Battery thermal runaway risk assessment method and system and storage medium

Abstract

The invention provides a battery thermal runaway risk assessment method and system and a storage medium, and relates to the technical field of battery safety monitoring. By introducing inverse standard deviation normalization weight, equivalent expression of parameters with different dimensions, scales or distribution characteristics in comprehensive scoring is realized, an expected weight target set theoretically is achieved, and fairness, interpretability and model robustness of multi-index evaluation are improved. The method is suitable for thermal runaway risk characteristic parameters of any dimension, and a multi-dimensional scoring matrix can be constructed according to parameters selected by specific application, including thermal release intensity, toxicity exposure, explosion characteristics, pressure change, gas composition and the like. In the evaluation process, a sorting technology method, a principal component analysis method or a median method based on ideal solution similarity can be matched for comparison verification. According to the method, high-resolution, high-comparability and high-reliability risk grading can be realized under different types of batteries, and the method is particularly suitable for thermal safety evaluation with large sample fluctuation and remarkable variable difference.

Description

Technical Field

[0001] The present invention relates to the field of battery safety monitoring technology, and specifically to a thermal runaway risk assessment method, system, and storage medium that achieves fair integration of multiple indicators through a dynamic weight allocation mechanism. The method is particularly suitable for cross-system safety performance comparison of ternary lithium batteries, lithium iron phosphate batteries, sodium ion batteries, etc. Background Art

[0002] With the rapid development of battery technology, lithium-ion batteries have deeply penetrated into core energy scenarios such as electric vehicles and energy storage power stations, while new technologies such as sodium-ion batteries and high-nickel ternary systems are rapidly rising due to their energy density and cost advantages. However, the risk of thermal runaway that comes with the diversification of battery systems has become increasingly prominent - the chain reaction formed by the interweaving of gas explosion, toxic release and heat accumulation has become a major hidden danger threatening personal safety and facility integrity. Different types of batteries have significant differences in performance structure, chemical system and thermal stability. Therefore, whether their thermal safety is better than that of existing mature battery systems has become a core issue of current research. Therefore, how to establish a scientific, objective and comparable evaluation mechanism to systematically compare and quantitatively determine the hazards of different types of batteries during thermal runaway is of great significance.

[0003] However, existing battery thermal runaway risk assessment methods mostly rely on single thermal parameters (such as starting temperature, maximum temperature rise, etc.) or empirical judgments, and fail to fully cover multi-dimensional indicators such as gas release, explosiveness, toxicity and heat release intensity. They lack a unified mathematical modeling and weight processing mechanism, resulting in the assessment results being easily affected by subjective factors and difficult to adapt to the evaluation needs brought about by the diversity of new materials and new systems. Safety design lags behind technological iteration, posing a risk of misleading industry decision-making.

[0004] Therefore, there is an urgent need to build a thermal runaway risk assessment method that is driven by multiple indicators, data-adaptive, and physically interpretable to improve the accuracy and practicality of battery thermal safety assessment. Summary of the Invention

[0005] 1. Technical Issues

[0006] The present invention aims to provide a battery thermal runaway risk assessment method, system, and storage medium to solve the following problems:

[0007] 1. High variance parameters are overly dominant when integrating multiple indicators;

[0008] 2. Risk assessment results between different battery systems are not comparable;

[0009] 3. The model is not robust enough when the sample distribution shifts.

[0010] 2. Technical Solution

[0011] In order to solve the above technical problems, in a first aspect, the present invention provides a battery thermal runaway risk assessment method, comprising the following steps:

[0012] S1. Obtain multi-dimensional risk parameters during battery thermal runaway , , and , The parameters are constructed into a multi-dimensional risk indicator matrix;

[0013] S2. Perform variance statistics on each parameter index to obtain the variance value of each parameter index ;

[0014] S3. According to variance Calculate the weight of each parameter , which is and proportional to;

[0015] S4. For each battery sample, multiply the parameter value by its weight and sum them to obtain a weighted comprehensive score;

[0016] S5. Grade the battery thermal runaway risk based on the weighted comprehensive score.

