Lithium battery thermal runaway early warning system and method based on multi-source parameter monitoring

Abstract

The present invention belongs to the technical field of battery monitoring, and discloses a lithium battery thermal runaway early warning system and method based on multi-source parameter monitoring. The method includes: collecting multi-source battery parameters in real time and marking them as real-time multi-source battery parameters; processing the real-time multi-source battery parameters to extract battery characteristic parameters; performing fusion analysis on the real-time multi-source battery parameters and the extracted battery characteristic parameters to predict the thermal runaway probability of the lithium battery; judging whether to generate an evaluation instruction according to the thermal runaway probability; if an evaluation instruction is generated, evaluating the thermal runaway type of the lithium battery and initiating an active prevention and control measure. The present invention realizes high-precision and timely monitoring and early warning of lithium battery thermal runaway accidents, reduces the thermal runaway risk, effectively improves the operation safety and reliability of lithium batteries, and thus protects the life safety of new energy vehicle users.

Description

Technical Field

[0001] The present invention relates to the technical field of battery monitoring. More specifically, the present invention relates to a lithium battery thermal runaway early warning system and method based on multi-source parameter monitoring. Background Art

[0002] In recent years, with the rapid development of new energy vehicles, lithium batteries, as their core power source, have been widely used due to their advantages such as high power density and high energy density. However, lithium batteries also have certain safety hazards under high energy density, and battery thermal runaway is one of the important safety risks. Thermal runaway refers to the uncontrollable continuous increase in the internal temperature of the battery due to a series of exothermic chemical reactions, ultimately leading to the complete loss of function of the battery and even causing fire and explosion. Once thermal runaway occurs, it may trigger serious chain reactions and dangerous consequences, threatening the lives and safety of users of new energy vehicles. Traditional battery thermal runaway early warning methods often have problems such as single monitoring parameters and insufficient early warning sensitivity, and it is difficult to comprehensively reflect the occurrence mechanism of thermal runaway.

[0003] Of course, there are also lithium battery thermal runaway early warning methods with multi-source monitoring parameters. For example, the patent with the publication number CN118151017A discloses a lithium battery thermal runaway early warning method and medium. The method includes: collecting relevant parameter data of the lithium battery, where the relevant parameter data includes ambient temperature, current, surface temperature, and voltage; inputting the surface temperature and voltage into a pre-trained surface temperature and voltage prediction model to output the predicted surface temperature and voltage; inputting the predicted surface temperature and voltage into a pre-trained three-level thermal runaway early warning model to predict the internal temperature of the lithium battery, and obtaining a thermal runaway early warning of the corresponding level according to the predicted internal temperature of the lithium battery. This invention has advantages such as accurate early warning, reduced model complexity, and improved calculation speed.

[0004] However, the above technology mainly predicts the internal temperature of the lithium battery through the surface temperature and voltage to achieve thermal runaway early warning, ignoring other parameters that affect thermal runaway, such as gas emissions and electrolyte concentration, and cannot comprehensively capture the complex dynamics inside the lithium battery. And only through the internal temperature of the lithium battery, it is impossible to accurately evaluate the type of thermal runaway, and it is difficult to provide a basis for subsequent maintenance personnel to trace the root cause.

[0005] In view of this, the present invention proposes a lithium battery thermal runaway early warning system and method based on multi-source parameter monitoring to solve the above problems. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art and achieve the above object, the present invention provides the following technical solution: A lithium battery thermal runaway early warning method based on multi-source parameter monitoring, including:

[0007] Collect multi-source battery parameters in real time and mark them as real-time multi-source battery parameters;

[0008] Process the real-time multi-source battery parameters to extract battery characteristic parameters;

[0009] Perform fusion analysis on the real-time multi-source battery parameters and the extracted battery characteristic parameters to predict the thermal runaway probability of the lithium battery;

[0010] According to the thermal runaway probability, determine whether to generate an evaluation instruction; if an evaluation instruction is generated, evaluate the thermal runaway type of the lithium battery;

[0011] If an evaluation instruction is generated, initiate active prevention and control measures.

[0012] Furthermore, the battery characteristic parameters are the change rates of each parameter in the real-time multi-source battery parameters;

[0013] The method for extracting the battery characteristic parameters includes:

[0014] Obtain historical multi-source battery parameters, where the historical multi-source battery parameters are the multi-source battery parameters collected last time; subtract each parameter in the real-time multi-source battery parameters from the corresponding parameter in the historical multi-source battery parameters to obtain the change amount of each parameter; divide the change amount of each parameter by the collection interval to obtain the change rate of each parameter in the real-time multi-source battery parameters, and use it as the battery characteristic parameter.

[0015] Furthermore, the method for obtaining the collection interval includes:

[0016] Obtain the historical thermal runaway probability and the historical probability change rate; the historical thermal runaway probability is the thermal runaway probability predicted after collecting the multi-source battery parameters last time, and the historical probability change rate is the difference in the thermal runaway probability predicted after collecting the multi-source battery parameters in the previous two times, divided by the value obtained by dividing the corresponding collection interval when collecting the multi-source battery parameters last time;

[0017] Preset an initial collection interval; the time interval between the first collection of multi-source battery parameters and the second collection of multi-source battery parameters is the initial collection interval; perform weighted summation on the historical thermal runaway probability and the historical probability change rate to calculate the adjustment coefficient; multiply the adjustment coefficient by the initial collection interval to obtain the collection interval.

[0018] Furthermore, the method for predicting the thermal runaway probability of the lithium battery includes:

[0019] Use the real-time multi-source battery parameters and the battery characteristic parameters as analysis data, and input the analysis data into the trained probability prediction model to predict the thermal runaway probability of the lithium battery;

[0020] The training process of the probability prediction model includes:

[0021] Pre-collect g sets of analysis data, and set corresponding thermal runaway probabilities for each of the g sets of analysis data, where g is an integer greater than 1;

[0022] Convert the analysis data and the corresponding thermal runaway probabilities into a corresponding set of feature vectors; each set of feature vectors is used as the input of the probability prediction model. The probability prediction model outputs a set of predicted thermal runaway probabilities corresponding to each set of analysis data, and uses the actual thermal runaway probability corresponding to each set of analysis data as the prediction target. The actual thermal runaway probability is the thermal runaway probability pre-collected corresponding to the analysis data; use minimizing the sum of prediction errors of all analysis data as the training target; train the probability prediction model until the sum of prediction errors converges and then stop training; the probability prediction model is a deep neural network model.

