Method and apparatus for parameterizing an electrochemical cell model of a battery based on data from multiple batteries

Abstract

The invention relates to a computer-implemented method for parameterizing an electrochemical cell model of a specific battery of a device, having the following steps: providing current operating characteristic points of the specific battery and operating characteristic points of other batteries of a plurality of further devices for different evaluation time periods, wherein the operating characteristic points characterize the operation of the respective battery over one or more evaluation time periods according to a plurality of operating characteristics, wherein the operating characteristics result from curves of operating variables of the respective battery; selecting operating characteristic points of the other batteries that are similar to the current operating characteristic points of the specific battery; providing model parameters of the cell model that are associated with the current operating characteristic points and the selected operating characteristic points; deriving at least one fused model parameter from the model parameters of the current operating characteristic points of the specific battery and the selected operating characteristic points of the other batteries; signaling the at least one fused model parameter or updating the cell model with the at least one fused model parameter.

Description

Technical Field

[0001] This invention relates to electrical devices with batteries that operate independently of the power grid, and also to measures for determining the State of Health (SOH) of an energy storage device. In addition to mobile energy storage devices, the invention also relates to stationary energy storage devices. Background Technology

[0002] Electrical equipment and machinery that operate independently of the power grid, such as electrically powered motor vehicles, are typically powered by batteries.

[0003] A battery's state of aging decreases significantly over its lifespan, leading to a reduction in maximum storage capacity. The metric for battery aging depends on the individual battery's load, which in turn depends on the usage behavior of the devices operating on the battery, external environmental conditions, and the type of vehicle battery.

[0004] While the current aging state of an energy storage device can be determined based on historical operating condition curves using empirical aging state models, these models are inaccurate under certain circumstances. This inaccuracy of conventional aging state models makes predicting aging state curves difficult. However, predicting the aging state curve of an energy storage device is an important technical variable because it allows for the economical assessment of the device's remaining value.

[0005] To determine the aging state of a battery, an electrochemical battery model can be used. This electrochemical battery model is parameterized by a large number of model parameters and is based on differential equations that model the internal equilibrium state.

[0006] To adapt the model parameters of an electrochemical battery model, curves representing the battery's operating variables, such as battery current, battery voltage (terminal voltage), battery temperature, and state of charge, can be used, provided the battery has a known or measured aging state. Because electrochemical battery models are highly dependent on the battery's aging state, the battery model is reparameterized or its parameters are updated at regular intervals. Summary of the Invention

[0007] According to the present invention, a method and a corresponding apparatus are provided for parameterizing the battery model of the battery of a battery-operated machine.

[0008] Other design schemes are described in different implementation methods.

[0009] Based on the first aspect, a computer-implemented method for parameterizing an electrochemical cell model of a specific battery of a device is proposed, comprising the following steps: - Provides the current operating characteristic points of a specific battery and the operating characteristic points of other batteries of multiple other devices for different evaluation time periods, wherein the operating characteristic points characterize the operation of the corresponding battery within one or more evaluation time periods based on multiple operating characteristics, wherein the operating characteristics are derived from the curves of the operating variables of the corresponding battery. - Select operating characteristic points of other batteries that are similar to the current operating characteristic points of the specific battery; - Provides the model parameters of the battery model associated with the current operating feature point and the selected operating feature point; - Calculate at least one fused model parameter based on the model parameters of the current operating characteristic points of a specific battery and the selected operating characteristic points of other batteries; - Signal at least one fused model parameter or update the battery model with at least one fused model parameter.

[0010] The state of aging of rechargeable energy storage devices, especially device batteries, is not typically measured directly. This would require a series of sensors near the energy storage device, which would make the manufacturing of such devices expensive and costly, and would also increase structural space. Furthermore, there are no commercially available measurement methods suitable for everyday use to determine the state of aging in devices. Therefore, the current state of aging is usually determined by relying on physical aging models of the device. These physical aging models are inaccurate under certain conditions and typically have a model bias of up to 5%.

