RELIABLE ESTIMATION OF THE STATE OF CHARGE OF A VEHICLE'S CELLULAR BATTERY
By identifying critical and non-critical cells within a battery and applying differential sampling periods and filters, the method addresses inefficiencies in existing SoC estimation, achieving precise and efficient state of charge estimation for batteries with many cells.
Patent Information
- Application Number
- FR2022013504
- Authority / Receiving Office
- FR · FR
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2022-12-16
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2042-12-16
AI Technical Summary
Existing methods for estimating the state of charge (SoC) of a battery with a large number of cells are inefficient and imprecise, leading to computational overload and significant estimation errors, particularly in the presence of aging and dispersion within the battery.
A method that identifies critical and non-critical cells within the battery, applying different sampling periods and filters to estimate the SoC, using extended Kalman filters and equivalent electrical circuit models to reduce computational load while ensuring precise and rapid estimation.
This approach significantly reduces calculation load while providing accurate and rapid estimation of the SoC for both critical and non-critical cells, ensuring high accuracy and reliability in battery state estimation.
Smart Images

Figure 00000017_0000 
Figure 00000018_0000
Abstract
Description
Title of the invention: RELIABLE ESTIMATION OF THE STATE OF CHARGE OF A CELLULAR BATTERY OF A VEHICLE Technical field of the invention
[0001] The invention relates to vehicles comprising at least one rechargeable cellular battery, and more precisely to the estimation of the current state of charge (or SoC (“State of Charge”)) of such batteries. State of the art
[0002] Certain vehicles, possibly of the automobile type, include a rechargeable cellular battery constituting a main (or traction) battery because it is responsible for powering at least one electric motor of their powertrain (or powertrain).
[0003] Here, the term "cellular battery" means a battery comprising numerous electrical energy storage cells, possibly electrochemical (for example of the lithium-ion (or Li-ion) or Ni-Mh or Ni-Cd type). For example, such a cellular battery may be of the low or medium voltage type (typically at least 400 V).
[0004] Furthermore, the term "battery" here means the assembly comprising the cells, a support structure for these cells, a tray (or housing) housing the support structure with the elements, as well as generally an electronic control and protection box.
[0005] Such a cellular battery is usually associated with a battery case which comprises internal voltage / current / temperature measuring means and a battery calculator. The latter centralizes the current measurements, the voltage measurements and the internal temperature measurements (inside the cellular battery), and determines or estimates parameters of this cellular battery based on these measurements, and in particular its internal resistance, its minimum voltage and its current state of charge (or SoC (State of Charge)).
[0006] This state of charge (or SoC) is an important parameter in an electric GMP vehicle because it makes it possible to estimate the mileage range of this vehicle.
[0007] At least five techniques can be implemented to estimate the SoC of a (cellular) battery.
[0008] A first technique consists of estimating the SoC of the battery using the consumption count (in Ah). However, relying solely on current information is not optimal due to the accumulation of errors in the current sensors over time and the impossibility of correcting the capacity error when the battery ages.
[0009] A second technique consists of estimating the SoC of the battery by means of an extended Kalman filter (or EKF) applied only to the minimum and maximum cell voltages. In this case, the battery cell voltages are composed of the open circuit voltage (or OCV) and the contribution of the direct current resistance (or DCR). Given the aging and dispersion within the battery at the beginning of its life, it is rare that the minimum voltage (respectively the maximum voltage) leads to a minimum SoC (respectively a maximum SoC). Therefore, this second solution is not accurate and may lead to poor anticipation of functional limitations for the battery cell voltages.
[0010] A third technique consists of estimating the SoC of the battery using the extended Kalman filter (EKF) and the so-called recursive least squares (or RLS) technique applied to an equivalent electric circuit model (or EEC) and a hysteresis model for all the cells of the battery in the same sampling period. By estimating the SoC of each cell of the battery using an EKF / RLS algorithm each time, the limits of the computing capacity of the battery calculator can be quickly reached, particularly when the number of cells is very large (typically for an 800 V battery).
[0011] A fourth technique is to estimate the battery SoC using the Extended Kalman Filter (EKF) and Recursive Least Squares (RLS) technique applied to an Equivalent Electrical Circuit (EEC) model and a hysteresis model for batches of battery cells with a sequential approach to fit the computational constraints of the battery calculator. This technique is an improved version of the third technique that does not completely solve the problem because the computational load will still increase with the number of cells but only less than before.