[0023] Further, the method for determining whether to generate an evaluation instruction includes:

[0024] Preset a probability threshold, and compare the thermal runaway probability with the probability threshold;

[0025] If the thermal runaway probability is greater than or equal to the probability threshold, generate an evaluation instruction;

[0026] If the thermal runaway probability is less than the probability threshold, generate a to-be-evaluated instruction;

[0027] If a to-be-evaluated instruction is generated, obtain the historical probability change rate used each time the calculation acquisition interval is calculated, and mark it as the historical change rate; mark the currently predicted thermal runaway probability as the current probability, subtract the historical thermal runaway probability from the current probability, and then divide by the acquisition interval to obtain the current change rate; sort each historical change rate and the current change rate in the order of the corresponding calculation time, and generate a sorting table; obtain the first n values in the sorting table in reverse order, and mark them as the analysis change rates, where n is an integer greater than 1; according to the forward order of the sorting table, subtract each analysis change rate from the next analysis change rate in turn to obtain n - 1 change rate differences; preset a difference threshold, compare each change rate difference with the difference threshold respectively, and count the number of change rate differences corresponding to the values greater than the difference threshold, and mark it as the increase number; compare the increase number with and if the increase number is greater than or equal to then generate an evaluation instruction; if the increase number is less than then do not generate an evaluation instruction.

[0028] Further, the method for evaluating the thermal runaway type of the lithium battery includes:

[0029] Preset a type database, which includes a parameter set corresponding to each thermal runaway type of the lithium battery. The parameter set includes multi-source battery parameters and battery characteristic parameters, where one thermal runaway type corresponds to m parameter sets, and m is an integer greater than 1;

[0030] Calculate the similarity between each parameter in the analysis data and the corresponding parameter in each parameter set in the type database respectively; sequentially add the similarities corresponding to each parameter set as the total similarity of each parameter set; sort the total similarities from largest to smallest, and obtain the thermal runaway type corresponding to the parameter set corresponding to the largest total similarity.

[0031] Further, the multi-source battery parameters include physical data and chemical data; the physical data includes temperature, voltage, and current; the temperature is the temperature of the region where the internal physical reaction of the lithium battery cell occurs; the chemical data includes gas concentration, gas pressure, and electrolyte concentration; the gas concentration includes hydrogen concentration and carbon dioxide concentration; the gas pressure is the total pressure formed by the gas inside the lithium battery; the electrolyte concentration is the concentration of lithium salt in the electrolyte;

[0032] The active prevention and control measures include module isolation, monomer isolation, and gas isolation;

[0033] The method of module isolation is: divide the inside of the lithium battery into multiple independent modules, build heat insulation boards between each module with heat insulation materials, and when an evaluation instruction is generated, turn on the heat insulation board through an electronic switch;

[0034] The method of monomer isolation is: set isolation devices between each battery unit inside the lithium battery, and when an evaluation instruction is generated, cut off the connection between each battery unit through the isolation device;

[0035] The method of gas isolation is: set a gas valve inside the lithium battery, and when an evaluation instruction is generated, start the gas valve to discharge the hot gas inside the lithium battery.

[0036] Further, the method further includes re-evaluating the thermal runaway type of the lithium battery;

[0037] The method for re-evaluating the thermal runaway type of the lithium battery includes:

[0038] Sort the total similarities from largest to smallest, retain the first half of the total similarities in all total similarities according to the positive order, and mark the parameter sets corresponding to the retained total similarities as the analysis sets; represent the analysis data as an analysis matrix C, and calculate the covariance matrix of each analysis set; calculate the Mahalanobis distance between the analysis data and each group of analysis sets according to the covariance matrix, and mark it as the similarity distance; sort each similarity distance from smallest to largest, and obtain the thermal runaway type corresponding to the parameter set corresponding to the smallest similarity distance value.

[0039] Further, the method for calculating the covariance matrix of each analysis set includes:

[0040] Add the same data in the analysis sets corresponding to each type of thermal runaway successively, and then divide by Obtain the mean data corresponding to each type of thermal runaway, and represent it as a 1×14 matrix, denoted as B; subtract the corresponding parameter in the mean data from each parameter in each analysis set corresponding to each type of thermal runaway to obtain the standardized parameters corresponding to each analysis set; represent the standardized parameters corresponding to each analysis set corresponding to each type of thermal runaway as a matrix, and denote it as A i , A i is the matrix of the i-th analysis set,

[0041] The expression of the covariance matrix ∑i is:

[0042] In the formula, ∑ i is the covariance matrix of the i-th analysis set, A i T is the transpose of the matrix A i ;

[0043] The expression of the Mahalanobis distance is:

[0044] In the formula, E i (C) is the Mahalanobis distance between the analysis matrix C and the corresponding matrix of the i-th analysis set, is the inverse matrix of the covariance matrix of the i-th analysis set.

[0045] For the lithium battery thermal runaway early warning system based on multi-source parameter monitoring, implementing the lithium battery thermal runaway early warning method based on multi-source parameter monitoring includes:

[0046] A parameter acquisition module, which is used to acquire multi-source battery parameters in real time and mark them as real-time multi-source battery parameters;

[0047] A parameter processing module, which is used to process the real-time multi-source battery parameters and extract battery characteristic parameters;

[0048] A fusion analysis module, which is used to perform fusion analysis on the real-time multi-source battery parameters and the extracted battery characteristic parameters to predict the thermal runaway probability of the lithium battery;

[0049] A type evaluation module, which is used to judge whether to generate an evaluation instruction according to the thermal runaway probability; if an evaluation instruction is generated, then evaluate the thermal runaway type of the lithium battery;

[0050] An active prevention and control module, which is used to start active prevention and control measures if an evaluation instruction is generated.

[0051] ​Technical effects and advantages of the lithium battery thermal runaway early warning system and method based on multi-source parameter monitoring:

[0052] 1. By comprehensively collecting multi-dimensional physical and chemical parameters of lithium batteries, the characteristics of parameter change rates are proposed, and the acquisition interval of parameters is dynamically adjusted according to the characteristic parameters, which is more in line with the state change rate of lithium batteries, improving the timeliness of early warning; using deep learning technology to fully explore the implicit associations between multi-source battery parameters, accurately predicting the probability of heat loss; and then generating corresponding evaluation instructions, accurately judging and identifying the types of thermal runaway, providing a basis for subsequent maintenance personnel to trace the root cause of thermal runaway accidents, and starting multiple prevention and control measures to form a closed-loop management; realizing high-precision and timely monitoring and early warning of lithium battery thermal runaway accidents, reducing the risk of thermal runaway, effectively improving the operation safety and reliability of lithium batteries, and thus protecting the lives of new energy vehicle users.

