Method, apparatus, and computer program product for determining capacity loss of a battery energy storage

Abstract

High-precision coulometric measurements for aging estimation of lithium-ion batteries are improved in that the capacity value of the battery is determined from measurement data including current measurements by means of two different calculation methods, which in principle should have the same result. Here, a current calibration, which indicates the relationship between the actual current strength and the measured current strength, enters to different extents into the two calculation methods, whereby the results of the calculation methods can not be equal when the current calibration is erroneous. The current calibration is determined by an optimization method in which the results of the two calculation methods are as good as possible in agreement and thus the error is reduced or completely avoided.

Description

Technical Field

[0001] The present invention relates to a method for determining the capacity loss of a battery energy store, in particular for determining the predicted aging of a battery energy store, a device for carrying out the method, and a computer program product. Background Art

[0002] Lithium-ion batteries, also referred to below as Li-ion batteries, are used as energy storage devices in mobile and stationary applications due to their high power and energy density. To operate these electrochemical energy storage devices safely, reliably, and for as long as possible without maintenance, it is necessary to have the most accurate possible knowledge of key operating states, particularly regarding the state of charge (SOC) and the state of health (SOH).

[0003] It is known that battery aging, in particular so-called cyclic aging, can be negatively affected by rapid charging at high and low temperatures, depending on the state of charge and depth of discharge, as well as the charging and discharging power. Therefore, it is possible that the same type of battery cell can undergo significantly different numbers of load cycles, depending on the aforementioned parameters.

[0004] To determine the predictable aging profile, the prior art uses measurements during the design phase of the battery system to determine the aging characteristics of the battery cells used. The actual aging rate with a realistic load profile is usually not tested. Rather, in so-called shock tests (Rafftests), the aging rate or cycling stability is determined over a compressed load profile. These results are used to parameterize empirical aging models, which in turn derive the aging profile in use. The determination of the future aging profile based on physical and / or chemical measurements, which depends on the load profile, operating point, and ambient conditions, is difficult to implement due to the nonlinearity of the underlying physical and chemical processes and their complex interactions.

[0005] High-Precision Coulometry (HPC) can improve the measurement of capacity loss and the estimation of aging in actual operation. In HPC measurements, load cycles are performed and the capacity loss is determined based on the trend of the capacity measurement.

[0006] In HPC measurements, relatively small differences in the amount of charge added to and removed from the battery are determined by integrating the measured current intensity. Therefore, HPC measurements fundamentally require very high measurement accuracy, particularly in current measurement. This high measurement accuracy can only be achieved with test equipment designed specifically for HPC measurements, if available.

[0007] Typical test equipment used in battery production is simply not useful due to high measurement uncertainties. Summary of the Invention

[0008] Therefore, the technical problem to be solved by the present invention is to provide a method and a device which can improve the intrinsic accuracy of HPC measurements.

[0009] The above-mentioned technical problem is solved by the method according to the invention for determining capacity loss, the device according to the invention and the computer program product according to the invention.

[0010] In the method according to the present invention for determining at least one average capacity loss of a battery energy storage device, the results of a series of measurements on the battery energy storage device comprising a plurality of load cycles are provided. A load cycle comprises a charging phase and a discharging phase. The results include values ​​of the current measurements for the load cycles (i.e., the charging and discharging).

[0011] Furthermore, a first discharge capacity and a second discharge capacity of the battery energy storage device are determined based on the results of the measurement series using a first calculation rule and a second calculation rule. Calculation rules are used in this case such that the calibration of the current measurement is incorporated differently into the first and second calculation rules. The calibration is the relationship between the actual current and the current measurement value.

[0012] Finally, an optimization method is performed such that a calibration of the current measurement is determined with which a maximum agreement of the determined first and second charge capacities is achieved.

[0013] The device according to the invention for carrying out the method according to the invention includes a memory for recording the results of a series of measurements on a battery energy store having a plurality of load cycles. The load cycles include a charging phase and a discharging phase, and the results include current measurement values.

