Battery fault prediction method and device, electronic equipment and storage medium

Abstract

The invention discloses a battery fault prediction method and device, electronic equipment and a storage medium, and belongs to the technical field of lithium batteries, and the method comprises the steps: taking the charging and discharging parameters of a battery pack as the input, taking the ideal voltage of the battery pack as the output, and constructing a neural network model; obtaining a charge-discharge time sequence parameter set of the battery pack to be detected and a second curve of the battery monomers in the battery pack to be detected, wherein the second curve is a voltage change curve of the battery monomers; sequentially inputting time sequence parameters in a time sequence parameter set of charging and discharging into the neural network model to obtain a first curve, wherein the first curve is an ideal voltage change curve of the battery pack; calculating class correlation coefficients of the first curve and the second curve; and according to the class correlation coefficient and a preset fault threshold value, positioning a fault cell and predicting a fault type. By comparing the class correlation between the voltage of the single battery and the ideal voltage in the charging and discharging process, the fault single battery is positioned, the fault type is predicted, and the fault prediction in the initial stage of the battery fault is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium batteries, and in particular to a battery failure prediction method, device, electronic equipment and storage medium. Background Art

[0002] Lithium-ion batteries are widely used in different scenarios such as electronic consumer products, electric vehicles, distributed energy storage, and large-scale energy storage due to their advantages such as high energy density and long cycle life. Especially in the application of electric vehicles, distributed energy storage, and large-scale energy storage, in order to meet the needs of current, voltage, power, and energy, it is often necessary to connect a large number of monomers in series and parallel to form battery packs, battery packs, and even battery clusters. This will result in a large number of connection components, which greatly increases the complexity of the system, resulting in an increased probability of various types of failures and increased safety hazards. Therefore, implementing accurate and reliable fault diagnosis is the key to ensuring the safe, stable, and reliable operation of the battery system.

[0003] In the existing technology, there are relevant standards that put forward requirements for the BMS fault alarm system. The battery management system (BMS) is responsible for status monitoring, active balancing, charge and discharge control, fault alarm, etc. of the lithium battery system to ensure the safe and efficient operation of the battery. However, as the most basic guarantee for the safe operation of the battery system, the fault alarm system has developed to a relatively serious stage when the alarm occurs, causing great losses to the battery system, and also puts forward higher requirements for the structure and fire protection of the battery system.

[0004] Therefore, there is an urgent need for a battery fault prediction method that can locate the faulty cell and predict the fault by analyzing the voltage change curve during the charging and discharging process in the early stage of the fault. Summary of the invention

[0005] In view of this, it is necessary to provide a battery fault prediction method, device, electronic device and storage medium, which can locate the faulty cell and predict the fault by analyzing the voltage change curve during the charging and discharging process in the early stage of the fault.

[0006] In order to solve the above technical problems, on the one hand, the present invention provides a battery failure prediction method, comprising: The charging and discharging parameters of the battery pack are used as input, and the ideal voltage of the battery pack is used as output to construct a neural network model; Acquire a charge and discharge timing parameter set of the battery pack to be tested and a second curve of a battery cell in the battery pack to be tested, wherein the second curve is a voltage variation curve of the battery cell; Inputting the timing parameters in the charging and discharging timing parameter set into the neural network model in sequence to obtain a first curve, wherein the first curve is an ideal voltage change curve of the battery pack; calculating a class correlation coefficient between the first curve and the second curve; According to the class correlation coefficient and the preset fault threshold, the faulty battery cell is located and the fault type is predicted.

[0007] In one possible implementation, the charging and discharging parameters of the battery pack are used as input, and the ideal voltage of the battery pack is used as output to construct a neural network model, including: An initial neural network model is constructed by taking the charge and discharge parameters at each sampling time during the charge and discharge process of the battery pack as input and the ideal voltage of the battery pack at the sampling time as output, wherein the charge and discharge parameters include the battery pack SOC value, ambient temperature and charge and discharge current; Obtain the voltage value, battery pack SOC value, charge and discharge current and ambient temperature of a battery pack with good performance at each sampling time during the charge and discharge process; Iteratively training the initial neural network model according to the battery pack SOC value, charge and discharge current, ambient temperature and voltage value at each sampling time to generate a first model; Based on the AdamW optimization algorithm, the first model is optimized to generate a battery pack neural network model.

[0008] In one possible implementation, the calculation formula of the AdamW optimization algorithm is: , in, is the neural network weight independent variable at the current moment, is the neural network weight independent variable at the previous moment, is the learning rate, is the adaptive adjustment parameter of the learning rate, is the cumulative gradient, is the weight decay parameter.

