Method for reconstructing instantaneous values ​​of electrical quantities relating to an electrical energy storage system

Abstract

Method for reconstructing instantaneous values ​​of electrical quantities relating to an electrical energy storage system, the method comprising, at an instant following a successful recovery of the state of charge of said system, a review of different possible decompositions into numerical values ​​of voltage and current of a current value, known for the purposes of reconstruction, of power developed by the storage system, to determine (E1), for said possible decompositions, the resulting state of charge values ​​for the storage system taking into account a previous state of charge, and a comparison (E2) between the vectors each consisting of one of said determined state of charge values ​​and the associated voltage and current values, and vectors known from a prior characterization (E0) of the energy storage system and each consisting, for an accessible physical state of the storage system,of a state-of-charge value associated with voltage and current values, said comparison being carried out to determine the most probable decomposition among said possible decompositions. Abbreviated figure: 15,

Description

Title of the invention: Method for reconstructing instantaneous values ​​of electrical quantities relating to an electrical energy storage system. Technical context

[0001] The invention falls within the field of battery energy storage systems - in English Battery energy storage System, acronym BESS.

[0002] These storage systems have been developing since lithium batteries became available in large numbers and with significant capacities, and are of interest for the effort to decarbonize energy production, since they allow the storage of electrical energy and thus the decoupling of the moment of production and the moment of consumption, such a decoupling being useful for supplying consumers with intermittent production systems.

[0003] These battery energy storage systems can be stationary when they are used to supply electricity to essentially immobile, land-based equipment, such as industrial equipment or homes. They are then connected to a territorial land-based distribution network, or at least to a local network, for example, a domestic network. Conversely, these battery energy storage systems can be mobile when they are installed in a vehicle, in which case they primarily power the traction motor, the vehicle being a car or another type of vehicle. Thus, the storage capacity can range from a few kWh to several GWh.

[0004] The storage system is a dipole and provides a direct current (DC), which is, if needed and is often the case, converted into alternating current (AC) by an inverter, also called a power converter.

[0005] Storage systems interface in a terrestrial or onboard network with other sources of electrical power, such as production equipment like any type of controlled power plant or an intermittent source like a photovoltaic or wind power unit, or an electrical machine operating in generator mode. Storage systems also interface with electrical power sinks, such as consumers in homes or industry, or any type of electrical machine operating in motor mode. Depending on the circumstances, an element can be either a consumer or a producer. The storage system can, depending on the circumstances, supply electrical power, and thus release stored energy, or take advantage of excess available power to recharge itself.

[0006] The interfacing is achieved via transformers or converters that transmit the conditioned electrical power (current type, number of phases, voltage) between the partners, with an efficiency that is generally well known and stable over time. On the side of the partners identified as primarily consumers, the interface of the transformer or converter is referred to as the delivery point.

[0007] The battery can be placed in ambient air, or benefit from air conditioning in an enclosed space, or possibly from a cooling and / or heating system without an enclosed space. Charging and discharging only cause heating if the current is high, but external conditions cannot themselves lead to low or high temperatures, whereas the cells operate optimally within a given temperature range and must be protected against excessive temperatures.

[0008] Temperature management is therefore an issue, particularly in the field of batteries for vehicles where the circulating currents are significant in relation to the capacity of the batteries used, but also for stationary batteries used in less temperate conditions and which it has been decided not to equip with powerful air conditioning, in particular to avoid having to oversize them.

[0009] The energy storage system must be monitored over time to identify its evolution and to best control its operation.

[0010] To achieve this, the battery or storage system traditionally has a battery management system (BMS), which is an electronic system that controls, charges, and discharges the battery by monitoring the voltages, temperatures, and states of charge of the individual cells. A BMS protects the battery by preventing it from operating outside its safe operating range. It also balances the cells and, moreover, communicates overall battery data to any external supervisor that may require it. This overall data includes, in particular, the instantaneous voltage across the battery terminals, its temperature (which may, for example, be a temperature taken at a representative point on the battery itself or its immediate environment, or an extreme or average temperature among several temperatures measured within the battery itself), and a current (which may be an outgoing current, an incoming current, or zero current).

