Method for detecting a fault state of a battery cell, detection device and motor vehicle

Abstract

The invention relates to a method for detecting a fault state (F) of at least one battery cell (16a, 16) of a battery (18) having a plurality of battery cells (16, 16a), wherein the cell voltages (Ui; UZ, UX) of the respective battery cells (16, 16a) of the plurality of battery cells (16, 16a) are detected at a specific measuring time, and a comparison value (ΔUn) is determined based on at least one of the cell voltages (Ui; UZ, UX), which is compared with a provided first reference value (R1), wherein the fault state (F) is detected based on the result of the comparison. In this case, a dispersion value (ΔUn) is determined, which represents the dispersion of at least a part (Ui, Un) of the cell voltages (Ui; UZ, UX) detected at the specific measuring time, and the fault state (F) is determined based on the dispersion value (ΔUn).

Description

Technical Field

[0001] The invention relates to a method for detecting a fault state of at least one battery cell of a battery having a plurality of battery cells, wherein a cell voltage of a corresponding battery cell of the plurality of battery cells is detected at a measuring time, and a comparison value is determined based on at least one of the cell voltages, the comparison value is compared with a provided first reference value, wherein the fault state is detected based on the result of the comparison. The invention also relates to a detection device and a motor vehicle having such a detection device. Background Art

[0002] Batteries, especially motor vehicle batteries, are known from the prior art, which can be configured as power batteries and thus as high-voltage batteries, for example. Such high-voltage batteries typically include a plurality of individual battery cells. These cells can be combined into cell modules. In the past, it was believed that all cells and cell modules could be provided with sufficient mass and could be operated in the vehicle without major mechanical or thermal influences. However, the following understanding has been obtained from various attempts and vehicle tests, namely that cell defects may also occur only during operation. In particular, cell defects may occur due to contaminants in the manufacture of cells during the construction of cell modules, such as particles in the cells, for example due to damage to the cells from the outside, or due to damage to the underbody protection of the motor vehicle when passing over an obstacle and due to damage to the cells from the outside. Therefore, it is sought to be able to identify possible cell defects, which are referred to as fault states, as early as possible in order to avoid more serious consequences, such as battery combustion.

[0003] DE 10 2020 109 133 A1 describes a deviation detection system for an energy storage system. A first group of sensors is assigned to an energy storage module having one or more energy storage devices, wherein the sensors in the first group generate current measurement results representing one or more parameters of the energy storage module. A control circuit is configured to receive the current measurement results and determine a reference value and / or reference fluctuation of a specific parameter associated with the energy storage module based at least in part on the current measurement results. Based on the current measurement results generated by the sensor, the control circuit compares the monitoring value or monitoring fluctuation of the specific parameter with the reference value or reference fluctuation of the specific parameter, and detects deviations greater than a specified tolerance range. In addition, the monitoring of the battery is decisively based on temperature measurement. In addition, other parameters such as voltage, power or current intensity can also be considered. In this case, multiple temperature sensors can be provided in the corresponding energy storage module, and the temperature sensors are assigned to the corresponding energy storage devices in the relevant energy storage module. Here, the temperature change process of each temperature sensor can be recorded and compared with the reference temperature change process. This reference temperature can be limited by a specified tolerance range at an upper limit and a lower limit. This may be static or may also be dynamic and fluctuate over time based on operating conditions, age, etc.

[0004] Although it is advantageous to dynamically determine such a tolerance range, since it allows for fluctuations that are determined by the situation, there is the problem of how such a dynamic adaptation can be implemented in detail. In addition, this is also significantly simpler to implement with respect to the detected battery temperature than, for example, when using the cell voltage as a monitoring parameter for identifying critical cell states or occurring cell defects. Cells operate reproducibly depending on environmental conditions, such as dynamic load conditions or uneven heat generation. However, there may be different characteristics and strong differences (Spreizung) between the cells in the battery. For example, cells may have slightly different capacities. This causes the voltages of the cells to drift relative to each other during discharge, for example under load. However, this drift is not the reason for assuming a cell defect. In addition, the extent of this difference is caused by aging on the one hand and also strongly depends on the operating state of the battery on the other hand. Therefore, a robust diagnosis for all cells can only be designed very roughly. In other words, for example, the limit value for the maximum permissible deviation of the cell voltage from, for example, a determined reference value must be designed to be very large in order to exclude possible misdiagnosis and avoid unnecessary warnings to the driver before a cell defect or a battery fire. However, in some cases, even small deviations in the cell voltage can already indicate a cell defect. Thus, a cell defect which initially causes only a small voltage change can remain unnoticed for a long time. Summary of the invention

[0005] It is therefore an object of the present invention to provide a method, a detection device and a motor vehicle which allow a fault state of a battery cell to be detected as reliably and early as possible.