[0011] Physical aging models are based on electrochemical battery models, which have a series of differential equations for determining equilibrium parameters and are parameterized via model parameters. The parameterization of model parameters in electrochemical battery models is typically performed by measuring the battery's total capacity or internal resistance to infer the battery's aging state. With the aid of optimizers, the model parameters of electrochemical battery models can now be determined in a known manner based on operating variables, such as battery voltage (terminal voltage), battery current, battery temperature, and state-of-charge curves, along with specific descriptions of the aging state.

[0012] Furthermore, due to the inaccuracy of physical aging condition models, these models can only describe the current aging state of the energy storage device. The prediction of aging state results in highly fluctuating and inaccurate predictions, rendering them unusable. These predictions are particularly related to the energy storage device's operating mode, the magnitude and amount of charge inflow and outflow in the device's battery, and consequently, usage behavior and operating parameters.

[0013] State of Occurrence (SOH) is a key variable used to describe the remaining battery capacity or remaining battery charge when a device battery is used as an energy storage device. SOH is a measure of the aging of the energy storage device. In the case of a device battery, battery module, or single battery cell, SOH can be described as either the Capacity Retention Rate (SOH-C) or the increase in internal resistance (SOH-R). The Capacity Retention Rate (SOH-C) is the ratio of the measured instantaneous capacity to the initial capacity of a fully charged battery. The relative change in internal resistance (SOH-R) increases with battery aging.

[0014] The aforementioned method now proposes to derive model parameters for an electrochemical battery model for a specific battery based on data from multiple other batteries. To this end, clustering methods are employed to select operating feature points from other batteries that are similar to the current operating feature points of the specific battery, and associated model parameters for the electrochemical battery model are determined for each of the selected operating feature points. A fusion method is then used to combine the model parameters of the current operating feature points of the specific battery with those of the selected operating feature points from other batteries to obtain fused model parameters for the battery model of the specific battery.

[0015] In this way, the model parameters of the electrochemical battery model can be determined more accurately, even under specific battery or other battery operating variables or measurement precision, such as aging state descriptions obtained by measurement. The model parameters of the electrochemical battery model are now suitable for: determining the aging state in an electrochemical aging state model or using the electrochemical battery model in an idle characteristic curve model.

[0016] Furthermore, it can be proposed that: at least a portion of the current operating characteristic points of a specific battery and a portion of the operating characteristic points of other batteries be provided with an aging state, and an operating characteristic point similar to the current operating characteristic point of a specific battery be selected based on the aging state of the corresponding operating characteristic point.

[0017] It can be proposed that the aging state is determined based on measurements, particularly based on the coulomb counting method, where the model parameters of the battery at the relevant operating characteristic points are updated according to the curves and measurements of the relevant battery operating variables.

[0018] Alternatively, an aging state model can be provided based on the battery model to model the corresponding aging state at the current operating characteristic point of a particular battery and the operating characteristic points of other batteries, wherein the selection of operating characteristic points similar to the current operating characteristic point of a particular battery is performed according to the aging state of the corresponding operating characteristic point.

[0019] Furthermore, clustering methods, particularly k-Means clustering, can be used to select operating characteristic points for other batteries that are similar to the current operating characteristic points of a particular battery.

[0020] According to one implementation, the method can be executed in a central unit, where fused model parameters are transmitted to a device with a battery so that the battery model can be updated there.

[0021] Furthermore, the fused model parameters can be obtained by combining the model parameters with the Kalman filter or the Lumberjack observer, or by averaging them separately, or by averaging them separately in a similarity-weighted manner, wherein the model parameters are associated with the current operating feature point of the battery and the selected operating feature points of the other batteries, respectively.