[0011] Battery management systems are often remote in terrestrial storage systems. They can be implemented in a cloud computing system, using remote computer servers hosted in data centers connected to the Internet to store, manage, and process data. They can be included in an energy management system (EMS) that performs calculations of electrical quantities based on known quantities and the grid topology, and ensures proper grid operation.

[0012] Communication to the management system of time profiles of battery voltage and current is useful for making prognoses of health states in order to facilitate predictive maintenance of BESS battery energy storage systems.

[0013] In certain circumstances, some instantaneous data measured in the storage system fail to reach the management system. The causes for such a situation can be varied: a sensor may have malfunctioned temporarily, a wireless connection may have been momentarily interrupted, or a computer error related to software or a hardware component may have occurred. The cause is generally temporary, either because it is resolved by automatic processes implemented by software, or because human intervention is possible quickly. Sometimes the cause may persist longer, particularly in the case of a storage system that is difficult to access, which can occur in mountainous environments or on an island. In such cases, it becomes necessary to make assumptions about the missing data in order to continue operations without interruption, with good performance and, of course, security.

[0014] Handling missing data for data-driven methods is generally a challenge. A distinction is made between data imputation and data reconstruction. Data imputation is the addition of realistic data to replace missing data. Data reconstruction involves reconstructing a complete series, which has been, for example, augmented by imputation, to remove noise and manage outliers.

[0015] Methods for managing missing data can thus be classified into two categories: statistical methods and machine learning methods. Statistical methods are simple and efficient but have inaccuracies. Machine learning methods can be more precise but are complex to implement and require significant computing power, hence their offline use.

[0016] The article M. Xiang, et al. (2020) State-of-Health Prognosis for Lithium-Ion Batteries Considering the Limitations in Measurements via Maximal Information Entropy and Collective Sparse Variational Gaussian Process', IEEE Access, 8, pp. 188199-188217 presents a statistical method for data imputation for health-of-health (SOH) prognosis and management. It uses linear statistical data interpolation and a search for maximum information entropy.

[0017] Document DE102021203729 presents a learning measure for predicting a state of health (SOH) of an energy storage device.

[0018] In both cases, either the method is time-consuming to implement, or it provides results that are not fully satisfactory.

[0019] As explained, it sometimes happens that data relating to the storage system is missing or incorrect in the storage system monitoring files. The method proposed below overcomes this problem by using the fact that the power at the point of delivery is known via an uninterrupted channel, which can therefore be used to fill the gap in measured data.

[0020] The invention therefore consists of a method using other accessible data of the system and thus less data-intensive than, for example, learning methods and also more precise. Features of the invention and advantages

[0021] For this purpose, a method for reconstructing instantaneous values ​​of electrical quantities relating to an electrical energy storage system is proposed.

[0022] The method includes, at a time following a successful recovery of the state of charge of said system, a review of different possible decompositions into numerical values ​​of voltage and current of a current value, known for the purposes of reconstruction, of power developed by the storage system.

[0023] This review is carried out to determine, for the said possible decompositions and by integrating the intensity over time, the resulting state of charge values ​​for the storage system taking into account a previous state of charge.

[0024] The process then includes a comparison between

[0025] - the vectors, each consisting of one of the said determined charge state values and associated voltage and current values,

[0026] - and known vectors from a prior characterization of the storage system of energy and each consisting, for an accessible physical state of the storage system, of a state of charge value and associated voltage and current values.

[0027] The said comparison is carried out to determine the most probable decomposition among the said possible decompositions.

[0028] The invention is advantageous because the method used is very precise, while being low in computational consumption, which is remarkable in view of the state of the art where only methods very computationally intensive were available, or methods that were numerically unreliable.

[0029] According to optional and advantageous features:

[0030] - said storage system can be interfaced with transmission equipment of power so that an instantaneous power balance revealing the current value of power developed by the storage system is accessible to an electrical energy storage system management system.

[0031] - said comparison can be made by minimizing the difference in state value of charge.

