A method of battery internal short circuit evaluation

Abstract

A kind of battery internal short circuit evaluation method, including data acquisition module, data preprocessing module, feature extraction and fusion module, parameter and consumption effect analysis module, fusion diagnosis module, verification and optimization module, data acquisition module is used for battery data acquisition, data preprocessing module is used for data preprocessing, feature extraction and fusion module are used for the extraction and fusion of parameter feature and consumption feature, parameter and consumption effect analysis module are used for the analysis of parameter effect and consumption effect, fusion diagnosis module is used to diagnose battery internal short circuit state, verification and optimization module is used to verify and optimize the evaluation method of battery internal short circuit.The present application proposes multi-source dynamic collaborative enhancement type battery internal short circuit monitoring algorithm to carry out feature extraction and feature fusion, proposes battery internal short circuit three-order collaborative analysis algorithm parameter effect analysis, proposes battery internal short circuit multi-scale dynamic coupling analysis algorithm consumption effect analysis, provides more optimal scheme for a kind of battery internal short circuit evaluation method.

Description

Technical Field

[0001] This invention relates to the fields of battery safety monitoring and multiphysics coupling analysis, specifically to a method for assessing short circuits within a battery. Background Technology

[0002] Battery safety monitoring technology refers to a technical system that achieves early warning and accurate diagnosis of short-circuit safety risks within batteries by real-time acquisition and analysis of multi-dimensional physical parameters. Its core objective is to identify micro-short-circuit signs before thermal runaway occurs, thus gaining a critical time window for safety protection. Battery safety monitoring technology includes traditional experimental simulation methods (such as mechanical puncture tests and overcharge / over-discharge tests), simulation modeling techniques (such as multiphysics coupling models), and early warning algorithms (such as those based on voltage / temperature thresholds and electrochemical impedance spectroscopy). The collaboration of these technologies provides a more efficient monitoring method for short circuits within batteries.

[0003] Multiphysics coupling analysis is an interdisciplinary approach that uses mathematical modeling and simulation to study the interactions of multiple physical processes (electric, thermal, and mechanical) within a battery. In battery short-circuit assessment, its core value lies in revealing the fault evolution mechanism, achieving precise location, and overcoming the limitations of single-field analysis. Multiphysics coupling analysis includes pure electrochemical models, thermal abuse models, electro-thermal coupling models, and thermo-mechanical coupling models. The collaboration of these technologies provides a more accurate analytical method for battery short-circuit assessment. Summary of the Invention

[0004] To address the aforementioned problems, this invention aims to provide a method for assessing internal short circuits in batteries.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for assessing internal short circuits in a battery, comprising a data acquisition module, a data preprocessing module, a feature extraction and fusion module, a parameter and consumption effect analysis module, a fusion diagnosis module, and a verification and optimization module. The data acquisition module is used to acquire high-precision electrical parameters, thermal parameters, and mechanical parameters. The data preprocessing module is used to preprocess the acquired data. The feature extraction and fusion module proposes a multi-source dynamic collaborative enhanced internal short circuit monitoring algorithm to extract and fuse parameter features and consumption features respectively. The parameter and consumption effect analysis module includes a parameter effect analysis unit and a consumption effect analysis unit. The parameter effect analysis unit proposes a third-order collaborative analysis algorithm for internal short circuits in the battery to analyze parameter effects. The consumption effect analysis unit proposes a multi-scale dynamic coupling analysis algorithm for internal short circuits in the battery to analyze consumption effects. The fusion diagnosis module is used to fuse the parameter effect analysis results and the consumption effect analysis results to diagnose the internal short circuit state of the battery. The verification and optimization module is used to verify and optimize the assessment method for internal short circuits in the battery.

[0006] Furthermore, the data acquisition module is used to collect high-precision electrical, thermal, and mechanical parameters. It acquires single-cell terminal voltage data through a 16-bit ADC, acquires current data through dual-mode detection using a shunt and Hall sensor, acquires impedance spectrum data through online EIS technology, acquires temperature data through an NTC thermistor, and acquires battery expansion force change data through micro-strain gauges mounted on the battery casing surface.

[0007] Furthermore, the data preprocessing module performs noise reduction preprocessing on the collected electrical parameters through wavelet packet transform, captures local abnormal heating preprocessing on the collected thermal parameters through spatial gradient calculation, and performs signal correction preprocessing on the mechanical parameters through temperature drift compensation.

[0008] Furthermore, the feature extraction and fusion module proposes a multi-source dynamic collaborative enhanced battery short-circuit monitoring algorithm to extract and fuse parameter features and consumption features respectively.