[0006] This object is achieved by a method, a detection device and a motor vehicle having the features according to the respective independent claims. Advantageous embodiments of the invention are the subject matter of the dependent claims, the description and the drawings.

[0007] In the method according to the invention for detecting a fault state of at least one battery cell of a battery having a plurality of battery cells, a cell voltage of a corresponding battery cell of the plurality of battery cells is detected, in particular measured, at a specific measuring time, and a comparison value is determined from at least one of the cell voltages, the comparison value is compared with a provided first reference value, wherein the fault state is detected as a function of the result of the comparison. In this case, a dispersion value is determined, which represents the dispersion of at least a part of the cell voltages detected at the specific measuring time, and the fault state is determined as a function of the dispersion value.

[0008] Thus, the dispersion of the cell voltage can be advantageously taken into account when detecting a fault state. As already described at the outset, in different operating states of the battery, for example during driving, during charging or during regeneration, strongly different voltage differences of the cell voltage and thus various dispersions of these individual cell voltage values ​​can occur, so that, for example, a strong deviation of a single cell voltage value from a reference value, for example, an average value, is not itself a reason for assuming that a fault state exists. By taking into account the dispersion of the cell voltage, it is now advantageously possible to significantly better and case-dependently determine whether a fault state of a battery cell exists. To give an example, it can be assumed that in the case of a very strong dispersion of the cell voltages of the individual battery cells around an average value, a significantly larger deviation between the respective cell voltage and, for example, a reference value or an average value can be permitted before a fault state is detected. Conversely, in the case of a cell voltage having only a very small dispersion with respect to an average value, the maximum permissible deviation from this average value can be selected to be significantly smaller, so that in this case a small deviation from the average value can already lead to the detection of a fault state. Thus, a fault state can advantageously be detected reliably and early in the same way in all possible operating states of the battery. Therefore, the limit value for the maximum permissible deviation of the cell voltage can be selected in accordance with the corresponding operating conditions of the battery. Thus, a robust diagnosis of all cells is advantageously achieved. Therefore, when a cell defect or a fault state is identified, a service warning can also be displayed in time and workshop measures can be initiated. A serious cell defect may initially manifest itself as an aging-related deterioration of the battery size. Therefore, it is particularly advantageous to also consider all the remaining cells of the battery when judging that a cell has a fault state. In the case of a purely aging-induced deterioration of the battery size, attempts are often made to overcome the overall defect by balancing, whereby the remaining cells also age more strongly. When the battery parameters deteriorate based on the actual cell defect, usually only the relevant cell shows a change in its battery parameters, especially its cell voltage relative to other cells, which can now be detected accurately and early in the case of considering the dispersion related to the situation and the operating state. In addition, another advantage of the described method is that in principle, the detection of the cell voltage at a single measurement time is sufficient to detect the fault state. In other words, it is not necessary to consider the time variation curve of the cell voltage or the value related to the cell voltage, although this is still feasible.

[0009] The determined discreteness value can be used, for example, to determine a comparison value and / or a first reference value. For example, the comparison value compared with the reference value can be determined based on the discreteness value. However, it can also be provided that the first reference value is determined based on the discreteness value.

[0010] The battery is preferably a high-voltage battery for a motor vehicle. As also described at the beginning, the battery can have a plurality of battery cells, such as lithium-ion cells. These battery cells can optionally be combined into a battery module. However, in the present case, the battery can also be a battery module of such a high-voltage battery for a motor vehicle. Such a high-voltage battery can accordingly include a plurality of such battery modules. In other words, the described method can be implemented globally for all battery cells included in the high-voltage battery or individually for the corresponding battery modules of such a high-voltage battery. In addition, in order to determine whether a single cell defect exists, in particular in addition to the detected single cell voltage, other battery parameters and single cell parameters can also be considered, such as the temperature in the battery module or battery, the installation position, the state of charge, etc.