[0032] - a successful recovery can also be a successful recovery of a value of temperature of the storage system, the prior characterization includes a temperature variation of the storage system, and the comparison is carried out between the vectors each consisting of one of the said determined charge state values ​​and the associated voltage and current values, as well as the last known or estimated temperature, with the known vectors of the prior characterization which also contain a temperature value.

[0033] - storage system management can be carried out remotely by Storage system audit report.

[0034] - the power decomposition can be a product between the delivered current, and voltage between terminals, said circulating current being increased in absolute value and said voltage between terminals being decreased and increased by values ​​relating to the normal use of the energy storage system.

[0035] - the energy storage system can be a stationary system, connected to a network terrestrial distribution via an inverter.

[0036] - said inverter can also connect a photovoltaic production system electricity audit terrestrial distribution network.

[0037] - the method can be used recursively to reconstruct several voltages, missing intensities and charge states at successive times.

[0038] The invention also relates to a device for reconstructing instantaneous values ​​of electrical quantities relating to an electrical energy storage system, the device comprising means for, at an instant following a successful recovery of the state of charge of said system, reviewing different possible decompositions into numerical values ​​of voltage and current of a current value, known for the purposes of reconstruction, of power developed by the storage system, in order to determine, for said possible decompositions, the resulting state of charge values ​​for the storage system taking into account a previous state of charge, and means for conducting a comparison between the vectors each consisting of one of said determined state of charge values ​​and the associated voltage and current values, and vectors known from a prior characterization of the energy storage system and each consisting,for an accessible physical state of the storage system, a state of charge value, and the associated voltage and current values, said comparison is carried out to determine the most probable decomposition among said possible decompositions.

[0039] Optionally and advantageously, said storage system may be interfaced with power transmission equipment so that an instantaneous power balance reveals the current value of power developed by the system of storage is accessible to an electrical energy storage system management system.

[0040] And said comparison can be made in the device by minimizing the difference between the state of charge values. List of figures

[0041] Fig. 1 represents an example of a system to which the invention applies.

[0042] Figure 2 represents an example of a temporal profile used for a mode of realization of the invention, the values ​​represented being the voltage across the battery terminals and the intensity of the current coming out of the battery.

[0043] Fig. 3 represents another aspect of this profile, the value shown being the state of charge of the battery.

[0044] Fig. 4 represents another aspect of this profile, the value shown being the power supplied by the battery.

[0045] Figure 5 represents another aspect of this profile, the value shown being a battery temperature.

[0046] Figure 6 represents the evolution of the voltage across the battery terminals as a function of time under a given and constant charge or discharge current.

[0047] Fig. 7 represents the observed relationship, which is not unambiguous, between the state of charge and the voltage across the terminals.

[0048] The [Fig.8] is a map of the state of charge as a function of the intensity of the outgoing or incoming current and the voltage across the terminals.

[0049] Fig.9 identifies on the time profile of figures 2 to 5, and more specifically in relation to the power value which has been detailed in Fig.4, an instant which serves as an example in the following figures.

[0050] The [Fig. 10] represents the voltage current pairs compatible with the power observed at the instant identified in [Fig.9].

[0051] Fig. 11 represents the load states for the different pairs of Fig. 10.

[0052] Figure 12 represents the comparison of the calculated state of charge in relation to the couples in Figures 10 and 11 with the states of charge measured for the same couples.

[0053] Fig. 13 represents an example of a time profile reconstructed using the invention - the value represented being the voltage.

[0054] Figure 14 represents the current value on this same example.

[0055] Figure 15 schematically illustrates the process used. Description related to the figures

[0056] [Fig. 1] In [Fig. 1], an electrical power management system is shown. 10 comprising a photovoltaic PV array, producing electricity in current A photovoltaic array (PV) and a battery, both operating on direct current, are connected to an onshore AC power grid via a single 100-watt inverter. The PV array and the battery are connected separately to the inverter. Power can be transferred from the PV array to the battery or vice versa with an efficiency of approximately 1, through the DC section of the 100-watt inverter.

[0057] This is one embodiment of the invention, which is not limited to such an arrangement. Here, the photovoltaic array and the battery, as long as the latter's state of charge is not too low, supply electrical power to the distribution network, which includes consumers. The connection between the distribution network and the inverter is therefore called the delivery point. The network also potentially includes producers. The battery can also be recharged, when its state of charge is not at its maximum, by the power supplied by the photovoltaic array, or by the network when the latter includes producers, or by both simultaneously.