[0012] A fifth technique is to estimate the SoC of the battery by means of a virtual average cell by using the average value of the cell voltages and calculating the SoC of the virtual average cell and deducing the SoC of all the cells of the battery by calculating the differences in SoC of each cell with respect to the virtual average cell. This is efficient and can considerably reduce the computational load of the battery calculator. However, the virtual average cell is difficult to define and does not correspond to any physical reality. As a result, large errors in estimating the SoC can occur, which is problematic, especially for low SoC values and / or in the presence of low temperatures for which the values of the internal direct current resistors (DCRs) vary considerably.
[0013] There is therefore no known technique for estimating the SoC of a battery comprising a large number of cells with sufficient precision using a conventional battery calculator.
[0014] The invention therefore aims in particular to improve the situation. Presentation of the invention
[0015] For this purpose, it proposes in particular a method for estimating the state of charge of a battery comprising N cells and equipping a vehicle.
[0016] This estimation method is characterized by the fact that it comprises a step in which:
[0017] - we identify among the N cells M so-called critical cells, then
[0018] - we determine for each of these M identified critical cells and according to a first sampling period, a state of charge by applying to a chosen cell model a first chosen filter using measured impedance and capacity of the cell considered, then a so-called nominal cell is selected from the M critical cells identified, then
[0019] - for each of NM non-critical cells, we determine, according to a second sampling period greater than the first sampling period and by applying to the chosen cell model a second chosen filter, a new state of charge difference depending on a previous state of charge difference, a current measured in the battery, a measured capacity of the non-critical cell considered and the measured capacity of the nominal cell, then a new state of charge by summing the determined state of charge of the nominal cell and the new state of charge difference, and
[0020] - the state of charge of the battery by exploiting the new determined states of charge of the NM non-critical cells and the determined states of charge of the M critical cells.
[0021] Thanks to the invention, it is possible to significantly reduce the computational load of the estimation device while guaranteeing a precise and very rapid estimation of the charge states of the M most critical cells at each first sampling period and ensuring a rapid and precise estimation of the charge states of the NM remaining non-critical cells and of the charge state of the cell battery at each second sampling period.
[0022] The estimation method according to the invention may include other characteristics which may be taken separately or in combination, and in particular:
[0023] - in its step, the first filter chosen can be an extended Kalman filter (or EKF) with two states, a first state being the state of charge of the critical cell considered and a second state being a polarization voltage of the critical cell considered;
[0024] - in its step, the second filter chosen can be an extended Kalman filter (or EKF) to a state, this state being the state of charge of the non-critical cell considered;
[0025] - in its step the chosen cell model can be an electric circuit model n-RC equivalent, with n > 1;
[0026] - in its step, we can use a second sampling period which is equal at K times the first sampling period, where K is a number of groups into which the NM non-critical cells are distributed, and it is possible to determine during each first sampling period the new charge states of the non-critical cells of one of the K groups so as to have the NM new charge states of the NM non-critical cells at the end of each second sampling period;
[0027] - in its step, we can select the M critical cells by determining among the N cells those with the largest M / 2 upper operating limits and those with the smallest M / 2 lower operating limits relative to predefined operating limits;
[0028] - in the presence of the last option, in its step, we can determine among the M / 2 critical cells having the largest M / 2 upper operating limits and among the M / 2 critical cells having the smallest M / 2 lower operating limits the one that is most critical and which becomes the nominal cell.
[0029] The invention also provides a computer program product comprising a set of instructions which, when executed by processing means, is capable of implementing an estimation method of the type presented above, in a vehicle comprising a battery comprising N cells, to estimate a state of charge of this battery.
[0030] The invention also proposes a state of charge estimation device intended to equip a vehicle comprising a battery comprising N cells.
[0031] This estimation device is characterized by the fact that it comprises at least one processor and at least one memory arranged to carry out the operations consisting of:
[0032] - to identify among the N cells M cells called critical, then
[0033] - to be determined for each of these M identified critical cells and according to a first sampling period, a state of charge by applying to a chosen cell model a first chosen filter using measured impedance and capacity of the cell considered, then selecting a so-called nominal cell from among the M critical cells identified, then
[0034] - for each of NM non-critical cells, to be determined, according to a second sampling period greater than the first sampling period and by applying to the chosen cell model a second chosen filter, a new difference state of charge as a function of a previous difference in state of charge, a current measured in the battery, a measured capacity of the non-critical cell considered and the measured capacity of the nominal cell, then a new state of charge by summing the determined state of charge of the nominal cell and the new difference in state of charge, and
[0035] - to determine according to this second sampling period the state of charge of the battery by exploiting the new determined states of charge of the NM non-critical cells and the determined states of charge of the M critical cells.
