Battery thermal runaway intelligent grading early warning method

Abstract

The application relates to a battery thermal runaway intelligent grading early warning method, and relates to the technical field of battery thermal runaway early warning, and comprises the following steps: loading pre-control parameters and environmental temperature of a target battery cell, and constructing a battery cell dynamic image; obtaining a first battery cell temperature extreme value through a battery cell temperature extreme value prediction model; when the first battery cell temperature extreme value is smaller than a thermal runaway temperature threshold value, loading abnormal pressure timing information and cycle times of the target battery cell, and constructing a battery cell static image; when the abnormal pressure timing information is not empty or / and the cycle times are greater than a cycle times threshold value, searching a same-model battery cell sample set; statistically calculating a proportion of a thermal runaway sample in the same-model battery cell sample set, and setting the proportion as a target battery cell thermal runaway risk coefficient; and executing a grading early warning strategy. The application solves the problem that traditional battery thermal runaway management relies on temperature collection of a battery pack temperature sensor, early warning is only performed when the temperature reaches a preset value, and thermal runaway is difficult to accurately predict, early warning is lagged, and there is a safety hazard.

Description

TECHNICAL FIELD

[0001] The application relates to the field of battery thermal runaway early warning, in particular to a battery thermal runaway intelligent grading early warning method. BACKGROUND

[0002] With the improvement of the requirements for safe operation of batteries in new energy storage, electric vehicles and other fields, accurate early warning of battery thermal runaway has become a key technical requirement to ensure the stability of equipment and avoid safety accidents.

[0003] At present, the traditional battery thermal runaway management mode does not fully cope with the influence of complex chemical reactions, physical processes inside the battery and external environmental temperature fluctuations, and cannot break through the limitation of relying only on battery pack temperature sensors to collect temperature and achieving preset values to warn, which not only makes it difficult to accurately predict thermal runaway, but also increases the safety risk in the use of batteries due to the lag of early warning. SUMMARY

[0004] The application provides a battery thermal runaway intelligent grading early warning method, which improves the status of the traditional battery thermal runaway management relying on battery pack temperature sensors to collect temperature, and only executing early warning when the temperature reaches the preset value, which leads to the difficulty in accurately predicting thermal runaway, the lag of early warning and the existence of safety hazards.

[0005] The application embodiment discloses the following technical scheme:

[0006] The application embodiment provides a battery thermal runaway intelligent grading early warning method, which comprises the following steps:

[0007] Load the pre-control parameters and the environmental temperature of the target battery cell, and construct a battery cell dynamic image;

[0008] Process the battery cell dynamic image through a battery cell temperature extreme value prediction model to obtain a first battery cell temperature extreme value, wherein the battery cell temperature extreme value prediction model is obtained by machine learning training of a battery cell dynamic historical image set and a battery cell temperature extreme value detection value set, wherein the number of abnormal pressure times is zero, and the number of cycles is less than or equal to a cycle threshold value;

[0009] When the first battery cell temperature extreme value is less than a thermal runaway temperature threshold value, load the abnormal pressure timing information and the number of cycles of the target battery cell, and construct a battery cell static image;

[0010] When the abnormal pressure timing information is not empty, or / and the number of cycles is greater than the cycle threshold value, search a same type battery cell sample set meeting the battery cell static image and the battery cell dynamic image;

[0011] Statistically, the proportion of thermal runaway samples in the same type battery cell sample set is set as a target battery cell thermal runaway risk coefficient;

[0012] Based on the target battery cell thermal runaway risk coefficient, a corresponding level coefficient grading early warning strategy is executed.

[0013] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0014] The present application proposes a battery thermal runaway intelligent grading early warning method, which realizes accurate prediction and differentiated control of battery thermal runaway risk through the cooperative operation of constructing dynamic and static double images of battery cells, training battery cell temperature extreme value prediction model, quantifying thermal runaway risk coefficient, and executing grading early warning strategy. First, load the pre-control parameters and environmental temperature of the target battery cell to construct the dynamic image reflecting the real-time running state of the battery cell; then process the dynamic image of the battery cell through the temperature extreme value prediction model trained by the data of the battery cell without pressure damage and low cycle loss to obtain the first battery cell temperature extreme value; if the first battery cell temperature extreme value is less than the thermal runaway threshold, load the abnormal pressure timing information and cycle number of the battery cell to construct the static image of the battery cell, when the battery cell has abnormal pressure record or cycle number exceeds the threshold, retrieve the sample set of the same type battery cell and calculate the proportion of thermal runaway samples, and set it as the thermal runaway risk coefficient; finally, based on the pre-defined risk threshold, send prompt information, generate maintenance prompt and trigger prohibited operation warning for different risk levels of low, medium and high; if the first battery cell temperature extreme value exceeds the thermal runaway threshold, directly set the thermal runaway risk coefficient to 1 and start the highest level warning.

[0015] The technical solutions of the present application solve the problems of excessive attention to the overall state of the battery pack and neglect of fine analysis of a single battery cell in traditional thermal runaway management, and provide technical support for safe operation and efficient operation and maintenance of the battery system. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0017] Figure 1 A flowchart of a battery thermal runaway intelligent grading early warning method provided in the embodiments of the present application is shown in the figure.

[0018] Figure 2 A flowchart of a battery thermal runaway intelligent grading early warning method provided in the embodiments of the present application is shown in the figure. DETAILED DESCRIPTION

[0019] The application provides a battery thermal runaway intelligent grading early warning method, which is used for solving the technical problems that the battery thermal runaway management in the prior art depends on the temperature collection of the battery pack temperature sensor, and the early warning is only performed when the temperature reaches a preset value, so that the thermal runaway is difficult to accurately predict and there are safety hazards.

[0020] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0021] In the description of the application, the terms "first", "second" are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0022] In the description of the application, the term "for example" is used to indicate "as an example, illustration or description". Any embodiment described as "for example" in the application is not necessarily interpreted as more preferred or more advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the application. In the following description, details are listed for the purpose of explanation. It should be understood that those skilled in the art can realize the application without using these specific details. In other examples, well-known structures and processes will not be described in detail to avoid unnecessary details making the description of the application obscure. Therefore, the application is not intended to be limited to the shown embodiments, but is consistent with the broadest scope of principles and characteristics disclosed in the application.

[0023] Embodiments, as shown in the accompanying drawings Figure 1 The application provides a battery thermal runaway intelligent grading early warning method, which comprises the following steps:

[0024] S110: load the pre-control parameters of the target battery cell and the environmental temperature, and construct a battery cell dynamic image;

[0025] In the embodiments of the present application, in order to break the limitation of traditional battery pack temperature monitoring, the real-time operation characteristics of the battery cell are analyzed to provide data basis for subsequent temperature extreme value prediction. The pre-control parameters and environmental temperature of the target battery cell need to be loaded first, and then the dynamic image of the battery cell is constructed to realize the fine characterization of the current running state of the battery cell.

[0026] Specifically, first, the pre-control parameter category of the target battery cell is determined. The pre-control parameter includes the rated voltage, rated capacity, and internal resistance reference value of the battery cell calibrated at the factory, and is the basis for reflecting the inherent properties of the battery cell and judging whether the operation deviates from the normal range.

[0027] At the same time, the real-time environmental temperature data of the area where the battery cell is located is collected from the environmental temperature sensor arranged separately for the target battery cell in the battery pack. The fluctuation of the environmental temperature directly affects the internal chemical reaction rate and heat dissipation efficiency of the battery cell, and is a key external factor leading to the temperature change of the battery cell.

[0028] Further, after obtaining the pre-control parameters and environmental temperature data, the two types of data need to be integrated and processed synchronously. In the integration process, the time stamp consistency of the data needs to be ensured, that is, each set of pre-control parameters and corresponding environmental temperature data corresponds to the same monitoring time, so as to avoid the inconsistency between the dynamic image of the battery cell and the actual running state of the battery cell due to time deviation.