[0053] 2. By analyzing the covariance laws between different parameters, grasping the internal linkage laws between multi-source battery parameters; and by calculating and analyzing the Mahalanobis distance between the data and each analysis set, reflecting both the parameter correlation degree and the similarity distance, enhancing the depth and rigor of type judgment; compared with the method of separately judging the similarity of each parameter, it can more comprehensively and microscopically evaluate the internal logical relationship behind multi-source information, effectively improving and standardizing the ability to identify the types of thermal runaway accidents, providing a more accurate basis for subsequent maintenance personnel to trace the root cause, and thus providing better technical protection for the operation safety of lithium batteries. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 Schematic diagram of the lithium battery thermal runaway early warning system based on multi-source parameter monitoring according to Embodiment 1 of the present invention;

[0055] Figure 2 Schematic diagram of the lithium battery thermal runaway early warning system based on multi-source parameter monitoring according to Embodiment 2 of the present invention;

[0056] Figure 3 Flowchart of the lithium battery thermal runaway early warning method based on multi-source parameter monitoring according to Embodiment 3 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0058] Embodiment 1

[0059] Please refer to Figure 1As shown in the figure, the lithium battery thermal runaway early warning system based on multi-source parameter monitoring in this embodiment includes a parameter acquisition module, a parameter processing module, a fusion analysis module, a type evaluation module, and an active prevention and control module; each module is connected by wired and / or wireless means to achieve data transmission between modules;

[0060] The parameter acquisition module is used to collect multi-source battery parameters in real time and mark them as real-time multi-source battery parameters.

[0061] The multi-source battery parameters include physical data and chemical data;

[0062] The physical data includes temperature, voltage, and current; the temperature is the temperature of the area where the internal physical reaction of the lithium battery cell occurs; the temperature is obtained by a temperature sensor built into the lithium battery; the voltage is obtained by a voltage sensor installed at the positive and negative terminals of the cell housing inside the lithium battery; the current is obtained by a current sensor installed at the connection wire of the positive and negative electrodes of the cell housing inside the lithium battery.

[0063] The chemical data includes gas concentration, gas pressure, and electrolyte concentration;

[0064] The gas concentration includes hydrogen concentration and carbon dioxide concentration; the gas pressure is the total pressure formed by the gas inside the lithium battery; the electrolyte concentration is the concentration of lithium salts (such as LiPF6, LiBF4, etc.) in the electrolyte; the gas concentration is obtained by a MEMS gas sensor installed between the lithium battery housing and the cover; the gas pressure is obtained by a pressure sensor (such as a piezoelectric sensor or a strain gauge) installed on the lithium battery housing near the gas cavity; the electrolyte concentration is obtained by an electrochemical sensor installed on the lithium battery electrode plate.

[0065] It should be noted that the increase in temperature may be an initial signal of thermal runaway, and high temperature will accelerate the internal chemical reaction of the lithium battery, leading to further temperature increase and thermal runaway; the voltage change may reflect the abnormal state of the lithium battery, such as overcharge, over-discharge, or short circuit, triggering the risk of thermal runaway; excessive current may cause the lithium battery to overheat, increasing the risk of thermal runaway; the generation of hydrogen and carbon dioxide and the electrolyte concentration are all related to the internal chemical reaction of the lithium battery. When the concentrations of hydrogen and carbon dioxide increase and the electrolyte concentration changes, it indicates that the internal chemical reaction rate of the lithium battery accelerates, which may cause the lithium battery to overheat, resulting in thermal runaway; the increase in gas pressure will cause the lithium battery housing to rupture or leak, which may trigger the thermal runaway of the lithium battery.

[0066] The parameter processing module is used to process the real-time multi-source battery parameters and extract battery characteristic parameters.

[0067] The battery characteristic parameter is the change rate of each parameter in the real-time multi-source battery parameters.

[0068] The method for extracting the battery characteristic parameters includes:

[0069] Obtain historical multi-source battery parameters, where the historical multi-source battery parameters are the multi-source battery parameters collected last time; subtract each parameter in the real-time multi-source battery parameters from the corresponding parameter in the historical multi-source battery parameters to obtain the change amount of each parameter; divide the change amount of each parameter by the collection interval to obtain the change rate of each parameter in the real-time multi-source battery parameters, and use it as the battery characteristic parameter.

[0070] The method for obtaining the collection interval includes:

[0071] Obtain the historical thermal runaway probability and the historical probability change rate; the historical thermal runaway probability is the thermal runaway probability predicted after collecting the multi-source battery parameters last time, and the historical probability change rate is the difference in the thermal runaway probability predicted after collecting the multi-source battery parameters in the previous two times, divided by the value obtained by the corresponding collection interval when collecting the multi-source battery parameters last time; for example, a total of three collections of multi-source battery parameters are performed, and the real-time multi-source battery parameters are the multi-source battery parameters collected in the third time. Subtract the thermal runaway probability predicted after collecting the multi-source battery parameters in the second time from the thermal runaway probability predicted after collecting the multi-source battery parameters in the first time to obtain the probability difference; obtain the collection interval corresponding to the second collection of multi-source battery parameters and mark it as the secondary interval, that is, the time interval between the first collection of multi-source battery parameters and the second collection of multi-source battery parameters; divide the probability difference by the secondary interval to obtain the historical probability change rate;

[0072] Preset an initial collection interval, which is preset by those skilled in the art according to the change rate of the multi-source battery parameters under the normal operating state of the lithium battery; the time interval between the first collection of multi-source battery parameters and the second collection of multi-source battery parameters is the initial collection interval; perform a weighted sum of the historical thermal runaway probability and the historical probability change rate to calculate the adjustment coefficient; multiply the adjustment coefficient by the initial collection interval to obtain the collection interval; the expression of the adjustment coefficient is: In the formula, TJ is the adjustment coefficient, LG is the historical thermal runaway probability, LB is the historical probability change rate, and ω1 and ω2 are both preset weight coefficients.

