Fault detection methods, detection devices and electronic equipment for lithium-ion batteries and lithium-ion battery packs

Abstract

A fault detection method includes: acquiring multiple sample data sets of a normal lithium-ion battery and multiple corresponding reference voltage data; training a deep learning algorithm using the sample data to obtain a first mapping relationship and a second mapping relationship; and calculating a voltage prediction error threshold based on the voltage prediction values ​​obtained from the first and second mapping relationships. Multiple test data sets and multiple corresponding real voltage data of the lithium-ion battery under test are collected; using the first and second mapping relationships, a first error, a second error, and a voltage prediction error are calculated; and a fault judgment result of the lithium-ion battery under test is determined through a preset logical relationship.

Description

[Technical Field]

[0001] This application relates to the field of lithium-ion battery technology, and in particular to a fault detection method, detection device and electronic equipment for lithium-ion batteries and lithium-ion battery packs. [Background Technology]

[0002] Lithium-ion batteries have achieved widespread application due to their high energy density and long cycle life. However, the frequent occurrence of safety incidents related to thermal runaway (TR) has raised concerns about the safety of lithium-ion batteries. The spread of localized hot spots triggered by early internal short circuit (ISC) faults throughout the battery is one of the main causes of TR.

[0003] Therefore, before ISC was developed into TR, effective testing of lithium-ion batteries was of great significance for their safe and reliable use. [Summary of the Invention]

[0004] The inventors of this application discovered during the development of lithium-ion battery fault detection that: due to the highly complex internal electrochemical reaction mechanism of lithium-ion batteries, it is difficult to accurately capture changes in battery parameters caused by ISC (Independent Fluctuation Characteristic). Existing threshold-based or model-based ISC diagnosis and detection methods have many defects and shortcomings, and cannot adequately meet the needs of practical applications.

[0005] To this end, the inventors of this application provide a fault detection method for lithium-ion battery packs. The fault detection method includes: acquiring multiple sample data sets of a normal lithium-ion battery and multiple corresponding reference voltage data sets; each sample data set includes: sample voltage data, sample current data, sample temperature data collected by the normal lithium-ion battery within a preset time period, and simulated voltage data simulated by a pre-constructed electrochemical mechanism model; one reference voltage data set is voltage data collected by the normal lithium-ion battery within a preset time period after the end time of the corresponding sample data set; training a preset deep learning algorithm using the sample voltage data, sample current data, and sample temperature data as input information, and the reference voltage data as output information, to obtain a first mapping relationship; training the deep learning algorithm using the simulated voltage data, sample current data, and sample temperature data as input information, and the reference voltage data as output information, to obtain a second mapping relationship; and training the deep learning algorithm based on the reference voltage data and the first... A first error threshold is calculated based on the sampled voltage prediction value output by a mapping relationship, and a second error threshold is calculated based on the simulated voltage prediction value output by the reference voltage data and the second mapping relationship. The first and second error thresholds are integrated into a voltage prediction error threshold using a preset numerical integration method. Multiple test data sets and multiple corresponding real voltage data sets of the lithium-ion battery under test are collected. Each test data set includes voltage data, current data, and temperature data collected by the lithium-ion battery under test within a preset time period. A first error is calculated based on the real voltage data and the first voltage prediction value output by the first mapping relationship, and a second error is calculated based on the real voltage data and the second voltage prediction value output by the second mapping relationship. The first and second errors are integrated into a voltage prediction error using the numerical integration method. Based on the first error, the second error, the voltage prediction error, and the voltage prediction error threshold, a fault judgment result for the lithium-ion battery under test is determined through a preset logical relationship.

[0006] The aforementioned fault detection method flexibly predicts voltage sequences of different time lengths by adjusting the length of the input data during training and learning the mapping relationship, and features high voltage prediction accuracy. Furthermore, based on higher accuracy and faster diagnosis of internal short-circuit faults in the battery, it achieves high diagnostic accuracy under various operating conditions, enabling early detection and warning.

[0007] In conjunction with the first aspect, in one possible implementation, acquiring multiple sample data sets and multiple corresponding reference voltage data of a normal lithium-ion battery specifically includes: collecting sample voltage data, sample current data, and sample temperature data of the normal lithium-ion battery as it changes over time under different operating conditions; obtaining simulated voltage data of the normal lithium-ion battery under the different operating conditions through a pre-constructed electrochemical mechanism model; dividing the sample voltage data, sample current data, sample temperature data, and simulated voltage data into multiple data segments at preset time steps, starting from the start time of the operating time; the time step is equal to the preset time period length; grouping the data segments of sample voltage data, sample current data, sample temperature data, and simulated voltage data within the same time period into a sample data set; and collecting voltage data of the lithium-ion battery within a preset time period after the end time of each sample data set, using this as the reference voltage data corresponding to the sample data set.

[0008] In conjunction with the first aspect, or any of the above possible implementations of the first aspect, in another possible implementation, the electrochemical mechanism model is an electrochemical mechanism model that includes thermal field simulation, including: single-particle models, quasi-two-dimensional models, and multi-dimensional multi-field electrochemical models, etc., which are models constructed from electrochemical theory.

[0009] In combination with the first aspect, or any of the above possible implementations of the first aspect, in yet another possible implementation, the deep learning algorithm includes: convolutional neural networks, fully connected neural networks, recurrent neural networks, long short-term memory neural networks, attention mechanisms, and Transformers.

[0010] In conjunction with the first aspect, or any of the possible implementations of the first aspect described above, in yet another possible implementation, the first error threshold is calculated using the following formula: Where N is the length of the preset time period, B is the correlation coefficient, C is the charging or discharging rate of the battery, E1 is the first error threshold, and U FP1 U is the predicted value of the sampled voltage. F The reference voltage data is used; the second error threshold is calculated using the following formula: Where N is the length of the preset time period, B is the correlation coefficient, C is the charging or discharging rate of the battery, E2 is the second error threshold, and U FP2 For the simulated voltage prediction value, U F The reference voltage data is used; the preset numerical integration method is weighted average.

[0011] In conjunction with the first aspect, or any of the possible implementations of the first aspect described above, in another possible implementation, the step of determining the fault judgment result of the lithium-ion battery under test based on the first error, the second error, the voltage prediction error, and the voltage prediction error threshold through a preset logical relationship specifically includes: determining that the lithium-ion battery under test has not failed when at least two of the first error, the second error, and the voltage prediction error are less than or equal to the voltage prediction error threshold; determining that the lithium-ion battery under test has failed when at least two of the first error, the second error, and the voltage prediction error are greater than the voltage prediction error threshold; and determining the fault level of the lithium-ion battery under test based on the values ​​by which the first error, the second error, and the voltage prediction error exceed the voltage prediction error threshold when the lithium-ion battery under test is determined to have failed.

[0012] Secondly, embodiments of this application provide a fault detection method for a lithium-ion battery pack. The lithium-ion battery pack includes multiple lithium-ion batteries, and the fault detection method includes: using the fault detection method for lithium-ion batteries described above to obtain a fault judgment result for each lithium-ion battery in the lithium-ion battery pack under test; analyzing the differences between different lithium-ion batteries in the lithium-ion battery pack based on the fault detection method; and determining the fault state of the lithium-ion battery pack under test based on the fault judgment result of each lithium-ion battery and the differences between different lithium-ion batteries.

[0013] In conjunction with the second aspect, in one possible implementation, the fault detection method based on the lithium-ion battery analyzes the differences between different lithium-ion batteries in the lithium-ion battery pack, specifically including: acquiring historical detection data of one target lithium-ion battery among multiple lithium-ion batteries; using the historical detection data to retrain the first and second mapping relationships in the fault detection method of the lithium-ion battery; calculating the voltage prediction error corresponding to each lithium-ion battery based on the retrained first and second mapping relationships; and determining the fault state of the lithium-ion battery pack under test based on the fault judgment result of each lithium-ion battery and the differences between different lithium-ion batteries, specifically including: in the... When the fault judgment result of the target lithium-ion battery is that a fault has occurred, and the difference between different lithium-ion batteries does not meet the standard, the fault state of the lithium-ion battery pack under test is determined to be a level two alarm state; when the fault judgment result of the target lithium-ion battery is that a fault has occurred, and the difference between different lithium-ion batteries meets the standard, the fault state of the lithium-ion battery pack under test is determined to be a level one alarm state; wherein, if the voltage prediction error corresponding to each of the other batteries is greater than a preset first value, and the voltage prediction error corresponding to the target lithium-ion battery calculated by retraining the first and second mapping relationships using historical detection data of other lithium-ion batteries is greater than a preset second value, the difference between different lithium-ion batteries is determined to not meet the standard.

