Monitoring a battery block

Abstract

The invention relates to a battery block for monitoring a battery. A system and method for monitoring the behavior of battery blocks (b1,..., bn) connected in series. For each of a plurality of points in time (t0, t-1, t-2), and for each battery block, a deviation value ([Delta] Qb1,..., [Delta] Qbn), which is a voltage deviation of the voltage measurement with respect to an average (M) of the voltage measurements (Ub1,..., Ubn) of the battery block, or a corresponding charge deviation, is calculated. The deviation values ([delta] Qb1,..., [delta] Qbn) of the battery blocks are aggregated. The aggregation deviation value (10) is evaluated relative to a reference aggregation deviation value (12) to detect a potential abnormal state of the battery block. The reference aggregation deviation value (12) is obtained based on a corresponding aggregation deviation value (10) of a corresponding battery block of a corresponding battery (14).

Description

[0001] This invention relates to monitoring the behavior of a plurality of battery blocks connected in series in a battery, for detecting an abnormal state in one of the plurality of battery blocks. Invention Overview

[0002] Batteries, especially electrochemical batteries (such as lithium-ion batteries), are increasingly being used for battery storage. The aim is to provide economical, safe, and sustainable operation of battery storage devices.

[0003] The purpose of this invention is to provide a novel method for monitoring the behavior of multiple battery blocks connected in series in a battery.

[0004] The desired outcome is that this method allows for monitoring of battery cell behavior based solely on electrical parameters readily available in the battery management system of a battery storage device.

[0005] The desired outcome is that this method allows for the establishment of non-invasive monitoring of battery block behavior.

[0006] Specifically, it is desirable that this method allows for monitoring of the battery cell behavior based solely on voltage measurements taken at corresponding time points.

[0007] In addition, it is desirable to be able to automatically detect abnormal conditions in one of multiple battery cells, especially potentially safety-related conditions and / or conditions that reflect manufacturing defects.

[0008] Furthermore, it is expected that this method can detect abnormal states of battery blocks without relying on ideal or perfectly normal batteries as a reference.

[0009] The invention is specified in the independent claims. Further embodiments are specified in the dependent claims.

[0010] According to one aspect of the present invention, a method is provided for monitoring the behavior of a plurality of battery blocks connected in series in a battery, for detecting a potential abnormal state of one of the plurality of battery blocks. The method includes: for each of a plurality of time points, providing a voltage measurement result of each of the plurality of battery blocks measured at that time point. The method further includes: for each of the plurality of time points, and for each of the plurality of battery blocks, calculating a deviation value for the battery block. The deviation value is a voltage deviation of the voltage measurement result of the battery block relative to the average of the voltage measurement results of the plurality of battery blocks (at that time point), or a charge deviation of the battery block corresponding to the voltage deviation. The method further includes: for each of the plurality of battery blocks and for a first time frame, aggregating the deviation values ​​of the battery blocks calculated for time points belonging to the first time frame to determine an aggregated deviation value. That is, the aggregated deviation value is determined by aggregating the deviation values ​​of the battery blocks belonging to the time points of the first time frame. The method further includes: for each of the plurality of battery blocks, evaluating the aggregated deviation value to detect a potential abnormal state of the battery block. The evaluation includes: evaluating the aggregated deviation value relative to at least one reference aggregated deviation value. The reference polymerization deviation value can be obtained based on the corresponding polymerization deviation value of the corresponding cell block of the corresponding cell.

[0011] A battery can also be referred to as a "monitored battery." During the battery's lifespan, battery cells connected in series typically undergo the same series of charging, energy storage, and discharging cycles. Furthermore, battery cells often have similar structures and ages, and similar electrical parameters (such as capacity). When calculating the deviation value of a battery cell, the average of the voltage measurements of the battery cells is used as a "self-generated" reference on a local basis. The voltage measurements are derived from battery cells subjected to the same current at a given point in time, because the battery cells are connected in series. When performed individually, the difference between the voltage measurement of a single battery cell and the average voltage measurement of the battery cells may not show any significant characteristics. However, by aggregating the deviation values ​​and evaluating them relative to at least one reference aggregated deviation value already obtained based on the corresponding corresponding battery cell, the effects of different states of charge are balanced within the first time frame. Furthermore, seasonal effects and effects due to the battery's construction type are considered equivalent, as these effects are expected to affect the monitored battery and the corresponding battery in a similar manner. Therefore, based on voltage measurement results, and using the average of the voltage measurement results of the battery pack as a "self-generated" reference, and using another "self-generated" external reference in the form of a reference aggregated deviation value derived from the corresponding battery, a potential abnormal state of one of the series-connected battery packs can be detected.

