Advanced battery architecture

Abstract

The present invention relates to a battery with NxM electrochemical cells, comprising M branches connected in parallel, and, in each branch, N cells connected in series and of respective ranks 1, 2, …, N, characterized in that it further comprises: - N-1 equipotential lines, each line connecting cells of the same rank across all M branches, - NxM switches, each switch being connected in series with a cell between two equipotential lines, and - a unit of measurement (UM) for the operation of each cell, capable of controlling each switch to short-circuit a cell whose measured operation indicates a cell failure. Abstract figure: Figure 5

Description

Title of the invention: Battery with improved architecture Technical field

[0001] The present disclosure relates to an electrochemical cell type battery. Prior art

[0002] An electrochemical cell is composed of two electrodes immersed in an electrolyte (a substance that can exist in the form of mobile ions). The electrodes, anode and cathode, are electronic conductors. Electrochemical equilibrium at each of the electrode-electrolyte interfaces is reached for a particular voltage value (called the "electrode potential"). At equilibrium, the potential difference between the two electrodes defines the "nominal voltage" of the cell.

[0003] With reference to [fig. 1], when charging or discharging the electrochemical cell, the electrode potential U varies. Relaxation times are necessary to return to chemical equilibrium. The potential of the accumulator then depends on its state of charge and the temperature, for example. The electrochemical couple is chosen so that the reaction is reversible.

[0004] However, over time, aging processes tend to alter the energy efficiency of the accumulator (deformation of the crystal lattice, auxiliary chemical reactions, etc.).

[0005] The energy that an electrochemical cell is capable of restoring at a time t of its life during a complete discharge is the capacity of the cell, expressed in Wh (unit homogeneous to an energy) or in Ah (if we divide the energy by the nominal voltage). The capacity of a cell can depend on the current regime, but in any case, the capacity decreases with the age of the cell (typically the number of charges / discharges).

[0006] The voltage values ​​of elementary electrochemical cells are low (generally less than 4 V). In order to reach industrially usable voltage levels (12 V, 48 V, or others) several cells are placed in series. This implies that the same current Iseries passes through all the cells in series as illustrated in [fig.2]. The voltage that can be globally reached Vbat is then the sum of the voltages Vcell at the terminals of each cell.

[0007] However, the charge levels of the different cells may differ, for example due to initial manufacturing differences, or due to uneven aging of the cells, an imbalance in the circuit, or other reasons.

[0008] In this case, the same current value imposed by a load on several cells placed in series results in different voltage values ​​depending on the level of charge and health status of each cell. The total voltage of the cell series can deviate greatly from the expected theoretical voltage.

[0009] This imbalance limits the useful capacity of the cells and degrades the overall capacity of the battery. In addition, if the imbalance between the charge levels of the cells is significant, the least charged cells risk suffering local overcharges, which degrades the health of the battery, and can lead to its destruction with risks of thermal runaway or explosion depending on the technologies.

[0010] In order to guarantee a certain quantity of stored energy, several series of electrochemical cells are connected in parallel as illustrated in [fig.3]. This type of battery then comprises M branches B1 to BM and, in each branch, N cells of respective ranks RI, R2, ..., RN.

[0011] Thus, the series share the same voltage and the series with the lowest voltage then imposes its voltage on all the series in parallel and therefore on the battery.

[0012] The battery discharge current is composed of the sum of the currents of the different series, with an unequal distribution between the different series depending on the value of the charge current and depending on the charge levels in each series of cells. This current imbalance aggravates the imbalance between series and even within cells of the same series. Some series can discharge into other less charged series, which risks damaging them by thermal runaway.