[0014] Thirdly, this application provides a fault detection device for lithium-ion batteries. The fault detection device includes: a sample data acquisition module, used to acquire multiple sample data sets and multiple corresponding reference voltage data sets of a normal lithium-ion battery; each sample data set includes: sample voltage data, sample current data, sample temperature data collected by the normal lithium-ion battery within a preset time period, and simulated voltage data simulated by a pre-constructed electrochemical mechanism model; one of the reference voltage data sets is voltage data collected by the normal lithium-ion battery within a preset time period after the end time of the corresponding sample data set; a mapping relationship training module, used to train a preset deep learning algorithm with the sample voltage data, sample current data, and sample temperature data as input information and the reference voltage data as output information to obtain a first mapping relationship; and to train the deep learning algorithm with the simulated voltage data, sample current data, and sample temperature data as input information and the reference voltage data as output information to obtain a second mapping relationship; and a threshold calculation module, used to calculate the output sample voltage based on the reference voltage data and the first mapping relationship. The system comprises: a prediction value, a first error threshold, and a second error threshold; a voltage prediction error threshold, calculated based on the reference voltage data and the second mapping relationship; a test data acquisition module, used to acquire multiple test data sets and multiple corresponding real voltage data of the lithium-ion battery under test; each test data set includes voltage data, current data, and temperature data acquired by the lithium-ion battery under test within a preset time period; an error calculation module, used to calculate a first error based on the real voltage data and the first voltage prediction value calculated by the first mapping relationship, and a second error based on the real voltage data and the second voltage prediction value calculated by the second mapping relationship; the first error and the second error are integrated into a voltage prediction error using the numerical integration method; and a judgment module, used to determine the fault judgment result of the lithium-ion battery under test based on the first error, the second error, the voltage prediction error, and the voltage prediction error threshold, through a preset logical relationship.

[0015] Fourthly, this application provides an electronic device. The electronic device includes a processor and a memory, wherein the memory stores a non-volatile storage medium, and when the non-volatile storage medium is invoked by the processor, the processor executes the fault detection method described above.

[0016] The beneficial effects of the related devices provided in the second, third and fourth aspects of this application can be referred to the beneficial effects of the technical solution in the first aspect, and will not be repeated here. [Attached Image Description]

[0017] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0018] Figure 1 A flowchart illustrating a fault detection method for a lithium-ion battery provided in an embodiment of this application;

[0019] Figure 2 A flowchart illustrating a method for obtaining sample data and a reference voltage as provided in an embodiment of this application;

[0020] Figure 3 A flowchart illustrating the method for determining the fault judgment result of a lithium-ion battery according to an embodiment of this application;

[0021] Figure 4 This is a flowchart of a fault detection method for a lithium-ion battery pack provided in an embodiment of this application.

[0022] Figure 5 A flowchart illustrating a method for analyzing differences between different lithium-ion batteries in a lithium-ion battery pack, provided as an embodiment of this application.

[0023] Figure 6 A flowchart of a method for determining the fault judgment result of a lithium-ion battery pack provided in an embodiment of this application;

[0024] Figure 7 A flowchart illustrating a method for determining whether differences between different lithium-ion batteries conform to a standard, provided in an embodiment of this application.

[0025] Figure 8A A flowchart illustrating the mapping relationship training phase of the fault detection method provided in this application embodiment;

[0026] Figure 8B A flowchart illustrating the testing and usage phase of the fault detection method provided in this application embodiment;

[0027] Figure 9 A schematic diagram of the sample data and reference voltage data provided in the embodiments of this application;

[0028] Figure 10 A comparison chart of the predicted voltage value calculated using the first mapping relationship ALG1 and the actual voltage value obtained by the acquisition, provided for an embodiment of this application;

[0029] Figure 11 A comparison chart of the predicted voltage value calculated using the second mapping relationship ALG2 and the actual voltage value obtained by the acquisition, provided for an embodiment of this application;

[0030] Figure 12 This is a schematic diagram illustrating the logical judgment relationship for determining the fault state of a lithium-ion battery pack based on voltage prediction error, as provided in an embodiment of this application.

[0031] Figure 13 This is a functional block diagram of the fault detection device provided in the embodiments of this application;

[0032] Figure 14 This is a schematic diagram of the control terminal provided in an embodiment of this application.

Detailed Implementation Methods

[0033] To facilitate understanding of this application, a more detailed description is provided below with reference to the accompanying drawings and specific embodiments. It should be noted that when an element is described as being "fixed to" another element, it can be directly on the other element, or one or more intermediate elements may exist between them. When an element is described as being "connected" to another element, it can be directly connected to the other element, or one or more intermediate elements may exist between them. The terms "upper," "lower," "inner," "outer," "bottom," etc., used in this specification indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. Furthermore, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0034] Unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The term "and / or" as used in this specification includes any and all combinations of one or more of the associated listed items.

[0035] Furthermore, the technical features involved in the different embodiments of this application described below can be combined with each other as long as they do not conflict with each other.

[0036] Typical fault detection methods for lithium-ion batteries are mainly divided into threshold-based methods and model-based methods. Among them, threshold-based methods refer to the detection method that triggers and provides corresponding fault detection results when the measured parameters exceed or fall below the preset threshold by setting warning thresholds for key battery parameters (such as voltage, temperature, internal resistance, etc.).

[0037] Model-based methods refer to detection methods that establish mathematical models describing battery behavior, compare actual measured data with model predictions, and use parameter identification and state estimation to determine whether a lithium-ion battery has a fault.

[0038] In developing this application, the applicant discovered that threshold-based methods require a long time to diagnose faults in lithium-ion batteries. Furthermore, due to the dynamic characteristics of batteries, the judgment criteria change under different operating conditions and throughout their lifespan, making it difficult to set accurate thresholds. While model-based methods generally offer higher accuracy, their capabilities in areas such as diagnostic time, diagnostic accuracy, and maximum identification of internal short-circuit equivalent resistance remain limited, failing to meet practical testing needs.

[0039] The lithium-ion battery fault detection method provided in this application uses simulated voltage data, sampled real voltage data, current data, and temperature data as training data to train a selected deep learning model. This trains a first mapping relationship reflecting the current voltage, current, and temperature with the future voltage, and a second mapping relationship reflecting the simulated voltage with the real voltage. Based on the predicted voltage obtained from the first and second mapping relationships, the method diagnoses internal short-circuit faults in the battery with higher accuracy and faster speed, achieving high diagnostic accuracy under various operating conditions, thus enabling early detection and warning.

[0040] Moreover, when training and learning the first and second mapping relationships mentioned above, it is convenient to flexibly predict voltage sequences of different time lengths by adjusting the length of the input data, so that the predicted voltage has good accuracy.

[0041] Figure 1 This is a flowchart illustrating a fault detection method for lithium-ion batteries provided in an embodiment of this application. The fault detection method is broadly divided into a mapping relationship training phase and a test and use phase. When performing actual fault detection on a lithium-ion battery, after the fault detection method obtains the required first and second mapping relationships, the steps involved in the mapping relationship training phase are no longer repeated. Instead, the already stored first and second mapping relationships are directly invoked, and the steps of the test and use phase are executed to complete the fault detection.

[0042] In some embodiments, such as Figure 1 As shown, the mapping relationship training phase of this fault detection method includes the following steps:

[0043] S11. Obtain multiple sample data sets of normal lithium-ion batteries and multiple corresponding reference voltage data.

[0044] The sample data set refers to a combination of multiple data points used to record and describe the operating state of a lithium-ion battery within a specific time period. This data serves as training data and is further used in subsequent model training processes.

[0045] In this embodiment, each sample data set includes: sample voltage data, sample current data, sample temperature data of lithium-ion batteries collected within a preset time period, and simulated voltage data simulated by a pre-built electrochemical mechanism model.

[0046] Specifically, the sample voltage data represents the actual voltage variation of the lithium-ion battery over time. The sample current data represents the actual charging and discharging current of the lithium-ion battery. The sample temperature data represents the temperature variation of the lithium-ion battery over time. The simulated voltage data represents the theoretical value of the lithium-ion battery voltage calculated using this electrochemical mechanism model.

[0047] Reference voltage data refers to the continuous observation data collected after the completion of the sample data set. It serves as the voltage baseline data, a reference for the predicted values ​​output by the model. A reference voltage data set consists of voltage data collected from a normal lithium-ion battery within a preset time period after the end of the corresponding sample data set.

[0048] S12. Using sample voltage data, sample current data, and sample temperature data as input information, and using reference voltage data as output information, train the deep learning algorithm to obtain the first mapping relationship.