[0009] Furthermore, the multi-source dynamic collaborative enhanced battery short-circuit monitoring algorithm is as follows: Assume the collected multi-source data is... ,in, Electrical parameter data, i.e. ,in, This is the first electrical parameter data. This is the second electrical parameter data. For the first Electrical parameter data, For thermal parameter data, i.e. ,in, This is the first thermal parameter data. This is the second thermal parameter data. For the first thermal parameter data, For mechanical parameter data, i.e. ,in, This is the first mechanical parameter data. This is the second mechanical parameter data. For the first Mechanical parameter data, assuming a sampling period of [number] days. Then the sampled multi-source data is ,in, This is the multi-source data matrix of the sampled battery. For the first Electrical parameter data from the second sampling. For the first The thermal parameter data from the second sampling For the first To improve the signal-to-noise ratio of micro-short circuit transient features using sampled mechanical parameter data, a method is proposed to enhance the transient feature extraction capability by strengthening differential operations. This enhanced differential operation is... , ,in, For the data in the first Differential eigenvalues ​​of the next sample, The standardized coefficient is used to eliminate the influence of the sampling period. For difference operators, for Difference operations, For the first Electrical parameter data after the second sampling For the first Electrical parameter data after the second sampling For the first Electrical parameter data after the second sampling For the first To achieve adaptive feature weighting of the electrical parameter data after sampling, a SOH correlation function is proposed to enhance the contribution of key parameters. The multiphysics weighted energy is... , , ,in, These are weighted fusion eigenvalues ​​of multi-physics fields, used to comprehensively characterize the degree of anomaly in multi-physics field signals. These are the weighting coefficients for electrical parameters. These are the weighting coefficients for the thermal parameters. These are the weighting coefficients for the mechanical parameters. The values ​​of the three weighting coefficients vary depending on the input parameters. In order to be in The normalized rate of change of voltage at time In order to be in The normalized magnitude of the temperature gradient vector at time step. In order to be in The normalized rate of change of mechanical strain over time. For dynamic weighting coefficients, The slope of the weight decay. For the first To assess the health status of the next charge-discharge cycle and quantify signal disorder, a moving average is proposed instead of a fixed reference to reduce baseline drift caused by cycle aging. Therefore, multimodal data feature extraction is performed as follows: ,in, For the first Information entropy feature value of the second sample To adjust the sliding window size, The offset index within the window. For historical sampling data, The mean of the sliding window. To calculate the squared deviation of each data point from the window mean;

[0010] Then, feature fusion processing is performed. To capture multi-physics cooperative anomalies, a multi-modal data frequency domain coupling analysis is proposed to reduce computational load. The coupling power spectral density is... ,in, For the coupled power spectral density, This is the upper limit of the fault-sensitive frequency band. This is the lower limit of the fault-sensitive frequency band. For frequency, Multiphysics weighted fusion eigenvalues To enhance early signal characteristics, a differential operator is proposed, which modifies the fixed bandwidth to aging-related dynamic focusing to improve the spectral bandwidth's sensitivity to weak faults. The dynamic focusing bandwidth is... ,in, For dynamic focusing bandwidth, The center frequency of the spectrum, coefficient This is a scaling factor used to map dimensionless aging metrics to bandwidth adjustment. This is the normalized cumulative historical voltage change, i.e. ,in, For the first Characteristic voltage changes in each cycle. The algorithm for multi-source dynamic collaborative enhanced battery short-circuit monitoring, which takes the absolute value of voltage change as an example, first proposes to enhance the differential operation to improve the transient feature extraction capability. Then, it proposes the SOH correlation function to achieve feature weight adaptation. Next, it proposes to change the fixed benchmark to the moving average to reduce the baseline drift caused by cyclic aging. Finally, it proposes multi-modal data frequency domain coupling analysis and improves the fixed bandwidth to aging-related dynamic focusing for feature fusion, so as to realize the extraction and fusion of battery parameter features and consumption features.

[0011] Furthermore, the parameter effect analysis unit proposes a third-order collaborative analysis algorithm for battery internal short circuits to analyze parameter effects.

[0012] Furthermore, the third-order collaborative analysis algorithm for short circuits within the battery is as follows: To address the insensitivity of traditional low-order differential methods to weak signals, a third-order differential feature enhancement algorithm is proposed to improve sensitivity to weak signals. The third-order differential feature value is... ,in, It is the third-order differential eigenvalue, that is, the battery multi-source data at time t. The intensity of transient changes For the first Second sampling, The sampling period is These are the standardized coefficients for the third-order differential method. for The third-order differential operation, namely , For the first Electrical parameter data from the second sampling. For the first Multi-source battery data from the second sampling For the first Multi-source battery data from the second sampling For the first To address the issue of false alarms due to single parameters in multi-source battery data sampling, a multi-physics collaborative verification method is proposed to achieve short-circuit localization. The verification results are calculated as follows: ,in, For the verification results of the calculation, This is the basic threshold for the battery. In order to be in The normalized magnitude of the temperature gradient vector at time step. In order to be in The normalized mechanical strain rate of change at time, with thresholds of 2.0 and 5.0 as example values ​​for dimensionless parameters, requires experimental calibration for specific battery models and sensors in practical applications. To achieve full lifecycle threshold adjustment, an aging adaptive decision is proposed, namely... ,in, An adaptive threshold for aging. The aging sensitivity coefficient, The normalized cumulative voltage change over time is used to calculate the voltage decay rate after charging is completed. ,in, The attenuation rate factor, The moment when charging ends. The moment when relaxation ends. The rate of change of voltage. To determine the absolute value of the voltage change rate, a three-dimensional thermal model is constructed for three-dimensional thermal residual analysis. ,in, Along the length of the battery. In the direction of battery width, In the direction of battery thickness, For temperature residual, For actual measured temperature, The temperature rise coefficient has dimensions of . , This is the operating current. The spatial thermal resistance distribution is determined, and then online EIS monitoring of low-frequency impedance is performed. Characteristic time constants are identified using DRT, and impedance DRT analysis is used to calculate the following: , ,in, The characteristic time constant of the SEI film, For the measured impedance spectrum, It is a time constant. Let be the distributed relaxation time function. For the model impedance spectrum, It is the L2 norm. As a regularization coefficient, the third-order collaborative analysis algorithm for short circuits in batteries first proposes third-order differential feature enhancement to improve sensitivity to weak signals, then proposes multi-physics field collaborative verification to achieve short circuit location, then proposes aging adaptive decision to achieve threshold adjustment throughout the entire life cycle, and finally performs voltage decay rate calculation, three-dimensional thermal model construction, and characteristic time constant calculation through DRT identification to achieve accurate analysis of parameter effects.

