Lithium ion battery fault diagnosis method and system, storage medium and computing equipment

Abstract

The invention relates to the field of lithium ion batteries, and discloses a lithium ion battery fault diagnosis method and system, a storage medium and computing device.The lithium ion battery fault diagnosis method comprises the steps that operation data in the running process of an electric vehicle is obtained in real time, the change amplitude, duration and change trend of voltage, current and temperature in the operation data are analyzed, and the voltage sudden change point of a single body is marked; the method comprises the following steps: acquiring fault vehicle full life cycle data collected by a BMS, taking voltage, temperature, current and time as characteristic values, taking a fault type as a label quantity, and constructing a fault type identification model through a machine learning method; and extracting characteristic values and inputting the characteristic values into the fault type identification model in the process of acquiring the operation data of the voltage sudden change point of the single body, outputting the probability of each fault type, and if the fault probability is higher than a set confidence interval, judging that a fault occurs and triggering an alarm. According to the invention, real-time monitoring, fault type identification and fault single body alarm of single body voltage abrupt change in the battery pack can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium-ion batteries, and in particular to a lithium-ion battery fault diagnosis method, system, storage medium and computing device based on voltage mutation. Background Art

[0002] With the growing global demand for clean energy, lithium-ion batteries, as efficient, clean, and high-energy-density energy storage devices, have been widely used in many fields. However, lithium-ion batteries pose many safety risks during use. In particular, automotive lithium-ion batteries face major safety challenges due to their complex operating conditions. Among them, abnormal reduction in single-cell voltage is a key issue. Because lithium-ion battery packs are typically composed of multiple single cells connected in series or parallel, the performance and status of each single cell may be inconsistent during the charge and discharge process. For example, some single cells may experience a reduction in single-cell voltage due to factors such as aging, internal resistance differences, uneven heat dissipation, and abnormal connections.

[0003] Traditional lithium-ion battery cell voltage monitoring methods mostly monitor voltage changes only under static conditions. When a cell voltage falls below a set threshold, an alarm is triggered (for example, see patent publication number CN119283637A). However, this approach has significant limitations. It cannot monitor abnormal cell voltage changes under a wider range of operating conditions, nor can it determine the cause of abnormal voltage changes. Ensuring the reliability of battery troubleshooting and reducing maintenance difficulty has become a pressing technical challenge. Summary of the Invention

[0004] In response to the above problems, the purpose of the present invention is to provide a lithium-ion battery fault diagnosis method, system, storage medium and computing device, which can realize real-time monitoring of cell voltage mutations in a battery pack, fault type identification and fault cell alarm.

[0005] To achieve the above objectives, the present invention, in its first aspect, adopts a technical solution: a lithium-ion battery fault diagnosis method, comprising: acquiring real-time operating data from electric vehicles during driving, analyzing the amplitude, duration, and trend of voltage, current, and temperature changes in the operating data, and marking the voltage mutation points of each battery cell; acquiring full lifecycle data of the faulty vehicle collected by the battery management system (BMS), using voltage, temperature, current, and time as feature values ​​and the fault type as a label, and constructing a fault type identification model through machine learning methods; extracting feature values ​​from the operating data of the battery cell voltage mutation points and inputting them into the fault type identification model, outputting the probability of each fault type, and determining that a fault has occurred if the fault probability is higher than a set confidence interval, triggering an alarm. By real-time monitoring of battery cell voltage mutations and combining them with the fault type identification model, battery cell voltage faults can be accurately identified and diagnosed.

[0006] Furthermore, the voltage, current, and temperature variation in the operating data are analyzed in terms of their magnitude, duration, and trend, and the voltage mutation points of the cells are marked, including:

[0007] Process the real-time collected operating data and remove operating data parameters that exceed normal thresholds;

[0008] The amplitude and duration of voltage changes are calculated using a sliding window algorithm. If both the amplitude and duration of voltage changes exceed the set threshold, the time window is marked as an abnormal window.

