Method and apparatus for estimating internal resistance of a battery cell

Abstract

The invention relates to a method for estimating an internal resistance (20) of a battery cell, the method comprising the following measures: detecting a voltage (10) at the battery cell; detecting a current (11) at the battery cell; determining a detected voltage differential (15) by differentiating the detected voltage (10) by means of a filter; determining a modeled voltage differential (13) by differentiating by means of a filter from the detected current (11) and a current internal resistance estimate; determining a correction factor (17) from the detected voltage differential (15) and the modeled voltage differential (13); and estimating a new current internal resistance estimate from the determined correction factor (17) and a previous current internal resistance estimate; providing the new current internal resistance estimate as an estimated internal resistance (20) of the battery cell. The invention also relates to a device (1) for estimating an internal resistance of a battery cell.

Description

Technical Field

[0001] This invention relates to a method and apparatus for estimating the internal resistance of a single battery cell. Background Technology

[0002] As battery cells age over time, this negatively impacts performance. Aging is particularly manifested in the increase in internal resistance and the decrease in capacitance of the cells. Cell aging can be tracked by examining internal resistance. However, internal resistance cannot be directly measured; it can only be estimated from the voltage at the cell and the current flowing through it. However, various factors can make this estimation difficult: measurement noise, imbalances in internal resistance during charging and discharging, the RC-like behavior of the cell, and the fact that the equivalent circuit models typically used in modeling do not perfectly reflect the electrochemical processes within the cell. Summary of the Invention

[0003] The present invention is based on the following objective: to create a method and apparatus for estimating the internal resistance of a single battery cell, wherein the internal resistance of the single battery cell can be reliably estimated using the method and apparatus.

[0004] According to the present invention, the task is solved by a method having the features of the present invention and an apparatus having the features of the present invention.

[0005] In particular, a method for estimating the internal resistance of a battery cell is provided, the method comprising: detecting the voltage at the battery cell; detecting the current at the battery cell; determining a differential portion of the detected voltage by differentiating the detected voltage using a filter; determining a modeled differential portion of the voltage by differentiating the detected current and a current internal resistance estimate using a filter; determining a correction coefficient by the detected differential portion of the voltage and the modeled differential portion of the voltage; and estimating a new current internal resistance estimate by the determined correction coefficient and a previous current internal resistance estimate; and providing the new current internal resistance estimate as an estimated internal resistance of the battery cell.

[0006] Furthermore, a device for estimating the internal resistance of a battery cell has been created, the device comprising: an interface configured to receive a voltage and a current detected at the battery cell; and a control device configured to determine a differential portion of the detected voltage by differentiating the detected voltage with the aid of a filter; determine a modeled differential portion of the voltage by differentiating the detected current and a current internal resistance estimate with the aid of a filter; determine a correction coefficient by the detected differential portion of the voltage and the modeled differential portion of the voltage; estimate a new current internal resistance estimate by the determined correction coefficient and a previous current internal resistance estimate; and provide the new current internal resistance estimate as an estimated internal resistance of the battery cell.

[0007] The method and apparatus enable training of the internal resistance of a battery cell, particularly starting from a reference value. The training is based, in particular, on a differentially determined internal resistance, i.e., the internal resistance is determined primarily as a differential internal resistance. The reference value here initially corresponds primarily to the nominal internal resistance of the battery cell. Specifically, the reference value is set as a first current internal resistance estimate and then updated to a new current internal resistance estimate in each real-time iteration using the measures of the method. The update is performed as follows: a modeled voltage is estimated from the detected current and the current internal resistance estimate. Differentials are formed from the detected voltage and the modeled voltage by differentiation using a (differentiated) filter. This is based on the idea that the internal resistance can be determined and trained particularly well precisely when the voltage (or current) changes. A correction coefficient is determined from the ratio of the differentials relative to each other. A new current internal resistance estimate is estimated from the determined correction coefficient and the (previous) current internal resistance estimate. The new current internal resistance estimate is provided as an estimated internal resistance of the battery cell, particularly in the form of an analog or digital signal.