[0058] In favorable conditions, the voltage across the battery terminals and the charging or discharging current of the battery are known at all times.

[0059] The invention is concerned with the situation in which data from the battery are unknown, while the data from the photovoltaic field, terminal voltage and output current, and those from the distribution network, again terminal voltage and current, are known at all times.

[0060] In terms of power, if v is the efficiency of the power inverter 100, and applying the convention that the power developed by the battery Pbatt is negative in discharge, we have the relation, which expresses the fact that the power supplied to the network Préseau is the sum of the powers transmitted to it by the inverter from the photovoltaic field (PPV) and the battery.

[0061] vPPV" ^^Batt P network

[0062] The battery power (or battery charging power) is therefore expressed as

[0063] n _ P network p 1 batt ~ ' v + rPV

[0064] Although the power developed at any instant by the battery is the product of the voltage across its terminals and the intensity of the current flowing through it, according to the equality

[0065] Pbatt(t) = Ibatt(t) x Ubatt(t,I)

[0066] Since the voltage U depends on the current I, and these two quantities are unknown, there are many pairs of values ​​(U, I) allowing the equation to be verified even when limiting the reasoning as required to the intervals of voltages and currents admissible by the battery.

[0067] The instantaneous values ​​of the power developed by the battery Pbatt(t) known thanks to the instantaneous power at the delivery point and the efficiencies.

[0068] But in the scenario of interest in which the invention is situated, the instantaneous values ​​of voltage across the terminals Ubatt (t) and current delivered Ibatt (t) are unknown - these are missing data.

[0069] The voltage across the battery terminals is temperature-dependent. To account for this, an additional dimension is added to all the matrices in the calculations presented below. However, to simplify the description, this temperature dependence is not systematically mentioned in the following description.

[0070] We go back to Ubatt and Ibatt using the state of charge SOC (for State of charge) of the battery.

[0071] A preliminary step is the generation of the SOCread state of charge value matrix as a function of the voltage value U and the current value I, these values ​​having been determined experimentally during a prior characterization of the storage system, before, for example, making it available to the operator.

[0072] The suffix "read" refers to the fact that the charge states in question have been measured, in this case during a prior characterization phase.

[0073] A later step, during the operation phase, and in a situation where data is missing, is the identification for a given instantaneous battery power Pbatt(t), corresponding to that known for example from the delivery point, of all the current and voltage pairs (I, U) which allow us to have Pbatt = U*I, that is to say to find this delivered power.

[0074] An associated step is the calculation, by coulometry (integration of the current over time), of the corresponding state of charge SOCcaic for each voltage U, current I pair. The suffix "cale" refers to the fact that the state of charge in question is derived from a calculation, in this case by coulometry, and has not been measured. The calculation can be carried out in several fairly similar ways, essentially taking into account the current being studied and the time step (the time elapsed since the last known value, if it has been measured and recorded, or recognized, assuming it has been calculated, of the state of charge).

[0075] A subsequent step, during the operating phase, is the comparison, for each voltage and current pair (Ux,Ix) of the measured state of charge SOCread (Ux,Ix) with the calculated state of charge SOCcaic(Ux,Ix)

[0076] Two matrices are subtracted, and the minimum value of the matrix resulting from this subtraction is sought. The coordinates of this minimum are the coordinates of the pair unique U,I leading to the power developed by the battery Pbatt at the correct state of charge SOC, taking into account the prior characterization of the system.

[0077] The process is validated by checking the values ​​at different points and reconstructing the time profile.

[0078] [Fig.2] A temporal profile represented in figures is chosen to discuss the process 2 to 5. For this example, we are working with a time profile whose results come from laboratory tests on a given lithium-ion (Li-ion) cell at a fixed temperature of 25°C. The length of the profile is represented on the x-axis and is 140 h. Figure 2 represents on the y-axis the current delivered by the battery from -30 A to +30 A (the curve is marked I), and the voltage across the battery terminals from 2.8 to 4.4 V (the curve is marked U).