[0013] The most problematic case is in fact that of the beginning of thermal runaway of a cell. This is a safety issue. Depending on the technology, if a cell undergoes the beginning of thermal runaway, the other cells in the same series and in other series supply current to this runaway leading to the destruction of the cell. The destruction of the cell then spreads to the adjacent cells. In the case of Li-Ion, for example, the combustion of a cell produces its own oxidizer (which is not oxygen). The phenomenon can spread to the entire battery. In the case of aqueous electrolytes (Pb or Ni-Cd technologies), the drying of the electrolyte causes the release of hydrogen produced by one electrode. The released hydrogen is unable to recombine with the oxygen produced by the other electrode. Hydrogen is an explosive gas, whose lower explosive limit is very low (4%).

[0014] Cell assemblies in the form of conventional battery packs use an arrangement of the type illustrated in [fig.3], where the cells are arranged in series in "legs" to form the battery output voltage and several legs are connected in parallel to increase the stored energy.

[0015] Electrochemical cells come from different technologies: lead, Nickel-Cadmium, Lithium-Ion, or others.

[0016] This conventional architecture then poses several problems.

[0017] The limitation of a cell results in a limitation in current and voltage of all the cells in series in the same branch.

[0018] The limitation of a series of cells leads to the voltage limitation of all the series in parallel, which results in a limitation of the available energy and therefore in a reduction of the autonomy of the battery and its inability to meet a need for electrical power. In the event of such a battery failure, there is currently no known solution for identifying the cell(s) causing the failure, and it is therefore necessary to replace the entire battery.

[0019] Indeed, it is not currently planned to carry out the diagnosis at the level of an isolated cell to know its state. It is possible to establish a diagnosis of the complete battery but it is necessary to disconnect it and therefore lose its availability.

[0020] Furthermore, due to the runaway phenomenon, the failure of a single cell leads to the failure of the battery. This failure can also lead to risks of runaway (fire, explosion, or other) or in the best case scenario the decommissioning of the entire battery.

[0021] Even if it were possible to protect a battery with a faulty cell by seeking to disconnect in particular the cell identified as faulty, the entire branch to which this cell belongs must be disconnected (whereas cells connected in series with the faulty cell would potentially be valid and operational). Thus, a fault in one cell causes the isolation of the entire series and therefore the loss of a significant fraction of the battery capacity. Summary

[0022] The present disclosure improves the situation.

[0023] A battery with NxM electrochemical cells is proposed, comprising M branches mounted in parallel, and, in each branch, N cells mounted in series and of respective ranks 1, 2, ..., N. In particular, the battery further includes: - Nl equipotential lines, each line connecting cells of the same rank on all M branches, - NxM switches, each switch being mounted in series with a cell between two equipotential lines, and - a unit for measuring the operation of each cell, capable of controlling each switch to short-circuit a cell whose measured operation shows a cell failure.

[0024] In one embodiment, each switch is of the electromechanical type. Alternatively, each switch is of the semiconductor type. Of course, the switches of the battery can be mixed between electromechanical type switches and semiconductor type switches.

[0025] The measurement unit may comprise at least one measurement sensor per cell. This may be, for example, a temperature sensor, or an electrical parameter such as a voltage across the terminals of the cell or an intensity of current delivered, for example (or even an energy or an electrical power).

[0026] Thus, it can be provided that the measuring unit comprises at least one voltage sensor at the terminals of each cell.

[0027] However, the number of electrical parameter sensors to be provided can be advantageously limited, and, in one embodiment, the measurement unit comprises a set of only N+M sensors of an electrical parameter, this set comprising one electrical parameter sensor per branch, and one electrical parameter sensor per equipotential line.

[0028] In such an embodiment, the measuring unit is arranged to: - measure the electrical parameter for each branch and compare the measurement to an expected value for each branch, and - measure the electrical parameter for each equipotential line and compare the measurement to an expected value for each line, a gap between the measurement and the expected value both on an i-th branch and on a j-th line characterizing a failure of the cell of rank j on branch i.

[0029] The measuring unit can then be further arranged to, in the event of a deviation identified both on an i-th branch and on a j-th line, control the opening of the switch of the cell of rank j on branch i.