[0012] The corresponding battery can be a battery that is structurally identical to the battery being monitored, or a battery of a similar type to the battery being monitored. The corresponding battery includes batteries other than the (monitored) battery that provides voltage measurement results.

[0013] The corresponding aggregated deviation value for the corresponding cell block of the corresponding battery is obtained in the same manner as the aggregated deviation value determined for the corresponding block among multiple cell blocks of the monitored battery. In other words, the corresponding aggregated deviation value can be determined in a manner corresponding to the steps of calculating the deviation value and aggregating the deviation values ​​of the corresponding cell blocks of the monitored battery. Similarly, the deviation value of the cell block of the corresponding battery is obtained in the same manner as the deviation value of the corresponding block among multiple cell blocks of the monitored battery. The deviation value of the cell block of the corresponding battery can also be referred to as the corresponding deviation value.

[0014] Therefore, at least one reference aggregate deviation value can be obtained based on corresponding measurements of multiple batteries of similar types (“corresponding batteries”), particularly batteries that are structurally identical to the monitored battery. This is particularly advantageous because, for example, the behavior of a specific battery block (e.g., battery block #n in a series-connected battery block) can be evaluated relative to the behavior of battery block #n of multiple structurally identical batteries in the field. Although battery block #n may have a specific behavior, this specific behavior may be due to structural differences between series-connected battery blocks and may appear in a similar manner in multiple structurally identical batteries. However, when the behavior of battery block #n differs from the behavior of battery block #n of multiple structurally identical batteries, an anomalous state can be detected. Thus, the detection of potential anomalous states of a corresponding battery block in a battery block is significantly improved.

[0015] The corresponding aggregate deviation value for the corresponding battery block of the corresponding battery can be obtained for the corresponding time frame corresponding to the corresponding first time frame or, where applicable, the second time frame. In particular, two time frames that correspond to each other can have the same duration and / or can overlap for most of their time.

[0016] The term "battery block" is understood to refer to a unit within a battery or battery storage device, where unit voltage measurements are available. A battery block may include one or more battery units. In particular, a battery block may include multiple battery units connected in parallel. The term "battery block" may also include a single battery unit.

[0017] The battery may include battery modules or battery strings, wherein the plurality of battery modules are connected in series. The battery may be part of a battery storage device. The battery storage device may include a battery management system.

[0018] The term "voltage measurement result" is understood to mean a value obtained by measuring voltage. Providing a voltage measurement result can include measuring voltage to obtain the voltage measurement result. For example, a voltage measurement result can be provided by a battery management system. The voltage measurement result is measured at a corresponding point in time. For example, the voltage measurement may be performed at a specific point in time or may have already occurred at that point in time.

[0019] The calculated deviation value is correlated with the corresponding voltage measurement time point. In other words, the deviation value "appears" at that time point, or at the voltage measurement result taken at the same time point.

[0020] The calculated deviation value is a signed deviation value; that is, the sign of the voltage deviation relative to the average value of the voltage measurements is retained. Where applicable, the sign of the charge deviation corresponds to the sign of the voltage deviation.

[0021] Charge deviation can be an estimated charge deviation. It can be estimated based on voltage deviation. For example, the charge deviation of a corresponding battery cell can be estimated based on the voltage deviation of the battery cell and on voltage-charge characteristics. Voltage-charge characteristics can be the voltage-charge characteristics of a battery cell or multiple battery cells. Charge deviation can be calculated based on the voltage-charge characteristics of a battery cell or multiple battery cells and on voltage measurements of the battery cell. Voltage-charge characteristics can define the relationship between voltage and charge at different voltages. Voltage-charge characteristics can be ideal or estimated. Voltage-charge characteristics can be voltage-charge curves, particularly voltage-SoC curves that describe the relationship between voltage and state of charge (SoC). For example, at a specific voltage (voltage measurement result), the slope or local behavior of the voltage-charge curve describes the relationship between voltage deviation and corresponding charge deviation at that specific voltage.

