FR3123156A1 - BATTERY MANAGEMENT SYSTEM COMPRISING MEANS FOR DETECTING LIMITING CELLS, VEHICLE AND METHOD BASED ON SUCH A SYSTEM - Google Patents

Abstract

The invention relates to a battery management system (BMS) for a battery pack (B) comprising cells (CB) connected in series, the system (BMS) being configured to measure cell voltages separately, characterized in that it comprises a model (M) for determining at least one limiting cell (CL1, CL0) from operating parameters of the battery cells (DCH, RGN, PAU, CHG, Ui, SOC, T), and in this that it is configured to identify at least one limiting cell (CL1, CL0). In particular, identification is made on the basis of DC capacitance and resistance. The invention further relates to a motor vehicle and a battery management method based on such a system. Fig. 1

Description

BATTERY MANAGEMENT SYSTEM COMPRISING MEANS FOR DETECTING LIMITING CELLS, VEHICLE AND METHOD BASED ON SUCH A SYSTEM

The present invention relates to the field of motor vehicle battery management system software functions (or BMS: for “Battery Management System” in English). The invention also relates to motor vehicle traction battery diagnostic means.

A battery function management system and its security allow optimal use of the motor vehicle by limiting failures or other risks for the user.

Battery diagnosis is an important element in optimizing its use as accurately as possible, reducing maintenance requirements and prematurely preventing a major malfunction.

The solutions proposed in the prior art are generally of three types:
- an estimate of the capacity of each cell with complex algorithms (for example of the Kalman filter type in particular);
- a method based on a durability model per cell;
- an estimation of the limitations of the battery pack only with the minimum/maximum voltage of all the cells without any notion of physical continuity.

These solutions are not satisfactory because they are complex, depend on relaxation frequencies, and on the accuracy of a current sensor. In addition, these solutions are costly in terms of memory backup.

In addition, they generate a loss of continuity of the physical representativeness of the limiting cell.

Thus, an objective of the present invention is to propose a simplified battery management optimization solution.

To achieve this objective, the invention proposes a battery management system for a battery pack comprising cells connected in series, the system being configured to measure cell voltages separately, characterized in that it comprises a model for determining at least one limiting cell from operating parameters of the battery cells, and in that it is configured to identify at least one limiting cell.

Advantageously, the system makes it possible to identify limiting cells and to optimize battery management accordingly.

According to a variant, the determination model determines capacity indices and direct current resistance indices from operating parameters of the battery cells. This allows for easily comparable objective values.

According to a variant, the determination model is based on the formula:

where

is the voltage across a cell;

is the open circuit voltage at the terminals of said cell;

is the DC resistance across said cell;

is the current across said cell.

The model is therefore precise and simplified by taking into account parameters that are easy to evaluate.

According to a variant, the determination model determines occurrences during which at least one cell has the minimum voltage during a rolling phase.

According to a variant, said limiting cell is a cell having the lowest capacitance or the highest direct current resistance. This is an easy means of determination that can use a known architecture.

According to a variant, said parameters comprise a rolling phase with discharging current, a rolling phase with regeneration, a pause phase, a loading phase, a voltage index, a state of charge, a temperature. This makes it possible to take into account said parameters in order to have an accurate identification of said limiting cell.

According to a variant, the battery management system is configured to dedicate specific monitoring to said limiting cell.

According to a variant, said tracking comprises a state of charge or state of health tracking having more calculation resources dedicated to said limiting cell.

The invention further relates to a motor vehicle comprising a battery management system according to the invention.

Another object of the invention relates to a battery management method for a battery pack comprising cells connected in series, the method comprising steps for
- identifying at least one limiting cell from operating parameters of the battery cells, and
- dedicate specific monitoring to said limiting cell.

