Lithium battery thermal runaway early warning and protection method based on adaptive neural network

Abstract

The invention relates to the technical field of lithium battery safety management, and particularly discloses a lithium battery thermal runaway early warning and protection method based on an adaptive neural network, which comprises the following steps: acquiring temperature, internal resistance and voltage data of a lithium battery in real time, extracting multi-dimensional features based on discrete wavelet transform, singular value decomposition and principal component analysis algorithms, and performing early warning and protection on the lithium battery thermal runaway. The influence of internal resistance and voltage on abnormal temperature fluctuation is quantified; constructing a feature vector through the comprehensive influence coefficient, inputting the feature vector into an adaptive fuzzy neural network prediction model, outputting a temperature anomaly fluctuation coefficient, and accurately predicting the thermal runaway risk of the lithium battery; according to the lithium battery thermal runaway prediction method, multi-level protection measures such as early warning response triggering, system automatic charging and discharging strategy adjusting, power output reducing and heat dissipation system starting are taken based on the prediction result, development of the thermal runaway risk is restrained, multi-dimensional data fusion and an intelligent algorithm are combined, and the lithium battery thermal runaway prediction accuracy and the risk control real-time performance are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery safety management, and particularly to a method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network. Background Art

[0002] With the wide application of lithium batteries in new energy electric vehicles, portable electronic devices and energy storage systems, their safety issues have become the focus and difficulty of research. Lithium batteries are widely used due to their high energy density and good cycle performance. However, in practical applications, lithium batteries are prone to thermal runaway under extreme conditions such as overcharging, over-discharging, high temperature, low temperature or mechanical damage, resulting in a rapid increase in the internal temperature of the battery, which may further lead to safety accidents such as fires or explosions. Therefore, how to monitor the operating state of lithium batteries in real time, accurately identify the thermal runaway risk and take protective measures in time has become an important technical issue to ensure the safe use of lithium batteries.

[0003] The existing technologies have the following deficiencies: The existing technologies cannot simultaneously achieve accurate prediction and real-time protection of the thermal runaway risk of lithium batteries, especially there are obvious deficiencies in the correlation analysis and dynamic response optimization of complex multi-dimensional data. This leads to difficulties in ensuring the real-time performance, accuracy and stability in the assessment and protection of lithium battery thermal runaway risk, thus increasing the possibility of serious safety accidents. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network to solve the problems in the above background.

[0005] The purpose of the present invention can be achieved by the following technical solutions: A method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network includes the following steps: S1: Real-time collect the working temperature data of the lithium battery, analyze the working temperature fluctuation of the lithium battery, identify the characteristics of abnormal temperature fluctuation, and judge whether there is thermal runaway during the working process of the lithium battery; S2: Real-time collect the internal resistance and voltage data of the lithium battery, analyze the change trends of the internal resistance and voltage, identify the characteristic patterns related to the thermal runaway risk, and calculate the influence degree of the changes in the internal resistance and voltage of the lithium battery on the temperature of the lithium battery; S3: According to the analysis result of the influence degree, divide the influence degree into serious influence and slight influence; S4: Based on the serious influence, construct a thermal runaway prediction model according to the internal resistance and voltage of the lithium battery, predict the thermal runaway of the lithium battery, and perform early warning response and protection according to the prediction result.

[0006] As a further solution of the present invention: The determination of whether there is thermal runaway during the operation of the lithium battery specifically includes: Within a monitoring period, use a temperature sensor to collect the working temperature data of the lithium battery in real time. According to the degree of temperature fluctuation of the lithium battery, calculate the temperature abnormal fluctuation coefficient, and determine whether the temperature abnormal fluctuation coefficient is greater than or equal to a preset threshold. If so, there is thermal runaway; if not, there is no thermal runaway.