[0022] The first time frame can have a predetermined length or duration. The length of the first time frame can be at least 14 days, preferably at least 20 days, and more preferably at least 28 days. Therefore, voltage measurements at many different charge states of the battery can typically be included in the voltage measurement results. The length of the first time frame can be less than 2 months, preferably less than 1 month. This allows for the detection of the development of abnormal states in the battery pack on a suitable timescale.

[0023] The number of time points belonging to the first time frame can be at least 100,000, preferably at least 400,000.

[0024] Multiple time points can be sequentially spaced, for example, at intervals of one second or longer. The time interval can be at least one second or at least five seconds. The time interval can be one hour or less, 10 minutes or less, 5 minutes or less, or preferably 60 seconds or less.

[0025] In this embodiment, the average value of the voltage measurements of the multiple battery cells is the median value of the voltage measurements of the multiple battery cells.

[0026] However, the average value can also be the arithmetic mean or other central tendency of the voltage measurements of the battery cells at the corresponding time points.

[0027] The aggregate deviation value can also be referred to as the total deviation value.

[0028] In one embodiment, the aggregation includes calculating the sum of the absolute values ​​of the deviation values ​​of the corresponding battery blocks calculated for time points belonging to the first time frame. In other words, the aggregated deviation value is the sum of the absolute values ​​of the deviation values ​​of the corresponding battery blocks calculated for time points belonging to the first time frame. Alternatively, the aggregated deviation value may be calculated as the sum of the squares of the absolute values ​​of the deviation values ​​of the corresponding battery blocks calculated for time points belonging to the first time frame.

[0029] In an embodiment, evaluating the aggregation deviation value with respect to at least one reference aggregation deviation value includes: scaling and / or adjusting the determined aggregation deviation value based on at least one reference aggregation deviation value, or includes comparing the determined aggregation deviation value with at least one of the at least one reference aggregation deviation value.

[0030] The method may include detecting a potential abnormal state of the battery block when a scaled and / or adjusted determined aggregation deviation value is greater than a threshold. That is, the battery block is detected to have a potential abnormal state.

[0031] The method may include detecting a potential abnormal state of the battery block when a scaled and / or adjusted determined aggregation deviation value is greater than an upper threshold or less than a lower threshold. That is, a potential abnormal state of the battery block is detected.

[0032] Adjusting the determined polymerization deviation value may include subtracting a reference polymerization deviation value, for example, subtracting the median of the corresponding polymerization deviation values ​​for the corresponding battery block of the corresponding battery.

[0033] Scaling the determined aggregation deviation value (e.g., scaling the adjusted determined aggregation deviation value) may include dividing the aggregation deviation value by a reference aggregation deviation value, or dividing the aggregation deviation value by the difference between the upper limit reference aggregation deviation value (or the first upper limit threshold) and the lower limit reference aggregation deviation value (or the first lower limit threshold).

[0034] The method may include: detecting a potential abnormal state of the battery block when the determined polymerization deviation value is greater than a reference polymerization deviation value. That is, detecting that the battery block has a potential abnormal state.

[0035] At least one reference aggregation deviation value may be at least one first threshold.

[0036] At least one reference polymerization deviation value may be a (statistical) representation of the polymerization deviation value of the corresponding battery.

[0037] In particular, at least one reference aggregation deviation value may be a statistical representation of the aggregation deviation value of the corresponding battery block of the corresponding battery.

[0038] For example, at least one of the reference aggregation deviation values ​​(e.g., the upper limit aggregation reference deviation value) can be the n-th q-quantile of the corresponding aggregation deviation value of the corresponding cell block of the corresponding cell, where n / q is at least 0.5.