[0011] The cell voltage detected at a specific measuring time and in particular also at other specific measuring times of the respective battery cell is detected by the measurement and is provided in the form of a corresponding voltage value. Accordingly, in the following, the term voltage value is used in part instead of the cell voltage. In other words, the terms "cell voltage" and "voltage value" can be considered synonymous in the following.

[0012] In another advantageous design of the present invention, a discreteness value for a subset of the cell voltage detected at the measuring time is determined, wherein the subset does not include at least the maximum voltage value and / or the minimum voltage value of the voltage value detected at the measuring time. In other words, the discreteness value is determined only for a portion of the detected cell voltage in the defined subset. In particular, the voltage extremes are excluded from the subset. This has a great advantage that the actual dispersion of the cell voltage will not be distorted by the extreme values ​​of the defective cell. In other words, the cell voltage of the defective cell can deviate greatly from the average value of the remaining cell voltages. However, if such extreme cell voltages are considered not only when forming the average value but also when determining the discreteness value, this ultimately leads to a significantly larger discreteness value. Therefore, this negative impact of the voltage value of the defective cell can be advantageously avoided. It can be provided here that only the maximum voltage value among all voltage values ​​or only the minimum voltage value among all voltage values ​​are excluded from the subset. However, it is advantageous that at least the maximum voltage value and the minimum voltage value among all voltage values ​​detected at the same measuring time are excluded. Furthermore, not only the maximum voltage value but also two, three or usually n maximum voltage values ​​and m minimum voltage values ​​can be excluded from the subset, wherein n and m are respectively natural numbers and the sum of n and m is less than the number of battery cells included in the battery. Preferably, the sum of n and m is at most half of all battery cells. It is also preferred that n and m are of the same size. If n and m are respectively 25% of all battery cells, the maximum dispersion of the subset is exactly the interquartile range. Since at least 50% of all detected single cell voltages are included in the subset, a representative dispersion value can be determined. However, preferably, all single cell voltages that are significantly less than 50% are excluded from the subset, because it cannot be assumed that, for example, single cell defects occur simultaneously in 50% of all cells of the battery, at least not under normal circumstances where the motor vehicle has not been involved in an accident.

[0013] It is also particularly advantageous here to provide the dispersion value as the difference between the maximum voltage value and the minimum voltage value of the subset. The dispersion value thus represents the maximum dispersion amplitude of the subset in a certain way. The dispersion value can thus be determined mathematically particularly simply and quickly, which makes it particularly easy to implement the described method.

[0014] In a further very advantageous embodiment of the invention, at least one voltage extreme value representing the maximum voltage value and / or the minimum voltage value of the voltage values ​​detected at the measuring time is compared with a second reference value, wherein the comparison value represents the dispersion value and a fault state is detected if,

[0015] - if the maximum dispersion value does not exceed the first reference value, and at least one voltage extreme value differs from the second reference value by at least one first limit value;

[0016] If the dispersion value is greater than the reference value, at least one voltage extreme value differs from a second reference value by at least a second limit value, the second limit value being greater than the first limit value.

[0017] In other words, if the dispersion of the cell voltages within the subset is large, a greater deviation from the reference value is permitted for detecting a fault state than if the dispersion of the cell voltages within the subset is small. If, for example, the dispersion value determined for the subset—as described above—is, for example, below 10 millivolts, then, for example, a maximum permissible deviation from the second reference value of 20 millivolts can be specified, from which a fault state is detected. If one of the voltage values ​​differs from the second reference value by more than 20 millivolts, then, for example, a fault state can be considered to have been detected. If, for example, the dispersion value of the subset is greater than the 10 millivolts mentioned, then the second limit value can be, for example, 400 millivolts instead of the 20 millivolts mentioned above. If, for example, the cell voltage differs from the second reference value by more than 400 millivolts, then a fault state can also be considered to have been detected. On the other hand, if the deviation is less than 400 millivolts, then in this case no fault state is detected. In this way, a value suitable for the respective case can be determined for the limit value based on the dispersion value of the subset, which value specifies the maximum permissible deviation from the reference value, i.e., the second reference value. As a result, a possible fault state of a battery cell can advantageously be detected reliably and early in any desired operating situation, including dynamic ones.

[0018] It is also very advantageous if the second reference value is the average value of the voltage values ​​at least included in the subset. The reference value can also be the average value of all voltage values ​​at the time of the measurement. It is particularly advantageous if the reference value is not or is not provided solely on the basis of a model. This is because a direct comparison with the properties of other battery cells is thereby obtained. Thus, aging-related effects, which typically occur approximately the same in all cells, are also advantageously taken into account.

