Current sensor inaccuracy detection and determination method for improving model accuracy

Abstract

The present invention relates to a method of selecting a current sensor from at least two current sensors installed in a battery structure that provides a relatively accurate current measurement profile as an input variable as a basis for model calculations within the cloud. Current measurement value change curves measured by the at least two current sensors are transmitted to a pre-processing stage through the main BCU data, wherein current sensor identification is carried out. The current measurement value curves measured by the at least two current sensors are continuously compared. The measured current measurement value curves of the at least two current sensors are fed to a battery model, which determines a battery voltage curve for each phase on the basis of the measured current measurement value curves. The determined battery voltage profile is then compared with the current measured value profiles measured by the at least two current sensors, on the basis of which a determined battery voltage profile is obtained, which has a minimum deviation from the measured voltage profile and is used for subsequent model calculations in the cloud.

Description

Technical Field

[0001] This invention relates to a method for selecting a current sensor from at least two current sensors mounted in a battery structure that provide relatively accurate current measurements, which serve as input parameters for model calculations within a cloud. Furthermore, this invention relates to an application of the method for detecting and determining current sensor inaccuracies in order to improve the accuracy of the computational model. Background Technology

[0002] DE 10 2012 205 401 A1 discloses an apparatus comprising a battery having a plurality of individual cells and a sensor system for redundantly determining the battery current flowing through the battery electrodes, wherein the sensor system comprises: a current measuring device for detecting the battery current and providing a current measurement value; a voltage measuring device for individually detecting a time-varying curve of the voltage of at least one individual cell and providing at least one time-varying curve of the voltage value; and an evaluation unit configured to determine the battery current from the at least one time-varying curve of the voltage measurement value, provide a corresponding current value, and compare the current value with the current measurement value.

[0003] DE 10 2018 127 817 A1 discloses a method for diagnosing an electrical system. The electrical system includes a traction battery, a motor, a frequency converter configured to control the motor, and a plurality of current sensors configured to detect current characteristics of the traction battery, or current characteristics through or within the frequency converter. The frequency converter is controlled according to a reference variable, wherein, during normal operation of the system, predetermined loads are connected to the system at predetermined time intervals, and / or the control of the frequency converter is altered to change the electrical power loss in the motor and / or the frequency converter. The current sensors are diagnosed using their outputs.

[0004] DE 10 2014 208 680 A1 discloses a method for monitoring a current sensor when determining the total current output from a battery in a circuit, wherein the circuit has a total line segment and parallel line segments connected in series with the total line segment, wherein the total line segment has a total line and the parallel line segments have at least two parallel line paths. The total current flows in the total line, and a portion of the current flows in each of the at least two parallel line paths, wherein each of the at least two line paths has a current sensor for determining the corresponding portion of the current. The circuit has a current sensor for determining the total current in the total line. In a scaling step, an assigned scaling factor is determined for each line path, the scaling factor describing the inverse relationship between the corresponding portion of the current in the corresponding line path and the total current in the total line, and in a monitoring step, the portion of the current measured by the current sensor in the corresponding line path is determined. In the total line, the measured total current is determined by the current sensor. The total current is calculated from the measured portion of the current and the assigned scaling factor. To monitor the current sensor, the calculated total current is compared with each other and with the measured total current separately.

[0005] Modern battery packs typically include at least two current sensors. This is necessary to compensate for the failure of either current sensor, preventing vehicle breakdowns. However, these current sensors are required not only for redundancy but also for mutual reliability verification. For cost reasons, current sensors are often chosen such that one current sensor (the more expensive of the two) has a higher resolution and different resolutions across different current ranges. They may also differ in accuracy. However, neither current sensor possesses the high resolution and accuracy used in laboratory measurements, which are applied and calibrated, for example, on battery models. Current sensor values ​​are transmitted from the BCU to the cloud in cloud-connected electric vehicles and used there for various algorithms, such as State of Harm (SOH) determination, or for optimizing charging strategies. Typically, the two sensors are located at different points in the battery pack and measure current using different methods (shunting effect or Hall effect) to prevent external influences from harming the battery. Ideally, the current values ​​measured by these sensors should be identical. However, using two sensors also means that they may display different values ​​due to differences in the resolution and accuracy of the measured parameters, or, for example, one sensor being more aged than the other and therefore its tolerance gradually deviating more from the target tolerance range. Sensor offset errors can also lead to different measurement results. It is impossible to determine which sensor captured the true current value based solely on the two sensors.