[0022] According to another aspect, an apparatus is proposed for parameterizing an electrochemical cell model of a specific cell in a device, the apparatus having, particularly in a central cell communicatively connected to multiple cells, the following steps: - Provides the current operating characteristic points of a specific battery and the operating characteristic points of other batteries of multiple other devices for different evaluation time periods, wherein the operating characteristic points characterize the operation of the corresponding battery within one or more evaluation time periods based on multiple operating characteristics, wherein the operating characteristics are derived from the curves of the operating variables of the corresponding battery. - Select operating characteristic points of other batteries that are similar to the current operating characteristic points of the specific battery; - Provides the model parameters of the battery model associated with the current operating feature point and the selected operating feature point; - Calculate at least one fused model parameter based on the model parameters of the current operating characteristic points of a specific battery and the selected operating characteristic points of other batteries; - Signal at least one fused model parameter or update the battery model with at least one fused model parameter.

[0023] In addition, devices with specific batteries can correspond to motor vehicles, electric bicycles, airplanes, and especially drones, machine tools, autonomous robots and / or home appliances. Attached Figure Description

[0024] The embodiments are explained in more detail below with reference to the accompanying drawings. Wherein: Figure 1 A schematic diagram is shown of a system for providing driver-specific and vehicle-specific operating variables for determining the aging state of the vehicle battery in a central unit. Figure 2 A flowchart illustrating a method for providing an aging state model in the central unit is shown; and Figure 3 The diagram illustrates the operating characteristics and aging status of batteries in multiple vehicles. Detailed Implementation

[0025] The method according to the invention is described below using vehicle batteries as energy storage devices in multiple motor vehicles that operate as devices of the same type of battery. In the motor vehicle, a physical aging state model of the corresponding vehicle battery can be implemented in the control unit. The aging state model can be continuously or periodically updated or retrained in the central unit based on the operating variables of the vehicle batteries from the fleet.

[0026] The example above represents multiple fixed or mobile devices, such as vehicles (electric vehicles, electric bicycles, etc.), facilities, machine tools, home appliances, IoT devices, etc., that have their own power supply independent of the power grid and are connected to a central unit (cloud) via appropriate communication connections (such as LAN, Internet).

[0027] Figure 1 System 1 is shown for collecting fleet data in central unit 2 to create, run, and evaluate aging condition models. The aging condition models are used to determine the aging condition of electrical energy storage devices, such as vehicle batteries in motor vehicles. Figure 1 A convoy 3 with multiple motor vehicles 4 is shown.

[0028] exist Figure 1 One of the motor vehicles 4 is shown in more detail. The motor vehicle 4 has a vehicle battery 41 as a rechargeable electrical energy storage device, an electric drive motor 42, and a control unit 43. The control unit 43 is connected to a communication module 44, which is adapted to transmit data between the respective motor vehicle 4 and the central unit 2 (cloud).

[0029] The central unit 2 has a data processing unit 21 and a database 22, in which the methods described below can be executed, and the database is used to store the aging state curves of the batteries of multiple vehicles 4 of the fleet 3.

[0030] The vehicle 4 sends an operating variable F to the central unit 2. This operating variable describes at least the following variables, which are related to the aging state of the vehicle battery. In the case of the vehicle battery, the operating variable F can describe instantaneous battery current, instantaneous battery voltage, instantaneous battery temperature, and instantaneous state of charge (SOC) at the group level, module level, and / or individual cell level. The operating variable F is detected using a fast time grid from 2Hz to 100Hz and can be periodically transmitted to the central unit 2 in uncompressed and / or compressed form. For example, time series can be transmitted to the central unit 2 in blocks at intervals of 10 minutes to several hours.

[0031] Based on the operating variable F, operating characteristics can be generated in the central unit 2 or, in other embodiments, in the corresponding vehicle 4. These operating characteristics relate to an evaluation time period. To determine the aging state, the evaluation time period can range from several hours (e.g., 6 hours) to several weeks (e.g., one month). A commonly used value for the evaluation time period is one week.

