efficiency, we draw more energy than what is actually stored, loss of the converter at the charge. The discharged power is multiplied by the efficiency, it is the losses of the converter at the discharge.

[0184] Constraint 7: Use of renewable energy (PV for photovoltaic) as a priority when PV production covers consumption, no discharge or withdrawal on these time steps. If production does not cover consumption, we cannot withdraw more than is necessary for direct consumption. This limits purchasing / resale on the network between peak and off-peak hours.

[0185] Constraint 8: It is necessary to limit the value of the SOC on the last slot otherwise we see

[0186]

[0187]

[0188]

[0189]

[0190]

[0191]

[0192]

[0193]

[0194]

[0195] a strong discharge to maximize gains. In reality there is no end of data series, so this constraint is only necessary for simulations. A power ramp limit can be added to limit the end of simulation effects. Constraint 9: In this constraint we limit the total amount of energy exchanged Eexchange during the self-consumption slot. The total amount of energy exchanged during a self-consumption slot is the sum of the charged energy and the discharged energy or P*dt. Depending on the state of the SOC at the start of the self-consumption slot we define different upper and lower bounds. For example, if the battery has a SOC lower than the SOC defined as the lower bound, the total energy exchange during the slot must be positive but must not exceed 20% of the battery storage capacity. More generally, constraint 9 can be written as follows: If SoCj SoC upper bound - A1 < F . < R1 ^Battrie — exchange — " 1 Battrie With Al and B1 as constants. If SoCj S()Clower bound Battery " ^exchunge 2s ^Battery Otherwise - CBattery < Eexchange < B3*CBattery Constraint 9 also includes BMS operating parameters, including the following quantities: Upper SoCborne, Lower SoCborne, Al, A2, A3, and B1, B2, B3.

[0196] The step of modifying the value of at least one operating parameter may for example comprise the modification of at least any one of the above quantities.

[0197] Preferably, it is planned to charge the battery from the network only during a self-consumption slot, but the battery can be charged and discharged on the network when it is in FCR service.< / f>

Claims

Claims

1. Method for controlling a collective electrical energy storage system, or battery, the battery being connected to: - an electrical network selectively allowing a charge or a discharge of said battery; - a device for charging said battery, separate from the electrical network, and for example by renewable energy; - a device for discharging said battery, or device for self-consumption of electrical energy, separate from the electrical network, the battery comprising a battery management system (BMS), configured to selectively activate: - a discharge of the battery on the electrical network, for frequency support (FCR) thereof; or - a discharge of the battery on the device for self-consumption of electrical energy;or - charging the battery by the charging device of said battery or charging by the network, characterized in that the battery management system (BMS) comprises a set of operating parameters, the method comprises steps consisting of: - Determining the state of health (SOH) of the battery; - Comparing the determined state of health (SOH) with a reference value, and - Modifying the value of at least one operating parameter according to the result of the comparison.;

2. Method according to claim 1, in which the set of operating parameters comprises at least one of the parameters among: - a value (SOCmin) of minimum state of charge of the battery - a value (SOCmax) of maximum state of charge of the battery, the modification step comprising at least one of the steps among: - increasing the value (SOCmin) of minimum state of charge of the battery, and - the decrease in the value (SOCmax) of the maximum state of charge of the battery.

3. Method according to claim 2, wherein: - the step of increasing the value (SOCmin) of the minimum state of charge of the battery comprises adding an adjustment value to the value (SOCmin) recorded in a memory, the adjustment value being a function of at least one of: • The value of the state of health (SOH) of the battery determined, and • the value of the difference between the state of health (SOH) determined and the reference value; - the step of decreasing the value (SOCmax) of the maximum state of charge of the battery comprises removing an adjustment value from the value (SOCmax) recorded in a memory, the adjustment value being a function of at least one of: • The value of the state of health (SOH) of the battery determined, and • the value of the difference between the state of health (SOH) determined and the reference value.

4. A method according to any preceding claim, comprising a step of minimizing the value of an objective function, the objective function (FO) comprising a relationship between a predetermined set of operating parameters and variables, and corresponding to a predetermined optimization of the charging or discharging of the battery; such that the objective function is determined by the following relationship: minFO- min( / la} / 2) With f J = Ts*PsautRi*dt - PinjctfTr fdt - RFCRfP Battery *FCRj) ] L.- ?souri * dt 1 1 y-, N 1- v Peon^ / dt \ \ “t ■ ® / / with: * ^soutR, i “ Ppvi + in jeti + Pc}ti - Poéchi + ^conso, i ' - a, a predetermined coefficient whose value may be dynamic, and which represents, depending on its value, an allocation or a other, for example in favor of self-consumption in order to encourage it; where: T(s,i) is a parameter relating to the purchase tariff (withdrawal) on the network on time slot i; T_(r,i) is a parameter relating to the resale tariff (injection) of surplus to the network on time slot i; R_(FCR,i) is a parameter relating to the FCR income on time slot i for 1 MW sold; P_(Battery) is a parameter relating to the maximum power sold for the FCR; P_(conso,i) is a parameter relating to the consumption (power) on time slot i; P_(soutR,i) is a variable relating to the power withdrawn from the network on time slot i; P_(injct,i) is a variable relating to the power injected into the network on time slot i; FCR_i is a binary variable relating to the selection or not of time slot i for FCR;The method comprising a step of modifying the value of the coefficient a, as a function of at least one of: - The value of the state of health (SOH) of the battery determined, and - the value of the difference between the state of health (SOH) determined and the reference value.;

5. A method according to any preceding claim, comprising a step of determining the aging of the battery, comprising one of: - calendar aging; - cyclic aging; - the sum of cyclic aging and calendar aging.

6. Method according to any one of the preceding claims, in which the total quantity of energy exchanged Eexchange during the self-consumption slot is greater than A*Cbattery and less than B*C battery, aVCC . - Catterie the capacity of the battery,

8. - A and B are either two constants whose value is predetermined, or two variables whose value is a function of at least one of: - The determined battery state of health (SOH) value, and - the value of the difference between the determined state of health (SOH) and the reference value.

7. A battery management system (BMS), configured to implement the method according to any one of the preceding claims. An electric battery comprising a battery management system (BMS) according to claim 7.

Images