[0019] As already mentioned, the described method advantageously enables the detection of fault states in any operating state of the battery. Accordingly, the determined measuring time can be in an operating state of the battery and / or the motor vehicle, such as a stationary state, such as a switched-off state of the motor vehicle, when the motor vehicle or the battery is not currently being charged, or a charging state, a discharging state of the battery, such as during driving operation of the motor vehicle, wherein different driving states can also be included, such as during acceleration of the motor vehicle, braking of the motor vehicle and a regeneration phase.

[0020] In another embodiment of the invention, the standard deviation or variance for all cell voltages detected at the time of measurement is determined as a dispersion value. Therefore, in this example, the cell voltage is not excluded. The variance can be defined as the average of the squares of the differences from the mean value, and the standard deviation is the square root of the variance. Although, as mentioned above, possible distortions of this standard deviation or variance due to possibly defective cells cannot be excluded, the dispersion of the cell voltage can also be advantageously taken into account in the corresponding operating state, so that fault states can be detected more accurately.

[0021] However, if the standard deviation or variance is used as the dispersion value, it is advantageous to examine this dispersion value not only for the current state but also in the course of time. Accordingly, another very advantageous embodiment of the invention provides that the cell voltage is determined for a plurality of consecutive measuring times and that the corresponding dispersion value is determined for each measuring time, wherein the fault state is detected based on the time change of the dispersion value, in particular, when the time change of the dispersion value is determined as a comparison variable and when the comparison variable exceeds a first reference value, it is considered that the fault state has been detected.

[0022] In this way, it is advantageously possible to detect, for example, sudden changes in the standard deviation or variance over time. Such sudden changes are very strongly indicative of a possible fault state of a battery cell, which can therefore also be reliably detected.

[0023] In this case, it is also very advantageous if a plurality of successive measuring times lie within an operating time window that is assigned to a certain operating state of the battery in which the battery current of the battery changes over time less than a predetermined or predeterminable limit value. In other words, it is advantageous in this case if the variance or standard deviation is used as the dispersion value if the cell voltage is measured in an operating state of the battery in which no dynamic range is included, but the battery current is at least approximately constant or even zero. This can be the case, for example, when the motor vehicle is stationary or in a state of rest of the battery, or for example also during a charging state for charging the battery, in particular with a constant charging current or a charging current that changes only slowly or at least not abruptly over time. This can also be a suitable operating state even if the motor vehicle, for example, stops at a red light during driving operation and therefore draws from the battery for supplying electrical consumers of the motor vehicle an at least approximately constant discharge current during this phase. As a result, strong changes in the dispersion value due to dynamic conditions, such as those present during driving, acceleration or regeneration, can be excluded.

[0024] The invention also relates to a detection device for detecting a fault state of at least one battery cell of a battery having a plurality of battery cells, wherein the detection device is designed to detect a cell voltage of a corresponding battery cell of the plurality of battery cells at a certain measuring time, determine a comparison value based on at least one of the cell voltages, compare the comparison value with a provided first reference value, and detect the fault state based on a result of the comparison. In this case, the detection device is designed to determine a dispersion value (ΔUn) representing the dispersion of at least a part of the cell voltages detected at the certain measuring time, and to determine the fault state based on the dispersion value.

[0025] The advantages described for the method according to the invention and its embodiments apply in the same manner to the detection device according to the invention.

[0026] Furthermore, the invention relates to a motor vehicle having a detection device according to the invention or one of its embodiments.

[0027] Embodiments of the detection device according to the invention also belong to the invention, which have the features already described in conjunction with the embodiments of the method according to the invention. For this reason, the corresponding embodiments of the detection device according to the invention are not described again here.

[0028] The motor vehicle according to the invention is preferably designed as a car, in particular a passenger car or a truck, or as a passenger car or a motorcycle.