[0013] Furthermore, the consumption effect analysis unit proposes a multi-scale dynamic coupling analysis algorithm for short circuits within the battery to analyze the consumption effect.

[0014] Furthermore, the specific algorithm for multi-scale dynamic coupling analysis of short circuits within the battery is as follows: the coulombic efficiency for each cycle is calculated as follows: , ,in, For the first Secondary cycle coulomb efficiency. For discharge capacity, For charging capacity, This is the capacity decay amount. This is the current full charge capacity. To determine the initial full charge capacity and identify active material loss, differential capacity calculations are performed. , ,in, For differential capacity, For the first A discrete capacity point, For the first A discrete capacity point, For the first Discrete voltage points, For the first Discrete voltage points, This represents the characteristic peak voltage offset. This is the peak voltage of the current cycle. Using the initial reference peak voltage and the ambient temperature under static conditions, the abnormal loss is calculated using the energy balance equation, i.e. , ,in, Self-discharge rate This is the start time of the settling process. This is the end of the settling period. The rate of change of voltage with respect to time, For energy loss, For equivalent leakage current, For electrical parameter data, the OCV-SOC relationship changes as follows: ,in, For the OCV curve offset integral, This is the current open-circuit voltage-SOC curve. For reference, the OCV-SOC curve, In the charging state, the multi-scale dynamic coupling analysis algorithm for short circuits in the battery first proposes a coulomb efficiency calculation method for capacity decay monitoring, then proposes differential capacity analysis to identify active material loss, then proposes abnormal loss calculation through energy balance equation, and finally proposes OCV-SOC relationship change to diagnose consumption effect, thereby achieving accurate analysis of consumption effect.

[0015] Furthermore, the fusion diagnostic module is used to fuse parameter effect analysis results and consumption effect analysis results to diagnose the internal short circuit state of the battery. First, a correlation matrix containing all characteristic parameters is constructed. An improved DS evidence theory is used to handle uncertain information, and decision fusion is achieved through fuzzy logic reasoning. The early warning system implements a three-level response mechanism: a level one early warning is triggered when a single parameter exceeds a threshold; a level two early warning is triggered when multiple parameters are abnormal; and a level three early warning is triggered only when both parameter and consumption abnormalities occur simultaneously. The verification and optimization module is used to verify and optimize the evaluation method for internal short circuits in the battery. The system continuously optimizes its performance through closed-loop verification. Under controlled laboratory conditions, internal short circuits of different severity are artificially created to verify the accuracy of the diagnostic algorithm. The accumulated fault cases will continuously enrich the sample library, gradually optimizing the evaluation method.

[0016] Compared with the prior art, the beneficial effects of the present invention are:

[0017] 1. This invention provides a method for assessing short circuits within a battery. A multi-source dynamic collaborative enhanced battery short circuit monitoring algorithm is proposed to extract and fuse parameter features and consumption features. The innovation of this invention lies in the following: First, the multi-source dynamic collaborative enhanced battery short circuit monitoring algorithm proposes to strengthen differential operations to enhance transient feature extraction capabilities. Then, it proposes a SOH correlation function to achieve adaptive feature weights. Next, it proposes changing the fixed benchmark to a moving average to reduce the baseline drift caused by cyclic aging. Finally, it proposes multi-modal data frequency domain coupling analysis and improves the fixed bandwidth to aging-related dynamic focusing for feature fusion, thereby achieving the extraction and fusion of battery parameter features and consumption features.

[0018] 2. A third-order collaborative analysis algorithm for short circuits within the battery is proposed to analyze parameter effects. The innovation of this invention lies in the fact that the third-order collaborative analysis algorithm for short circuits within the battery first proposes third-order differential feature enhancement to improve the sensitivity to weak signals, then proposes multi-physics field collaborative verification to achieve short circuit location, then proposes aging adaptive decision to achieve threshold adjustment throughout the entire life cycle, and finally performs voltage decay rate calculation, three-dimensional thermal model construction, and characteristic time constant calculation through DRT identification, thereby achieving accurate analysis of parameter effects.