[0009] In a possible implementation, the timing parameters include a battery pack SOC value, an ambient temperature, a charge and discharge current, and a sampling time. The timing parameters in the charge and discharge timing parameter set are sequentially input into the neural network model to obtain a first curve, including: Inputting the battery pack SOC value, ambient temperature, and charge and discharge current in the timing parameters into the neural network model to obtain the ideal voltage value at each sampling time point; A first curve is generated according to the ideal voltage at each sampling time point and its corresponding sampling time.

[0010] In a possible implementation, the calculation formula of the class correlation coefficient between the first curve and the second curve is: , in, is the class correlation coefficient between the ideal voltage change curve of the battery pack and the voltage change curve of the battery cell, The ideal voltage change curve of the battery pack is The ideal voltage value for each sampling time is The voltage variation curve of the battery cell The voltage value at each sampling time, is the total number of sampling times.

[0011] In a possible implementation, determining the battery inconsistency fault type according to the class correlation coefficient and a preset fault threshold includes: When the class correlation coefficient is within a first preset fault threshold interval, it is determined that the battery has no inconsistent fault; When the class correlation coefficient is within a second preset fault threshold range, determining that the battery fault is a first-level inconsistency fault; When the class correlation coefficient is within a third preset fault threshold interval, it is determined that the battery fault is a secondary inconsistency fault.

[0012] In one possible implementation, it also includes: Determine the voltage change rate and temperature change rate of the battery cell within a preset time period during the charging and discharging process; Based on a preset voltage change rate alarm threshold and a temperature change rate alarm threshold, an alarm is issued to the battery cell according to the voltage change rate and the temperature change rate of the battery cell, and the alarm duration is counted; When the alarm duration exceeds a preset alarm duration threshold, the battery cell is diagnosed as a sudden internal short circuit fault.

[0013] In a second aspect, the present invention further provides a battery failure prediction device, comprising: A model building module, which is used to take the charging and discharging parameters of the battery pack as input and the ideal voltage of the battery pack as output to build a neural network model; A data acquisition module, used to acquire a charge and discharge timing parameter set of a battery pack to be tested and a second curve of a battery cell in the battery pack to be tested, wherein the second curve is a voltage variation curve of the battery cell; An ideal curve generating module, used for sequentially inputting the timing parameters in the charging and discharging timing parameter set into the neural network model to obtain a first curve, wherein the first curve is an ideal voltage variation curve of the battery pack; A class correlation coefficient calculation module, used to calculate the class correlation coefficient between the first curve and the second curve; The fault location and prediction module is used to locate the faulty battery cell and predict the fault type according to the class correlation coefficient and the preset fault threshold.

[0014] In a third aspect, the present invention also provides an electronic device, comprising a memory and a processor, wherein the memory is used to store programs and data; the processor is coupled to the memory, and is used to execute the program stored in the memory, to implement the battery fault prediction method as described above, and / or to implement the battery fault prediction as described above.

[0015] In a fourth aspect, the present invention further provides a computer storage medium for storing computer-readable programs or instructions, which, when executed by a processor, can implement the battery fault prediction method as described above.

[0016] The beneficial effects of the present invention are as follows: first, the charging and discharging parameters of the battery pack are used as input, and the ideal voltage of the battery pack is used as output to construct a neural network model; then, a charging and discharging timing parameter set of the battery pack to be tested and a second curve of the battery cell in the battery pack to be tested are obtained; then, the timing parameters in the charging and discharging timing parameter set are sequentially input into the neural network model to obtain a first curve, which is the ideal voltage change curve of the battery pack; finally, the class correlation coefficient of the first curve and the second curve is determined, and the faulty battery cell is located and the fault type is predicted according to the class correlation coefficient and the preset fault threshold. The present invention predicts the ideal voltage of the battery pack according to the charging and discharging parameters of the battery pack by constructing a model, generates an ideal voltage curve, and then obtains the voltage change curve of the battery cell during the charging and discharging process, calculates the class correlation coefficient between the two curves, locates the faulty battery cell according to the class correlation coefficient, and predicts the fault type. The present invention predicts battery faults by analyzing the class correlation between the ideal voltage curve of the battery pack and the voltage change curve of the battery cell during the charging and discharging process, and can accurately locate and predict faults in the early stage of the fault, thereby improving the safety and stability of the battery pack. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 A schematic flow chart of an embodiment of a battery failure prediction method provided by the present invention; Figure 2 For the present invention Figure 1 A schematic flow chart of an embodiment of step S101; Figure 3 For the present invention Figure 1 A schematic flow chart of an embodiment of step S102; Figure 4 For the present invention Figure 1 A schematic flow chart of an embodiment of step S105; Figure 5 A schematic diagram of a flow chart of an embodiment of a sudden internal short circuit fault provided by the present invention; Figure 6 A schematic diagram of the structure of an embodiment of a battery failure prediction electronic device provided by the present invention; Figure 7 A schematic diagram of the structure of an embodiment of the battery fault prediction storage medium provided by the present invention. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0020] In the description of the embodiments of the present invention, unless otherwise specified, “plurality” means two or more than two.

