Abnormal cause determination method and device of abnormal battery and electronic equipment

Abstract

The invention discloses a method and a device for determining an abnormal reason of an abnormal battery and electronic equipment. The method comprises the steps that a battery pack is obtained, and the battery pack comprises a plurality of single batteries; acquiring voltage time sequence data respectively corresponding to the plurality of single batteries; according to the multiple pieces of voltage time sequence data, a voltage deviation rate matrix corresponding to the voltage data matrix is determined, and elements in the voltage deviation rate matrix represent the voltage deviation rates of the single batteries under the corresponding numbers at the corresponding moments; determining an abnormal battery from the plurality of single batteries according to the voltage deviation rate matrix; and inputting the first battery parameter of the abnormal battery and the second battery parameter of the normal battery into a target abnormal reason identification model, and determining an abnormal reason causing the battery to be abnormal. According to the method and the device, the technical problems that the abnormality judgment step is complicated when whether the abnormal battery exists in the plurality of single batteries or not is determined in the prior art, and the abnormality reason is difficult to accurately determine after the abnormality is determined are solved.

Description

Technical Field

[0001] The present invention relates to the field of batteries, and in particular to a method, device and electronic equipment for determining the abnormal cause of an abnormal battery. Background Art

[0002] In actual operation, electric vehicles face a variety of working conditions and different driving behaviors, which makes the collected data complex and diverse with many interference factors. In the related technology, when determining whether there are abnormal cells among multiple single cells, it is necessary to compare the differences in data between single cells, which requires a large amount of calculation. If it does not conform to the characteristics of normal distribution, it will inevitably bring difficulties to the judgment of abnormal single cells. In addition, due to the individual differences of single cells, the steps of judging abnormalities are more complicated, resulting in time-consuming and laborious technical problems.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present invention provide a method, device and electronic device for determining the abnormal cause of an abnormal battery, so as to at least solve the technical problems in the related art that when determining whether there is an abnormal battery among multiple single cells, the steps for judging the abnormality are complicated, and after the abnormality is determined, it is difficult to accurately determine the cause of the abnormality.

[0005] According to one aspect of an embodiment of the present invention, a method for determining an abnormal cause of an abnormal battery is provided, comprising: obtaining a battery pack, wherein the battery pack includes a plurality of single cells; obtaining voltage timing data corresponding to the plurality of single cells respectively; determining a voltage shift rate matrix corresponding to a voltage data matrix based on the plurality of voltage timing data, wherein the elements in the voltage shift rate matrix represent the voltage shift rates of single cells with corresponding numbers at corresponding moments; determining an abnormal battery from the plurality of single cells based on the voltage shift rate matrix; inputting a first battery parameter of the abnormal battery and a second battery parameter of a normal battery into a target abnormal cause identification model to determine an abnormal cause causing the battery to be abnormal, wherein the normal battery is a battery among the plurality of single cells except the abnormal battery, and the target abnormal cause identification model is a model obtained by training an initial abnormal cause identification model based on a loss function using a plurality of groups of sample data, wherein the plurality of groups of sample data include: sample battery parameters corresponding to sample abnormal batteries and sample battery parameters corresponding to sample normal batteries.

[0006] Optionally, the target abnormal cause identification model is obtained by training with the multiple groups of sample data based on a loss function in the following manner: the initial abnormal cause identification model is obtained, and a loss function for model training is constructed, wherein the loss function includes a first loss function term, a second loss function term and a third loss function term, the first loss function term is used to maintain the similarity of battery parameters of the same type of batteries in the feature space, the second loss function term is used to maintain the heterogeneity of battery parameters of different types of batteries in the feature space, and the third loss function term is used to classify the abnormal causes of the abnormal batteries; based on the loss function, the initial abnormal cause identification model is trained with the multiple groups of sample data to obtain the target abnormal cause identification model.

[0007] Optionally, the first battery parameter of the abnormal battery and the second battery parameter of the normal battery are input into a target abnormal cause identification model to determine the abnormal cause of the battery abnormality, including: inputting the first battery parameter and the second battery parameter into a first abnormal identification module of the target abnormal cause identification model to determine whether the abnormal result of the battery abnormality is due to an abnormal internal parameter of the battery; when the abnormal result indicates that the battery abnormality is not caused by an abnormal internal parameter of the battery, inputting the first battery parameter and the second battery parameter into a second abnormal identification module of the target abnormal cause identification model to determine whether the abnormal result of the battery abnormality is due to a connection line fault corresponding to the abnormal battery, so as to troubleshoot the cause of the abnormality of the abnormal battery through the connection line.

[0008] Optionally, determining a voltage offset rate matrix corresponding to the voltage data matrix based on multiple voltage timing data includes: constructing a voltage data matrix corresponding to the multiple single cells based on multiple voltage timing data, wherein the voltage data matrix comprises multiple row vectors and multiple column vectors, each of the multiple row vectors represents the voltage value of single cells with different numbers at the same time, each of the multiple column vectors represents the voltage value of single cells with the same number at different times, and the elements in the voltage data matrix represent the voltage value of single cells with corresponding numbers at corresponding times; determining a voltage offset rate matrix corresponding to the voltage data matrix based on the average voltage value corresponding to the battery pack.

[0009] Optionally, determining a voltage offset rate matrix corresponding to the voltage data matrix includes: determining an offset rate of each element in the voltage data matrix compared to the average voltage value to obtain an offset rate corresponding to each element; and determining the voltage offset rate matrix based on the offset rates corresponding to each element.

[0010] Optionally, determining the average voltage value corresponding to the battery pack includes any one of the following: determining the average voltage value corresponding to the battery pack based on the voltage data matrix; determining the average voltage value corresponding to the battery pack based on the predicted voltage value of the battery pack; determining the average voltage value corresponding to the battery pack based on the average voltage value of the battery pack in a leading time period and a voltage change parameter corresponding to the battery pack, wherein the leading time period is a time period before the corresponding moment of obtaining the voltage timing data.

[0011] Optionally, abnormal batteries are determined from the multiple single cells based on the voltage shift rate matrix, including: determining an average single cell shift rate corresponding to the multiple single cells based on the voltage shift rate matrix; determining a standard deviation of single cell shift rates corresponding to the multiple single cells respectively based on the average single cell shift rate; and determining abnormal batteries from the multiple single cells based on the standard deviation of single cell shift rates corresponding to the multiple single cells respectively.

[0012] Optionally, after determining an abnormal battery from the multiple single batteries based on the voltage shift rate matrix, it also includes: when there are multiple abnormal batteries, determining the distribution positions and environmental parameters of the multiple abnormal batteries; based on the distribution positions and environmental parameters of the multiple abnormal batteries, determining whether there is an influence result of abnormal influencing factors.

[0013] According to one aspect of an embodiment of the present invention, there is provided an electronic device, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement any of the above-described methods for determining the cause of an abnormal battery.

[0014] According to one aspect of an embodiment of the present invention, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute any of the above-mentioned methods for determining the cause of an abnormal battery.