[0029] At the same time, the collected data is verified for effectiveness, such as checking whether the pre-control parameters are within the reasonable error range calibrated at the factory, and whether there are jump values in the environmental temperature data caused by sensor abnormalities. If data abnormalities are found, the data reacquisition mechanism is triggered immediately to obtain accurate data from the standby sensor, so as to ensure the data source reliability of the battery cell image construction.

[0030] Further, the pre-control parameters and real-time environmental temperature data of the target battery cell after time stamp synchronization verification and effectiveness screening are associated and bound to form a set of data that can fully reflect the inherent performance reference and external environmental influence of the battery cell at a specific time. Based on the data set, the dynamic image of the battery cell is constructed to ensure that the image can accurately reflect the current running basic state and environmental adaptation of the target battery cell.

[0031] For example, for a certain type of ternary lithium battery cell, the rated capacity in the pre-control parameter is 200 Ah, and the internal resistance reference value is 80 mΩ. The environmental temperature of the battery cell at a certain time is 25°C. After integrating the above data, the dynamic image of the battery cell can initially present the real-time state that the ternary lithium battery cell is currently running at the rated capacity of 200 Ah and the internal resistance of 80 mΩ as the reference performance in the 25°C environment. If the subsequent environmental temperature rises to 35°C, the updated dynamic image of the battery cell will also reflect the change in environmental temperature.

[0032] The constructed battery cell dynamic image is not a static data set, but is updated in real time with the monitoring period. Every interval of a preset monitoring time (for example, 1 minute), the latest pre-control parameters and real-time ambient temperature of the target battery cell are automatically reloaded, the battery cell dynamic image is iteratively updated to ensure that the image is always consistent with the current running state of the battery cell, and accurate dynamic data support is provided for subsequent accurate calculation of the first battery cell temperature extreme value by the battery cell temperature extreme value prediction model.

[0033] S120: processing the battery cell dynamic image by a battery cell temperature extreme value prediction model to obtain a first battery cell temperature extreme value, wherein the battery cell temperature extreme value prediction model is obtained by machine learning training of a battery cell dynamic historical image set and a battery cell temperature extreme value detection value set with zero abnormal pressure times and a cycle number less than or equal to a cycle number threshold;

[0034] In the embodiments of the present application, in order to accurately obtain the temperature extreme value that the target battery cell may reach in the future and provide a reliable basis for subsequent risk judgment, the battery cell temperature extreme value prediction model obtained by specific training is used to process the battery cell dynamic image to realize the forward-looking evaluation of the thermal state of the battery cell.

[0035] Specifically, first, the preset battery cell model is obtained by reading the target battery cell factory identification and the battery cell model information stored in the battery BMS system.

[0036] Further, the obtained preset battery cell model is input into the threshold calibration table, and the pre-defined pressure threshold and cycle number threshold corresponding to the model battery cell are accurately located through the preset "battery cell model-threshold" association mapping relationship in the table, so as to clearly define the standard for subsequent screening of battery cell dynamic historical images and judging abnormal pressure state.

[0037] Further, the battery cell data with zero abnormal pressure times and a cycle number less than or equal to a cycle number threshold is selected. Wherein, the number of times that the abnormal pressure meets the condition that the pressure monitoring value is greater than or equal to the pressure threshold is zero, which ensures that the selected battery cell is not damaged by pressure, and the cycle number is controlled within the threshold, which ensures that the battery cell is not over-aged, so as to exclude interference factors and ensure that the data can reflect the temperature characteristics of the battery cell under normal operation.

[0038] Further, the battery cell dynamic historical images meeting the above standard are obtained according to the preset battery cell model to form a battery cell dynamic historical image set; at the same time, the corresponding initial battery cell temperature extreme value detection value set is collected with the same preset battery cell model and dynamic historical image as the constraint, the centralized trend analysis is performed on the set, the first battery cell temperature extreme value detection value representing the stable temperature performance of the battery cell is extracted, and the battery cell temperature extreme value detection value set is integrated.

[0039] Further, the machine learning is used to train the set of dynamic history portraits of the battery cell and the set of temperature extreme value detection values of the battery cell, to construct a battery cell temperature extreme value prediction model. During the training process, the parameters in the dynamic history portrait are taken as the input features, and the corresponding temperature extreme value detection value is taken as the output label. The model parameters are repeatedly iterated and optimized to ensure that the model can accurately predict the temperature extreme value according to the dynamic state of the battery cell.

[0040] Finally, the constructed dynamic portrait of the battery cell is input into the trained battery cell temperature extreme value prediction model. The model analyzes the pre-control parameters, ambient temperature and other information in the portrait, calculates and outputs the first battery cell temperature extreme value. The extreme value can reflect the future thermal state change trend of the battery cell in advance, and provide data support for subsequent judgment of whether to trigger further risk assessment.

[0041] The step S120 in the method provided by the embodiment of the present application comprises:

[0042] obtaining a preset battery cell model;

[0043] inputting the preset battery cell model into a threshold calibration table, and matching a predefined pressure threshold and a cycle number threshold;

[0044] under the constraint of the preset battery cell model, collecting a first battery cell dynamic history portrait with zero abnormal pressure times and a cycle number less than or equal to the cycle number threshold, wherein the abnormal pressure represents that the number of times that the pressure monitoring value is greater than or equal to the pressure threshold is zero;

[0045] under the constraint of the preset battery cell model and the first battery cell dynamic history portrait, collecting a set of initial battery cell temperature extreme value detection values with zero abnormal pressure times and a cycle number less than or equal to the cycle number threshold, performing centralized trend analysis, and obtaining a first battery cell temperature extreme value detection value;

[0046] adding the first battery cell dynamic history portrait into the set of battery cell dynamic history portraits, and adding the first battery cell temperature extreme value detection value into the set of battery cell temperature extreme value detection values.

[0047] In the embodiment of the present application, in order to ensure that the training data of the battery cell temperature extreme value prediction model can accurately reflect the temperature characteristics of the battery cell under normal operation, high-quality training data needs to be selected through explicit constraint conditions, to construct an accurate set of battery cell dynamic history portraits and a set of battery cell temperature extreme value detection values, so as to improve the accuracy of the model in predicting the target battery cell temperature extreme value, and provide reliable data support for subsequent thermal runaway risk judgment.

[0048] Specifically, first, the factory specification parameters of the target battery cell, the battery cell model record stored in the battery management system (BMS), or the battery cell model information directly collected from the label on the battery pack should be read to obtain a preset battery cell model.

[0049] Further, the obtained preset battery cell model is input into a threshold calibration table, and through stored data pre-associated with the preset battery cell model in the table, a pre-defined pressure threshold and a cycle number threshold corresponding to the battery cell model are accurately matched.

[0050] In the method provided by the embodiments of the present application, the construction step of the threshold calibration table comprises:

[0051] The initial pressure threshold is configured as zero, and the cycle number threshold is j, where j represents a positive integer, the initial value of j is equal to 1, and the maximum value of j is equal to the rated cycle number;

[0052] With the preset battery cell model and the preset battery cell dynamic image as constraints, a first battery cell temperature extreme detection value set with zero abnormal pressure number and cycle number less than or equal to the cycle number threshold is collected, centralized trend analysis is performed, and a first fitting value of the battery cell temperature extreme is obtained;

[0053] With the preset battery cell model and the preset battery cell dynamic image as constraints, a second battery cell temperature extreme detection value set with zero abnormal pressure number and cycle number equal to j+1 cycle number is collected, centralized trend analysis is performed, and a second fitting value of the battery cell temperature extreme is obtained;

[0054] When a first temperature extreme deviation between the second fitting value of the battery cell temperature extreme and the first fitting value of the battery cell temperature extreme is greater than or equal to a temperature extreme deviation threshold, a pressure threshold is configured based on the cycle number threshold, the cycle number threshold, the pressure threshold and the preset battery cell model are stored in association, and are added to the threshold calibration table;

[0055] When the first temperature extreme deviation between the second fitting value of the battery cell temperature extreme and the first fitting value of the battery cell temperature extreme is less than the temperature extreme deviation threshold, j is incremented by one, and the cycle is executed.