[0009] Furthermore, the amplitude and duration of the voltage change are calculated using a sliding window algorithm. If both the amplitude and duration of the voltage change exceed the set threshold, the time window is marked as an abnormal window. Specifically:

[0010] Obtain the voltage data of each cell in the current window, perform statistics on the voltage data in the window according to the set window size and sliding step size, and calculate the voltage mean and standard deviation in each time window;

[0011] The deviation between the voltage value of each cell in the current window and the window mean is calculated. If the deviation exceeds the preset mutation threshold, it is considered that a voltage mutation has occurred at this point and marked as an abnormal time window.

[0012] Furthermore, the deviation between the voltage value of each cell in the current window and the window mean is calculated. If the deviation exceeds the preset mutation threshold, it is considered that a voltage mutation has occurred at this point and marked as an abnormal time window, including:

[0013] Calculate the difference between the cell voltage value and the window mean voltage at the moment of mutation as the voltage drop amplitude V drop , and calculate the duration of the monomer voltage drop at the moment of the mutation as the voltage mutation duration T drop ;

[0014] If the vehicle is in a stationary state or charging state, determine V drop and T drop whether both are greater than a set first voltage mutation threshold and a first duration threshold;

[0015] If the vehicle is in driving state, determine V drop and T drop whether both are greater than a set second voltage mutation threshold and a second duration threshold;

[0016] If any state is greater than the set threshold, it is determined that there is a single cell voltage mutation phenomenon at the current moment and marked as an abnormal time window.

[0017] Furthermore, a fault type identification model is constructed through machine learning methods, including:

[0018] Based on the known faulty battery pack information, historical operating parameter data and fault type determination results are obtained respectively;

[0019] Perform data cleaning on the acquired battery pack operating parameter data, and normalize the cleaned voltage, current, and temperature data;

[0020] Extract feature values ​​from the normalized data. Use the fault type as a label and use correlation analysis to output the feature with the highest correlation with the fault type.

[0021] Select a machine learning model, input the feature values ​​into the model for model training, and perform model evaluation to output a fault type identification model with prediction accuracy that meets the conditions.

[0022] Furthermore, in the operating data of the voltage mutation point of the single cell, the characteristic value is extracted and input into the fault type identification model, and the probability of each fault type is output. If the fault probability is higher than the set confidence interval, it is determined that a fault has occurred and an alarm is triggered, including:

[0023] Acquire real-time current, temperature, and voltage data in the abnormal window, extract the eigenvalues ​​required to build the fault type identification model, and normalize the eigenvalues;

[0024] The normalized characteristic values ​​are input into the fault type identification model to output the fault probability of each fault type;

[0025] If the failure probability is higher than the set alarm threshold, an alarm is triggered to comprehensively analyze the changing characteristics of multiple parameters such as voltage, current and temperature, and accurately judge the fault type at the moment of voltage mutation when the voltage suddenly drops. Based on the fault type determined by the model, the possible fault cause and fault module are uploaded to the remote monitoring platform, and the alarm information and related battery status data are displayed in real time on the platform.

[0026] Furthermore, the fault type at the moment of sudden voltage drop is accurately determined, including:

[0027] If the cell voltage suddenly drops at the moment of voltage mutation, but the temperature and current do not show abnormal changes, the model will determine the fault type as a connection fault based on relevant characteristics;

[0028] If the cell voltage suddenly drops at the moment of voltage mutation, accompanied by an abnormal increase in temperature, but the current remains unchanged, the model will determine that the fault type is an internal short circuit based on these characteristics.

[0029] In the second aspect, the technical solution adopted by the present invention is: a lithium-ion battery fault diagnosis system, which includes: a real-time data analysis module, which obtains the operating data of the electric vehicle in real time during driving, analyzes the change amplitude, duration and change trend of the voltage, current and temperature in the operating data, and marks the voltage mutation point of the single cell; a fault type identification module, which obtains the full life cycle data of the faulty vehicle collected by the BMS, takes the voltage, temperature, current and time as characteristic values, and the fault type as the label quantity, and constructs a fault type identification model through machine learning methods; a fault single cell alarm module, which extracts characteristic values ​​from the operating data of the voltage mutation point of the single cell and inputs them into the fault type identification model, outputs the probability of each fault type, and if the fault probability is higher than the set confidence interval, it is determined that a fault has occurred and an alarm is triggered.