[0008] The advantages of the method and apparatus are that they can reliably determine the internal resistance of a single battery cell. The use of a filter for differentiation enables not only the determination of the differential part but also the reduction of measurement noise.

[0009] The method is performed repeatedly, in particular, to continuously obtain the current internal resistance estimate and in this way to understand the aging of the battery cell through the estimated internal resistance.

[0010] Voltage and current are detected, especially with the help of sensors designed for this purpose. Sensors can also be part of the device.

[0011] The device, particularly the control unit, can be constructed individually or in combination as a combination of hardware and software, for example, as program code that executes on a microcontroller or microprocessor. Alternatively, the portion can be configured individually or in combination as an application-specific integrated circuit (ASIC) or a field-programmable gate array (FPGA).

[0012] The method is explained below based on the formula. It is assumed that changes in current or load cause changes in the detected voltage, and that these changes are related to the internal resistance of the battery cell. R It is directly proportional to 0. Internal resistance can be determined using the following method for differential resistance. R diff The formula is determined as follows: in, dU For the voltage differential, and dI This is the differential part of the current.

[0013] With the help of a differential filter FILTER (·), determined by the detected voltage U and the detected current I Determine the voltage derivative of the detected voltage and the voltage derivative of the modeled voltage: If a jump occurs in the current, the following must be applied: The correction factor can then be determined by the ratio of the voltage differentials. α k : Here, In particular, the nominal internal resistance of the battery cell is used as a reference value, that is, especially the internal resistance indicated by the battery cell at the time of delivery. This is the actual internal resistance of the battery cell. This actual internal resistance is estimated using the formula described above and provided as a current estimate of the internal resistance.

[0014] Under ideal conditions, new correction coefficients can be obtained for each subsequent iteration using the following method. f .

[0015] In this way, for each iteration step, the current internal resistance estimate can be estimated from the detected voltage and the detected current.

[0016] In one embodiment, the filter is configured as an asymmetric filter with a finite impulse response. This filter is particularly suitable for achieving not only differentiation but also noise reduction. Noise reduction is achieved, in particular, through the low-pass characteristics of one or more filters. The asymmetric filter with a finite impulse response is specifically designed such that it differentiates at low frequencies and reduces noise (low-pass characteristics).

[0017] In one embodiment, the correction coefficients are estimated using at least one estimation filter when determining them. This reduces short-term fluctuations. In particular, the at least one estimation filter has a smoothing effect over time.

[0018] In an improved embodiment, the at least one estimation filter is configured as a Kalman filter. The advantage is that, in addition to the estimation result, the Kalman filter can additionally provide the confidence level of the estimation result, thereby providing additionally reliable conclusions regarding the estimated internal resistance of the battery cell.

[0019] In one embodiment, at least one voltage derivative is compared to a preset threshold, wherein when the value of one of the voltage derivatives reaches or exceeds the preset threshold, measures for determining the estimated internal resistance are triggered. This has the advantage that the measures for determining the new current internal resistance are performed (only) under ideal conditions for training the internal resistance of the battery cell, i.e., when there are large voltage and / or current changes that result in a good signal-to-noise ratio. For example, a suitable threshold can be determined empirically and / or through simulation.

[0020] In one embodiment, the power flow direction is distinguished, wherein the current direction is determined for this purpose, and charging internal resistance estimates and / or discharging internal resistance estimates are determined based on the determined current direction, and the charging internal resistance estimates and / or discharging internal resistance estimates are provided as the internal resistance of the battery cell related to the power flow direction. Thus, the internal resistance can be determined based on the power flow. This approach is essentially similar to the above approach for both cases. However, it is specifically configured to determine a current internal resistance estimate for each power flow direction (charging / discharging). In particular, the power flow direction is determined, especially based on the current direction (at the corresponding voltage). Thus, even when a change in the sign of the current occurs, one or more correction coefficients can be trained. For example, if a jump in current from -100A to 50A occurs at a voltage with constant polarity, the first two-thirds of this jump can be assigned to the negative power flow direction (e.g., charging), and the latter one-third can be assigned to the positive power flow direction (e.g., discharging). The first two-thirds of the signal jump is used to estimate the internal resistance under negative power flow conditions, and the latter one-third is used to estimate the internal resistance under positive power flow conditions. For example, the individual values ​​of an asymmetric filter with a finite impulse response can be evaluated accordingly. For this purpose, the asymmetric filter with a finite impulse response is specifically positioned in time at the midpoint of the signal transition. Since the coefficients are negative in the first part of the asymmetric filter and positive (or vice versa) in the second part, the corresponding portions can be used to determine the moving average of the current values ​​before and after the signal transition in the case of power flow transformation. The moving average can reduce signal noise, thereby improving the signal-to-noise ratio.