[0014] The device also includes a calculation unit. This calculation unit is designed to calculate a first discharge capacity and a second discharge capacity of the battery energy storage device based on the results of the measurement series using a first calculation rule and a second calculation rule. The current measurement is calibrated differently in the first and second calculation rules, with the calibration being the relationship between the actual current and the current measurement value.

[0015] Furthermore, the computing unit is designed to carry out an optimization method in order to determine a calibration of the current measurement with which a maximum agreement between the determined first charge capacity and the second charge capacity is achieved.

[0016] The computer program product according to the invention can be directly loaded into the memory of a programmable computing unit. The computer program product comprises program code means for carrying out the method according to the invention when the computer program product is executed in the computing unit.

[0017] It goes without saying that the optimization is expediently performed in a computer-assisted manner in a computing unit.

[0018] The method according to the invention eliminates inaccuracies in the measurement of battery properties, in particular inaccuracies in the form of charge measurements, which arise from faulty, i.e., inaccurate, calibration of the current measurement. The invention recognizes that it is possible to determine at least two charge values ​​from the current measurements recorded during a load cycle measurement, which are nominally identical, but in which the calibration of the current measurement influences the result to varying degrees. This allows the influence of the calibration to be eliminated.

[0019] Advantageously, this improves measurement accuracy, particularly in HPC testers. Thus, even devices that were calibrated too imprecisely at the factory can be used in this manner and accurately calibrated using the method according to the present invention (during battery measurements). Furthermore, convincing HPC measurements can even be performed on "standard testers," with which meaningful measurements with the necessary accuracy would not be possible without the method according to the present invention.

[0020] The result of the measurement series includes current measurement values ​​suitable for approximating a time-integrated sum. In other words, this involves a plurality of measurement values. The measurement values ​​are typically present as electrical signals, for example, as digital signals. The result of the measurement series may also include voltage values, in particular voltage values ​​correlated with the current values.

[0021] An optimization method is understood to be a method that, for example, minimizes an objective function, wherein certain set parameters of the objective function can be varied for the purpose of minimization. Optimization methods generally operate iteratively and cannot be replaced by analytical solutions. In the present case, for example, the objective function to be minimized could be the absolute value of the difference between the first and second charge capacities. The parameters to be varied are those that form the calibration for the current measurement.

[0022] The provision of measurement results can be understood as meaning that the measurement results are already available and are therefore merely recorded and processed. In this case, the actual measurement is performed beforehand and can therefore be separated in space and time. However, it is also possible for the measurement results to be provided and processed immediately during the measurement process, i.e., for the measurement to be performed simultaneously with the method.

[0023] The present invention also provides an advantageous design solution of the method according to the present invention. Here, the embodiments according to the present invention can be combined with one or more other features. Therefore, the following features can also be additionally provided for this method:

[0024] The first discharge capacity of a load cycle can be calculated based on the discharge drift of the load cycle and the Coulombic efficiency of the load cycle. For this purpose, for example, Equation 3 below can be used as a calculation rule. The discharge drift is understood to be the difference between two consecutive charge states in the lower charge state of a charge cycle, i.e., after a discharge, and can be calculated, for example, using Equation 5. The Coulombic efficiency or Coulomb coefficient refers to the ratio of the energy captured by the discharge to the energy supplied by the previous charge; for this purpose, Equation 4 can be used.

[0025] Alternatively, the first discharge capacity of the load cycle can be calculated based on the capacity value after the charging phase of the load cycle and the capacity value before the charging phase.

[0026] The second discharge capacity of a load cycle can be determined based on the first discharge capacity of the preceding additional load cycle and the sum of the capacity losses of the load cycles between the load cycle and the additional load cycle. The corresponding equation for this calculation is given in Equation 6. The capacity loss is understood to be the difference between the charge drift and the used discharge drift, as given in Equation 7. Similarly to the discharge drift, the charge drift is understood to be the difference between two consecutive charge states in the upper charge state of the charge cycle, i.e., after charging, and can be calculated, for example, using Equation 8.