[0030] In one embodiment, the battery may comprise, for each cell, a control track for the switch of this cell, the control track being connected on the one hand to the switch of this cell and on the other hand to the measurement unit. The measurement unit is then arranged to open, via a corresponding track, the switch of a cell of rank j and branch i, identified as faulty.

[0031] Furthermore, the measurement unit can be arranged to carry out active balancing in the event of detection of an overload of one of the NxM cells, by controlling the opening of at least one switch of a cell of the same rank as the overloaded cell, but of a different branch, the switch of the overloaded cell remaining closed.

[0032] Such an embodiment of the measuring unit also makes it possible to use cells with different characteristics in the battery (for example of different types and / or manufacturers). Thus, in such an embodiment, the battery comprises cells that are different from each other, and the measuring unit is arranged to carry out active balancing compensating for the differences between cells, by controlling the opening of a part of the switches of the battery, the switches of a complementary part remaining closed.

[0033] In one embodiment of the battery, the measuring unit comprises a plurality of sensors capable of measuring at least one parameter specific to each of the cells, a processor and a memory, the memory storing expected values ​​of said parameters, and the processor being capable of: - compare for each cell a current measurement of the parameter of this cell with the expected value of this parameter stored in memory, - and, in the event of a discrepancy between the stored value and the current measurement for a given cell, control the opening of the switch for the given cell.

[0034] The present disclosure also relates to a computer program intended to be executed by a processor of a measuring unit of such a battery, and comprising instructions for: - obtain measurements of at least one parameter specific to each of the cells, - store in memory the expected values ​​of said parameters, - compare for each cell a current measurement of the parameter of this cell with the expected value of this parameter stored in memory, - and, in the event of a discrepancy between the stored value and the current measurement for a given cell, control the opening of the switch for the given cell.

[0035] According to another aspect, there is provided a computer program comprising instructions for implementing all or part of a method as defined herein when this program is executed by a processor. According to another aspect, there is provided a non-transitory, computer-readable recording medium on which such a program is recorded.

[0036] Thus, the present disclosure proposes a battery with better performance in terms of lifetime and operational availability. The architecture for connecting the electrochemical cells is based on a matrix circuit, equipped with individual switches for each cell in order to be able to decouple a faulty cell, and sensors to detect such a failure.

[0037] By such an implementation, it is possible to: - diagnose each cell separately while maintaining critical battery functions (total maintenance or in degraded mode), - thus making the battery less sensitive to the failure of one or more cells, - carry out active cell balancing while maintaining critical functions to increase battery life, - predict the residual lifespan of an isolated cell to anticipate its replacement, - being able to replace a faulty cell independently of the others, which avoids the problem of obsolescence over time, - thus preventing an underperforming cell from damaging other cells in the battery, - and thus prevent a cell from running away (for example if attempts to balance a cell are unsuccessful due to a permanent defect in the cell).

[0038] Furthermore, it is possible to replace a faulty cell with a different cell (which does not come, for example, from the same batch, or from the same manufacturer) because the measurement unit can control the operation of this heterogeneous cell in a way that is compatible with the other cells. It is thus possible to deal with obsolescence problems without being dependent on a particular manufacturer. Thus, the active balancing that the measurement unit can perform can allow different cells to be used in the battery.

[0039] Such an architecture makes it possible to isolate this faulty cell while awaiting its replacement while maintaining in service the battery in general but also the cells in series with the faulty cell so as to reduce the total capacity loss of the battery to the capacity loss of a single cell. Brief description of the drawings

[0040] Other characteristics, details and advantages will appear on reading the detailed description below, and on analyzing the attached drawings, in which: Fig.l

[0041] [fig.l] illustrates the charging and discharging of an electrochemical cell. Fig. 2

[0042] [fig-2] illustrates a series connection of several electrochemical cells. Fig. 3

[0043] [fig.3] illustrates a parallel connection of several series of electrochemical cells. Fig. 4

[0044] [fig-4] illustrates a battery architecture proposed in one embodiment. Fig. 5