[0017] Preferably, the weight in step S3 The calculation satisfies:

[0018]

[0019] in, For the The weight of each risk parameter, For the The variance of the risk parameter, is the total number of risk parameters.

[0020] Preferably, the expert scoring matrix ( , ), and the subjective and objective weights are integrated through the following formula:

[0021]

[0022] in, For the Expert scoring values ​​of risk parameters, For the first Expert scoring value of each risk parameter.

[0023] Preferably, in step S1, at least two of the multi-dimensional risk parameters are required for comprehensive evaluation.

[0024] Preferably, in step S4, the weighted comprehensive score is used to generate a visual result, wherein the bubble size of the bubble chart represents the score value, and the bubble color represents the battery type or risk level.

[0025] Preferably, the weighted comprehensive score is compared and verified with the results of classical evaluation methods, wherein the classical evaluation methods include at least one of a ranking technique based on the similarity of ideal solutions, principal component analysis, and a median method.

[0026] Preferably, the variance The calculation is based on a database under uniform experimental conditions, which should be consistent:

[0027] In a second aspect, the present invention provides a battery thermal runaway risk assessment system, wherein the system is configured to implement the battery thermal runaway risk assessment method according to the first aspect, including:

[0028] Parameter acquisition module, which obtains multi-dimensional risk parameters during battery thermal runaway;

[0029] Variance calculation module, which calculates the variance of each parameter based on the multidimensional risk parameter ;

[0030] The weight distribution module, based on the variance according to Calculate weights and support input of expert scoring matrix ( , );

[0031] The scoring module generates a weighted comprehensive score and risk level based on the weights.

[0032] Preferably, the weight allocation module performs:

[0033] or .

[0034] In a third aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the battery thermal runaway risk assessment method described in the first aspect.

[0035] 3. Beneficial Effects

[0036] The present invention establishes a database to clarify the variance contribution balance of each risk indicator in the comprehensive score of different battery types, fundamentally eliminating the systematic deviation in which high-volatility parameters dominate the results and low-volatility parameters are suppressed in the process of processing raw data in traditional evaluations. Thanks to this, the contribution rate of each indicator in the results of standardized linear regression has steadily converged from the extreme differentiation state to the equilibrium interval of the initial expected weight, significantly improving the fairness basis of risk assessment. This method enables different types of batteries such as ternary lithium, lithium iron phosphate, and sodium batteries to be objectively evaluated and ranked objectively according to the expected weights in a true sense under a unified standard. The more far-reaching impact is that through 、 The triple comparison validation using the median method and the proposed model demonstrates the unique stability of the proposed model in scenarios with sample distribution shifts. This ultimately achieves a paradigm shift from "subjective experience-driven" to "data-adaptive-driven," providing highly reliable decision-making support for power battery selection and energy storage system safety design. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for describing the embodiments or the prior art.

[0038] Figure 1 for Scoring results and standardized linear regression graph;

[0039] Figure 2 for Scoring results and standardized linear regression graph;

[0040] Figure 3 The median method scoring results and standardized linear regression graphs are shown;

[0041] Figure 4 Schematic diagram of the scoring results of the evaluation method of the present invention;

[0042] In the figure, It measures how well the regression model fits the target variable. DETAILED DESCRIPTION

[0043] The following examples are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0044] Example 1:

[0045] This embodiment provides a battery thermal runaway risk assessment method, comprising the following steps:

[0046] S1. Obtain multi-dimensional risk parameters during battery thermal runaway , , and , The parameters are constructed into a multi-dimensional risk indicator matrix;

[0047] S2. Perform variance statistics on each parameter index to obtain the variance value of each parameter index ;

[0048] S3. According to variance Calculate the weight of each parameter , which is and proportional to;

[0049] S4. For each battery sample, multiply the parameter value by its weight and sum them to obtain a weighted comprehensive score;

[0050] S5. Grade the battery thermal runaway risk based on the weighted comprehensive score.