[0021] Other features of the device design are derived from the description of the method design. The advantages of the device are the same as those described in the method design.

[0022] Furthermore, a means of transport is specifically proposed, comprising at least one device according to one of the described embodiments. The means of transport is particularly a motor vehicle. However, in principle, the means of transport can also be other land, rail, water, air, or space vehicles. In principle, the method and device can also be used in other mobile or stationary installations. Attached Figure Description

[0023] The present invention will now be explained in more detail with reference to the accompanying drawings, based on preferred embodiments. Herein: Figure 1 A schematic diagram illustrating one embodiment of a device for estimating the internal resistance of a single battery cell; Figure 2a A schematic graph of the detected current over time is shown; Figure 2b A schematic graph of the modeled voltage differential over time is shown; Figure 3 A schematic diagram illustrating one embodiment of the method is shown; Figure 4a , 4b A schematic line diagram is shown to illustrate the methods described in this disclosure; Figure 5a , 5b A schematic line diagram illustrating one implementation of the method is shown. Detailed Implementation

[0024] exist Figure 1 The diagram shows a schematic representation of one embodiment of a device 1 for estimating the internal resistance 20 of a single battery cell. Device 1 is configured to implement the method described herein. The method is explained with reference to device 1.

[0025] Device 1 includes interface 2, which is configured to receive voltage 10 and current 11 detected at a battery cell. A combined interface 2 may also be used. Voltage 10 is detected using voltage sensor 50 at the battery cell. Current 11 is detected using current sensor 51 at the battery cell.

[0026] In addition, device 1 also includes a control device 3. Control device 3 includes a computing device 4 and a memory 5. For example, computing device 4 is a microprocessor or microcontroller that executes program code to implement the methods described in this disclosure. However, hardware components with fixed wiring may also be provided, which partially or completely implement the methods. Device 1 may also be part of a battery controller (not shown).

[0027] The control device 3 is configured to determine the detected voltage derivative 15 by differentiating the detected voltage 10 using a filter. Specifically, the filter is an asymmetric filter with a finite impulse response. Furthermore, the control device 3 is configured to determine the modeled voltage derivative 13 by differentiating the detected current 11 and the current internal resistance estimate 9 using a filter. Specifically, the filter is an asymmetric filter with a finite impulse response.

[0028] Furthermore, the control device 3 determines the correction coefficient 17 using the detected voltage differential 15 and the modeled voltage differential 13. Additionally, the control device 3 estimates a new current internal resistance estimate 9 using the determined correction coefficient 17 and the previous current internal resistance estimate 9, and provides the new current internal resistance estimate 9 as an estimated internal resistance 20 for the battery cell, for example, as an internal resistance signal 21, for example, at the interface 6. The estimated internal resistance 20 or the internal resistance signal 21 can be supplied to the battery controller 52.

[0029] It can be configured such that when the correction coefficient 17 is determined, the correction coefficient is estimated using at least one estimation filter.

[0030] As an improvement, the at least one estimation filter can be configured as a Kalman filter.

[0031] It can be configured to compare at least one of the voltage derivatives 13 and 15 with a preset threshold 14, wherein measures for determining the estimated internal resistance 20 are triggered when the value of one of the voltage derivatives 13 and 15 reaches or exceeds the preset threshold 14.