[0019] 3. This invention proposes a multi-scale dynamic coupling analysis algorithm for short circuits within batteries to analyze the consumption effect. The innovation of this invention lies in the fact that the multi-scale dynamic coupling analysis algorithm for short circuits within batteries first proposes a coulomb efficiency calculation method for capacity decay monitoring, then proposes differential capacity analysis to identify active material loss, then proposes abnormal loss calculation through the energy balance equation, and finally proposes the change of OCV-SOC relationship to diagnose the consumption effect, thereby achieving accurate analysis of the consumption effect. Attached Figure Description

[0020] The invention will be further illustrated with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the invention. For those skilled in the art, other drawings can be obtained based on the following drawings without any creative effort.

[0021] Figure 1 This is a schematic diagram of the structure of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] A method for assessing internal short circuits in a battery includes a data acquisition module, a data preprocessing module, a feature extraction and fusion module, a parameter and consumption effect analysis module, a fusion diagnosis module, and a verification and optimization module. The data acquisition module collects high-precision electrical, thermal, and mechanical parameters. The data preprocessing module preprocesses the collected data. The feature extraction and fusion module proposes a multi-source dynamic collaborative enhanced internal short circuit monitoring algorithm to extract and fuse parameter features and consumption features. The parameter and consumption effect analysis module includes a parameter effect analysis unit and a consumption effect analysis unit. The parameter effect analysis unit proposes a third-order collaborative analysis algorithm for internal short circuits to analyze parameter effects, and the consumption effect analysis unit proposes a multi-scale dynamic coupling analysis algorithm for internal short circuits to analyze consumption effects. The fusion diagnosis module fuses the parameter effect analysis results and consumption effect analysis results to diagnose the internal short circuit state of the battery. The verification and optimization module verifies and optimizes the assessment method for internal short circuits in the battery.

[0024] Preferably, the data acquisition module is used to acquire high-precision electrical, thermal, and mechanical parameters. It acquires single-cell terminal voltage data through a 16-bit ADC (±0.1mV accuracy), acquires current data through dual-mode detection of a shunt and Hall sensor, acquires impedance spectrum data through online EIS technology, acquires temperature data through an NTC thermistor, and acquires battery expansion force change data through micro-strain gauges mounted on the battery casing surface.

[0025] Preferably, the data preprocessing module performs noise reduction preprocessing on the acquired electrical parameters through wavelet packet transform (using Db6 wavelet basis for 5-level decomposition, and soft thresholding of high-frequency coefficients to retain the effective frequency band of 0.01-1Hz), performs local abnormal heating capture preprocessing on the acquired thermal parameters through spatial gradient calculation, and performs signal correction preprocessing on the mechanical parameters through temperature drift compensation (eliminating spurious strain signal data).

[0026] Preferably, the feature extraction and fusion module proposes a multi-source dynamic collaborative enhanced battery short-circuit monitoring algorithm to extract and fuse parameter features and consumption features respectively.

[0027] Specifically, the multi-source dynamic collaborative enhanced battery short-circuit monitoring algorithm is as follows: Assume the collected multi-source data is... ,in, Electrical parameter data, i.e. ,in, This is the first electrical parameter data. This is the second electrical parameter data. For the first Electrical parameter data, For thermal parameter data, i.e. ,in, This is the first thermal parameter data. This is the second thermal parameter data. For the first thermal parameter data, For mechanical parameter data, i.e. ,in, This is the first mechanical parameter data. This is the second mechanical parameter data. For the first Mechanical parameter data, assuming a sampling period of [number] days. Then the sampled multi-source data is ,in, This is the multi-source data matrix of the sampled battery. For the first Electrical parameter data from the second sampling. For the first The thermal parameter data from the second sampling For the first To improve the signal-to-noise ratio of micro-short circuit transient features using sampled mechanical parameter data, a method is proposed to enhance the transient feature extraction capability by strengthening differential operations. This enhanced differential operation is... , ,in, For the data in the first Differential eigenvalues ​​of the next sample, The standardized coefficient is used to eliminate the influence of the sampling period. For difference operators, for Difference operations, For the first Electrical parameter data after the second sampling For the first Electrical parameter data after the second sampling For the first Electrical parameter data after the second sampling For the first The differential operations on the electrical, thermal, and mechanical parameter data after the second sampling are consistent with the differential operations on the electrical parameters. To achieve adaptive feature weights, a SOH correlation function is proposed to enhance the contribution of key parameters. The multiphysics weighted energy is... , , ,in, These are weighted fusion eigenvalues ​​of multi-physics fields, used to comprehensively characterize the degree of anomaly in multi-physics field signals. These are the weighting coefficients for electrical parameters. These are the weighting coefficients for the thermal parameters. These are the weighting coefficients for the mechanical parameters. The values ​​of the three weighting coefficients vary depending on the input parameters. In order to be in The normalized rate of change of voltage at time In order to be in The normalized magnitude of the temperature gradient vector at time step. In order to be in The normalized rate of change of mechanical strain over time. These are dynamic weighting coefficients. In practical applications, the dimensional scaling factors need to be calibrated based on experimental data to balance the contributions of each physical field. The slope of the weight decay. For the first To assess the health status of the next charge-discharge cycle and quantify signal disorder, a moving average is proposed instead of a fixed reference to reduce baseline drift caused by cycle aging. Therefore, multimodal data feature extraction is performed as follows: ,in, For the first Information entropy feature value of the second sample To adjust the sliding window size, The offset index within the window. For historical sampling data, The mean of the sliding window. To calculate the squared deviation of each data point from the window mean;