[0051] Specifically, the currently commercial Square lithium iron phosphate battery, Square ternary nickel cobalt manganese oxide lithium ion battery, For example, for prismatic sodium-ion batteries, an expert scoring matrix of (1, 1, 1) is used. This matrix allows domain experts to vote on the importance of different risk parameters, such as explosiveness and toxicity. For example, if an expert believes that the toxicity parameter is more harmful in an actual accident, their weight can be manually increased to prevent the algorithm from underestimating the toxicity impact due to data volatility. This fusion of subjective and objective experience prevents underestimation of key risks due to data distribution characteristics, enhancing the engineering practicality and scenario-adaptive flexibility of the assessment results.

[0052] First, the target battery cell is fully charged and placed in an inert gas sealed chamber. The thermal runaway process is triggered by unilateral heating, and the pressure change rate inside the chamber is monitored in real time. When the pressure change rate is continuously lower than The main outgassing stage is determined to be over when the gas cloud is fully degassing. At this time, the total volume of the gas cloud, the peak pressure and the concentration of gas components are collected. The gas components include key components such as hydrogen, methane, carbon monoxide, and carbon dioxide.

[0053] Based on the gas parameters collected in the above experiments, the gas cloud explosion risk ERL and gas cloud toxicity of the gas cloud released by the battery are calculated respectively. Heat of burning gas clouds .

[0054] The gas cloud explosion risk ERL is determined by the gas cloud upper explosion limit UEL and the gas cloud lower explosion limit LEL. The specific calculation formula is as follows:

[0055]

[0056] Based on the gas and The effective dose fraction of the gas toxicity quantitative evaluation calculated by the concentration is referenced Standard calculations;

[0057] The total heat release capacity of the mixed gas is calculated based on the total mass of the gas and the unit combustion heat:

[0058]

[0059] in, For the corresponding quality , ;

[0060] Take the equivalent conversion factor ;

[0061] is the explosion coefficient, which is taken when the ground explodes ;

[0062] is the mass of the mixed gas generated by battery thermal runaway, ;

[0063] The combustion heat of the mixed gas generated by the thermal runaway of the battery, ;

[0064] Unit mass The heat of explosive explosion, .

[0065] As shown in Table 1, the parameters of 30 groups of samples such as ternary lithium batteries, sodium ion batteries, and lithium iron phosphate batteries obtained under unified experimental conditions are constructed into a standardized database. These parameters include material system, test environment, capacity, and calculated 、 、 values, forming a thermal runaway feature dataset covering different battery systems.

[0066]

[0067] Table 1 Database of gas cloud characteristics of thermal runaway gas production in existing battery inert atmosphere tank experiments

[0068] Then, enter the core weight allocation stage, 、 、 The variance statistics of each parameter are performed, a three-dimensional risk parameter matrix is ​​constructed, and its ranking technology based on the similarity of the ideal solution is calculated. , principal component analysis As well as the scores in classic evaluation methods such as the median method, to form a comparative verification with the method of this embodiment.

[0069] In order to verify the impact of each evaluation parameter on the final result, each parameter index is standardized and a standardized linear regression is performed.

[0070]

[0071]

[0072] in is the standardized parameter index, Indicates the Battery samples, Indicates the variables;

[0073] Indicates the The battery samples are in The original value of the variable;

[0074] Indicates the The sample mean of the variable;

[0075] Indicates the The standard deviation of the variables;

[0076] a. 、 Corresponding representation 、 、 The regression coefficient of the parameter indicator, i.e., the weight, reflects its influence on the final score;

[0077] is the constant term of the regression model, controlling the overall offset;

[0078] For the comprehensive rating.