[0030] In a third aspect, the technical solution adopted by the present invention is: a computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, and when the instructions are executed by a computing device, the computing device executes any one of the above methods.

[0031] In a fourth aspect, the technical solution adopted by the present invention is: a computing device, comprising: one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the above methods.

[0032] The present invention has the following advantages due to the adoption of the above technical solution:

[0033] 1. The present invention can monitor voltage mutations in real time and quickly identify fault moments through the BMS system; and with the help of multi-parameter fusion and machine learning algorithms, it significantly improves the accuracy of fault type identification.

[0034] 2. The machine learning algorithm used in this invention possesses powerful self-learning capabilities, and through continuous iteration, it can continuously improve the model's prediction accuracy. This not only enables fault alarms but also provides fault cause analysis, providing an important reference for troubleshooting and troubleshooting, effectively preventing further damage to the battery pack or more serious safety incidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a flow chart of a lithium-ion battery fault diagnosis method based on voltage mutation according to an embodiment of the present invention;

[0036] Figure 2 This is a flow chart of a real-time data analysis module according to an embodiment of the present invention;

[0037] Figure 3This is a schematic diagram of the fault judgment module model training process in an embodiment of the present invention;

[0038] Figure 4 This is a schematic diagram of the fault type alarm module flow in an embodiment of the present invention;

[0039] Figure 5 It is a schematic structural diagram of a lithium-ion battery fault diagnosis system based on voltage mutation in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] With the continuous development of battery technology and the increasing complexity of application scenarios, the requirements for battery safety and reliability are also getting higher and higher. The abnormal voltage reduction phenomenon caused by different fault causes has different degrees of influence on battery safety performance, which increases the difficulty of fault detection and maintenance. In order to solve these technical problems, the present invention integrates various technical means such as battery management system (BMS) technology, data analysis and processing technology, and multi-parameter fusion model construction. Based on the real-time data of power vehicle operation and the historical data of the entire life cycle of the faulty vehicle, a lithium-ion battery fault diagnosis method, system, storage medium and computing device are provided, which can diagnose the fault type in real time, realize the alarm of the abnormal voltage reduction time of the single cell and the cause of the fault, and ensure the safe and stable operation of the battery system.

[0041] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of the present invention.

[0042] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0043] In one embodiment of the present invention, a lithium-ion battery fault diagnosis method is provided for real-time monitoring of voltage changes of cells in a battery pack, identifying the fault type and locating the faulty cell. Figure 1 As shown, the method includes the following steps:

[0044] 1) Real-time acquisition of operating data of electric vehicles during driving, including acquisition time, current, voltage of each cell in the battery pack, temperature of each battery pack in the battery pack, etc., analyze the change amplitude, duration and change trend of parameters such as voltage, current and temperature, and mark the voltage mutation point of the cell.

[0045] In this embodiment, a big data query method is used to obtain data such as the collection time, current, voltage of each cell in the battery pack, and temperature of each battery pack during the driving process of the electric vehicle from the national monitoring and management platform of new energy vehicles in real time.

[0046] 2) Obtain the full life cycle data of the faulty vehicle collected by the BMS, use parameters such as voltage, temperature, current, time as feature values, and the fault type as the label quantity, and build a fault type recognition model through machine learning methods.

[0047] 3) Extract characteristic values ​​from the acquired voltage mutation point operating data of the single cell and input them into the fault type identification model, outputting the probability of each fault type. If the fault probability is higher than the set confidence interval, it is determined that a fault has occurred and an alarm is triggered.

[0048] When in use, the present invention realizes real-time monitoring of cell voltage mutations in a battery pack, fault type identification and fault cell alarm through the coordinated work of real-time data analysis, fault type identification and fault cell alarm.