[0029] The present invention also includes combinations of features of the described embodiments. Therefore, the present invention also includes implementations which respectively have a combination of features of a plurality of the described embodiments, as long as these embodiments are not described as mutually exclusive. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The following describes an embodiment of the present invention. To this end, it is shown that:

[0031] Figure 1 A schematic diagram showing a motor vehicle having a detection device for detecting a fault state of a battery cell according to an embodiment of the present invention;

[0032] Figure 2 A diagram showing the voltage profile of the cell voltage of a sound battery cell with large voltage differences among all cells;

[0033] Figure 3 A schematic diagram showing a mechanically damaged battery cell of a battery;

[0034] Figure 4 Show Figure 3A diagram showing a comparison of a voltage change curve of a mechanically damaged battery cell with a voltage change curve of other battery cells;

[0035] Figure 5 A diagram showing a comparison of voltage profiles of a plurality of battery cells with a voltage profile of a defective battery cell;

[0036] Figure 6 A schematic diagram showing a plurality of voltage curves of battery cells including a defective battery cell during a balancing process;

[0037] Figure 7 Show Figure 6 Schematic diagram of the end of the equilibrium process;

[0038] Figure 8 a schematic diagram showing the voltage profile of a battery cell during a charging process with a subsequent rest phase, including the voltage profile of a defective battery cell; and

[0039] Fig. 9 A schematic diagram showing a detection device for detecting a fault state according to an exemplary embodiment of the present invention is shown. DETAILED DESCRIPTION

[0040] The embodiment set forth below is a preferred embodiment of the present invention. In the embodiment, the described parts of the embodiment are respectively the features of the present invention that can be considered independently of each other, and these features also improve the present invention independently of each other. Therefore, the present disclosure is intended to include the combination of the features of the embodiment except the combination shown. In addition, the embodiment can also be supplemented by other features in the described features of the present invention.

[0041] In the figures, the same reference numerals respectively denote elements with the same function.

[0042] Figure 1A schematic diagram of a motor vehicle 10 according to an exemplary embodiment of the present invention is shown, which has a detection device 12 for detecting a fault state F of a battery cell 16 of a battery 18 of the motor vehicle. In this case, the battery 18 can be designed as a high-voltage battery of the motor vehicle 10 or as a battery module of such a high-voltage battery. In this case, the battery 18 also includes a plurality of battery cells 16. One of these battery cells 16 is additionally denoted here by the reference numeral 16a in order to be able to better illustrate and describe the method for detecting a fault state F of such a battery cell 16. Accordingly, how a fault state of such a battery cell 16 can be detected is explained with reference to this first battery cell 16a, but can also be applied in the same way to all other battery cells 16. If a fault state F of such a battery cell 16 is detected by the detection device 12, a signal S can be output, for example a warning signal to the driver of the motor vehicle or a signal for taking specific measures, such as switching off the battery, switching off the battery cell, etc.

[0043] The cells 16 operate reproducibly depending on environmental conditions, such as dynamic load distribution, uneven heat generation, etc. However, there may be different characteristics and strong differences between the cells 16 within the battery 18. Therefore, a robust diagnosis of all cells 16 is often very difficult. The following describes different voltage differences of the individual cell voltages of the battery cells 16 in different situations based on some examples.

[0044] Figure 2 For example, the voltage variation curve of each cell voltage UZ is plotted over a time t in seconds, in particular over a period of about one minute. Figure 2 As shown in FIG. 1 , none of the cells 16 has defects. As can be seen, even in the case where the cells 16 have no defects, sometimes very large voltage differences occur among all the cells 16 depending on the operating state. Therefore, there are no abnormal values ​​in this example. Figure 2 As is also clearly shown in FIG. 1 , there are ranges in which the individual cell voltage curves are very far apart from each other, but in other ranges they still extend approximately overlapping each other. Especially when the individual cell voltages UZ vary greatly over time, the corresponding differences can only be poorly resolved. This makes the detection of possible cell defects difficult.

[0045] Figure 3The schematic diagram of a portion of a battery 18 with a battery cell 16, which is again denoted by reference numeral 16a, is shown, and a defective battery cell 16a is also shown. In this example, the battery cell 16a is mechanically damaged from the outside and has a previous damage in the form of an intrusion of up to 3.4 millimeters in particular. Such mechanical damage to the cell from the outside can lead to a cell defect, which can be detected in a changed cell voltage UZ. This is shown in FIG. Figure 4 Shown in.

[0046] Figure 4 Shown here Figure 3 1 is a graphical representation of the voltage curve UZ of the individual cell voltages of the battery 18 in FIG. 1 , wherein the voltage curve of the damaged cell 16a is represented by UX. The time t is plotted here in hours. The battery 18 with the damaged cell 16a is subjected to a plurality of charge and discharge cycles, wherein the cell defect of the cell 16a is detected precisely when switching from discharge to charge, in particular in the strongly deviating cell voltage UX. It can also be seen here that the deviation of this cell voltage UX from the remaining cell voltages UZ is not so obvious in other ranges, for example during charging or discharging. It can already be seen here that it is difficult to define, for example, a limit value according to which such a fault state of the cell 16a can be detected and which is simultaneously valid for all possible operating ranges of the battery 18.