[0007] As a further solution of the present invention: The process of obtaining the temperature abnormal fluctuation coefficient is as follows: By collecting the working temperature data of the lithium battery in real time, decomposing and analyzing the data based on the discrete wavelet transform algorithm, extracting the frequency domain characteristics of temperature fluctuation, and calculating the temperature abnormal fluctuation coefficient to evaluate the thermal runaway risk, including: Normalize the collected temperature data; Select a wavelet basis function and the number of decomposition layers, and decompose the normalized temperature data into low-frequency coefficients and high-frequency coefficients at multiple frequency levels; Calculate the energy of each layer for the high-frequency wavelet coefficients; Sum up the energies of all high-frequency wavelet coefficient layers to obtain the total energy of the high-frequency wavelet coefficient layers. Calculate the ratio of the total energy of the high-frequency wavelet coefficient layers to the total energy of all wavelet transform layers to obtain the temperature abnormal fluctuation coefficient.

[0008] As a further solution of the present invention: The calculation of the influence degree of the change in the internal resistance and voltage of the lithium battery on the temperature of the lithium battery specifically includes: By collecting the internal resistance data and voltage data of the lithium battery in real time, using the time series analysis method to extract the change trends of the internal resistance and voltage, constructing a multi-dimensional feature space to capture the feature patterns related to the thermal runaway risk, and through trend analysis and fluctuation characteristic decomposition, deeply quantify the abnormal changes of the internal resistance and voltage respectively, calculate the internal resistance influence factor and the voltage influence factor, and perform a normalized comprehensive calculation process on the internal resistance influence factor and the voltage influence factor to obtain the comprehensive influence coefficient. Compare the comprehensive influence coefficient with a preset threshold to determine the influence degree of the change in the internal resistance and voltage of the lithium battery on the temperature of the lithium battery.

[0009] As a further solution of the present invention: The acquisition logic of the internal resistance influence factor is as follows: Collect the internal resistance data of the lithium battery in real time. The internal resistance data represents the change of the internal resistance of the lithium battery over time during operation; Normalize the collected internal resistance data; The internal resistance data is segmented into multiple segments according to a preset time window. The internal resistance data within each window constitutes a vector, and the internal resistance data within all time windows is combined into a matrix. Each column of the matrix represents the characteristics of the internal resistance change within a time window, and each row represents the internal resistance value at each time point within the corresponding time window; Perform singular value decomposition on the matrix, including the left singular vector matrix, the diagonal matrix, and the right singular vector matrix; Calculate the energy proportion of each singular value; Based on the energy characteristics after singular value decomposition, the internal resistance influence factor is the sum of the energy proportions of all singular values.

[0010] As a further solution of the present invention: It is characterized in that the acquisition logic of the voltage influence factor is as follows: Collect the voltage data of the lithium battery in real time and perform standardization processing; After obtaining the standardized voltage data, construct the covariance matrix of the voltage data; Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues and corresponding eigenvectors; According to the calculated eigenvalues and eigenvectors, select the eigenvectors corresponding to the first principal components to form the principal component matrix; According to the results of principal component analysis, calculate the contribution degree of each principal component; Calculate by multiplying the contribution degree of each principal component by its corresponding eigenvalue and summing them to obtain the voltage influence factor.

[0011] As a further solution of the present invention: According to the analysis results of the influence degree, the influence degree is divided into serious influence and slight influence, specifically including: Judge whether the comprehensive influence coefficient is greater than or equal to a preset threshold. If so, it is recorded as a serious influence. If not, it is recorded as a slight influence.

[0012] As a further solution of the present invention: Construct the thermal runaway prediction model to predict the thermal runaway of the lithium battery, specifically including: Construct a comprehensive feature vector from the internal resistance influence factor and the voltage influence factor of the lithium battery as the input of the runaway prediction model. The output of the model is the temperature abnormal fluctuation coefficient of the lithium battery. The goal of the model is to minimize the prediction error function through training to make the predicted temperature abnormal fluctuation coefficient consistent with the actual temperature abnormal fluctuation coefficient; when the change amplitude of the error is less than the set threshold, stop training to obtain the final model parameters, where the runaway prediction model is an adaptive fuzzy neural network.