[0056] In the embodiments of the present application, in order to match exclusive and accurate pressure thresholds and cycle number thresholds for different battery cell models, a threshold calibration table needs to be constructed through cycle testing and deviation judgment, so as to ensure that each preset battery cell model has a corresponding judgment standard that can accurately define the state of no pressure damage and no over-aging, and to provide a basis for subsequent screening of training data.

[0057] Specifically, first, the initial pressure threshold is configured as zero, and the initial value of the cycle number threshold j is 1, and the maximum value of j is set as the rated cycle number of the battery cell.

[0058] The initial pressure threshold is set to zero, which aims to test from the most stringent pressure judgment standard, and gradually find the critical value that can distinguish the influence of pressure on the temperature of the battery cell; the cycle number threshold starts from 1, which aims to track the influence of cycle number on temperature from the early stage of the life cycle of the battery cell to ensure that the full cycle range of normal use of the battery cell is covered, and the maximum value of j is set to the rated cycle number, which is to avoid exceeding the designed service life of the battery cell to cause the data to lose reference significance.

[0059] After completing the initial parameter configuration, data is collected with the preset battery model and preset battery dynamic image as constraints. The preset battery dynamic image includes pre-control parameters and environmental temperature data within a controllable range, and this constraint can ensure that the collected battery temperature extreme value data is only affected by the cycle number, excluding the interference of other external environments or parameter abnormalities; the zero abnormal pressure frequency further ensures that the battery is not damaged by pressure, and the temperature extreme value collected at this time can truly reflect the correlation between cycle number and temperature.

[0060] Further, based on the above constraints, a first battery temperature extreme value detection value set is collected with the abnormal pressure frequency being zero and the cycle number being less than or equal to the cycle number threshold (i.e., not exceeding the current set j value, j initial is 1 and not exceeding the rated cycle number of the battery cell), and centralized trend analysis (i.e., calculating the median) is performed on the first battery temperature extreme value detection value set to obtain a first fitted value of the battery temperature extreme value, so as to eliminate accidental errors of single temperature detection and accurately reflect the stable temperature extreme value characteristics of the battery cell within the cycle number range.

[0061] For example, for a battery cell of a preset model of "ternary lithium-18650", the cycle number threshold j = 700, 50 battery cells of this model, with the abnormal pressure frequency being zero and the cycle number being 1-700, are collected, 25 temperature extreme values of each battery cell are collected, 1250 temperature data are obtained, the median is 41°C after sorting, and the first fitted value of the battery temperature extreme value is obtained.

[0062] Further, also based on the above constraints, a second battery temperature extreme value detection value set is collected with the abnormal pressure frequency being zero and the cycle number being equal to j+1 cycle number, and centralized trend analysis (also calculating the median) is performed on the second battery temperature extreme value detection value set to obtain a second fitted value of the battery temperature extreme value.

[0063] For example, continuing the above "ternary lithium-18650" battery cell case, the cycle number threshold j+1 = 701, 45 battery cells of this model, with the abnormal pressure frequency being zero and the cycle number being exactly 701, are collected, 25 temperature extreme values of each battery cell are collected, 1125 temperature data are obtained, the median is 44°C after sorting, and the second fitted value of the battery temperature extreme value is obtained.

[0064] Further, a difference between the second cell temperature extreme fitting value and the first cell temperature extreme fitting value is calculated to obtain a first temperature extreme deviation, and the deviation is compared with a predefined temperature extreme deviation threshold.

[0065] When the first temperature extreme deviation is greater than or equal to the temperature extreme deviation threshold, it indicates that the cell temperature extreme has a significant change after the cycle number is increased from ≤j to j+1, and the current cycle number threshold j can effectively define the boundary of the cell not being over-aged, and at this time, a corresponding pressure threshold needs to be further configured based on the cycle number threshold.

[0066] In the method provided by the embodiments of the present application, the configuration of the pressure threshold based on the cycle number threshold comprises:

[0067] loading a pressure consistency deviation, wherein the pressure consistency deviation is a predefined pressure fault tolerance deviation;

[0068] summing the pressure consistency deviation and the initial pressure threshold to obtain a feature pressure;

[0069] collecting a third cell temperature extreme detection value set in which the number of abnormal pressure receiving times is not zero, the pressure of abnormal pressure receiving is less than the feature pressure, and the cycle number is less than or equal to the cycle number threshold, performing centralized trend analysis, and obtaining a third cell temperature extreme fitting value, with the preset cell model and the preset cell dynamic portrait as constraints;

[0070] When a second temperature extreme deviation between the third cell temperature extreme fitting value and the first cell temperature extreme fitting value is greater than or equal to a temperature extreme deviation threshold, the initial pressure threshold is set as the pressure threshold;

[0071] Otherwise, when the second temperature extreme deviation between the third cell temperature extreme fitting value and the first cell temperature extreme fitting value is less than the temperature extreme deviation threshold, the initial pressure threshold is updated using the feature pressure, and the cycle is performed.

[0072] In the embodiments of the present application, in order to determine the critical value that can accurately distinguish the influence of pressure on temperature for different cell models, the final pressure threshold needs to be gradually optimized and determined through the cooperation calculation of the pressure consistency deviation and the initial pressure threshold, data collection and deviation comparison, so as to ensure that the pressure threshold can cover a reasonable pressure fault tolerance range and accurately identify the pressure situation that has a significant impact on the cell temperature.

[0073] Specifically, first, a pressure consistency deviation is loaded. The pressure consistency deviation is a pressure fault-tolerant deviation predefined in combination with the production process of the battery cell, the material characteristics, and the actual use scenario. For example, for a certain type of square lithium iron phosphate battery, considering the slight pressure fluctuation in the battery cell assembly process and the normal deformation in use, the pressure consistency deviation is set to 5 kPa. This deviation value needs to ensure that within the premise of not affecting the normal temperature performance of the battery cell, it allows a certain range of pressure fluctuations to avoid misjudging slight and harmless pressure changes as abnormal pressure.

[0074] Further, after loading the pressure consistency deviation, it is summed with the initial pressure threshold to obtain the characteristic pressure. The initial pressure threshold is configured as zero at the beginning of the threshold calibration table construction, and at this time the characteristic pressure is the pressure consistency deviation. If the initial pressure threshold is updated subsequently, the characteristic pressure is adjusted synchronously with the change of the initial pressure threshold.

[0075] Further, a third battery cell temperature extreme value detection value set is collected with the preset battery cell model and the preset battery cell dynamic image as constraints. Similarly, the preset battery cell dynamic image needs to include the pre-control parameters and the environmental temperature within the controllable range to exclude the interference of abnormal pre-control parameters and environmental temperature fluctuations on the battery cell temperature; the number of abnormal pressure is not zero and the abnormal pressure is less than the characteristic pressure; the cycle number is less than or equal to the cycle number threshold.

[0076] At the same time, centralized trend analysis (also the median) is performed on the collected third battery cell temperature extreme value detection value set to obtain a third fitting value of the battery cell temperature extreme value.

[0077] Further, the second temperature extreme value deviation between the third fitting value of the battery cell temperature extreme value and the first fitting value of the battery cell temperature extreme value is compared to determine the size relationship between the deviation and the temperature extreme value deviation threshold.

[0078] The second temperature extreme value deviation is the difference between the third fitting value of the battery cell temperature extreme value and the first fitting value of the battery cell temperature extreme value, which reflects the influence of slight pressure changes on the battery cell temperature.

[0079] Specifically, if the second temperature extreme value deviation is greater than or equal to the temperature extreme value deviation threshold (for example, 3°C), it indicates that the pressure range corresponding to the current initial pressure threshold has a significant impact on the battery cell temperature, and further expanding the pressure range may lead to abnormal temperature fluctuations. Therefore, the initial pressure threshold is set as the final pressure threshold.