The invention will be further detailed by the description of non-limiting embodiments, and on the basis of the appended figures illustrating variants of the invention, in which:
- schematically illustrates a battery management system according to a preferred variant of the invention;
- schematically illustrates cell capacities of a battery cell pack;
- schematically illustrates DC resistances of cells of said battery cell pack;
- schematically illustrates identification of limiting cells in discharge from 100% to 10% under three different conditions;
- schematically illustrates an identification of limiting cells in discharge from 100% to 70% under said three different conditions; and
- schematically illustrates identification of limiting cells in deep discharge of less than 20% under said three different conditions.

A BMS battery management system for battery pack B is provided. Battery pack B includes battery cells CB connected together in series. The BMS system is configured to measure Ui cell voltages separately, and use this information for battery B management.

According to the invention, the management system comprises a model M for determining at least one limiting cell CL1, CL0 from operating parameters of the battery cells CB.

An exploitation of basic information, namely the individual voltages of the CB cells, which is guided by the knowledge of a macroscopic model M of lithium-ion battery voltage B.

The cross-referencing of this information on certain controlled phases of use leads to the identification of the cells having limiting quantities CL1, CL0 for pack B. We will speak in particular of limiting cells CL1, CL0.

In the preferred variant, the limiting cell CL1, CL0 is the one having a low capacitance K-, or a high direct current resistance DCR+. This determination is made from operating parameters of the battery cells CB.

Preferably, the determination model M is based on the formula:

Where

is the voltage across a cell;

is the open circuit voltage at the terminals of said cell;

is the DC resistance across said cell;

is the current across said cell.

In an alternative variant, the determination model determines occurrences during which at least one cell has the minimum tension during a rolling phase.

Thus, the management system BMS is configured to identify said limiting cell CL1, CL0.

In the preferred variant, the determination model M makes it possible to determine capacity indices K+, K− and direct current resistance indices DCR from operating parameters of the battery cells CB.

In particular, the operating parameters include a rolling phase with discharging current DCH, a rolling phase with regeneration RGN (recharging), a pause phase PAU, a loading phase CHG, a voltage index Ui, a state of charge SOC, a temperature T.

More precisely, in a variant, the invention is based on the separation of the identification of indices of cells with minimum and maximum voltages in different phases of use corresponding to some of said parameters: running with discharging current (DCH discharge), running with RGN regeneration (recharging current while driving), a pseudo-pause and PAU pause (current in absolute value close to 0) and charging in so-called "plug-in" mode CHG (i.e. "charger connected" ).

Based on these parameters, the limiting cell CL0, CL1 is identified as a cell with the lowest capacitance K- (for CL0) or the highest DC resistance DCR+ (for CL1).

Once the limiting cell CL0, CL1 has been identified, the management system can be configured to dedicate a specific monitoring to said limiting cell CL1, CL0, in particular a monitoring of the state of charge or the state of health having more resources of dedicated calculations for said limiting cell CL1, CL0.

Simple tracking of an index change and/or quantification of occurrences of minimum or maximum voltages on a deep discharge or long tap load indicates the limiting cells CL0, CL1.

Firstly, the index of the CB cell appearing in both the list of high occurrence cells for the minimum cell index in discharge DCH and the minimum cell index in regeneration (RGN) indicates the information of the capacity the lower K- on deep discharge.

Second, the index of the CB cell having the minimum voltage on a pause/pseudo-pause phase (current close to 0) would also indicate the lowest capacitance information K- on a globally discharging profile.

Third, the index of the CB cell appearing in both the list of cells with high occurrence in minimum discharge and maximum regeneration would indicate a cell with the highest DC resistance DCR+.

Fourthly, a change of index (or a strong occurrence on a histogram) of the maximum cell under load in connection (index that could be initialized to that of the maximum direct current resistance DCR+) would indicate the CL0 cell having the capacity the lower K- on a load with a long connection (always on the assumption of a pack that respects a minimum balance).

Fifthly, in the same way, a change of index of the minimum cell CL0 or a strong occurrence on the histogram at the end of the connection charge would indicate the strongest K+ capacity on a long connection charge.