[0013] As a further solution of the present invention: Perform early warning response and protection measures according to the prediction results, specifically including: The coefficient of abnormal temperature fluctuation of a lithium battery based on adaptive fuzzy neural network prediction. When the coefficient of abnormal temperature fluctuation exceeds the set preset threshold, the system will automatically trigger an early warning response, including sending an alarm signal, notifying the management personnel, and displaying the current state and degree of abnormality of the battery through a data visualization interface, ensuring that the battery management system can promptly identify potential thermal runaway risks; Based on the early warning response, an automatic activation of protective measures is constructed. The system automatically adjusts the working environment of the battery according to the real-time monitoring data of temperature, internal resistance, and voltage, including increasing the cooling efficiency through the heat dissipation system, activating the over-temperature protection mechanism, and restricting the charge and discharge rate of the battery.

[0014] The beneficial effects of the present invention: (1) Through highly sensitive temperature sensors, internal resistance testers, and voltage sensors deployed inside the battery, multi-dimensional data of the battery in a dynamic working state are continuously collected, including real-time change information of temperature, internal resistance, and voltage. Subsequently, advanced discrete wavelet transform is used to decompose the frequency domain characteristics of the temperature data, effectively extracting its high-frequency fluctuation characteristics and revealing potential abnormal fluctuation patterns; singular value decomposition is used to perform multi-dimensional degradation analysis on the internal resistance data to accurately capture its change trend and main influencing patterns; at the same time, principal component analysis is used to extract the fluctuation characteristics of the voltage data and quantify the correlation strength between key characteristics and thermal runaway. By normalizing the influencing factors of internal resistance and voltage, a comprehensive influence coefficient is further constructed as the core input of the prediction model. To improve the prediction accuracy, the present invention introduces an adaptive fuzzy neural network model, which combines the reasoning ability of fuzzy logic and the learning ability of neural networks. Through the input of the comprehensive influence coefficient, accurate modeling and prediction of the non-linear complex characteristics during the thermal runaway of lithium batteries are achieved.

[0015] (2)The present invention not only has excellent accuracy in predicting the risk of thermal runaway, but also further improves the full-process safety management mechanism from risk identification to active response. Through the thermal runaway prediction model based on the adaptive fuzzy neural network, the system can output the coefficient of abnormal temperature fluctuation of the battery in real time, and can quickly detect potential safety hazards. When the prediction result exceeds the preset threshold, the system immediately triggers a multi-level early warning response mechanism, including sending out real-time alarm signals, transmitting risk information to the battery management system, and displaying abnormal features and severity to the operator through the data visualization interface. This early warning mechanism can ensure that managers and the system can timely grasp the state of the battery, laying a foundation for the rapid implementation of subsequent protection measures. Compared with the traditional battery management scheme with passive response, the present invention realizes the full-life cycle safety management from "detection - prediction - early warning - protection", improving the safety and stability of lithium batteries under various complex working conditions. At the same time, this intelligent and multi-level early warning and protection strategy effectively reduces the occurrence probability of serious safety accidents such as fires and explosions, providing reliable technical support for the application of lithium batteries in key fields such as industry, transportation, and energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The present invention will be further described below in conjunction with the accompanying drawings.

[0017] Figure 1 It is a specific step flow block diagram of a method for thermal runaway early warning and protection of lithium batteries based on an adaptive neural network according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with 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 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.

[0019] Please refer to Figure 1 as shown, the present invention is a method for thermal runaway early warning and protection of lithium batteries based on an adaptive neural network, including the following steps: S1: Collect the working temperature data of the lithium battery in real time, analyze the working temperature fluctuation of the lithium battery, identify the characteristics of abnormal temperature fluctuation, and judge whether there is thermal runaway during the working process of the lithium battery; S2: Collect the internal resistance and voltage data of the lithium battery in real time, analyze the change trends of the internal resistance and voltage, identify the characteristic patterns related to the thermal runaway risk, and calculate the influence degree of the changes in the internal resistance and voltage of the lithium battery on the temperature of the lithium battery; S3: According to the analysis result of the influence degree, divide the influence degree into serious influence and slight influence; S4: Based on the severe impact, construct a thermal runaway prediction model according to the internal resistance and voltage of the lithium battery, predict the thermal runaway of the lithium battery, and carry out early warning response according to the prediction results.