[0049] The "first mapping relationship" is a time-series prediction model based on historical data. By taking sample voltage, sample current, and sample temperature data as inputs and using reference voltage data as the expected output, it can learn the voltage evolution law of lithium-ion batteries under different operating conditions, thereby achieving accurate prediction of future voltage change trends.

[0050] S13. Using simulated voltage data, sample current data, and sample temperature data as input information, and reference voltage data as output information, train the deep learning algorithm to obtain the second mapping relationship.

[0051] The second mapping relationship is a model that reflects the conversion relationship between theoretical prediction and actual measurement. It is obtained by training the model with the prediction results (simulated voltage data) of the simulation theoretical model along with the operating condition information (current and temperature data) as input and the reference voltage data as the expected output.

[0052] Specifically, the preset deep learning algorithms include, but are not limited to: convolutional neural networks, fully connected networks, recurrent neural networks, long short-term memory neural networks, attention mechanisms, and Transformers.

[0053] In steps S12 and S13 above, the reliability and accuracy of lithium-ion battery state assessment are effectively improved through the mutual complementation and cooperation between the two different mapping relationships.

[0054] In some embodiments, different types of deep learning algorithms are selected for steps S12 and S13. In other embodiments, steps S12 and S13 use the same type of deep learning algorithm.

[0055] S14. Calculate the predicted value of the output sampled voltage based on the reference voltage data and the first mapping relationship, and calculate the first error threshold.

[0056] The first error threshold represents the deviation range between the voltage prediction value output by the first mapping relationship and the actual measured voltage value. It can serve as an indicator to evaluate the accuracy of predicting the actual operating state of the battery.

[0057] Specifically, the first error threshold is calculated using the following formula (1):

[0058]

[0059] Where N is the length of the preset time period, B is the correlation coefficient, C is the charging or discharging rate of the battery, E1 is the first error threshold, and U FP1 U is the predicted value of the sampled voltage. F This is for reference voltage data.

[0060] S15. Calculate the output simulated voltage prediction value based on the reference voltage data and the second mapping relationship, and calculate the second error threshold.

[0061] The second error threshold is the range of deviation between the voltage value predicted by the electrochemical mechanism model and the actual measured voltage value. It reflects the ability of the electrochemical mechanism model to explain and simulate the actual behavior of lithium-ion batteries.

[0062] Specifically, the second error threshold is calculated using the following formula (2):

[0063]

[0064] Where N is the length of the preset time period, B is the correlation coefficient, C is the charging or discharging rate of the battery, E2 is the second error threshold, and U FP2 For the simulated voltage prediction value, U F This is for reference voltage data.

[0065] S16. By using a preset numerical integration method, the first error threshold and the second error threshold are integrated into a voltage prediction error threshold.

[0066] Among them, "numerical integration method" refers to a data processing method that combines multiple independent numerical indicators into a comprehensive indicator through specific mathematical operations. Specifically, it employs methods such as simple arithmetic mean, weighted arithmetic mean, geometric mean, and harmonic mean.

[0067] Specifically, the preset numerical integration method is a weighted average. This weighted average method integrates the first and second error thresholds by assigning different weight coefficients to them, reflecting their importance and influence on the voltage prediction error threshold.

[0068] The above-described method, which integrates the first and second error thresholds into a unified voltage prediction error threshold, comprehensively considers prediction errors in two different dimensions, thus obtaining a more comprehensive evaluation standard. Furthermore, by rationally allocating weights, the advantages and limitations of the first and second error thresholds are balanced, improving the reliability and accuracy of the voltage prediction error threshold.

[0069] Please continue reading. Figure 1 The testing and application phase of this fault detection method includes the following steps:

[0070] S21. Collect multiple test data sets and multiple corresponding real voltage data of the lithium-ion battery under test.

[0071] The test data set is a combination of multiple data points used to record and describe the operating state of the lithium-ion battery under test within a specific time period. It serves as the initial data collected for subsequent fault diagnosis. The actual voltage data refers to the continued observation data collected after the test data set has been acquired.

[0072] In this embodiment, each test data set includes: voltage data, current data, and temperature data of the lithium-ion battery under test collected within a preset time period. The actual voltage data is collected in the same way as the reference voltage data, thereby ensuring that the test data set and the corresponding actual voltage data have the same data structure as the sample data set and the corresponding reference voltage data.

[0073] S22. Calculate the first error based on the real voltage data and the first voltage prediction value calculated from the first mapping relationship.

[0074] The first error reflects the accuracy of predicting future voltage changes based on historical data and is one of the indicators for judging whether the battery's operating state is abnormal.

[0075] Specifically, the first error is calculated in the same way as the first error threshold is calculated in step S14. In other words, the first error is also calculated using the above formula (1), by substituting the first voltage prediction value and the actual voltage data.

[0076] S23. Calculate the second error based on the actual voltage data and the second voltage prediction value calculated from the second mapping relationship.

[0077] The second error reflects the degree of matching between the simulation theoretical model and the actual behavior of lithium-ion batteries, indicating whether the lithium-ion battery deviates from its normal physical characteristics. It can also serve as an indicator of whether the battery's operating state is abnormal.

[0078] Specifically, the second error is calculated in the same way as the second error threshold is calculated in step S15. In other words, the second error is also calculated by substituting the second voltage prediction value and the actual voltage data into the above formula (2).

[0079] S24. By using a numerical integration method, the first error and the second error are integrated into a voltage prediction error.

[0080] Among them, voltage prediction error is a comprehensive index that integrates the first error based on time-series prediction and the second error based on simulation theoretical model prediction. It utilizes the complementary advantages of the two prediction methods to obtain more comprehensive error assessment results and provide a more reliable basis for fault judgment.

[0081] S25. Based on the first error, the second error, the voltage prediction error, and the voltage prediction error threshold, the fault judgment result of the lithium-ion battery under test is determined through a preset logical relationship.

[0082] The "logical relationship" includes several judgment rules at multiple levels. Based on the comparison results between the first error, the second error, and the voltage prediction error calculated in steps S22 to S24 above, and the voltage prediction error threshold determined in step S16, a series of multi-level judgment rules are used to comprehensively consider different indicators and form and determine the final fault judgment result of the lithium-ion battery.

[0083] The fault detection method provided in this application provides a high-precision predicted voltage value by flexibly predicting voltage sequences of different time lengths when training and learning the mapping relationship by adjusting the length of the input data.

[0084] In addition, based on the error of the predicted voltage value, corresponding logical judgments are made to diagnose short circuit faults in the battery under various different operating conditions, quickly obtain high-precision diagnostic results, realize early detection and warning, and well meet the needs of practical applications.

[0085] Figure 2 This is a flowchart illustrating a method for obtaining a sample data set and corresponding reference voltage data according to an embodiment of this application. In some embodiments, such as Figure 2 As shown, the method for obtaining multiple sample data sets and multiple corresponding reference voltage data includes the following steps:

[0086] S111. Collect sample voltage data, sample current data, and sample temperature data of normal lithium-ion batteries under different operating conditions as the operating time changes.

[0087] "Operating conditions" refers to the combination of test conditions and charge / discharge conditions during the testing of lithium-ion batteries. Specifically, these test conditions are not limited to specific temperatures, charge / discharge rates, ambient humidity, irradiation, etc., while charging methods include constant current charging, constant current-constant voltage charging, multi-stage constant current charging, pulse charging, etc.

[0088] A "normal lithium-ion battery" refers to a battery that has not experienced any malfunctions and is used as a benchmark for judging battery failure. The specific type of lithium-ion battery, positive electrode material, or negative electrode material is selected based on the actual needs of the situation.

[0089] "Running time" refers to the time interval from a specific start time t0 to the end time of data acquisition. For example, the start time t0 is the moment when a lithium-ion battery begins charging or discharging.

[0090] In addition, to ensure the richness and comprehensiveness of the training data, and to enable the trained model to better adapt to different usage scenarios, data was collected under multiple different operating conditions. In other words, the operating condition is also one of the parameters in the sample data.

[0091] S112. By using a pre-constructed electrochemical mechanism model, simulated voltage data of normal lithium-ion batteries under different operating conditions are obtained.

[0092] The "simulated voltage data" refers to the theoretical voltage values ​​calculated using an electrochemical mechanism model. These simulated voltage data under different operating conditions correspond to the sample voltage data collected in step S111. In other words, for each sample voltage data point, simulated voltage data under the same operating conditions and at the same time point is calculated using an electrochemical mechanism model.

[0093] Specifically, this electrochemical mechanism model is one that incorporates thermal field simulation. An "electrochemical mechanism model that incorporates thermal field simulation" refers to a theoretical model that can simultaneously consider electrochemical reaction processes and heat transfer processes. Based on the fundamental working principle of lithium-ion batteries, this model simulates various physicochemical processes within the battery by solving a series of partial differential equations describing electrochemical reactions, mass transport, and heat transfer.