[0039] For example, at least one of the reference aggregation deviation values ​​(e.g., the lower limit aggregation reference deviation value) can be the (qn)th q quantile of the corresponding aggregation deviation value of the corresponding battery block of the corresponding battery.

[0040] Preferably, n / q is at least 0.75, at least 0.8, or at least 0.9. For example, n / q can be less than 0.95. For example, n / q can be 0.9.

[0041] Therefore, for a battery block, the aggregation deviation value is evaluated only by considering those corresponding aggregation deviation values ​​of the corresponding battery block for the corresponding battery, which are above (and / or below) a corresponding threshold that truncates the lowest (and / or correspondingly, the highest) aggregation deviation value for the corresponding aggregation deviation value. This improves the robustness of the evaluation.

[0042] The term "q-quantile" refers to the q-quantile of the distribution of the corresponding aggregate deviation values ​​for the corresponding battery cell. If q = 100, then the nth q-quantile is the nth percentile.

[0043] In an embodiment, the step of aggregating deviation values ​​and evaluating the aggregated deviation values ​​(to detect potential abnormal states of the battery block) is repeated for at least one second time frame. The evaluation may include detecting a potential abnormal state of the battery block when the deviation of a corresponding determined aggregated deviation value relative to at least one of corresponding at least one reference aggregated deviation values ​​crosses (exceeds or falls below) a threshold within a predetermined duration of the monitoring time.

[0044] In an embodiment, evaluating the aggregate deviation value (to detect potential abnormal states of the battery block) further includes: evaluating the ratio of the square of the aggregate deviation value of the battery block to the sum of the squares of the aggregate deviation values ​​of the other battery blocks (connected in series in the battery). Therefore, calculating this ratio is similar to calculating the signal-to-noise ratio, where the "signal" is the aggregate deviation value of the battery block being evaluated, and the "noise" corresponds to the aggregate deviation values ​​of the other battery blocks. In an embodiment, the ratio of evaluating the square of the aggregate deviation value may be the ratio of evaluating the square of a scaled and / or adjusted aggregate deviation value.

[0045] Evaluating this ratio allows for better detection of anomalous states in specific situations. For example, a cell in a cluster may have a polymerization deviation value that appears suspicious when evaluated relative to its corresponding cell in the field (i.e., relative to at least one reference polymerization deviation value). However, this could be due to variations between cells caused by the cell construction process, and the monitored cell may still be in a normal, regular state. By evaluating this ratio, the polymerization deviation value is compared to the polymerization deviation values ​​of other cells with the same (monitored) cells. Therefore, such variations can be excluded from being detected as anomalous states. Thus, the effectiveness of the evaluation can be improved.

[0046] In one embodiment, the evaluation ratio includes comparing the ratio with a second threshold.

[0047] The method may include: detecting a potential abnormal state of the battery block when the ratio is greater than a second threshold. That is, detecting that the battery block has a potential abnormal state.

[0048] In an embodiment, the step of aggregating deviation values ​​and evaluating the ratio is repeated for at least one second time frame, wherein the evaluation includes detecting a potential abnormal state of the battery block when the corresponding ratio crosses (exceeds and / or correspondingly falls below) a third threshold within a predetermined duration of the monitoring time.

[0049] In an embodiment, the step of aggregating deviation values ​​and evaluating the ratio is repeated for at least one second time frame, wherein the evaluation includes: detecting a positive deviation of the corresponding ratio relative to a fourth threshold; and determining whether the positive deviation increases over the monitoring time.

[0050] For example, the increase in monitoring time can be determined by, for instance, the steps of aggregating deviation values ​​and evaluating the ratio for at least one second time frame (e.g., for multiple new second time frames), each second time frame advancing in time relative to a previous time frame; detecting positive deviations for each time frame; and determining the increase in deviation, for example, based on linear regression of the detected deviations. The increase in deviation can also be determined by other algorithms.

[0051] In this document, the increase (or decrease) in monitoring time can include an increase (or decrease) in time in the direction toward the latest time frame (end) (in particular from the first time frame toward at least one second time frame (end)).