[0047] exist Figure 5Another example is shown in FIG. Here again, the variation curves of the cell voltage UZ of multiple intact battery cells 16 and the variation curve of the cell voltage UX of a defective battery cell are shown. The charging process begins in the first time phase. The charging phase is indicated by the reference numeral L. In addition, three driving operations F1, F2, and F3 with different loads follow the charging phase L. At the beginning of the charging process L, the voltage variation curves of all cells do not have voltage deviations relative to each other. Only at the end of charging, that is, at the end of the charging phase L, the first voltage deviation of the cell 16a from the remaining cells 16 is displayed. The voltage of the defective cell 16a is again indicated by the reference numeral UX, while the voltage variation curves of the remaining cells are indicated by UZ. It is in the scope of this first driving operation F1 that a small standard deviation is generally displayed. Therefore, when charging L, all cell voltages UX and UZ are still close to each other. At the end of charging or before driving operation F1, a cell 16a suddenly deteriorates significantly. This is particularly obvious because the remaining cells differ from each other by only, for example, 10 millivolts. Under load, i.e. in the driving operating state F2, a greater voltage difference is observed in all cells 16, 16a, but this is not common in the case of cold or unevenly heated cells 16. Under light load, for example in the driving state range F3, the cell voltages UZ are close to each other again, wherein the cell 16a also deviates further with its cell voltage UX. As can be seen, a cell defect of a cell 16a cannot be identified equally well in all operating states. For example, in the driving state range F2, a significant deviation of the cell voltage UZ of the defective cell 16a from the remaining cells 16 cannot be identified. However, detection in the driving ranges F1 and F3 is possible in principle. This knowledge is used to determine a suitable limit value for detecting such a defective cell 16a, as will be explained in more detail later.

[0048] first, Figure 6 , Figure 7 and Figure 8 Other examples of voltage variation curves of individual battery cells 16 are also shown, including the voltage variation curve UX of a defective battery cell. Figure 7 Show Figure 6 A detailed view of a part of the embodiment of the present invention, wherein the part is denoted by reference numeral 20. Figure 6 The cell balancing is also shown, wherein the balancing time increases uniformly and initially without any abnormalities until the last charging process of the cell, which is at Figure 7 Here it can be seen that the balancing system attempts to balance all cells 16 , wherein the cell 16 a has deviated significantly from the remaining voltage UZ with its voltage UX since the last charging process and can no longer be charged.

[0049] Figure 8 The charging process L is shown together with the subsequent rest phase R, in which the battery 18 is no longer subjected to load. The rest phase R lasts approximately 30 hours. In this rest phase R, the cell 16a shows an increased self-discharge rate compared to the remaining cells 16 since the last charging process.

[0050] In order to detect these defects accurately and early, there are now several possibilities. For this purpose, on the one hand, battery parameters such as temperature and voltage UZ, UX and parameters derived from these battery parameters such as standard deviation and / or interval or interquartile range can be measured or calculated and stored.

[0051] For example, the battery variables can be compared with earlier values ​​by means of the standard deviation under constant, similar conditions, for example a new charging process with the same charging current, when the vehicle is started with a low discharge current, optionally also taking into account the state of charge or the battery temperature. If, taking into account the standard deviation, a cell suddenly deviates greatly from the other cells compared to the time of the last diagnosis, it can be assumed that a cell defect has occurred.

[0052] It is particularly advantageous to first use the interquartile range of a certain subset of measurements or to use a general interval of the cell voltage UZ in order to determine constant conditions, such as a charging process with a constant current, a stationary vehicle with a small discharge current, where comparability of the battery parameters is possible and the deviations are relatively small. Under constant conditions and therefore, for example, with a small interquartile range or a small standard deviation, small outliers can already indicate a cell defect. In contrast, for dynamic conditions, significantly larger threshold values ​​can be applied. This allows possible cell defects to be detected in the same way for different conditions and states.