[0032] Operating characteristics may include, for example, characteristics relating to the evaluation period and / or cumulative characteristics and / or statistical variables derived over the entire service life to date. Specifically, operating characteristics may include, for example, histogram data on the state of charge curve, temperature, battery voltage, battery current, particularly histogram data on the battery temperature distribution with respect to the state of charge, the charging current distribution with respect to temperature, and / or the discharging current distribution with respect to temperature, cumulative total capacity (Ah), average capacity increase during charging processes (especially for charging processes where the charge increase exceeds a certain threshold share of the total battery capacity (e.g., 20%), maximum differential capacity (dQ / dU: charge change divided by battery voltage change), etc. Electrochemical state (layer thickness, concentration, cyclizable lithium, etc.) may also be used as operating characteristics if known.

[0033] Other information can be obtained from the operating characteristics: time-load patterns, such as charging and driving cycles, which are determined by the use patterns (such as fast charging or strong acceleration or braking processes with regenerative braking at high current intensity), vehicle battery usage time, charge accumulated over operating time and discharge accumulated over operating time, maximum charging current, maximum discharging current, charging frequency, average charging current, average discharging current, power throughput during charging and discharging, (especially average) charging temperature, (especially average) dispersion of the state of charge, etc.

[0034] Physical aging state models are implemented in each of the motor vehicles 4. These aging state models can be used periodically, i.e., after the corresponding assessment period has expired, to determine the current aging state of the vehicle battery 41 based on operational characteristics. In other words, it is feasible to determine the aging state of the relevant vehicle battery 41 based on operational characteristics derived from the operational variable curves of one of the motor vehicles 4 in the fleet 3.

[0035] The aging model is based on an electrochemical battery model of single-cell and single-cell chemistry. The battery model consists of a series of differential equations, which take the battery's behavior and, in particular, the electrochemical equilibrium state as model parameters. For example, model parameters may include layer thickness, such as SEI thickness, changes in cyclizable lithium due to anode-cathode side reactions, rapid electrolyte depletion, slow electrolyte depletion, loss of active material in the anode, loss of active material in the cathode, etc. The battery model calculates the internal physical battery state based on the curve of the operating variable F to provide a physical aging state in the form of capacity retention rate (SOH-C) and / or as the rate of increase in internal resistance (SOH-R).

[0036] Electrochemical battery models can be used in aging state models or charging state models, which are based on idling characteristic curves determined by the battery model.

[0037] exist Figure 2 The flowchart shown describes a method for determining model parameters of an electrochemical battery model based on fleet data of batteries in multiple motor vehicles 4. The method may be performed partially or entirely in the central unit 2.

[0038] The method is implemented as software and / or hardware in the control unit 43 of the central unit 2.

[0039] In step S1, operating variables F are continuously received from multiple vehicles 4 in the central unit 2, such that curves of the operating variables are provided in the central unit 2. Furthermore, physical aging state models based on electrochemical cell models are implemented in each of the vehicles 4. The electrochemical cell models are initially parameterized using model parameters.

[0040] Furthermore, in step S2, a complete charge or discharge cycle is executed in vehicle 4 within the range of the coulomb counting method, either triggered by a specific operating state or randomly. This achieves a description of the aging state SOH in the form of maximum capacity (capacity retention rate SOH-C) or as the rate of increase in internal resistance (SOH-R). This description is, in principle, based on an electrochemical battery model, which is used through an aging state model to update the parameterization of the model parameters of the electrochemical battery model using methods known per se. The aging state SOH is also transmitted to the central unit 2.

[0041] In step S3, for each vehicle battery 41, it is checked whether the aging state of the vehicle battery 41 has decreased by a preset, incremental relative value, for example, a decrease of 2%. If this is the case (either of two options: yes), the method continues to step S4 for the relevant specific vehicle battery 41; otherwise, it jumps back to step S1. The standard triggers an update of the model parameters of the electrochemical battery model.