[0094] In practical applications, the electrochemical mechanism model that includes thermal field simulation includes, but is not limited to: the Single Particle Model (SPM), which simplifies the electrode into a single representative particle, mainly considers the solid-phase diffusion process, has a smaller computational load, and is suitable for low-rate operation; the Pseudo Two-Dimensional Model (P2D), which combines the concentration and potential distribution in the electrode thickness direction to describe the transport process inside the battery; and the multidimensional multi-field electrochemical model that further introduces temperature field calculations based on the P2D model, which can simulate the coupled response of the battery under various complex operating conditions, as well as other suitable models formed by electrochemical theory.

[0095] S113. Starting from the beginning of the running time, according to the preset time step, the sample voltage data, sample current data, sample temperature data and simulated voltage data are divided into multiple data segments, and the data segments of sample voltage data, sample current data, sample temperature data and simulated voltage data in the same time period are combined into a sample data group.

[0096] Here, "time step" represents the length of the time interval for a data segment. The time step is set to an appropriate length according to needs and different considerations (e.g., battery dynamic response characteristics, data sampling frequency, and prediction accuracy requirements). For example, 1 second, 10 seconds, 100 seconds, 1000 seconds, or other suitable time lengths.

[0097] "Same time period" indicates the temporal correspondence between different data segments. For example, a data segment of sample voltage data records battery voltage data collected within a certain time interval. Data segments of sample current data, sample temperature data, and simulated voltage data within the same time period refer to current data, temperature data, and simulated voltage data calculated by an electrochemical mechanism model, respectively, collected within the same time interval.

[0098] Therefore, each sample data set contains complete information on the working state of lithium-ion batteries within a specific time interval, including both actual collected data and corresponding theoretical calculations, providing structured training data.

[0099] S114. Starting from the end time of each sample data group, collect the voltage data of the lithium-ion battery within a preset time period after the end time of the sample data group, and use it as the reference voltage data corresponding to the sample data group.

[0100] The “end time” is both the end point of the acquisition time for a certain sample data set and the starting point of the acquisition of reference voltage data, thus establishing a connection between the sample data set and its corresponding reference voltage data in the time series.

[0101] The reference voltage data consists of actual voltage values ​​collected within a preset time period after the end of the sample data set, serving as a benchmark for verifying the data and its accuracy. Specifically, the preset time period for collecting the reference voltage data is set to have equal durations as described above, facilitating subsequent comparative analysis.

[0102] In other embodiments, multiple sets of test data and multiple corresponding real voltage data of the lithium-ion battery under test are obtained in the following manner:

[0103] First, the voltage data, current data, and sample temperature of the lithium-ion battery under test are collected over a period of time.

[0104] Then, starting from the beginning of the running time, the voltage data, current data, and temperature data are divided into multiple data segments according to the preset time step, and the data segments of voltage data, current data, and temperature data in the same time period are combined into a test data group.

[0105] Finally, starting from the end time of each test data set, the voltage data of the lithium-ion battery under test is collected within a preset time period after the end time of the test data set, which is used as the actual voltage data corresponding to the test data set.

[0106] In other words, multiple test data sets and multiple corresponding real voltage data can be acquired by performing steps S111, S113 and S114 above in a similar manner, thereby forming the structured input information required for the first mapping relationship and the second mapping relationship.

[0107] Figure 3 This is a flowchart illustrating a method for determining the fault diagnosis result of a lithium-ion battery under test, provided in an embodiment of this application. In some embodiments, such as... Figure 3 As shown, the method for determining the fault diagnosis result includes the following steps:

[0108] S251. Determine the difference between the first error and the voltage prediction error threshold, the difference between the second error and the voltage prediction error threshold, and the difference between the voltage prediction error and the voltage prediction error threshold.

[0109] The difference quantifies the degree of deviation of various error indicators from the voltage prediction error threshold. The difference between the first error and the voltage prediction error threshold reflects the difference between the tested lithium-ion battery and a standard normal lithium-ion battery under the prediction of the first mapping relationship. The difference between the second error and the voltage prediction error threshold reflects the difference between the tested lithium-ion battery and a standard normal lithium-ion battery under the prediction of the second mapping relationship. The difference between the voltage prediction error and the voltage prediction error threshold is a comprehensive evaluation indicator that takes into account both the results of the first and second mapping relationships.

[0110] S252. Determine whether at least two of the first error, the second error, and the voltage prediction error are less than or equal to the voltage prediction error threshold. If yes, proceed to step S253; if no, proceed to step S254.

[0111] Specifically, if only one error indicator exceeds the voltage prediction error threshold, it may be due to inaccurate model prediction or temporary interference factors. However, if two or all error indicators exceed the voltage prediction error threshold, there is a greater certainty that the lithium-ion battery under test differs from a normal lithium-ion battery. This judgment mechanism effectively reduces the false positive rate and improves the reliability of fault diagnosis.

[0112] S253. Confirm that the lithium-ion battery under test is not faulty.

[0113] If there is no difference in battery behavior, such as electrochemical characteristics and performance, between the lithium-ion battery under test and a normal lithium-ion battery, the lithium-ion battery under test is considered to be normal and no possible internal malfunction has occurred.

[0114] S254. It is determined that the lithium-ion battery under test has malfunctioned.

[0115] If the battery behavior, such as the electrochemical characteristics and working performance, of the lithium-ion battery under test differs from that of a normal lithium-ion battery, it is considered that the lithium-ion battery under test has malfunctioned or is abnormal.

[0116] In other embodiments, after determining that the lithium-ion battery under test has failed, the fault level of the lithium-ion battery under test is further determined based on the values ​​of the first error, the second error, and the voltage prediction error exceeding the voltage prediction error threshold, thereby displaying and presenting the current fault state of the lithium-ion battery under test in more detail.

[0117] For example, when the absolute value of the difference is less than 50% of the voltage prediction error threshold, it indicates a minor fault, suggesting battery performance degradation. When the difference is between 50% and 100% of the voltage prediction error threshold, it indicates a moderate fault, suggesting a significant decrease in battery performance. When the difference is greater than 100% of the voltage prediction error threshold, it indicates a serious fault, suggesting a potential safety hazard in the battery.

[0118] Based on the fault detection method for lithium-ion batteries provided in one or more of the above embodiments, the embodiments of this application further apply it to the fault detection of lithium-ion battery packs that include multiple lithium-ion batteries, to determine the current state of the lithium-ion battery pack.

[0119] Figure 4 A fault detection method for a lithium-ion battery pack provided in embodiments of this application. In some embodiments, such as Figure 4 As shown, the fault detection method for this lithium-ion battery pack includes the following steps:

[0120] S31. Determine the fault diagnosis result of each lithium-ion battery in the lithium-ion battery pack under test.

[0121] After completing the training phase and determining the first mapping relationship, the second mapping relationship, and the voltage prediction error threshold, the fault judgment results of each lithium-ion battery are obtained by performing the above steps S21 to S23.

[0122] Specifically, the fault diagnosis result is a simple binary judgment: either the lithium-ion battery has malfunctioned, or the lithium-ion battery has not malfunctioned.

[0123] S32. Analyze the differences between different lithium-ion batteries in this lithium-ion battery pack.

[0124] The phrase "differences between different lithium-ion batteries" refers to the inconsistencies in electrochemical characteristics, operating states, and performance among individual lithium-ion batteries within the same lithium-ion battery pack. In this application, these differences are quantified and described using the lithium-ion battery fault detection methods described in one or more of the above embodiments.

[0125] S33. Based on the fault judgment results of each lithium-ion battery and the differences between different lithium-ion batteries, determine the fault status of the lithium-ion battery pack under test.

[0126] In assessing the fault status of lithium-ion battery packs, this approach integrates the fault assessment results of individual lithium-ion batteries with the differences between different lithium-ion batteries, forming an effective complementary and cross-validation mechanism to compensate for potential blind spots in single indicators. This allows for a more accurate differentiation of problems within the entire lithium-ion battery pack, providing comprehensive and reliable fault diagnosis results.

[0127] Figure 5 A flowchart illustrating a method for analyzing differences between different lithium-ion batteries in a lithium-ion battery pack, provided as an embodiment of this application. In some embodiments, such as Figure 5 As shown, this difference analysis method includes the following steps:

[0128] S321. Among N lithium-ion batteries, obtain the historical detection data of one of the target lithium-ion batteries.

[0129] The target lithium-ion battery is arbitrarily selected from N lithium-ion batteries in a lithium-ion battery pack. N is a positive integer representing the number of lithium-ion batteries in the lithium-ion battery pack. The value of N is determined by the actual lithium-ion battery pack being tested, for example, 3, 4, 5, or other positive integers.