[0079] The minimum and maximum currents of this time profile are between C / 25 and C / 3 during charging and -C / 25 and -C / 3 during discharging. C / 25 and C / 3 are current values ​​at which the battery capacity would be consumed in 25 hours and 3 hours, respectively.

[0080] [Fig.3] In [Fig.3], the cell's charge state is represented on the ordinate from 0 to 100%. The ordinates are identical to those of [Fig.2] (this is also the case in figures 4 and 5).

[0081] [Fig.4] In [Fig.4], the power developed by the battery is represented in ordered from -100 W to +100 W.

[0082] [Fig.5] In [Fig.5], the battery temperature is represented on the ordinate, from 24 to 26°C.

[0083] [Fig.6] In parallel, a database of read state values ​​is generated SOCread charge as a function of the voltage across the battery terminals Ubatt and the current flowing at the output of the battery Ibatt (current delivered), from separate tests carried out in charge and discharge at constant current between C / 25 and C / 3.

[0084] During such a test, both the state of charge and the voltage increase, when charging, and both the state of charge and the voltage decrease, when discharging.

[0085] Performing a large number of independent tests for a constant current value allows us to obtain averaged and therefore more reliable data.

[0086] For example, twelve different current values ​​are used (which can therefore make 24 curves, these values ​​being used in discharge and charge).

[0087] The results for C / 25 and C / 3 are presented in [Fig. 6] where the x-axis represents time from 0 to 10,000 s (approximately 3 h). Curve 1 is the charging curve at C / 25, curve 2 is the charging curve at C / 3, curve 3 is the discharging curve at C / 25, and curve 4 is the discharging curve at C / 3.

[0088] [Fig.7] From these data, the state of charge SOC of the battery can be represented in function of voltage. In [Fig. 7] the state of charge is on the x-axis from 0 to 100%, the Voltage on the ordinate ranges from 2.8 V to 4.2 V. It is clear that the relationship is not unambiguous.

[0089] [Fig.8] A surface of SOCread charge state values ​​is also generated as a function The current (Ibatt) flowing across the battery terminals and the voltage (Ubatt) between the battery terminals are plotted using measurement points, with interpolation between these points smoothing the surface. The state of charge is plotted on the y-axis, while the two x-axis values ​​are voltage (in volts, from 2.8 to 4.2) and current during discharge and charge (in amperes, from -25 to +25).

[0090] [Fig.9] The method then includes taking charge of a given battery. For Given a battery power value Pbatt, from any battery time profile, we calculate all possible pairs (I,U) which allow us to verify Pbatt = IxU.

[0091] The chosen point is represented on [Fig.9] (which reproduces the curve of [Fig.4]): it is a point for which the power developed by the battery is -89.1886 W.

[0092] [Fig. 10] The choices of the two parameters I and U are limited by the minimum voltage Umin and the maximum voltage Umax of the battery, and the maximum current Imax and the minimum current Imin. The current extrema can be those chosen previously to generate the load state surface SOCread or depend on the minimum and maximum currents of the usage time profile and / or possibly of the cell.

[0093] The points obtained are shown in [Fig. 10], where the x-axis represents current and the y-axis represents voltage. The points shown are discharge points: the current is between -22.4 A and -21.2 A.

[0094] The apparently linear nature, in [Fig. 10], of the relationship between U and I is linked to the small amplitude of the values ​​- in fact U and I vary up to a multiplicative factor, like the inverse of each other, P being fixed and constituting their product.

[0095] [Fig. 11] The state of charge (SOC) is then calculated for each pair (U,I) generated in the previous step. The point identified by a developed power Pbatt and of interest is considered the first missing point in the profile. Thus, by retrieving standard, reliable data, the state of charge SOC of the battery at the previous point is known. It is therefore possible to calculate, using coulometry, the new state of charge SOC corresponding to the missing battery power Pbatt. This calculation is performed for each pair (U,I).

[0096] A series of state of charge values ​​is obtained. This is represented in [Fig. 11]: on the abscissa are the numbers of the successive experimental pairs (U,I), which here number approximately 120, and on the ordinate are the state of charge values ​​SOC around 87%, and in an interval of length approximately 7.106%.