[0020] In S1: Real-time collect the working temperature data of the lithium battery, analyze the working temperature fluctuation of the lithium battery, identify the characteristics of abnormal temperature fluctuation, and judge whether there is thermal runaway during the working process of the lithium battery, specifically including: Within a monitoring period, use a temperature sensor to collect the working temperature data of the lithium battery in real time. According to the degree of working temperature fluctuation of the lithium battery, calculate the abnormal temperature fluctuation coefficient, and judge whether the abnormal temperature fluctuation coefficient is greater than or equal to the preset threshold. If so, there is thermal runaway; if not, there is no thermal runaway. The acquisition process of the abnormal temperature fluctuation coefficient is as follows: By collecting the working temperature data of the lithium battery in real time, decompose and analyze the data based on the discrete wavelet transform algorithm, extract the frequency domain characteristics of temperature fluctuation, and calculate the abnormal temperature fluctuation coefficient to evaluate the thermal runaway risk, including: Temperature data preprocessing, normalize the collected temperature data to eliminate the influence of data amplitude differences. Wavelet decomposition and coefficient extraction, select the wavelet basis function and the decomposition layer number , decompose the normalized temperature data into low-frequency coefficients and high-frequency coefficients at multiple frequency levels, and the calculation expression is: ; In the formula, represents the temperature data collected in real time, represents the th wavelet decomposition layer, represents the total number of wavelet decomposition layers, represents the temperature data, represents the low-frequency coefficient corresponding to the wavelet decomposition of the th layer, represents the high-frequency coefficient corresponding to the wavelet decomposition of the th layer, represents the acquisition time point; represents the acquisition time point; Calculate the energy of each layer for the high-frequency wavelet coefficients to quantify the fluctuation amplitude, and the calculation expression is: ; In the formula, represents the energy corresponding to the wavelet decomposition of the th layer; Calculate the total energy of all high-frequency wavelet coefficient layers, and calculate the ratio with the total energy of all wavelet transform layers to obtain the abnormal temperature fluctuation coefficient.

[0021] It should be noted that in the prior art, the abnormal evaluation of temperature fluctuations usually adopts simple statistical methods, such as directly calculating the variance or range of temperature data, lacking fine-grained analysis of multi-band fluctuation characteristics. The present invention, through the calculation method combining discrete wavelet transform and energy ratio, not only refines the frequency components of temperature fluctuations, but also effectively extracts high-frequency characteristics directly related to thermal runaway, improving the accuracy and reliability of risk assessment.

[0022] In S2, the internal resistance and voltage data of the lithium battery are collected in real time, the change trends of the internal resistance and voltage are analyzed, the characteristic patterns related to the thermal runaway risk are identified, and the influence degree of the change of the internal resistance and voltage of the lithium battery on the temperature of the lithium battery is calculated, specifically including: By collecting the internal resistance data and voltage data of the lithium battery in real time, using the time series analysis method to extract the change trends of the internal resistance and voltage, constructing a multi-dimensional feature space to capture the characteristic patterns related to the thermal runaway risk, through trend analysis and fluctuation characteristic decomposition, deeply quantifying the abnormal changes of the internal resistance and voltage respectively, calculating the internal resistance influence factor and the voltage influence factor, and performing a normalized comprehensive calculation process on the internal resistance influence factor and the voltage influence factor to obtain a comprehensive influence coefficient, and comparing the comprehensive influence coefficient with a preset threshold to judge the influence degree of the change of the internal resistance and voltage of the lithium battery on the temperature of the lithium battery; The process of obtaining the internal resistance influence factor is as follows: The internal resistance data of the lithium battery are collected in real time, and the internal resistance data represent the change of the internal resistance of the lithium battery over time during operation. In order to eliminate the influence caused by different measurement ranges, the collected internal resistance data are normalized to make the ranges of the internal resistance data consistent, so as to ensure the accuracy of subsequent analysis.

