Method and device for determining internal battery states and model parameters of an electrochemical battery model for a device battery of a technical device

Abstract

The invention relates to a computer-implemented method for parameterizing an electrochemical battery model of a device battery (41) in a technical device (4) with updated model parameters for determining an internal battery state, in particular for determining a charging curve, for monitoring for anomalies and / or for determining an aging state, comprising the following steps: - recording (S1) time courses of operating variables of a plurality of device batteries (41); - For a specific device battery (41), selecting (S2) a subset of similar device batteries (41) from the plurality of device batteries that have a similar overall condition to the specific device battery (41); - For the model parameters of the battery models of the subset of device batteries (41), determining (S4) a lower and an upper parameter limit for the respective model parameters depending on the distribution of the values ​​of the model parameters in the subset of device batteries (41) at a calendar age of the specific device battery (41); - performing (S5) a fitting procedure for the battery model to the specific device battery (41) depending on the parameter limits of the model parameters based on the temporal courses of the operating variables in order to obtain the updated model parameters.

Description

Technical area

[0001] The invention relates to the determination of internal battery states and model parameters of an electrochemical battery model for a portable battery, which can be designed, for example, as a "continuum model," using a fitting method. The model parameters can, for example, form the basis for condition monitoring, an aging state determination, and a charging curve calculation for the portable battery. Technical background

[0002] The power supply for off-grid electrical devices and machines, such as electrically powered vehicles, is usually provided by device batteries or vehicle batteries. These provide electrical energy to operate the devices.

[0003] Portable batteries degrade over their service life and depending on their load and use. This so-called aging leads to a continuously decreasing maximum power or storage capacity. The state of aging corresponds to a measure of the aging of energy storage devices. According to convention, a new portable battery can have an aging state (in terms of its capacity, SOH-C) of 100%, which decreases noticeably over the course of its service life. The measure of the aging of a portable battery (change in the state of aging over time) depends on the individual load on the portable battery, i.e. in the case of vehicle batteries, the driver's usage behavior, external ambient conditions and the vehicle battery type.

[0004] Device batteries in off-grid electrical devices, particularly mobile devices, are continuously monitored to detect sudden battery failures (sudden death), thermal fault events, rapid cyclic aging, or other fault events (collectively referred to herein as anomalies) in a timely manner and to warn the user of the occurrence of such anomalies. Such anomalies include the detection of current or future anomalies based on measured temporal patterns of device battery operating parameters. Such anomaly detection is often performed cloud-based.

[0005] To perform anomaly detection that is as cloud-independent as possible, anomaly detection models can be implemented in the technical devices. These models evaluate operating parameters and appropriately signal anomalies in the event of deviations from conventional operating patterns. The anomaly detection models use an electrochemical battery model to determine and evaluate battery states.

[0006] The behavior of portable batteries can generally be modeled using an electrochemical battery model. Such models are characterized by the fact that they establish the relationship between externally measurable battery parameters, especially the terminal voltage, and the internal battery states, e.g., the concentration of lithium in the electrolyte and solid state. The internal battery states result from the assumed model parameters. Typical representatives of this model class are reduced-order electrochemical continuum models, such as Newman models. Electrochemical battery models can be used to determine optimized charging current trajectories (charging curves) for the fastest and most gentle charging of the battery cells with the most consistent life expectancy, for battery cell condition monitoring, for determining an aging state, and so on. Disclosure of the invention

[0007] According to the invention, a method for parameterizing an electrochemical battery model of a device battery in a technical device with updated model parameters for determining an internal battery state according to claim 1 and by the device according to the independent claim are provided.

[0008] Further embodiments are specified in the dependent claims.

[0009] According to a first aspect, a computer-implemented method is provided, in particular for determining a charging curve, for monitoring for anomalies and / or for determining an aging state, comprising the following steps: - Recording of temporal courses of operating variables of a large number of device batteries; - For a particular device battery, selecting a subset of similar device batteries from the plurality of device batteries that have a similar overall condition to the particular device battery; - For the model parameters of the battery models of the subset of device batteries, determining a lower and upper parameter limit for the respective model parameters depending on the distribution of the values ​​of the model parameters in the subset of device batteries at a calendar age of the specific device battery; - Performing a fitting procedure for the battery model to the specific device battery depending on the parameter limits of the model parameters based on the time courses of the operating variables in order to obtain the updated model parameters.

[0010] Conventional electrochemical battery models are mathematical models and have a multitude of model parameters. These battery models serve to mathematically describe the terminal voltage behavior and internal electrochemical processes in battery cells. The model parameters of such battery models are typically fitted by comparing the measured and simulated operating parameters using a conventional optimization procedure. Such fitting is performed regularly at recalibration times, as the model parameters must be adjusted to reflect the progressive degradation of the device battery.