[0021] The descriptions of "first", "second", etc. involved in the embodiments of the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the technical features defined as "first" or "second" may explicitly or implicitly include at least one of the features.

[0022] Reference to an "embodiment" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0023] The present invention provides a battery fault prediction method, device, electronic device and storage medium, which are described below respectively.

[0024] Figure 1 A schematic flow chart of an embodiment of a battery failure prediction method provided by the present invention is shown in FIG. Figure 1 As shown, the battery failure prediction method includes: S101, using the charging and discharging parameters of the battery pack as input and the ideal voltage of the battery pack as output to construct a neural network model; S102, obtaining a charge and discharge timing parameter set of the battery pack to be tested and a second curve of a battery cell in the battery pack to be tested, where the second curve is a voltage variation curve of the battery cell; S103, sequentially inputting the timing parameters in the charging and discharging timing parameter set into the neural network model to obtain a first curve, where the first curve is an ideal voltage change curve of the battery pack; S104, calculating the class correlation coefficient between the first curve and the second curve; S105 . Locate the faulty battery cell and predict the fault type based on the class correlation coefficient and the preset fault threshold.

[0025] It should be noted that the battery pack in this example is used in lithium battery systems including battery management systems BMS such as electric ships and electric vehicles, and a battery charging and discharging integrated machine is used to charge and discharge the lithium battery pack; the battery management system BMS is used to collect and monitor the charging and discharging parameter data of the battery pack and each battery cell, and a neural network model is built through a model building tool on a mobile terminal, and the charging and discharging parameters of the battery cell and the corresponding time series are analyzed through data analysis software to draw a voltage change curve, analyze and calculate the class correlation coefficient between the curves, and analyze and compare the class correlation coefficients; the final calculation results, as well as the results of locating the faulty battery cell and predicting the fault type are displayed through a mobile terminal, and the mobile terminal can be a variety of electronic devices with a display screen that supports data analysis and model building, including but not limited to smart phones, tablet computers, laptop portable computers, and desktop computers. The battery fault prediction method provided in this embodiment can be implemented by an application, a small program, or a web page on a mobile terminal.

[0026] It should be further explained that, at a specific ambient temperature, the battery pack is charged and discharged at a constant current, and the charge and discharge parameters of the battery pack and each battery cell are collected based on the battery management system BMS. The charge and discharge parameters of the battery pack include the SOC of the battery pack, the battery pack voltage and the sampling time, and the charge and discharge parameters of the battery cell include the battery cell voltage and the sampling time; the collected data are cleaned and normalized; the interval of each sampling time is set according to the experimental needs, and can be set to 0.5 seconds, 1 second, 2 seconds or 5 seconds, etc. In this embodiment, the sample is sampled once every 1 second, and the shorter the sampling time interval, the more accurate the calculation result.

[0027] Specifically, based on the LSTM neural network, the ideal voltage prediction model is constructed with the battery pack SOC, ambient temperature, and charge and discharge current as inputs and the ideal voltage of the battery pack as output. The ideal voltage at the sampling time under the ideal state is predicted according to the battery pack SOC, ambient temperature, and charge and discharge current obtained at the current sampling time of the battery pack to be tested. The ideal voltage is the voltage at the sampling time when the consistency and battery performance of the battery pack are good. According to the time series, the specific charge and discharge process is sampled and predicted in sequence, and the voltage change curve under the ideal state in the specific charge and discharge process can be obtained; at the same time, the voltage change curve of each battery cell is obtained by monitoring each battery cell through the BMS, and the class correlation coefficient and Euclidean distance between the voltage change curve of each battery cell and the voltage change curve under the ideal state are analyzed, so that the deviation degree of the voltage change of each battery cell can be intuitively obtained, and then the battery cell is predicted whether there is a fault and the cause of the fault, so as to locate the faulty battery cell, and the faulty battery cell is further analyzed through the size of the class correlation coefficient, and its fault type is predicted, so as to achieve accurate prediction of the fault type.