[0028] Then, feature fusion processing is performed. To capture multi-physics cooperative anomalies, a multi-modal data frequency domain coupling analysis is proposed to reduce computational load. The coupling power spectral density is... ,in, For the coupled power spectral density, This is the upper limit of the fault-sensitive frequency band. This is the lower limit of the fault-sensitive frequency band. For frequency, Multiphysics weighted fusion eigenvalues To enhance early signal characteristics, a differential operator is proposed, which modifies the fixed bandwidth to aging-related dynamic focusing to improve the spectral bandwidth's sensitivity to weak faults. The dynamic focusing bandwidth is... ,in, For dynamic focusing bandwidth, The center frequency of the spectrum, coefficient This is a scaling factor, and its unit is already [missing information]. This is used to map dimensionless aging metrics to bandwidth adjustment amounts. This is the normalized cumulative historical voltage change, i.e. ,in, For the first Characteristic voltage changes in each cycle. The algorithm for multi-source dynamic collaborative enhanced battery short-circuit monitoring, which takes the absolute value of voltage change as an example, first proposes to enhance the differential operation to improve the transient feature extraction capability. Then, it proposes the SOH correlation function to achieve feature weight adaptation. Next, it proposes to change the fixed benchmark to the moving average to reduce the baseline drift caused by cyclic aging. Finally, it proposes multi-modal data frequency domain coupling analysis and improves the fixed bandwidth to aging-related dynamic focusing for feature fusion, so as to realize the extraction and fusion of battery parameter features and consumption features.

[0029] Preferably, the parameter effect analysis unit proposes a third-order collaborative analysis algorithm for short circuits within the battery to analyze parameter effects.

[0030] Specifically, the third-order collaborative analysis algorithm for battery short circuits is as follows: To address the problem of traditional low-order differential methods being insensitive to weak signals, a third-order differential feature enhancement algorithm is proposed to improve sensitivity to weak signals. The third-order differential feature value is... ,in, It is the third-order differential eigenvalue, that is, the battery multi-source data at time t. The intensity of transient changes For the first Second sampling, The sampling period is These are the standardized coefficients for the third-order differential method. for The third-order differential operation, namely , For the first Electrical parameter data from the second sampling. For the first Multi-source battery data from the second sampling For the first Multi-source battery data from the second sampling For the first To address the issue of false alarms due to single parameters in multi-source battery data sampling, a multi-physics collaborative verification method is proposed to achieve short-circuit localization. The verification results are calculated as follows: ,in, For the verification results of the calculation, This is the basic threshold for the battery. In order to be in The normalized magnitude of the temperature gradient vector at time step. In order to be in The normalized mechanical strain rate of change at time, with thresholds of 2.0 and 5.0 as example values ​​for dimensionless parameters, requires experimental calibration for specific battery models and sensors in practical applications. To achieve full lifecycle threshold adjustment, an aging adaptive decision is proposed, namely... ,in, An adaptive threshold for aging. The aging sensitivity coefficient, The normalized cumulative voltage change over time is used to calculate the voltage decay rate after charging is completed. ,in, The attenuation rate factor, The moment when charging ends. The moment when relaxation ends. The rate of change of voltage. To determine the absolute value of the voltage change rate, a three-dimensional thermal model is constructed for three-dimensional thermal residual analysis. ,in, The length direction of the battery (the axis from the negative terminal to the positive terminal). The width direction of the battery (the horizontal plane parallel to the electrode). The thickness direction of the battery (vertical axis from the bottom of the casing to the top cover). For temperature residual, For actual measured temperature, This is the operating current. The spatial thermal resistance distribution is determined, and then online EIS monitoring of low-frequency impedance is performed. Characteristic time constants are identified using DRT, and impedance DRT analysis is used to calculate the following: , ,in, The characteristic time constant of the SEI film, For the measured impedance spectrum, It is a time constant. Let be the distributed relaxation time function. For the model impedance spectrum, It is the L2 norm. As a regularization coefficient, the third-order collaborative analysis algorithm for short circuits in batteries first proposes third-order differential feature enhancement to improve sensitivity to weak signals, then proposes multi-physics field collaborative verification to achieve short circuit location, then proposes aging adaptive decision to achieve threshold adjustment throughout the entire life cycle, and finally performs voltage decay rate calculation, three-dimensional thermal model construction, and characteristic time constant calculation through DRT identification to achieve accurate analysis of parameter effects.

[0031] Preferably, the consumption effect analysis unit proposes a multi-scale dynamic coupling analysis algorithm for short circuits within the battery to analyze the consumption effect.