[0045] [fig.5] illustrates an embodiment for monitoring and diagnosing cells of the battery. Fig. 6

[0046] [fig.6] illustrates an implementation of active balancing in the case of an overloaded cell. Fig. 7

[0047] [fig-7] illustrates an embodiment of a measuring unit according to an exemplary embodiment. Fig. 8

[0048] [fig.8] illustrates an example of algorithmic processing for the detection of a de cell failure. Description of the embodiments

[0049] With reference to [fig.4], a battery architecture according to an embodiment of the invention is matrix-based, as presented previously with reference to [fig.3] (with M branches in parallel and, in each branch, N cells). However, here, a switch is associated with each cell, this switch being controllable to connect the cells in series in each branch. In addition, sensors for measuring a parameter characterizing the operation of each cell (for example the voltage at the terminals of each cell) are provided for monitoring the operation of each cell, with the possibility of injecting a current for the diagnosis of each cell. Indeed, in a battery with NxM electrochemical cells, comprising M branches connected in parallel, if N cells connected in series and of respective ranks 1, 2, ..., N are provided in each branch, the architecture of [fig.4] comprises, in addition to the NxM switches, Nl equipotential lines each connecting the cells of the same rank on all of the M branches.

[0050] Thus, the injection of test current can be carried out in a branch and / or in an equipotential line to identify a faulty or aging cell for example.

[0051] In such an architecture, the NxM switches are each mounted in series with a cell between two equipotential lines.

[0052] With reference to [fig.5], the battery then comprises a unit UM for measuring the operation of each cell, capable of controlling each switch to short-circuit a cell whose measured operation shows a failure of the cell (beyond a tolerance threshold for example). For this purpose, the measuring unit is connected, on the one hand, to the aforementioned sensors, and, on the other hand, to the switches, for example via electrical control tracks of each switch.

[0053] Such a measurement unit UM can apply an algorithmic processing of active balancing, security and / or diagnosis of the state of health of each cell, the instructions of which are then applied via the use of the switches.

[0054] Each of the cells of the battery can be monitored, for example continuously (by a simple measurement). In the event of detection of an anomaly, the measuring unit UM disconnects the cell in question by sending an opening order to the switch associated with this cell. Once this cell is disconnected, the measuring unit UM controls all the tests and measurements (charges / discharges) allowing the cell to be diagnosed.

[0055] The current from other cells in the same series uses the equipotentials (horizontal lines) to bypass the disconnected cell.

[0056] When the measurement unit diagnoses a faulty cell health status despite attempts at active balancing (procedure detailed later) and recovery of this cell (for example a cell detected as underperforming despite a complete recharge cycle), the measurement unit sends a command to the switch in series with this cell to disconnect the latter. This prevents the worsening of the fault and the runaway of the cell. Furthermore, the measurement unit can send and / or record a signal to warn of the failure of the cell so that it can be replaced at an appropriate time (during a future maintenance visit, or by planning exceptional maintenance if a critical number of cells are faulty).

[0057] Thus, only the fraction of energy of the faulty cell is lost, the other cells continuing to operate. The participation of the cells in series is distributed over the other cells of the same level (i.e. on the same equipotential line), as illustrated by the double arrows in [fig.5].

[0058] In the event of a detected anomaly, the measurement unit can then proceed with a procedure called "active balancing" detailed below. The active balancing mechanism is based on the same logic as that previously presented with reference to [fig.5]. However, as illustrated in [fig.6], it involves overloading a faulty cell relative to cells of the same level when the battery discharges. The measurement unit can decide to overload this cell by disconnecting the cells of the same level (same equipotential line) to lower their charge level and thus ensure balancing at the level of the cells of the same equipotential. The measurement unit can control this operation instantly or by sending timed control orders, proportional to the charge of each cell.It is also possible to arbitrate such balancing on several cells of the same equipotential line by opening only a few of the switches of this line (or even just one).