[0079] The final result is as follows Figures 1 to 3 As shown, among which a, 、 The indicators vary greatly. Because different indicators have different dimensions and distribution ranges, if they are not weighted, the scoring results will be biased towards variables with large fluctuations, causing the actual contribution of each parameter in the evaluation results to deviate from the initial weighting.

[0080] Therefore, in order to achieve the same effect of each parameter index on the final result, this embodiment 、 、 Perform variance statistics on the three indicators to obtain the variance value of each indicator , based on the principle of equal variance contribution, the inverse standard deviation normalization method is used to calculate the weight of each indicator. The specific weight formula is:

[0081]

[0082] The derivation process of the formula is as follows:

[0083] The variance changes of the overall rating results are as follows:

[0084]

[0085] Assuming that each scoring result is independent, the above formula can be further written as:

[0086]

[0087] So the contribution of each variable to the variance is:

[0088]

[0089] So assuming each variable contributes equally the contribution is as follows:

[0090]

[0091]

[0092] Finally, the normalized weight expression is formed:

[0093]

[0094] in, For the The weight of each risk parameter, For the The variance of the risk parameter, is the total number of risk parameters, The number of indicators used to weight participation scores.

[0095] Then, the three standardized parameter indicators of each battery sample are multiplied by their corresponding weights and summed to obtain a comprehensive risk score. The comprehensive risk score It can be used as a quantitative basis for the battery thermal runaway hazard level.

[0096] All battery samples are sorted according to the comprehensive risk score and visualized through bubble charts. The bubble diameter and The colors are divided into risk levels according to the preset thresholds. The results are as follows Figure 4Compared with the unweighted method, the equal variance contribution method in this embodiment significantly reduces the scoring deviation caused by different indicator volatility, unifying the impact of all indicators on the final result to the same value of 0.33. In particular, it improves some low variance indicators, such as The underestimated system scores improve the accuracy and rationality of risk identification.

[0097] Example 2:

[0098] This embodiment provides a battery thermal runaway risk assessment system, which is configured to implement the battery thermal runaway risk assessment method described in Example 1, including a parameter acquisition module for acquiring multidimensional risk parameters in the battery thermal runaway process; a variance calculation module for calculating the variance of each parameter based on the multidimensional risk parameters. ; Weight distribution module, used according to the variance according to Calculate weights; and generate a score based on the weights to generate a weighted composite score and risk level.

[0099] Specifically, the parameter acquisition module receives the raw data input of the battery thermal runaway experiment through the sensor interface, including the UEL and LEL values ​​output by the gas concentration sensor, the gas temperature data recorded by the thermal imager, and the cavity pressure changes monitored by the pressure sensor. The built-in processor of the parameter acquisition module calculates the gas cloud explosion risk in real time. and Equivalent heat release :

[0100]

[0101]

[0102] Directly obtain the basis Standard calculated toxic exposure dose The output of the variance calculation module is connected to the input of the variance calculation module through a high-speed data bus. 、 、 Parameter groups are transmitted in real time.

[0103] Variance calculation module Chip passed After receiving the parameter group transmitted by the parameter acquisition module, the interface performs variance statistics calculation on the current parameter group. The pin is directly connected to the input register of the weight distribution module, and the calculated variance value is written in real time in the form of a digital signal.

[0104] Weight distribution module After the kernel reads the variance value from the input register, it executes the fixed core algorithm:

[0105]

[0106] Its output is through The serial bus connects the input buffer of the score generation module and 、 、 The corresponding three-way weight values ​​are packaged and transmitted.

[0107] Rating generation module The processor performs two tasks simultaneously: on the one hand, The channel reads the current sample from the shared memory of the parameter acquisition module 、 、 value, on the other hand through The serial bus receives the data transmitted by the weight distribution module The weighted composite score is then calculated.