[0032] Specifically, the multi-scale dynamic coupling analysis algorithm for short circuits within the battery is as follows: the coulombic efficiency for each cycle is calculated as follows: , ,in, For the first Secondary cycle coulomb efficiency. For discharge capacity, For charging capacity, This is the capacity decay amount. This is the current full charge capacity. To determine the initial full charge capacity and identify active material loss, differential capacity calculations are performed. , ,in, For differential capacity, For the first A discrete capacity point, For the first A discrete capacity point, For the first Discrete voltage points, For the first Discrete voltage points, This represents the characteristic peak voltage offset. This is the peak voltage of the current cycle. Using the initial reference peak voltage and the ambient temperature under static conditions, the abnormal loss is calculated using the energy balance equation, i.e. , ,in, Self-discharge rate This is the start time of the settling process. This is the end of the settling period. The rate of change of voltage with respect to time, For energy loss, For equivalent leakage current, For electrical parameter data, the OCV-SOC relationship changes as follows: ,in, For the OCV curve offset integral, This is the current open-circuit voltage-SOC curve. For reference, the OCV-SOC curve, In the charging state, the multi-scale dynamic coupling analysis algorithm for short circuits in the battery first proposes a coulomb efficiency calculation method for capacity decay monitoring, then proposes differential capacity analysis to identify active material loss, then proposes abnormal loss calculation through energy balance equation, and finally proposes OCV-SOC relationship change to diagnose consumption effect, thereby achieving accurate analysis of consumption effect.

[0033] Preferably, the fusion diagnostic module is used to diagnose the internal short circuit state of the battery by fusing the results of parameter effect analysis and consumption effect analysis. First, a correlation matrix containing all characteristic parameters is constructed. An improved DS evidence theory is used to process uncertain information. Decision fusion is achieved through fuzzy logic reasoning. The early warning system implements a three-level response mechanism: a level one early warning is triggered when a single parameter exceeds a threshold; a level two early warning is triggered when multiple parameters are abnormal; and a level three early warning is triggered only when both parameter abnormalities and consumption abnormalities (such as voltage drop accompanied by abnormal self-discharge) occur simultaneously. The verification and optimization module is used to verify and optimize the evaluation method for internal short circuits in the battery. The system continuously optimizes its performance through closed-loop verification. Under controlled laboratory conditions, internal short circuits of different severity are artificially created to verify the accuracy of the diagnostic algorithm. The accumulated fault cases will continuously enrich the sample library, gradually optimizing the evaluation method.

[0034] This invention proposes a method for assessing internal short circuits in batteries, enabling early warning and accurate diagnosis of internal short circuits. Through the integration of data acquisition, data preprocessing, feature extraction and fusion, parameter and consumption effect analysis, fusion diagnosis, and verification and optimization modules, a method for assessing internal short circuits in batteries is provided. A multi-source dynamic collaborative enhanced internal short circuit monitoring algorithm is proposed to extract and fuse parameter and consumption features separately. The innovation of this invention lies in the following: First, the multi-source dynamic collaborative enhanced internal short circuit monitoring algorithm first proposes strengthening differential operations to enhance transient feature extraction capabilities; then, it proposes a SOH correlation function to achieve adaptive feature weights; then, it proposes changing the fixed benchmark to a moving average to reduce the baseline drift caused by cyclic aging; then, it proposes multi-modal data frequency domain coupling analysis and improves the fixed bandwidth to aging-related dynamic focusing for feature fusion, realizing the extraction and fusion of battery parameter and consumption features. A third-order collaborative analysis algorithm for internal short circuits in batteries is proposed to analyze parameter effects. The innovation of this invention lies in the following: The third-order collaborative analysis algorithm for internal short circuits in batteries first proposes third-order differential feature enhancement to improve sensitivity to weak signals; then, it proposes... This invention employs multi-physics field collaborative verification to locate short circuits, then proposes aging adaptive decision-making to adjust thresholds throughout the battery lifecycle, and finally calculates voltage decay rate, constructs a three-dimensional thermal model, and calculates characteristic time constants through DRT identification to achieve accurate analysis of parameter effects. A multi-scale dynamic coupling analysis algorithm for battery short circuits is proposed to analyze consumption effects. The innovation of this invention lies in its approach: firstly, it proposes a coulomb efficiency calculation method for capacity decay monitoring; secondly, it proposes differential capacity analysis to identify active material loss; thirdly, it proposes abnormal loss calculation through energy balance equations; and finally, it proposes changes in the OCV-SOC relationship to diagnose consumption effects. This achieves accurate analysis of consumption effects, effectively improving the performance of a battery short circuit assessment method, providing more comprehensive and accurate technical support for battery short circuit assessment methods, and offering better decision support for scientific and efficient battery short circuit assessment methods. Furthermore, this invention relates to battery safety monitoring and multi-physics field coupling analysis technology, providing an accurate and efficient battery short circuit assessment method and contributing significant application value to battery short circuit assessment.