[0011] Fitting methods are used to adapt a mathematical model to a given set of data. The model parameters are chosen to minimize the discrepancy between the observed data and the values ​​predicted by the model. The goal is to estimate the optimal model parameters that provide the "best" fit to the data. An error function is often defined that quantifies the discrepancy between the observed and modeled data. The most commonly used methods are the least squares method and the maximum likelihood method.

[0012] The battery model parameters are adjusted over the course of the battery's life cycle to reflect the progressive aging or degradation of the device battery. To this end, the model parameters are adjusted using nonlinear fitting methods based on temporal operating variables, which can include, in particular, the temporal profiles of cell voltage, cell current, cell temperature, and state of charge. This ensures that the battery model accurately represents a cell voltage curve depending on the cell current, temperature, and state of charge. The least squares method is particularly used here, with the number of parameters to be adjusted typically very high, between 20 and 30.

[0013] The application of the conventional fitting method to adapt the model parameters to the current measurement series of operating parameter curves has the disadvantage, particularly when recalibration intervals apply between recalibration times, that a large number of model parameter combinations lead to comparable terminal voltage behavior. In other words, the fitting method, which is based on minimizing the squared error, often leads to convergence within a local minimum, which, however, does not necessarily correspond to the global minimum of the cost function. Internal battery states, which result as model states from the evaluation of the battery model parameterized by the model parameters, can thus deviate from the real battery states intended to be described by the modeled internal battery states.This effect is aggravated by an insufficient quality of the provided operating variables due to measurement inaccuracies or if the required optimal operating conditions for carrying out the recalibration, i.e. the reparameterization of the battery model, are not present.

[0014] Particularly with longer recalibration intervals, one problem is that the parameter values ​​determined during recalibration can jump, since the large number of model parameters means that several local minima can usually be reached through optimization. All of these model parameter combinations are characterized by the same or very similar terminal voltage behavior with regard to underlying calibration data, i.e. measured temporal operating parameter profiles. This can lead to significantly different model parameter combinations with each recalibration. This means that a change in the sign of the gradient can occur when a model parameter changes monotonically (decreasing or increasing) with the aging of the device battery and can be assigned to an internal battery state.However, this contradicts the fundamental expectation that the model parameter values ​​should change continuously and monotonically during aging and thus indicates an at least temporarily implausible description of the battery state expected at the respective time.

[0015] The central concept of the above procedure lies in adjusting the permissible parameter limits or restricting the value space of the model parameters to enable an optimized setting of the model parameters. This helps to minimize the intrinsic uncertainty of the fitting procedure, which arises from the possibility of convergence to different local minima.

[0016] The above method provides for the parameterization of a battery model of a specific portable battery, determining the parameter limit values ​​for the model parameters based on an evaluation of the model parameters or an aging state derived from the model parameter. To this end, a subset of device batteries is first selected from a large number of other device batteries that have a similar overall condition to the specific device battery for which the model parameters are to be recalibrated. To this end, the temporal profiles of the model parameters and the aging states of the large number of other device batteries are compared with the temporal profiles of the model parameters and the aging state of the specific device battery, and a number of the other device batteries are selected as a subset using a similarity analysis.

[0017] In particular, the overall condition of a device battery can be determined by a feature point that is determined by an aging condition and / or one or more of the model parameters of the associated battery model and the calendar age of the device battery.

[0018] Similar portable batteries can be identified using a clustering method of feature points defined by the internal battery states, the aging state, the chronological age, and / or the battery's operating history. Similar portable batteries can be identified as such if, with regard to their feature points, they have a Euclidean distance from each other or from a centroid of the cluster they form that is smaller than a predefined threshold. For this purpose, the elements of the feature points can be standardized and / or weighted as appropriate. Batteries from the multitude of other portable batteries of the same type are considered to be similar, in particular, if they exhibit the greatest similarity (= smallest Euclidean distance) with the device battery under investigation in terms of the temporal progression of the aging state calculated by the aging model since the beginning of the device battery's life.Other characteristics that are suitable for identifying similar batteries are the cumulative energy throughput since the beginning of life and load histograms, which provide information about the state of charge and temperature range in which the battery was loaded to what extent in relation to power consumption.