[0036] The invention also proposes a vehicle, possibly of the automobile type, and comprising, on the one hand, a battery comprising N cells, and, on the other hand, an estimation device of the type presented above. Brief description of the figures
[0037] Other characteristics and advantages of the invention will appear on examining the detailed description below, and the appended drawings, in which:
[0038] [Fig-1] schematically and functionally illustrates an example of the embodiment of a vehicle comprising a GMP transmission chain with an electric motor associated with a cellular battery and associated with a battery calculator, and an estimation device according to the invention,
[0039] [Fig.2] schematically and functionally illustrates an example of the embodiment of a battery calculator comprising an exemplary embodiment of an estimation device according to the invention, and
[0040] [Fig.3] schematically illustrates an example of an algorithm implementing a estimation method according to the invention. Detailed description of the invention
[0041] The invention aims in particular to propose an estimation method, and an associated estimation device DE, intended to allow the estimation of the current state of charge (or SoC (State of Charge)) of a cellular battery of a vehicle V.
[0042] In the following, it is considered, by way of non-limiting example, that the vehicle V is of the automobile type. It is for example a car, as illustrated in [Fig.l]. But the invention is not limited to this type of vehicle. It relates in fact to any type of vehicle comprising a cellular battery whose state of charge is to be estimated. Thus, it relates to land vehicles (utility vehicles, camper vans, minibuses, coaches, trucks, motorcycles, road machinery, construction machinery, agricultural machinery, leisure machinery (snowmobile, kart), tracked vehicles, trains and trams, for example), aircraft and boats.
[0043] Furthermore, it is considered in the following, by way of non-limiting example, that the vehicle V comprises a powertrain transmission chain (or GMP) of the all-electric type (and therefore whose drive is provided exclusively by at least one MME electric motor). But the GMP could be of the hybrid type (thermal and electric).
[0044] [Fig.l] schematically shows a vehicle V comprising an electric GMP transmission chain (and therefore an electric motor MME), a cellular battery BC associated with a battery calculator CB, a main electrical circuit CEP connected to a charging connector CR, a charger CH, a converter CV, and an estimation device DE according to the invention.
[0045] The main electrical circuit (or "high voltage") CEP is connected, on the one hand, to the cellular battery BC, and, on the other hand, to electronic equipment, such as for example the converter CV and the electric motor MME. It also allows the recharging of the cellular battery BC by electrical energy from an external power source and temporarily connected to the charging connector CR of the vehicle V. This main electrical circuit CEP therefore comprises at least one power supply circuit PI ensuring the coupling between the cellular battery BC and at least the electric motor MME and converter CV, and a recharging circuit P2 connected to the recharging connector CR and allowing the recharging of the cellular battery BC when the latter (CR) is temporarily connected to an external power source.
[0046] In the example illustrated non-limitingly in [Fig.l] the recharging circuit P2 makes it possible to recharge the cellular battery BC not only in direct current (or mode 4), but also in alternating current (or mode 2 or 3), under the control of at least one charger computer CC (here forming part of the charger CH and managing in particular the exchanges of information with the external power source during a recharging phase).
[0047] But in alternative embodiments not illustrated, the charging circuit P2 could only allow direct current charging (or mode 4) or only alternating current charging (or mode 2 or 3). It is recalled that alternating current charging is done via the CV converter.
[0048] It will be noted, as illustrated non-limitingly in [Fig.l], that the CV converter can be part of the CH charger.
[0049] The transmission chain has a GMP which is, here, purely electric and therefore which comprises, in particular, an electric motor MME. Here, the term "electric motor" means an electric machine arranged so as to provide torque to move the vehicle V when it is supplied with electrical energy, as well as possibly to recover torque in the transmission chain.
[0050] The electric motor MME (here an electric motor) is here coupled to the cellular battery BC via the power supply circuit PI of the main electrical circuit CEP, in order to be supplied with electrical energy, as well as possibly to supply this cellular battery BC with electrical energy, for example during a regenerative braking phase.
