Intelligent Battery Management Method, Device, Equipment and Storage Medium

Abstract

The present invention relates to the technical field of battery management, and discloses an intelligent battery management method, device, equipment and storage medium. The method includes: through multi-source data acquisition and feature separation of the first sensing parameter and the second sensing parameter of the target intelligent battery, obtaining the basic state feature, and performing frequency domain decomposition and spatio-temporal reconstruction on the basic state feature to obtain battery multi-modal data; furthermore, through multi-level combined feature extraction and fusion, realizing the deep coupling of electrochemical features and structural features, obtaining a fusion feature vector, and performing separation of two sensing parameter channels, calculation of dual-channel weights and prediction of multi-target time series on the fusion feature vector to obtain multi-dimensional state evaluation indicators, obtaining the comprehensive state evaluation indicator of the battery, so as to determine the corresponding fault risk level based on the comprehensive state evaluation indicator, and finally generating the battery management result of the target intelligent battery. The accuracy and reliability of the state evaluation and fault diagnosis of the intelligent battery are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and particularly to an intelligent battery management method, device, equipment and storage medium. Background Art

[0002] In the field of intelligent battery technology, the safety and reliability assessment of batteries is an important link in R & D and quality control, and the real-time monitoring and fault diagnosis of battery states are the most challenging stages. Accurately evaluating the operating state of a battery pack is crucial for manufacturing enterprises, R & D institutions and testing centers, which directly affects the performance optimization, life prediction and safety management of the battery.

[0003] Currently, the commonly used battery management methods mainly rely on the monitoring and analysis of traditional electrochemical parameters. For example, fault models are established by collecting data such as voltage, current and temperature, or machine learning techniques are used to evaluate the battery state. However, these methods still face challenges in dealing with multi-dimensional parameter fusion, non-linear feature extraction and dynamic response analysis. In particular, the evolution characteristics of the internal structure of the battery are ignored, such as the monitoring of non-traditional parameters such as mechanical stress changes, material interface evolution and structural integrity. These factors have an important impact on battery state assessment and fault diagnosis. That is, the existing battery management methods fail to make full use of the collaborative analysis of multi-source sensing data, resulting in low accuracy and reliability of state assessment and fault diagnosis. Summary of the Invention

[0004] The main object of the present invention is to solve the problem that the existing battery management methods fail to make full use of the collaborative analysis of multi-source sensing data, resulting in low accuracy and reliability of state assessment and fault diagnosis.

[0005] In the first aspect of the present invention, an intelligent battery management method is provided. The intelligent battery management method includes: collecting first multi-source original sensing parameter data corresponding to a first sensing parameter and second multi-source original sensing parameter data corresponding to a second sensing parameter of a target intelligent battery, and respectively performing spatio-temporal correlation and multiple preprocessings on the first multi-source original sensing parameter data and the second multi-source original sensing parameter data to obtain first multi-modal sensing data and second multi-modal sensing data; extracting multi-level combined features from the first multi-modal sensing data and the second multi-modal sensing data to obtain a multi-battery parameter combined feature set, and performing multi-dimensional feature fusion on each of the battery parameter combined feature sets to obtain a fusion feature vector of the intelligent battery state; separating the fusion feature vector into a first sensing parameter channel and a second sensing parameter channel to obtain an electrochemical state feature stream and a structural evolution feature stream, and calculating cross-attention weights of the two channels and predicting multi-objective time series for the electrochemical state feature stream and the structural evolution feature stream to obtain multi-dimensional state evaluation indexes of the target intelligent battery; performing time series anomaly detection and determination of abnormal spatial distribution on the multi-dimensional state evaluation indexes to obtain a multi-dimensional fault feature vector, and calculating a propagation path and dividing a risk level for the multi-dimensional fault feature vector to obtain a fault risk level of the target intelligent battery; performing safety threshold classification calculation and conversion of management instructions on the fault risk level and the multi-dimensional state evaluation indexes to generate a battery management result of the target intelligent battery.