[0130] S322. Using historical detection data of the target lithium-ion battery, retrain the first and second mapping relationships used in the fault detection method to obtain new first and second mapping relationships.

[0131] Retraining refers to transforming historical detection data into structured training data in the same form as in step S11 through appropriate data processing methods, and then re-executing steps S12 and S13 to update the model parameters in the first and second mapping relationships.

[0132] It is understandable that the new first and second mapping relationships obtained after such retraining represent the battery behavior of the target lithium-ion battery.

[0133] S323. Based on the new first and second mapping relationships, and through the steps of the test and use phase in the above-mentioned fault detection method, calculate the voltage prediction error corresponding to each lithium-ion battery.

[0134] The first error, second error, and voltage prediction error calculated for other lithium-ion batteries based on the new first and second mapping relationships actually represent the differences between them and the target lithium-ion battery. In other words, the first error, second error, and voltage prediction error are used to quantify or describe the differences between the two lithium-ion batteries.

[0135] S324. After selecting a new target lithium-ion battery from the remaining N-1 lithium-ion batteries, repeat steps S321 to S323 until the voltage prediction error of each lithium-ion battery relative to other different lithium-ion batteries is obtained.

[0136] Among them, the above steps S321 to S324 describe how each lithium-ion battery is used as the target lithium-ion battery in turn by going through the process one by one, and different first mapping relationships and second mapping relationships are trained to obtain the differences between each lithium-ion battery in the complete lithium-ion battery pack.

[0137] This approach avoids the bias that a single reference standard may bring, and can more accurately reflect the real differences between different lithium-ion batteries, as well as the uniqueness and degree of difference of a particular lithium-ion battery within a lithium-ion battery pack.

[0138] Figure 6 A flowchart illustrating a method for determining the fault state of a lithium-ion battery pack under test, provided in an embodiment of this application. In some embodiments, such as... Figure 6 As shown, determining the fault state of the lithium-ion battery pack under test includes the following steps:

[0139] S331. Among N lithium-ion batteries, select one as the lithium-ion battery to be judged.

[0140] The process of determining the fault state of the lithium-ion battery pack under test is carried out by testing each lithium-ion battery individually. N lithium-ion batteries are tested and judged sequentially. The lithium-ion battery currently selected for judgment is called the "lithium-ion battery to be judged".

[0141] S332. Determine the fault judgment result of the lithium-ion battery to be judged. If the fault judgment result is that a fault has occurred, proceed to step S333; if the fault judgment result is that no fault has occurred, proceed to step S336.

[0142] The fault determination result is a binary result, including both a faulty lithium-ion battery and a non-faulty lithium-ion battery. Specifically, the method described in steps S251 to S254 above is used to determine the specific fault determination result.

[0143] S333. Determine whether the difference between the lithium-ion battery to be judged and other lithium-ion batteries meets the standard. If yes, proceed to step S334; otherwise, proceed to step S335.

[0144] Here, "standard" refers to a pre-defined set of logical judgment rules used to determine and measure differences. The standard consists of one or more judgment levels and integrates various different indicators.

[0145] S334. Determine the fault status of the lithium-ion battery pack under test as a Level 1 alarm status.

[0146] S335. The fault status of the lithium-ion battery pack under test is determined to be a level 2 alarm status.

[0147] The "Level 1 Alarm Status" and "Level 2 Alarm Status" are classifications of the fault status of lithium-ion battery packs, indicating the severity of the current fault. In other words, if a single lithium-ion battery is determined to be faulty, and this battery also exhibits significant differences from other lithium-ion battery packs, it indicates a higher risk, and the alarm status of the lithium-ion battery pack is set to the higher Level 2 alarm status.

[0148] If the fault diagnosis result of a single lithium-ion battery is no fault, but the lithium-ion battery is significantly different from other lithium-ion battery packs, it indicates that there is an imbalance and difference within the lithium-ion battery pack. In this case, the alarm status of the lithium-ion battery pack under test is set to a relatively low level one alarm status.

[0149] S336. Among the remaining N-1 lithium-ion batteries, select one again as the lithium-ion battery to be judged, and execute steps S332 to S335.

[0150] S337. When the fault judgment results of N lithium-ion batteries are all no fault, the lithium-ion battery pack under test is determined to be fault-free.

[0151] If all lithium-ion batteries have been traversed and it is determined that the differences between these lithium-ion batteries are not significant and that they are all operating normally without any faults, then the lithium-ion battery pack is considered to be in a fault-free or normal state and no alarm needs to be issued.

[0152] Figure 7 A flowchart illustrating a method for determining whether the differences between a lithium-ion battery to be judged and other lithium-ion batteries conform to a standard, provided for embodiments of this application. In some embodiments, such as Figure 7 As shown, the method for determining whether a difference meets the standard includes the following steps:

[0153] S3331. Using the historical test data of the lithium-ion battery to be judged, after retraining the first and second mapping relationships, calculate the voltage prediction error between the lithium-ion battery to be judged and other lithium-ion batteries.

[0154] S3332. When the voltage prediction error of the lithium-ion battery to be judged is less than a preset first value, and the voltage prediction error of each other lithium-ion battery is greater than a preset second value, step S3333 is executed.

[0155] The first and second values ​​are both preset values, set by technicians according to actual needs, and presented in any suitable form, including but not limited to percentages or absolute values.

[0156] The first value is used to measure whether the predicted voltage error is small, while the second value is used to measure whether the predicted voltage error is large. In other words, when the target lithium-ion battery is the lithium-ion battery to be judged, the voltage prediction error calculated based on this standard for the target lithium-ion battery is small, while the voltage prediction error calculated based on this standard for other lithium-ion batteries is large. Specifically, the first value is smaller than the second value.

[0157] S3333: Re-train the first and second mapping relationships using historical test data of other different lithium-ion batteries, and recalculate the voltage prediction error of the lithium-ion battery to be judged and other lithium-ion batteries.

[0158] S3334. When the voltage prediction error of the lithium-ion battery to be judged is greater than the second value, and the voltage prediction error of other lithium-ion batteries is less than the first value, it is determined that the difference between different lithium-ion batteries does not meet the standard.

[0159] In particular, steps S3332 and S3334 establish judgment conditions in two directions, thus forming a more rigorous standard for measuring differences. Only when the judgment conditions of steps S3332 and S3334 are met simultaneously is it confirmed that the differences between different lithium-ion batteries do not meet the standard, that is, it is considered with a high degree of confidence that there are significant differences between different lithium-ion batteries. Otherwise, it is considered that the differences between different lithium-ion batteries meet the standard, and there are no significant differences between them.

[0160] Such two-way verification can corroborate each other, reduce the risk of misjudgment, and ensure the stability and credibility of the difference judgment results.

[0161] The following provides a specific example that fully explains and describes the above-mentioned fault detection method. This specific example demonstrates in detail the complete fault detection process, from acquiring sample data sets, training to obtain the first and second mapping relationships, determining the fault judgment result of a single lithium-ion battery, to comprehensively determining the fault state of the lithium-ion battery pack based on the fault judgment results of a single lithium-ion battery and the differences between different lithium-ion batteries.

[0162] Figure 8A This is a flowchart illustrating the mapping relationship training phase of the fault detection method provided in this application embodiment. Figure 8B This is a flowchart illustrating the testing and usage phase of the fault detection method provided in this application embodiment. Figure 8AAs shown, the steps in the training phase of this mapping relationship include:

[0163] 1) Training yields the first and second mapping relationships:

[0164] S101. Perform charge-discharge cycles on normal lithium-ion batteries under different temperatures, humidity, charging schemes, and discharging schemes to obtain voltage (U), current (I), and temperature (T) under different aging states and charge-discharge conditions.

[0165] The charging schemes include, but are not limited to, constant current charging, constant current and constant voltage charging, multi-stage constant current charging, and pulse charging. The standard lithium-ion battery is a suitable type of lithium-ion battery selected based on the specific needs of the application scenario.

[0166] S102. Construct a quasi-two-dimensional model incorporating thermal field simulation, and obtain the simulated battery voltage (U0) under the same charge-discharge conditions as in step S101 using this quasi-two-dimensional model. S ).

[0167] S103. Starting from the start of operation, divide the U, I, and T data during the charging and discharging process of the lithium-ion battery into a data segment every 30 seconds.

[0168] Among them, such as Figure 9 As shown, data segments of U, I, and T within the same time period are grouped into a data set and labeled as X1.

[0169] S104. Starting from the end time of each X1, continue to collect the voltage data (U) of the lithium-ion battery for the next 30 seconds. F ).

[0170] Please continue reading for more details. Figure 9 Voltage data U collected 30 seconds after each X1 F They are combined into a single data set and labeled Y.