[0028] This embodiment constructs a model to predict the ideal voltage of the battery pack according to the battery pack charging and discharging parameters, generates an ideal voltage curve, obtains the voltage change curve of the battery cell during the charging and discharging process, calculates the class correlation coefficient between the two curves, locates the faulty battery cell according to the class correlation coefficient, and predicts the fault type. The present invention predicts battery faults by analyzing the class correlation of the voltage change curves of the battery pack and the battery cell during the charging and discharging process, and can accurately locate and predict faults in the early stage of the fault, thereby improving the safety and stability of the battery pack.

[0029] In some embodiments of the present invention, Figure 2 As shown, Figure 2 The present invention provides Figure 1 The flowchart of an embodiment of step S101 in the embodiment includes: S201, taking the charge and discharge parameters at each sampling time during the charge and discharge process of the battery pack as input and the ideal voltage of the battery pack at the sampling time as output, constructing an initial neural network model, wherein the charge and discharge parameters include the battery pack SOC value, ambient temperature and charge and discharge current; S202, obtaining the voltage value, battery pack SOC value, charge and discharge current and ambient temperature of a battery pack with good performance at each sampling time during the charge and discharge process; S203, iteratively training the initial neural network model according to the battery pack SOC value, charge and discharge current, ambient temperature and voltage value at each sampling time to generate a first model; S204. Based on the AdamW optimization algorithm, the first model is optimized to generate a battery pack neural network model.

[0030] It should be noted that the SOC of a battery pack describes the remaining state of charge of the battery pack. During the charging process, as the SOC increases, the chemical reaction inside the battery gradually tends to saturation, and the voltage will rise accordingly. During the discharge process, as the SOC decreases, the battery releases energy and the voltage gradually decreases. The ambient temperature has a significant impact on the performance of the battery. At high temperatures, the chemical reaction rate inside the battery is accelerated, and the charge transfer is smoother, which leads to an increase in the battery voltage. On the contrary, at low temperatures, the chemical reaction rate inside the battery slows down, charge transfer is blocked, and the battery voltage decreases. The size of the charge and discharge current directly affects the average voltage of the battery. During the charging process, a larger charge current will cause more heat and voltage drop inside the battery, thereby reducing the average voltage of the battery. Similarly, during the discharge process, a larger discharge current will also cause the average voltage of the battery to drop faster. In each time period, the battery's SOC, ambient temperature, and charge and discharge current jointly affect the average voltage of the battery.

[0031] Specifically, an initial neural network model is constructed, the input layer contains three neurons, which are the battery pack SOC value, ambient temperature and charge and discharge current, and the output layer contains one neuron, which outputs the ideal voltage of the battery pack; normal battery packs with good battery performance and consistency are screened out, and constant current charge and discharge operations are performed on them at a specific ambient temperature; based on the BMS system, the SOC, charge and discharge current, ambient temperature and battery voltage of the battery pack during the charge and discharge process are monitored; during the charge and discharge process, sampling is performed according to a preset sampling time, and after all data are cleaned and normalized, they are divided into a training set and a test set; the initial neural network model is iteratively trained through the training set, and a suitable loss function is selected, such as mean square error as the loss function of the regression problem, but not limited to the loss function; the neural network model is optimized using the AdamW optimization algorithm to generate a final neural network model, which can obtain the ideal voltage of the battery pack in this time period according to the battery pack SOC value, charge and discharge current and ambient temperature at the sampling time during the charge and discharge process.

[0032] This embodiment constructs a neural network model to predict the ideal voltage of the battery pack according to the SOC, ambient temperature and charge and discharge current of the battery pack, and uses the AdamW optimization algorithm to optimize the neural network model, thereby improving the accuracy of the model prediction. At the same time, the trained model is used to predict the ideal voltage, which greatly improves the efficiency of the model prediction and improves the accuracy and efficiency of battery fault prediction.

[0033] In some embodiments of the present invention, the calculation formula of the AdamW optimization algorithm is: , in, is the neural network weight independent variable at the current moment, is the neural network weight independent variable at the previous moment, is a decimal constant, is the learning rate, is the adaptive adjustment parameter of the learning rate, is the cumulative gradient, is the weight decay parameter.

[0034] Specifically, in the AdamW optimization algorithm, is a decimal constant used to prevent the denominator from being zero and ensure the stability of the value. is the cumulative gradient, i.e. the first-order moment estimate after variation correction, which represents the moving average of the gradient and helps to accelerate convergence. It is the deviation-corrected second-order moment estimate, which represents the moving average of the square of the gradient and is used to adjust the learning rate to achieve the effect of adaptive learning rate. The weight decay parameter is used to control the strength of weight decay, which helps prevent the model from overfitting. The AdamW optimization algorithm improves the convergence speed and generalization ability of the model by improving the processing method of weight decay.

