Method and apparatus for predicting device battery pack error

Abstract

The invention relates to a computer-implemented method for providing a risk value for a predicted error in a system battery (41) of a technical system (4) by means of an error evaluation model having at least one error factor assignment table, the method comprises the following steps: detecting (S1) a temporal operating variable change process of the at least one device battery (41); -carrying out (S2) an abnormality detection as a function of the operating variable variation over time; detecting (S3, S4) an error-dependent variable when an abnormality is detected; evaluating (S6) an error evaluation model as a function of the error-related variable in order to determine an error type of the predicted error; assigning error factors to error types by means of at least one provided error factor assignment table of the error evaluation model; determining (S7) a risk value from the error factors; and signaling (S8) the risk value.

Description

Technical Field

[0001] The present invention relates to predicting errors in device batteries, particularly based on abnormal evaluation. Background Art

[0002] The technical functions of vehicle components are usually monitored. In this case, the relevant operating variables are usually evaluated based on rules in order to determine anomalies or errors, for example by means of threshold value comparisons. After such an error occurs, the error can be signaled accordingly to the user of the technical device or can be stored in an error memory for later evaluation.

[0003] However, to date, if abnormal or atypical behavior occurs, the user is asked to carry out corresponding repairs or maintenance. However, due to the different possible error types, it is necessary to grade the action instructions (Handlungsanweisung) in the form of corresponding signal notifications. Not every abnormality that occurs requires an immediate visit to the workshop, but it may be sufficient to record it accordingly so that it can be evaluated and, if necessary, repaired during the next visit to the workshop. Other errors require a visit to the workshop as quickly as a few days, while in the case of critical errors (such as thermal runaway of the battery pack), operation must be stopped immediately and the user must be kept away from the technical equipment to avoid danger.

[0004] Errors caused by unusual or atypical behavior of the corresponding components are usually foreseen early and can be identified by corresponding rule-based evaluations. This makes it possible to predict errors so that the user is informed of impending defects in good time.

[0005] For example, a method for predictive error detection is known from the publication DE 10 2020 212 277 A1, which describes a method for predictive diagnosis of an electric drive system by evaluating known and unknown error types.

[0006] In addition, publication US2021 / 0405104 A1 discloses a computer-implemented method for predicting a battery pack failure, the method comprising: calculating a plurality of corresponding prediction values ​​using a plurality of corresponding algorithms, wherein each of the plurality of corresponding algorithms uses historical battery pack data and current battery pack data; calculating the overall reliability of the battery pack failure prediction using the plurality of corresponding prediction values; and providing a prompt for the predicted battery pack failure based on the calculated overall safety. Summary of the invention

[0007] According to the invention, a method for predictive detection and evaluation of errors in a battery of a device is provided according to claim 1 and a corresponding device according to the independent parallel claim.

[0008] Further refinements are described in the dependent claims.

[0009] According to a first aspect, a computer-implemented method is provided for providing a risk value for a predicted error in a device battery of a technical device by means of an error assessment model, wherein the error assessment model has an error factor allocation table, the method comprising the following steps:

[0010] - Detecting the temporal course of operating parameters of at least one device battery pack

[0011] -Perform anomaly identification based on the temporal changes in operating parameters;

[0012] - when an anomaly is identified, detecting error-related parameters;

[0013] - evaluating the error assessment model based on error-related parameters to determine the error type of the predicted error;

[0014] - assigning error factors to error types with the aid of a provided error factor assignment table of the error assessment model;

[0015] -Determine the risk value based on the error factor;

[0016] -Signaling the risk value.

[0017] As described above, abnormal and untypical behavior of the device battery can be detected by, for example, rule-based evaluation of parameters of a battery model describing the current state of the device battery, by comparison of threshold values ​​of operating variables. If abnormal or unexpected behavior occurs, a monitoring mode is initiated in which one or more error-related variables are recorded, in particular using a sampling grid (Abtastraster) with a higher sampling rate.