[0015] In an embodiment of the present invention, a battery pack is obtained, wherein the battery pack includes a plurality of single cells; voltage timing data corresponding to the plurality of single cells are obtained; a voltage offset rate matrix corresponding to the voltage data matrix is ​​determined based on the plurality of voltage timing data, wherein the elements in the voltage offset rate matrix represent the voltage offset rates of the single cells with corresponding numbers at corresponding moments; an abnormal battery is determined from the plurality of single cells based on the voltage offset rate matrix; a first battery parameter of the abnormal battery and a second battery parameter of the normal battery are input into a target abnormal cause identification model to determine an abnormal cause causing the battery to be abnormal, wherein the normal battery is a battery other than the abnormal battery among the plurality of single cells, and the target abnormal cause identification model is based on a loss function, and a plurality of groups of sample data are used to train an initial abnormal cause identification model to obtain a model, wherein the plurality of groups of sample data include: sample battery parameters corresponding to the sample abnormal battery, and sample battery parameters corresponding to the sample normal battery. It can be seen that when determining abnormal batteries from single cells, only the voltage timing data of the single cells are considered to obtain the voltage offset rate, so that the abnormal batteries can be determined from multiple single cells. There is no need to compare the data differences between single cells, and the amount of calculation is small. Moreover, through the target abnormal cause identification model trained by the loss function, the cause of the abnormality is accurately detected, thereby solving the technical problem in the related technology that when determining whether there are abnormal batteries in multiple single cells, the steps of judging the abnormality are complicated, and after the abnormality is determined, it is difficult to accurately determine the cause of the abnormality. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0017] Figure 1 is a flow chart of a method for determining an abnormality cause of an abnormal battery according to an embodiment of the present invention;

[0018] Figure 2 4 is a structural block diagram of a device for determining an abnormal cause of an abnormal battery according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0020] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0021] Example 1

[0022] According to an embodiment of the present invention, an embodiment of a method for determining the abnormal cause of an abnormal battery is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0023] Figure 1 is a flow chart of a method for determining an abnormal cause of an abnormal battery according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0024] Step S102, obtaining a battery pack, wherein the battery pack includes a plurality of single cells;

[0025] In step S102 provided in the present application, a battery pack is obtained, and the battery pack includes a plurality of single cells.

[0026] Among them, the battery pack is a component that integrates multiple single cells, usually used to store and provide electrical energy. It is designed according to specific application requirements, such as electric vehicles, energy storage systems, portable electronic devices and other fields. The battery pack contains complex structures, such as battery modules (consisting of multiple single cells connected in series or parallel, or connected in a special structure), battery management system (BMS, used to monitor battery status, protect battery safety, optimize battery performance, etc.), thermal management system (used to control the temperature in the battery pack to prevent overheating or overcooling from damaging battery performance), shell and connecting wires, etc.

[0027] Among them, the single cell is the basic unit of the battery pack and is the smallest unit that can be charged and discharged independently. Common types of single cells include: lithium-ion batteries, nickel-metal hydride batteries, lead-acid batteries, etc. Among them, lithium-ion batteries are widely used in electric vehicles and portable electronic devices due to their high energy density, long cycle life, and no memory effect. The single cell contains key materials such as positive electrode, negative electrode, electrolyte and separator, which generate electrical energy through chemical reactions. During the charging and discharging process, lithium ions migrate between the positive and negative electrodes to achieve the storage and release of electrical energy. The performance of the single cell directly affects the performance of the entire battery pack, such as energy density, power density, cycle life, and safety. Therefore, it is extremely critical to identify abnormal batteries in a timely manner.

[0028] Step S104, obtaining voltage time series data corresponding to a plurality of single cells;

[0029] In step S104 provided in the present application, voltage timing data corresponding to a plurality of single cells are obtained.

[0030] Among them, the voltage time series data corresponding to multiple single cells are involved. The voltage of each single cell will change with the process of charging, discharging, and standing. Recording these changes (i.e., voltage time series data) can help us understand the specific performance of each battery. Time series data refers to a set of data points recorded in chronological order, which reflects the trend of voltage changes over time. These data are usually stored in the form of time series to facilitate subsequent data analysis and processing.

[0031] It should be noted that to obtain these data, a voltage monitoring circuit or sensor can be integrated into the battery management system to collect the voltage value of each single battery in real time. The data acquisition system (such as a data logger, microcontroller, etc.) will read these voltage values ​​regularly (such as every second, every minute, etc.) and store them in a local or remote database.

[0032] It should also be noted that the acquired battery timing data may be the voltage timing data corresponding to a plurality of single cells collected after the battery state changes. For example, when the battery state changes to charging, discharging, charging at a predetermined rate, discharging at a predetermined rate, and charging / discharging at different rates, through this setting, it is possible to observe whether there is any abnormality in the process of dynamic state adjustment of the single cell.

[0033] Step S106, determining a voltage shift rate matrix corresponding to the voltage data matrix according to the plurality of voltage time series data, wherein the elements in the voltage shift rate matrix represent the voltage shift rates of the single cells with corresponding numbers at corresponding moments;

[0034] In step S106 provided in the present application, a voltage offset rate matrix corresponding to the voltage data matrix is ​​determined based on a plurality of voltage timing data.

[0035] Among them, the voltage offset rate matrix is ​​involved. The voltage offset rate matrix is ​​a matrix that represents the offset rate of each single cell voltage relative to the average voltage of the battery pack. The offset rate is usually expressed as a percentage or decimal form, which reflects the relative difference between the single cell voltage and the average voltage. A positive offset rate means that the single cell voltage is higher than the average voltage, and a negative offset rate means that it is lower than the average voltage.

[0036] The calculation of the offset rate is also involved. Based on the voltage value of each single cell at each moment, the average voltage value of the battery pack at that moment can be calculated first. Then, the voltage value of the single cell is subtracted from the average voltage value, and the result is divided by the average voltage value (or multiplied by 100 to convert to a percentage) to obtain the voltage offset rate of the single cell at that moment. The offset rates of all single cells at all moments are filled into the corresponding positions of the voltage offset rate matrix.

[0037] The voltage deviation rate matrix provides an intuitive tool for evaluating battery consistency. By observing the elements in the matrix, you can quickly identify which single cells have voltages that deviate more from the average voltage, which may mean that these cells have performance degradation, internal faults, or unbalanced charging.

[0038] In addition, the voltage offset rate matrix can also be used to monitor the dynamic changes of the battery pack during the charging and discharging process, helping to understand the behavioral characteristics of the battery pack under different working conditions. In the battery management system, the voltage offset rate matrix is ​​an important basis for implementing balanced control, fault diagnosis and performance evaluation. By real-time monitoring and analysis of the voltage offset rate matrix, the battery pack can be managed in a refined manner to improve the overall performance and reliability of the battery pack. At the same time, the voltage offset rate matrix is ​​also one of the important input data for battery health management and remaining life prediction. By analyzing and mining historical offset rate data, a battery health status assessment model can be established to predict the remaining service life of the battery, providing a scientific basis for battery maintenance and replacement.

[0039] Step S108, determining an abnormal battery from a plurality of single batteries according to the voltage deviation rate matrix;

[0040] In step S108 provided in the present application, an abnormal battery is determined from a plurality of single batteries according to a voltage deviation rate matrix.

[0041] Among them, abnormal batteries are involved. Abnormal batteries refer to single cells in the battery pack whose voltage deviation rate significantly deviates from the normal range or shows unstable and irregular changes. These batteries may have performance degradation or malfunction due to internal failures, aging, uneven charging, etc.

[0042] In the voltage offset rate matrix, each element represents the voltage offset rate of a single cell at a specific moment. By comparing the offset rates of different single cells at the same moment, we can find out which cells have voltages that deviate more from the average voltage. At the same time, we can also observe the offset rate change trend of each single cell at different moments. If the offset rate of a battery is continuously high or low, or shows abnormal fluctuations, then this battery may be abnormal.

[0043] It should be noted that the method for determining an abnormal battery may include the following:

[0044] A reasonable voltage deviation rate threshold (which can be an absolute value threshold or a relative threshold) can be set, and a single battery with a deviation rate exceeding the threshold is considered abnormal. This method is simple and direct, but the setting of the threshold needs to be based on actual data and experience.

[0045] Statistical methods (such as mean, standard deviation, median, etc.) can also be used to calculate the distribution of voltage deviation rate, and batteries that are far from the center of the distribution or located at the edge of the distribution are considered abnormal. This method is more objective and scientific, but requires certain statistical knowledge.

[0046] You can also observe how the voltage drift rate changes over time. If the drift rate of a battery continues to rise or fall, or the fluctuation range increases abnormally, it may indicate that there is a problem with the battery.