[0051] The cellular battery BC here powers the electric motor MME, it constitutes a main (or traction) battery. It comprises N electrical energy storage cells (not illustrated), possibly electrochemical (for example of the lithium-ion (or Li-ion) or Ni-Mh or Ni-Cd type). Also for example, the cellular battery BC can be of the low voltage type (typically 450 V), or medium voltage (typically from 600 V to 1000 V). But it could be of the high voltage type. It will be noted that the invention is all the more useful as the number N of cells is large (and in particular greater than 100).
[0052] As mentioned above, the invention proposes in particular an estimation method intended to allow the estimation of the state of charge SoCBC of the battery BC of the vehicle V.
[0053] This (estimation) method can be implemented at least partially by the estimation device DE (illustrated at least partially in FIGS. 1 and 2) which comprises for this purpose at least one processor PR1, for example a digital signal processor (or DSP ("Digital Signal Processor")), and at least one memory MD. This estimation device DE can therefore be produced in the form of a combination of electrical or electronic circuits or components (or "hardware") and software modules (or "software"). For example, it can be a microcontroller.
[0054] The memory MD is live in order to store instructions for the implementation by the processor PR1 of at least part of the estimation method. The processor PR1 may comprise integrated (or printed) circuits, or several integrated (or printed) circuits connected by wired or wireless connections. An integrated (or printed) circuit is understood to mean any type of device capable of carrying out at least one electrical or electronic operation.
[0055] In the example illustrated non-limitingly in Figures 1 and 2, the estimation device DE is part of the battery calculator CB. But this is not obligatory. Indeed, the estimation device DE could comprise its own dedicated calculator, which is then coupled to the battery calculator CB, or could be part of another calculator on board the vehicle V, for example.
[0056] As illustrated non-limitingly in [Fig. 3], the (estimation) method, according to the invention, comprises a step 10-40 which is implemented at least each time the vehicle V is used (and in particular each time its GMP is in operation).
[0057] Step 10-40 of the method firstly comprises a sub-step 10 in which one (the estimation device DE) identifies among the N cells of the cellular battery BC M cells called critical. It will be understood that M is strictly less than N (for example, M can be between N / 20 and N / 5). It will also be understood that the remaining NM cells are called non-critical.
[0058] For example, in sub-step 10 one (the estimation device DE) can select the M critical cells by determining among the N cells those which have the largest M / 2 upper operating limits and those which have the smallest M / 2 lower operating limits with respect to predefined operating limits (by the manufacturer of the cellular battery BC).
[0059] Several predefined operating limits can be used. In the following, it is considered, by way of non-limiting example, that the operating limits are relative to the respective states of charge (or SoCs) of the M critical cells. It will be noted that the minimum and / or maximum states of charge which are used to initially determine the functional limits can be deduced from voltage measurements carried out on the cells and more precisely from measurements of the open circuit voltages (or OCV).
[0060] But the operating limits could be relative to the capacities (respectively SOHC ("State Of Health of Capacity")), or impedances (respectively SOHR ("State Of Health of Resistance")), or respective internal temperatures of the M critical cells. In this case, they result from measurements of capacities or impedances (or resistances).
[0061] Step 10-40 of the method also comprises a sub-step 20 in which one (the estimation device DE) determines according to a first sampling period T whether at least one state of charge SoO(k) (as well as possibly a bias voltage V'rc) for each of the M critical cells identified and designated by the exponent i. For this purpose, one (the estimation device DE) applies to a chosen cell model a first chosen filter which uses the measured impedance Z; and capacitance Capa; of the critical cell i considered.
[0062] The parameters of each cell (and in particular of the critical cells), and in particular the impedance and the capacity, are here considered as available within the vehicle V, and therefore the means implemented to obtain them are not described below.
[0063] For example, the first sampling period Tsi may be between 10 ms and 200 ms. As an illustrative example, the first sampling period Tsi may be equal to 100 ms.
[0064] In sub-step 20, one (the estimation device DE) also selects a so-called nominal cell from among the M critical cells identified in sub-step 10.
[0065] For example, in sub-step 20 one (the estimation device DE) can determine among the M / 2 critical cells having the M / 2 upper operating limits the largest and among the M / 2 critical cells having the M / 2 smallest lower operating limits the one that is most critical. In this case, the cell that is most critical among the M becomes the nominal cell and corresponds to the current physical reality in the cell battery BC for the first sampling period Tsi considered.