[0035] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A system for assessing short circuits within a battery, characterized in that, The system includes a data acquisition module, a data preprocessing module, a feature extraction and fusion module, a parameter and consumption effect analysis module, a fusion diagnosis module, and a verification and optimization module. The data acquisition module collects high-precision electrical, thermal, and mechanical parameters. The data preprocessing module preprocesses the collected data from various sources. The feature extraction and fusion module proposes a multi-source dynamic collaborative enhanced battery internal short-circuit monitoring algorithm to extract and fuse parameter and consumption features. The specific multi-source dynamic collaborative enhanced battery internal short-circuit monitoring algorithm is as follows: Assuming the collected multi-source data is... ,in, Electrical parameter data, i.e. ,in, This is the first electrical parameter data. This is the second electrical parameter data. For the first Electrical parameter data, For thermal parameter data, i.e. ,in, This is the first thermal parameter data. This is the second thermal parameter data. For the first thermal parameter data, For mechanical parameter data, i.e. ,in, This is the first mechanical parameter data. This is the second mechanical parameter data. For the first Mechanical parameter data, assuming a sampling period of [number] days. Then the sampled multi-source data is ,in, This is the multi-source data matrix of the sampled battery. For the first Electrical parameter data from the second sampling. For the first The thermal parameter data from the second sampling For the first To improve the signal-to-noise ratio of micro-short circuit transient features using sampled mechanical parameter data, a method is proposed to enhance the transient feature extraction capability by strengthening differential operations. This enhanced differential operation is... , ,in, For the data in the first Differential eigenvalues ​​of the next sample, The standardized coefficient is used to eliminate the influence of the sampling period. For difference operators, for Difference operations, For the first Electrical parameter data after the second sampling For the first Electrical parameter data after the second sampling For the first Electrical parameter data after the second sampling For the first To achieve adaptive feature weighting of the electrical parameter data after sampling, a SOH correlation function is proposed to enhance the contribution of key parameters. The multiphysics weighted energy is... , , ,in, These are weighted fusion eigenvalues ​​of multi-physics fields, used to comprehensively characterize the degree of anomaly in multi-physics field signals. These are the weighting coefficients for electrical parameters. These are the weighting coefficients for the thermal parameters. These are the weighting coefficients for the mechanical parameters. The values ​​of the three weighting coefficients vary depending on the input parameters. In order to be in The normalized rate of change of voltage at time In order to be in The normalized magnitude of the temperature gradient vector at time step. In order to be in The normalized rate of change of mechanical strain over time. For dynamic weighting coefficients, The slope of the weight decay. For the first To assess the health status of the next charge-discharge cycle and quantify signal disorder, a moving average is proposed instead of a fixed reference to reduce baseline drift caused by cycle aging. Therefore, multimodal data feature extraction is performed as follows: ,in, For the first Information entropy feature value of the second sample To adjust the sliding window size, The offset index within the window. For historical sampling data, The mean of the sliding window. To calculate the squared deviation of each data point from the window mean; Then, feature fusion processing is performed. To capture multi-physics cooperative anomalies, a multi-modal data frequency domain coupling analysis is proposed to reduce computational load. The coupling power spectral density is... ,in, For the coupled power spectral density, This is the upper limit of the fault-sensitive frequency band. This is the lower limit of the fault-sensitive frequency band. For frequency, Multiphysics weighted fusion eigenvalues To enhance early signal characteristics, a differential operator is proposed, which modifies the fixed bandwidth to aging-related dynamic focusing to improve the spectral bandwidth's sensitivity to weak faults. The dynamic focusing bandwidth is... ,in, For dynamic focusing bandwidth, The center frequency of the spectrum, coefficient This is a scaling factor used to map dimensionless aging metrics to bandwidth adjustment. This is the normalized cumulative historical voltage change, i.e. ,in, For the first Characteristic voltage changes in each cycle. The algorithm for multi-source dynamic collaborative enhanced battery short circuit monitoring, which takes the absolute value of voltage change as an example, first proposes to enhance the differential operation to improve the transient feature extraction capability, then proposes the SOH correlation function to achieve feature weight adaptation, then proposes to change the fixed benchmark to the moving average to reduce the baseline drift caused by cyclic aging, and then proposes multi-modal data frequency domain coupling analysis and improves the fixed bandwidth to aging-related dynamic focusing to perform feature fusion, thereby realizing the extraction and fusion of battery parameter features and consumption features. The parameter and consumption effect analysis module includes a parameter effect analysis unit and a consumption effect analysis unit. The parameter effect analysis unit proposes a third-order collaborative analysis algorithm for battery short circuits to analyze parameter effects. The specific algorithm is as follows: To address the insensitivity of traditional low-order differential methods to weak signals, a third-order differential feature enhancement algorithm is proposed to improve sensitivity to weak signals. The third-order differential feature value is... ,in, It is the third-order differential eigenvalue, that is, the battery multi-source data at time t. The intensity of transient changes For the first Second sampling, The sampling period is These are the standardized coefficients for the third-order differential method. for The third-order differential operation, namely , For the first Electrical parameter data from the second sampling. For the first Multi-source battery data from the second sampling. For the first Multi-source battery data from the second sampling For the first To address the issue of false alarms due to single parameters in multi-source battery data sampling, a multi-physics collaborative verification method is proposed to achieve short-circuit localization. The verification results are calculated as follows: ,in, For the verification results of the calculation, This is the basic threshold for the battery. In order to be in The normalized magnitude of the temperature gradient vector at time step. In order to be in The normalized mechanical strain rate of change at time, with thresholds of 2.0 and 5.0 as example values ​​for dimensionless parameters, requires experimental calibration for specific battery models and sensors in practical applications. To achieve full lifecycle threshold adjustment, an aging adaptive decision is proposed, namely... ,in, An adaptive threshold for aging. The aging sensitivity coefficient, The normalized cumulative voltage change over time is used to calculate the voltage decay rate after charging is completed. ,in, The attenuation rate factor, The moment when charging ends. The moment when relaxation ends. The rate of change of voltage. To determine the absolute value of the voltage change rate, a three-dimensional thermal model is constructed for three-dimensional thermal residual analysis. ,in, Along the length of the battery. In the direction of battery width, In the direction of battery thickness, For temperature residual, For actual measured temperature, The temperature rise coefficient has dimensions of . , This is the operating current. The spatial thermal resistance distribution is determined, and then online EIS monitoring of low-frequency impedance is performed. Characteristic time constants are identified using DRT, and impedance DRT analysis is used to calculate the following: , ,in, The characteristic time constant of the SEI film, For the measured impedance spectrum, It is a time constant. Let be the distributed relaxation time function. For the model impedance spectrum, It is the L2 norm. As the regularization coefficient, the third-order collaborative analysis algorithm for short circuits in the battery first proposes third-order differential feature enhancement to improve the sensitivity to weak signals, then proposes multi-physics field collaborative verification to achieve short circuit location, then proposes aging adaptive decision to achieve threshold adjustment throughout the entire life cycle, and finally performs voltage decay rate calculation, three-dimensional thermal model construction, and characteristic time constant calculation through DRT identification to achieve accurate analysis of parameter effects. The battery consumption effect analysis unit proposes a multi-scale dynamic coupling analysis algorithm for battery short circuits to analyze the consumption effect. The specific algorithm is as follows: the coulombic efficiency for each cycle is calculated as follows: , ,in, For the first Secondary cycle coulomb efficiency. For discharge capacity, For charging capacity, This is the capacity decay amount. This is the current full charge capacity. To determine the initial full charge capacity and identify active material loss, differential capacity calculations are performed. , ,in, For differential capacity, For the first A discrete capacity point, For the first A discrete capacity point, For the first Discrete voltage points, For the first Discrete voltage points, This represents the characteristic peak voltage offset. This is the peak voltage of the current cycle. Using the initial reference peak voltage and the ambient temperature under static conditions, the abnormal loss is calculated using the energy balance equation, i.e. , ,in, Self-discharge rate This is the start time of the settling process. This is the end of the settling period. The rate of change of voltage with respect to time, For energy loss, For equivalent leakage current, For electrical parameter data, the OCV-SOC relationship changes as follows: ,in, For the OCV curve offset integral, This is the current open-circuit voltage-SOC curve. For reference, the OCV-SOC curve, For charging state, the multi-scale dynamic coupling analysis algorithm for battery internal short circuit first proposes a coulombic efficiency calculation method for capacity decay monitoring, then proposes differential capacity analysis to identify active material loss, followed by abnormal loss calculation using the energy balance equation, and finally proposes OCV-SOC relationship changes to diagnose consumption effects, thereby achieving accurate analysis of consumption effects; the fusion diagnosis module is used to fuse parameter effect analysis results and consumption effect analysis results to diagnose the battery internal short circuit state, and the verification and optimization module is used to verify and optimize the battery internal short circuit evaluation method.