[0047] It can be seen that when determining abnormal batteries from single cells, only the voltage value and the average voltage value of the single cell are considered to obtain the voltage deviation rate, so that the abnormal battery can be determined from multiple single cells. There is no need to compare the data differences between single cells, and the amount of calculation is small. This solves the technical problem in the related technology that when determining whether there are abnormal batteries among multiple single cells, the steps of judging the abnormality are complicated, and after the abnormality is determined, it is difficult to accurately determine the cause of the abnormality.

[0048] Step S110, inputting the first battery parameter of the abnormal battery and the second battery parameter of the normal battery into the target abnormal cause identification model to determine the abnormal cause of the battery abnormality, wherein the normal battery is a battery other than the abnormal battery among the multiple single batteries, and the target abnormal cause identification model is based on a loss function, and multiple groups of sample data are used to train the initial abnormal cause identification model to obtain a model, and the multiple groups of sample data include: sample battery parameters corresponding to the sample abnormal battery, and sample battery parameters corresponding to the sample normal battery;

[0049] In step S110 provided in the present application, a target abnormality cause identification model obtained by training with multiple groups of sample data based on a loss function is used.

[0050] Among them, the first battery parameter of the abnormal battery is determined. That is, the key internal parameters of the abnormal battery can be collected and recorded, which parameters generally include but are not limited to: Voltage: the current voltage value of the abnormal battery and its historical change trend. Internal resistance: the size of the internal resistance of the battery, which reflects the energy loss of the battery during the charging and discharging process. Temperature: the temperature of the battery, high temperature or low temperature may affect the performance of the battery. Capacity: the remaining capacity or health state of the battery, which reflects the long-term performance of the battery. These parameters can be directly read by the battery management system or measured by special testing equipment. The second battery parameter of the normal battery can also be determined, and the parameter can be the same as the above parameter type. That is, the same parameters of a group of normal batteries (i.e., other single cells except the abnormal battery) are collected as a comparison benchmark. The parameters of these normal batteries will be used to evaluate the degree of deviation of the abnormal battery parameters, so as to determine whether the abnormality is caused by abnormal internal battery parameters. After that, the first battery parameter and the second battery parameter can be compared to input the first battery parameter of the abnormal battery and the second battery parameter of the normal battery into the model to determine the cause of the abnormality by analyzing whether the abnormal battery parameters deviate significantly from the normal range, or by other means.

[0051] Among them, the training process is based on a loss function and multiple sets of sample data to iteratively optimize the initial abnormal cause identification model. The sample data includes: sample battery parameters corresponding to sample abnormal batteries: these parameters reflect the abnormal battery status in history or experiments. Sample battery parameters corresponding to sample normal batteries: in contrast, these parameters represent the status of normal batteries. The target abnormal cause identification model may be a classification model or a regression model, depending on how the abnormal cause is expressed (either a discrete category or a continuous quantity). During the model training process, the loss function is used to quantify the difference between the model prediction results and the actual results. By minimizing this difference, the model can gradually learn the mapping relationship from battery parameters to abnormal causes. Once the training is completed, the target abnormal cause identification model can be quickly applied to new battery data to help quickly locate and solve battery abnormal problems, thereby improving the reliability and safety of the battery system.

[0052] Through the above steps S102-S110, a battery pack is obtained, wherein the battery pack includes a plurality of single cells; voltage timing data corresponding to the plurality of single cells are obtained; based on the plurality of voltage timing data, a voltage shift rate matrix corresponding to the voltage data matrix is ​​determined, wherein the elements in the voltage shift rate matrix represent the voltage shift rates of the single cells with corresponding numbers at corresponding moments; based on the voltage shift rate matrix, abnormal batteries are determined from the plurality of single cells; the first battery parameter of the abnormal battery and the second battery parameter of the normal battery are input into the target abnormal cause identification model to determine the abnormal cause causing the battery abnormality, wherein the normal battery is a battery other than the abnormal battery among the plurality of single cells, and the target abnormal cause identification model is based on a loss function, and multiple groups of sample data are used to train the initial abnormal cause identification model to obtain a model, wherein the multiple groups of sample data include: sample battery parameters corresponding to the sample abnormal battery, and sample battery parameters corresponding to the sample normal battery. It can be seen that when determining abnormal batteries from single cells, only the voltage timing data of the single cells are considered to obtain the voltage offset rate, so that the abnormal batteries can be determined from multiple single cells. There is no need to compare the data differences between single cells, and the amount of calculation is small. Moreover, through the target abnormal cause identification model trained by the loss function, the cause of the abnormality is accurately detected, thereby solving the technical problem in the related technology that when determining whether there are abnormal batteries in multiple single cells, the steps of judging the abnormality are complicated, and after the abnormality is determined, it is difficult to accurately determine the cause of the abnormality.

[0053] As an optional embodiment, the target abnormal cause identification model is obtained by training with multiple groups of sample data based on a loss function in the following manner: an initial abnormal cause identification model is obtained, and a loss function for model training is constructed, wherein the loss function includes a first loss function term, a second loss function term, and a third loss function term, the first loss function term is used to maintain the similarity of battery parameters of the same type of batteries in the feature space, the second loss function term is used to maintain the heterogeneity of battery parameters of different types of batteries in the feature space, and the third loss function term is used to classify the abnormal causes of abnormal batteries; based on the loss function, the initial abnormal cause identification model is trained with multiple groups of sample data to obtain the target abnormal cause identification model.

[0054] In this embodiment, an initial abnormal cause identification model is obtained, which may be built based on a machine learning algorithm (such as a neural network, a support vector machine, etc.). It is trained based on a loss function to obtain the target. The loss function is the core of the model training process, which defines the degree of difference between the model prediction result and the actual result. In this scenario, the loss function is specially designed to contain three loss function terms to meet specific training needs. The first loss function term is used to maintain the similarity of battery parameters of the same type of battery in the feature space. This means that the model should learn a feature representation so that batteries of the same type (for example, both normal and abnormal) are close to each other in the feature space. This helps the model better understand and distinguish different types of batteries. The second loss function term is used to maintain the heterogeneity of battery parameters of different types of batteries in the feature space. In contrast to the first loss function term, this term supports the feature representation learned by the model to distinguish different types of batteries. Even if the battery parameters may be similar in the original space, the model should be able to distinguish them in the feature space. The third loss function term is used to classify the abnormal causes of abnormal batteries. This is one of the main goals of model training. The third loss function term measures the accuracy of the model in classifying the abnormal causes of abnormal batteries. By minimizing this loss, the model is able to learn the mapping from battery parameters to anomaly causes.

[0055] The feature space is the data representation space learned by the model. In this space, the distance and position relationship between data points (i.e., battery parameters of batteries) reflects their similarities and differences. By considering both the similarities of the same type of batteries and the heterogeneity of different types of batteries, the model is able to learn more robust and generalized feature representations. This helps improve the performance of the model on new data.

[0056] Then, multiple sets of sample data (including battery parameters of sample abnormal batteries and sample normal batteries) are used to train the initial abnormal cause identification model. During the training process, the model will continuously adjust its parameters to minimize the value of the loss function. This usually involves the application of optimization algorithms such as gradient descent. After enough training iterations, the model will converge to a stable state, at which time the value of the loss function reaches a minimum or close to a minimum. So that the final target abnormal cause identification model can accurately identify the cause of battery abnormality.

[0057] As an optional embodiment, the first battery parameter of the abnormal battery and the second battery parameter of the normal battery are input into a target abnormal cause identification model to determine the abnormal cause of the battery abnormality, including: inputting the first battery parameter and the second battery parameter into a first abnormal identification module of the target abnormal cause identification model to determine whether the abnormal result of the battery abnormality is due to the abnormal internal parameters of the battery; when the abnormal result indicates that the battery abnormality is not caused by the abnormal internal parameters of the battery, inputting the first battery parameter and the second battery parameter into a second abnormal identification module of the target abnormal cause identification model to determine whether the abnormal result of the battery abnormality is due to a connection line fault corresponding to the abnormal battery, so as to check the cause of the abnormal battery abnormality through the connection line.