[0097] [Fig. 12] For each voltage and current pair (Ux,Ix) we then seek SOCread(Ux,Ix) represented by points 200 and compare it to SOCcaic(Ux,Ix), represented by points 205.

[0098] We then subtract these two SOC vectors and determine the U,I values ​​associated with the minimum of the result, which correspond to the most probable U and I values, the closest approximations to reality, among those for which the coulometry calculations have been made and taking into account the characterization which has been made previously of the storage system.

[0099] The values ​​associated with the minimum are the coordinates of the unique pair of voltage and current values ​​(U,I) corresponding to the power developed by the battery P batt at the correct state of charge SOC.

[0100] We check that the results are satisfactory for other discrete points of the profile, and we perform calculations of the deviations and errors.

[0101] In the table below, the first column indicates a point number, the second the power output in watts, the third the effective voltage in volts, the fourth the effective current in amperes, and the last two the calculated voltage and current values. It can be seen that these are always very close to the actual values. J___________________ ]_______32^2? ^.rwï.zx. î 4.14^3 ......................... . 'T. X x J * VX ? -vX* x <

[0102] This table shows that for any absent point, following in time a point for which the values ​​are known, the invention can recalculate the pair (U,I) and therefore the state of charge SOC with a very low error. And by proceeding step by step, a profile is reconstructed.

[0103] [Fig. 13] Fig. 13 shows a portion of the time profile selected in the example, reconstructed according to the method. These are the voltage values.

[0104] [Fig. 14] The [Fig. 14] shows the current values ​​reconstructed according to the same process and therefore corresponding to the voltage values ​​of the [Fig. 14].

[0105] Without the use of the method, voltage and current values ​​are much less well controlled in the absence of data feedback from the battery, and quickly, useful information is no longer accessible or constructible.

[0106] According to an improved embodiment, errors in voltage and current that increase during relaxation phases, in which the power developed is zero, are corrected. Thanks to the method, as soon as the The power developed by the battery is again non-zero, and the voltage and current data are again correctly obtained according to the principles of the invention. Thus, an improvement is made to account for the voltage evolution during the relaxation phases and thereby correct the prediction. During these phases, the power developed Pbatt = 0 W, and the current Ibatt = 0 A. To determine the evolution of the voltage Ubatt over time, a relaxation voltage map learned in advance from the initial calibration time profile is used. This map depends, in particular, on the value and sign of the current preceding this relaxation during a pause. Therefore, several maps are established in advance, indexed by a negative or positive current value, and the appropriate map is chosen based on the current value observed just before the relaxation phase.

[0107] According to one variant, compatible with the preceding ones, an additional dimension or variable corresponding to a temperature, between -40°C and +60°C, is added to the data recorded and calculated in matrix form, along with a temperature sensor on the battery. Thus, temperature is also taken into account. The temperature considered is that of the inverter or an element of the environment, such as the photovoltaic array, particularly if the battery charge and discharge currents are less than, for example, C / 3, in which case the battery does not heat up on its own and adopts the ambient temperature, which can thus be obtained by the inverter or another element communicating flawlessly with the EMS energy manager, ensuring all or part of the BMS function.

[0108] The invention applies to a system comprising a battery and other components which can be a source of power (for example a renewable energy source, or a territorial distribution network supplying power) or conversely a sink of power (for example a local consumer or a territorial customer network), or alternatively a source or sink of power depending on the periods considered.

[0109] Once the missing data has been reconstructed according to the invention, the battery management system (BMS), or the energy manager of the EMS system, can define, using the reconstructed data, a power command to be developed by the battery for the instant or instants following the data reconstruction.

[0110] [Fig.15] In [Fig.15], the process has been represented according to one embodiment of the invention.

[0111] A preliminary mapping is carried out once and for all during a step E0 to characterize the energy storage system and to know its state of charge as a function of temperature values, terminal voltage and delivered current, according to the principles discussed in relation to figures 6 to 8.

[0112] Then, in a circular fashion, cycles of steps El and E2 are carried out as follows.