[0006] Currently, when the reliability check yields no results or there is no sensor error, the average of the two values ​​is considered for further calculations. Therefore, small inaccuracies are difficult to detect; however, the situation is different when the deviation is large. This leads to increasingly larger inaccuracies in the SOC value, which is directly related to the current value, and thus results in inaccurate determination of the usable driving range. The inaccuracy of the current value also directly affects the determination of the SOH value, and therefore also directly affects the battery replacement time. Summary of the Invention

[0007] This invention relates to a method for selecting a current sensor from at least two current sensors mounted in a battery structure that provides a relatively accurate current measurement change curve, said relatively accurate current measurement change curve being used as an input parameter for model calculations within the cloud, the method comprising the following steps: a) The current measurement curves I_Senor 1 and I_Senor 2, measured by at least two current sensors, are transmitted from the main BCU to the preprocessing stage, where current sensor identification is performed. b) Continuously compare the current measurement change curves I_Senor 1 and I_Senor 2 of at least two current sensors. c) Input the measured current change curves I_Senor 1 and I_Senor 2 from at least two current sensors to the battery model, which determines the battery voltage change curves for each stage based on the measured current change curves I_Senor 1 and I_Senor 2. d) Compare the battery voltage change curve obtained according to method step c) with the current measurement change curves I_Senor 1 and I_Senor 2 measured by at least two current sensors. e) Use the measured current change curves I_Senor 1 and I_Senor 2 from the two current sensors to perform further model calculations in the cloud, and obtain the battery voltage change curve determined according to step c) based on the current sensors, wherein the battery voltage change curve has minimal deviation from the measured voltage change curve.

[0008] Furthermore, in an advantageous embodiment of the method proposed according to the invention, the measured current measurement change curves I_Senor 1, I_Senor 2, battery temperature T, and battery voltage U are supplied to the pretreatment stage via the main BCU according to method step a).

[0009] Furthermore, in the method proposed according to the invention, it is advantageously configured that the execution of method steps a) to e) is performed before the model calculation for determining SOH (State of Health).

[0010] Furthermore, in the method proposed according to the invention, it is advantageously configured that, according to step b), the current measurement change curves I_Senor 1 and I_Senor 2 of at least two current sensors are divided into static and dynamic stages. Thus, in the case of electric vehicle operation and in the case of charging the electric vehicle, the different effects of the dynamic process can be advantageously separated and evaluated.

[0011] Furthermore, in the method proposed according to the present invention, the battery voltage change curve is simulated in the battery model according to the static and dynamic stages of the measured current measurement change curves I_Senor 1 and I_Senor 2, in step c) of the method.

[0012] In an advantageous improvement of the method proposed according to the invention, step d) involves comparing the battery voltage change curve determined by means of a battery model with the measured voltage change curve.

[0013] Furthermore, in the method proposed according to the present invention, the comparison is performed according to method step d) to obtain: - The current measurement curves I_Senor 1 and I_Senor 2 measured by at least two current sensors deviate from each other by, for example, a threshold of 50 mA, or - The high current indicator shows too high and the low current indicator shows too low, or - At least two current sensors display the same measured current change curves I_Senor 1 and I_Senor 2.

[0014] Furthermore, in the method proposed according to the present invention, the voltage simulated in the battery model and the measured voltage change curves of at least two current sensors are gradually adapted to the current curve when deviations occur, according to method step d).