[0066] Step 10-40 of the method also comprises a sub-step 30 in which one (the estimation device DE) determines for each of the NM non-critical cells, according to a second sampling period Ts2, greater than the first sampling period Tsb and by applying to the chosen cell model a second chosen filter, a new difference in state of charge ASoO(k). The exponent j here designates one of the NM non-critical cells. Each difference in state of charge ASoO(k) of a non-critical cell j is a function of the previous (and therefore last) difference in state of charge ASoO(kl), of the current I measured in the cell battery BC (i.e. the current flowing through it), of the measured capacity Capai of the non-critical cell j considered and of the measured capacity Capanom of the nominal cell.
[0067] Then, in this sub-step 30, we (the estimation device DE) always determine for each of the NM non-critical cells, according to the second sampling period Ts2, a new state of charge SoO(k) by summing the determined state of charge of the nominal cell SoCnom(k) and the new difference in state of charge ASoO(k), i.e. SoO(k) = SoCn°”(k) + ASoO(k).
[0068] Step 10-40 of the method also comprises a sub-step 40 in which the estimation device DE determines, according to the second sampling period Ts2, the state of charge SoCBC of the battery BC by exploiting the new determined states of charge SoO(k) of the NM non-critical cells and the determined states of charge SoC1(k) of the M critical cells.
[0069] In other words, only M critical cells i (i = 1 to M) are subject to calculations of their state of charge SoO(k) (as well as possibly their polarization voltage V' RC) with a first filter and according to the first sampling period Tsb while the remaining NM non-critical cells j (j = 1 to NM) are subject to calculations of at least their state of charge SoO(k) less frequently according to the second sampling period Ts2 (> Tsi), which makes it possible to obtain an estimate of the state of charge SoCBC of the cellular battery BC at each second sampling period Ts2.This allows to closely monitor the M critical cells using more complete cell battery models with many states to achieve high accuracy and reliability and to significantly reduce the computational (or CPU) load of the estimation device DE (and thus here of the battery calculator CB), while ensuring an accurate and very fast estimation of the states of charge SoC'(k) of the M most critical cells at each first sampling period Ts[ and in . ensuring fast and accurate estimation of the states of charge SoO(k) of the remaining non-critical NM cells and the state of charge SoCBC of the cell battery BC at every second sampling period Ts2.
[0070] While waiting for a complete refresh of the state of charge of the NM non-critical cells which is done every Ts2 with the entire resolution of the second filter, the state of charge of the NM non-critical cells can continue to be updated every Tsi with the equation SoO(k) = SoCnom(k) + ASoCj(prior), where ASoO(prior) is the SOC difference corresponding to the last multiple instant of Ts2 (i.e. the SOC difference separating the state of charge of the non-critical cell (j) from that of the state of charge of the nominal cell). Thus, one can also permanently benefit from the updating of the SoCnom of the nominal cell in all the first sampling periods Tsi, and one can optimize this with the second filter every second sampling period Ts2.
[0071] Thus we constantly benefit from all the Tsl of the update of the SOC of the nominal cell and we come to optimize this with the second filter each Ts2.
[0072] For example, in sub-step 20 of step 10-40, the first filter chosen may be a two-state extended Kalman filter (or EKF). In this case, the first state may be the state of charge SoC'(k) of the critical cell i considered and the second state may be the bias voltage V'RC of the critical cell i considered. But this is not obligatory. Indeed, other first filters with at least two states may be used, provided that at least two of these states are the state of charge SoO(k) and the bias voltage V'rc.
[0073] Also for example, in sub-step 30 of step 10-40, the second filter chosen may be a one-state extended Kalman filter (or EKF). In this case, the single state may be the charge state SoO(k) of the non-critical cell j considered. But this is not obligatory. Indeed, other second one-state filters may be used.
[0074] Also for example, in sub-steps 20 and 30 of step 10-40, the chosen cell model may be an n-RC equivalent electrical circuit model, with n > 1, well known to those skilled in the art. Thus, it will be possible to use a 1-RC equivalent electrical circuit model or, to improve the accuracy, a 2-RC equivalent electrical circuit model and more generally an n-RC equivalent electrical circuit model (with n > 2). It is also possible to add to an n-RC equivalent electrical circuit model (with n > 1) a simple hysteresis model to increase the accuracy of the state of charge estimations, in particular in the presence of flat SoC-OCV curves around 60% to 30% of SoC for LFP (“Lithium Iron Phosphate” or (LiFePO4)) type cells.