[0027] For calibration calculation and optimization, an offset can be used, which indicates the difference between the actual current value and the measured current value at a current of 0 A. Furthermore, a slope value can be used, which indicates the proportionality factor between the actual current and the measured current. This corresponds to depicting the current curve using a first-order polynomial, i.e., a straight line. In this case, the actual current can be, for example, the raw electrical signal from the measurement.

[0028] This calibration is simple and universal and represents a typical ratio for current measurements. The offset corresponds to the zero value of the AD converter used, while the slope value or gain reflects the resistor used as a shunt for the current measurement, whose magnitude is subject to sample scatter and therefore does not conform exactly to the specification.

[0029] Alternatively, a curvature factor can be used for calculation and optimization of the calibration, which indicates the nonlinear, in particular quadratic, relationship between the actual current and the measured current. Together with the offset and gain, this creates a quadratic polynomial that allows a more precise adaptation to the actual measurement.

[0030] Alternatively, for calibration calculation and optimization, the relationship between the actual current and the measured current can be modeled as a piecewise linear relationship with three or more support points. This approach allows for more accurate modeling than a simple straight line and, by increasing the number of support points, allows, in principle, any degree of accuracy, regardless of the actual form of the relationship between the actual current and the measured value.

[0031] The results of the measurement series are preferably generated using a high-precision coulometric measuring device, which is designed specifically for this type of measurement and can ideally compensate for any inaccuracies in the current measurement that still exist using the method according to the invention.

[0032] Preferably, a load cycle is used in which charging and discharging always takes place between a definable lower voltage and a definable upper voltage of the battery energy store.

[0033] In this case, it is preferably ensured that in each successive load cycle a constant temperature prevails within the load cycle, thereby reducing temperature-related inaccuracies in the measurement.

[0034] Preferably, load cycles are performed during the measurement until the capacity loss is approximately constant over two or more consecutive load cycles. The determined capacity loss is considered to be approximately constant if the slope of a tangent matching the course of the capacity loss is less than 10% of the average value of the slopes of the last 10% of the measured capacity losses. Alternatively, the capacity loss is considered to be approximately or substantially constant if the absolute variation of at least two consecutive capacity losses is, in particular, less than 5%.

[0035] For further use of the measurement results, an average capacity loss can be determined. The average capacity loss is obtained as the average of a number of capacity losses for different load cycles. The average capacity loss describes the aging rate for the selected load cycle and is expressed as capacity loss per cycle. Advantageously, this allows a quantitative evaluation of the measurement data of a high-precision coulometric measurement device with respect to the aging rate of the battery. A quantitative evaluation is possible because the absolute value of the capacity can be determined based on the determination of the average capacity loss. The load cycle used here defines a specific operating point by selecting voltage limits, which is characterized by the average state of charge (SOC) and the depth of cycle (DOD).

[0036] Advantageously, the average capacity loss is determined computer-aided by a sliding linear fit on the values ​​of the capacity loss and finding the minimum slope in the resulting straight line equation. Based on the fit of all capacity losses, the data set including the determined capacity loss is continuously shortened and a new straight line is fitted. This fitting is performed until a certain minimum remaining length of the data set, i.e. the capacity loss. Subsequently, the straight line equations are sorted in increasing order according to the values ​​of their slopes. The measurement can be considered valid when at least two of these slopes are numerically less than 10% of the average value of the last 10% of the capacity loss. For example, if the average value of the last twenty capacity losses (in particular when at least 200 capacity losses are measured) is 5 mAh / load cycle, the slopes of the two best-fitting tangents ("fits") should be less than 0.05 mAh / load cycle.

[0037] It is particularly advantageous if the capacity loss is not taken into account for determining the residual capacity until after the start-up phase of the load cycle. The capacity loss determined at the beginning of the measurement, i.e. during the start-up process, is subject to error and should therefore not be included in the determination of the average capacity loss. It has been found that the start-up phase ends when at least two of the slopes of the straight lines applied to the capacity loss in the fit are numerically less than 10% of the average value of the last 10% of the measured capacity losses. Alternatively, the capacity loss is considered to be almost constant if two consecutive capacity losses and / or a floating-point average value over at least twenty capacity losses have a capacity loss variation of less than 5%. Advantageously, this method ensures that the residual capacity can be determined quickly but reliably based on the capacity loss.