[0052] In the above embodiments where the steps of aggregating the deviation value and evaluating the aggregation deviation value and / or evaluating the ratio are repeated for at least one second time frame, one or more of the following features may be applied: The steps of aggregating the deviation value and evaluating the aggregated deviation value and / or the evaluation ratio can be repeated for multiple new second time frames, each second time frame advancing in time relative to the previous time frame.

[0053] The second time frame can be an updated time frame. For example, a new second time frame can be processed daily. Each second time frame can overlap with a previous time frame or the first time frame via time overlap. Each first time frame and the second time frame can have the same length or duration.

[0054] The monitoring time can be extended over a period of time longer than the corresponding time frame.

[0055] The steps of aggregating and evaluating the aggregated deviation value can be repeated for new / updated time frames, for example, daily, or after several days, preferably at least 1 and / or preferably less than 8 days. Each new time frame can overlap with a previous time frame. For example, each new time frame can overlap with a previous time frame by at least one day or the aforementioned number of days.

[0056] In any of the foregoing embodiments, the method may include one or more of the following features: This method can be implemented by a computer.

[0057] The method may include: collecting at least one corresponding deviation value for a corresponding cell block specific to the corresponding cell from the corresponding cell. The method may also include determining at least one reference aggregation deviation value based on the corresponding deviation value for the corresponding cell block specific to the corresponding cell.

[0058] Voltage measurement results from multiple battery cells of the monitored battery can be received via a communication network. Voltage measurement results from a corresponding battery cell can also be received via the communication network. Reference aggregate deviation values ​​and / or corresponding aggregate deviation values ​​from a corresponding battery cell can be received via the communication network. The communication network can be a telecommunications network, such as the Internet or a mobile network.

[0059] In this embodiment, the potential abnormal state of the battery block is a potential safety-related state and / or a state reflecting a manufacturing defect.

[0060] This method may include detecting potential safety-related conditions and detecting conditions reflecting manufacturing defects. Specifically, safety-related conditions can be detected independently of detecting conditions reflecting manufacturing defects. In particular, a potential safety-related condition can be detected when a deviation from expected behavior, or a corresponding rate of deviation, increases with monitoring time.

[0061] In aspects of the present invention, a battery monitoring system is provided, comprising means for performing the steps of the methods described herein. In at least one embodiment, the system includes computing means for receiving and processing voltage measurement results. The system may include means for receiving and / or calculating a reference aggregate deviation value. Attached Figure Description

[0062] Preferred embodiments of the invention will now be described with reference to the accompanying drawings, in which: Figure 1 This is a schematic diagram of a method and system for monitoring the behavior of multiple battery cells connected in series in a battery; and Figure 2 This is a schematic diagram of the steps of another method for monitoring the behavior of multiple battery blocks connected in series in a battery.

[0063] Figure 1 A method and system 100 for monitoring the behavior of multiple battery blocks b1, ..., bn connected in series in a battery is schematically illustrated for detecting a potential abnormal state of one of the multiple battery blocks b1, ..., bn. The battery may be part of a battery storage device.

[0064] For multiple time points t0, t -1 t -2 For each time point in ..., provide the voltage measurement result U for each of the multiple battery blocks b1, ..., bn measured at that time point. b1 ... U bn For example, measurements can be taken every 1 to 60 seconds or every 5 minutes within a first time frame T1 of, for example, 28 days.

[0065] For the first time frame T1, perform the following operations: For multiple time points t0, t1 belonging to the first time frame T1 -1 t -2 For each time point in the series b1, ..., bn, calculate the voltage measurement result U for each of the multiple battery blocks b1, ..., bn. b1 ... U bn The median value M is used to calculate the average value of the battery cells. Additionally, the voltage deviation ΔU of the battery cells is calculated. b1 ... ΔU bnVoltage deviation value ΔU b1 ... ΔU bn This refers to the voltage measurement result U of the battery pack at the corresponding time point. b1 ... U bn The voltage measurement results U of multiple battery blocks b1, ..., bn b1 ... U bn The average voltage deviation M.

[0066] Based on voltage deviation value ΔU b1 ... ΔU bn The charge deviation value ΔQ is calculated based on the voltage-charge characteristics (e.g., using a voltage-SoC curve). b1 ... ΔQ bn The corresponding deviation value of the form.