[0023] The internal resistance data are segmented into multiple segments according to a preset time window. The internal resistance data within each window form a vector, and the internal resistance data within all time windows are combined into a matrix. Each column of this matrix represents the internal resistance change characteristics within a time window, and each row represents the internal resistance value at each time point within this time window. In this way, the internal resistance data are presented in the form of a matrix, which is convenient for subsequent analysis.

[0024] Perform singular value decomposition on the matrix, and the calculation expression is: ; In the formula, represents the left singular vector matrix, represents the diagonal matrix, containing the singular values , represents the right singular vector matrix; Calculate the energy proportion of each singular value, and the calculation expression is: ; In the formula, represents the The energy proportion of a singular value, denotes the th singular value, denotes the number of singular values, denotes the total number of singular values; For the energy feature after singular value decomposition, the internal resistance influence factor is the sum of the energy proportions of all singular values, and a function is used to quantify the influence of abnormal fluctuations on temperature. The calculation expression is: ; In the formula, denotes the internal resistance influence factor, denotes the number of main singular values selected.

[0025] It should be noted that: In the present invention, the main patterns in the internal resistance change are extracted, and the internal resistance influence factor is calculated through singular values and energy proportions to further quantify the influence of the internal resistance on temperature anomalies.

[0026] The process for obtaining the voltage influence factor is as follows: Real-time collect the voltage data of the lithium battery and perform standardization processing to remove the mean and variance in the data to ensure the unity of the data; After obtaining the standardized voltage data, construct the covariance matrix of the voltage data to quantify the relationship between each moment in the voltage time series. The calculation expression of the covariance matrix is: ; In the formula, denotes the covariance matrix, denotes the acquisition time point, denotes the total number of acquisition time points, denotes the standardized voltage data, denotes the mean of the voltage data, denotes the transpose operation.

[0027] Perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalues and the corresponding eigenvectors . The eigenvalue represents the degree of variation of this feature, and the eigenvector represents the direction of this feature; among them, denotes the number of principal components; According to the calculated eigenvalues and eigenvectors, select the eigenvectors corresponding to the first principal components to form the principal component matrix . The purpose of selecting the first principal components is to extract the main change patterns in the data. The selection basis for each principal component is the magnitude of the eigenvalue, and the principal components with larger eigenvalues are preferentially selected; According to the results of the principal component analysis, calculate the contribution degree of each principal component. The calculation expression is as follows: ; In the formula, represents the contribution degree of the th principal component, represents the total number of principal components; calculate by multiplying the contribution degree of each principal component by its corresponding eigenvalue and summing them to obtain the voltage influence factor.

[0028] The calculation expression of the comprehensive influence coefficient is as follows: ; In the formula, represents the comprehensive influence coefficient, and represent preset proportionality coefficients, and and are both greater than 0, represents the internal resistance influence factor, represents the voltage influence factor.

[0029] It should be noted that: the comprehensive influence coefficient reflects the comprehensive influence coefficient of the lithium battery resistance and voltage on the thermal failure of the lithium battery, and when the value of the comprehensive influence coefficient is larger, the corresponding influence degree is higher.

[0030] It should be noted that: in the present invention, the temperature data is monitored and collected in real time through the thermistor sensor deployed inside the lithium battery, the internal resistance of the lithium battery is accurately measured through the internal resistance tester deployed inside the lithium battery using the frequency response analysis technology, and the voltage data is monitored and collected in real time through the voltage sensor deployed inside the lithium battery.

[0031] In S3, according to the analysis result of the influence degree, divide the influence degree into serious influence and slight influence, specifically including: Compare the comprehensive influence coefficient with the preset threshold. If the comprehensive influence coefficient is greater than or equal to the preset threshold, it indicates that the influence degree of the corresponding internal resistance and voltage on the thermal runaway of the lithium battery is high, which is recorded as a serious influence. If the comprehensive influence coefficient is less than the preset threshold, it indicates that the influence degree of the corresponding internal resistance and voltage on the thermal runaway of the lithium battery is low, which is recorded as a slight influence.