[0053] Fig. 9Here, a schematic diagram of an embodiment of a detection device of a motor vehicle 10 is shown, which detection device can be used, for example, to detect such a single cell defect F. The detection device 12 uses the individual single cell voltages Ui of all the battery cells 16 of the battery 18, together with the single cell 16a assumed to be defective at the beginning, as input variables. In this example, the single cell voltage associated with the defective battery cell and the remaining single cell voltages are no longer distinguished, but the method for all single cell voltages Ui of the battery 18 is generally described. Here, the minimum single cell voltage Umin of all single cell voltages Ui is provided at the first input 22, and the maximum single cell voltage Umax is provided at the second input 24. In addition, an average value ū is determined from all single cell voltages Ui, which is also referred to as the second reference value ū in the scope of the present invention and is provided at the third input 26. In addition, a subset Un is determined from the single cell voltages Ui. In order to determine this subset, at least the maximum single cell voltage Umax and the minimum single cell voltage Umin are excluded. However, it is also possible to exclude multiple maximum voltage values ​​and minimum voltage values ​​from this subset Un. In addition, the magnitude between the maximum deviation of the maximum and minimum cell voltages of the subset Un is also determined in the fourth input module 28. This deviation represents the size of the dispersion of the subset Un and is therefore referred to as the dispersion value ΔUn below. In a special case, this dispersion value can be exactly the interquartile range. This dispersion value ΔUn is compared with a previously determined limit value, which is also referred to as the first reference value R1 below, by means of a proportional operator 30. In parallel with this, the numerical distance of the minimum voltage value Umin from the average value ū is determined on the one hand, and the numerical distance of the maximum voltage value Umax from the average value ū is determined by respective addition operators 32, 34. From these two obtained distances, the maximum distance is selected by the maximum operator 36. Therefore, as a result, this provides the maximum voltage interval from the average value ū in all input voltages Ui. If the comparison performed by the comparison operator 30 shows that the dispersion value ΔUn is less than or equal to the first reference value R1, the above-mentioned maximum deviation Δmax at the output of the maximum operator 36 is compared with the first threshold value S1. This comparison is again performed by the proportional operator 38. If the limit value S1, for example 20 millivolts, is exceeded, a fault is detected and a corresponding signal is output at the output 40. In addition, the reference numeral 42 denotes a logical AND operator, which ensures that two conditions must be met for this purpose, namely that the maximum deviation Δmax exceeds the first threshold value S1 and the interquartile range or dispersion value ΔUn is less than or equal to the first reference value R1. If these two conditions are not met, the deviation Δmax exceeding the first limit value S1 will not lead to fault detection. However, in any case, the maximum deviation Δmax is also compared with the second threshold value S2.Therefore, the second threshold value S2 also represents a limit value S2, by which the voltage extreme value, i.e. the minimum voltage value Umin or the maximum voltage value Umax, differs from the voltage average value ū, i.e. the second reference value ū, at most, or in this example, it is just no longer allowed for the voltage extreme value to deviate from the voltage average value, so that no fault state F is detected. This comparison is again carried out by another comparison operator 43. The threshold value S2 is significantly greater than the threshold value S1 and is, for example, 400 millivolts. If the deviation Δmax is greater than the second threshold value S2, a signal S is also output at the output 40, which indicates that a fault is present. Here, reference numeral 44 represents a logical OR operator.

[0054] In the case of a small dispersion represented by the dispersion value ΔUn, small deviations from the mean value ū can advantageously lead to the triggering of the signal S, while in the case of a large dispersion ΔUn, only large deviations of the voltage value Ui from the mean value ū also lead to the output of such a signal S. As a result, the limit value for triggering can also be selected to be significantly smaller, for example 20 mV, in a range with constant conditions and therefore with a small interquartile range, because in this range a small trigger already indicates a cell defect. For dynamic conditions, for example during driving or regeneration, a significantly larger threshold value is used, for example the described 400 mV. This makes it possible to provide an adaptive and always early detection of a cell defect.

[0055] Overall, these examples show how the identification of suddenly occurring single-body defects can be provided in a particularly efficient and adaptive manner by means of the present invention.