[0049] In the above step 1), if Figure 2 As shown, analyzing the variation amplitude, duration, and variation trend of voltage, current, and temperature in the operating data and marking the voltage mutation point of the battery cell includes the following steps:

[0050] 1.1) Process the real-time collected operating data and remove operating data parameters that exceed the normal threshold;

[0051] Specifically, the data is first checked for consistency and rationality. For example, data is cleaned based on the physically reasonable range of current and voltage. If the value range of a single cell voltage is less than 0V or greater than 5V, the data is considered invalid and discarded. Similarly, if the current or temperature data contains typical invalid values ​​such as 65535 or 254, these values ​​are also discarded to avoid misjudgments.

[0052] 1.2) The amplitude and duration of voltage change are calculated using a sliding window algorithm. If both the amplitude and duration of voltage change exceed the set threshold, the time window is marked as an abnormal window.

[0053] In step 1.2) above, the amplitude and duration of the voltage change are calculated using a sliding window algorithm. If both the amplitude and duration of the voltage change exceed the set threshold, the time window is marked as an abnormal window. Specifically, the following steps are included:

[0054] 1.2.1) Get the voltage data of each cell in the current window, and calculate the voltage data (x i ) are used for statistics, and the voltage mean (μ) and standard deviation (σ) in each time window are calculated respectively.

[0055] In this embodiment, the calculation formulas for the voltage mean (μ) and standard deviation (σ) are:

[0056]

[0057] Where n is the total number of voltage data in the window.

[0058] 1.2.2) Calculate the deviation between the voltage value of each cell in the current window and the window mean. If the deviation exceeds the preset mutation threshold, it is considered that a voltage mutation has occurred at this point and marked as an abnormal time window.

[0059] In this embodiment, the Z-score method is used to judge the mutation. The deviation Z of each monomer voltage value in the current window and the window mean is calculated. i for:

[0060]

[0061] For example, in this embodiment, the mutation threshold is set to 3, that is, when the voltage mutation |Zi|>3, it is determined that the voltage value in the time window may have a mutation.

[0062] In this embodiment, the deviation between the voltage value of each cell in the current window and the window mean is calculated. If the deviation exceeds a preset mutation threshold, it is considered that a voltage mutation has occurred at this point and marked as an abnormal time window. The following steps are included:

[0063] (1) Calculate the difference between the cell voltage value and the window mean voltage at the moment of the sudden change as the voltage drop amplitude V drop , and calculate the duration of the monomer voltage drop at the moment of the mutation as the voltage mutation duration T drop ;

[0064] (2) If the vehicle is in a stationary state or charging state, determine V drop and T drop whether both are greater than a set first voltage mutation threshold and a first duration threshold;

[0065] (3) If the vehicle is in motion, determine V drop and Tdrop whether both are greater than a set second voltage mutation threshold and a second duration threshold;

[0066] (4) If any state in step (2) to step (3) is greater than the set threshold, it is determined that there is a single cell voltage mutation phenomenon at the current moment, and it is marked as an abnormal time window.

[0067] In this embodiment, optionally, the first voltage mutation threshold is 20 mV, the first duration threshold is 5 minutes, the second voltage mutation threshold is 300 mV, and the second duration threshold is 3 minutes.

[0068] In the above step 2), a fault type recognition model is constructed based on the fault vehicle database, and the historical operating parameter data of the vehicle is obtained. The voltage, temperature, current, time and other parameters are used as feature values, and the fault type is used as the label quantity. The fault type recognition model is constructed through machine learning methods, such as Figure 3 As shown, the following steps are included:

[0069] 2.1) Based on the known faulty battery pack information, obtain historical operating parameter data and fault type determination results.

[0070] In this example, the BMS collects data such as time series, voltage, current, temperature, insulation resistance, and mileage throughout the vehicle's life cycle. The fault type at each point in time is annotated as a label quantity. Label quantities include normal, internal short circuit, connection fault, and electrolyte leakage.

[0071] 2.2) Perform data cleaning on the acquired battery pack operating parameter data. Data cleaning includes processing outliers and missing values ​​in the data; and normalizing the cleaned voltage, current, and temperature data.

[0072] In this embodiment, the acquired battery pack operating parameter data is processed, including outlier and missing value processing. The outlier processing method is similar to that used for real-time data, eliminating invalid voltage, current, and temperature parameters. Missing value processing involves using linear interpolation to fill in missing data for a time less than a set time threshold. Data missing for a time exceeding the set time threshold is directly eliminated to avoid affecting the accuracy of the dataset.