[0113] Coulometry calculations are carried out during a step El using the power developed communicated by the delivery point or any network partner, to determine the possible states of load, according to the principles discussed in Figures 10 and 11.

[0114] Then, based on the results of these calculations, the difference between the possible load states calculated in step El and those present in the characterization of step E0 is minimized according to the principles discussed in relation to [Fig. 12] and during a step E2. This allows reconstructed values ​​to be determined.

[0115] The coulometry calculations of step El are performed using a load state transmitted at least once, then during subsequent recurrences, are performed with the latest load state value reconstructed on date.

Claims

Demands

1. A method for reconstructing instantaneous values ​​of electrical quantities relating to an electrical energy storage system, the method comprising, at an instant following a successful recovery of the state of charge of said system, a review of different possible decompositions into numerical values ​​of voltage and current of a current value, known for the purposes of reconstruction, of power developed by the storage system, to determine (E1), for said possible decompositions, the resulting state of charge values ​​for the storage system taking into account a previous state of charge, and a comparison (E2) between the vectors each consisting of one of said determined state of charge values ​​and the associated voltage and current values, and vectors known from a prior characterization (EO) of the energy storage system and each consisting, for an accessible physical state of the storage system,of a state of charge value and the associated voltage and current values, said comparison being carried out to determine the most probable decomposition among said possible decompositions.

2. A method for reconstructing values ​​according to claim 1, characterized in that said storage system is interfaced with a power transmission equipment (100) such that an instantaneous power balance revealing the current value of power developed by the storage system is accessible to an electrical energy storage system management system.

3. Method for reconstructing values ​​according to claim 1 or claim 2, characterized in that said comparison (E2) is made by minimizing the difference between the load state values.

4. A method for reconstructing values ​​according to any one of claims 1 to 3, characterized in that the successful recovery is also a successful recovery of a temperature value of the storage system, the prior characterization includes a temperature variation of the storage system, and the comparison is carried out between vectors, each consisting of one of said determined charge state values ​​and the associated voltage and current values, as well as the last known or estimated temperature, with the vectors known from the prior characterization which also contain a temperature value.

5. A method for reconstructing values ​​according to any one of claims 1 to 4, characterized in that the management of the storage system is carried out in a manner delocalized with respect to said storage system.

6. A method for reconstructing values ​​according to any one of claims 1 to 5, characterized in that the power decomposition is a product of current delivered and voltage between terminals, said current being increased in absolute value and said voltage between terminals being decreased and increased by values ​​relating to the normal use of the energy storage system.

7. A method for reconstructing values ​​according to any one of claims 1 to 6, characterized in that the energy storage system is a stationary system, connected to a terrestrial distribution network by an inverter.

8. A method for reconstructing values ​​according to claim 7, characterized in that said inverter also connects a photovoltaic electricity production system to said terrestrial distribution network.

9. A method for reconstructing values ​​according to any one of claims 1 to 8, characterized in that the method is used recursively to reconstruct several missing voltages, currents and charge states at successive times.

10. A device for reconstructing instantaneous values ​​of electrical quantities relating to an electrical energy storage system, the device comprising means for, at an instant following a successful recovery of the state of charge of said system, performing a review of different possible decompositions into numerical values ​​of voltage and current of a current value, known for the purposes of reconstruction, of power developed by the storage system, in order to determine, for said possible decompositions, the resulting state of charge values ​​for the storage system taking into account a previous state of charge, and means for conducting a comparison between the vectors each consisting of one of said determined state of charge values ​​and the associated voltage and current values, and vectors known from a prior characterization of the energy storage system and each consisting, for an accessible physical state of the storage system, of a state of charge value and associated voltage and current values, the said comparison being carried out to determine the most probable decomposition among the said possible decompositions.

11. Value reconstruction device according to claim 10, characterized in that said storage system is interfaced with a power transmission equipment (100) such that an instantaneous power balance revealing the current value of power developed by the storage system is accessible to an electrical energy storage system management system.

12. A value reconstruction device according to claim 10 or claim 11, characterized in that said comparison is made by minimizing the difference between the load state values.

Images