[0015] Furthermore, in the method proposed according to the invention, it is advantageously set that, taking into account the charging and discharging capacity, especially during the charging phase of an electric vehicle, the percentage increase or decrease causes, after several iterations, the voltage change curve to be consistent with the delivered or removed charge carrier, at least to the optimal value that causes minimal deviation.

[0016] Furthermore, the key feature of the method proposed according to the present invention is that it evaluates and detects progressive errors for at least two current sensors, and in particular, it enables prediction of the failure time or remaining lifetime of one or both of the at least two current sensors.

[0017] Furthermore, in the method proposed according to the invention, it is advantageously configured to perform a model calibration on the battery model in the cloud based on a comparison of the measured and simulated battery voltage change curves of the more accurate current sensor among at least two current sensors with the measured current value change curves I_Senor 1 and I_Senor 2, the model calibration tracking the aging characteristics of at least two current sensors and being taken into account in the calculation.

[0018] Finally, it is advantageous in the method proposed according to the invention that, in cases where one of the at least two current sensors has been repeatedly proven to provide more accurate current measurement change curves I_Senor 1, I_Senor 2, the data transmission from the main BCU to the cloud is reduced to the preferred current sensor among the two, at least two current sensors, and the data transmission rate is reduced.

[0019] Furthermore, the present invention relates to a method for using the aforementioned method to detect and determine the inaccuracy of a current sensor in order to improve the accuracy of a computational model.

[0020] Advantages of this invention: According to the present invention, a method is proposed that enables the detection of a sensor with greater inaccuracy from two sensors, particularly a current sensor. Furthermore, the object of the present invention is to determine the sensor inaccuracy and to provide confidence levels for parameters calculated from the corresponding current sensor up to the point of replacement.

[0021] The method proposed according to the present invention enables the early identification of progressive current sensor inaccuracies. Furthermore, a feasible scheme is provided for selecting a more accurate current sensor among current sensors, based on the measurements of the more accurate current sensor, enabling more precise model calculations to determine the SOC or SOH values, and the model calculations provide significantly more accurate results.

[0022] Furthermore, the method proposed according to the present invention enables the determination of sensor inaccuracies, which in turn enables the evaluation of the corresponding sensor, i.e., sensor scoring.

[0023] In the future, after evaluation and verification, only the more accurate values ​​of the current sensors will be transmitted to the cloud, which means significant data simplification.

[0024] The method proposed according to the invention advantageously eliminates the need for reference curves, reference tables, and families of reference characteristic curves. Furthermore, data selection enables significantly better model calibration and tracking of sensor aging occurring during operation within the corresponding model.

[0025] Furthermore, the method proposed according to the invention is suitable for integration in service environments. The method proposed according to the invention can be advantageously used for mobile batteries, as well as for stationary batteries, or batteries from consumer applications with cloud connectivity.

[0026] The method proposed according to the invention has the potential to significantly contribute to improving the accuracy of calculations dependent on current sensor values. It also enables significant improvements in model calibration, particularly in calibrating stored battery models, which further tracks changes in signal characteristics (offset / dynamics / aging). The method proposed according to the invention can also detect and evaluate progressive errors up to the predicted failure time point. Attached Figure Description

[0027] Embodiments of the present invention will be described in detail with reference to the accompanying drawings and the following description. Wherein: Figure 1 An exemplary, schematic structure of a battery pack with first and second current sensors is shown. Figure 2 A schematic diagram of the method proposed according to the present invention, performed in the cloud, is shown, and Figure 3 It shows Figure 2 The flowchart of the method is shown schematically in the figure. Detailed Implementation

[0028] Depend on Figure 1 It is understood that the battery structure 10 has first to sixth switches 16-26. According to... Figure 1 The battery structure 10, in addition to the main BCU 12, also includes a first branch line 30, a second branch line 32, and a third branch line 34. The first branch line 30 is connected via a first switch 16 and a fourth switch 22, while the second branch line 32 is connected via a second switch 18 and a fifth switch 24. Finally, according to... Figure 1 The third branch 34 of the battery structure 10 is connected via a third switch 20 and a sixth switch 26. The battery pack 28 is indicated by dashed lines. Within the battery pack 28, for example, a first battery module 36 and a second battery module 38 are connected in parallel. It is also possible to install a larger or smaller number of battery modules within the battery pack 28; the number is merely conceived as an example.