[0058] In this embodiment, a first abnormality identification module and a second abnormality identification module of a target abnormality cause identification model are set. In the first abnormality identification module, it is analyzed whether the abnormal battery parameters deviate significantly from the normal range. If one or more parameters of the abnormal battery are significantly different from those of the normal battery, it can be preliminarily determined that the abnormality may be caused by abnormal internal parameters of the battery. Based on the above comparison, determine whether the abnormal result is an abnormal result of the battery abnormality due to abnormal internal parameters of the battery. If the abnormal result points to abnormal internal parameters of the battery, it may be necessary to further disassemble and analyze the battery or replace the battery. If the abnormal result indicates that the battery abnormality is not caused by abnormal internal parameters of the battery, the second abnormality identification module is entered for processing, then the external connection lines related to the abnormal battery can be focused on next. These connection lines may include connection lines between batteries, communication lines between batteries and battery management systems, etc.

[0059] Among them, the cause is checked through the connection line. The possible problems of the connection line are checked through the following steps: For example, check whether the connection line is physically damaged, such as broken, corroded or loose. Measure the resistance of the connection line to determine whether the resistance value is within the normal range. If the connection line involves signal transmission (such as communication between the battery management system and the battery), signal testing is required to verify the integrity and accuracy of the signal transmission.

[0060] Based on the results of the connection line inspection, the specific cause of the abnormal battery is determined. Once the cause is determined, appropriate measures can be taken to solve the problem, such as repairing or replacing damaged connection lines, adjusting the settings of the battery management system, etc.

[0061] By processing the above steps through the modules in the model, the cause of the abnormality can be determined more quickly and accurately, so as to avoid dangerous situations caused by battery abnormalities that are difficult to determine in time.

[0062] As an optional embodiment, a voltage offset rate matrix corresponding to a voltage data matrix is ​​determined based on multiple voltage timing data, including: constructing a voltage data matrix corresponding to multiple single cells based on multiple voltage timing data, wherein the voltage data matrix comprises multiple row vectors and multiple column vectors, each of the multiple row vectors represents the voltage value of single cells with different numbers at the same time, each of the multiple column vectors represents the voltage value of single cells with the same number at different times, and the elements in the voltage data matrix represent the voltage value of single cells with corresponding numbers at corresponding times; determining a voltage offset rate matrix corresponding to the voltage data matrix based on the average voltage value corresponding to the battery pack.

[0063] In this embodiment, a voltage data matrix corresponding to multiple single cells is constructed based on multiple voltage time series data. The voltage data matrix is ​​a matrix including multiple elements, which contains the voltage values ​​of all single cells at all recorded time points. The number of rows and columns of the matrix corresponds to the number of single cells and the number of time points, respectively. Each element of the matrix (i.e., a value in the matrix) represents the voltage value of a specific single cell at a specific moment.

[0064] In the voltage data matrix, each row vector represents the voltage value of all single cells at a time point. In other words, if each row of the matrix is ​​regarded as a "snapshot" at a time point, then this "snapshot" records the voltage state of each single cell at that time point. By comparing the row vectors at different time points, the change trend of each single cell voltage over time can be observed.

[0065] Similarly, each column vector represents the voltage sequence of a single cell at all recorded time points. This sequence reflects the voltage fluctuation of the single cell during the entire observation period. By analyzing the column vector, the performance stability, attenuation or abnormal fluctuation of a single cell can be evaluated.

[0066] The construction of the voltage data matrix enables a deeper analysis of the battery pack data. For example, cluster analysis can be used to identify groups of cells with similar performance, trend analysis can be used to predict the remaining life of the battery, or correlation analysis can be used to evaluate the mutual influence between cells.

[0067] Then, according to the average voltage value corresponding to the battery pack, the voltage offset rate matrix corresponding to the voltage data matrix is ​​determined, wherein the elements in the voltage offset rate matrix represent the voltage offset rate of the single cell with the corresponding number at the corresponding moment; wherein, the average voltage value of the battery pack is involved, and the average voltage value of the battery pack refers to the average value of the voltage values ​​of all single cells at a certain moment. It reflects the overall voltage level of the battery pack and is an important indicator for evaluating the performance of the battery pack. Then, the voltage offset rate can be accurately calculated to construct an accurate voltage offset matrix.

[0068] As an optional embodiment, determining a voltage offset rate matrix corresponding to a voltage data matrix includes: determining an offset rate of each element in the voltage data matrix compared to an average voltage value to obtain an offset rate corresponding to each element; and determining a voltage offset rate matrix based on the offset rate corresponding to each element.

[0069] In this embodiment, each element in the voltage data matrix is ​​determined. It is necessary to clarify the structure of the voltage data matrix. Each row of this matrix represents the voltage value of a single cell at different time points, and each column represents the voltage value of all single cells at a specific time point. Therefore, each element in the matrix corresponds to the voltage value of a specific single cell at a specific time point.

[0070] The average voltage value can be calculated through the above, that is, before calculating the offset rate, it is necessary to determine the average voltage value of the battery pack at each time point. This is usually obtained by adding the voltage values ​​of all single cells at the same time point and then dividing it by the total number of single cells. In this way, a sequence of average voltage values ​​corresponding to the time dimension of the voltage data matrix is ​​obtained.

[0071] After that, the offset rate of each element can be calculated. Iterate through each element in the voltage data matrix. For each element, perform the following steps to calculate its offset rate: Find the average voltage value corresponding to the time point where the element is located. Then, subtract the average voltage value from the voltage value of the element to get the voltage difference. Finally, divide the voltage difference by the average voltage value (or multiply it by 100 to convert it into a percentage) to get the offset rate of the element. This offset rate represents the degree of deviation of the voltage value of the single cell at a specific time point relative to the average voltage of the entire battery pack.

[0072] Then determine the voltage offset rate matrix. Through the above steps, a corresponding offset rate is calculated for each element in the voltage data matrix. Next, these offset rates are reorganized into a new matrix, namely the voltage offset rate matrix, according to their positions in the voltage data matrix. This matrix has the same number of rows and columns as the voltage data matrix, but its elements are no longer voltage values, but offset rates.

[0073] Among them, the voltage deviation rate matrix provides an intuitive and quantitative way to evaluate the performance consistency of each single cell in the battery pack. By observing the elements in the deviation rate matrix, it is possible to quickly identify which single cells have voltages that deviate more from the average voltage, and then analyze the possible problems of these cells. This is very important for fault diagnosis and balancing control of the battery management system.

[0074] As an optional embodiment, determining the average voltage value corresponding to the battery pack includes any one of the following: determining the average voltage value corresponding to the battery pack based on a voltage data matrix; determining the average voltage value corresponding to the battery pack based on a predicted voltage value of the battery pack; determining the average voltage value corresponding to the battery pack based on the average voltage value of the battery pack in a leading time period and a voltage change parameter corresponding to the battery pack, wherein the leading time period is a time period before the corresponding moment of obtaining the voltage timing data.

[0075] In this embodiment, various methods of determining the average voltage value corresponding to the battery pack are described.

[0076] Optionally, the average voltage value corresponding to the battery pack can be determined based on the voltage data matrix. This method calculates the average voltage value directly based on the voltage data matrix obtained in real time. The voltage data matrix contains the voltage values ​​of all single cells at different time points. For each time point, the voltage values ​​of all single cells at that time point are added together, and then divided by the total number of single cells to obtain the average voltage value of the battery pack at that time point.

[0077] This method is suitable for real-time monitoring systems and can instantly reflect the current status of the battery pack. During the driving process of electric vehicles, by continuously collecting voltage data and calculating the average voltage value, the performance changes of the battery pack can be monitored in real time.