[0080] Conversely, if the second temperature extreme deviation is less than the temperature extreme deviation threshold, it indicates that the pressure range corresponding to the current initial pressure threshold is still within the fault-tolerant range that the battery can withstand, and slight pressure changes have no significant impact on the temperature, so the initial pressure threshold needs to be updated using the characteristic pressure, and the steps of "calculating a new characteristic pressure-collecting new data-analyzing the deviation" are performed again until the initial pressure threshold is found that can make the second temperature extreme deviation reach or exceed the temperature extreme deviation threshold.

[0081] For example, for a battery of a preset model "ternary lithium-21700", the cycle threshold has been determined as 800 times, the initial pressure threshold is 0 kPa, and the pressure consistency deviation is 5 kPa. First, the characteristic pressure is calculated as 0+5=5 kPa.

[0082] Further, with the model battery, controllable pre-control parameters (rated voltage 3.7V, rated capacity 5.0Ah) and environmental temperature (25℃±2℃) as constraints, 40 battery samples with abnormal pressure times not equal to zero, pressure less than 5 kPa and cycle times ≤800 times are collected, 30 temperature extremes are collected for each sample, and centralized trend analysis is performed to obtain the third fitting value of the battery temperature extreme 43℃.

[0083] Further, the previously obtained first fitting value of the battery temperature extreme is 41℃, the second temperature extreme deviation is 2℃, which is less than the temperature extreme deviation threshold of 3℃, so the initial pressure threshold is updated to 5 kPa. The characteristic pressure is calculated again as 5+5=10 kPa, and the same condition battery samples with pressure less than 10 kPa are collected, and centralized trend analysis is performed to obtain the third fitting value of the battery temperature extreme 45℃, which is 4℃ away from the first fitting value of the battery temperature extreme, which is greater than the temperature extreme deviation threshold. At this time, 5 kPa is set as the final pressure threshold.

[0084] Finally, the pressure threshold determined by the above step-by-step optimization can not only avoid misjudging slight pressure fluctuations as abnormal, but also accurately identify pressure conditions that have a significant impact on the temperature of the battery, ensuring that subsequent screening of the dynamic historical portrait of the battery can accurately exclude battery samples with pressure damage, and providing reliable training data for the battery temperature extreme prediction model.

[0085] Further, after determining the final cycle threshold and pressure threshold, the two thresholds need to be associated and bound with the corresponding preset battery model to form a complete corresponding relationship of "battery model-cycle threshold-pressure threshold".

[0086] At the same time, the first temperature extreme deviation between the second fitting value of the battery temperature extreme and the first fitting value of the battery temperature extreme needs to be determined according to the previous results.

[0087] Specifically, if the first temperature extreme deviation is greater than or equal to the temperature extreme deviation threshold value, the threshold value calibration process of the current type of battery cell is completed.

[0088] Conversely, if the first temperature extreme deviation is less than the temperature extreme deviation threshold value, it indicates that the current cycle threshold value j has not yet reached the boundary that can cause a significant change in the temperature of the battery cell, and the value of j needs to be increased by 1 (i.e., from the initial 1 to 2, and then from 2 to 3, etc.), and the entire process of "collecting data with the new j as the cycle threshold value - calculating the first and second fitting values - comparing the first temperature extreme deviation" is re-executed until the cycle threshold value that can make the first temperature extreme deviation meet the threshold value requirement is found, and after configuring the pressure threshold value based on the cycle threshold value, the complete correlation is added to the threshold value calibration table to ensure that each pre-set battery cell type can find its exclusive determination threshold value in the threshold value calibration table.

[0089] Further, when the pressure threshold value and the cycle threshold value corresponding to the pre-set battery cell type are stored in the threshold value calibration table, the pre-set battery cell type can be input into the threshold value calibration table, and the pre-defined pressure threshold value and cycle threshold value exclusive to the type of battery cell can be automatically matched according to the pre-established "battery cell type - threshold value" correlation in the table.

[0090] Further, after matching the corresponding threshold value, the battery cell samples consistent with the pre-set battery cell type are screened out while ensuring that the samples meet the conditions of zero abnormal pressure times and cycle times less than or equal to the cycle threshold value.

[0091] Among them, the abnormal pressure needs to be determined by the number of times the pressure monitoring value is greater than or equal to the matched pressure threshold value, so as to exclude the battery cells that have been damaged by pressure or over-aging, and the dynamic historical data of these qualified battery cells is collected to form the first battery cell dynamic historical portrait, which needs to completely cover the key dynamic information such as the pre-control parameters and the environmental temperature of the battery cell under normal operating conditions.

[0092] Further, the pre-set battery cell type and the first battery cell dynamic historical portrait are used as dual constraints to further collect the initial battery cell temperature extreme detection value set.

[0093] During the collection process, the operating conditions consistent with the first battery cell dynamic historical portrait need to be maintained to ensure that the abnormal pressure times are zero and the cycle times do not exceed the pressure threshold value, and by collecting the battery cell temperature extreme values under the same working conditions multiple times, the accidental error of single data is reduced. The collected initial battery cell temperature extreme detection value set is subjected to centralized trend analysis (also the median is calculated), and the first battery cell temperature extreme detection value representing the stable temperature characteristics of this type of battery cell is extracted.

[0094] Further, the collected first battery cell dynamic history images are added one by one into the battery cell dynamic history image set, and the corresponding first battery cell temperature extreme value detection value is added into the battery cell temperature extreme value detection value set, so as to ensure that the entries of the two types of data sets correspond one by one, and form a model training sample pair with regular structure and accurate data.

[0095] Further, the model is trained by combining machine learning, that is, a battery cell temperature extreme value prediction model is constructed based on a gradient boosting decision tree framework, so as to accurately capture the complex correlation between the battery cell dynamic parameters and the temperature extreme value.

[0096] Specifically, the model core parameters are set as follows: the learning rate is set to 0.05, which is used to control the step size of parameter update in each iteration, so as to avoid the model from falling into local optimum due to too fast update; the number of leaf nodes of a single tree is limited to 31, so as to prevent overfitting caused by too complex tree structure; the maximum depth of the tree is set to 6, which further constrains the model complexity and balances the fitting ability and the generalization ability; the minimum sample number of the leaf node is 20, which ensures that each leaf node has sufficient data support and improves the prediction stability.

[0097] At the same time, the L1 regularization coefficient is set to 0.1 and the L2 regularization coefficient is set to 0.2, so as to reduce the sensitivity of the model to noise data through regularization; the objective function is set to regression type, so as to adapt to the prediction requirement of continuous numerical values such as temperature extreme value.

[0098] During the training process, first, data preprocessing is performed, that is, abnormal samples with temperature exceeding-20℃-60℃ and cycle number being negative are removed, and missing values are filled with the mean value of adjacent time stamp features, so as to ensure the purity and integrity of the training data set and avoid interference of abnormal data or missing data on the model learning rule.

[0099] Further, the data set is divided into a training set and a validation set in a ratio of 7:3. The training set is used for learning and fitting of model parameters, and the validation set is used for real-time monitoring of the performance change in the model training process, so as to discover overfitting or underfitting problems in time.

[0100] Further, during the model training, a 5-fold cross-validation strategy is adopted, the root mean square error (RMSE) is used as the loss function, the parameters are iteratively optimized by the gradient descent algorithm, the training is stopped when the RMSE of the validation set does not decrease for 5 consecutive rounds, so as to avoid overfitting of the model, and to ensure that the model has good generalization ability and can accurately predict the temperature extreme value of unknown battery cell data.

[0101] Further, after completing the basic training, an independent test set (20% of the total data) is used to evaluate the actual prediction performance of the model, and the test set RMSE≤1.5℃ is required. If it does not meet the requirements, adjust the parameters (for example, increase the regularization coefficient) and repeat the training until the model meets the requirements to obtain the final battery cell temperature extreme value prediction model that can stably and accurately predict the battery cell temperature extreme value.

[0102] Further, after obtaining the battery cell temperature extreme value prediction model, the target battery cell dynamic image constructed previously is input into the battery cell temperature extreme value prediction model. The model combines the pre-control parameters in the image with the real-time environmental temperature, processes the data through the built-in algorithm, and finally outputs the first battery cell temperature extreme value that can reflect the future possible temperature of the target battery cell, providing key temperature basis for subsequent judgment of whether the battery cell has a thermal runaway risk.