[0038] In further advantageous embodiments and developments of the present invention, a constant temperature is present during the determination of the capacity loss in each successive load cycle. In other words, the temperature may be different in two consecutive determinations of the capacity loss. However, the temperature remains constant during the load cycle. Advantageously, load cycles recorded at different temperatures can be combined to determine the average capacity loss, as long as the temperature remains constant within a load cycle.

[0039] Batteries or battery cells can be operated in a temperature-controlled chamber. In this embodiment, the batteries or battery cells are arranged in a temperature-controlled chamber. In particular, the temperature-controlled chamber ensures sufficiently high temperature stability during the battery's load cycle. Alternatively, it is possible to stabilize the battery storage temperature using a contact temperature regulator and / or a cooling circuit. Advantageously, the use of temperature regulation ensures that the temperature remains constant during the determination of capacity loss. This advantageously increases the reliability of determining the remaining capacity of the battery storage.

[0040] In advantageous embodiments and extensions of the present invention, the lower voltage is selected from a first voltage range, and the upper voltage is selected from a second voltage range. The second voltage range is preferably higher than the first voltage range. Particularly advantageously, both the first and second voltage ranges can be selected from the entire operating voltage range of the battery storage device. In other words, a full cycle need not be performed. Therefore, it is possible to use the permissible voltage range of the battery storage device according to the product specifications, or a voltage range that exceeds this range. Advantageously, measuring capacity loss without performing a full cycle, i.e., a complete charge and discharge, allows for a shorter measurement duration. Furthermore, the battery storage device is not subjected to a significant load by this measurement, which advantageously prevents rapid aging.

[0041] Particularly advantageously, the floating average value for determining the average capacity loss is determined from at least 20 capacity losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Other features, characteristics and advantages of the present invention will be apparent from the following description with reference to the accompanying drawings, in which:

[0043] Figure 1 shows an apparatus for determining the average capacity loss and the remaining capacity using a high-precision coulometric measuring device,

[0044] Figure 2 shows the voltage-time graph of the load cycle,

[0045] Figure 3 A graph showing the time course of the absolute charge balance,

[0046] Figure 4 shows a graph with a current calibration characteristic curve before and after carrying out the optimization method,

[0047] Figure 5 A graph showing the discharge capacity before performing the optimization method,

[0048] Figure 6 shows a graph of the discharge capacity after performing the optimization method,

[0049] Figure 7 Shown are graphs with values ​​of the coulombic efficiency before and after performing the optimization method. DETAILED DESCRIPTION

[0050] Figure 1The present invention shows an apparatus for determining average capacity loss and residual capacity using a high-precision coulometric measuring device 1. Apparatus 1 includes a battery reservoir 2 having at least one battery cell. The battery reservoir is located in a temperature-controlled chamber 3. Battery reservoir 2 is connected to a high-precision coulometric measuring device 4 via a power supply line 11. High-precision coulometric measuring device 4 is in turn connected to a computing unit 10 via a data line 12. High-precision coulometric measuring device 4 records the charge-time curve of battery reservoir 2 with very high accuracy. Battery reservoir 2 is operated in a periodic load cycle 100.

[0051] Figure 2 The figure shows a voltage-time graph recorded by a high-precision coulomb measuring device 4 during a periodic load cycle 100 of a battery energy store 2. The load cycle 100 includes a discharge from a first charge state 21 to a second charge state 22, wherein the first charge state 21 is at an upper voltage 25 and the second charge state 22 is at a lower voltage 26. Subsequently, in the load cycle 100, the battery energy store 2 is charged from the second charge state 22 to a third charge state 23. As a next step, the third charge state 23 is discharged to a fourth charge state 24 in the load cycle 100. In each individual charge / discharge step, the upper voltage 25 and the lower voltage 26 are maintained as voltage limits. Charging lasts for a charging time period t C The discharge lasts for the discharge time period t D .