[0078] Optionally, the average voltage value corresponding to the battery pack can also be determined based on the predicted voltage value of the battery pack. This method estimates the voltage value of the battery pack at a certain point in the future based on a certain prediction model or algorithm, and calculates the average voltage value accordingly. The prediction model may be constructed based on various factors such as historical data, battery characteristics, and operating conditions.

[0079] This method is very useful in scenarios where the future performance of a battery pack needs to be predicted, such as power prediction and charging strategy optimization in energy management systems. By predicting the average voltage value, the charging strategy can be adjusted in advance and energy distribution can be optimized, thereby improving the efficiency and safety of the battery pack. Moreover, by predicting, there is no need to collect and calculate the voltage value in real time, which can save a certain amount of calculation.

[0080] Optionally, the average voltage value corresponding to the battery pack is determined based on the average voltage value of the battery pack in the preceding time period and the voltage change parameters corresponding to the battery pack. This method combines historical data and voltage change parameters to predict or estimate the average voltage value of the battery pack at the current or future time point. First, obtain the average voltage value of the battery pack in the preceding time period (i.e., the period before the current time point); then, estimate the average voltage value at the current or future time point based on the voltage change parameters of the battery pack (such as voltage decay rate, the impact of temperature change on voltage, etc.).

[0081] This method is suitable for scenarios that require comprehensive consideration of historical data and real-time change factors. By combining the average voltage value and voltage change parameters of the preceding time period, the current state of the battery pack can be more accurately evaluated and its future performance can be predicted. This is of great significance for formulating long-term maintenance plans and optimizing battery pack configuration.

[0082] Specifically, an adaptive method can be selected to determine the average voltage value according to actual applications and scenarios.

[0083] As an optional embodiment, an abnormal battery is determined from a plurality of single cells based on a voltage shift rate matrix, including: determining an average single cell shift rate corresponding to the plurality of single cells based on the voltage shift rate matrix; determining a single cell shift rate standard deviation corresponding to the plurality of single cells respectively based on the single cell average shift rate; and determining an abnormal battery from the plurality of single cells based on a single cell shift rate standard deviation corresponding to the plurality of single cells respectively.

[0084] In this embodiment, a specific method of determining an abnormal battery from a plurality of single cells is described. First, the average offset rate of the single cells corresponding to the plurality of single cells can be determined. First, the offset rate data of each single cell at all time points needs to be extracted from the voltage offset rate matrix. Then, these data are averaged to obtain the average offset rate of each single cell. This average offset rate reflects the long-term deviation of the single cell from the average voltage of the battery pack.

[0085] After that, the standard deviation of the offset rate of each single cell corresponding to each of the multiple single cells can be determined. After obtaining the average offset rate of each single cell, the standard deviation of these offset rate data needs to be further calculated. The standard deviation is a statistic that measures the degree of data dispersion, which indicates the fluctuation of the offset rate of the single cell around its average value. The larger the standard deviation, the greater the fluctuation of the offset rate of the single cell at different time points, and there may be anomalies.

[0086] Then, based on the standard deviation of the offset rate of the single cells corresponding to the multiple single cells, the abnormal cells are determined from the multiple single cells. Finally, the abnormal cells are identified based on the offset rate standard deviation of each single cell. Usually, a standard deviation threshold is set, and the single cells whose offset rate standard deviation exceeds the threshold are considered abnormal. This threshold can be set based on the actual situation of the battery pack, historical data, and safety standards.

[0087] It should be noted that relying solely on the standard deviation of the offset rate to determine abnormal batteries may not be comprehensive enough. In practical applications, other indicators (such as voltage change rate, temperature, etc.) can also be combined to comprehensively evaluate the health status of single cells.

[0088] As an optional embodiment, after determining an abnormal battery from a plurality of single cells based on a voltage shift rate matrix, the method further includes: in the case where there are multiple abnormal batteries, determining the distribution positions and environmental parameters of the multiple abnormal batteries; and determining whether there are any results of influence of abnormal influencing factors based on the distribution positions and environmental parameters of the multiple abnormal batteries.

[0089] In this embodiment, the steps after the abnormal battery is determined are described. After this, the distribution positions of multiple abnormal batteries can be determined, and the specific position of each abnormal battery in the battery pack can be clarified. This usually involves a physical layout diagram or digital model of the battery pack, in which each single battery has a unique identifier and location coordinates. By matching the identifiers of these abnormal batteries with the location information in the layout diagram or model, their distribution positions in the battery pack can be determined.

[0090] Next, environmental parameters associated with abnormal battery distribution locations can be collected. These parameters may include, but are not limited to: Temperature: The temperature in different locations may be different, and high or low temperatures may have an adverse effect on battery performance. Humidity: Excessive humidity may cause internal corrosion or short circuits in the battery. Ventilation: Poor ventilation may cause heat accumulation and affect battery heat dissipation. Vibration and shock: Certain locations may be more susceptible to vibration and shock during vehicle driving, which may affect battery connections and internal structures.

[0091] Then analyze the relationship between the distribution location of abnormal batteries and environmental parameters. That is, conduct a comprehensive analysis of the distribution location of abnormal batteries and environmental parameters to determine whether there is a certain correlation or pattern. For example: if the abnormal batteries are mainly concentrated in a certain area of ​​the battery pack, and the temperature in this area is significantly higher than other areas, then it can be inferred that high temperature may be one of the causes of the abnormality of these batteries. If the abnormal batteries are distributed in poorly ventilated areas, then consider improving ventilation conditions to alleviate the problem. If there is no obvious correlation between the distribution of abnormal batteries and changes in environmental parameters, then other potential factors may need to be further considered, such as battery manufacturing defects, improper use, etc.

[0092] Based on the above analysis, it is possible to preliminarily determine whether there are abnormal influencing factors and their specific impact on battery performance. If abnormal influencing factors are identified, then corresponding measures need to be formulated to eliminate or mitigate their impact. For example: adjust the cooling system of the battery pack to improve temperature distribution. Optimize the layout of the battery pack to improve ventilation conditions. Perform individual inspections and maintenance on abnormal batteries and replace them if necessary.

[0093] As an optional embodiment, after determining the abnormal battery from multiple single cells based on the voltage shift rate matrix, it also includes: determining the voltage shift rate matrices corresponding to multiple time periods respectively; and determining the voltage shift trend data corresponding to the multiple single cells respectively based on the voltage shift rate matrices corresponding to the multiple time periods respectively.

[0094] In this embodiment, the steps after determining the abnormal battery are described. After that, the voltage shift rate matrix corresponding to multiple time periods can be determined. The original voltage data or voltage shift rate data can be divided into multiple time periods in chronological order. These time periods can be set according to actual needs, such as daily, hourly, minute-by-minute or shorter time intervals. For each time period, the voltage shift rate matrix is ​​recalculated, that is, the voltage shift rate of each single cell at each time point within the time period.

[0095] Then extract the voltage offset trend data corresponding to multiple single cells. After obtaining the voltage offset rate matrix for each time period, the trend data of the voltage offset rate over time can be extracted for each single cell. These data are usually presented in the form of a time series, showing the changing trend of the voltage offset rate of the single cell in different time periods. Then analyze the voltage offset trend data. By analyzing the voltage offset trend data, observe whether the changing trend of the voltage offset rate of the single cell is rising, falling or fluctuating. This helps to judge the health status and performance changes of the battery. In addition, the time points when the voltage offset rate suddenly changes or fluctuates abnormally can be identified, which may be related to internal battery failure, unbalanced charging or changes in external conditions. Moreover, compare the voltage offset trends between different single cells to find the correlation or difference between them. This helps to identify potential problem areas or performance bottlenecks in the battery pack.

[0096] Based on the above embodiments and optional embodiments, an optional implementation is provided, which is described in detail below.