[0075] It is recalled that in an equivalent 1-RC electrical circuit model the cell is electrically represented by a resistor R0 mounted in series with a set comprising a resistor RI connected in parallel with a capacitor.
[0076] Also for example, and as illustrated non-limitingly in [Fig. 3], in sub-step 30 of step 10-40 one (the estimation device DE) can use a second sampling period Ts2 which is equal to K times the first sampling period Tsb where K is a number of groups in which the NM non-critical cells are distributed. In this case, and as illustrated non-limitingly in [Fig. 3], one (the estimation device DE) can determine during each first sampling period the new charge states of the non-critical cells] of one of the K groups so as to have the NM new charge states SoCj(k) of the NM non-critical cells j at the end of each second sampling period Ts2.
[0077] It will be understood that thus, during each new first sampling period Tsi, the states of charge SoO of the non-critical cells j of one of the K groups are updated. Consequently, after K successive first sampling periods Ts[, there is an update of all the states of charge SoCj of the NM non-critical cells j which can then be used with the states of charge SoC of the M critical cells obtained during the very last first sampling period Tsb to obtain the estimate of the state of charge SoCBC of the cellular battery BC for the second sampling period Ts2 which has just elapsed.
[0078] Such processing by groups of non-critical cells j during each first sampling period Tsi advantageously makes it possible to further reduce the quantity of computing resources to be used.
[0079] It is important to note that at each first sampling period Tsi a new selection of M critical cells is made, and therefore the M critical cells can vary from one first sampling period Tsi to another. Similarly, at each first sampling period Tsi a new selection of a nominal cell is made and therefore the nominal cell can vary from one first sampling period Tsi to another.
[0080] An implementation of the invention is described below by means of equations, when the first filter is a two-state extended Kalman filter (or EKF), the second filter is a one-state extended Kalman filter (or EKF), and the cell model is a 1-RC equivalent electrical circuit model.
[0081] After selecting the M critical cells i from among the N cells of the cellular battery BC, a first two-state EKF filter is applied to the cell model for each of these M critical cells i using the following state equations of the state of charge SoC'(k) and bias voltage V'RC:
[0082] [Math.l] V^k] = OCV(SoC 1 [^] ) + x
[0083]
[0084]
[0085]
[0086]
[0087]
[0088]
[0089]
[0090]
[0091]
[0092]
[0093] where i G {1, 2,..., M], OCV is the open circuit voltage (which depends on the SoC and may also depend on internal temperature and aging), V[[Æ] is the terminal voltage of cell i, R'o and R'i are the resistances of the 1-RC equivalent electrical circuit model, r] is a predefined time constant, and I is the current flowing in the cell battery BC. To determine the charge states and polarization voltages y^(^ of the M critical cells i, the following equations can be used: [Math.2] SoC^f — SoCq + É —^dt ■' u Capa [Math.3] Vkc [ k ] = ( 1 - ) ( V, [ k -1 ] - OCV ( SoC[k -1] ) - -1] ) + R'^ Then, the nominal cell is selected from the M identified critical cells. Then, the state of charge of each of the NM cells is determined remaining non-critical using the following equations: [Math.4] A SoCJ[£] = SoCj[k] - [Math.5] A SbC ; |>] = A -1] + v / [£] = where I {1, 2, NM}. We can also determine the polarization voltage k ] of each of the NM remaining non-critical cells using the following equations: [Math.6] c [k ] = ( 1 - ) ( v{[ k -1 ] - OCV(SoC[k-1]) - R^k -1]) + R^I The last equation is an approximation obtained by using the following equation with the final voltage _ 1 j:
[0094] [Math.7] v ;iC W = ( 1 ] +
[0095] It will also be noted, as illustrated non-limitingly in [Fig. 2], that the battery calculator CB (or the calculator of the estimation device DE) may also comprise a mass memory MM1, in particular for storing the measured (or estimated) parameters of the cells, as well as any intermediate data involved in all its calculations and processing. Furthermore, this battery calculator CB (or the calculator of the estimation device DE) may also comprise an input interface IE for receiving at least the measured (or estimated) parameters of the cells and the requests for estimating the state of charge SoCBC of the cellular battery BC for use in calculations or processing, possibly after having been shaped and / or demodulated and / or amplified, in a manner known per se, by means of a digital signal processor PR2.In addition, this battery calculator CB (or the calculator of the estimation device DE) can also include an output interface IS, in particular for delivering a message containing the estimation of the state of charge SoC BC of the cellular battery BC.