[0029] In addition, from the perspective of Figure 1 As can be seen in the diagram, for example within the range of the second battery module 38 connected in parallel, each battery module shown includes a certain number of individual battery cells.

[0030] Each of the parallel-connected battery modules 36 and 38 is equipped with a modular BCU 52. Similarly, this applies to... Figure 1 The diagram shows the parallel connection of each battery module within the first battery module 36. Each module-type BCU 52, belonging to modules 40, 42, 44, 46, 48, and 50, is respectively connected to... Figure 1 The main BCU12 of the battery structure 10 shown in the diagram is connected via control circuitry.

[0031] In addition, from the perspective of Figure 1 As can be seen in the diagram, the battery structure 10 is equipped with a first current sensor 60 and a second current sensor 70.

[0032] Embodiments of the present invention: In the following description of embodiments of the invention, the same or similar elements are designated by the same reference numerals, and in some cases, repeated descriptions of these elements are omitted. The drawings are merely schematic representations of the subject matter of the invention.

[0033] from Figure 2 A schematic diagram of the method proposed according to the present invention can be seen in the diagram.

[0034] Within the cloud 100, various calculation or determination steps are executed, with input parameters transmitted to the cloud via the main BCU 12. The main BCU transmits the measured current value change curves I_Senor 1 84 and I_Senor 2 86, along with battery temperature 80 and battery voltage 82, as input parameters to the cloud 10, specifically to the preprocessing stage 88. The output parameters of the preprocessing stage are then transmitted to the model within the cloud 100 used for SOH determination 108 (State of Health). After the preprocessing stage 88 runs, more precise values ​​from at least two current sensors 60 and 70 are transmitted to the model used for SOH determination 108, enabling accurate calculations.

[0035] The preprocessing stage 88 includes current sensor identification 90 and preprocessing of data transmitted in the framework of data transmission 122, wherein a battery model 94 and an evaluation unit 96 are stored in the preprocessing stage.

[0036] In preprocessing stage 88, the two measured current change curves I_Senor 1 84 and I_Senor 2 86 of the data packet are continuously compared with each other, and the more accurate current sensor among at least two current sensors 60 and 70 is identified 90. The comparison of the measured current change curves I_Senor 1 84 and I_Senor 2 86 can lead to one of the following results: The measured current measurement curves I_Senor 1 84 and I_Senor 2 86 deviate from each other by more than a threshold, such as 50 mA, or the deviation is not constant throughout the measurement range, such as high current being displayed too high and low current being displayed too low, or the two current sensors 60 and 70 not showing a deviation. Therefore, either there is an offset deviation between the two current sensors 60 and 70, or the two current sensors 60 and 70 are measuring the correct current value.

[0037] To distinguish between inaccuracies in offset and inaccuracies in the dynamic sensor characteristics of current sensors 60 and 70, the measured current value change curves I_Senor 1 84 and I_Senor 2 86 are divided into static and dynamic phases. These static and dynamic phases already exist and have been evaluated in the current data package or one of the previous data packages. Dynamic current change curves are generated during driving operations, while static current change curves are generated during the current constant phase, as occurs during the charging process of the vehicle battery. Whether fast charging with high current values ​​or graded charging curves, or constant current constant voltage charging (CCCV-Laden) with a constant current and a hyperbolic-like decreasing current down to very low currents, the battery voltage change curves are simulated based on these data and compared with the measured voltage change curves, especially the measured battery voltage.