[0018] The error-related variables to be recorded are determined according to the assignment table as a function of the variable from which the anomaly is derived, for example, in a rule-based manner. Error-related variables may include, for example, balancing frequency or temperature behavior, charge state course, OCV course (open circuit voltage characteristic course) (particularly in low charge states), aging state course, charging behavior, battery pressure course, etc.

[0019] The monitored error-related variables can be analyzed by means of feature extraction. Thus, for example, the gradient of one or more error-related variables that are particularly relevant to the effect of an error can be determined, or histograms can be created, for example of the temporal distribution of the battery temperature and / or the distribution of the cell pressure of the battery cells and / or the distribution of the cell voltage (to detect a deep discharge), which can provide an explanation of the error effect. Error-related features are obtained, to which error factors can be assigned by means of an error factor assignment table.

[0020] The error factors may indicate the propagation speed of the error, the severity of the error, and the probability of propagation of the error. The propagation speed indicates how quickly the error that should be expected will propagate and may result in a complete or critical failure or hazard. The severity of the error indicates how the error affects the ability of the battery pack to operate. The propagation probability indicates the probability that the error that should be expected is actually caused by the identified abnormal or unexpected behavior. By multiplying these error factors together, the risk for the expected error can be determined in the form of a risk value.

[0021] These error factors can be classified with values ​​within a predetermined range, for example between 0 and 10. Thus, a value range between 0 and 1000 can be derived for the risk value in this example.

[0022] According to the size of the risk value Now, a recommendation for action can be issued to the user of the device battery pack. In the case of a lower risk, a power reduction can be activated, for example, by current limitation. This power reduction can also be signaled to the user accordingly.

[0023] In addition, in the case of a higher risk value, a corresponding action proposal for shutting down the technical device or repairing or maintaining the technical device can be signaled. In the case of a particularly critical error indicated by a very high risk value, the device battery can be immediately completely shut down.

[0024] As an extreme example, thermal runaway should be mentioned as an error type in which the propagation speed is assigned a high value, since a thermal runaway of a device battery can lead to a very rapid increase in temperature and to a fire in the device battery. A high value must also be assigned to the severity of the error involved, since such an error can lead to a fire in the device battery and ultimately to a fire in the technical device. Since a thermal runaway of a device battery cannot be stopped once it has occurred, the probability of propagation should also be assigned a correspondingly high value.

[0025] The accuracy of the risk assessment and the action recommendations derived therefrom respectively depend on an error factor allocation table which assigns a corresponding classification of error factors to error-related characteristics such as propagation speed, error severity and propagation probability.

[0026] It may be provided that at least one of the steps of performing the anomaly detection and evaluating the error evaluation model is performed in a central unit remote from the device.

[0027] The method can be carried out at least partially in a central unit remote from the device. For this purpose, a large number of technical devices are monitored accordingly and, in the event of an error, an appropriate evaluation is performed. For this purpose, operating variables can be transmitted from the technical devices to the central unit. There, abnormalities and unexpected behaviors that have occurred can be detected and all possible error-related variables can be detected accordingly by signaling the technical devices involved. If the error-related variable is an operating variable, it can be detected at an increased sampling rate and transmitted to the central unit.

[0028] The device can be monitored to create an error factor assignment table. If an error eventually occurs, the corresponding error type can be assigned to the previously detected anomaly. By evaluating the error-related parameters, the propagation speed, i.e., the speed of the error development (Entwicklung) can be determined, for example, by classifying the time period between the identification of the anomaly and the occurrence of the corresponding error into an error factor.

[0029] Depending on the degree of danger resulting from the error type or the remaining usability of the device battery, the severity of the error can be graded so that, for example, the remaining service life at the time of the error occurrence can be assigned to the corresponding category.