2. The system for assessing internal short circuits in a battery according to claim 1, characterized in that, The data acquisition module is used to collect high-precision electrical, thermal, and mechanical parameters. It acquires single-cell terminal voltage data through a 16-bit ADC, current data through dual-mode detection of a shunt and Hall sensor, impedance spectrum data through online EIS technology, temperature data through an NTC thermistor, and battery expansion force change data through micro-strain gauges mounted on the battery casing surface.

3. The system for assessing internal short circuits in a battery according to claim 1, characterized in that, The data preprocessing module performs noise reduction preprocessing on the collected electrical parameters through wavelet packet transform, captures local abnormal heating of the collected thermal parameters through spatial gradient calculation, and performs signal correction preprocessing on the mechanical parameters through temperature drift compensation.

4. The system for assessing internal short circuits in a battery according to any one of claims 1-3, and the method for assessing short circuits in a battery, characterized in that, The fusion diagnostic module is used to diagnose the internal short circuit state of the battery by fusing the results of parameter effect analysis and consumption effect analysis. First, a correlation matrix containing all characteristic parameters is constructed. An improved DS evidence theory is used to handle uncertain information. Decision fusion is achieved through fuzzy logic reasoning. The early warning system implements a three-level response mechanism: a level one early warning is triggered when a single parameter exceeds a threshold; a level two early warning is triggered when multiple parameters are abnormal; and a level three early warning is triggered only when both parameter and consumption abnormalities occur simultaneously. The verification and optimization module is used to verify and optimize the evaluation method for internal short circuits in the battery.

Images