[0052] Based on Figure 2 From the measurements shown in , it is now possible to determine how much accumulated charge has flowed during the various charging and discharging steps. The first charge quantity Q1 can be calculated using Equation 1, where I represents the current and T D Indicates the discharge time period:

[0053] Q1=∫I dt D Equation 1

[0054] Within the load cycle 100, the battery energy storage device 2 is then charged from the second charge state 22 to the third charge state 23 by means of the first charge 32. A second charge quantity Q2 is added to the battery energy storage device 2. Q2 can be calculated using Equation 2:

[0055] Q2=∫I dt C Equation 2

[0056] Within load cycle 100, battery energy storage 2 is subsequently discharged from third charge state 23 to fourth charge state 24 by means of second discharge 33. The withdrawn charge quantity Q3 can again be calculated analogously to equation 1 from the time period of the discharge and the associated current.

[0057] Figure 3A schematic and highly simplified diagram is shown in which the charge quantities Q1 ... 3 thus determined are plotted over time. For the further course, particular attention is paid to the discharge capacity Q 0,i , for example in Figure 3 Q indicated in 0,2 and Q 0,4 These discharge capacities can be determined in two different ways using the equations described below.

[0058] In the first calculation form, the discharge capacity Q 0,i Determined with the aid of equation 3. In all subsequent equations, the index i always denotes the time of charging in the i-th load cycle, and the index j always denotes the time of discharging following this charging time in the same load cycle.

[0059]

[0060] Here, CE i represents the coulombic efficiency of the corresponding cycle consisting of charge and discharge, i.e., according to

[0061] Equation 4: Ratio of the extracted charge to the previously fed charge.

[0062]

[0063] Δ D,j is understood as the discharge drift. This is calculated according to Equation 5 and gives two consecutive discharge states Q j and Q j-1 The difference:

[0064] Δ D,j =Q j -Q j-1 Equation 5

[0065] exist Figure 3 The discharge drift Δ is shown for various values ​​of j in D,j .

[0066] In the second calculation form, the nominally identical discharge capacity Q is determined using Equation 6 meas,i :

[0067]

[0068] The capacity loss Δ used here Kap,k By charge drift Δ C,i and discharge drift Δ D,j Calculated according to Equation 7:

[0069] Δ Kap,i =Δ D,j -Δ C,i Equation 7

[0070] The discharge drift has already been used and is calculated according to Equation 5 given above. The charge drift is calculated in a similar way according to Equation 8:

[0071] Δ C,i =Q i -Q i-1 Equation 8

[0072] Therefore, the value Q 0,i The calculation is essentially based on this variable present in the i-th load cycle. meas,i According to the initial value (here exemplarily Q 0,2 ) and the value obtained between the i-th load cycle. In the case of ideal, i.e. error-free current measurement, these two values ​​are identical, i.e.:

[0073] Q 0,i =Q meas,i Equation 9

[0074] However, these two values ​​differ from each other in reality due to the current calibration present in the current measurement, which is not completely accurate. The greater the difference between these values, the greater the error in the current calibration.

[0075] Equation 9 with f = Q 0,i -Q meas,i The form of is used as the basis for the optimization, where the function value f must be minimized. The variable to be varied for the optimization forms the current calibration. The current calibration is the mapping from the measured current value to the corrected measured value. If the value Q is achieved by the optimization 0,i With Q meas,i The corrected measured value corresponds very accurately to the actual current.

[0076] Current calibration can, for example, have an offset and a gain. The offset represents a constant value around which the measured current is shifted, while the gain indicates the slope of the current characteristic curve. Optimization starts with arbitrary values ​​for the offset and gain. Here, 0 is used as the starting value for the offset, meaning no offset.

[0077] The optimization is usually performed on a computer basis. For this purpose, known programs can be used, which only require the function to be minimized as well as boundary conditions and parameters to be specified.

[0078] The result of this optimization is Figure 4 Shown in. Figure 4 The straight line 41 before optimization is shown, which produces an offset and a gain, where it is a straight line passing through the origin and having a slope of 1 / 1000. Figure 4The resulting straight line 42 after optimization is shown. As shown, the optimization results in a slight deviation of approximately -30 mA offset and 1 / 1000 gain, ie the straight line is slightly rotated. Thus, using these values ​​for current calibration, Q is achieved. 0,i and Q meas,i to a large extent equal.