[0038] It is also possible to consider using a constant current for a predefined discharge, as in V2X systems (V2L, V2H, V2B, or V2G), to determine the accuracy of the sensor.

[0039] The two measured current change curves I_Senor 1 84 and I_Senor 2 86, the battery temperature 80, and the battery voltage 82 are fed to the battery model 94 in the preprocessing stage 88. The battery model 94 can be an equivalent circuit diagram model or an electrochemical model, which determines the individual battery voltage change curves for each stage based on the two transmitted measured current change curves I_Senor 1 84 and I_Senor 2 86. Based on the results, the values ​​are compared with the measured voltage change curves 128 (see [reference to...]). Figure 3(See diagram). The calculated battery voltage change curve is obtained from the battery model 94 based on the measurements of current sensors 60 and 70. The battery voltage change curve that is closest to the measured voltage change curve, that is, the one with the smallest deviation, is continued to be used in all further calculations, especially as I_Senor X 102 for SOC determination according to model 106. In addition, this more accurate voltage change curve with the smallest deviation is fed to the sensor prediction model 104, the SOC determination model 106, and especially the SOH determination model 108 and the SOH prediction model 110 implemented in the cloud 100. In the previous context, the term "closest" means that the deviation is the smallest during the corresponding voltage change curve stage, wherein possible outliers of the measured current measurements I_Senor 1 84 and I_Senor 2 86 have been eliminated in advance.

[0040] Based on the comparison 128 outlined above (see by...), Figure 3 (See diagram). The simulated voltage and the measured voltage process of at least two current sensors 60 and 70 are gradually fitted to the current curve in such a way that the deviation is above the threshold until the deviation between the measured and simulated voltage change curves can no longer be determined.

[0041] The equality of charging or discharging capacity can be considered as an additional condition. The optimal condition for this is one with… Figure 1 The battery structure 10 shown in the diagram represents the charging phase of an electric vehicle, and here in particular the constant current phase, which can occur, for example, in the case of fast charging, in the case of charging in the constant current phase during constant current and constant voltage charging, or in the case of charging with a graded curve.

[0042] Starting from the current curve, a percentage increase or decrease is made, and after several iterations, the optimal value is obtained that, under the best conditions, causes the voltage change curve to be consistent with the amount of charge delivered / removed, but at least causes the minimum deviation.

[0043] Regarding percentage deviation, at least two current sensors 60 and 70 are clustered or evaluated based on the results. The resulting error can be quantified and compared with pre-given parameters, such as those in technical customer data.

[0044] Using the method proposed according to the present invention, progressive errors can also be detected based on inaccuracies occurring in the past over time, and it can be predicted when the relevant current sensors 60, 70 are outside their predetermined specifications and should be replaced. Therefore, the failure time 112 or remaining lifespan can also be predicted using the method proposed according to the present invention. Accuracy parameters are important for further calculations within the cloud 100, as they indicate the confidence level of the corresponding parameters.

[0045] The method proposed according to the invention can also be used in dynamic current change curves, especially during the discharge phase of electric vehicles, that is, during the operation of electric vehicles. However, when the deviation is not constant throughout the measurement range and, for example, the high current is too high and the low current is too low, seeking the minimum deviation becomes more complicated.

[0046] It should be noted that comparing the calculated battery voltage change curve based on the input data of the more accurate current sensor among at least two current sensors 60 and 70 with the measured voltage change curve enables model calibration, especially for battery model 94 (see [link to relevant documentation]). Figure 3 (Position 140 in the model). This allows for continuous tracking of the aging characteristics of the corresponding current sensors 60 and 70 in the model.