[0019] Portable batteries can be removed from the subset of selected portable batteries if the aging state progression and / or the progression of the model parameters of the selected portable batteries are implausible. For this purpose, aging and model parameter trajectories can be created using polynomial or other regression methods to exclude portable batteries with implausible aging state progressions and / or one or more of the model parameters, or to remove those portable batteries from the subset whose aging state and / or model parameter trajectories do not correspond to the function progressions assumed in the regression. For example, the aging state and certain model parameters or internal battery states can only be monotonically increasing or decreasing.If both time periods with positive and time periods with negative gradients are determined within a course for the aging state and / or a model parameter, an implausible course can be assumed.

[0020] It may be provided that the smallest and largest value of the relevant model parameter of the battery models of the subset of portable batteries interpolated with respect to the calendar age of the specific portable battery are taken as the lower and upper parameter limits.

[0021] The historical model parameters of the other portable batteries in the subset determined at recalibration intervals can be interpolated to determine a distribution of each of the model parameters of the battery models in the subset of selected portable batteries for the current calendar age of the specific portable battery. This allows the respective parameter limits for the relevant model parameter to be determined in order to apply them in a fitting procedure for recalibrating the model parameters of the specific portable battery and to specify the search space for a suitable model parameter combination. The respective lowest and highest values ​​of the interpolation value of the relevant model parameter for the calendar age under consideration can be assumed as the lower and upper parameter limits.

[0022] It can be provided that the updated model parameters are used to determine the recalibrated internal battery states and / or the aging state of the specific device battery directly or after applying a computational model. The updated internal battery states can be used to monitor the specific device battery for an anomaly, in particular using a rule-based or data-based anomaly detection model, and / or to generate a charging curve. A charging curve adapted to the updated model parameters can be calculated, for example, by applying a current profile to the recalibrated model, under the application of which the battery states relevant for excessive battery aging attributable to the charging process (e.g. lithium concentrations at the electrode / electrolyte interface) do not exceed certain threshold values.This ensures that the device battery is loaded in a way that is aging- and condition-dependent, so that excessive aging caused by the charging process is reduced to a minimum.

[0023] According to a further aspect, an apparatus for carrying out the above method is provided. Brief description of the drawings

[0024] Embodiments are explained in more detail below with reference to the attached drawings. They show: Fig. 1 a schematic representation of a system with a plurality of vehicles of a vehicle fleet and a central unit for carrying out a recalibration of an electrochemical battery model, in particular for a charging curve optimization; and Fig. 2 a flowchart illustrating the method for recalibrating model parameters of an electrochemical battery model, in particular for use in monitoring the battery condition and determining a charging curve. Description of embodiments

[0025] The method according to the invention is described below using vehicle batteries as device batteries in a large number of motor vehicles as similar devices. For this purpose, an electrochemical battery model is parameterized in the central unit and used to monitor the overall condition of the vehicle battery, generate charging curves, and calculate an aging state. In the central unit, the battery models for each of the vehicle batteries are regularly updated or re-parameterized based on the operating parameters of the respective vehicle batteries in the vehicle fleet.

[0026] The above example is representative of a variety of stationary or mobile devices with off-grid energy supply, such as vehicles (electric vehicles, pedelecs, etc.), systems, machine tools, household appliances, IOT devices and the like, which are connected to a device-external central unit (cloud) via a corresponding communication connection (e.g. LAN, Internet).

[0027] Fig. Figure 1 shows a system 1 for collecting fleet data in a central unit 2 for creating and operating an electrochemical battery model. The model parameters and the resulting parameterized battery model are used to determine the aging state of battery cells, monitor the vehicle battery for anomalies, and determine a charging curve. Fig. 1 shows a vehicle fleet 3 with several motor vehicles 4. The electrochemical battery model is used to model a terminal voltage dependent on a battery current, a battery temperature and a state of charge at the cell level, module level and / or pack level.

[0028] One of the motor vehicles 4 is in Fig. 1. The motor vehicles 4 each have a vehicle battery 41, an electric drive motor 42, and a control unit 43. The control unit 43 is connected to a communication device 44, which is suitable for transmitting data between the respective motor vehicle 4 and a central unit 2 (a so-called cloud). The vehicle battery 43 has a plurality of battery cells 45.

[0029] The control unit 43 is particularly designed to record operating variables with a high temporal resolution, such as between 1 and 50 Hz, such as 10 Hz, and to transmit them to the central unit 2 via the communication device 44. In the case of a vehicle battery, the operating variables F can indicate a current battery current, a current battery voltage, a current battery temperature, and a current state of charge (SOC), both at the pack, module, and / or cell level. The operating variables can be regularly transmitted to the central unit 2 in uncompressed and / or compressed form. For example, the time series can be transmitted to the central unit 2 in blocks at intervals of 10 minutes to several hours using compression algorithms to minimize data traffic to the central unit 2.