[0059] Typically, such an embodiment can also be used in the case where the cells are of different natures (having different electrical characteristics and / or from different manufacturers). In this case, the measuring unit can dynamically decouple certain cells of the battery (and connect others together) in order to balance the charges / discharges in the battery.

[0060] [Fig. 7] illustrates a possible embodiment of a measurement unit UM. The battery comprises M branches B1, B2, ..., BM and Nl equipotential lines EQ1, EQ2, ..., EQN-1. Each cell Cij is thus identified according to the branch i to which it belongs and the line EQj to which it is connected. In addition, an electrical track is provided (for example in a circuit superimposed on the plane of [Fig. 7]), this track being then connected to an interface INTI that the measurement unit UM comprises in order to control the opening / closing of each switch. The measurement unit UM further comprises an interface INTM connected to each sensor that the battery BATT comprises for monitoring each cell. The measurement unit UM further comprises a memory MEM for storing expected values ​​of the measurements carried out on the cells, as well as a processor PROC for compare these expected values ​​with the respective measurements currently being carried out on the cells. In the event of a significant deviation for at least one of these measurements, observed on a particular cell, the processor PROC can control via the INTI interface the opening of the switch of this cell to short-circuit it from its series. In addition, the memory MEM can store a computer program comprising instructions to implement such an implementation when the processor PROC executes these instructions. The measurement unit UM also includes a visualization interface VIS (for example an array of LED diodes or other) to indicate a cell identified as faulty.

[0061] Such an embodiment can use one sensor per cell, i.e. NxM sensors in total. However, alternatively, it is possible to use fewer sensors, for example only N+M sensors, as explained in an exemplary embodiment below.

[0062] With reference to figures 7 and 8, for the M branches B1 to BM (step S1), an electrical parameter of each branch Bi is tested in step S2 successively (step S2). The switches of the other branches are open. The electrical parameter can be the voltage across the terminals of the total branch (with its N cells), or the current delivered by its N cells in series, or an energy or an electrical power. The value of this measurement VAL is compared with a value stored STO in the memory MEM as an expected value (step S3). If, for a given branch Bi, the difference between the stored value and the measured value exceeds a predetermined threshold THR1 (in absolute value), then the index i of this branch Bi is stored in memory (step S4).

[0063] The same tests can then be carried out on the equipotential lines, starting for example with line EQN-1 of [fig.7]. The switches of the cells under line EQN-1 are closed and all the others open. An electrical parameter can be measured for all the cells that remain connected (in parallel) at step S6 and the measured value compared to an expected value. If the difference between these values ​​is greater than a threshold THR2 (in absolute value) at step S7, then the index of the equipotential line is stored at step S8. The switches under line EQN-1 are then opened, the switches under line EQN-2 are closed, all the other switches being open, and steps S6 to S8 are repeated, and so on until only the switches under line EQ0 of [fig.7] are closed.Thus, successively, all the equipotential lines EQN-1 to EQ0 are tested (step S5), by storing the index j of a line whose measured electrical parameter deviates from an expected value (OK arrow at the output of test S7).

[0064] In step S9, it is thus possible to identify the failure of a cell Cij, of which both the branch index i and the row index j have been stored in memory (respectively in steps S4 and S8).

[0065] Thus, the algorithmic monitoring (and possibly balancing) processing can be implemented through the use of the matrix architecture of the type presented above. The possible functionalities of this processing include: - Measuring voltage and current levels in each cell, - Measuring a charge level by applying controlled charges / discharges to isolated cells, - Estimation of load level based for example on models, - Estimation of the health status of each cell (and prediction of the residual lifespan based on models), - Calculation of control orders to be sent for monitoring, active balancing, isolation, etc., for example by maintaining measurement redundancy; however, it is possible to alternate over time the cells whose operation is measured to keep a critical number of cells connected at all times (by equipotential line for example).

[0066] It is thus possible to replace existing battery boxes by enhancing them with these functionalities, and this in different sectors of an electrical network for example. A non-exhaustive list is given for information purposes: - auxiliary power supply for HTA / LV stations for communication and reconnection of stations, - emergency power supply in electrical power generation plants, - emergency power supply in general, - batteries for electric vehicles, etc.