[0067] For each of the multiple battery blocks b1, ..., bn and for the first time frame T1, the deviation value ΔQ of the battery block, which has already been calculated for the time point belonging to the first time frame T1, is... b1 ... ΔQ bn Aggregation is performed to determine the aggregation deviation value of 10. Specifically, the aggregation deviation value is calculated as follows: the deviation value ΔQ of the corresponding battery cell, which has already been calculated for the time point belonging to the first time frame T1. b1 ... ΔQ bn The sum of the absolute values ​​of .

[0068] For each of the multiple battery blocks b1, ..., bn, the aggregation deviation value 10 is scaled (indicated as 16) based on the reference aggregation deviation value 12. The reference aggregation deviation value 12 is obtained based on the corresponding aggregation deviation value of the corresponding battery block of the corresponding battery 14. The reference aggregation deviation value 12 is a statistical representation of the corresponding aggregation deviation value 10 of the corresponding battery block of the corresponding battery 14. Specifically, the reference aggregation deviation value 12 is the 90th percentile of the corresponding aggregation deviation value 10 of the corresponding battery block of the corresponding battery 14.

[0069] The (scaled) aggregate deviation value 10 is then evaluated to detect potential abnormal states of the battery block. The corresponding (scaled) aggregate deviation value 10 is compared with a threshold 18.

[0070] Specifically, the steps of repeatedly aggregating the deviation value and evaluating the aggregated deviation value 10 for multiple second time frames T2 are performed. For each second time frame T2, the end point of time frame T2 corresponds to the corresponding monitoring time t, and the corresponding aggregated deviation value 10 is assigned to that time t.

[0071] exist Figure 1The lower right section shows a graph of the (scaled) aggregate deviation value 10 obtained for each of the battery blocks b1, ..., bn within the monitoring time t.

[0072] When the positive deviation of the corresponding determined aggregation deviation value 10 relative to the corresponding reference aggregation deviation value 12 exceeds the threshold 18 within a predetermined duration of the monitoring time t, a potential abnormal state of the battery block is detected.

[0073] Figure 2 This is a schematic diagram of the steps of another method for monitoring the behavior of multiple battery cells connected in series in a battery. The first part of this method is similar to... Figure 1 The method is described below. The second part of the method is described below.

[0074] As shown in “S / N”, similar to the signal-to-noise ratio, the square of the aggregate deviation value 10 of the battery block is calculated as the ratio of the sum of the squares of the aggregate deviation values ​​of the other battery blocks in the plurality of battery blocks. Therefore, the (scaled) aggregate deviation value 10 of the battery block is related to the aggregate deviation values ​​10 of the other battery blocks in the plurality of battery blocks b1, ..., bn.

[0075] exist Figure 2 The lower right section shows a graph of the ratios obtained for each of the battery blocks b1, ..., bn within the monitoring time t.

[0076] like Figure 2 As shown in the example, one of the battery cells exhibits behavior exceeding a fourth threshold 22, indicated by a thin dashed line. Specifically, for the most recent portion of monitoring time t, this ratio shows a positive deviation relative to the fourth threshold 22. Furthermore, for the most recent portion of monitoring time t, including the current time t, the positive deviation is determined to generally increase with monitoring time t (as indicated by a thick dashed line). Based on this, a potential abnormal state of the battery cell is detected.