[0030] Error severity. For example, in the case of a complete failure and danger to the user and the technical device, the error severity can be estimated according to the above exemplary scale as 10. If only the power of the device battery has deteriorated and the device battery is still usable, the error severity can be classified according to the above exemplary scale as a low error severity value between 1 and 5, corresponding to the remaining service life.

[0031] Furthermore, the error evaluation model can be updated in the central unit based on the course of operating variables of a large number of device batteries in dependence on detected anomalies and subsequently occurring errors of a specific error type.

[0032] By observing a large number of device batteries, the same abnormal and unexpected behavior can be evaluated in such a way that the probability of how many of the device batteries thus determined will develop the error type that occurred after the determination of the anomaly in question can be determined. From this, a propagation probability can be derived, where a value of 10 can indicate a deterministic development of the error in question, while a value of 0 or 1 can indicate a rare development of the corresponding error type according to the above exemplary scale.

[0033] The method described above makes it possible to easily predict errors of certain error types with little computational effort and to give the user of the technical device instructions for action at an early stage. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The embodiments are explained in more detail below with reference to the accompanying drawings, wherein:

[0035] Figure 1 A schematic diagram of a system for providing driver and vehicle specific operating parameters for providing action instructions after an abnormality occurs in a vehicle battery pack is shown;

[0036] Figure 2 A flow chart is shown for illustrating a method for providing action instructions after an anomaly is identified. DETAILED DESCRIPTION

[0037] The method according to the invention is described below with reference to a vehicle battery pack as a device battery pack in a large number of motor vehicles as identical devices. In a central unit, an error evaluation model with an error factor allocation table is optimized and continuously improved. The error evaluation model can be evaluated in the central unit or the error factor allocation table can be transmitted to the control devices of the vehicles in the fleet, so that in the event of an anomaly detected with the aid of the error evaluation model, action instructions for possible errors can be transmitted to the users of the vehicles at an early stage.

[0038] The above examples represent a large number of fixed or mobile devices with energy supply that is not dependent on the power grid, such as vehicles (electric vehicles, electric assisted bicycles, etc.), facilities, machine tools, household appliances, IoT devices, etc. These devices are connected to a central unit (cloud) outside the device via corresponding communication connections (such as LAN, Internet).

[0039] Figure 1 A system 1 is shown for collecting fleet data for the purpose of creating and operating an error assessment model in a central unit 2. In the central unit 2, operating variables can be detected and an aging state model for determining the aging state of a vehicle battery in a motor vehicle, an electrochemical battery model for modeling an internal battery state of a vehicle battery, a battery performance model for modeling electrical parameters and / or statistical or aggregate methods can be applied for feature extraction.

[0040] Figure 1 A fleet 3 of multiple motor vehicles 4 is shown. One of the motor vehicles 4 is Figure 1 These motor vehicles 4 each have a vehicle battery 41 with battery cells 45, 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 the central unit 2 (the so-called cloud).

[0041] Control unit 43 is designed in particular to detect operating variables of vehicle battery 41 detected by means of battery management system 46 with a high time resolution of, for example, between 1 Hz and 50 Hz, for example 10 Hz, and to transmit them in a suitable manner to central unit 2 via communication device 44 .

[0042] The motor vehicle 4 sends operating variables F to the central unit 2, which at least indicate the following variables: these variables influence the battery state of the vehicle battery 41 or are influenced by the battery state and are required to determine the internal battery state, the aging state and to parameterize the electrochemical battery model. In the case of a vehicle battery, these operating variables F can indicate the instantaneous battery current, the instantaneous battery voltage, the instantaneous battery temperature and the instantaneous state of charge (SOC: State of Charge).

[0043] The operating variable F is detected as an operating variable variation process with a fast time grid of 0.1 Hz to 50 Hz and can be regularly transmitted in an uncompressed and / or compressed form to the central unit 2. For example, for the purpose of minimizing the data traffic to the central unit 2, the time series can be transmitted to the central unit 2 block by block at intervals of 10 minutes to several hours by using a compression algorithm.