[0032] In S4, based on the serious influence, construct a thermal runaway prediction model according to the internal resistance and voltage of the lithium battery, predict the thermal runaway of the lithium battery, and perform early warning response and protection according to the prediction result, specifically including: Construct a comprehensive feature vector from the internal resistance influence factor and voltage influence factor of the lithium battery as the input of the out-of-control prediction model. The output of the model is the temperature abnormal fluctuation coefficient of the lithium battery. The goal of the model is to minimize the prediction error function through training to make the predicted temperature abnormal fluctuation coefficient consistent with the actual temperature abnormal fluctuation coefficient. When the change amplitude of the error is less than the set threshold, stop training to obtain the final model parameters. Among them, the out-of-control prediction model is an adaptive fuzzy neural network, specifically including: Obtain the comprehensive feature vector constructed from the internal resistance influence factor and voltage influence factor of the lithium battery as the input of the neural network model. In the construction of the prediction model, an adaptive fuzzy neural network is used for modeling. By taking the influence factors of internal resistance and voltage as the input, the neural network model outputs the temperature abnormal fluctuation coefficient of the lithium battery. The adaptive fuzzy neural network can effectively capture the non-linear influence of internal resistance and voltage on the battery temperature change through the combination of fuzzy rules and neural network, and continuously optimize the fuzzy rules and network parameters through adaptive learning, so as to accurately predict the abnormal fluctuation of the battery temperature.

[0033] During the training process, by minimizing the prediction error function, the error between the predicted temperature abnormal fluctuation coefficient and the actual temperature abnormal fluctuation coefficient is minimized. The error function describes the gap between the predicted value and the actual value, and the network parameters are adjusted iteratively to minimize the error. When the change amplitude of the error during the training process is less than the set threshold, the training process ends, the model parameters will be fixed, and finally an optimized thermal runaway prediction model is obtained.

[0034] In practical applications, combined with the real-time monitored internal resistance and voltage data, input the comprehensive feature vector through the trained adaptive fuzzy neural network model to predict the temperature abnormal fluctuation coefficient of the lithium battery. If the predicted temperature abnormal fluctuation coefficient exceeds the set threshold, the system will trigger the thermal runaway warning mechanism and take corresponding protective measures.

[0035] Among them, early warning response and protection are carried out according to the prediction results, specifically including: Based on the temperature abnormal fluctuation coefficient of the lithium battery predicted by the adaptive fuzzy neural network, when this coefficient exceeds the set safety threshold, the system will automatically trigger an early warning response. This early warning response includes sending an alarm signal, notifying relevant personnel or control systems, and displaying the current state and abnormal degree of the battery through a data visualization interface to ensure that the battery management system can timely identify potential thermal runaway risks. The early warning response can also be linked with other safety measures in the battery management system, such as adjusting the charge and discharge strategy, reducing the battery power output or cutting off the power supply, etc., to reduce the probability of thermal runaway.

[0036] Based on the warning response, the present invention further proposes the automatic activation of protective measures. The system automatically adjusts the working environment of the battery according to the real-time monitoring data of temperature, internal resistance and voltage, such as increasing the cooling efficiency through the heat dissipation system or starting the over-temperature protection mechanism to limit the charge and discharge rate of the battery. The activation of the protection mechanism can effectively prevent the battery from further heating up and reduce the safety risks caused by thermal runaway. The implementation of the entire warning and protection process ensures the safety and stability of lithium batteries during use and significantly reduces the potential risks of fire or explosion.