[0073] It should be noted that the specific values of the weight coefficients in the formula can be set according to the actual situation. The weight coefficients reflect the influence degree of the historical thermal runaway probability and the historical probability change rate on the collection interval. Those skilled in the art can preset the corresponding weight coefficients according to the actual influence degree of the thermal runaway probability and the historical probability change rate on the collection interval, so as to accurately obtain the collection intervals of the parameters of the lithium battery in different states.

[0074] It should be understood that the greater the historical thermal runaway probability and the historical probability change rate, the faster the deterioration rate of the internal state of the lithium battery, and the faster the change rate of the multi-source battery parameters. Therefore, the corresponding acquisition interval should be smaller to timely grasp the dynamic changes of the internal state of the lithium battery and perform more accurate thermal runaway early warning. Otherwise, the opposite is true.

[0075] The fusion analysis module is used to perform fusion analysis on real-time multi-source battery parameters and the extracted battery characteristic parameters to predict the thermal runaway probability of the lithium battery.

[0076] The method for predicting the thermal runaway probability of a lithium battery includes:

[0077] Taking the real-time multi-source battery parameters and the battery characteristic parameters as analysis data, and inputting the analysis data into the trained probability prediction model to predict the thermal runaway probability of the lithium battery.

[0078] The specific training process of the probability prediction model includes:

[0079] Pre-collect g groups of analysis data, and set corresponding thermal runaway probabilities for the g groups of analysis data, where g is an integer greater than 1; the thermal runaway probability corresponding to the analysis data is determined by those skilled in the art during the historical lithium battery thermal runaway analysis process. Collect g groups of analysis data for experiments, continuously run the lithium battery multiple times under the conditions of each group of analysis data, and detect whether thermal runaway occurs during each operation. Determine the corresponding thermal runaway probability according to the experimental results; set the corresponding thermal runaway probabilities for the g groups of analysis data in sequence;

[0080] Convert the analysis data and the corresponding thermal runaway probability into a corresponding set of feature vectors; each set of feature vectors is used as the input of the probability prediction model, and the probability prediction model outputs a set of predicted thermal runaway probabilities corresponding to each group of analysis data, and takes the actual thermal runaway probability corresponding to each group of analysis data as the prediction target. The actual thermal runaway probability is the thermal runaway probability pre-collected corresponding to the analysis data; taking the minimization of the sum of the prediction errors of all analysis data as the training target; where the calculation formula for the prediction error is η K =(β K -ε K ) 2 where η K is the prediction error, K is the group number of the feature vector corresponding to the analysis data, β K is the predicted thermal runaway probability corresponding to the Kth group of analysis data, and ε K is the actual thermal runaway probability corresponding to the Kth group of analysis data; train the probability prediction model until the sum of the prediction errors reaches convergence and then stop training.

[0081] The above probability prediction model is specifically a deep neural network model, which includes an input layer, a hidden layer, and an output layer. Each hidden layer contains multiple neurons, and there are connections between each neuron and the neurons in the next layer. These connections contain weights that determine the importance and influence of data transmission in the neural network. An activation function is applied to each neuron between the hidden layer and the output layer. The activation function introduces non-linearity, allowing the network to learn more complex patterns and features.

[0082] A type evaluation module is used to determine whether to generate an evaluation instruction based on the thermal runaway probability. If an evaluation instruction is generated, the thermal runaway type of the lithium battery is evaluated.

[0083] The method for determining whether to generate an evaluation instruction includes:

[0084] A preset probability threshold, which is preset by those skilled in the art according to the actual situation of thermal runaway of lithium batteries.

[0085] Compare the thermal runaway probability with the probability threshold.

[0086] If the thermal runaway probability is greater than or equal to the probability threshold, an evaluation instruction is generated, indicating that the probability of thermal runaway of the lithium battery is relatively high, and it is necessary to evaluate the thermal runaway type in time and initiate active prevention and control measures to avoid further temperature rise of the lithium battery causing thermal runaway.

[0087] If the thermal runaway probability is less than the probability threshold, a to-be-evaluated instruction is generated, indicating that the probability of thermal runaway of the lithium battery is relatively low, but further analysis is still required. If the change rate of the thermal runaway probability is large multiple times, it still indicates an increase in the thermal runaway risk of the lithium battery.

[0088] If a to-be-evaluated instruction is generated, obtain the historical probability change rate used each time when calculating the acquisition interval and mark it as the historical change rate. Mark the currently predicted thermal runaway probability as the current probability. Subtract the historical thermal runaway probability from the current probability and then divide by the acquisition interval to obtain the current change rate. Sort each historical change rate and the current change rate in the order of the corresponding calculation time and generate a sorting table. Obtain the first n values from the sorting table in reverse order and mark them as the analysis change rates, where n is an integer greater than 1, and the specific value of n is set by those skilled in the art according to the actual situation. According to the forward order of the sorting table, subtract each analysis change rate from the next analysis change rate in turn to obtain n - 1 change rate differences. Exemplarily, if the sorting table has 8, 7, 5, 4 in reverse order, then the change rate differences are 8 - 7 = 1, 7 - 5 = 2, 5 - 4 = 1, and there are 3 change rate differences in total. Preset a difference threshold, compare each change rate difference with the difference threshold respectively, and count the number of change rate differences corresponding to the values greater than the difference threshold and mark it as the increase quantity. Compare the increase quantity with Compare, if the increase quantity is greater than or equal to an evaluation instruction is generated; if the number of increases is less than no evaluation instruction is generated; the difference threshold is preset by those skilled in the art according to the actual situation.

[0089] A method for evaluating the thermal runaway type of a lithium battery includes:

[0090] Preset a type database, where the type database includes a parameter set corresponding to each thermal runaway type of the lithium battery. The parameter set includes multi-source battery parameters and battery characteristic parameters. One thermal runaway type corresponds to m parameter sets, and m is an integer greater than 1; the type database is constructed by those skilled in the art by collecting the multi-source battery parameters and battery characteristic parameters corresponding to each thermal runaway type multiple times when historical lithium batteries undergo thermal runaway; the thermal runaway type is, for example, thermal runaway during the normal operation of the lithium battery (thermal runaway caused by overcharging, over-discharging, short circuit, etc.), thermal runaway caused by mechanical abuse (thermal runaway caused by mechanical damage such as extrusion and puncture of the lithium battery), etc.;

[0091] Calculate the similarity between each parameter in the analysis data and the corresponding parameter in each parameter set in the type database respectively. The similarity calculation methods are, for example, Euclidean distance, cosine similarity, Manhattan distance, etc. The similarity calculation method is prior art and will not be elaborated here; add up the similarities corresponding to each parameter set in turn as the total similarity of each parameter set; sort the total similarities of each parameter set from large to small, and obtain the thermal runaway type corresponding to the parameter set corresponding to the largest total similarity.