[0044] The central unit 2 has a data processing unit 21 and a database 22. A part of the method described below can be executed in the data processing unit 21, and the database 22 is used to store data points, model parameters, states, etc.

[0045] The central unit 2 can be designed to receive operating variable profiles F from all vehicles 3. Based on these operating variable profiles, the central unit 2 can detect anomalies of the vehicle battery 41 and errors in further temporal profiles in a rule-based and / or data-based manner. In the central unit 2 or the vehicle 3 concerned, an error evaluation model can be created for the detected anomalies or types of anomalies and with the help of subsequently occurring errors.

[0046] according to Figure 2The flow chart in describes in more detail a method for predicting a specific type of error of a vehicle battery pack in a vehicle operating on the battery pack.

[0047] First, in step S1, operating variables are monitored in the control unit 43, in particular in the battery control unit. These operating variables may include the battery voltage, the line current, the battery temperature and the charge state or the charge states of the batteries.

[0048] In step S2, the temporal variation of the operating parameter may be evaluated by means of one or more abnormality recognition models to identify an abnormality. The abnormality recognition model may identify an abnormality of the vehicle battery pack 41 based on an operating feature obtained by feature extraction of the temporal variation of the operating parameter. The abnormality recognition model may include a rule-based and / or data-based model, for example, a model trained on an autoencoder.

[0049] Accordingly, in step S3 it is checked whether there is any abnormal or unexpected behavior of the vehicle battery 41 .

[0050] If this is determined (option: yes), the method is continued with step S4 , otherwise the monitoring with step S1 is continued uninterruptedly.

[0051] In step S4, error-related variables are monitored in monitoring mode with an increased sampling rate. Error-related variables are derived from previously observed anomalies or observed unexpected behaviors according to an assignment table, wherein the assignment table assigns anomaly types to corresponding error-related variables, and the error-related variables may include, for example: balancing frequency, information on temperature behavior, charge state profile, open circuit voltage profile (OCV profile), aging state profile, charging behavior, in particular the dependency of the charge state on temperature, battery pressure profile, etc.

[0052] Based on the error-related parameters detected in this way, error-related characteristics can be determined in step S5 by evaluation or aggregation, for example, the gradient of one or more error-related parameters and / or operating parameters can be evaluated, or histograms can be created, such as histograms of the time distribution of the battery pack temperature and / or the distribution of the cell pressure of the battery pack cells and / or the distribution of the cell voltage (to identify deep discharge), which histograms can include explanations about the effects of the error.

[0053] According to the error factor allocation table of the provided error evaluation model, the error-related features can be allocated in step S6 according to the values ​​as a measure of the propagation speed, the error severity and the propagation probability. The allocation is carried out with the help of the error factor allocation table, which allocates the corresponding values ​​of the corresponding error factors to an error-related feature or a combination of a plurality of error-related features. The values ​​for the propagation speed, the error severity and the propagation probability can be within a predetermined value range, for example between 0 and 10. Other value ranges for the allocation of the corresponding error factors are also possible.

[0054] In the following step S7, the risk value is determined by multiplying the above-mentioned error factor.

[0055] Depending on the risk value, an intervention in the vehicle operation can be carried out in step S8 and / or a recommendation for action can be issued.

[0056] If the risk is low, for example, if the risk value is less than 100, then it can be done that the abnormality that has occurred is first signaled accordingly and the corresponding error type that should be expected is displayed. This can be coupled with the following action instruction: Go to the workshop as soon as possible.

[0057] In the case of risk values ​​in the middle range between 100 and 300, for example, power limitation of the device battery pack can be carried out. For example, current limitation of the charging and discharging current can be carried out to avoid overloading the vehicle battery pack. At the same time, an emergency instruction can be issued: the driver should go to a workshop.