[0079] Since the calibration changes the current value, the integral (or summation) used to form the charge value must be recalculated for optimization, namely equations 1 and 2. The current value entering the integral or summation is changed to I korr (t)=offset+gain*I(t). This change, ie the correction of the calibration, thus influences the resulting charge value, but cannot simply be performed directly on the charge value Q due to the integral formation.

[0080] In this embodiment, the optimization is performed based on measurements comprising 200 charging cycles. For the best quality of the current calibration, several value pairs Q are considered. 0,i and Q meas,i , ie a plurality of subscripts i. For example, the last 50 value pairs of a measurement series with 200 charging cycles can be considered.

[0081] Figure 5 For this example, Q before optimization is shown. 0,i 51 and Q meas,i 52. It can be seen that nominally the same value actually differs by about 0.15% due to the slightly inaccurate current calibration. Therefore, the current measurement does not achieve the 0.01% accuracy usually required for HPC measurements. Figure 6 Shows the Q after optimization 0,i 61 and Q meas,i 62, i.e. using improved current calibration. It is shown that the two paths are offset and the paths 61, 62 are almost completely superimposed on each other after optimization, thereby achieving high measurement accuracy. The accuracy achieved in this way meets the accuracy standard of 0.01%.

[0082] Figure 7 This shows that improved current calibration also improves the accuracy of other derived values. Curve 71 shows the behavior of the Coulombic efficiency according to Equation 4 before optimization. For a wide range of measurements, the Coulombic efficiency is approximately 102%, which is impossible because the battery cannot extract more energy than it was previously charged. Therefore, there is a significant measurement error. In contrast, after optimization, curve 72 results, where the Coulombic efficiency rises from the initial low value to almost 100% over the majority of the measurement. Therefore, at least here, there is no longer a significant measurement error.

[0083] Even in the case of Coulombic efficiency results below 100%, a comparison of absolute values ​​is often difficult, as the measurement errors due to imperfect current calibration are greater than the differences between the battery cells. Since the present invention corrects these errors for each measurement and thus almost completely eliminates them, even a comparison of the absolute values ​​of the Coulombic efficiency is possible.

[0084] Advantageously, the process shown includes an implicit verification of the measurement results by controlling the residuals of the optimizer. For example, the sum of the differences of the two equations 3 and 6, normalized to the mean value of the discharge capacity considered, can be implemented for all measurement points as the objective function for the optimization method. This ensures that for all measurements, regardless of battery capacity and depth of discharge, a uniform measure of the quality of the found optimal solution is provided. The optimizer, for example the MATLAB function "fmincon", can be set to, for example, 10 -6 The quality at which the optimization process is concluded as successful. If the target value is not reached, or the optimizer is interrupted for other reasons, the measurement should be carefully considered and, if necessary, discarded. Otherwise, the measurement can be considered valid and its results can be considered convincing.