[0042] In step S4, the corresponding model parameters of the electrochemical battery model can now be updated based on the aging state of vehicle 4 measured in step S2. In this way, the current model parameters of the battery model of vehicle 4's battery 41 of vehicle 3 are always present in the central unit 2.

[0043] In step S5, the operating characteristics of batteries 41 in multiple vehicles 4 are subsequently determined based on curves of operating variables, such as battery current, battery voltage, battery temperature, and state of charge. As described above, the operating characteristics are determined at different evaluation time points, and the modeled or measured aging state can be associated with each evaluation time point. Therefore, in database 22, for each vehicle battery 41 of a vehicle, there are the final determined model parameters of the electrochemical battery model, the measured or modeled aging state, and the corresponding operating characteristic points for each evaluation time period, which describes the overall operating characteristics of the corresponding vehicle battery 41.

[0044] Now, in step S6, based on the current operating characteristic points of a specific vehicle battery 41 of vehicle 4 in fleet 3, and the operating characteristic points of battery 41 and, if necessary, their respective states of aging (SOH), other batteries similar to the specific vehicle battery 41 and having similar battery states can be selected by means of a clustering method, wherein the model parameters of the battery model should be updated for the specific vehicle battery. This selection is performed from the set of operating characteristic points of each battery 41 from the corresponding start of commissioning and the evaluation period of the current model parameters, or from the corresponding time point.

[0045] exist Figure 3 The example plots operational features M for the aging state of multiple batteries 41. For example, two clusters are obtained. By using a clustering method, operational feature points similar to the current operational feature points of a particular vehicle battery 41 can be easily identified, which have been identified currently or in the past for other vehicle batteries.

[0046] Therefore, by using clustering methods, such as k-means clustering, to identify the operating characteristic points of other vehicle batteries, these operating characteristic points are similar to those of the observed specific battery. All operating characteristic points having an Euclidean distance from the current operating characteristic point that is less than a preset threshold are similar. Alternatively, the distance from the centroid of the identified cluster, in which the current operating characteristic point is located, can also be selected. Accordingly, the battery to which multiple batteries in the fleet belong is identified by means of the operating characteristic points, and the effective model parameters of the corresponding electrochemical battery model during the evaluation time period of the identified operating characteristic points are obtained.

[0047] The model parameters of the electrochemical battery model selected for each vehicle battery, which are continuously updated in step S4, can also be determined only for the selected operating characteristic points. This is done, in particular, by means of optimization methods, based on the operating variables of the associated batteries and the measured aging state during the evaluation period identified by the operating characteristic points. Therefore, for the selected operating characteristic points, the determination of the model parameters of the corresponding electrochemical battery model is triggered by specific changes in the aging state of a particular battery similar to that of the selected battery.

[0048] Now, for a specific battery and for the selected battery, the corresponding model parameters of the electrochemical battery model are obtained.

[0049] In the subsequent step S7, the model parameters thus obtained are fused together to obtain adapted model parameters for a specific battery. For example, for each observed operational point, the individual model parameters are fused using a Kalman filter or a Lumberjack observer to obtain updated model parameters. Alternatively, the model parameters are individually averaged or averaged in a weighted manner. The weights for this can be derived from the Euclidean distance to the centroid of the cluster obtained for the specific vehicle battery or from the distance to the operational feature point of the specific vehicle battery 41.

[0050] This approach is feasible by updating the model parameters based on fleet data from multiple batteries for a specific battery.

[0051] In step S8, the fused model parameters thus obtained are transmitted to the vehicle with the specific vehicle battery and there used in the electrochemical battery model to determine the aging state or the state of charge.