[0103] For example, for a certain type of lithium iron phosphate battery cell, the battery cell dynamic image contains a rated capacity of 50 Ah, an internal resistance reference value of 75 mΩ, and a real-time environmental temperature of 28℃. After inputting the battery cell dynamic image into the trained battery cell temperature extreme value prediction model, the model calculates the first battery cell temperature extreme value as 42℃. This value can be directly used for comparison with the thermal runaway temperature threshold to determine whether further loading of the battery cell static image analysis risk is needed.

[0104] Through the above steps, a data foundation is laid for subsequent scenario-based thermal runaway risk assessment strategies, avoiding the lag of traditional real-time temperature monitoring, and identifying potential battery cell thermal runaway hazards in advance through forward-looking temperature extreme value prediction, providing a guarantee for the accuracy and timeliness of battery thermal runaway early warning.

[0105] S130: When the first battery cell temperature extreme value is less than the thermal runaway temperature threshold, load the abnormal pressure timing information and cycle number of the target battery cell to construct a battery cell static image.

[0106] In the embodiments of the present application, in order to break through the limitations of traditional temperature monitoring, the battery cell pressure history and use loss information are supplemented to improve the description of the battery cell state. The abnormal pressure timing information and cycle number of the target battery cell are loaded to construct a battery cell static image, so as to realize fine evaluation of the long-term state of the battery cell and provide comprehensive data support for subsequent risk coefficient calculation.

[0107] Specifically, first, from the pressure sensor deployed on the target battery cell in the battery pack, the pressure monitoring value and the corresponding pressure monitoring timestamp of the battery cell are collected. The pressure sensor needs to monitor the pressure change of the battery cell in real time to ensure that each pressure data can be accurately associated with a specific time.

[0108] Further, the collected pressure monitoring value is compared with a predefined pressure threshold value. When the pressure monitoring value is greater than or equal to the pressure threshold value, it indicates that the battery cell is under pressure beyond the normal range, and the pressure monitoring value and the corresponding pressure monitoring timestamp are added to the abnormal pressure timing information.

[0109] Conversely, if the pressure monitoring value is less than the pressure threshold value, no recording is performed to ensure that the abnormal pressure timing information only contains pressure data that has a potential impact on the state of the battery cell, and to avoid invalid information interfering with subsequent analysis.

[0110] At the same time, the cycle number of the target battery cell is extracted from the battery BMS system. The battery BMS system records the charging and discharging cycles of the battery cell in real time, and the extracted cycle number accurately reflects the current use and wear of the battery cell, providing a basis for determining whether the battery cell has performance degradation due to long-term use, thereby increasing the risk of thermal runaway.

[0111] After obtaining the abnormal pressure timing information and the cycle number, the two types of information are integrated to construct a static image of the battery cell. The image clearly presents the past abnormal pressure records of the target battery cell (including the pressure value and time of each abnormal pressure) and the current cycle number, forming a static portrayal of the long-term state of the battery cell. This complements the previous dynamic image of the battery cell that reflects the real-time state, and together lays the foundation for subsequent retrieval of the same type of battery cell sample set and calculation of the thermal runaway risk coefficient.

[0112] The method provided in the embodiments of the present application comprises the following steps S130:

[0113] Collecting a pressure monitoring value and a pressure monitoring timestamp from a pressure sensor deployed on the target battery cell by the battery pack;

[0114] When the pressure monitoring value is greater than or equal to the pressure threshold value, the pressure monitoring value and the pressure monitoring timestamp are added to the abnormal pressure timing information;

[0115] Extracting the cycle number of the target battery cell from the battery BMS system.

[0116] In the embodiments of the present application, in order to further obtain the long-term damage risk and use and wear information of the target battery cell when the first battery cell temperature extreme value does not reach the thermal runaway threshold, the pressure data needs to be collected, the abnormal pressure records need to be screened, and the cycle number needs to be extracted to provide key data for constructing a static image of the battery cell.

[0117] Specifically, first, the pressure monitoring value and the corresponding pressure monitoring timestamp of the target battery cell are collected from the pressure sensor in which the battery pack is deployed. The pressure sensor needs to be deployed one-to-one with the target battery cell to ensure that the collected pressure data accurately reflect the pressure condition of the battery cell during use, rather than the pressure state of other battery cells or the entire battery pack.

[0118] At the same time, the pressure monitoring timestamp needs to be accurate to the second level to record the time node of each pressure change completely, providing a time dimension reference for subsequent tracing of the abnormal pressure occurrence period and analyzing the impact of pressure on the long-term performance of the battery cell.

[0119] Further, after obtaining the pressure monitoring value and the timestamp, the pressure monitoring value is compared with the predefined pressure threshold. If the pressure monitoring value is greater than or equal to the pressure threshold, it indicates that the pressure on the battery cell has exceeded the safe range, which may cause internal structure damage and increase the risk of thermal runaway. In this case, the pressure monitoring value and the corresponding pressure monitoring timestamp are added to the abnormal pressure timing information to form an abnormal pressure record arranged in chronological order.

[0120] On the contrary, if the pressure monitoring value is less than the pressure threshold, it is determined to be normal pressure, and no record is made. In this way, the abnormal pressure timing information only retains key data that has potential risks to the battery cell state, avoiding invalid data occupying storage resources or interfering with subsequent analysis.

[0121] At the same time, the cycle number of the target battery cell is extracted from the battery BMS system. The battery BMS system tracks and records the charging and discharging cycle process of each battery cell in real time. Each complete charging and discharging cycle (discharging from full charge to discharge cutoff voltage, and then charging to charge cutoff voltage) is counted as one cycle. The extracted cycle number needs to be bound to the unique identifier of the target battery cell to ensure that it is not confused with the cycle number of other battery cells.

[0122] The cycle number directly reflects the degree of use and wear of the battery cell. The more the cycle number, the more obvious the decay of the active material inside the battery cell, and the higher the risk of thermal runaway. Therefore, the cycle number is a key indicator for evaluating the aging state of the battery cell.

[0123] In actual application scenarios, the synchronization of pressure data collection and cycle number extraction needs to be ensured. For example, when the pressure monitoring value of a target battery cell reaches the pressure threshold at 14:30 and is recorded in the abnormal pressure timing information, the cumulative cycle number of the battery cell at 14:30 needs to be extracted from the BMS system at the same time. This associates the abnormal pressure record with the current cycle number, facilitating subsequent analysis of the impact of abnormal pressure at a specific cycle stage on the risk of thermal runaway of the battery cell.

[0124] In addition, during data collection, the validity of the data needs to be verified. If the pressure sensor fails, resulting in empty or significant jumps in the collected pressure monitoring values, a sensor failure alarm needs to be triggered and a backup pressure sensor needs to be enabled to re-collect.

[0125] At the same time, if there is data delay or error when extracting the cycle number from the BMS system, the last valid cycle number record needs to be retrieved through the system data backup module and marked as an abnormal data state. After the system recovers, the data is updated to ensure that the collected pressure data and cycle number are reliable, laying a foundation for subsequent construction of an accurate cell static image.

[0126] Finally, the abnormal pressure timing information and cycle number are integrated to associate and bind the abnormal pressure timing information and the extracted cycle number, forming a data set that can reflect the long-term pressure damage record and usage wear state of the target cell. Based on the data set, a cell static image is constructed to comprehensively present the long-term state information of the cell.

[0127] For example, for a target cell, the pressure monitoring values collected from the pressure sensor on May 10, 2025, at 10:20 are 9kPa (greater than the pressure threshold of 8kPa), and the pressure monitoring values on May 15, 2025, at 14:35 are 10kPa (greater than the pressure threshold of 8kPa). These two data are sorted by time to form abnormal pressure timing information.

[0128] At the same time, the cycle number of the cell on May 15, 2025, extracted from the BMS system is 720 times. The cell static image constructed after integrating the two can clearly present the long-term state of "the cell has two abnormal pressure records, and the current cycle number is 720 times", providing a basis for subsequent risk assessment.