[0030] The motor vehicles 4 send the operating variables F to the central unit 2, which at least indicate variables that influence the aging state of the vehicle battery 41 and are necessary for parameterizing the battery model. The model parameters correspond to internal battery states and / or can be used to determine the internal battery states of the battery cells 45.

[0031] Model parameters of the electrochemical battery model can be fitted or parameterized in the central unit based on operating parameter curves recorded during rest phases within short periods (a few minutes to a few hours) within a limited period of time, whereby electrochemical equilibrium parameters and kinetic model parameters can be derived, which can be interpreted as internal battery states such as electrolyte concentrations, reaction rates, layer thicknesses, porosity, etc.

[0032] The central unit 2 comprises a data processing unit 21 in which the method described below can be carried out, and a database 22 for storing data points, model parameters, states and the like.

[0033] A battery model is parameterized in the central unit 2 for each vehicle battery 41. The battery model can be specified in the form of a well-known Newman model. The Newman model is an electrochemical model for lithium-ion batteries that describes the transport and storage of Li ions in the porous electrodes, thereby predicting the expected terminal voltage of the battery depending on the Li concentration and the electrochemical potential in the electrode / electrolyte.

[0034] The battery model can be used regularly, e.g., after specified recalibration intervals, to recalibrate the model parameters of the battery model based on the temporal progression of the operating variables. The model parameters, when combined, allow the aging state of the vehicle battery 41 to be determined.

[0035] The state of health (SOH) is the key parameter for specifying the remaining battery capacity or remaining battery charge. The state of health represents a measure of the aging of the vehicle battery, a battery module, or a battery cell and can be specified as the capacity retention rate (SOH-C) or as the increase in internal resistance (SOH-R). The capacity retention rate SOH-C, i.e. the capacity-related state of health, is specified as the ratio of the measured instantaneous capacity to the initial capacity of the fully charged battery and decreases with increasing aging. Alternatively, the state of health can be specified as the increase in internal resistance (SOH-R) relative to the internal resistance at the beginning of the device battery's service life. The relative change in internal resistance SOH-R increases with increasing battery age.

[0036] In the central unit, the operating parameters are typically evaluated in a variety of ways. In particular, an electrochemical battery model is created for each vehicle battery, which best describes the respective vehicle battery 41 and, in particular, its internal battery states.

[0037] The battery model is typically designed as a reduced-order electrochemical continuum model, such as a Newman model, and allows the terminal voltage behavior and the internal electrochemical processes and battery states of the vehicle battery 41 or the battery cells contained therein to be described. Due to the degradation of the vehicle batteries 41, the associated battery model must be recalibrated regularly. For this purpose, time periods of the operating parameter curves are used as calibration data to adjust the model parameters using a known fitting method, which can be designed as an optimization based on the method of least squares.

[0038] The method described below uses the possibility of improving the optimization process by adapting the model parameters by specifying parameter limits in order to avoid incorrect parameterizations due to the diverging behavior of the optimization caused by the large number of model parameters to be parameterized.

[0039] Fig. 2 schematically shows a method carried out in the central unit 2.

[0040] In step S1, operating parameter profiles are received from a plurality of vehicle batteries 41, and for the subsequent process, a specific one of the vehicle batteries is selected for recalibration. Typically, the battery models for the vehicle batteries 41 are recalibrated at regular recalibration intervals, although the recalibration of all vehicle batteries 41 need not occur simultaneously.

[0041] In step S2, a subset of vehicle batteries 41 is first selected from the plurality of vehicle batteries 41 that have a similar overall condition to the specific vehicle battery 41. The overall condition of a vehicle battery 41 can be determined from a feature point from the aging condition, one or more of the model parameters of the associated battery model, and the chronological age of the vehicle battery 41.

[0042] For this purpose, a progression of the aging state and the model parameters can be created for each of the vehicle batteries 41 using a respective regression model based on their calendar age. Similar vehicle batteries 41 to the specific vehicle battery 41 can be found, for example, by determining the model parameters of the battery models and the aging states of the other vehicle batteries 41 for the corresponding calendar age for the calendar age of the specific vehicle battery 41 using the regression model, and thus determining feature points of all vehicle batteries 41 for the corresponding calendar age.Similar vehicle batteries 41 of the subset can be determined using a clustering analysis of the feature points or by evaluating the Euclidean distances of the feature points from the feature point of the specific vehicle battery 41 using a threshold comparison with a predetermined threshold.