[0037] Working principle of the present invention: By collecting the temperature, internal resistance and voltage data of lithium batteries in real time, combining multi-dimensional feature extraction and complex signal analysis techniques, a systematic thermal runaway risk assessment and prediction model is established. The working temperature data is obtained by using a temperature sensor, and its frequency domain characteristics are analyzed by discrete wavelet transform to extract the temperature abnormal fluctuation coefficient to judge the thermal runaway risk of the battery. Based on the internal resistance tester and voltage sensor to obtain the internal resistance and voltage data, the influence on temperature change is quantified through singular value decomposition and principal component analysis respectively, and the comprehensive influence coefficient is calculated to capture the key feature patterns related to thermal runaway. Subsequently, an adaptive fuzzy neural network model is constructed, with the comprehensive influence coefficient as the input, and by optimizing the fuzzy rules and network parameters, the temperature abnormal fluctuation coefficient is output to predict the thermal runaway risk level. The model is iteratively optimized by minimizing the error function, and when the change amplitude of the error is less than the set threshold, the final model parameters are output. Based on the prediction results, when the temperature abnormal fluctuation coefficient exceeds the threshold, the system automatically triggers a warning response, including alarms, data visualization and linkage protection mechanisms, and dynamically adjusts the charge and discharge strategies and heat dissipation management to control the risks. Through multi-level risk analysis, prediction and protection, the present invention significantly improves the safety and stability of lithium batteries and is applicable to a wide range of lithium battery monitoring and management scenarios.

[0038] The above formulas are all dimensionless and take their numerical calculations. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0039] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired or wireless (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, or a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0040] It should be understood that the term "and / or" in this document is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood with reference to the context before and after.

[0041] It should be understood that in various embodiments of the present application, the magnitudes of the serial numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0042] The above has described in detail one embodiment of the present invention, but the content described is only a preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the application of the present invention should still fall within the scope covered by the patent of the present invention.

Claims

1. A method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network, characterized in that, It includes the following steps: S1: Collect the working temperature data of the lithium battery in real time, analyze the working temperature fluctuation of the lithium battery, identify the characteristics of abnormal temperature fluctuations, and determine whether there is thermal runaway during the working process of the lithium battery; S2: Collect the internal resistance and voltage data of the lithium battery in real time, analyze the change trends of the internal resistance and voltage, identify the characteristic patterns related to the thermal runaway risk, and calculate the influence degree of the changes in the internal resistance and voltage of the lithium battery on the temperature of the lithium battery; S3: According to the analysis result of the influence degree, divide the influence degree into serious influence and slight influence; S4: Based on the serious influence, construct a thermal runaway prediction model according to the internal resistance and voltage of the lithium battery, predict the thermal runaway of the lithium battery, and perform early warning response and protection according to the prediction result.

2. The method for lithium battery thermal runaway early warning and protection based on an adaptive neural network according to claim 1, wherein The determination of whether there is thermal runaway during the working process of the lithium battery specifically includes: Within a monitoring period, use a temperature sensor to collect the working temperature data of the lithium battery in real time. According to the degree of working temperature fluctuation of the lithium battery, calculate the abnormal temperature fluctuation coefficient, and determine whether the abnormal temperature fluctuation coefficient is greater than or equal to the preset threshold. If so, there is thermal runaway; if not, there is no thermal runaway.

3. A method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network according to claim 2, characterized in that, The acquisition process of the abnormal temperature fluctuation coefficient is: By collecting the working temperature data of the lithium battery in real time, decomposing and analyzing the data based on the discrete wavelet transform algorithm, extracting the frequency domain characteristics of the temperature fluctuation, and calculating the abnormal temperature fluctuation coefficient to evaluate the thermal runaway risk, including: Normalize the collected temperature data; Select the wavelet basis function and the decomposition layer number, and decompose the normalized temperature data into low-frequency coefficients and high-frequency coefficients at multiple frequency levels; Calculate the energy of each layer for the high-frequency wavelet coefficients; Sum up the energies of all high-frequency wavelet coefficient layers to obtain the total energy of the high-frequency wavelet coefficient layers. Calculate the ratio of the total energy of the high-frequency wavelet coefficient layers to the total energy of all wavelet transform layers to obtain the abnormal temperature fluctuation coefficient.