[0092] It should be noted that the purpose of evaluating the thermal runaway type of the lithium battery is to understand the cause of the thermal runaway accident: different types of thermal runaway may stem from different incentives, and identifying the type helps to trace the root cause of the accident; optimize the battery design: according to the type analysis, it helps those skilled in the art to improve the lithium battery in terms of structural design and reduce the possibility of the recurrence of similar accidents; improve the accident handling efficiency: clarifying the thermal runaway type helps the maintenance personnel to maintain the lithium battery faster, control the scope of thermal runaway in time, and reduce economic losses and potential safety hazards.

[0093] An active prevention and control module is used to start active prevention and control measures to prevent heat spread if an evaluation instruction is generated.

[0094] The active prevention and control measures include module isolation, cell isolation, and gas isolation;

[0095] The method of module isolation is: divide the inside of the lithium battery into multiple independent modules, and build heat-insulating plates between each module with heat-resistant and heat-insulating materials. When an evaluation instruction is generated, the heat-insulating plates are opened through electronic switches (such as relays, circuit breakers, etc.) to block the spread of heat;

[0096] The method of monomer isolation is as follows: An isolation device, such as a thermal fuse or a solenoid valve, is provided between each battery cell inside the lithium battery. When an evaluation instruction is generated, the connection between each battery cell can be cut off through the isolation device to limit the heat diffusion;

[0097] The method of gas isolation is as follows: A gas valve is provided inside the lithium battery. When an evaluation instruction is generated, by activating the gas valve, the hot gas inside the lithium battery can be quickly discharged from the lithium battery to block the heat spread;

[0098] It should be noted that the purpose of activating the active prevention and control measures is to quickly block the heat conduction inside the lithium battery; compared with passive heat dissipation, active isolation can cut off the heat conduction path to the greatest extent and prevent heat spread; and a variety of isolation measures are reasonably combined to carry out superimposed prevention from multiple angles to improve the safety of the lithium battery.

[0099] In this embodiment, by comprehensively collecting multi-dimensional physical and chemical parameters of the lithium battery, the parameter change rate characteristics are proposed, and the acquisition interval of the parameters is dynamically adjusted according to the characteristic parameters, which is more in line with the state change rate of the lithium battery and improves the timeliness of early warning; the implicit associations between multi-source battery parameters are fully mined by using deep learning technology to accurately predict the probability of heat loss; and then the corresponding evaluation instructions are generated to accurately judge and identify the type of thermal runaway, providing a basis for subsequent maintenance personnel to trace the root cause of the thermal runaway accident, and activating multiple prevention and control measures to form a closed-loop management; realizing high-precision and timely monitoring and early warning of the lithium battery thermal runaway accident, reducing the risk of thermal runaway, effectively improving the operation safety and reliability of the lithium battery, and thus protecting the lives of new energy vehicle users.

[0100] Embodiment 2

[0101] Please refer to Figure 2 As shown, this embodiment further improves the design on the basis of Embodiment 1. In Embodiment 1, the type of thermal runaway is evaluated by calculating and analyzing the cosine similarity, Euclidean distance, etc. between each parameter in the data and the corresponding parameters in each parameter set in the type database respectively; however, this method does not fully consider the joint change form between different parameters; for example, when a short circuit or overcharge occurs in the lithium battery, both the temperature rise and the voltage drop will occur; only by independently analyzing the similarity of the temperature and voltage respectively and adding them up, the type of thermal runaway cannot be effectively evaluated, resulting in misjudgment of the type of thermal runaway; therefore, this embodiment provides a lithium battery thermal runaway early warning system based on multi-source parameter monitoring, which also includes a secondary evaluation module;

[0102] The secondary evaluation module is used to re-evaluate the type of thermal runaway of the lithium battery.

[0103] The method for re-evaluating the type of thermal runaway of the lithium battery includes:

[0104] Sort each sum of similarities from largest to smallest. Retain the first half of the sums of similarities in ascending order, and mark the set of parameters corresponding to the retained sums of similarities as the analysis set. Represent the analysis data as an analysis matrix C, calculate the covariance matrix of each analysis set to reflect the correlation between different parameters in the analysis data. Calculate the Mahalanobis distance between the analysis data and each group of analysis sets based on the covariance matrix, and mark it as the similarity distance. Sort each similarity distance from smallest to largest, and obtain the thermal runaway type corresponding to the set of parameters with the smallest similarity distance value.

[0105] The method for calculating the covariance matrix of each analysis set includes:

[0106] Add the same data in the analysis set corresponding to each thermal runaway type in turn, and then divide by Obtain the mean data corresponding to each thermal runaway type, represent it as a 1×14 matrix, and mark it as B. Subtract the corresponding parameter in the mean data from each parameter in each analysis set corresponding to each thermal runaway type to obtain the standardized parameter corresponding to each analysis set. For each analysis set corresponding to the standardized parameters is represented as a i , A i is the matrix of the i-th analysis set, where 14 is the number of parameter types in the standardized parameters,

[0107] The expression of the covariance matrix ∑i is:

[0108] In the formula, ∑ i is the covariance matrix of the i-th analysis set, A i T is the matrix A i transpose.

[0109] The expression of the Mahalanobis distance is:

[0110] In the formula, E i (C) is the Mahalanobis distance between the analysis matrix C and the matrix corresponding to the i-th analysis set, is the inverse matrix of the covariance matrix of the i-th analysis set.

[0111] In this embodiment, by analyzing the covariance law between different parameters, the internal linkage law between multi-source battery parameters is grasped; and by calculating the Mahalanobis distance between the analysis data and each analysis set, both the parameter correlation degree and the similarity distance are reflected, enhancing the depth and rigor of type judgment; compared with the method of separately judging the similarity of each parameter, it can more comprehensively and microscopically evaluate the internal logical relationship behind multi-source information, effectively improving and standardizing the identification ability of thermal runaway accident types, providing a more accurate basis for subsequent maintenance personnel to trace the root cause, and thus providing better technical guarantee for the operation safety of lithium batteries.