[0067] Typically, “lead-acid” type batteries of around forty Ah in the ITI remote control boxes (400 A switch remote control interface box) for the emergency power supply of the HTA panels of HTA / LV stations can thus be equipped with an architecture such as described above.

[0068] Nevertheless, other electrochemical cell technologies (Lead or Lithium-Ion technologies) also show functional limits, in particular problems of self-consumption and unbalancing, which can be resolved by a monitored architecture as presented above.

Claims

Claims

1. Battery with NxM electrochemical cells, comprising M branches mounted in parallel, and, in each branch, N cells mounted in series and of respective ranks 1, 2, ..., N, characterized in that it further comprises: - Nl equipotential lines, each line connecting cells of the same rank on all of the M branches, - NxM switches, each switch being mounted in series with a cell between two equipotential lines, and - a unit (UM) for measuring the operation of each cell, capable of controlling each switch to short-circuit a cell whose measured operation shows a failure of the cell, and in that the measuring unit is arranged to carry out active balancing in the event of detection of overload of one of the NxM cells, by controlling the opening of at least one switch of a cell of the same rank as the overloaded cell, but of a different branch, the switch of the overloaded cell remaining closed.

2. Battery according to claim 1, characterized in that each switch is of the electromechanical type.

3. Battery according to claim 1, characterized in that each switch is of the semiconductor type.

4. Battery according to one of the preceding claims, characterized in that the measuring unit comprises at least one measuring sensor per cell.

5. Battery according to claim 4, characterized in that the measuring unit comprises at least one voltage sensor at the terminals of each cell.

6. Battery according to one of claims 1 to 4, characterized in that the measuring unit comprises a set of N+M sensors of an electrical parameter, said set comprising one electrical parameter sensor per branch, and one electrical parameter sensor per equipotential line.

7. Battery according to claim 6, characterized in that the measuring unit is arranged to: - measure the electrical parameter for each branch and compare the measurement to an expected value for each branch, and - measure the electrical parameter for each equipotential line and compare the measurement to an expected value for each line, a gap between the measurement and the expected value both on an i-th branch and on a j-th line characterizing a failure of the cell of rank j on branch i.

8. Battery according to claim 7, characterized in that the measuring unit is further arranged to, in the event of a deviation identified both on an i-th branch and on a j-th line, control the opening of the switch of the cell of row j on branch i.

9. Battery according to one of the preceding claims, characterized in that it comprises, for each cell, a control track for the switch of this cell, the control track being connected on the one hand to the switch of this cell and on the other hand to the measurement unit, and in that the measurement unit is arranged to open, via a corresponding track, the switch of a cell of rank j and branch i, identified as faulty.

10. Battery according to one of the preceding claims, comprising cells which are different from each other, characterized in that the measuring unit is arranged to carry out active balancing compensating for the differences between cells, by controlling the opening of a part of the switches of the battery, the switches of a complementary part remaining closed.

11. Battery according to one of the preceding claims, characterized in that the measurement unit comprises a plurality of sensors capable of measuring at least one parameter specific to each of the cells, a processor (PROC) and a memory (MEM), the memory storing expected values ​​of said parameters, and the processor being capable of: - comparing for each cell a current measurement of the parameter of this cell with the expected value of this parameter stored in memory, - and, in the event of a difference between the stored value and the current measurement for a given cell, controlling the opening of the switch of the given cell.

12. Computer program comprising instructions for: - obtaining measurements of at least one parameter specific to each of the cells, - storing in memory expected values ​​of said parameters, - comparing for each cell a current measurement of the parameter of this cell with the expected value of this parameter stored in memory, - and, in the event of a difference between the stored value and the current measurement for a given cell, controlling the opening of the switch of the given cell, when the computer program is executed by a processor of a battery measuring unit according to claim 11.

Images