Claims

1. A method for detecting a fault state (F) of at least one battery cell (16a, 16) of a battery (18) having a plurality of battery cells (16, 16a), wherein: The method comprises detecting a cell voltage (Ui; UZ, UX) of a corresponding battery cell (16, 16a) of the plurality of battery cells (16, 16a) at a specific measuring time, determining a comparison value (ΔUn) based on at least one of the cell voltages (Ui; UZ, UX), comparing the comparison value with a provided first reference value (R1), wherein a fault state (F) is detected based on a result of the comparison, It is characterized in that A dispersion value (ΔUn) is determined, which represents the dispersion of at least a part (Ui, Un) of the cell voltages (Ui; UZ, UX) detected at a specific measuring time, and a fault state (F) is determined based on the dispersion value (ΔUn), wherein at least one voltage extreme value (Umax, Umin) is compared with a second reference value (ū), wherein the voltage extreme value represents the maximum voltage value (Umax) and / or the minimum voltage value (Umin) of the cell voltages (Ui; UZ, UX) detected at the measuring time, wherein the comparison value (ΔUn) represents the dispersion value (ΔUn) and the fault state (F) is detected in the following case, - if the dispersion value (ΔUn) does not exceed a first reference value (R1) at most and the at least one voltage extreme value (Umax, Umin) differs from a second reference value (ū) by at least a first limit value (S1); and If the dispersion value (ΔUn) is greater than the first reference value and the at least one voltage extreme value (Umax, Umin) differs from a second reference value (ū) by at least a second limit value (S2), the second limit value being greater than the first limit value (S1).

2. The method according to claim 1, It is characterized in that A dispersion value (ΔUn) is determined for a subset (Un) of the cell voltages (Ui; UZ, UX) detected at a measuring time, wherein the subset (Un) does not include at least a maximum voltage value (Umax) and / or a minimum voltage value (Umin) of the cell voltages (Ui; UZ, UX) detected at a measuring time.

3. The method according to any one of the preceding claims, It is characterized in that The dispersion value (ΔUn) is provided as the difference between the maximum voltage value and the minimum voltage value of the subset (Un).

4. The method according to claim 1, It is characterized in that The second reference value (ū) is an average value (ū) of the cell voltages (Ui; UZ, UX) at least comprised by the subset (Un).

5. The method according to claim 1 or 2, It is characterized in that As the dispersion value, the standard deviation or the variance of all the cell voltages (Ui; UZ, UX) detected at the measuring time is determined.

6. The method according to claim 1 or 2, It is characterized in that The cell voltages (Ui; Uz, UX) are determined for a plurality of successive measuring moments and a corresponding dispersion value is determined for each measuring moment, wherein a fault state (F) is detected based on a time variation of the dispersion value, wherein, when the time variation of the dispersion value is determined as a comparison variable, a fault state (F) is considered to be detected when the comparison variable exceeds a first reference value.

7. The method according to claim 1 or 2, It is characterized in that The plurality of consecutive measuring times lies within an operating time window which corresponds to a specific operating state of the battery (18) in which a change in the battery current of the battery (18) over time is less than a predefinable limit value.

8. A detection device (12) for detecting a fault state (F) of at least one battery cell (16a, 16) of a battery (18) having a plurality of battery cells (16, 16a), wherein: The detection device (12) is designed to detect the cell voltage (Ui; UZ, UX) of the corresponding battery cell (16, 16a) of the plurality of battery cells (16, 16a) at a specific measurement time, determine a comparison value (ΔUn) based on at least one of the cell voltages (Ui; UZ, UX), compare the comparison value (ΔUn) with a provided first reference value (R1), and detect a fault state (F) based on a result of the comparison, It is characterized in that The detection device (12) is designed to determine a dispersion value (ΔUn) representing the dispersion of at least a part of the cell voltage (Ui; UZ, UX) detected at a specific measuring time, and to determine a fault state (F) based on the dispersion value (ΔUn), wherein at least one voltage extreme value (Umax, Umin) representing the maximum voltage value (Umax) and / or the minimum voltage value (Umin) of the cell voltage (Ui; UZ, UX) detected at the measuring time is compared with a second reference value (ū), wherein the comparison value (ΔUn) represents the dispersion value (ΔUn) and the fault state (F) is detected in the following case, - if the dispersion value (ΔUn) does not exceed a first reference value (R1) at most and the at least one voltage extreme value (Umax, Umin) differs from a second reference value (ū) by at least a first limit value (S1); and If the dispersion value (ΔUn) is greater than the first reference value and the at least one voltage extreme value (Umax, Umin) differs from a second reference value (ū) by at least a second limit value (S2), the second limit value being greater than the first limit value (S1).

9. A motor vehicle (10) having a detection device (12) according to claim 8.

Images