[0032] This is symbolic in Figure 2a and 2b This is explained in the text. Figure 2a The current 11 (in A) detected over time 12 in seconds is shown. Figure 2b The absolute value (in volts per second) of the voltage derivative 13 modeled over time 12, measured in seconds, is shown. This voltage derivative is determined by the detected current 11 of the battery cell and the current internal resistance estimate, obtained by differentiation using a filter, particularly an asymmetric filter with a finite impulse response. At approximately 150 seconds, the absolute value of the modeled voltage derivative 13 exceeds a preset threshold 14. After reaching or exceeding the preset threshold 14, measures are performed to determine the estimated internal resistance 20, thereby providing an updated estimated internal resistance 20 of the battery cell. Figure 1 Therefore, in order to train the current internal resistance 20 in terms of signal-to-noise ratio, advantageous values ​​of voltage 10 and current 11 can be used.

[0033] This implementation method is in Figure 3The flowchart is illustrated as an illustration. For example, in control device 3 ( Figure 1 In this context, modules 100 to 103 are constructed for this purpose. Figure 3 In module 100, the detected current 11 is converted into a voltage using the current internal resistance estimate. This voltage is then converted into a modeled voltage derivative 13 using a differential filter, particularly an asymmetric filter with a finite impulse response. This modeled voltage derivative 13 is compared in module 101 with a preset threshold (see also...). Figure 2b If the threshold comparison in module 101 is reached or exceeded, then measures to determine a new current internal resistance estimate or an updated internal resistance 20 are initiated in module 103; otherwise, the threshold comparison with the current value for the detected current is repeated continuously.

[0034] To implement the aforementioned measures in module 103, a detected voltage differential 15 is supplied to module 103. This voltage differential is generated in module 102 by means of a differentiating filter, particularly an asymmetric filter with a finite impulse response, from the detected voltage 10. Furthermore, a modeled voltage differential 13 is supplied to module 103. Then, a (new and updated) correction coefficient is trained in module 103, which allows the estimation of the current internal resistance 20 of the battery cell.

[0035] It can be configured to distinguish the direction of power flow. For this purpose, control device 3 ( Figure 1 The current direction is determined. Based on the determined current direction, the control device 3 determines the estimated internal resistance during charging and / or the estimated internal resistance during discharging, and provides these as the internal resistances 20l and 20e (charging and discharging) of the battery cells in relation to the power flow direction.

[0036] exist Figure 4a and 4b The diagram shows a schematic line graph used to illustrate the method described in this disclosure. It shows the trend of the detected voltage derivative 15 and the voltage derivative 16 determined by means of the nominal internal resistance of the battery cell (as a reference internal resistance or initial internal resistance) over time 12 in seconds, and the trend of the modeled voltage derivative 13 (in V) over time 12 in seconds. It can be seen that at the beginning of the method, i.e., after approximately 160 seconds ( Figure 4a The modeled voltage derivative 13 initially matched the voltage derivative 16 but deviated from the detected voltage derivative 15, particularly at the peak. Conversely, after approximately 2270 seconds, the modeled voltage derivative 13 became consistent with the detected voltage derivative 15, primarily through adaptation of the current internal resistance estimate via correction coefficients. Therefore, the training effect is clearly evident.

[0037] exist Figure 5a and Figure 5b The diagram shows a schematic line drawing illustrating one embodiment of the method. It is configured to distinguish the power flow direction, wherein a current direction is determined for this purpose, and charging internal resistance estimates and / or discharging internal resistance estimates are determined based on the determined current direction, and these are provided as the internal resistance of the battery cell in relation to the power flow direction. Furthermore, in this embodiment, when determining correction coefficients 17l and 17e, the correction coefficients are estimated using a corresponding estimation filter, wherein the estimation filters are respectively constructed as Kalman filters.

[0038] exist Figure 5a The correction factors 17l and 17e for the estimated charging internal resistance over time 12 in seconds are shown. Figure 5b The correction factors 17l and 17e for the estimated discharge internal resistance over time 12 in seconds are shown. It can be seen that the values ​​of the correction factors 17l and 17e for the estimated charging internal resistance and the estimated discharging internal resistance are different from each other. Accordingly, the estimated charging internal resistance determined by correction factor 17l and the estimated discharging internal resistance determined by correction factor 17e are also different from each other.