[0035] In some embodiments of the present invention, Figure 3 As shown, Figure 3 The present invention provides Figure 1 The flowchart of an embodiment of step S102 is a flowchart of a method of performing the following steps: the timing parameter set includes a battery pack SOC value, an ambient temperature, a charge and discharge current, and a sampling time; the timing parameters in the charging and discharging timing parameter set are sequentially input into the neural network model to obtain a first curve, including: S301, inputting the battery pack SOC value, ambient temperature, and charge and discharge current in the timing parameters into the neural network model to obtain the ideal voltage value at each sampling time point; S302: Generate a first curve according to the ideal voltage at each sampling time point and its corresponding sampling time.

[0036] Specifically, the battery pack to be tested is charged and discharged at a specific ambient temperature at a constant current. Based on the BMS system, the SOC value of the battery pack is sampled at a preset sampling time. Based on the neural network model, the battery pack SOC value, ambient temperature and charge and discharge current at each sampling time are input to obtain the ideal voltage value at the corresponding sampling time. According to each sampling time and its corresponding ideal voltage value, the ideal voltage change curve of the battery pack during the charge and discharge process is obtained.

[0037] This embodiment predicts the ideal voltage value of the battery pack to be tested through a neural network model and generates an ideal voltage curve, thereby providing a data basis for battery fault prediction and improving the accuracy of fault prediction.

[0038] In some embodiments of the present invention, the calculation formula of the class correlation coefficient of the first curve and the second curve is: , in, is the class correlation coefficient between the ideal voltage change curve of the battery pack and the voltage change curve of the battery cell, The ideal voltage change curve of the battery pack is The ideal voltage value for each sampling time is The voltage variation curve of the battery cell The voltage value at each sampling time, is the total number of sampling times.

[0039] Specifically, the class correlation coefficient algorithm ICC is used to calculate the similarity of two data sets, and its formula is: , , , in, is the class correlation coefficient of the two-column data set, The first column of the data set data, The second column of the data set data, N means both the first column and the second column have N data. is the average value of all data in the first and second columns of the data set. is the variance of all data sets in the first column and the second column; According to the formula of the above-mentioned class correlation coefficient algorithm ICC, the correlation coefficient calculation formula between the battery cell voltage change curve that changes with time and the ideal voltage curve of the battery pack is finally obtained, which is used to analyze the deviation between the battery cell voltage and the ideal voltage of the battery pack at a certain moment of the battery pack to be tested. This class correlation coefficient is a value between [0,1], which is used to quantify the correlation between the battery cell voltage and the ideal voltage of the battery pack. The closer this value is to 1, the more ideal the performance of the battery cell is, and the higher the consistency with the ideal voltage of the battery pack, indicating that the performance of the battery cell is better. The closer this value is to 0, the more the voltage change of the battery cell deviates from the ideal voltage of the battery pack, and the higher the possibility of fault.

[0040] This embodiment uses the ICC algorithm to evaluate the correlation between the battery cell voltage change and the ideal voltage of the battery pack, providing data basis for the fault prediction of the battery cell.

[0041] In some embodiments of the present invention, Figure 4 As shown, Figure 4 The present invention provides Figure 1 The flowchart of an embodiment of step S105 in the embodiment includes: S401, when the class correlation coefficient is within the first preset fault threshold range, it is determined that the battery has no inconsistent fault; S402: When the class correlation coefficient is within the second preset fault threshold range, it is determined that the battery fault is a first-level inconsistency fault; S403: When the class correlation coefficient is within a third preset fault threshold range, determine that the battery fault is a level 2 inconsistency fault.

[0042] Specifically, when the class correlation coefficient is in the first preset fault threshold interval, that is, the correlation coefficient is greater than or equal to 0.75 and less than or equal to 1, it means that the voltage change of the battery cell is highly consistent with the ideal voltage change of the battery pack, and it is determined that the battery has no inconsistent fault; when the class correlation coefficient is in the second preset fault threshold interval, that is, the correlation coefficient is greater than or equal to 0.4 and less than 0.75, it means that the voltage change of the battery cell is inconsistent with the ideal voltage change of the battery pack, and the battery cell is determined to be a first-level inconsistent fault, that is, the battery cell has an inconsistent fault; when the class correlation coefficient is in the second preset fault threshold interval, that is, the correlation coefficient is greater than or equal to 0 and less than 0.4, it means that the deviation between the voltage change of the battery cell and the ideal voltage change of the battery pack is large, and the battery cell is determined to be a second-level inconsistent fault, that is, the battery cell has a serious fault.