[0047] If one of the two current sensors 60 and 70 used has been proven multiple times to be the preferred current sensor 60 or 70 for further calculations, then the data transmission 122 of the current sensors 60 and 70 from the main BCU 12 to the cloud 100 can be reduced to the preferred current sensor of the two current sensors 60 and 70, thereby reducing the data rate to be transmitted. However, the complete transmission of the measured current values ​​I_Senor 1 84 and I_Senor 2 86 of the inferior current sensor of the two current sensors 60 and 70 should be omitted. After a certain period of time, such as once a week, or if an anomaly or even a malfunction is identified in the preferred current sensor of the current sensors 60 and 70, it is necessary to transmit these sensor values ​​to the cloud 100 in the future.

[0048] The deterioration or premature aging of the assumed better current sensor among the two current sensors 60 and 70 due to environmental factors will also lead to the current sensor 60 and 70 no longer being considered for further calculations, and the value of the current sensor 60 and 70 that has been proven to be the assumed worse sensor will be used instead.

[0049] The method proposed according to the invention makes a significant contribution to the accuracy of calculations dependent on the measured current sensor values ​​I_Senor 1 84 and I_Senor 2 86. Improvements are also possible in model calibration, which further allows tracking of changes in signal characteristics in terms of offset, dynamics, and aging. The method proposed according to the invention enables the detection and scoring of progressive errors up to the determination of the predicted failure time point, i.e., the remaining failure time 112.

[0050] Figure 3 A flowchart illustrating a feasible method according to the present invention is shown schematically, starting from startup step 120. Following startup step 120, data transmission 122 is performed from the main BCU 12 to the cloud 100, wherein, in particular, the measured current measurement change curves I_Senor 1 84, I_Senor 2 86, battery temperature 80, and battery voltage 82 are transmitted to the cloud 100 or to the computational models 94, 104, 106, 108, and 110 existing there. The current curves are then partitioned, i.e., divided 124 into static and dynamic phases. Subsequently, the voltage change curves are modulated, and 126 the capacity change in terms of capacity increase or decrease is determined. Within the framework of comparisons already mentioned several times, the modulated battery voltage is compared with the measured voltage, and the modulated capacity is compared taking into account the measured voltage or the voltage determined in the model. Following this comparison 128, a query 130 is performed to determine whether the difference exceeds a threshold. If the query 130 is denied, then the average of the values ​​from the two sensors 60 and 70 is used to form 132 for further calculations in the cloud, specifically using I_Senor X102 to determine the SOH within model 108.

[0051] Conversely, if the above inquiry 130 is affirmed, then sensor identification 134 is performed in terms of whether the sensor is "good" or "bad," that is, the first current sensor 60 or the second current sensor 70 is evaluated.

[0052] Next, sensor error determination 136 can be performed, and after sensor identification 134, the "good" current sensor values ​​138 and 140 are used as a basis for further calculations and also considered for model calibration or aging tracking or monitoring. Time determination is performed in reference numeral 142, that is, once the specifications of the corresponding current sensor 60 or 70 are exceeded, the remaining failure time 112 until the current sensor 60 or 70 fails is determined.

[0053] After the timing routine 142 completes, the obtained value is transmitted to the main BCU 12 via data transmission 144, and then the termination step 146 is performed.

[0054] This invention is not limited to the embodiments described herein and the aspects highlighted therein. Rather, various modifications that are within the scope of the claims and are of skill to those skilled in the art can be implemented.