[0058] In the case of a risk value greater than 300, it can be stipulated, for example: Immediately stop the operation of the equipment battery pack. An example is thermal runaway, in which it should be expected that the vehicle battery pack may catch fire within a short period of time. Accordingly, the driver must be urgently instructed to stop the vehicle and leave it.

[0059] The error-related features can be assigned to the corresponding error factors with the help of an error factor assignment table. The error factor assignment table can be continuously supplemented and improved by evaluating the identified abnormal or unexpected behavior of the vehicle battery pack, which is then followed by a specific error. The assignment depends on which type of error occurs in the subsequent operation of the vehicle. For example, the time between the occurrence of an abnormality and the occurrence of an error of a specific error type can be assigned to the propagation speed. In particular, the duration period between the first occurrence of an abnormality and the occurrence of an error can be assigned to a corresponding value of the propagation speed. In addition, the type of functional impairment and the danger to the vehicle and the user can be graded accordingly. The value for the propagation probability can be derived from the following aspects: the frequency of a specific error type resulting from the identification of a specific abnormality. This can be done by observing a fleet with a large number of vehicles.

[0060] The continuous adaptation of the error evaluation model can take place in the central unit 2. The evaluation of vehicle-specific anomalies can take place in the central unit 2 or also in individual vehicles 4 of the fleet 3 after the parameters of the error evaluation model have been transmitted.

Claims

1. A computer-implemented method for providing a risk value for a predicted error in a device battery (41) of a technical device (4) by means of an error assessment model, wherein the error assessment model has at least one error factor allocation table, the method comprising the following steps: - detecting (S1) a temporal change process of an operating parameter of at least one device battery pack (41); - performing (S2) abnormality identification according to the temporal operation parameter variation process; - when an anomaly is identified, detecting (S3, S4) a parameter associated with the error; - evaluating (S6) the error assessment model based on the error-related parameter to determine the error type of the predicted error; - assigning error factors to error types by means of at least one error factor assignment table provided by the error evaluation model; - determining (S7) a risk value based on the error factor; - signalling (S8) said risk value.

2. The method according to claim 1, wherein: The error-related variables include: balancing frequency, temperature behavior, charge state profile, in particular OCV profile (open circuit voltage characteristic curve) in a low charge state, aging state profile, charging behavior and / or battery pressure profile.

3. The method according to claim 1 or 2, wherein: A feature extraction is performed using the error-related variables in order to obtain error-related features, wherein an error factor for the error type is determined from the error-related features by means of a provided error factor assignment table of the error evaluation model.

4. The method according to any one of claims 1 to 3, wherein: The error factor includes at least one of the following factors: a propagation speed of the error, a severity of the error, and a propagation probability of the error.

5. The method according to any one of claims 1 to 4, wherein the risk value is determined based on the multiplication of the error factors.

6. The method according to any one of claims 1 to 5, wherein an action recommendation is issued to a user of the device battery pack (41) based on the size of the risk value and the type of the error.

7. The method according to any one of claims 1 to 6, wherein: At least one of the steps of performing anomaly identification and evaluating the error assessment model is performed in a central unit (2) remote from the device.

8. The method according to any one of claims 1 to 7, wherein: Detecting the error-related parameter includes: detecting the operating parameter change process at a higher sampling rate, or detecting the operating parameter change process at a higher sampling rate after the abnormality is identified.

9. The method according to any one of claims 1 to 8, wherein: In the central unit (2), the error evaluation model and in particular the error factor allocation table are updated based on the course of operating variables of a plurality of device batteries in accordance with the detected anomalies and subsequently occurring errors of a specific error type.

10. An apparatus for performing the method according to any one of claims 1 to 9.

11. A computer program product comprising instructions which, when said program is executed by at least one data processing device, cause said data processing device to perform the steps of the method according to any one of claims 1 to 9.

12. A machine-readable storage medium comprising instructions which, 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 9.

Images