[0085] Reference Signs List

[0086] 1 Device for predicting remaining capacity

[0087] 2 Battery Energy Storage

[0088] 3. Temperature Control Room

[0089] 4 High-precision Coulomb measurement device

[0090] 10 computing units

[0091] 11 Power cord

[0092] 12 data cables

[0093] 13 Computer program products

[0094] 21 First charging state

[0095] 22 Second charging state

[0096] 23 Third charging state

[0097] 24 Fourth charging state

[0098] 25 Upper voltage

[0099] 26 Lower voltage

[0100] 100 load cycles

[0101] tC Charging time period

[0102] t D Discharge time period

[0103] Q 0,2 ,Q 0,4 ,Q meas,4 discharge capacity

[0104] Δ D,j Discharge drift

[0105] Δ C,i Charge drift

[0106] 41 Current calibration before optimization

[0107] 42 Current calibration after optimization

[0108] 51, 52 Discharge capacity before optimization

[0109] 61, 62 Discharge capacity after optimization

[0110] 71,72 Coulombic efficiency before and after optimization

Claims

1. A method for determining at least one average capacity loss of a battery energy storage device (2), comprising the following steps: - providing the results of a series of measurements on a battery energy storage device (2) having a plurality of load cycles (100), wherein the load cycles (100) comprise a charging phase and a discharging phase, and wherein the results comprise values ​​of current measurements, - determining the first discharge capacity and the second discharge capacity (Q) of the battery energy storage device (2) from the results of the measurement series by means of a first calculation rule and a second calculation rule 0,i , Q meas,i ), wherein the calibration of the current measurement enters differently into the first calculation rule and the second calculation rule, wherein the calibration is a calculation rule for correcting the value of the current measurement, - performing an optimization method such that a calibration of the current measurement is determined, with which the determined first and second discharge capacities (Q 0,i , Q meas,i ) for maximum consistency.

2. The method according to claim 1, wherein The first discharge capacity (Q) of the load cycle (100) is calculated based on the discharge drift of the load cycle (100) and the coulombic efficiency of the load cycle (100). 0,i ).

3. The method according to claim 1, wherein The first discharge capacity (Q) of the load cycle (100) is calculated based on the capacity value after the charging phase of the load cycle (100) and the capacity value before the charging phase. 0,i ).

4. The method according to claim 1, wherein The second discharge capacity (Q) of the load cycle is determined from the sum of the first discharge capacity of the preceding further load cycle and the capacity loss of the load cycle between the load cycle and the further load cycle. meas,i ).

5. The method according to any one of claims 1 to 4, wherein To calculate and optimize the calibration, an offset value indicating the difference between the actual current value and the measured current value when the current is 0 A and a slope value indicating the proportionality factor between the actual current and the measured current are used.

6. The method according to any one of claims 1 to 4, wherein For the calculation and optimization of the calibration, a curvature factor is used, which indicates the non-linear relationship between the actual current and the measured current.

7. The method according to any one of claims 1 to 4, wherein For calculation and optimization of the calibration, the relationship between the actual current and the measured current is modeled as a piecewise linear relationship with at least three support points.

8. The method according to any one of claims 1 to 4, wherein The results of the measurement series are generated with the aid of a high-precision coulometric measuring device (4).

9. The method according to any one of claims 1 to 4, wherein The charging and discharging of the load cycle (100) takes place between a lower voltage (26) and an upper voltage (25) of the battery energy store (2).

10. The method according to any one of claims 1 to 4, wherein In each successive load cycle (100), a constant temperature exists within the load cycle.

11. The method according to any one of claims 1 to 4, wherein A first charge transfer is determined as a difference between a first upper charge state (24) and a second upper charge state (22), and wherein the second charge transfer is determined as a difference between a first lower charge state (23) and a second lower charge state (21), and wherein the capacity loss is determined by the difference between the first charge transfer and the second charge transfer.

12. The method according to any one of claims 1 to 4, wherein The load cycles (100) are performed until the capacity loss is nearly constant over two or more consecutive load cycles (100).

13. The method according to any one of claims 1 to 4, wherein The method is carried out computer-assisted in a computing unit (10).

14. A device (1) for carrying out the method according to claim 1, comprising a memory and a computing unit (10), the memory being used to record the results of a series of measurements on a battery energy store (2) with a plurality of load cycles (100), wherein: The load cycle (100) comprises a charging phase and a discharging phase, wherein the result comprises a value of a current measurement, wherein the calculation unit (10) is designed to: - determining the first discharge capacity and the second discharge capacity (Q) of the battery energy storage device (2) from the results of the measurement series by means of a first calculation rule and a second calculation rule 0,i , Q meas,i ), wherein the calibration of the current measurement enters differently into the first calculation rule and the second calculation rule, wherein the calibration is a calculation rule for correcting the value of the current measurement, - performing an optimization method such that a calibration of the current measurement is determined, with which the determined first and second discharge capacities (Q 0,i , Q meas,i ) for maximum consistency.

15. A computer program product (13) that can be directly loaded into a memory of a programmable computing unit (10), the computer program product having program code means for executing a method according to any one of claims 1 to 13 when the computer program product (13) is executed in the computing unit (10).

Images