[0112] Embodiment 3

[0113] Please refer to Figure 3 As shown, for the parts not described in detail in this embodiment, refer to the descriptions in Embodiment 1 and Embodiment 2. A lithium battery thermal runaway warning method based on multi-source parameter monitoring is provided. The method includes:

[0114] Collect multi-source battery parameters in real time and mark them as real-time multi-source battery parameters;

[0115] Process the real-time multi-source battery parameters to extract battery characteristic parameters;

[0116] Perform fusion analysis on the real-time multi-source battery parameters and the extracted battery characteristic parameters to predict the thermal runaway probability of the lithium battery;

[0117] According to the thermal runaway probability, determine whether to generate an evaluation instruction; if an evaluation instruction is generated, evaluate the thermal runaway type of the lithium battery;

[0118] If an evaluation instruction is generated, activate the active prevention and control measures.

[0119] Further, the battery characteristic parameter is the change rate of each parameter in the real-time multi-source battery parameters;

[0120] The method for extracting the battery characteristic parameters includes:

[0121] Obtain historical multi-source battery parameters, where the historical multi-source battery parameters are the multi-source battery parameters collected last time; subtract each parameter in the real-time multi-source battery parameters from the corresponding parameter in the historical multi-source battery parameters to obtain the change amount of each parameter; divide the change amount of each parameter by the collection interval to obtain the change rate of each parameter in the real-time multi-source battery parameters, and use it as the battery characteristic parameter.

[0122] Further, the method for obtaining the collection interval includes:

[0123] Obtain the historical thermal runaway probability and the historical probability change rate; the historical thermal runaway probability is the thermal runaway probability predicted after the last collection of multi-source battery parameters, and the historical probability change rate is the difference in the thermal runaway probability predicted after the previous two collections of multi-source battery parameters, divided by the collection interval corresponding to the last collection of multi-source battery parameters.

[0124] Preset the initial collection interval; the time interval between the first collection of multi-source battery parameters and the second collection of multi-source battery parameters is the initial collection interval; perform a weighted sum on the historical thermal runaway probability and the historical probability change rate to calculate the adjustment coefficient; multiply the adjustment coefficient by the initial collection interval to obtain the collection interval.

[0125] Further, the method for predicting the thermal runaway probability of a lithium battery includes:

[0126] Use the real-time multi-source battery parameters and the battery characteristic parameters as analysis data, and input the analysis data into the trained probability prediction model to predict the thermal runaway probability of the lithium battery.

[0127] The training process of the probability prediction model includes:

[0128] Pre-collect g sets of analysis data, and set corresponding thermal runaway probabilities for the g sets of analysis data, where g is an integer greater than 1.

[0129] Convert the analysis data and the corresponding thermal runaway probability into a corresponding set of feature vectors; each set of feature vectors is used as the input of the probability prediction model. The probability prediction model outputs a set of predicted thermal runaway probabilities corresponding to each set of analysis data, and uses the actual thermal runaway probability corresponding to each set of analysis data as the prediction target. The actual thermal runaway probability is the thermal runaway probability pre-collected corresponding to the analysis data; use minimizing the sum of the prediction errors of all analysis data as the training target; train the probability prediction model until the sum of the prediction errors converges and then stop training; the probability prediction model is a deep neural network model.

[0130] Further, the method for determining whether to generate an evaluation instruction includes:

[0131] Preset a probability threshold, and compare the thermal runaway probability with the probability threshold.

[0132] If the thermal runaway probability is greater than or equal to the probability threshold, generate an evaluation instruction.

[0133] If the thermal runaway probability is less than the probability threshold, generate a to-be-evaluated instruction.

[0134] If a to-be-evaluated instruction is generated, obtain the historical probability change rate used each time the calculation acquisition interval is obtained, and mark it as the historical change rate; mark the thermally out-of-control probability predicted this time as the current probability, subtract the historical thermally out-of-control probability from the current probability, and then divide by the acquisition interval to obtain the current change rate; sort each historical change rate and the current change rate according to the chronological order of the corresponding calculation time, and generate a sorting table; obtain the first n values in the sorting table in reverse order and mark them as the analysis change rates, where n is an integer greater than 1; according to the forward order of the sorting table, subtract each analysis change rate from the next analysis change rate in turn to obtain n - 1 change rate differences; preset a difference threshold, compare each change rate difference with the difference threshold respectively, and count the number of change rate differences corresponding to the values greater than the difference threshold, and mark it as the increase quantity; compare the increase quantity with and if the increase quantity is greater than or equal to then generate an evaluation instruction; if the increase quantity is less than then no evaluation instruction is generated.

[0135] Further, the method for evaluating the thermally out-of-control type of the lithium battery includes:

[0136] Preset a type database, where the type database includes a parameter set corresponding to each thermally out-of-control type of the lithium battery. The parameter set includes multi-source battery parameters and battery characteristic parameters, and one thermally out-of-control type corresponds to m parameter sets, where m is an integer greater than 1;

[0137] Calculate the similarity between each parameter in the analysis data and the corresponding parameter in each parameter set in the type database respectively; add the similarities corresponding to each parameter set in turn as the similarity sum of each parameter set; sort the similarity sums of each parameter set from large to small, and obtain the thermally out-of-control type corresponding to the parameter set corresponding to the largest similarity sum.

[0138] Further, the multi-source battery parameters include physical data and chemical data; the physical data includes temperature, voltage, and current; the temperature is the temperature of the region where the internal physical reaction of the lithium battery cell occurs; the chemical data includes gas concentration, gas pressure, and electrolyte concentration; the gas concentration includes hydrogen concentration and carbon dioxide concentration; the gas pressure is the total pressure formed by the gas inside the lithium battery; the electrolyte concentration is the concentration of lithium salt in the electrolyte;

[0139] The active prevention and control measures include module isolation, monomer isolation, and gas isolation;

[0140] The method of module isolation is: divide the inside of the lithium battery into multiple independent modules, and build heat insulation boards between each module with heat insulation materials. When an evaluation instruction is generated, turn on the heat insulation board through an electronic switch;

[0141] The method for monomer isolation is as follows: An isolation device is set between each battery cell inside the lithium battery. When an evaluation instruction is generated, the connection between each battery cell is cut off through the isolation device.