In a context of continuous research for higher autonomy and the deployment of electrification to a wide range of vehicle sizes, the current B packs contain more and more elements in series, which would require more calculation means. complex if we continue to treat all the CB cells individually as in the prior art. The invention proposes a simplification of the state of charge SOC and state of health algorithms by identifying the limiting cells CL0, CL1.

The invention also allows maintenance or repair assistance with the localization of the limiting cell(s) CL0, CL1.

The invention also makes it possible to feed a mechanical/thermal design analysis on a set of vehicle fleets with the aim of improving the vehicles, and to avoid crises by anticipating detections of faults in the production of CB cells. .

For the search for the limiting cell CL0, CL1, one could follow the indices of the cells CB having the minimum and maximum voltage Ui within the pack B. The invention is based on the separation of the identification of the indices into different phases of use: driving current discharging (discharge -DCH), driving with regeneration RGN (current recharging during driving), pseudo pause and PAU pause (current in absolute value close to 0) and charging on connection (charger connected).

We can start from a state of relaxation where the battery was at sufficient rest to identify the CL0 cell with the lowest SOC state of charge. Subsequently, a change in the index during a discharge (running with a profile) would indicate either a higher DCR+ DC resistance or a lower SOC state of charge.

To identify the CL1 cell having the highest DCR+ direct current resistance due to degradation: a pseudo-pause PAU is used, i.e. a transition to zero current for a few seconds (for example a stop at a red light) followed by a current pulse to calculate the DCR resistances of each CB cell.

This calculation will be activated for example in an area around 25°C and 50% SOC to avoid the effect of temperature and state of charge SOC on the internal resistance DCR. This implies taking into account a thermal deviation and the error and/or difference in the state of charge SOC).

The DC resistance DCR can be obtained by the equation

A carrier idea of the invention consists in exploiting indices of minimum voltages in discharge (negative current) and minimum voltages in regeneration (positive current) during the rolling phase as well as those at almost zero current.

According to the analysis, the two conditions of minimum cell index (CL0) in DCH discharge and minimum cell index (CL0) in RGN regeneration will only be met when the SOC state of charge of the cell (CL0) is lower and therefore we can easily extract the lowest capacity information K- by combining this with the information in the initial state.

In addition, the index of the CL0 cell having the minimum voltage on a pause/pseudo-pause phase would also indicate a lower SOC state of charge on a globally discharging profile.

The goal is to determine the limiting cells mainly the cell with low capacity K- because it is this which would limit the end of a discharge/charge (balanced pack hypothesis).

This precise identification can be used to dedicate a monitoring of state of charge or state of health that consumes more computing resources for this cell (CL0, CL1) to be more precise.

At the start of life, the difference in capacitance would come from dispersions in CB cell production. On the other hand, throughout the lifetime of the cells within the same B pack, the CB cells do not undergo the same mechanical/thermal stresses and certain cells (CL0) will age more quickly.

The idea is to have a simple means of tracking the limiting cell CL0 that could be used for aftermarket analysis as well.

Instead of a simple index change detection, one could record the occurrences during which the CL0 cell has the minimum tension within the B pack throughout taxiing to form a classification of the CB cell indices by phase. A technique based on histograms could be considered.

The example illustrated in figures 2 to 6 concerns a pack of 24 CB cells (for ease of analysis). The analysis is done on a discharging cycle. The maximum DCR resistor spread is 25%, and the maximum capacitance spread is 6%.

The shows the capacities in Ah of all CB cells. The shows the dispersion of Resistance within pack B.

Figures 4-6 show a result of the identification of cell indices by rolling phase DCH discharge, Rgn regeneration and PAU pause/pseudo-pause. The references “Val” relate to the index values of the identified cells, and “BinEdg” relate to the positions of identified cells.

The identification can be repeated 3 times corresponding to the targeted depth of discharge window.

The concerns an entire profile discharge from 100% to 10%.