[0039] Furthermore, it is evident that, starting from the initial values ​​(=1) of the correction coefficients 17l and 17e, the corresponding values ​​gradually approach the value of the corresponding fundamental truth 19 over time.

[0040] The method and apparatus enable reliable estimation of the internal resistance of individual battery cells. A particular advantage is that correction coefficients can be trained whenever the signal-to-noise ratio is suitable. Furthermore, the internal resistance related to power flow can be estimated and provided, allowing for differentiation between internal resistance during charging and internal resistance during discharging.

[0041] List of reference numerals 1 device 2 interfaces 3 Control devices 4 Computing Device 5. Memory 6 interfaces 7. Current internal resistance estimate 10. Detected voltage 11. Detected current 12 hours 13. Voltage Differential Modeling 14. Preset threshold 15. Voltage differential detected 16. Voltage Differential (Reference Internal Resistance) 17 Correction coefficients 17L correction factor (charging) 17e correction factor (discharge) 19 Basic Truths 20 internal resistance 20L internal resistance (charging) 20e internal resistance (discharge) 21 internal resistance signal 50 Voltage Sensor 51 Current Sensor 52 Battery Controller Modules 100-103

Claims

1. A method for estimating the internal resistance (20) of a single battery cell, the method comprising the following measures: Detect the voltage at the individual battery cell (10); Detect the current at the individual battery cell (11); The differential part (15) of the detected voltage is determined by taking the differential of the detected voltage (10) with the aid of a filter; The voltage differential (13) modeled is determined by taking the differential of the detected current (11) and the current internal resistance estimate using a filter; The correction coefficient (17) is determined by the detected voltage derivative (15) and the modeled voltage derivative (13); and The new current internal resistance estimate is estimated from the determined correction coefficient (17) and the previous current internal resistance estimate; The new current internal resistance estimate is provided as the estimated internal resistance of the battery cell (20).

2. The method according to claim 1, characterized in that, The filter is designed as an asymmetric filter with a finite impulse response.

3. The method according to claim 1 or 2, characterized in that, When determining the correction coefficient (17), the correction coefficient is estimated using at least one estimation filter.

4. The method according to claim 3, characterized in that, The at least one estimation filter is constructed as a Kalman filter.

5. The method according to claim 1 or 2, characterized in that, At least one of the voltage derivatives (13, 15) is compared with a preset threshold (14), wherein when the value of one of the voltage derivatives (13, 15) reaches or exceeds the preset threshold (14), measures for determining the estimated internal resistance (20) are triggered.

6. The method according to claim 1 or 2, characterized in that, Distinguish the power flow direction, wherein the current direction is determined for this purpose and the estimated charging internal resistance and / or the estimated discharging internal resistance are determined based on the determined current direction, and the estimated charging internal resistance and / or the estimated discharging internal resistance are provided as the internal resistance (20) of the battery cell in relation to the power flow direction.

7. An apparatus (1) for estimating the internal resistance (20) of a single battery cell, the apparatus comprising: Interface (2), the interface being configured to receive voltage (10) and current (11) detected at the battery cell; as well as Control device (3), wherein the control device (3) is configured to determine the differential portion (15) of the detected voltage by taking the differential of the detected voltage (10) with the aid of a filter; The voltage differential (13) modeled is determined by differentiating the detected current (11) and the current internal resistance estimate using a filter. The correction coefficient (17) is determined by the detected voltage derivative (15) and the modeled voltage derivative (13); and The new current internal resistance estimate is estimated from the determined correction coefficient (17) and the previous current internal resistance estimate; and The new current internal resistance estimate is provided as the estimated internal resistance of the battery cell (20).

8. The device (1) according to claim 7, characterized in that, The filter is designed as an asymmetric filter with a finite impulse response.

9. The device (1) according to claim 7 or 8, characterized in that, When determining the correction coefficient (17), the correction coefficient is estimated using at least one estimation filter.

10. The device (1) according to claim 9, characterized in that, The at least one estimation filter is constructed as a Kalman filter.

Images