[0142] The method for gas isolation is as follows: A gas valve is set inside the lithium battery. When an evaluation instruction is generated, the hot gas inside the lithium battery is discharged from the lithium battery by starting the gas valve.

[0143] Furthermore, the method further includes: re-evaluating the thermal runaway type of the lithium battery;

[0144] The method for re-evaluating the thermal runaway type of the lithium battery includes:

[0145] Sort each sum of similarities from large to small. According to the positive order, retain all the sums of similarities in the first half of the sums of similarities, and mark the parameter set corresponding to the retained sum of similarities as the analysis set; represent the analysis data as an analysis matrix C, and calculate the covariance matrix of each analysis set; calculate the Mahalanobis distance between the analysis data and each group of analysis sets according to the covariance matrix, and mark it as the similarity distance; sort each similarity distance from small to large, and obtain the thermal runaway type corresponding to the parameter set corresponding to the smallest similarity distance value.

[0146] Furthermore, the method for calculating the covariance matrix of each analysis set includes:

[0147] Add the same data in the analysis set corresponding to each thermal runaway type in sequence, and then divide by Obtain the mean data corresponding to each thermal runaway type, and represent it as a 1×14 matrix, marked as B; subtract the corresponding parameter in the mean data from each parameter in each analysis set corresponding to each thermal runaway type to obtain the standardized parameter corresponding to each analysis set; for each analysis set corresponding to each thermal runaway type, represent the corresponding standardized parameters as a matrix, and mark it as A i A i is the matrix of the i-th analysis set,

[0148] The expression of the covariance matrix ∑i is:

[0149] In the formula, ∑ i is the covariance matrix of the i-th analysis set, A i T is the transpose of matrix A i ;

[0150] The expression of the Mahalanobis distance is:

[0151] wherein, E i (C) is the Mahalanobis distance between the analysis matrix C and the corresponding matrix of the i-th analysis set, is the inverse matrix of the covariance matrix of the i-th analysis set.

[0152] Example 4

[0153] The present application also provides an electronic device. The electronic device may include one or more processors and one or more memories. Among them, computer-readable code is stored in the memory, and when the computer-readable code is run by one or more processors, it can execute the method for warning thermal runaway of lithium batteries based on multi-source parameter monitoring as described above.

[0154] The method or system according to the embodiment of the present application can also be implemented by means of the architecture of the electronic device shown in the present application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. The storage device in the electronic device, such as ROM or a hard disk, can store the method for warning thermal runaway of lithium batteries based on multi-source parameter monitoring provided by the present application. Further, the electronic device may further include a user interface. Of course, the architecture shown in the present application is only exemplary, and when implementing different devices, one or more components shown in the electronic device of the present application can be omitted according to actual needs.

[0155] Example 5

[0156] One embodiment of the present application discloses a computer-readable storage medium. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are run by a processor, the method for warning thermal runaway of lithium batteries based on multi-source parameter monitoring according to the embodiment of the present application described with reference to the above drawings can be executed. The storage medium includes but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory, etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0157] In addition, according to the embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, and the non-transitory machine-readable storage medium stores machine-readable instructions, and the machine-readable instructions can be run by a processor to execute instructions corresponding to the method steps provided by the present application, for example: the method for warning thermal runaway of lithium batteries based on multi-source parameter monitoring. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0158] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the said claims.

[0159] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should all be included within the protection scope of the present invention.

Claims

1. A lithium battery thermal runaway early warning method based on multi-source parameter monitoring, characterized in that: include: Collect multi-source battery parameters in real time and mark them as real-time multi-source battery parameters; Process real-time multi-source battery parameters and extract battery characteristic parameters; Methods for extracting battery characteristic parameters include: Obtain historical multi-source battery parameters, subtract corresponding parameters in the historical multi-source battery parameters from each parameter in the real-time multi-source battery parameters, and obtain the change amount of each parameter; divide the change amount of each parameter by the acquisition interval, obtain the change rate of each parameter in the real-time multi-source battery parameters, and use it as the battery characteristic parameter; the multi-source battery parameters include physical data and chemical data; the physical data include temperature, voltage and current; the chemical data include gas concentration, gas pressure and electrolyte concentration; Methods for obtaining the collection interval include: Obtain the historical thermal runaway probability and the historical probability change rate; preset the initial collection interval; the time interval between the first collection of multi-source battery parameters and the second collection of multi-source battery parameters is the initial collection interval; perform weighted summation of the historical thermal runaway probability and the historical probability change rate to calculate the adjustment coefficient; multiply the adjustment coefficient by the initial collection interval to obtain the collection interval; The expression of the adjustment coefficient is: ; In the formula, is the adjustment coefficient, is the historical thermal runaway probability, is the historical probability change rate, , All are preset weight coefficients; Fusion analysis of real-time multi-source battery parameters and extracted battery characteristic parameters to predict the thermal runaway probability of lithium batteries; According to the probability of thermal runaway, determine whether to generate an evaluation instruction; if an evaluation instruction is generated, evaluate the thermal runaway type of the lithium battery; If an assessment instruction is generated, active prevention and control measures will be initiated.

2. The lithium battery thermal runaway early warning method based on multi-source parameter monitoring according to claim 1, characterized in that: The battery characteristic parameter is the change rate of each parameter in the real-time multi-source battery parameter; The historical multi-source battery parameters are the multi-source battery parameters collected last time.

3. The lithium battery thermal runaway early warning method based on multi-source parameter monitoring according to claim 2 is characterized in that: The historical thermal runaway probability is the thermal runaway probability predicted after the last multi-source battery parameter collection, and the historical probability change rate is the difference in thermal runaway probabilities predicted after the last two multi-source battery parameter collections, divided by the value obtained by the corresponding collection interval when the multi-source battery parameters were collected last time.