Regarding the dotted circle identified, the cell at x=16 was dominant for the minimum index in RGN regeneration because of its lower resistance. Nevertheless, we see that the difference widens with the depth of discharge with the CL0 cell at x=22.

We manage to confirm with the proposed method a correct identification of the CL0 cell with low K- capacity.

The maximum cell indices are also a source of information on limiting cells CL0, CL1:

For example, the index of the cell in both the list of minimum discharging cells and maximum regenerating cells would indicate a cell with the highest DCR resistance.

The invention allows the BMS battery management system to identify the limiting cell for the calculation of power limitations.

During a connection charge, a change of index (or a strong occurrence on the histogram) of the maximum cell in connection charge (index that could be initialized to that of the maximum DCR resistance) would indicate the CL1 cell having the strongest state of charge. By analogy to the discharge, this would indicate the lowest capacity K- on a long connection charge (still assuming a pack that respects a minimum balance).

In the same way, a change in the index of the minimum cell or a strong occurrence on the histogram at the end of the charge in connection would indicate the weakest state of charge SOC and therefore the strongest capacity K+ on a charge in connection long.

Identifying the strongest CL1 cell can help the BMS battery management system target cells to balance.

As a variant, the identification can be made on a batch of limiting cells and not on a single limiting cell. One could imagine that a batch of cells would be identified in the event of the appearance of more than one index which meets the criteria for defining the minimum capacity K- / maximum resistance DCR+.

Alternatively, one could consider offline learning via an artificial neural network, for example by varying the internal resistance dispersion DCR and the capacitance K on different cells CB of a battery pack B; play on different temperature conditions T, initial SOC charge states on driving and recharging in connection and identify a model M which outputs the limiting cells CL0, CL1.

Alternatively, several production aging and dispersion scenarios could be used for training.

The invention also relates to a motor vehicle V comprising a battery management system as described above.

Another object of the invention relates to a battery management method comprising steps for
- identifying at least one limiting cell CL1, CL0 from operating parameters of the battery cells CB, and
- dedicate specific monitoring to said limiting cell CL1, CL0.

More generally, the method can include any action of any element of a battery management system BMS as described above or of a battery B as described above.

Claims (9)

Battery management system (BMS) for a battery pack (B) comprising cells (CB) connected in series, the system (BMS) being configured to measure cell voltages separately,
characterized in that it comprises a model (M) for determining at least one limiting cell (CL1, CL0) from operating parameters of the battery cells (DCH, RGN, PAU, CHG, Ui, SOC, T ), and in that it is configured to identify at least one limiting cell (CL1, CL0). Battery management system (BMS) according to Claim 1, characterized in that the determination model determines capacity indices and direct current resistance (DCR) indices from operating parameters of the battery cells. Battery management system (BMS) according to Claim 2, characterized in that the determination model is based on the formula:

Where
is the voltage across a cell;
is the open circuit voltage at the terminals of said cell;
is the DC resistance across said cell;
is the current across said cell. Battery management system (BMS) according to Claim 1, characterized in that the determination model determines occurrences during which at least one cell has the minimum voltage during a driving phase. Battery management system (BMS) according to any one of claims 1 to 4, characterized in that said limiting cell is a cell having the lowest capacity (K) or the highest direct current resistance (DCR) . Battery management system (BMS) according to any one of Claims 1 to 3, characterized in that the said parameters comprise a driving phase with discharging current (DCH), a driving phase with regeneration (RGN), a pause (PAU), a charging phase (CHG), a voltage index (Ui), a state of charge (SOC), a temperature (T). Battery management system (BMS) according to any one of claims 1 to 6, characterized in that it is configured to dedicate specific monitoring to said limiting cell (CL1, CL0). Battery management system (BMS) according to claim 7, characterized in that said monitoring comprises a state of charge or state of health monitoring having more calculation resources dedicated to said limiting cell (CL1, CL0). Motor vehicle (V) comprising a battery management system according to any one of Claims 1 to 8.