[0096] It will also be noted that the invention also proposes a computer program product (or computer program) comprising a set of instructions which, when executed by processing means of the electronic circuit (or hardware) type, such as for example the processor PR1, is capable of implementing the estimation method described above to estimate in the vehicle V the state of charge SoCBC of the cellular battery BC.
Claims
Claims
1. Method for estimating a state of charge of a battery (BC) comprising N cells and equipping a vehicle (V), characterized in that it comprises a step (10-40) in which i) among said N cells, M so-called critical cells are identified, then ii) for each of said M identified critical cells and according to a first sampling period, a state of charge is determined by applying to a chosen cell model a first chosen filter using measured impedance and capacity of the cell considered, then a so-called nominal cell is selected from among said M identified critical cells, then iii) for each of NM non-critical cells, a new difference in state of charge is determined, according to a second sampling period greater than said first sampling period and by applying to said chosen cell model a second chosen filter, a new difference in state of charge as a function of a previous difference in state of charge,of a current measured in said battery (BC), of a measured capacity of the non-critical cell considered and of the measured capacity of said nominal cell, then a new state of charge by summing said determined state of charge of said nominal cell and said new difference in state of charge, and iv) determining according to said second sampling period said state of charge of the battery (BC) by exploiting said new determined states of charge of the NM non-critical cells and said determined states of charge of the M critical cells.,
2. Method according to claim 1, characterized in that in said step (10-40) said first selected filter is a two-state extended Kalman filter, a first state being the charge state of the critical cell considered and a second state being a bias voltage of the critical cell considered.
3. Method according to claim 1 or 2, characterized in that in said step (10-40) said second selected filter is a Kalman filter extended to a state, this state being the state of charge of the non-critical cell considered.
4. Method according to one of claims 1 to 3, characterized in that in said step (10-40) said chosen cell model is an n-RC equivalent electrical circuit model, with n > 1.
5. Method according to one of claims 1 to 4, characterized in that in said step (10-40) a second sampling period is used. equal to K times said first sampling period, where K is a number of groups into which said NM non-critical cells are distributed, and the new charge states of the non-critical cells of one of said K groups are determined during each first sampling period so as to have said NM new charge states of said NM non-critical cells at the end of each second sampling period.
6. Method according to one of claims 1 to 5, characterized in that in said step (10-40) said M critical cells are selected by determining among said N cells those which have the largest M / 2 upper operating limits and those which have the smallest M / 2 lower operating limits with respect to predefined operating limits.
7. Method according to claim 6, characterized in that in said step (10-40) one determines among said M / 2 critical cells having said M / 2 largest upper operating limits and among said M / 2 critical cells having said M / 2 smallest lower operating limits the one which is the most critical and which becomes said nominal cell.
8. Computer program product comprising a set of instructions which, when executed by processing means, is capable of implementing the estimation method according to one of claims 1 to 7, in a vehicle (V) comprising a battery (BC) comprising N cells, to estimate a state of charge of said battery (BC).
9. Estimation device (DE) for estimating a state of charge of a battery (BC) comprising N cells and equipping a vehicle (V), characterized in that it comprises at least one processor (PR1) and at least one memory (MD) arranged to carry out the operations consisting of i) identifying among said N cells M so-called critical cells, then ii) determining for each of said M identified critical cells and according to a first sampling period, a state of charge by applying to a chosen cell model a first chosen filter using measured impedance and capacity of the cell considered, then selecting a so-called nominal cell among said M identified critical cells, then iii) for each of NM non-critical cells, determining, according to a second sampling period greater than said first sampling period and by applying to said chosen cell model a second filter chosen, a new state of charge difference depending on a previous state of charge difference, a current measured in said battery (BC), a measured capacity of the non-critical cell considered and the measured capacity of said nominal cell, then a new state of charge by summing said determined state of charge of said nominal cell and said new state of charge difference, and iv) determining according to said second sampling period said state of charge of the battery (BC) by exploiting said new determined states of charge of the NM non-critical cells and said determined states of charge of the M critical cells.
10. Vehicle (V) comprising a battery (BC) comprising N cells, characterized in that it further comprises an estimation device (DE) according to claim 9.