[0097] In an optional embodiment of the present invention, an abnormal cell diagnosis and evaluation method based on the dynamic voltage offset rate method is provided, which can utilize the dynamic offset rate theory to solve the problems existing in the related art. For a relatively symmetrical distribution, it is relatively simple to screen and locate abnormal data, and in the normally distributed data, the inspection of abnormal data represented by the Raida criterion method is relatively simple and efficient, and has great advantages in application to engineering practice. However, during the operation of the battery, due to the large number of limiting factors of the battery pack, the battery data has a large correlation when the actual vehicle is running, so a new abnormal battery fault diagnosis method is needed for the data that does not conform to the normal distribution. The dynamic offset rate theory can solve this problem well. The dynamic offset rate theory is a method for processing abnormal values ​​of non-normally distributed data. Specifically, the performance of the battery cell is analyzed based on real-time data, and the abnormal cell is accurately located using the dynamic offset rate related theoretical algorithm, thereby realizing the analysis of the differences in the power battery pack.

[0098] The optional implementation of the present invention introduces the concept of a 10-minute dynamic offset rate, which is specifically defined as an indicator of the degree of deviation of the voltage of each battery cell at that moment from its overall moving average. The overall moving average refers to the overall voltage level of the battery pack within a certain period of time. The offset rate is used to reflect how much the battery cell deviates from the overall at that moment, and the dynamic iterative offset rate can be used to reflect the voltage trend of the battery cell. Applying the offset rate concept, the 10-minute offset rate formula is:

[0099]

[0100] Among them, Y represents the 10-minute offset rate, U represents the voltage data of the i-th cell in the 10th minute, It indicates the average value of all monomer voltages in the previous minute.

[0101] The moving average voltage of the reference data is used to represent the state value at the moment before the calculation time point. The voltage change degree at that moment is calculated by taking the difference between each cell voltage and the state value and then calculating the ratio, thereby reflecting the change degree of the cell voltage. It further indicates that the greater the deviation rate is from the moving average, the greater the possibility of abnormality of the cell battery. The change of the deviation rate is then expressed by calculating the standard deviation of all dynamic deviation rate values. A large standard deviation indicates that the voltage change range of the cell is relatively large and the abnormality is more obvious.

[0102] The following explains it:

[0103] S1: Obtain battery-related data through the big data cloud platform and perform data preprocessing on the voltage, time and other signal data collected in real time.

[0104] The big data cloud platform refers to a cloud platform built using big data technology to process and store power battery information. The required signal data such as voltage and time are obtained from the cloud platform, and data cleaning is performed on the data. Duplicate values ​​are directly deleted, and the processing of missing values ​​is divided into two cases: if the data is missing for 1 minute continuously, the missing value time period is directly removed without subsequent algorithms; if the data is not missing continuously, the mean value interpolation method is used to interpolate the missing values.

[0105] S2: Select reference data to form a single cell voltage data matrix at the corresponding moment.

[0106] The first 60 historical data points of the vehicle are selected as the benchmark data to form the single cell voltage data matrix U at the corresponding time:

[0107]

[0108] The first subscript of u is the time point, and the second subscript m is the number of single cells in the battery pack.

[0109] S3: Processing the cell voltage data matrix based on the dynamic offset rate algorithm to obtain the battery cell offset rate matrix at the corresponding moment.

[0110] It is divided into two steps:

[0111] S31: Calculate t 61 10 minutes before the time point (i.e. t1~t 60 ) The average value of all cell voltages That is, calculate the voltage mean of the reference matrix U;

[0112] S32: Calculate the 10-minute offset rate of each cell of the battery pack for the first time, t 61 At this moment, the voltage of the battery cells numbered 1, 2, 3, ..., m is [u 61,1 ,u 61,2 ,...,u 61,m ], calculate the 10-minute offset rate Y of battery cell number i 1,i , forming the 10-minute offset rate matrix Y of all single cells at that moment (1) =[Y 1,1 ,Y 1,2 ,...,Y 1,m ].

[0113]

[0114] Among them, Y 1,i The first subscript is the time for calculating the 10-minute offset rate of the battery pack, and the second subscript is the cell number.

[0115] Repeat the above two steps several times until the output is Y (n)(n is the number of times the 10-minute offset rate of each cell in the battery pack is calculated).

[0116] S4: Calculate the deviation rate index of different cells based on the battery cell deviation rate matrix, and then locate the abnormal cell.

[0117] It is divided into two steps:

[0118] S41: In calculating the 10-minute offset rate n times, a 10-minute offset rate matrix of each battery cell is formed. The 10-minute offset rate matrix of battery cell numbered i is Y i =[Y 1,i ,Y 2,i ,...,Y n,i ], from which the average deviation rate of the monomer can be calculated And the standard deviation of the deviation rate S i , the specific formula is as follows:

[0119]

[0120] Where j is a natural number.

[0121] S42: Arrange the obtained standard deviation values ​​from large to small, select the five battery cells with the largest 10-minute offset rate dispersion as abnormal values, find the corresponding battery cell numbers, and then the five battery cells are abnormal cells.

[0122] S5, inputting the first battery parameter of the abnormal battery and the second battery parameter of the normal battery into the target abnormal cause identification model to determine the abnormal cause causing the battery abnormality, wherein the normal battery is a battery other than the abnormal battery among multiple single cells, and the target abnormal cause identification model is based on a loss function, and multiple groups of sample data are used to train the initial abnormal cause identification model to obtain a model, and the multiple groups of sample data include: sample battery parameters corresponding to the sample abnormal battery, and sample battery parameters corresponding to the sample normal battery.

[0123] By determining the first battery parameter of the abnormal battery. That is, the key internal parameters of the abnormal battery can be collected and recorded, which parameters generally include but are not limited to: Voltage: the current voltage value of the abnormal battery and its historical change trend. Internal resistance: the size of the internal resistance of the battery, which reflects the energy loss of the battery during the charging and discharging process. Temperature: the temperature of the battery, high temperature or low temperature may affect the performance of the battery. Capacity: the remaining capacity or health state of the battery, which reflects the long-term performance of the battery. These parameters can be directly read by the battery management system or measured by special testing equipment. And determine the second battery parameter of the normal battery, the parameter can be the same as the above parameter type. That is, collect the same parameters of a group of normal batteries (i.e., other single cells except the abnormal battery) as a comparison benchmark. These normal battery parameters will be used to evaluate the degree of deviation of the abnormal battery parameters, so as to determine whether the abnormality is caused by abnormal internal battery parameters. After that, the first battery parameter and the second battery parameter can be compared to input the first battery parameter of the abnormal battery and the second battery parameter of the normal battery into the model to determine the cause of the abnormality by analyzing whether the abnormal battery parameters deviate significantly from the normal range, or by other means. Moreover, the target abnormality cause identification model trained by the loss function can accurately detect the cause of the abnormality, thereby solving the technical problem in the related technology that when determining whether there are abnormal batteries among multiple single batteries, the steps of judging the abnormality are complicated, and after the abnormality is determined, it is difficult to accurately determine the cause of the abnormality.

[0124] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0125] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods of various embodiments of the present invention.

[0126] Example 2

[0127] According to an embodiment of the present invention, there is also provided a device for implementing the above-mentioned method for determining the abnormal cause of an abnormal battery. Figure 2 is a structural block diagram of a device for determining an abnormal cause of an abnormal battery according to an embodiment of the present invention. Figure 2 As shown, the device includes: a first acquisition module 202, a second acquisition module 204, a first determination module 206, a second determination module 208 and a third determination module 210. The device is described in detail below.