[0171] S105, The U output from the electrochemical mechanism model including thermal field simulation. S The data is divided into corresponding segments according to the same time step as each X1.

[0172] Please continue reading for more details. Figure 9 U for each data segment S It is denoted as X2.

[0173] S106. Based on the above data sets X1 and Y, using U, I, and T as inputs, U F As output, the long short-term memory neural network algorithm with self-attention mechanism is trained to obtain the first mapping relationship ALG1.

[0174] The first mapping relationship ALG1 reflects the mapping relationship between the current voltage, current, and temperature and the voltage at a future time. Therefore, based on the first mapping relationship ALG1, the predicted future sampled voltage value UFP1 of the lithium-ion battery is calculated.

[0175] S107. Based on the above data sets X2 and X1, simulate the voltage U using a quasi-two-dimensional model that includes thermal field simulation. S I and T are used as inputs, and U is used as the input. F As output, a dual algorithm consisting of a long short-term memory neural network with self-attention mechanism and a fully connected neural network is trained to obtain the second mapping relationship ALG2.

[0176] The second mapping relationship, ALG2, reflects the mapping relationship between the simulated voltage and the actual voltage at future times. Therefore, based on this second mapping relationship ALG2, the future voltage UFP2 simulated by the electrochemical mechanism model is calculated.

[0177] For example, Figure 10 A comparison chart showing the predicted voltage value calculated using the first mapping relationship ALG1 and the actual voltage value obtained by acquisition is presented. Figure 11 A comparison chart is shown between the predicted voltage value calculated using the second mapping relationship ALG2 and the actual voltage value obtained from data acquisition. For example... Figure 10 and Figure 11 As shown, the error between the predicted voltage value and the actual voltage value of the two mapping relationships is very small, and the accuracy is very high.

[0178] 2) Determine the voltage prediction error threshold:

[0179] S201. Calculate the first error threshold using the following formula (1):

[0180]

[0181] Where N is the length of the data segment, B is the correlation coefficient (default setting is 1), and C is the charging or discharging rate of the battery.

[0182] S202. Calculate the second error threshold using the following formula (2):

[0183]

[0184] Where N is the length of the data segment, B is the correlation coefficient (default setting is 1), and C is the charging or discharging rate of the battery.

[0185] S203. The first error threshold E1 and the second error threshold E2 are weighted and averaged to obtain the voltage prediction error threshold E3.

[0186] like Figure 8B As shown, the testing phase includes the following steps:

[0187] 3) Testing lithium-ion batteries:

[0188] S301. Perform charge-discharge cycles on the lithium-ion battery under test at different temperatures, humidity levels, charging schemes, and discharging schemes to obtain the voltage (U) under different aging states and charge-discharge conditions. GZ ), current (I) GZ ) and temperature (T) GZ )data.

[0189] The charging scheme includes constant current charging, constant current and constant voltage charging, multi-stage constant current charging, and pulse charging. The lithium-ion battery under test can be selected from those exhibiting different faults such as lithium plating, thermal runaway, internal short circuit, and external short circuit, depending on the application scenario, to verify and test the reliability and accuracy of the fault detection method.

[0190] S302. Using the same method as steps S101 to S105, obtain fault test data that has the same data structure as the training data.

[0191] S303. Based on the acquired fault test data, the first voltage prediction value U is calculated using the first mapping relationship ALG1 and the second mapping relationship ALG2, respectively. fp3 Second voltage prediction value U fp4 .

[0192] S304, respectively, the first voltage prediction value U fp3 Second voltage prediction value U fp4 Substituting into the above formulas (1) and (2), the corresponding first error E is calculated. GZ1 Second error E GZ2 .

[0193] S305, Regarding the first error E GZ1 Second error E GZ2 The voltage prediction error E is obtained by performing a weighted average. GZ3 .

[0194] S306, in the first error E GZ1 Second error E GZ2 and voltage prediction error E GZ3 If any two values ​​exceed the voltage prediction error threshold E3, the lithium-ion battery being tested is determined to be faulty.

[0195] If only one value is greater than the voltage prediction error threshold E3, or if all three values ​​are not greater than the voltage prediction error threshold E3, then the battery behavior and characteristics of the lithium-ion battery being tested are consistent with those of a normal lithium-ion battery, and no fault has occurred.

[0196] S307. After determining that the lithium-ion battery being tested has failed, the fault is classified according to the magnitude of the value exceeding the voltage prediction error threshold E3.

[0197] Specifically, the fault classification results of the currently tested lithium-ion batteries are output by the corresponding battery management system or energy management system, and presented and displayed to the user.

[0198] 4) Testing lithium-ion battery packs:

[0199] In this embodiment, a lithium-ion battery pack consisting of three lithium-ion batteries connected in series is used as an example. Please continue reading. Figure 8B The testing process for lithium-ion battery packs includes:

[0200] S401. Using the steps S301 to S307 above, calculate the first error, the second error, and the voltage prediction error of the three lithium-ion batteries respectively.

[0201] Among them, such as Figure 12 As shown, the first error, second error, and voltage prediction error of the first lithium-ion battery are denoted as E. GZB11 E GZB12 and E GZB13 The first error, second error, and voltage prediction error of the second lithium-ion battery are denoted as E. GZB21 E GZB22 and E GZB23 The first error, second error, and voltage prediction error of the third lithium-ion battery are denoted as E. GZB31 E GZB32 and E GZB33 .

[0202] S402. The first mapping relationship ALG1 and the second mapping relationship ALG2 are retrained sequentially using the historical operating data of the first lithium-ion battery, the second lithium-ion battery and the third lithium-ion battery to obtain three new first mapping relationships ALG1 and second mapping relationships ALG2.

[0203] S403. Using the new first mapping relationship ALG1 and the first mapping relationship ALG2 in sequence, the voltage prediction error of the three lithium-ion batteries is calculated through the same steps S03 to S305.

[0204] Please continue reading for more details. Figure 12 Based on the historical operating data of the first lithium-ion battery, the first mapping relationship ALG1 and the second mapping relationship ALG2 were retrained. The voltage prediction errors of the first, second, and third lithium-ion batteries were calculated and labeled as E. B11 E B12 E B13 .

[0205] Based on the historical operating data of the second lithium-ion battery, the first mapping relationship ALG1 and the second mapping relationship ALG2 were retrained. The voltage prediction errors of the first, second, and third lithium-ion batteries were calculated and labeled as E. B21 E B22 E B23 .

[0206] Based on the historical operating data of the third lithium-ion battery, the first mapping relationship ALG1 and the second mapping relationship ALG2 were retrained. The voltage prediction errors of the first, second, and third lithium-ion batteries were calculated and labeled as E. B31 E B32 E B33 .

[0207] S404. Based on the error values ​​calculated in steps S401 to S403 above, determine the state of the lithium-ion battery pack under test and locate the faulty lithium-ion battery.

[0208] Please continue reading for more details. Figure 12 The status of the lithium-ion battery pack includes: the first lithium-ion battery is not faulty, the second lithium-ion battery is not faulty, the third lithium-ion battery is not faulty, a level one alarm status, and a level two alarm status. The logical relationship between the status of the lithium-ion battery pack, the location result of the faulty lithium-ion battery, and the above-mentioned multiple error values ​​is as follows: Figure 12 As shown.

[0209] Specifically, for one of the lithium-ion batteries, if at least two of the first error, the second error, and the voltage prediction error are greater than the voltage prediction error threshold E3, the lithium-ion battery is determined to be faulty. Otherwise, the lithium-ion battery is determined not to be faulty.

[0210] Furthermore, when a fault is determined in a particular lithium-ion battery, the issuance of a secondary alarm state is determined based on the following two conditions: First, if, when using this lithium-ion battery as a reference, the voltage prediction errors of the other two lithium-ion batteries are relatively large (e.g., greater than the aforementioned second value), while the voltage prediction error of this lithium-ion battery is relatively small (e.g., less than the aforementioned first value). Second, when using the other two lithium-ion batteries as a reference, while the voltage prediction error of this lithium-ion battery is relatively large (e.g., greater than the aforementioned second value), the voltage prediction errors of the other lithium-ion batteries are similarly distributed, all being relatively small errors (e.g., less than the aforementioned first value).

[0211] A level-two alarm is issued when both of the aforementioned conditions are met. Conversely, a level-one alarm is issued if either of the aforementioned conditions is not met.

[0212] Based on the fault detection method for lithium-ion batteries provided in the above embodiments, this application further provides a fault detection device for lithium-ion batteries. Figure 13 This is a functional block diagram of the fault detection device provided in the embodiments of this application.