[0006] Optionally, in the first implementation manner of the first aspect of the present invention, the step of respectively performing spatio-temporal correlation and multiple preprocessings on the first multi-source original sensing parameter data and the second multi-source original sensing parameter data to obtain first multi-modal sensing data and second multi-modal sensing data includes: segmenting the first multi-source original sensing parameter data by frequency to obtain an electrochemical low-frequency data segment, an electrochemical medium-frequency data segment, and an electrochemical high-frequency data segment, and segmenting the second multi-source original sensing parameter data by frequency to obtain a mechanical low-frequency data segment, a mechanical medium-frequency data segment, and a mechanical high-frequency data segment; performing Kalman filtering on the electrochemical low-frequency data segment and the mechanical low-frequency data segment to obtain state basic features, performing wavelet multi-scale decomposition on the electrochemical medium-frequency data segment and the mechanical medium-frequency data segment to obtain dynamic response features, and performing empirical mode decomposition on the electrochemical high-frequency data segment and the mechanical high-frequency data segment to obtain transient evolution features; performing spatio-temporal synchronous matching on the state basic features, the dynamic response features, and the transient evolution features to obtain initial multi-modal sensing data, and performing signal-to-noise ratio evaluation and weighted calculation of the corresponding original sensing parameters on the initial multi-modal sensing data to obtain first multi-modal sensing data and second multi-modal sensing data.

[0007] Optionally, in the second implementation manner of the first aspect of the present invention, the extraction of multi-level combined features from the first multi-modal sensing data and the second multi-modal sensing data to obtain a multi-battery parameter combined feature set, and the multi-dimensional feature fusion of each of the battery parameter combined feature sets to obtain a fusion feature vector of the intelligent battery state includes: performing hierarchical feature extraction on the first multi-modal sensing data to obtain a voltage-current feature set and a temperature-internal resistance feature set, and performing hierarchical feature extraction on the second multi-modal sensing data to obtain a vibration-acoustic emission feature set and an ultrasonic displacement feature set; performing time-series pairing on the voltage-current feature set and the vibration-acoustic emission feature set to obtain a multi-dimensional feature relationship matrix, and performing cross-correlation calculation and non-linear mapping on the multi-dimensional feature relationship matrix to obtain a first battery parameter combined feature set; performing spatial registration on the temperature-internal resistance feature set and the ultrasonic displacement feature set to obtain a structural parameter relationship matrix, and performing time-frequency feature mapping on the structural parameter relationship matrix to obtain a second battery parameter combined feature set; performing multi-level wavelet packet decomposition on the first battery parameter combined feature set to obtain multi-scale battery reaction stress features, and performing empirical mode decomposition on the second battery parameter combined feature set to obtain multi-scale thermal field deformation features, and performing energy reconstruction and feature selection on the multi-scale battery reaction stress features to obtain electrochemically strain-coupled features, and performing mode screening and feature fusion on the multi-scale thermal field deformation features to obtain temperature-deformation coupled features; performing hierarchical contribution degree calculation on the electrochemically strain-coupled features and the temperature-deformation coupled features to obtain a multi-level coupling weight matrix, and based on the multi-level coupling weight matrix, performing weighted fusion on the electrochemically strain-coupled features and the temperature-deformation coupled features to obtain a fusion feature vector of the intelligent battery state.

[0008] Optionally, in the third implementation manner of the first aspect of the present invention, the separation of the fusion feature vector into a first sensing parameter channel and a second sensing parameter channel to obtain an electrochemical state feature stream and a structural evolution feature stream includes: calculating the attention weight of the fusion feature vector corresponding to the first sensing parameter, extracting the initial electrochemical feature, and performing feature dimension compression and electrochemical feature reconstruction on the initial electrochemical feature to obtain an electrochemical feature subset of the target intelligent battery; extracting double-branch temporal convolution features from the electrochemical feature subset to obtain an electrochemical response matrix, and performing temporal reconstruction and performance feature enhancement on the electrochemical response matrix to obtain an electrochemical state feature stream; calculating the local similarity of the fusion feature vector corresponding to the second sensing parameter, extracting the initial structural feature of the corresponding feature component, and performing structural feature reconstruction and screening of key structural features on the initial structural feature to obtain a structural evolution feature subset of the target intelligent battery; extracting double-branch spatial convolution features from the structural evolution feature subset to obtain a structural response matrix, and performing spatial reconstruction and feature enhancement on the structural response matrix to obtain a structural evolution feature stream.

[0009] Optionally, in the fourth implementation manner of the first aspect of the present invention, the calculation of the cross-attention weight of the two channels and the prediction of the multi-objective time series of the electrochemical state feature stream and the structural evolution feature stream to obtain the multi-dimensional state evaluation index of the target intelligent battery includes: calculating the self-attention weight in the time series dimension for the corresponding charge-discharge feature branch and polarization feature branch in the electrochemical state feature stream to obtain an electrochemically enhanced feature, and calculating the self-attention weight in the spatial dimension for the corresponding stress feature branch and deformation feature branch in the structural evolution feature stream to obtain a structurally enhanced feature; performing cross-attention mapping calculation on the electrochemically enhanced feature and the structurally enhanced feature in the two channels to obtain a feature weight mapping matrix, and performing sliding time series segmentation and prediction of the battery performance trend on the feature weight mapping matrix to obtain a state evolution sequence; based on a preset branch evaluation model, evaluating the multi-branch state of the state evolution sequence to obtain an initial battery state evaluation value, and performing confidence calibration and index optimization on the initial battery state evaluation value to obtain the multi-dimensional state evaluation index of the target intelligent battery.