Claims

1. A method of monitoring the behavior of a plurality of battery blocks (bl,..., bn) connected in series in a battery for detecting a potential abnormal state of one of the plurality of battery blocks (bl,..., bn), the method comprising: For each of a plurality of time points (t0, t -1 , t -2 ,...), a voltage measurement (U b1 ,..., U bn ) of each of the plurality of battery blocks (b1,..., bn) measured at said time point is provided; for each of the plurality of time points (t0, t -1 , t -2 ,...), and for each of the plurality of battery blocks (b1,..., bn), a deviation value (AU b1 ,..., AU bn ; AQ b1 ,..., AQ bn ) of the battery block is calculated, wherein the deviation value is a voltage deviation of the voltage measurement of the battery block with respect to an average value (M) of the voltage measurements (U b1 ,..., U bn ) of the plurality of battery blocks (b1,..., bn), or a charge deviation of the battery block corresponding to the voltage deviation; for each of the plurality of battery blocks (bl,..., bn) and for a first time frame (Tl), aggregate the deviation values (AQ b1 ,..., AQ bn ) of the battery block that have been computed for time points belonging to the first time frame (Tl) to determine an aggregated deviation value (10); and for each battery block of the plurality of battery blocks (bl,..., bn), evaluating the aggregated deviation value (10) for detecting a potential abnormal state of the battery block, the evaluation comprising: evaluating the aggregated deviation value (10) with respect to at least one reference aggregated deviation value (12), wherein the at least one reference aggregated deviation value (12) is obtained based on a corresponding aggregated deviation value (10) of a corresponding battery block of a corresponding battery (14).

2. The method of claim 1, wherein, the average value (M) of the voltage measurement results (U b1 ,..., U bn ) of the plurality of battery blocks (b1,..., bn) is a median value of the voltage measurement results of the plurality of battery blocks (b1,..., bn).

3. The method according to any of the preceding claims, wherein, The aggregation comprises calculating the sum of the absolute values of the deviation values (AQ b1 ,..., AQ bn ) of the respective battery block that have been calculated for the time points belonging to the first time frame (T1).

4. The method according to any of the preceding claims, wherein, the evaluating the aggregated deviation value (10) with respect to at least one reference aggregated deviation value (12) comprises: scaling (16) and / or adjusting the determined aggregated deviation value (10) based on the at least one reference aggregated deviation value (12), or comprises: comparing the determined aggregated deviation value (10) with at least one of the at least one reference aggregated deviation value (12).

5. The method according to any of the preceding claims, wherein, the at least one reference aggregated deviation value (12) is a statistical representation of the corresponding aggregated deviation value (10) of the corresponding battery block of the corresponding battery (14).

6. The method according to any one of the preceding claims, wherein, at least one of the at least one reference aggregated deviation value (12) is an n-th q-quantile of the corresponding aggregated deviation value (10) of the corresponding battery block of the corresponding battery (14), wherein n / q is at least 0.

75.

7. The method according to any one of the preceding claims, wherein, at least one of the at least one reference aggregated deviation value (12) is an n-th q-quantile of the corresponding aggregated deviation value (10) of the corresponding battery block of the corresponding battery (14), wherein n / q is at least 0.

75.

8. The method of any of the preceding claims, wherein, the charge deviation of a respective battery block is estimated based on the voltage deviation of the battery block and based on a voltage-charge characteristic.

9. The method according to any one of the preceding claims, wherein, the steps of aggregating the deviation values and evaluating the aggregated deviation value (10) for detecting a potential abnormal state of the battery block are repeated for at least one second time frame (T2), wherein the evaluation comprises: detecting a potential abnormal state of the battery block when a deviation of a respective determined aggregated deviation value (10) with respect to at least one of the respective at least one reference aggregated deviation value (12) crosses a threshold value (18) within a predetermined duration of a monitoring time (t).

10. The method of any of the preceding claims, wherein, the evaluating the aggregated deviation value (10) for detecting a potential abnormal state of the battery block further comprises: evaluating a ratio of a square of the aggregated deviation value of the battery block to a sum of squares of aggregated deviation values of other battery blocks of the plurality of battery blocks.

11. The method of claim 10, wherein, the steps of aggregating the deviation values and evaluating the ratio are repeated for at least one second time frame (T2), wherein the evaluation comprises: detecting a potential abnormal state of the battery block when a respective ratio crosses a third threshold value (22) within a predetermined duration of a monitoring time (t).

12. The method of claim 10 or 11, wherein, The steps of aggregating the deviation values and evaluating the ratios are repeated for at least one second time frame (T2), wherein the evaluating comprises detecting a positive deviation of a respective ratio with respect to a fourth threshold value (22) and determining whether the positive deviation increases over the monitoring time.

13. A battery monitoring system (100), the system comprising means for performing the steps of the method according to any one of claims 1 to 12.

Images