[0129] In the method provided by the embodiments of the present application, when the abnormal pressure timing information is empty and the cycle number is less than or equal to the cycle number threshold, the monitoring program is returned.

[0130] In the embodiments of the present application, when the abnormal pressure timing information is empty, it means that all the pressure monitoring values collected from the pressure sensor of the target cell are less than the pre-defined pressure threshold, i.e. the cell has never been subjected to pressure exceeding the safe range during use, eliminating the possibility of thermal runaway risk caused by pressure damage.

[0131] At the same time, the cycle number being less than or equal to the cycle number threshold indicates that the current usage wear degree of the cell does not exceed the safe boundary, and the internal active material decay is within the normal range, without increasing the risk of thermal runaway due to excessive aging.

[0132] Therefore, by comprehensively determining the two conditions, it can be determined that the target battery is currently in a healthy state, without further performing the risk coefficient calculation and early warning process, and thus returning to the monitoring program, by continuously collecting real-time data such as temperature and pressure of the battery, the state change thereof is dynamically tracked, and once the subsequent abnormal pressure or the cycle number exceeds the threshold value, the thermal runaway risk evaluation process is restarted, so as to ensure the safety of the battery while avoiding unnecessary consumption of computing resources.

[0133] In the method provided by the embodiments of the present application, when the first battery temperature extreme value is greater than the thermal runaway temperature threshold value, the target battery thermal runaway risk coefficient is configured as 1.

[0134] In the embodiments of the present application, when the first battery temperature extreme value is greater than the thermal runaway temperature threshold value, it means that the temperature of the target battery has exceeded the critical range of safe operation, and the internal chemical reaction is extremely likely to enter an out-of-control state, and the thermal runaway risk has reached the highest level.

[0135] At this time, the target battery thermal runaway risk coefficient is configured as 1, which directly defines the emergency risk state of the battery by clear numerical judgment (risk coefficient 1 represents the highest risk level), avoids delaying the early warning opportunity due to the complex sample retrieval and proportion statistics process, and ensures that the highest level of early warning response can be triggered at the first time, provides a clear basis for subsequent rapid adoption of emergency measures such as shutdown and cooling, and maximally reduces the probability of thermal runaway accidents.

[0136] S140: when the abnormal pressure timing information is not empty, or / and the cycle number is greater than the cycle number threshold value, retrieve a same-type battery sample set meeting the battery static image and the battery dynamic image;

[0137] In the embodiments of the present application, in order to avoid the one-sidedness of risk judgment relying only on the data of the target battery, a same-type battery sample set meeting the battery static image and the battery dynamic image conditions needs to be retrieved to construct a comparable sample benchmark, and provide data support for subsequent calculation of the thermal runaway risk coefficient.

[0138] Specifically, first, the determination criteria of the retrieval trigger condition are determined. Among them, “abnormal pressure timing information is not empty” means that there is at least one record of a pressure monitoring value greater than or equal to the pressure threshold value in the historical data collected from the pressure sensor of the target battery, which indicates that the battery has experienced pressure exceeding the safe range and may have internal structure damage; “cycle number is greater than cycle number threshold value” means that the current use and wear of the battery has exceeded the preset safe boundary, and the internal active material decay may cause the thermal stability to decrease. As long as any one of the two conditions is met, or both conditions are met, the sample set retrieval step needs to be started.

[0139] Further, the double image constraints required for retrieval are determined. Among them, the static image of the battery cell contains the abnormal pressure timing information and the cycle number of the target battery cell, reflecting the damage and loss state accumulated during the long-term use of the battery cell; the dynamic image of the battery cell covers the current pre-control parameters and real-time environmental temperature of the target battery cell, reflecting the current running basis and external environmental influence of the battery cell. The two types of images together constitute the core screening conditions for retrieval, to ensure that the sample battery cells retrieved are highly similar to the target battery cell in terms of "long-term state" and "real-time running conditions", thereby ensuring the reference value of the subsequent risk assessment.

[0140] Further, after the constraints are clear, the retrieval of the sample set of the same type of battery cell is carried out. Specifically, first, the preset battery cell model is taken as the screening basis to preliminarily screen out the samples that are completely consistent with the target battery cell model from the battery cell sample database.

[0141] Further, secondary screening is carried out based on the static image of the battery cell. For example, if the target battery cell has 2 abnormal pressure records and the current cycle number is 750 (greater than the cycle number threshold of 700), then the same type of sample with abnormal pressure records (number not limited) or cycle number greater than 700 is screened out.

[0142] Finally, the third screening is carried out in combination with the dynamic image of the battery cell, and the samples with pre-control parameter error within ±5% and current environmental temperature difference within ±3°C of the target battery cell are matched, to exclude the interference of samples caused by too large control parameter or environmental temperature difference.

[0143] During the retrieval process, the integrity and timeliness of the sample database need to be ensured. The sample database needs to continuously include the same type of battery cell data in different use stages and different running environments, including normal running samples and thermal runaway failure samples, and periodically update the sample information, and eliminate invalid or repeated sample data.

[0144] For example, for a target battery cell with a preset model of "lithium iron phosphate-square 50Ah", the static image of the battery cell shows that there is 1 abnormal pressure record with a pressure value of 9kPa (pressure threshold is 8kPa), and the cycle number is 820 (cycle number threshold is 800), and the dynamic image of the battery cell shows that the current pre-control parameter is the rated voltage 3.2V, the rated capacity 50Ah, and the environmental temperature is 26°C.

[0145] During the retrieval, all "lithium iron phosphate-square 50Ah" type battery cell samples are first screened out, then the samples with "abnormal pressure records or cycle number > 800" are screened out from them, and finally the samples with "pre-control parameter 3.2V±5%, 50Ah±5%, environmental temperature 26°C±3°C" are matched, to finally form a sample set containing 120 same type battery cells, including 15 thermal runaway failure samples and 105 normal running samples.

[0146] In addition, the retrieved same-type battery sample set also needs to be verified for effectiveness. The verification content includes the integrity, authenticity and relevance of the sample data. If invalid samples are found, they need to be removed from the sample set in time to ensure the high quality and reliability of the sample set, and lay an accurate data foundation for subsequent statistics of the proportion of thermal runaway samples and calculation of the target battery thermal runaway risk coefficient.

[0147] S150: Statistics of the proportion of thermal runaway samples in the same-type battery sample set is set as the target battery thermal runaway risk coefficient;

[0148] In the embodiments of the present application, in order to avoid relying on subjective experience to judge the risk level of battery thermal runaway, the proportion of thermal runaway samples in the same-type battery sample set is calculated and set as the target battery thermal runaway risk coefficient, so as to realize the accurate quantification of the target battery thermal runaway risk and provide numerical basis for subsequent implementation of the grading warning strategy.

[0149] Specifically, first, the thermal runaway samples and non-thermal runaway samples in the same-type battery sample set are distinguished. The thermal runaway samples need to meet the conditions of once appearing temperature surge, smoke and other thermal runaway phenomena and having complete fault records. The non-thermal runaway samples are normal operation or compliant retirement samples without thermal runaway signs. When classifying, the historical operation log and fault report of each sample need to be checked one by one to ensure accurate classification.

[0150] Further, the calculation is performed according to the formula "thermal runaway risk coefficient = number of thermal runaway samples / total number of effective samples in the sample set". For example, a same-type battery sample set contains 120 effective samples, of which 9 are thermal runaway samples. The calculated battery thermal runaway risk coefficient is 0.075 (i.e. 9 / 120 = 7.5%). The value directly reflects the thermal runaway risk level of the target battery.

[0151] At the same time, the calculation result needs to be verified. If the sample set is too small, the reference value of the result needs to be evaluated, and if necessary, the search range needs to be expanded to supplement samples. If the thermal runaway samples are concentrated under certain conditions (such as cycle number far exceeding the cycle number threshold), the thermal runaway risk coefficient needs to be adjusted according to the actual state of the target battery to ensure that the quantification result can match the real risk situation of the target battery and provide reliable support for subsequent warning.