[0010] Optionally, in the fifth implementation manner of the first aspect of the present invention, the time series anomaly detection and determination of the abnormal spatial distribution of the multi-dimensional state evaluation index are performed to obtain a multi-dimensional fault feature vector, and the propagation path calculation and risk level division of the multi-dimensional fault feature vector are performed to obtain the fault risk level of the target intelligent battery, including: calculating the deviation of preset multi-monitoring parameters for the multi-dimensional state evaluation index to obtain a state change sequence, and classifying and abnormally positioning the state change sequence according to multi-level battery abnormal working thresholds to obtain an abnormal state sequence; performing spatial distribution mapping and correlation feature extraction on the abnormal state sequence to obtain a multi-dimensional fault feature vector, and constructing a fault correlation network for the multi-dimensional fault feature vector to obtain an initial propagation path matrix; calculating the diffusion rate and propagation direction of the initial propagation path matrix to obtain fault dynamic evolution parameters, and performing spatio-temporal sequence reconstruction on the fault dynamic evolution parameters to obtain a battery fault propagation matrix; calculating multi-dimensional risk indexes for the fault propagation matrix to obtain a fault risk score, and performing level division and priority sorting on the fault risk score to obtain the fault risk level of the target intelligent battery.

[0011] Optionally, in the sixth implementation manner of the first aspect of the present invention, the safety threshold grading calculation and management instruction conversion of the fault risk level and the multi-dimensional state evaluation index are performed to generate the battery management result of the target intelligent battery, including: performing safety factor mapping and working interval conversion on the fault risk level and the multi-dimensional state evaluation index to obtain battery working constraint parameters; performing charge and discharge parameter conversion of the corresponding power limit coefficient on the battery working constraint parameters to obtain a power management instruction, and performing calculation of the cooling channel flow distribution and heat dissipation power distribution corresponding to the temperature threshold on the battery working constraint parameters to obtain a temperature management instruction, and performing calculation of the active balancing circuit parameters corresponding to the state difference of the single battery on the battery working constraint parameters to obtain an equalization management instruction; performing priority sorting and battery state regulation on the power management instruction, the temperature management instruction and the equalization management instruction to generate the battery management result of the target intelligent battery.

[0012] In a second aspect of the present invention, an intelligent battery management device is provided. The intelligent battery management device includes: a preprocessing module configured to collect first multi-source raw sensing parameter data corresponding to a first sensing parameter and second multi-source raw sensing parameter data corresponding to a second sensing parameter of a target intelligent battery, and perform spatio-temporal correlation and multiple preprocessings on the first multi-source raw sensing parameter data and the second multi-source raw sensing parameter data respectively to obtain first multi-modal sensing data and second multi-modal sensing data; a feature fusion module configured to extract multi-level combined features from the first multi-modal sensing data and the second multi-modal sensing data to obtain a multi-battery parameter combination feature set, and perform multi-dimensional feature fusion on each of the battery parameter combination feature sets to obtain a fusion feature vector of the intelligent battery state; an index evaluation module configured to separate the fusion feature vector into a first sensing parameter channel and a second sensing parameter channel to obtain an electrochemical state feature stream and a structural evolution feature stream, and perform calculation of cross-attention weights for the two channels and prediction of multi-objective time series on the electrochemical state feature stream and the structural evolution feature stream to obtain multi-dimensional state evaluation indexes of the target intelligent battery; a risk division module configured to perform time series anomaly detection and determination of abnormal spatial distribution on the multi-dimensional state evaluation indexes to obtain a multi-dimensional fault feature vector, and perform calculation of propagation paths and division of risk levels on the multi-dimensional fault feature vector to obtain a fault risk level of the target intelligent battery; an instruction conversion module configured to perform safety threshold classification calculation and conversion of management instructions on the fault risk level and the multi-dimensional state evaluation indexes to generate a battery management result of the target intelligent battery.

[0013] In a third aspect of the present invention, an intelligent battery management device is provided, including: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor invokes the instructions in the memory so that the intelligent battery management device executes each step of the above-mentioned intelligent battery management method.