[0073] Furthermore, the voltage, current and temperature data after cleaning are normalized. In this embodiment, the Z-score normalization method is used for normalization, where μ is the mean and σ is the standard deviation.

[0074]

[0075] Normalization is performed because different metrics have different scales. For example, voltage ranges from 0 to 5000, while temperature ranges from 10 to 100. From a numerical perspective, variables with large numerical variations have a greater absolute impact when performing comprehensive evaluations. Therefore, it is necessary to normalize the extracted feature data of different types, consolidating the data to the range of [0, 1] or [-1, 1] to eliminate the impact of different data scales on subsequent analysis. Common normalization methods include min-max normalization and Z-score normalization.

[0076] 2.3) Perform feature extraction on the normalized data, extracting features such as the mean, variance, minimum, maximum, and voltage change rate of each time window; the maximum, minimum, and temperature change rate of each time window; and the mean and fluctuation amplitude of each time window. Furthermore, the fault type in each time window is labeled as a label, and correlation analysis is performed to output the feature with the highest correlation with the fault type.

[0077] Specifically, characteristic parameters with higher correlation and importance are selected through characteristic analysis. The Pearson correlation coefficient analysis method adopted in this embodiment outputs characteristic parameters with higher correlation with the fault type.

[0078] 2.4) Select a machine learning model and input the feature values ​​into the model for model training, and then perform model evaluation to output a fault type identification model with prediction accuracy that meets the requirements.

[0079] In this embodiment, a gradient boosting tree (XGBoost) model is used for training. First, the feature value dataset is divided into a training set and a test set in a ratio of 7:3. Subsequently, the training set is input into the gradient boosting tree model for training, and the hyperparameters are optimized through grid search and cross-validation to output a model with high prediction accuracy. Next, the test set data is input into the model for verification. Based on evaluation indicators such as accuracy and recall rate, the optimal model is finally output as the fault type recognition model.

[0080] In this embodiment, the fault types include, but are not limited to, internal short circuit, connection fault, electrolyte leakage, and other faults that may cause abnormal changes in battery cell voltage.

[0081] In the above step 3), by integrating the multi-parameter fault type identification model, it is accurately determined whether there is a voltage mutation phenomenon in the single cell voltage, and the corresponding fault type is efficiently output.

[0082] Specifically, such as Figure 4As shown, in the obtained operating data of the voltage mutation point of the battery cell, the characteristic value is extracted and input into the fault type identification model, and the probability of each fault type is output. If the fault probability is higher than the set confidence interval, it is determined that a fault has occurred and an alarm is triggered. The following steps are included:

[0083] 3.1) Obtain real-time current, temperature, and voltage data for the abnormal window, extract the eigenvalues ​​required to build the fault type identification model, and normalize the eigenvalues;

[0084] 3.2) Input the normalized feature values ​​into the fault type identification model and output the fault probability of each fault type;

[0085] 3.3) If the failure probability is higher than the set alarm threshold, an alarm is triggered. The changing characteristics of multiple parameters such as voltage, current, and temperature are comprehensively analyzed to accurately determine the fault type at the moment of sudden voltage drop. Based on the fault type determined by the model, the possible fault cause and fault module are uploaded to the remote monitoring platform. The alarm information and related battery status data are displayed in real time on the platform, facilitating remote monitoring and processing by management personnel.

[0086] In this embodiment, the alarm threshold p is set to 0.9. If the probability of a fault type exceeds 0.9, an alarm is triggered. The model can comprehensively analyze the changing characteristics of multiple parameters such as voltage, current, and temperature, and accurately determine the fault type when the voltage drops abnormally.

[0087] In this embodiment, the fault type at the moment of sudden voltage drop is accurately determined, including:

[0088] If the cell voltage suddenly drops at the moment of voltage mutation, but the temperature and current do not show abnormal changes, the model will determine the fault type as a connection fault based on relevant characteristics;

[0089] If a cell voltage suddenly drops during a voltage mutation, accompanied by an abnormally high temperature, while the current remains constant, the model will identify the fault as an internal short circuit based on these characteristics. The abnormal temperature increase refers to comparing the temperature change with a set temperature threshold. If the temperature change exceeds the set temperature threshold, the cell temperature is considered abnormally high at that moment.