[0043] This embodiment accurately locates the faulty battery cell, predicts the fault type and fault degree, and improves the accuracy of fault prediction based on the quasi-correlation coefficient between the voltage variation curve of the battery cell and the ideal voltage curve of the battery pack obtained by analysis and calculation.

[0044] In some embodiments of the present invention, it also includes the identification of sudden internal short circuit faults, such as Figure 5 As shown, Figure 5 A schematic flow chart of an embodiment of a sudden internal short circuit fault provided by the present invention includes: S501, determining the voltage change rate and temperature change rate of the battery cell within a preset time period during the charging and discharging process; S502, based on the preset voltage change rate alarm threshold and temperature change rate alarm threshold, alarm the battery cells according to the voltage change rate and temperature change rate of the battery cells, and count the alarm duration; S503: When the alarm duration exceeds a preset alarm duration threshold, the battery cell is diagnosed as a sudden internal short circuit fault.

[0045] Specifically, the BMS monitors the voltage and temperature changes of the battery cells during the charge and discharge process, calculates the voltage change rate and temperature change rate within a preset time period, sets a voltage change rate alarm threshold, a temperature change rate alarm threshold and an alarm duration threshold, and when the voltage change rate exceeds the voltage change rate alarm threshold and the temperature change rate exceeds the temperature change rate alarm threshold, the system triggers an alarm, and when the alarm duration exceeds the alarm duration threshold, the battery cell is diagnosed as a sudden internal short circuit fault.

[0046] Furthermore, in order to further diagnose the faulty battery cell, after determining that the battery cell has a primary inconsistency fault or a secondary inconsistency fault, the cause of the fault is further analyzed. During the discharge process, the voltage drops rapidly in the early stage, and the voltage change rate is large. After a period of discharge, the voltage enters a slowly changing stage, that is, a plateau period, and the voltage change rate is small. In the later stage of discharge, the voltage gradually decreases. Therefore, first, the battery cell voltage change curve during the discharge process is divided into an early stage, a plateau period, and a late stage according to the above-mentioned voltage change interval combined with the voltage change rate; then, the Euclidean distance between the first curve and the second curve in the early stage, the plateau period, and the late stage is calculated respectively, and the first curve is quantified by the Euclidean distance. The straight-line distance between the first curve and the second curve measures the degree of deviation between the battery cell voltage change curve and the ideal voltage in the corresponding interval; finally, based on the fault simulation experiment, the relationship between the fault cause and the Euclidean distance change characteristics is set to diagnose the cause of the inconsistency fault. Among them, the connection fault is manifested as a large and stable Euclidean distance value in the early, plateau and late stages; the aging fault is manifested as an insignificant Euclidean distance in the early and plateau stages, that is, the Euclidean distance value is small, and at the end of discharge, the Euclidean distance gradually increases; the initial SOC inconsistency is manifested as a large Euclidean distance in the early stage, a gradually decreasing Euclidean distance in the plateau stage, and a gradually increasing Euclidean distance in the late stage.

[0047] This embodiment predicts the sudden internal short circuit fault of the battery cell by analyzing the voltage change rate and temperature change rate of the battery cell.

[0048] In order to better implement the battery fault prediction method in the embodiment of the present invention, based on the battery fault prediction method, correspondingly, Figure 6 As shown, the embodiment of the present invention further provides a battery fault prediction device 600 including: A model building module 601 is used to build a neural network model by taking the charging and discharging parameters of the battery pack as input and the ideal voltage of the battery pack as output; A data acquisition module 602 is used to acquire a charge and discharge timing parameter set of a battery pack to be tested and a second curve of a battery cell in the battery pack to be tested, wherein the second curve is a voltage variation curve of the battery cell; An ideal curve generating module 603, used for sequentially inputting the timing parameters in the charging and discharging timing parameter set into the neural network model to obtain a first curve, wherein the first curve is an ideal voltage variation curve of the battery pack; A class correlation coefficient calculation module 604, used to calculate the class correlation coefficient between the first curve and the second curve; The fault location and prediction module 605 is used to locate the faulty battery cell and predict the fault type according to the class correlation coefficient and the preset fault threshold.

[0049] The battery fault prediction device 600 provided in the above embodiment can implement the technical solution described in the above battery fault prediction method embodiment. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above battery fault prediction method embodiment, which will not be repeated here.