[0152] S160: Based on the target battery thermal runaway risk coefficient, a grading warning strategy corresponding to the coefficient is executed.

[0153] In the embodiments of the present application, in order to avoid over-reaction or under-reaction caused by adopting a unified early warning method for different risk levels of battery cells, a grading early warning strategy is needed to be executed by loading a predefined risk threshold and combining the risk coefficient of the target battery cell, so as to realize accurate control of the thermal runaway risk and balance safety guarantee and use efficiency.

[0154] Specifically, first, a predefined first risk coefficient threshold and a second risk coefficient threshold are loaded. The second risk coefficient threshold is greater than the first risk coefficient threshold, and the two thresholds are determined based on a large amount of thermal runaway failure data and safe operation experience of battery cells of the same type, so as to divide three risk intervals of low, medium and high, and provide clear boundaries for subsequent grading response.

[0155] Further, the thermal runaway risk coefficient of the target battery cell is compared with the first and second risk coefficient thresholds, and the corresponding early warning strategy is matched.

[0156] Specifically, if the thermal runaway risk coefficient of the target battery cell is less than or equal to the first risk coefficient threshold, it indicates that the current thermal runaway risk of the battery cell is low, and there is no need to suspend use, only a potential thermal runaway risk prompt information needs to be sent to remind the staff to pay attention to the subsequent state change of the battery cell.

[0157] In addition, if the thermal runaway risk coefficient of the target battery cell is greater than the first risk coefficient threshold and less than the second risk coefficient threshold, it indicates that the battery cell has certain thermal runaway hidden danger, and a battery cell maintenance prompt needs to be generated to check the potential problems through professional maintenance.

[0158] In addition, if the thermal runaway risk coefficient of the target battery cell is greater than or equal to the second risk coefficient threshold, it means that the thermal runaway risk of the battery cell has reached a high level, and an operation prohibition warning needs to be generated immediately to forcibly stop the operation of the battery system where the battery cell is located, and a sound and light alarm is sent to notify the personnel to evacuate, so as to minimize the risk of accidents.

[0159] This step binds the quantified thermal runaway risk coefficient of the target battery cell with the grading early warning strategy, so that the early warning measures are more suitable for the actual risk situation of the battery cell, thereby ensuring safety while avoiding excessive intervention to low-risk battery cells to affect normal use, and improving the flexibility and accuracy of overall battery management.

[0160] As shown in the accompanying Figure 2 The step S160 in the method provided by the embodiments of the present application includes:

[0161] loading a predefined first risk coefficient threshold and a second risk coefficient threshold, wherein the second risk coefficient threshold is greater than the first risk coefficient threshold;

[0162] when the thermal runaway risk coefficient of the target battery cell is less than or equal to the first risk coefficient threshold, sending a potential thermal runaway risk prompt information;

[0163] generating a battery cell maintenance prompt when the target battery cell thermal runaway risk coefficient is greater than the first risk coefficient threshold and less than the second risk coefficient threshold;

[0164] generating a work prohibition warning when the target battery cell thermal runaway risk coefficient is greater than or equal to the second risk coefficient threshold.

[0165] In the embodiments of the present application, in order to take differentiated measures according to different degrees of target battery cell thermal runaway risk, avoid over-intervention on low-risk battery cells or insufficient response to high-risk battery cells, and lead to safety hazards or resource waste, it is necessary to divide risk levels by loading predefined risk coefficient thresholds and match corresponding graded warning strategies to achieve precise control of battery cell thermal runaway risk.

[0166] Specifically, first, a predefined first risk coefficient threshold and a second risk coefficient threshold are loaded.

[0167] The first risk coefficient threshold and the second risk coefficient threshold are determined based on a large number of thermal runaway failure data of battery cells of the same type, long-term operation experience and safety standards, for example, combined with historical failure statistics of a certain type of ternary lithium battery, it is found that when the risk coefficient is ≤0.05 (5%), the probability of battery cell thermal runaway is extremely low; when the risk coefficient is between 0.05-0.15 (5%-15%), the battery cell has potential risks but does not affect emergency use; when the risk coefficient is ≥0.15 (15%), the risk of battery cell thermal runaway is significantly increased, therefore the first risk coefficient threshold is set to 0.05 and the second risk coefficient threshold is set to 0.15, and the second risk coefficient threshold is greater than the first risk coefficient threshold, thereby forming a clear risk level division boundary.

[0168] Further, the target battery cell thermal runaway risk coefficient is compared with the first and second risk coefficient thresholds, and then the corresponding warning strategy is executed.

[0169] Specifically, when the target battery cell thermal runaway risk coefficient is less than or equal to the first risk coefficient threshold, it means that the battery cell is currently in a low-risk state and does not need to be suspended for use or disassembled for maintenance, at this time a potential thermal runaway risk prompt information can be sent.

[0170] For example, a battery management system (BMS) pushes a text or pop-up prompt to the terminal device of the operator, the content includes "the target battery cell thermal runaway risk is low (risk coefficient 0.03), it is recommended to strengthen temperature and pressure data monitoring once every 24 hours", which not only reminds the worker to pay attention to the state of the battery cell, but also avoids excessive intervention affecting normal work.

[0171] Conversely, when the target battery cell thermal runaway risk coefficient is greater than the first risk coefficient threshold and less than the second risk coefficient threshold, it indicates that the battery cell has certain thermal runaway risks, and if not handled in time, it may lead to risk escalation. At this time, a battery cell maintenance prompt needs to be generated. The maintenance prompt needs to specify specific operation and maintenance requirements, such as marking "target battery cell thermal runaway risk medium (risk coefficient 0.12), battery appearance inspection, internal resistance test and charge-discharge performance evaluation need to be completed within 72 hours", and automatically assigning a maintenance work order to the corresponding operation and maintenance team to ensure that the hidden danger can be investigated within a controllable time and prevent the risk from further expanding.

[0172] Conversely, when the target battery cell thermal runaway risk coefficient is greater than or equal to the second risk coefficient threshold, it means that the battery cell thermal runaway risk has reached a high level and a safety accident may occur at any time. At this time, a prohibited operation warning must be generated.

[0173] The prohibited operation warning needs to have mandatory intervention, such as immediately cutting off the charge-discharge circuit of the battery module where the battery cell is located, suspending the operation of related equipment, and notifying on-site personnel to evacuate to a safe area through sound and light alarms, emergency messages, etc. At the same time, link the fire emergency device to prepare, and minimize personnel casualties and property losses caused by thermal runaway accidents.

[0174] For example, for a certain type of lithium iron phosphate square battery cell, the first risk coefficient threshold is 0.04 and the second risk coefficient threshold is 0.12. If the target battery cell thermal runaway risk coefficient is 0.03 (≤0.04), send a "suggestion to strengthen daily monitoring" prompt message; if the risk coefficient is 0.08 (0.04<0.08<0.12), generate a "48-hour maintenance" work order; if the risk coefficient is 0.15 (≥0.12), immediately trigger a prohibited operation warning, cut off the circuit and issue an alarm.

[0175] At the same time, when executing the hierarchical warning strategy, the warning information also needs to be recorded and traced. For example, associate the unique identifier of the target battery cell, the risk coefficient calculation basis, the warning trigger time and the subsequent processing results for each warning to form a complete warning file, so as to provide data support for the adjustment of the warning strategy for the same type of battery cell, ensure the scientificity and adaptability of the execution of the hierarchical warning strategy, and further continuously improve the accuracy and reliability of the battery thermal runaway warning.