[0090] In this embodiment, the alarm information includes the fault time, the faulty unit, and the fault type.

[0091] In summary, the present invention utilizes multi-parameter fusion and machine learning algorithms to improve fault identification accuracy, optimizes model prediction accuracy through continuous iteration, implements fault alarm and cause analysis, provides a reference for maintenance and troubleshooting, and prevents battery pack damage and safety accidents.

[0092] like Figure 5 As shown, in one embodiment of the present invention, a lithium-ion battery fault diagnosis system is provided, which includes:

[0093] Real-time data analysis module, which obtains the operating data of electric vehicles in real time, analyzes the change amplitude, duration and trend of voltage, current and temperature in the operating data, and marks the voltage mutation point of battery cells;

[0094] The fault type identification module obtains the full life cycle data of the faulty vehicle collected by the BMS, uses voltage, temperature, current, and time as feature values ​​and the fault type as the label, and builds a fault type identification model through machine learning methods;

[0095] The faulty cell alarm module extracts characteristic values ​​from the acquired operating data of the voltage mutation point of the battery cell and inputs them into the fault type identification model, outputting the probability of each fault type. If the fault probability is higher than the set confidence interval, it is determined that a fault has occurred and an alarm is triggered.

[0096] In the above embodiment, analyzing the variation amplitude, duration, and variation trend of voltage, current, and temperature in the operating data and marking the voltage mutation point of the battery cell includes:

[0097] Process the real-time collected operating data and remove operating data parameters that exceed normal thresholds;

[0098] The amplitude and duration of voltage changes are calculated using a sliding window algorithm. If both the amplitude and duration of voltage changes exceed the set threshold, the time window is marked as an abnormal window.

[0099] In the above embodiment, the amplitude and duration of the voltage change are calculated using a sliding window algorithm. If both the amplitude and duration of the voltage change exceed the set threshold, the time window is marked as an abnormal window. Specifically:

[0100] Obtain the voltage data of each cell in the current window, perform statistics on the voltage data in the window according to the set window size and sliding step size, and calculate the voltage mean and standard deviation in each time window;

[0101] The deviation between the voltage value of each cell in the current window and the window mean is calculated. If the deviation exceeds the preset mutation threshold, it is considered that a voltage mutation has occurred at this point and marked as an abnormal time window.

[0102] In this embodiment, the deviation between the voltage value of each cell in the current window and the window mean is calculated. If the deviation exceeds a preset mutation threshold, it is considered that a voltage mutation has occurred at that point and marked as an abnormal time window, including:

[0103] Calculate the difference between the cell voltage value and the window mean voltage at the moment of the current voltage mutation as the voltage sag amplitude V drop , and calculate the duration of the monomer voltage reduction at the current voltage mutation moment as the voltage mutation duration T drop ;

[0104] If the vehicle is in a stationary state or charging state, determine V drop and T drop whether both are greater than a set first voltage mutation threshold and a first duration threshold;

[0105] If the vehicle is in driving state, determine V drop and T drop whether both are greater than a set second voltage mutation threshold and a second duration threshold;

[0106] If any state is greater than the set threshold, it is determined that there is a single cell voltage mutation phenomenon at the current moment and marked as an abnormal time window.

[0107] In the above embodiment, the fault type identification model is constructed by a machine learning method, including:

[0108] Based on the known faulty battery pack information, historical operating parameter data and fault type determination results are obtained respectively;

[0109] Perform data cleaning on the acquired battery pack operating parameter data, and normalize the cleaned voltage, current, and temperature data;

[0110] Extract feature values ​​from the normalized data. Use the fault type as a label and use correlation analysis to output the feature with the highest correlation with the fault type.

[0111] Select a machine learning model, input the feature values ​​into the model for model training, and perform model evaluation to output a fault type identification model with prediction accuracy that meets the conditions.