[0014] In a fourth aspect of the present invention, a computer-readable storage medium is provided, wherein instructions are stored in the computer-readable storage medium, and when the instructions are run on a computer, the computer is made to execute each step of the above-mentioned intelligent battery management method.

[0015] The above intelligent battery management method, device, equipment and storage medium. In the embodiments of the present invention, by collecting the first multi-source original sensing parameter data corresponding to the first sensing parameter of the target intelligent battery and the second multi-source original sensing parameter data corresponding to the second sensing parameter, and respectively performing spatio-temporal correlation and multiple preprocessings on the first multi-source original sensing parameter data and the second multi-source original sensing parameter data, the first multi-modal sensing data and the second multi-modal sensing data are obtained; extracting multi-level combined features from the first multi-modal sensing data and the second multi-modal sensing data to obtain a multi-battery parameter combined feature set, and performing multi-dimensional feature fusion on each battery parameter combined feature set to obtain a fusion feature vector of the intelligent battery state; separating the fusion feature vector into a first sensing parameter channel and a second sensing parameter channel to obtain an electrochemical state feature stream and a structural evolution feature stream, and calculating the cross-attention weights of the two channels and predicting multi-objective time series for the electrochemical state feature stream and the structural evolution feature stream to obtain multi-dimensional state evaluation indexes of the target intelligent battery; performing time series anomaly detection and determination of abnormal spatial distribution on the multi-dimensional state evaluation indexes to obtain a multi-dimensional fault feature vector, and calculating the propagation path and dividing the risk level of the multi-dimensional fault feature vector to obtain the fault risk level of the target intelligent battery; performing safety threshold grading calculation and conversion of management instructions on the fault risk level and the multi-dimensional state evaluation indexes to generate a battery management result of the target intelligent battery. Compared with the prior art, in this application, by collecting multi-source data and separating features of traditional electrochemical parameters and non-traditional physical parameters of the target intelligent battery, basic state features are obtained, and then these features are decomposed in the frequency domain and reconstructed spatio-temporally to obtain multi-modal battery data; furthermore, through multi-level combined feature extraction and fusion, deep coupling of electrochemical features and structural features is achieved to obtain a fusion feature vector; furthermore, based on the two-channel cross-attention mechanism and multi-objective time series prediction, comprehensive state evaluation indexes of the battery are obtained, so as to generate a battery management result of the target intelligent battery based on the comprehensive state evaluation indexes. Through hierarchical feature extraction and state evaluation, the problem of accurate identification of battery state monitoring is solved. Especially in aspects such as performance degradation, structural evolution and fault diagnosis, the dynamic characteristics and degradation mechanism of the battery are fully considered, effectively improving the evaluation accuracy; and a multi-level feature fusion and risk warning strategy is adopted, which not only realizes the collaborative analysis of multi-source parameters, but also enhances the reliability of fault diagnosis; in addition, through fault propagation analysis and management instruction optimization, intelligent management and safety control of the battery pack are realized.

[0016] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention are achieved and obtained by the structures specifically pointed out in the specification, claims and drawings.

[0017] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following specifically provides preferred embodiments and, in conjunction with the accompanying drawings, gives a detailed description as follows. Description of the Drawings

[0018] Figure 1 Schematic diagram of the first embodiment of the intelligent battery management method in the embodiment of the present invention;

[0019] Figure 2 Schematic diagram of an embodiment of the intelligent battery management device in the embodiment of the present invention;

[0020] Figure 3 Schematic diagram of an embodiment of the intelligent battery management device in the embodiment of the present invention. Detailed Embodiment

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. 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 scope of protection of the present invention.

[0022] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes other unlisted steps or units, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0023] To facilitate the understanding of this embodiment, the following describes the specific process of the embodiment of the present invention. Please refer to Figure 1 , the first embodiment of the intelligent battery management method in the embodiment of the present invention includes:

[0024] 101. Collect the first multi-source original sensing parameter data corresponding to the first sensing parameter of the target intelligent battery and the second multi-source original sensing parameter data corresponding to the second sensing parameter, and perform spatio-temporal correlation and multiple preprocessings on the first multi-source original sensing parameter data and the second multi-source original sensing parameter data respectively to obtain the first multi-modal sensing data and the second multi-modal sensing data;