Claims

1. A computer-implemented method for parameterizing an electrochemical cell model of a specific cell (41) of a device (4), comprising the following steps: - Provide (S5) the current operating characteristic points of a specific battery (41) and the operating characteristic points of other batteries of multiple other devices (4) for different evaluation time periods, wherein the operating characteristic points characterize the operation of the corresponding battery (41) within one or more evaluation time periods according to multiple operating characteristics (M), wherein the operating characteristics (M) are derived from the curve of the operating variable (F) of the corresponding battery (41). - Select (S6) an operating characteristic point of the other battery (41) that is similar to the current operating characteristic point of the specific battery; - Provide the model parameters of the battery model associated with the current operating feature point and the selected operating feature point; - Calculate (S7) at least one fused model parameter based on the model parameters of the current operating characteristic point of the specific battery (41) and the selected operating characteristic points of the other batteries; - Notify (S8) of at least one fused model parameter or update the battery model with at least one fused model parameter.

2. The method according to claim 1, further comprising the following steps: - Provide aging status (S4) for at least a portion of the current operating characteristic points of the particular battery and the operating characteristic points of the other batteries (41). - Select (S6) an operating feature point similar to the current operating feature point of the specific battery based on the aging state of the corresponding operating feature point.

3. The method according to claim 2, wherein the aging states are determined based on measurements, wherein the model parameters of the battery (41) for the relevant operating characteristic points are updated according to the curves and measurements of the relevant operating variables of the battery (41).

4. The method according to claim 3, wherein the aging state is determined based on the coulomb counting method.

5. The method according to claim 2, wherein an aging state model is provided based on the battery model to model corresponding aging states at the current operating characteristic point of the particular battery (41) and the operating characteristic points of the other batteries (41), wherein the selection of operating characteristic points similar to the current operating characteristic point of the particular battery is performed according to the aging state of the corresponding operating characteristic point.

6. The method according to any one of claims 1 to 5, wherein the selection of operating characteristic points of the other batteries (41) similar to the current operating characteristic points of the particular battery (41) is performed by means of a clustering method.

7. The method according to claim 6, wherein the clustering method is k-means clustering.

8. The method according to any one of claims 1 to 5, wherein the method is performed in a central unit (2), wherein the fused model parameters are transmitted to a device (4) having the particular battery (41) so as to update the battery model there.

9. The method according to any one of claims 1 to 5, wherein the fused model parameters are obtained by combining the model parameters with a Karman filter or a Luneburger observer, or by averaging them, or by averaging them in a similarity-weighted manner, wherein the model parameters are associated with the current operating feature point of the battery (41) and the selected operating feature points of the other batteries (41), respectively.

10. The method according to any one of claims 1 to 5, wherein the device having the particular battery corresponds to a motor vehicle, an electric bicycle, an aircraft, a machine tool, an autonomous robot, and / or a household appliance.

11. The method of claim 10, wherein the aircraft is a drone.

12. An apparatus for parameterizing an electrochemical cell model of a battery (41) in a device (4), wherein the apparatus is configured in a central unit (2) communicatively connected to a plurality of batteries (41) for: - Provide the current operating characteristic points of a specific battery (41) and the operating characteristic points of other batteries of multiple other devices (4) for different evaluation time periods, wherein the operating characteristic points characterize the operation of the corresponding battery (41) within one or more evaluation time periods according to multiple operating characteristics (M), wherein the operating characteristics (M) are derived from the curve of the operating variable (F) of the corresponding battery (41). - Select an operating characteristic point of the other battery (41) that is similar to the current operating characteristic point of the specific battery; - Provide the model parameters of the battery model associated with the current operating feature point and the selected operating feature point; - Calculate at least one fused model parameter based on the model parameters of the current operating characteristic point of the specific battery (41) and the selected operating characteristic points of the other batteries; - Signal at least one fused model parameter or update the battery model with at least one fused model parameter.

13. A computer program product comprising instructions that, when executed by at least one data processing device, cause the data processing device to perform the steps of the method according to any one of claims 1 to 11.

14. A machine-readable storage medium comprising instructions that, when executed by at least one data processing device, cause the data processing device to perform the steps of the method according to any one of claims 1 to 11.

Images