[0043] In step S3, selected vehicle batteries can be removed from the subset if the aging state curves and / or the model parameter curves of the selected vehicle batteries are implausible. Implausibilities are essentially characterized by changes in the sign of the aging state and model parameter gradients (both the aging state and the model parameters should exhibit a homogeneously increasing or decreasing curve) or the general sign of the aging state and model parameter gradients (for many model parameters, it is known a priori, e.g., based on domain knowledge, whether they increase or decrease over the course of aging).

[0044] In step S4, the parameter limits are determined from the distribution of the values ​​of each model parameter of the subset of vehicle batteries. In particular, the smallest and largest value of the respective model parameter of the battery models of the subset of vehicle batteries can be assumed as the lower and upper parameter limits. For this purpose, for example, the curves of the model parameters of the selected device batteries of the subset can be interpolated over their calendar age in order to determine a distribution of the values ​​of each of the model parameters for the current calendar age of the specific device battery. This makes it possible to determine the respective parameter limits for the respective model parameter in order to apply them in a fitting process for recalibrating the model parameters of the specific device battery 41 and to specify the search space for a suitable model parameter combination.The lowest and highest values ​​of the interpolation value of the respective model parameter for the calendar age considered can be assumed as the lower and upper parameter limits.

[0045] In a subsequent step S5, a fitting process for the battery model of the specific vehicle battery can now be carried out on the basis of the operating variable curves recorded for the specific vehicle battery, wherein the parameter limits of the model parameters are specified accordingly in order to limit the value space of the model parameters for the fitting process.

[0046] As a result of the fitting procedure, recalibrated model parameters of the battery model are obtained, from which both the internal battery states and the aging state of the specific vehicle battery for the current calendar age of the vehicle battery 41 can be determined directly or after applying a calculation model.

[0047] These are stored with respect to the specific vehicle battery 41, for example, to provide a digital twin in the central unit. The method can now be repeated for a different vehicle battery because the previously recalibrated battery models of the vehicle battery can be taken into account in the subset of the selected vehicle battery, if necessary.

[0048] The model parameters of a battery model can be used to monitor the vehicle battery 41 for anomalies, in particular using rule-based or data-based anomaly detection models, and / or to generate a charging curve.

Claims

[1] Computer-implemented method for parameterizing an electrochemical battery model of a device battery (41) in a technical device (4) with updated model parameters for determining an internal battery state, in particular for determining a charging curve, for monitoring for anomalies and / or for determining an aging state, comprising the following steps: - recording (S1) time courses of operating variables of a plurality of device batteries (41); - For a specific device battery (41), selecting (S2) a subset of similar device batteries (41) from the plurality of device batteries that have a similar overall condition to the specific device battery (41); - For the model parameters of the battery models of the subset of device batteries (41), determining (S4) a lower and an upper parameter limit for the respective model parameters depending on the distribution of the values ​​of the model parameters in the subset of device batteries (41) at a calendar age of the specific device battery (41); - performing (S5) a fitting procedure for the battery model to the specific device battery (41) depending on the parameter limits of the model parameters based on the temporal courses of the operating variables in order to obtain the updated model parameters. [2] Method according to claim 1, wherein the smallest and largest value of the respective model parameter of the battery models of the subset of the device batteries (41) interpolated with respect to the calendar actuator of the specific device battery (41) are assumed as the lower and upper parameter limits. [3] Method according to one of claims 1 to 2, wherein the overall state of a device battery (41) is determined by a feature point which is determined by an aging state and / or one or more of the model parameters of the associated battery model and the calendar age of the device battery (41). [4] Method according to one of claims 1 to 3, wherein similar device batteries (41) are determined in particular by a Euclidean distance to one another or, after a clustering analysis, by a Euclidean distance to a centroid of a specific cluster, wherein the Euclidean distance is less than a predetermined threshold value. [5] Method according to one of claims 1 to 4, wherein device batteries (41) are removed from the subset of selected device batteries (41) if the ageing state profile and / or the profile of the model parameters of the selected device batteries (41) are implausible. [6] Method according to one of claims 1 to 5, wherein the updated model parameters are used to determine the internal battery states and / or the aging state of the specific device battery (41) directly or after applying a calculation model. [7] Method according to claim 6, wherein the updated model parameters are used to monitor the specific device battery (41) for an anomaly, in particular using a rule-based or data-based anomaly detection model, and / or to generate a charging curve. [8] Apparatus for carrying out one of the methods according to one of claims 1 to 7. [9] Computer program product comprising instructions which, when the program is executed by at least one data processing device, cause the device to carry out the steps of the method according to one of claims 1 to 7. [10] Machine-readable storage medium comprising instructions which, when executed by at least one data processing device, cause the device to carry out the steps of the method according to one of claims 1 to 7.

Images