[0128] The first acquisition module 202 is used to acquire a battery pack, wherein the battery pack includes a plurality of single cells; the second acquisition module 204 is connected to the first acquisition module 202 and is used to acquire voltage timing data corresponding to the plurality of single cells respectively; the first determination module 206 is connected to the second acquisition module 204 and is used to determine a voltage offset rate matrix corresponding to the voltage data matrix based on the plurality of voltage timing data, wherein the elements in the voltage offset rate matrix represent the voltage offset rate of the single cell with the corresponding number at the corresponding moment; the second determination module 208 is connected to the first determination module 206 and is used to determine the voltage offset rate matrix from the voltage data matrix based on the voltage offset rate matrix. An abnormal battery is determined from multiple single cells; a third determination module 210 is connected to the above-mentioned second determination module 208, and is used to input the first battery parameter of the abnormal battery and the second battery parameter of the normal battery into the target abnormal cause identification model to determine the abnormal cause of the battery abnormality, wherein the normal battery is a battery other than the abnormal battery among multiple single cells, and the target abnormal cause identification model is based on a loss function, and multiple groups of sample data are used to train the initial abnormal cause identification model to obtain a model, and the multiple groups of sample data include: sample battery parameters corresponding to the sample abnormal battery, and sample battery parameters corresponding to the sample normal battery.

[0129] It should be noted here that the above-mentioned first acquisition module 202, second acquisition module 204, first determination module 206, second determination module 208 and third determination module 210 correspond to steps S102 to S110 in the method for determining the abnormal cause of an abnormal battery, and the instances and application scenarios implemented by the multiple modules are the same as the corresponding steps, but are not limited to the contents disclosed in the above-mentioned embodiment 1.

[0130] Example 3

[0131] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising: a processor; a memory for storing instructions executable by the processor, wherein the processor is configured to execute instructions to implement any of the following methods for determining an abnormal cause of an abnormal battery: obtaining a battery pack, wherein the battery pack includes a plurality of single cells; obtaining voltage timing data corresponding to the plurality of single cells respectively; determining a voltage offset rate matrix corresponding to a voltage data matrix based on the plurality of voltage timing data, wherein the elements in the voltage offset rate matrix represent the voltage offset rates of the single cells under corresponding numbers at corresponding moments; determining an abnormal battery from a plurality of single cells based on the voltage offset rate matrix; inputting a first battery parameter of the abnormal battery and a second battery parameter of the normal battery into a target abnormal cause identification model to determine an abnormal cause causing the battery to be abnormal, wherein the normal battery is a battery other than the abnormal battery among the plurality of single cells, and the target abnormal cause identification model is based on a loss function, and a plurality of groups of sample data are used to train an initial abnormal cause identification model to obtain a model, wherein the plurality of groups of sample data include: sample battery parameters corresponding to the sample abnormal battery, and sample battery parameters corresponding to the sample normal battery.

[0132] Optionally, the target abnormal cause identification model is obtained by training with multiple groups of sample data based on the loss function in the following manner: an initial abnormal cause identification model is obtained, and a loss function for model training is constructed, wherein the loss function includes a first loss function term, a second loss function term, and a third loss function term, the first loss function term is used to maintain the similarity of battery parameters of batteries of the same type in the feature space, the second loss function term is used to maintain the heterogeneity of battery parameters of batteries of different types in the feature space, and the third loss function term is used to classify the abnormal causes of abnormal batteries; based on the loss function, the initial abnormal cause identification model is trained with multiple groups of sample data to obtain the target abnormal cause identification model.

[0133] Optionally, the first battery parameter of the abnormal battery and the second battery parameter of the normal battery are input into a target abnormal cause identification model to determine the abnormal cause of the battery abnormality, including: inputting the first battery parameter and the second battery parameter into a first abnormal identification module of the target abnormal cause identification model to determine whether the abnormal result of the battery abnormality is due to abnormal internal parameters of the battery; when the abnormal result indicates that the battery abnormality is not caused by abnormal internal parameters of the battery, inputting the first battery parameter and the second battery parameter into a second abnormal identification module of the target abnormal cause identification model to determine whether the abnormal result of the battery abnormality is due to a connection line fault corresponding to the abnormal battery, so as to check the cause of the abnormal battery abnormality through the connection line.

[0134] Optionally, a voltage offset rate matrix corresponding to the voltage data matrix is ​​determined based on multiple voltage timing data, including: constructing a voltage data matrix corresponding to multiple single cells based on multiple voltage timing data, wherein the voltage data matrix comprises multiple row vectors and multiple column vectors, each of the multiple row vectors represents the voltage value of single cells with different numbers at the same time, each of the multiple column vectors represents the voltage value of single cells with the same number at different times, and the elements in the voltage data matrix represent the voltage value of single cells with corresponding numbers at corresponding times; and determining the voltage offset rate matrix corresponding to the voltage data matrix based on the average voltage value corresponding to the battery pack.

[0135] Optionally, determining a voltage offset rate matrix corresponding to the voltage data matrix includes: determining an offset rate of each element in the voltage data matrix compared to an average voltage value to obtain an offset rate corresponding to each element; and determining the voltage offset rate matrix based on the offset rate corresponding to each element.

[0136] Optionally, determining the average voltage value corresponding to the battery pack includes any one of the following: determining the average voltage value corresponding to the battery pack based on a voltage data matrix; determining the average voltage value corresponding to the battery pack based on a predicted voltage value of the battery pack; determining the average voltage value corresponding to the battery pack based on the average voltage value of the battery pack in a leading time period and a voltage change parameter corresponding to the battery pack, wherein the leading time period is a time period before the corresponding moment of obtaining the voltage timing data.

[0137] Optionally, an abnormal battery is determined from a plurality of single cells based on a voltage shift rate matrix, including: determining an average single cell shift rate corresponding to the plurality of single cells based on the voltage shift rate matrix; determining a single cell shift rate standard deviation corresponding to the plurality of single cells respectively based on the single cell average shift rate; and determining an abnormal battery from the plurality of single cells based on a single cell shift rate standard deviation corresponding to the plurality of single cells respectively.

[0138] Optionally, after determining an abnormal battery from a plurality of single batteries based on the voltage shift rate matrix, the method further includes: in the case where there are multiple abnormal batteries, determining the distribution positions and environmental parameters of the multiple abnormal batteries; and determining whether there is an influence result of an abnormal influencing factor based on the distribution positions and environmental parameters of the multiple abnormal batteries.

[0139] Example 4

[0140] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided. When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is able to execute any of the above-mentioned methods for determining the abnormal cause of an abnormal battery: obtaining a battery pack, wherein the battery pack includes a plurality of single cells; obtaining voltage timing data corresponding to the plurality of single cells respectively; determining a voltage offset rate matrix corresponding to the voltage data matrix based on the plurality of voltage timing data, wherein the elements in the voltage offset rate matrix represent the voltage offset rates of the single cells under the corresponding numbers at the corresponding moments; determining an abnormal battery from the plurality of single cells based on the voltage offset rate matrix; inputting a first battery parameter of the abnormal battery and a second battery parameter of the normal battery into a target abnormal cause identification model to determine the abnormal cause causing the battery to be abnormal, wherein the normal battery is a battery other than the abnormal battery among the plurality of single cells, and the target abnormal cause identification model is based on a loss function, and multiple groups of sample data are used to train an initial abnormal cause identification model to obtain a model, wherein the multiple groups of sample data include: sample battery parameters corresponding to the sample abnormal battery, and sample battery parameters corresponding to the sample normal battery.

[0141] Optionally, the target abnormal cause identification model is obtained by training with multiple groups of sample data based on the loss function in the following manner: an initial abnormal cause identification model is obtained, and a loss function for model training is constructed, wherein the loss function includes a first loss function term, a second loss function term, and a third loss function term, the first loss function term is used to maintain the similarity of battery parameters of batteries of the same type in the feature space, the second loss function term is used to maintain the heterogeneity of battery parameters of batteries of different types in the feature space, and the third loss function term is used to classify the abnormal causes of abnormal batteries; based on the loss function, the initial abnormal cause identification model is trained with multiple groups of sample data to obtain the target abnormal cause identification model.