[0112] In the above embodiment, characteristic values ​​are extracted from the acquired operating data of the voltage mutation point of the battery cell and input into the fault type identification model, and the probability of each fault type is output. If the fault probability is higher than the set confidence interval, it is determined that a fault has occurred and an alarm is triggered, including:

[0113] Acquire real-time current, temperature, and voltage data in the abnormal window, extract the eigenvalues ​​required to build the fault type identification model, and normalize the eigenvalues;

[0114] The normalized characteristic values ​​are input into the fault type identification model to output the fault probability of each fault type;

[0115] If the failure probability is higher than the set alarm threshold, an alarm is triggered to comprehensively analyze the changing characteristics of multiple parameters such as voltage, current and temperature, and accurately judge the fault type at the moment of voltage mutation when the voltage suddenly drops. Based on the fault type determined by the model, the possible fault cause and fault module are uploaded to the remote monitoring platform, and the alarm information and related battery status data are displayed in real time on the platform.

[0116] In this embodiment, the fault type at the moment of sudden voltage drop is accurately determined, including:

[0117] If the cell voltage suddenly drops at the moment of voltage mutation, but the temperature and current do not show abnormal changes, the model will determine the fault type as a connection fault based on relevant characteristics;

[0118] If the cell voltage suddenly drops at the moment of voltage mutation, accompanied by an abnormal increase in temperature, but the current remains unchanged, the model will determine that the fault type is an internal short circuit based on these characteristics.

[0119] The system provided in this embodiment is used to execute the above-mentioned method embodiments. Please refer to the above-mentioned embodiments for specific processes and detailed contents, which will not be repeated here.

[0120] In one embodiment of the present invention, a computing device is provided. The computing device may be a terminal, which may include: Department Processor, Communications Interface, Memory, Display and Input device. The processor, communication interface, and memory communicate with each other via a communication bus. The processor is used to provide computing and control capabilities. The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. When the computer program is executed by the processor, it implements the methods in the above embodiments; the internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The communication interface is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be achieved through WIFI, a management network, NFC (near field communication) or other technologies. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computing device, or an external keyboard, touchpad or mouse. The processor can call the logical instructions in the memory.

[0121] In addition, the logical instructions in the above-mentioned memory can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0122] In one embodiment of the present invention, a computer program product is provided, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the methods provided by the above-mentioned method embodiments.

[0123] In one embodiment of the present invention, a non-transitory computer-readable storage medium is provided. The non-transitory computer-readable storage medium stores server instructions. The computer instructions enable a computer to execute the methods provided in the above embodiments.

[0124] The above embodiment provides a computer-readable storage medium, whose implementation principle and technical effects are similar to those of the above method embodiment, and will not be repeated here.

[0125] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0126] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0127] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A lithium-ion battery fault diagnosis method, characterized in that: include: Real-time acquisition of electric vehicle operating data during driving, analysis of the voltage, current, and temperature variation, duration, and trend in the operating data, and marking of voltage mutation points of individual cells; Obtain the full life cycle data of the faulty vehicle collected by the BMS, use voltage, temperature, current, and time as feature values, and the fault type as the label quantity, and build a fault type recognition model through machine learning methods; In the operating data of the voltage mutation point of the single cell, the characteristic value is extracted and input into the fault type identification model, and the probability of each fault type is output. If the fault probability is higher than the set confidence interval, it is determined that a fault has occurred and an alarm is triggered.

2. The lithium-ion battery fault diagnosis method according to claim 1, wherein: Analyze the change amplitude, duration, and trend of voltage, current, and temperature in the operating data, and mark the voltage mutation points of the monomers, including: Process the real-time collected operating data and remove operating data parameters that exceed normal thresholds; The amplitude and duration of voltage changes are calculated using a sliding window algorithm. If both the amplitude and duration of voltage changes exceed the set threshold, the time window is marked as an abnormal window.

3. The lithium-ion battery fault diagnosis method according to claim 2, wherein: The voltage change amplitude and duration are calculated using a sliding window algorithm. If both the voltage change amplitude and duration exceed the set threshold, the time window is marked as an abnormal window. Specifically: Obtain the voltage data of each cell in the current window, perform statistics on the voltage data in the window according to the set window size and sliding step size, and calculate the voltage mean and standard deviation in each time window; The deviation between the voltage value of each cell in the current window and the window mean is calculated. If the deviation exceeds the preset mutation threshold, it is considered that a voltage mutation has occurred at this point and marked as an abnormal time window.