4. A method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network according to claim 1, characterized in that, The calculation of the influence degree of the changes in the internal resistance and voltage of the lithium battery on the temperature of the lithium battery specifically includes: By collecting the internal resistance data and voltage data of the lithium battery in real time, using the time series analysis method to extract the change trends of the internal resistance and voltage, constructing a multi-dimensional feature space to capture the characteristic patterns related to the thermal runaway risk, and through trend analysis and fluctuation characteristic decomposition, deeply quantify the abnormal changes of the internal resistance and voltage respectively, calculate the internal resistance influence factor and the voltage influence factor, and perform a normalized comprehensive calculation process on the internal resistance influence factor and the voltage influence factor to obtain the comprehensive influence coefficient. Compare the comprehensive influence coefficient with the preset threshold to determine the influence degree of the changes in the internal resistance and voltage of the lithium battery on the temperature of the lithium battery.

5. A method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network according to claim 4, characterized in that, The acquisition logic of the internal resistance influence factor is: Collect the internal resistance data of the lithium battery in real time. The internal resistance data represents the change of the internal resistance of the lithium battery over time during the working process; normalize the collected internal resistance data; The internal resistance data is segmented into multiple segments according to a preset time window. The internal resistance data within each window constitutes a vector, and the internal resistance data within all time windows is combined into a matrix. Each column of the matrix represents the characteristics of the internal resistance change within a time window, and each row represents the internal resistance value at each time point within the corresponding time window; Perform singular value decomposition on the matrix, including the left singular vector matrix, the diagonal matrix, and the right singular vector matrix; Calculate the energy proportion of each singular value; Based on the energy characteristics after singular value decomposition, the internal resistance influence factor is the sum of the energy proportions of all singular values.

6. A method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network according to claim 4, characterized in that, The acquisition logic of the voltage influence factor is as follows: Real-time collect the voltage data of the lithium battery and perform standardization processing; After obtaining the standardized voltage data, construct the covariance matrix of the voltage data; Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues and corresponding eigenvectors; According to the calculated eigenvalues and eigenvectors, select the eigenvectors corresponding to the first principal components to form a principal component matrix; According to the results of principal component analysis, calculate the contribution degree of each principal component; Calculate by multiplying the contribution degree of each principal component by its corresponding eigenvalue and summing them to obtain the voltage influence factor.

7. A method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network according to claim 1, characterized in that, According to the analysis results of the influence degree, the influence degree is divided into serious influence and slight influence, specifically including: Judge whether the comprehensive influence coefficient is greater than or equal to the preset threshold. If so, it is recorded as a serious influence. If not, it is recorded as a slight influence.

8. A method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network according to claim 1, characterized in that, Construct a thermal runaway prediction model to predict the thermal runaway of the lithium battery, specifically including: Construct a comprehensive feature vector from the internal resistance influence factor and the voltage influence factor of the lithium battery as the input of the runaway prediction model. The output of the model is the temperature abnormal fluctuation coefficient of the lithium battery. The goal of the model is to minimize the prediction error function through training to make the predicted temperature abnormal fluctuation coefficient consistent with the actual temperature abnormal fluctuation coefficient; when the change amplitude of the error is less than the set threshold, stop training to obtain the final model parameters, where the runaway prediction model is an adaptive fuzzy neural network.

9. A method for early warning and protection of lithium battery thermal runaway based on an adaptive neural network according to claim 1, characterized in that, Perform early warning response and protection measures according to the prediction results, specifically including: Based on the temperature abnormal fluctuation coefficient of the lithium battery predicted by the adaptive fuzzy neural network, when the temperature abnormal fluctuation coefficient exceeds the set preset threshold, the system will automatically trigger an early warning response, including sending an alarm signal, notifying the management personnel, and displaying the current state and abnormal degree of the battery through the data visualization interface to ensure that the battery management system can timely identify potential thermal runaway risks; On the basis of the early warning response, construct the automatic activation of protection measures. The system automatically adjusts the working environment of the battery according to the real-time monitoring data of temperature, internal resistance, and voltage, including increasing the cooling efficiency through the heat dissipation system, starting the over-temperature protection mechanism, and restricting the charge and discharge rate of the battery.

Images