[0176] The embodiments of the present application achieve the following technical effects through the specific implementation methods described above:

[0177] The application provides a battery thermal runaway intelligent grading early warning method. First, the pre-control parameters of a target battery cell and the environmental temperature are loaded, a battery cell dynamic image is constructed through data synchronization integration and effectiveness verification, and the real-time running state of the battery cell is accurately described. Then, the battery cell dynamic image is input into a temperature extreme value prediction model trained by non-pressure damage and low cycle loss battery cell data to obtain a first battery cell temperature extreme value. If the first battery cell temperature extreme value is greater than a thermal runaway temperature threshold, the thermal runaway risk coefficient is directly configured as 1. If the first battery cell temperature extreme value is less than the thermal runaway temperature threshold, a battery cell static image is constructed by loading the abnormal pressure timing information and the cycle number of the battery cell to present the long-term damage and loss state of the battery cell. When the battery cell has an abnormal pressure record or the cycle number exceeds the threshold, a sample set of the same type of battery cell matched with the battery cell dynamic image and the battery cell static image is searched, and the proportion of thermal runaway samples is counted as the thermal runaway risk coefficient of the target battery cell. Finally, a predefined first risk coefficient threshold and a second risk coefficient threshold are loaded, and potential risk prompts, maintenance prompts or operation prohibition early warnings are sent according to the risk coefficient, so that the thermal runaway risk is differentially controlled.

[0178] The method provided by the embodiments of the application solves the problems that the traditional thermal runaway management excessively focuses on the overall state of the battery pack and ignores the fine analysis of a single battery cell, and that there is a serious lag in early warning when the temperature is abnormal, realizes full-dimensional evaluation from the real-time state of the battery cell to the long-term risk, provides reliable technical support for the safe operation and efficient operation and maintenance of the battery system, effectively balances the safety guarantee and use efficiency, and reduces the probability of thermal runaway accidents.

[0179] It should be noted that the above-mentioned sequence of the embodiments of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. Moreover, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or can be advantageous.

[0180] The above only describes the preferred embodiments of the application and does not limit the application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the application should be included in the protection scope of the application.

[0181] The specification and drawings are, of course, to be regarded in an illustrative rather than a restrictive sense. It is to be understood that any such modifications, variations, combinations or equivalents that fall within the scope of the application are intended to be embraced herein.

Claims

1. A battery thermal runaway intelligent grading early warning method, characterized in that, The method comprises the following steps: loading the pre-control parameters and the ambient temperature of the target battery cell, and constructing a battery cell dynamic image; processing the battery cell dynamic image through a battery cell temperature extreme value prediction model to obtain a first battery cell temperature extreme value, wherein the battery cell temperature extreme value prediction model is obtained through machine learning using a battery cell dynamic historical image set and a battery cell temperature extreme value detection value set with zero abnormal pressure times and a cycle number less than or equal to a cycle number threshold value; wherein the battery cell temperature extreme value prediction model is obtained through machine learning using a battery cell dynamic historical image set and a battery cell temperature extreme value detection value set with zero abnormal pressure times and a cycle number less than or equal to a cycle number threshold value, comprising: obtaining a preset battery cell model; inputting the preset battery cell model into a threshold calibration table to match a predefined pressure threshold value and a cycle number threshold value; with the preset battery cell model as a constraint, collecting a first battery cell dynamic historical image with zero abnormal pressure times and a cycle number less than or equal to a cycle number threshold value, wherein the number of times of abnormal pressure, which represents a pressure monitoring value greater than or equal to the pressure threshold value, is zero; with the preset battery cell model and the first battery cell dynamic historical image as constraints, collecting an initial battery cell temperature extreme value detection value set with zero abnormal pressure times and a cycle number less than or equal to a cycle number threshold value, and performing centralized trend analysis to obtain a first battery cell temperature extreme value detection value; adding the first battery cell dynamic historical image to the battery cell dynamic historical image set and adding the first battery cell temperature extreme value detection value to the battery cell temperature extreme value detection value set; when the first battery cell temperature extreme value is less than a thermal runaway temperature threshold value, loading abnormal pressure timing information and cycle number of the target battery cell to construct a battery cell static image; when the abnormal pressure timing information is not empty or / and the cycle number is greater than the cycle number threshold value, retrieving a same-model battery cell sample set that meets the battery cell static image and the battery cell dynamic image; statistically analyzing the proportion of thermal runaway samples in the same-model battery cell sample set and setting it as a target battery cell thermal runaway risk coefficient; based on the target battery cell thermal runaway risk coefficient, performing a hierarchical early warning strategy corresponding to the coefficient.

2. The method of claim 1, wherein, It also comprises: when the first battery cell temperature extreme value is greater than the thermal runaway temperature threshold value, setting the target battery cell thermal runaway risk coefficient to 1.

3. The method of claim 1, wherein, The threshold calibration table construction step comprises: configuring an initial pressure threshold value as zero and a cycle number threshold value as j, wherein j represents a positive integer, the initial value of j is equal to 1, and the maximum value of j is equal to the rated cycle number; with the preset battery cell model and a preset battery cell dynamic image as constraints, collecting a first battery cell temperature extreme value detection value set with zero abnormal pressure times and a cycle number less than or equal to a cycle number threshold value, and performing centralized trend analysis to obtain a first fitted value of the battery cell temperature extreme value; with the preset battery cell model and a preset battery cell dynamic image as constraints, collecting a second battery cell temperature extreme value detection value set with zero abnormal pressure times and a cycle number equal to j+1 cycle numbers, and performing centralized trend analysis to obtain a second fitted value of the battery cell temperature extreme value; When the first temperature extreme value deviation between the second cell temperature extreme value fitting value and the first cell temperature extreme value fitting value is greater than or equal to a temperature extreme value deviation threshold, based on the cycle number threshold, a pressure threshold is configured, the cycle number threshold, the pressure threshold and the preset cell model are associated and stored, and the threshold calibration table is added; When the first temperature extreme value deviation between the second cell temperature extreme value fitting value and the first cell temperature extreme value fitting value is less than the temperature extreme value deviation threshold, j is incremented by one, and the loop is executed.

4. The method of claim 3, wherein, Configuring a pressure threshold based on the cycle number threshold comprises: Loading a pressure consistency deviation, wherein the pressure consistency deviation is a predefined pressure fault tolerance deviation; Summing the pressure consistency deviation and the initial pressure threshold to obtain a characteristic pressure; With the preset cell model and the preset cell dynamic image as constraints, a third cell temperature extreme value detection value set in which the number of abnormal pressure receiving is not zero and the pressure of abnormal pressure receiving is less than the characteristic pressure and the cycle number is less than or equal to the cycle number threshold is collected, centralized trend analysis is performed, and a third cell temperature extreme value fitting value is obtained; When the second temperature extreme value deviation between the third cell temperature extreme value fitting value and the first cell temperature extreme value fitting value is greater than or equal to the temperature extreme value deviation threshold, the initial pressure threshold is set as the pressure threshold; Otherwise, when the second temperature extreme value deviation between the third cell temperature extreme value fitting value and the first cell temperature extreme value fitting value is less than the temperature extreme value deviation threshold, the initial pressure threshold is updated using the characteristic pressure, and the loop is executed.

5. The method of claim 1, wherein, When the abnormal pressure receiving time sequence information is empty and the cycle number is less than or equal to the cycle number threshold, the monitoring program is returned.

6. The method of claim 1, wherein, Loading the abnormal pressure receiving time sequence information and the cycle number of the target cell, and constructing a cell static image comprises: Collecting pressure monitoring values and pressure monitoring time stamps from pressure sensors deployed on the target cell by the battery pack; When the pressure monitoring value is greater than or equal to the pressure threshold, the pressure monitoring value and the pressure monitoring time stamp are added to the abnormal pressure receiving time sequence information; The cycle number of the target cell is extracted from the battery BMS system.

7. The method of claim 1, wherein, Based on the target cell thermal runaway risk coefficient, a corresponding level coefficient grading early warning strategy is executed, comprising: Loading a predefined first risk coefficient threshold and a second risk coefficient threshold, wherein the second risk coefficient threshold is greater than the first risk coefficient threshold; When the target cell thermal runaway risk coefficient is less than or equal to the first risk coefficient threshold, a potential thermal runaway risk prompt information is sent; When the target cell thermal runaway risk coefficient is greater than the first risk coefficient threshold and less than the second risk coefficient threshold, a cell maintenance prompt is generated; When the target cell thermal runaway risk coefficient is greater than or equal to the second risk coefficient threshold, a work prohibition warning is generated.

Images