[0050] In the embodiment of the present invention, the battery fault prediction device can be an independent server, or a server network or server cluster composed of servers. For example, the battery fault prediction device described in the embodiment of the present invention includes but is not limited to a computer, a network host, a single network server, a plurality of network server sets or a cloud server composed of multiple servers. The cloud server is composed of a large number of computers or network servers based on cloud computing.

[0051] like Figure 7 As shown, the present invention also provides a battery fault prediction electronic device 700. The battery fault prediction electronic device 700 includes a processor 701, a memory 702 and a display 703. Figure 7 Only some components of the battery failure prediction electronic device 700 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0052] In some embodiments, the processor 701 may be a central processing unit (CPU), a microprocessor or other data processing chip, and is used to run program codes or process data stored in the memory 702, such as the battery fault prediction method of the present invention.

[0053] In some embodiments of the present invention, the processor 701 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, the processor 701 may be local or remote. In some embodiments, the processor 701 may be implemented in a cloud platform. In one embodiment, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-cloud, etc., or any combination thereof.

[0054] In some embodiments, the memory 702 may be an internal storage unit of the battery fault prediction electronic device 700, such as a hard disk or memory of the battery fault prediction electronic device 700. In other embodiments, the memory 702 may also be an external storage device of the battery fault prediction electronic device 700, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the battery fault prediction electronic device 700.

[0055] Furthermore, the memory 702 may include both an internal storage unit and an external storage device of the battery failure prediction electronic device 700. The memory 702 is used to store application software installed in the battery failure prediction electronic device 700 and various data.

[0056] In some embodiments, the display 703 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 703 is used to display information of the battery failure prediction electronic device 700 and to display a visual user interface. The components 901-903 of the battery failure prediction electronic device 700 communicate with each other via a system bus.

[0057] In some embodiments of the present invention, when the processor 701 executes the battery fault prediction program in the memory 702, the following steps may be implemented: The charging and discharging parameters of the battery pack are used as input, and the ideal voltage of the battery pack is used as output to construct a neural network model; Acquire a charge and discharge timing parameter set of a battery pack to be tested and a second curve of a battery cell in the battery pack to be tested, wherein the second curve is a voltage variation curve of the battery cell; Inputting the timing parameters in the charging and discharging timing parameter set into the neural network model in sequence to obtain a first curve, wherein the first curve is an ideal voltage change curve of the battery pack; Calculating the class correlation coefficient between the first curve and the second curve according to the voltage values ​​at each sampling time on the first curve and the second curve; According to the class correlation coefficient and the preset fault threshold, the faulty battery cell is located and the fault type is predicted.

[0058] It should be understood that: when the processor 701 executes the battery fault prediction program in the memory 702, in addition to the above functions, other functions may also be implemented. For details, please refer to the description of the corresponding method embodiment above.

[0059] Furthermore, the embodiment of the present invention does not specifically limit the type of the battery fault prediction electronic device 700 mentioned. The battery fault prediction electronic device 700 may be a portable battery fault prediction electronic device such as a mobile phone, a tablet computer, a personal digital assistant (PDA), a wearable device, a laptop computer, etc. Exemplary embodiments of portable battery fault prediction electronic devices include but are not limited to portable battery fault prediction electronic devices equipped with IOS, Android, Microsoft or other operating systems. The above-mentioned portable battery fault prediction electronic device may also be other portable battery fault prediction electronic devices. It should also be understood that in some other embodiments of the present invention, the battery fault prediction electronic device 700 may not be a portable battery fault prediction electronic device, but a desktop computer with a touch-sensitive surface (such as a touch panel).

[0060] In this embodiment, the battery fault prediction electronic device is applied to electric ships and electric vehicles including a battery management system BMS, etc., to perform fault prediction on the lithium battery system of the electric ship or electric vehicle. When fault prediction is needed, the electronic device is connected to the battery management system BMS of the lithium battery system, and data in the battery management system BMS is obtained through the electronic device. The electronic device analyzes and processes the obtained data, outputs a model, further analyzes the results of the model output, obtains the final battery fault prediction result, and displays the prediction result through the electronic device.

[0061] Accordingly, an embodiment of the present invention also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, the steps or functions of the battery fault prediction method provided in the above-mentioned method embodiments can be implemented.

[0062] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the computer program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0063] The battery fault prediction method, device, equipment and storage device provided by the present invention are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for technical personnel in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A battery failure prediction method, characterized in that: include: The charging and discharging parameters of the battery pack are used as input, and the ideal voltage of the battery pack is used as output to construct a neural network model; Acquire a charge and discharge timing parameter set of the battery pack to be tested and a second curve of a battery cell in the battery pack to be tested, wherein the second curve is a voltage variation curve of the battery cell; Inputting the timing parameters in the charging and discharging timing parameter set into the neural network model in sequence to obtain a first curve, wherein the first curve is an ideal voltage change curve of the battery pack; calculating a class correlation coefficient between the first curve and the second curve; According to the class correlation coefficient and the preset fault threshold, the faulty battery cell is located and the fault type is predicted.