[0213] In some embodiments, such as Figure 13 As shown, the fault detection device 10 includes: a sample data acquisition module 11, a mapping relationship training module 12, a threshold calculation module 13, a test data acquisition module 14, an error calculation module 15, and a judgment module 16.

[0214] The sample data acquisition module 11 is used to acquire multiple sample data sets and multiple corresponding reference voltage data sets of a normal lithium-ion battery. Each sample data set includes: sample voltage data, sample current data, sample temperature data, and simulated voltage data simulated by a pre-constructed electrochemical mechanism model collected by the normal lithium-ion battery within a preset time period. A reference voltage data set is the voltage data collected by the normal lithium-ion battery within a preset time period after the end time of the corresponding sample data set. The mapping relationship training module 12 is used to train a preset deep learning algorithm using the sample voltage data, sample current data, and sample temperature data as input information and the reference voltage data as output information to obtain a first mapping relationship; and to train the deep learning algorithm using the simulated voltage data, sample current data, and sample temperature data as input information and the reference voltage data as output information to obtain a second mapping relationship. The threshold calculation module 13 is used to calculate the output sampled voltage prediction value based on the reference voltage data and the first mapping relationship, calculate a first error threshold, and calculate the output simulated voltage prediction value based on the reference voltage data and the second mapping relationship, calculate a second error threshold; and integrate the first error threshold and the second error threshold into a voltage prediction error threshold through a preset numerical integration method. The test data acquisition module 14 is used to acquire multiple test data sets and multiple corresponding real voltage data of the lithium-ion battery under test; each test data set includes: voltage data, current data, and temperature data acquired by the lithium-ion battery under test within a preset time period. The error calculation module 15 is used to calculate a first error based on the real voltage data and the first voltage prediction value calculated from the first mapping relationship, and calculate a second error based on the real voltage data and the second voltage prediction value calculated from the second mapping relationship; and integrate the first error and the second error into a voltage prediction error through the numerical integration method. The judgment module 16 is used to determine the fault judgment result of the lithium-ion battery under test based on the first error, the second error, the voltage prediction error, and the voltage prediction error threshold through a preset logical relationship.

[0215] This fault detection device flexibly predicts voltage sequences of different time lengths by adjusting the length of the input data, and features high voltage prediction accuracy. Furthermore, based on the predicted voltage, it diagnoses internal short-circuit faults in the battery with higher accuracy and faster speed, achieving high diagnostic accuracy under various operating conditions, thus enabling early detection and warning.

[0216] In some embodiments, the sample data acquisition module 11 is specifically used for: collecting sample voltage data, sample current data, and sample temperature data of the normal lithium-ion battery as it changes over time under different operating conditions; obtaining simulated voltage data of the normal lithium-ion battery under different operating conditions through a pre-constructed electrochemical mechanism model; dividing the sample voltage data, sample current data, sample temperature data, and simulated voltage data into multiple data segments at intervals according to a preset time step, with the start time of the operating time as the starting point; the time step being equal to the length of the preset time period; forming a sample data group by grouping the data segments of sample voltage data, sample current data, sample temperature data, and simulated voltage data of the same time period; and collecting voltage data of the lithium-ion battery within a preset time period after the end time of each sample data group as the starting point, using it as reference voltage data corresponding to the sample data group.

[0217] Specifically, the electrochemical mechanism model is an electrochemical mechanism model that incorporates thermal field simulation, including: single-particle model, quasi-two-dimensional model, and multi-dimensional multi-field electrochemical model. The deep learning algorithm includes: convolutional neural network, fully connected network, recurrent neural network, long short-term memory neural network, attention mechanism, and Transformer.

[0218] In some embodiments, the threshold calculation module 13 is specifically used to: calculate a first error threshold by formula (1), calculate a second error threshold by formula (2), and integrate the first error threshold and the second error threshold into a voltage prediction error threshold by weighted averaging.

[0219] In some embodiments, the determination module 16 is specifically configured to: determine that the lithium-ion battery under test has not failed when at least two of the first error, the second error, and the voltage prediction error are less than or equal to the voltage prediction error threshold; determine that the lithium-ion battery under test has failed when at least two of the first error, the second error, and the voltage prediction error are greater than the voltage prediction error threshold; and determine the fault level of the lithium-ion battery under test based on the values ​​of the first error, the second error, and the voltage prediction error exceeding the voltage prediction error threshold when the lithium-ion battery under test is determined to have failed.

[0220] It should be noted that the fault detection device provided in this application embodiment can be executed by any suitable type of electronic computing platform. This electronic computing platform can implement or execute one or more of the above functions using electronic hardware, computer software programs, or a combination of both. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0221] Figure 14 This is a schematic diagram of a control terminal provided in an embodiment of this application. The control terminal can be used to execute one or more steps of the fault detection method described in one or more of the above embodiments. The control terminal is implemented using various forms of digital computers, such as workstations, servers, blade servers, mainframe computers, and other suitable electronic computing platforms.

[0222] It should be noted that, Figure 14 The components, their connections and relationships, and their functions shown are for illustrative purposes only and are not intended to impose limitations on the specific implementation of the control terminal.

[0223] like Figure 14 As shown, the control terminal 20 includes a processor 21, a memory 22, a storage device 23, a high-speed interface 25 connected to the memory 22 and multiple high-speed expansion ports 24, and a low-speed interface 27 connected to a low-speed expansion port 26 and the storage device 23.

[0224] Each of the processor 21, memory 22, storage device 23, high-speed interface 25, high-speed expansion port 24, and low-speed interface 27 is interconnected using various buses and mounted on a common motherboard or other suitable means.

[0225] Processor 21 processes computer program instructions stored in memory 22 or on storage device 23 for displaying graphical information on an external input / output device (e.g., display device 28 coupled to high-speed interface 25).

[0226] In some embodiments, multiple processors and / or multiple buses are used in conjunction with multiple memories and multiple types of memory. Furthermore, multiple electronic devices are connected, each providing a portion of the necessary operation (e.g., as a server group, blade server cluster, or multiprocessor system).

[0227] The memory 22 stores information within the control terminal. It may be one or more volatile memory cells, non-volatile memory cells, or another form of computer-readable media, such as a magnetic disk or optical disk.

[0228] Storage device 23 is capable of providing large-capacity storage for the control terminal. It contains computer-readable media, such as floppy disk devices, hard disk devices, optical disk devices or magnetic tape devices, flash memory or other similar solid-state storage devices, or device arrays, including devices or other configurations in a storage area network.

[0229] The computer program instructions are stored in an information carrier. When executed by one or more processing devices (e.g., processor 21), the computer program instructions implement the fault detection method described in one or more of the above embodiments.

[0230] High-speed interface 25 manages bandwidth-intensive operations for controlling the terminal, while low-speed interface 27 manages less bandwidth-intensive operations. In some embodiments, high-speed interface 25 is coupled to memory 22, display 28 (e.g., via a graphics processor or accelerator), and high-speed expansion port 24 that accepts various expansion cards. Low-speed interface 27 is coupled to storage device 23 and low-speed expansion port 26.

[0231] The low-speed expansion port 26 includes a communication port (e.g., USB, Ethernet, wireless Ethernet) and can be coupled to one or more input / output devices, such as a keyboard, pointing device, scanner, or a networked device such as a switch (e.g., via a network adapter).

[0232] For example, such as Figure 14 As shown, in order to provide interaction with the user, the control terminal 20 also has a display device 28 (e.g., a cathode ray tube or liquid crystal monitor) for displaying information to the user and a pointing device 29 (e.g., a mouse) for the user to provide input to the computer. Other types of interactive devices can also be used to provide interaction with the user;

[0233] Of course, the feedback provided to the user is any form of sensory feedback (such as visual feedback, auditory feedback, or tactile feedback); and input from the user is received in any form, including acoustic, voice, or tactile input.

[0234] The fault detection method described in one or more embodiments of this application is implemented in digital electronic circuits, integrated circuits, specially designed ASICs (Application-Specific Integrated Circuits), computer hardware, firmware, software, and / or combinations thereof. It includes embodiments of one or more computer programs that execute and / or interpret on a programmable system comprising at least one programmable processor, which may be dedicated or general-purpose, coupled to receive data and instructions from a storage system, at least one input device, and at least one output device, and to transmit data and instructions to the storage system, at least one input device, and at least one output device.

[0235] These computer programs (also referred to as programs, software, software applications, or code) include machine instructions for a programmable processor and are implemented using high-level programming and / or goal-oriented programming languages ​​and / or assembly / machine languages. In this embodiment, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, device, and / or apparatus (e.g., disk, optical disk, memory, programmable logic device (PLD)) used to provide machine instructions and / or data to a programmable processor, including machine-readable media that receive machine instructions as machine-readable signals. The terms "machine-readable signal" refer to any signal used to provide machine instructions and / or data to a programmable processor.