[0142] Optionally, the first battery parameter of the abnormal battery and the second battery parameter of the normal battery are input into a target abnormal cause identification model to determine the abnormal cause of the battery abnormality, including: inputting the first battery parameter and the second battery parameter into a first abnormal identification module of the target abnormal cause identification model to determine whether the abnormal result of the battery abnormality is due to abnormal internal parameters of the battery; when the abnormal result indicates that the battery abnormality is not caused by abnormal internal parameters of the battery, inputting the first battery parameter and the second battery parameter into a second abnormal identification module of the target abnormal cause identification model to determine whether the abnormal result of the battery abnormality is due to a connection line fault corresponding to the abnormal battery, so as to check the cause of the abnormal battery abnormality through the connection line.

[0143] Optionally, a voltage offset rate matrix corresponding to the voltage data matrix is ​​determined based on multiple voltage timing data, including: constructing a voltage data matrix corresponding to multiple single cells based on multiple voltage timing data, wherein the voltage data matrix comprises multiple row vectors and multiple column vectors, each of the multiple row vectors represents the voltage value of single cells with different numbers at the same time, each of the multiple column vectors represents the voltage value of single cells with the same number at different times, and the elements in the voltage data matrix represent the voltage value of single cells with corresponding numbers at corresponding times; and determining the voltage offset rate matrix corresponding to the voltage data matrix based on the average voltage value corresponding to the battery pack.

[0144] Optionally, determining a voltage offset rate matrix corresponding to the voltage data matrix includes: determining an offset rate of each element in the voltage data matrix compared to an average voltage value to obtain an offset rate corresponding to each element; and determining the voltage offset rate matrix based on the offset rate corresponding to each element.

[0145] Optionally, determining the average voltage value corresponding to the battery pack includes any one of the following: determining the average voltage value corresponding to the battery pack based on a voltage data matrix; determining the average voltage value corresponding to the battery pack based on a predicted voltage value of the battery pack; determining the average voltage value corresponding to the battery pack based on the average voltage value of the battery pack in a leading time period and a voltage change parameter corresponding to the battery pack, wherein the leading time period is a time period before the corresponding moment of obtaining the voltage timing data.

[0146] Optionally, an abnormal battery is determined from a plurality of single cells based on a voltage shift rate matrix, including: determining an average single cell shift rate corresponding to the plurality of single cells based on the voltage shift rate matrix; determining a single cell shift rate standard deviation corresponding to the plurality of single cells respectively based on the single cell average shift rate; and determining an abnormal battery from the plurality of single cells based on a single cell shift rate standard deviation corresponding to the plurality of single cells respectively.

[0147] Optionally, after determining an abnormal battery from a plurality of single batteries according to the voltage shift rate matrix, the method further includes: in the case where there are multiple abnormal batteries, determining the distribution positions and environmental parameters of the multiple abnormal batteries; and determining whether there is an influence result of an abnormal influencing factor according to the distribution positions and environmental parameters of the multiple abnormal batteries.

[0148] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0149] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0150] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0151] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0152] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0153] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0154] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for determining the abnormal cause of an abnormal battery, characterized in that: include: Obtaining a battery pack, wherein the battery pack includes a plurality of single cells; Acquire voltage time series data corresponding to each of the plurality of single cells; Determine a voltage shift rate matrix corresponding to the voltage data matrix according to the plurality of voltage time series data, wherein the elements in the voltage shift rate matrix represent the voltage shift rates of the single cells with corresponding numbers at corresponding moments; determining an abnormal battery from the plurality of single batteries according to the voltage deviation rate matrix; The first battery parameter of the abnormal battery and the second battery parameter of the normal battery are input into the target abnormal cause identification model to determine the abnormal cause of the battery abnormality, wherein the normal battery is a battery among the multiple single cells except the abnormal battery, and the target abnormal cause identification model is a model obtained by training the initial abnormal cause identification model based on a loss function using multiple groups of sample data, and the multiple groups of sample data include: sample battery parameters corresponding to sample abnormal batteries, and sample battery parameters corresponding to sample normal batteries.

2. The method according to claim 1, characterized in that The method further includes: using the multiple groups of sample data for training based on a loss function in the following manner to obtain the target abnormal cause identification model: Acquire the initial abnormal cause identification model, and construct a loss function for model training, wherein the loss function includes a first loss function term, a second loss function term, and a third loss function term, the first loss function term is used to maintain the similarity of battery parameters of the same type of batteries in the feature space, the second loss function term is used to maintain the heterogeneity of battery parameters of different types of batteries in the feature space, and the third loss function term is used to classify the abnormal causes of the abnormal battery; Based on the loss function, the initial abnormal cause identification model is trained using the multiple groups of sample data to obtain the target abnormal cause identification model.

3. The method according to claim 1, characterized in that Inputting the first battery parameter of the abnormal battery and the second battery parameter of the normal battery into the target abnormality cause identification model to determine the abnormality cause causing the battery abnormality, including: Inputting the first battery parameter and the second battery parameter into the first abnormality identification module of the target abnormality cause identification model to determine whether the abnormal result of the battery abnormality is caused by the abnormality of the internal battery parameters; When the abnormal result indicates that the battery abnormality is not caused by the abnormal internal parameters of the battery, the first battery parameter and the second battery parameter are input into the second abnormality identification module of the target abnormal cause identification model to determine whether the abnormal result of the battery abnormality is caused by a connection line fault corresponding to the abnormal battery, so as to investigate the cause of the abnormality of the abnormal battery through the connection line.

4. The method according to claim 1, characterized in that: Determining a voltage offset rate matrix corresponding to the voltage data matrix according to a plurality of voltage time series data includes: According to the plurality of voltage time series data, a voltage data matrix corresponding to the plurality of single cells is constructed, wherein the voltage data matrix comprises a plurality of row vectors and a plurality of column vectors, each of the plurality of row vectors represents a voltage value of single cells with different numbers at the same time, each of the plurality of column vectors represents a voltage value of single cells with the same number at different times, and an element in the voltage data matrix represents a voltage value of a single cell with a corresponding number at a corresponding time; A voltage offset rate matrix corresponding to the voltage data matrix is ​​determined according to an average voltage value corresponding to the battery pack.

5. The method according to claim 4, characterized in that Determining a voltage offset rate matrix corresponding to the voltage data matrix includes: Determine a deviation rate of each element in the voltage data matrix compared to the average voltage value, and obtain a deviation rate corresponding to each element; The voltage offset rate matrix is ​​determined according to the offset rate corresponding to each element.

6. The method according to claim 4, characterized in that Determining the average voltage value corresponding to the battery pack includes any one of the following: Determining an average voltage value corresponding to the battery pack according to the voltage data matrix; Determining an average voltage value corresponding to the battery pack according to the predicted voltage value of the battery pack; The average voltage value corresponding to the battery pack is determined based on the average voltage value of the battery pack in the preceding time period and the voltage variation parameter corresponding to the battery pack, wherein the preceding time period is the time period before the corresponding moment of obtaining the voltage timing data.

7. The method according to claim 4, characterized in that Determining an abnormal battery from the plurality of single batteries according to the voltage deviation rate matrix includes: Determining an average single cell offset rate corresponding to the plurality of single cells according to the voltage offset rate matrix; Determining the single cell battery offset rate standard deviations corresponding to the plurality of single cells respectively according to the single cell battery average offset rate; An abnormal battery is determined from the plurality of single batteries according to the standard deviations of the single battery offset rates respectively corresponding to the plurality of single batteries.

8. The method according to any one of claims 1 to 6, characterized in that: After determining an abnormal battery from the plurality of single batteries according to the voltage shift rate matrix, the method further includes: In the case where there are multiple abnormal batteries, determining the distribution locations and environmental parameters of the multiple abnormal batteries; According to the distribution positions and environmental parameters of the plurality of abnormal batteries, it is determined whether there is an impact result of an abnormal impact factor.

9. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the method for determining the abnormal cause of an abnormal battery as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method for determining the abnormality cause of an abnormal battery as claimed in any one of claims 1 to 7.

Images