4. The lithium battery thermal runaway early warning method based on multi-source parameter monitoring according to claim 3 is characterized in that: The method for predicting the thermal runaway probability of a lithium battery comprises: The real-time multi-source battery parameters and battery characteristic parameters are used as analysis data, and the analysis data are input into the trained probability prediction model to predict the thermal runaway probability of the lithium battery; The training process of the probabilistic prediction model includes: G groups of analysis data are collected in advance, and corresponding thermal runaway probabilities are set for the g groups of analysis data, where g is an integer greater than 1; The analysis data and the corresponding thermal runaway probability are converted into a corresponding set of feature vectors; each set of feature vectors is used as the input of a probability prediction model, and the probability prediction model takes a set of predicted thermal runaway probabilities corresponding to each set of analysis data as output, and takes the actual thermal runaway probability corresponding to each set of analysis data as a prediction target, and the actual thermal runaway probability is the pre-collected thermal runaway probability corresponding to the analysis data; minimizing the sum of prediction errors of all analysis data is used as a training target; the probability prediction model is trained until the sum of prediction errors reaches convergence and the training is stopped; the probability prediction model is a deep neural network model.

5. The lithium battery thermal runaway early warning method based on multi-source parameter monitoring according to claim 4 is characterized in that: The method for determining whether to generate an evaluation instruction comprises: Preset a probability threshold and compare the probability of thermal runaway with the probability threshold; If the probability of thermal runaway is greater than or equal to the probability threshold, an evaluation instruction is generated; If the probability of thermal runaway is less than the probability threshold, an instruction to be evaluated is generated; If an instruction to be evaluated is generated, the historical probability change rate used in each calculation of the collection interval is obtained and marked as the historical change rate; the thermal runaway probability predicted this time is marked as the current probability, the current probability is subtracted from the historical thermal runaway probability, and then divided by the collection interval to obtain the current change rate; each historical change rate and current change rate are sorted according to the order of the corresponding calculation time, and a sorting table is generated; the first n values ​​in the sorting table are obtained in reverse order and marked as analysis change rates, where n is an integer greater than 1; according to the positive order of the sorting table, each analysis change rate is subtracted from the next analysis change rate to obtain The difference of change rate is calculated; a difference threshold is preset, each difference of change rate is compared with the difference threshold, and the number of difference of change rate whose value is greater than the difference threshold is counted and marked as the number of increase; the number of increase is compared with For comparison, if the increase is greater than or equal to , an evaluation instruction is generated; if the increase is less than , no evaluation instructions are generated.

6. The lithium battery thermal runaway early warning method based on multi-source parameter monitoring according to claim 5, characterized in that: The method for evaluating the thermal runaway type of a lithium battery comprises: A preset type database includes a parameter set corresponding to each thermal runaway type of a lithium battery, the parameter set includes multi-source battery parameters and battery characteristic parameters, wherein one thermal runaway type corresponds to m parameter sets, where m is an integer greater than 1; Calculate the similarity between each parameter in the analysis data and the corresponding parameter in each parameter set in the type database respectively; add the similarities corresponding to each parameter set in turn to obtain the sum of the similarities of each parameter set; sort each similarity sum from large to small, and obtain the thermal runaway type corresponding to the parameter set corresponding to the front similarity sum.

7. The lithium battery thermal runaway early warning method based on multi-source parameter monitoring according to claim 6, characterized in that: The temperature is the temperature of the area where the physical reaction occurs inside the lithium battery cell; the gas concentration includes the hydrogen concentration and the carbon dioxide concentration; the gas pressure is the total pressure formed by the gas inside the lithium battery; the electrolyte concentration is the concentration of lithium salt in the electrolyte; The active prevention and control measures include module isolation, monomer isolation and gas isolation; The module isolation method is: the lithium battery is divided into multiple independent modules, and a heat insulation board is constructed between each module using heat insulation materials. When an evaluation instruction is generated, the heat insulation board is turned on by an electronic switch; The method of isolating the single cell is as follows: an isolation device is provided between each battery cell inside the lithium battery, and when an evaluation instruction is generated, the connection between each battery cell is cut off by the isolation device; The method of gas isolation is: a gas valve is set inside the lithium battery, and when an evaluation instruction is generated, the hot gas inside the lithium battery is discharged from the lithium battery by starting the gas valve.

8. The lithium battery thermal runaway early warning method based on multi-source parameter monitoring according to claim 7, characterized in that: The method further includes: re-evaluating the thermal runaway type of the lithium battery; Methods for re-evaluating the thermal runaway type of lithium batteries include: Sort each similarity sum from large to small, retain the similarity sums in the first half of all similarity sums in positive order, and mark the parameter set corresponding to the retained similarity sums as the analysis set; represent the analysis data as the analysis matrix C, and calculate the covariance matrix of each analysis set; calculate the Mahalanobis distance between the analysis data and each group of analysis sets based on the covariance matrix, and mark them as similarity distances; sort each similarity distance from small to large, and obtain the thermal runaway type corresponding to the parameter set corresponding to the similarity distance with the smallest value.

9. The lithium battery thermal runaway early warning method based on multi-source parameter monitoring according to claim 8, characterized in that: The method for calculating the covariance matrix for each analysis set includes: Add the same data in the analysis set corresponding to each thermal runaway type in sequence and divide by , obtain the mean data corresponding to each thermal runaway type and express it as , marked as B; each parameter in each analysis set corresponding to each thermal runaway type is subtracted from the corresponding parameter in the mean data to obtain the standardized parameter corresponding to each analysis set; each thermal runaway type is corresponding to The standardized parameters corresponding to the analysis set are expressed as a The matrix is ​​labeled , That is the matrix of the i-th analysis set, ; Covariance matrix The expression is: ; In the formula, is the covariance matrix of the ith analysis set, For the matrix The transpose of The expression of Mahalanobis distance is: ; In the formula, is the Mahalanobis distance between the analysis matrix C and the matrix corresponding to the i-th analysis set, is the inverse matrix of the covariance matrix of the ith analysis set.

10. A lithium battery thermal runaway early warning system based on multi-source parameter monitoring, implementing a lithium battery thermal runaway early warning method based on multi-source parameter monitoring as claimed in any one of claims 1 to 9, characterized in that: include: A parameter acquisition module, used for real-time acquisition of multi-source battery parameters and marking them as real-time multi-source battery parameters; A parameter processing module is used to process real-time multi-source battery parameters and extract battery characteristic parameters; Fusion analysis module, used to perform fusion analysis on real-time multi-source battery parameters and extracted battery characteristic parameters to predict the probability of thermal runaway of lithium batteries; A type assessment module is used to determine whether to generate an assessment instruction based on the probability of thermal runaway; if an assessment instruction is generated, the thermal runaway type of the lithium battery is assessed; The active prevention and control module is used to initiate active prevention and control measures if an evaluation instruction is generated.

Images