[0236] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and not to limit them; under the concept of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of this application as described above, which are not provided in detail for the sake of brevity; although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A fault detection method for lithium-ion batteries, characterized in that, include: Acquire multiple sample data sets of normal lithium-ion batteries and multiple corresponding reference voltage data; Each sample data set includes: sample voltage data, sample current data, sample temperature data, and simulated voltage data of the normal lithium-ion battery collected within a preset time period, and a reference voltage data is the voltage data of the normal lithium-ion battery collected within a preset time period after the end time of the corresponding sample data set. Using the sample voltage data, the sample current data, and the sample temperature data as input information, and the reference voltage data as output information, the deep learning algorithm is trained to obtain the first mapping relationship; Using the simulated voltage data, the sample current data, and the sample temperature data as input information, and the reference voltage data as output information, the deep learning algorithm is trained to obtain the second mapping relationship; Based on the reference voltage data and the first mapping relationship, the output sampled voltage prediction value is calculated, the first error threshold is calculated, and based on the reference voltage data and the second mapping relationship, the output simulated voltage prediction value is calculated, and the second error threshold is calculated. The first error threshold and the second error threshold are integrated into a voltage prediction error threshold using a preset numerical integration method. Multiple test data sets and multiple corresponding real voltage data are collected from the lithium-ion battery under test; each test data set includes: voltage data, current data and temperature data collected from the lithium-ion battery under test within a preset time period; Based on the actual voltage data and the first voltage prediction value calculated from the first mapping relationship, a first error is calculated, and based on the actual voltage data and the second voltage prediction value calculated from the second mapping relationship, a second error is calculated. The first error and the second error are integrated into a voltage prediction error using the numerical integration method described above. Based on the first error, the second error, the voltage prediction error, and the voltage prediction error threshold, the fault judgment result of the lithium-ion battery under test is determined through a preset logical relationship.

2. The fault detection method according to claim 1, characterized in that, The acquisition of multiple sample data sets and multiple corresponding reference voltage data of normal lithium-ion batteries specifically includes: Sample voltage data, sample current data, and sample temperature data of the normal lithium-ion battery under different operating conditions were collected as a function of operating time. The simulated voltage data of the normal lithium-ion battery under different operating conditions were obtained by using a pre-constructed electrochemical mechanism model. Starting from the beginning of the running time, the sample voltage data, sample current data, sample temperature data, and simulated voltage data are divided into multiple data segments according to a preset time step; the time step is equal to the length of the preset time period. The sample data group consists of data segments of sample voltage data, sample current data, sample temperature data, and simulated voltage data from the same time period. Starting from the end time of each sample data group, voltage data of the lithium-ion battery within a preset time period after the end time of the sample data group is collected, which serves as reference voltage data corresponding to the sample data group.

3. The fault detection method according to claim 1, characterized in that, The electrochemical mechanism model is an electrochemical mechanism model that includes thermal field simulation, including: single-particle model, quasi-two-dimensional model and multi-dimensional multi-field electrochemical model.

4. The fault detection method according to claim 1, characterized in that, The deep learning algorithms include: convolutional neural networks, fully connected networks, recurrent neural networks, long short-term memory neural networks, attention mechanisms, and transformers.

5. The fault detection method according to claim 1, characterized in that, The first error threshold is calculated using the following formula: Where N is the length of the preset time period, B is the correlation coefficient, C is the charging or discharging rate of the battery, E1 is the first error threshold, and U FP1 U is the predicted value of the sampled voltage. F Reference voltage data; The second error threshold is calculated using the following formula: Where N is the length of the preset time period, B is the correlation coefficient, C is the charging or discharging rate of the battery, E2 is the second error threshold, and U FP2 For the simulated voltage prediction value, U F Reference voltage data; The preset numerical integration method is: weighted average.

6. The fault detection method according to claim 1, characterized in that, The step of determining the fault judgment result of the lithium-ion battery under test based on the first error, the second error, the voltage prediction error, and the voltage prediction error threshold, through a preset logical relationship, specifically includes: When at least two of the first error, the second error, and the voltage prediction error are less than or equal to the voltage prediction error threshold, it is determined that the lithium-ion battery under test has not malfunctioned. If at least two of the first error, the second error, and the voltage prediction error are greater than the voltage prediction error threshold, the lithium-ion battery under test is determined to be faulty. When it is determined that the lithium-ion battery under test has failed, the fault level of the lithium-ion battery under test is determined based on the first error, the second error, and the voltage prediction error exceeding the voltage prediction error threshold.

7. A fault detection method for a lithium-ion battery pack, wherein the lithium-ion battery pack comprises multiple lithium-ion batteries, characterized in that, The method includes: Using the fault detection method for lithium-ion batteries as described in any one of claims 1-6, the fault judgment result of each lithium-ion battery in the lithium-ion battery pack under test is obtained respectively. Based on the fault detection method for lithium-ion batteries, the differences between different lithium-ion batteries in the lithium-ion battery pack are analyzed. The fault status of the lithium-ion battery pack under test is determined based on the fault judgment results of each lithium-ion battery and the differences between different lithium-ion batteries.

8. The fault detection method according to claim 7, characterized in that, The fault detection method based on the lithium-ion battery analyzes the differences between different lithium-ion batteries in the lithium-ion battery pack, specifically including: Among the multiple lithium-ion batteries, historical detection data of one target lithium-ion battery is obtained; Using the historical detection data, the first and second mapping relationships in the fault detection method for lithium-ion batteries are retrained; Based on the retrained first and second mapping relationships, the voltage prediction error for each lithium-ion battery is calculated respectively. The step of determining the fault state of the lithium-ion battery pack under test based on the fault judgment results of each lithium-ion battery and the differences between different lithium-ion batteries specifically includes: When the fault judgment result of the target lithium-ion battery is that a fault has occurred, and the differences between different lithium-ion batteries do not meet the standard, the fault status of the lithium-ion battery pack under test is determined to be a level two alarm status. When the fault judgment result of the target lithium-ion battery is that a fault has occurred, and the differences between different lithium-ion batteries meet the standard, the fault status of the lithium-ion battery pack under test is determined to be a level one alarm status. Specifically, if the voltage prediction error of the target lithium-ion battery obtained by retraining the first and second mapping relationships using historical test data of other lithium-ion batteries is greater than the preset second value, and the voltage prediction error of the remaining lithium-ion batteries obtained by retraining the first and second mapping relationships using historical test data of the target lithium-ion battery is greater than the preset second value, then the difference between different lithium-ion batteries is determined to be inconsistent with the standard.

9. A fault detection device for lithium-ion batteries, characterized in that, include: The sample data acquisition module is used to acquire multiple sample data sets and multiple corresponding reference voltage data of normal lithium-ion batteries; Each sample data set includes: sample voltage data, sample current data, sample temperature data, and simulated voltage data of the normal lithium-ion battery collected within a preset time period, and a reference voltage data is the voltage data of the normal lithium-ion battery collected within a preset time period after the end time of the corresponding sample data set. The mapping relationship training module is used to train a preset deep learning algorithm with the sample voltage data, the sample current data, and the sample temperature data as input information and the reference voltage data as output information to obtain a first mapping relationship; and to train the deep learning algorithm with the simulated voltage data, the sample current data, and the sample temperature data as input information and the reference voltage data as output information to obtain a second mapping relationship. The threshold calculation module is used to calculate the output sampled voltage prediction value based on the reference voltage data and the first mapping relationship, calculate the first error threshold, and calculate the output simulated voltage prediction value based on the reference voltage data and the second mapping relationship, calculate the second error threshold; and integrate the first error threshold and the second error threshold into a voltage prediction error threshold through a preset numerical integration method. The test data acquisition module is used to acquire multiple test data sets and multiple corresponding real voltage data of the lithium-ion battery under test; each test data set includes: voltage data, current data and temperature data of the lithium-ion battery under test within a preset time period. An error calculation module is used to calculate a first error based on the actual voltage data and a first voltage prediction value calculated and output from the first mapping relationship, and to calculate a second error based on the actual voltage data and a second voltage prediction value calculated and output from the second mapping relationship; the first error and the second error are integrated into a voltage prediction error through the numerical integration method. The judgment module is used to determine the fault judgment result of the lithium-ion battery under test based on the first error, the second error, the voltage prediction error and the voltage prediction error threshold, through a preset logical relationship.

10. An electronic device, characterized in that, include: A processor and a memory, wherein the memory stores a non-volatile storage medium, and when the non-volatile storage medium is invoked by the processor, the processor executes the fault detection method for a lithium-ion battery as described in any one of claims 1-6.

Images