[0108] The final output is through The interface is connected to the display device to visualize the scoring results in the form of a bubble chart, with the bubble diameter mapped The color space is divided into three zones of red, yellow and green according to the preset risk level thresholds, and the structured assessment report is transmitted to the cloud server through the Ethernet interface.

[0109] Example 3:

[0110] This embodiment provides a computer-readable storage medium on which is stored a computer program compiled by the battery thermal runaway risk assessment method described in Example 1, which can be recognized and executed by a processor. The storage medium includes but is not limited to hard disk, solid-state drive, or cloud storage server.

[0111] In summary, the present invention reconstructs the mathematical basis for battery thermal runaway risk assessment through the original principle of equal variance contribution.

[0112] Based on the formula It enforces the balance of variance contributions of various indicators in the comprehensive score, and thoroughly solves the systematic defects of traditional methods in which high-volatility parameters dominate the score and the contribution of low-volatility parameters is suppressed, providing a risk assessment paradigm with rigorous theory and reliable implementation for the field of battery safety.

[0113] Although the present invention has been described in detail above using general descriptions and specific embodiments, some modifications or improvements can be made on the basis of the present invention. However, these modifications or improvements made without departing from the spirit of the present invention are within the scope of protection claimed by the present invention.

Claims

1. A battery thermal runaway risk assessment method, characterized in that: The following steps are involved: S1. Obtain multi-dimensional risk parameters during battery thermal runaway , , and , The parameters are constructed into a multi-dimensional risk indicator matrix; S2. Perform variance statistics on each parameter index to obtain the variance value of each parameter index ; S3. According to variance Calculate the weight of each parameter , which is and proportional to; S4. For each battery sample, multiply the parameter value by its weight and sum them to obtain a weighted comprehensive score; S5. Grade the battery thermal runaway risk based on the weighted comprehensive score.

2. The battery thermal runaway risk assessment method according to claim 1, characterized in that: The weight in step S3 The calculation satisfies: in, For the The weight of each risk parameter, For the The variance of the risk parameter, is the total number of parameter indicators.

3. The battery thermal runaway risk assessment method according to claim 1 or 2, characterized in that: In step S3, the expert scoring matrix is ​​introduced ( , ), and the subjective and objective weights are integrated through the following formula: in, For the Expert scoring values ​​of risk parameters, For the first Expert scoring value of each risk parameter.

4. The battery thermal runaway risk assessment method according to claim 1, characterized in that: In step S1, at least two of the multi-dimensional risk parameters are required for comprehensive evaluation.

5. The battery thermal runaway risk assessment method according to claim 1, characterized in that: In step S4, the weighted comprehensive score is used to generate a visual result, where the bubble size of the bubble chart represents the score value, and the bubble color represents the battery type or risk level.

6. The battery thermal runaway risk assessment method according to claim 1, characterized in that: Also includes: The weighted comprehensive score is compared and verified with the results of classical evaluation methods, wherein the classical evaluation methods include at least one of a ranking technique based on the similarity of ideal solutions, principal component analysis, and a median method.

7. The battery thermal runaway risk assessment method according to claim 1, characterized in that: In step S2, the variance The calculation of is based on a database under uniform experimental conditions, and the experimental conditions should be consistent.

8. Battery thermal runaway risk assessment system, characterized by: The system is configured to implement the battery thermal runaway risk assessment method according to any one of claims 1 to 7, comprising: Parameter acquisition module, which obtains multi-dimensional risk parameters during battery thermal runaway; Variance calculation module, which calculates the variance of each parameter based on the multidimensional risk parameter ; The weight distribution module, based on the variance according to Calculate weights and support input of expert scoring matrix ( , ); The scoring module generates a weighted comprehensive score and risk level based on the weights.

9. The battery thermal runaway risk assessment system according to claim 8, characterized in that: The weight distribution module performs: or .

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the battery thermal runaway risk assessment method according to any one of claims 1 to 7 is implemented.

Images