2. The battery failure prediction method according to claim 1, characterized in that: The charging and discharging parameters of the battery pack are used as input, and the ideal voltage of the battery pack is used as output to construct a neural network model, including: An initial neural network model is constructed by taking the charge and discharge parameters at each sampling time during the charge and discharge process of the battery pack as input and the ideal voltage of the battery pack at the sampling time as output, wherein the charge and discharge parameters include the battery pack SOC value, ambient temperature and charge and discharge current; Obtain the voltage value, battery pack SOC value, charge and discharge current and ambient temperature of a battery pack with good performance at each sampling time during the charge and discharge process; Iteratively training the initial neural network model according to the battery pack SOC value, charge and discharge current, ambient temperature and voltage value at each sampling time to generate a first model; Based on the AdamW optimization algorithm, the first model is optimized to generate a battery pack neural network model.

3. The battery failure prediction method according to claim 2, characterized in that: The calculation formula of the AdamW optimization algorithm is: ， in, is the neural network weight independent variable at the current moment, is the neural network weight independent variable at the previous moment, is the learning rate, is the adaptive adjustment parameter of the learning rate, is the cumulative gradient, is the weight decay parameter.

4. The battery failure prediction method according to claim 1, characterized in that: The timing parameters include the battery pack SOC value, ambient temperature, charge and discharge current and sampling time. The timing parameters in the charge and discharge timing parameter set are sequentially input into the neural network model to obtain a first curve, including: Inputting the battery pack SOC value, ambient temperature, and charge and discharge current in the timing parameters into the neural network model to obtain the ideal voltage value at each sampling time point; A first curve is generated according to the ideal voltage at each sampling time point and its corresponding sampling time.

5. The battery failure prediction method according to claim 4, characterized in that: The calculation formula of the class correlation coefficient between the first curve and the second curve is: ， in, is the class correlation coefficient between the ideal voltage change curve of the battery pack and the voltage change curve of the battery cell, The ideal voltage change curve of the battery pack is The ideal voltage value for each sampling time is The voltage variation curve of the battery cell The voltage value at each sampling time, is the total number of sampling times.

6. The battery failure prediction method according to claim 5, characterized in that: According to the class correlation coefficient and the preset fault threshold, predict the fault type, including: When the class correlation coefficient is within a first preset fault threshold interval, it is determined that the battery has no inconsistent fault; When the class correlation coefficient is within a second preset fault threshold range, determining that the battery fault is a first-level inconsistency fault; When the class correlation coefficient is within a third preset fault threshold interval, it is determined that the battery fault is a secondary inconsistency fault.

7. The battery failure prediction method according to claim 1, characterized in that: Also includes: Determine the voltage change rate and temperature change rate of the battery cell within a preset time period during the charging and discharging process; Based on a preset voltage change rate alarm threshold and a temperature change rate alarm threshold, an alarm is issued to the battery cell according to the voltage change rate and the temperature change rate of the battery cell, and the alarm duration is counted; When the alarm duration exceeds a preset alarm duration threshold, the battery cell is diagnosed as a sudden internal short circuit fault.

8. A battery failure prediction device, characterized in that: include: A model building module, which is used to take the charging and discharging parameters of the battery pack as input and the ideal voltage of the battery pack as output to build a neural network model; A data acquisition module, used to acquire a charge and discharge timing parameter set of the battery pack to be tested and a second curve of a battery cell in the battery pack to be tested, wherein the second curve is a voltage variation curve of the battery cell; An ideal curve generating module, used for sequentially inputting the timing parameters in the charging and discharging timing parameter set into the neural network model to obtain a first curve, wherein the first curve is an ideal voltage variation curve of the battery pack; A class correlation coefficient calculation module, used to calculate the class correlation coefficient between the first curve and the second curve; The fault location and prediction module is used to locate the faulty battery cell and predict the fault type according to the class correlation coefficient and the preset fault threshold.

9. A battery failure prediction electronic device, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the steps of the battery fault prediction method according to any one of claims 1 to 7 are implemented.

10. A storage medium, characterized in that: The storage medium stores computer program instructions, and when the computer program instructions are executed by a computer, the computer is enabled to execute the battery failure prediction method according to any one of claims 1 to 7.

Images