4. The lithium-ion battery fault diagnosis method according to claim 3, wherein: Calculate the deviation between the voltage value of each cell in the current window and the window mean. If the deviation exceeds the preset mutation threshold, it is considered that a voltage mutation has occurred at that point and marked as an abnormal time window, including: Calculate the difference between the cell voltage value and the window mean voltage at the moment of mutation as the voltage drop amplitude V drop , and calculate the duration of the monomer voltage drop at the moment of the mutation as the voltage mutation duration T drop ; If the vehicle is in a stationary state or charging state, determine V drop and T drop whether both are greater than a set first voltage mutation threshold and a first duration threshold; If the vehicle is in driving state, determine V drop and T drop whether both are greater than a set second voltage mutation threshold and a second duration threshold; If any state is greater than the set threshold, it is determined that there is a single cell voltage mutation phenomenon at the current moment and marked as an abnormal time window.

5. The lithium-ion battery fault diagnosis method according to claim 1, wherein: A fault type identification model is constructed using machine learning methods, including: Based on the known faulty battery pack information, historical operating parameter data and fault type determination results are obtained respectively; Perform data cleaning on the acquired battery pack operating parameter data, and normalize the cleaned voltage, current, and temperature data; Extract feature values ​​from the normalized data. Use the fault type as a label and use correlation analysis to output the feature with the highest correlation with the fault type. Select a machine learning model, input the feature values ​​into the model for model training, and perform model evaluation to output a fault type identification model with prediction accuracy that meets the conditions.

6. The lithium-ion battery fault diagnosis method according to claim 1, wherein: In the operating data of the voltage mutation point of the single cell, the characteristic value is extracted and input into the fault type identification model, and the probability of each fault type is output. If the fault probability is higher than the set confidence interval, it is determined that a fault has occurred and an alarm is triggered, including: Acquire real-time current, temperature, and voltage data in the abnormal window, extract the eigenvalues ​​required to build the fault type identification model, and normalize the eigenvalues; The normalized characteristic values ​​are input into the fault type identification model to output the fault probability of each fault type; If the failure probability is higher than the set alarm threshold, an alarm is triggered to comprehensively analyze the changing characteristics of multiple parameters such as voltage, current and temperature, and accurately judge the fault type at the moment of voltage mutation when the voltage suddenly drops. Based on the fault type determined by the model, the possible fault cause and fault module are uploaded to the remote monitoring platform, and the alarm information and related battery status data are displayed in real time on the platform.

7. The lithium-ion battery fault diagnosis method according to claim 6, wherein: Accurately determine the fault type at the moment of sudden voltage drop, including: If the cell voltage suddenly drops at the moment of voltage mutation, but the temperature and current do not show abnormal changes, the model will determine the fault type as a connection fault based on relevant characteristics; If the cell voltage suddenly drops at the moment of voltage mutation, accompanied by an abnormal increase in temperature, but the current remains unchanged, the model will determine that the fault type is an internal short circuit based on these characteristics.

8. A lithium-ion battery fault diagnosis system, characterized in that: include: Real-time data analysis module, which obtains the operating data of electric vehicles in real time, analyzes the change amplitude, duration and trend of voltage, current and temperature in the operating data, and marks the voltage mutation point of the monomer; The fault type identification module obtains the full life cycle data of the faulty vehicle collected by the BMS, uses voltage, temperature, current, and time as feature values ​​and the fault type as the label, and builds a fault type identification model through machine learning methods; The faulty cell alarm module extracts characteristic values ​​from the operating data of the cell's voltage mutation point and inputs them into the fault type identification model, outputting the probability of each fault type. If the fault probability is higher than the set confidence interval, it is determined that a fault has occurred and an alarm is triggered.

9. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform any one of the methods of claims 1 to 7 .

10. A computing device, characterized in that include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the methods according to claims 1 to 7.

Images