Claims

1. Method for selecting a current sensor (60, 70) from at least two current sensors (60, 70) mounted at a battery structure (10) which provides a relatively accurate current measurement profile I_Sensor 1 (84), I_Sensor 2 (86) as an input quantity as a basis for model calculations (94, 104, 106, 108, 110) within a cloud (100), the method having the following method steps: a) Data transmission (122) of current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86) measured by the at least two current sensors (60, 70) to a pre-processing stage (88) by means of a main BCU (12), in which a current sensor qualification (90) is carried out, b) Continuous comparison of the measured current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86) of the at least two current sensors (60, 70), c) Delivery of the measured current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86) of the at least two current sensors (60, 70) to a battery model (94), which determines a battery voltage profile on the basis of the measured current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86) in a determination phase, d) Comparison (128) of the battery voltage profile determined in accordance with method step c) with a measured voltage profile, e) Use of the current measurement profile I_Sensor 1 (84), I_Sensor 2 (86) of the current sensor of the at least two current sensors (60, 70) on the basis of which the battery voltage profile determined in accordance with method step c) having the smallest deviation with respect to the measured current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86) for further calculations of a model (104, 106, 108, 110) in the cloud (100). The current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86), the battery temperature T (80) and the battery voltage U (82) are delivered to the pre-processing stage (88) within the cloud (100) by means of the main BCU (12) in accordance with method step a). The execution of the method steps a) to e) is provided before a model calculation (108) for SOH (State of health) determination. The measured current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86) of the at least two current sensors (60, 70) are divided into a static phase and a dynamic phase in accordance with method step b). The measured current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86) of the at least two current sensors (60, 70) are delivered to a battery model (94) in accordance with method step c). ​ 2. The method of claim 1, wherein, ​ 3. The method according to claims 1 and 2, characterized in that, ​ 4. The method according to claims 1 to 3, characterized in that, ​ 5. The method according to claims 1 to 4, characterized in that, According to method step c) the battery voltage profile is simulated in the battery model (94) from the measured current measurement progress I_Sensor 1 (84), I_Sensor 2 (86) in the static and dynamic phase.

6. The method according to claims 1 to 5, characterized in that, According to method step d) the battery voltage profile determined in the battery model (94) is compared (128) with the measured voltage profile.

7. The method according to claims 1 to 6, characterized in that, According to method step d) the comparison (128) leads to: - the current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86) of the at least two current sensors (60, 70) deviate from each other by a threshold value of e.g. 50 mA, or - the high current shows too high and the low current shows too low, or - the at least two current sensors (60, 70) show the same current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86).

8. The method according to claims 1 to 7, characterized in that, According to method step d) the current profile is gradually adapted in case of a deviation (130) of the simulated voltage, in particular the battery voltage of the dynamic and static phase of the at least two current sensors (60, 70) and the measured voltage profile from the battery model (94) and the measured voltage profile.

9. The method according to claims 1 to 8, characterized in that, The increase or decrease of the percentage leads to a conformity of the voltage profile with the amount of charge delivered or taken, at least to an optimum of minimal deviation after several iterations, taking into account the charged and discharged capacity, in particular during the charging phase of an electric vehicle.

10. The method according to the preceding claim, characterized in that, The evaluation is carried out on the at least two current sensors (60, 70) and a progressive error is detected, in particular a prediction (142) is made in terms of the expected time of failure (112) of one or both of the at least two current sensors (60, 70).

11. The method according to claims 1 to 10, characterized in that, Based on the comparison (128) of the battery voltage profile determined from the more precise current sensor of the at least two current sensors (60, 70) with the measured current measurement profiles I_Sensor 1 (84), I_Sensor 2 (86) a model calibration, in particular of the battery model (94) of the cloud (100) is carried out, which tracks the aging characteristics of the at least two current sensors (60, 70) and is taken into account when calculating in the cloud (100).

12. The method according to claims 1 to 11, characterized in that, In case one of the at least two current sensors (60, 70) has repeatedly proven to provide a more precise current measurement progress I_Sensor 1 (84), I_Sensor 2 (86) the data transmission (122) from the main BCU (12) to the cloud (100) is reduced to the preferred current sensor of the at least two current sensors (60, 70) and the transmission data rate is reduced.

13. Use of the method according to any one of claims 1 to 12 for detecting and determining the inaccuracy of a current sensor (60, 70) in order to improve the accuracy of a calculation model (94; 104, 106, 108, 110).

Images