[0025] Embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is a theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0026] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0027] In this embodiment, the above first sensing parameter is a basic parameter for monitoring the battery working state (such as voltage parameter, current parameter, temperature parameter, and internal resistance parameter), which mainly reflects the electrochemical performance and thermodynamic characteristics of the battery; the above second sensing parameter refers to structural response parameters (vibration parameter, acoustic emission parameter, ultrasonic parameter, and displacement parameter), which mainly reflects the mechanical structure changes of the battery. In addition, the first sensing parameter and the second sensing parameter can also be appropriately increased or decreased according to actual needs; respectively collect the battery voltage data, battery current data, battery temperature data, and battery internal resistance data corresponding to the first sensing parameter of the target intelligent battery, and generate first multi-source original sensing parameter data based on the battery voltage data, the battery current data, the battery temperature data, and the battery internal resistance data; respectively collect the battery vibration data, battery acoustic emission data, battery ultrasonic data, and battery displacement data corresponding to the second sensing parameter of the target intelligent battery, and generate second multi-source original sensing parameter data based on the battery vibration data, the battery acoustic emission data, the battery ultrasonic data, and the battery displacement data; perform frequency segmentation on the first multi-source original sensing parameter data to obtain an electrochemical low-frequency data segment, an electrochemical intermediate-frequency data segment, and an electrochemical high-frequency data segment, and perform frequency segmentation on the second multi-source original sensing parameter data to obtain a mechanical low-frequency data segment, a mechanical intermediate-frequency data segment, and a mechanical high-frequency data segment; perform Kalman filtering on the electrochemical low-frequency data segment and the mechanical low-frequency data segment to obtain state basic features, perform wavelet multi-scale decomposition on the electrochemical intermediate-frequency data segment and the mechanical intermediate-frequency data segment to obtain dynamic response features, and perform empirical mode decomposition on the electrochemical high-frequency data segment and the mechanical high-frequency data segment to obtain transient evolution features; perform spatio-temporal synchronous matching on the state basic features, the dynamic response features, and the transient evolution features to obtain initial multi-modal sensing data, and perform signal-to-noise ratio evaluation and weighted calculation on the initial multi-modal sensing data for the corresponding original sensing parameters to obtain first multi-modal sensing data and second multi-modal sensing data.

[0028] In practical applications, first, traditional electrochemical parameter data (i.e., the first multi-source original sensing parameter data) corresponding to the first sensing parameter of the battery is collected through a high-precision sensor network. This includes collecting battery voltage data using a voltage sensor with a sampling rate of 1 kHz and an accuracy of 0.1 mV, collecting battery current data using a Hall current sensor with a sampling rate of 1 kHz and an accuracy of 0.1%, collecting temperature field data using a distributed PT100 temperature sensor at a sampling rate of 100 Hz, and collecting internal resistance data by sweeping the frequency in the range of 0.1 Hz - 1 kHz using an AC impedance tester. At the same time, structural response parameter data (i.e., the second multi-source original sensing parameter data) corresponding to the second sensing parameter of the battery is collected using multi-source physical sensors, including collecting vibration data using a three-axis accelerometer at a sampling rate of 10 kHz, collecting acoustic emission signals using a piezoelectric sensor array in the range of 100 kHz - 1 MHz, collecting ultrasonic data using an ultrasonic transducer array in the frequency band of 20 - 100 kHz, and collecting displacement data using a fiber Bragg grating sensor at a sampling rate of 50 Hz and an accuracy of 1 μm. Then, the collected traditional electrochemical parameter data and structural response parameter data are subjected to spatio-temporal alignment and frequency segmentation. That is, based on the spatial arrangement and response characteristics of the sensors, a spatio-temporal mapping relationship of the data is established to achieve synchronous alignment of multi-source data. The electrochemical parameter data is divided into a low-frequency band (0 - 100 Hz), a medium-frequency band (100 Hz - 10 kHz), and a high-frequency band (>10 kHz) according to frequency characteristics. Similarly, the structural response parameter data is also divided into three corresponding frequency bands. For the low-frequency data segment, an improved Kalman filtering algorithm is used to eliminate slow-varying noise and system drift and extract the steady-state characteristics of the battery. For the medium-frequency data segment, the db4 wavelet basis function is used for 5-layer multi-scale decomposition to separate the dynamic response characteristics of the battery. For the high-frequency data segment, the empirical mode decomposition method is used to extract the transient characteristics and non-linear response characteristics of the battery. Then, the characteristics obtained from processing different frequency bands are subjected to spatio-temporal synchronous matching and feature fusion. That is, by establishing a time reference and a spatial registration relationship, the state-based characteristics, dynamic response characteristics, and transient evolution characteristics are aligned and integrated to form initial multi-modal sensing data. Then, the signal-to-noise ratio of the integrated data is evaluated, and the data quality indicators of each sensing parameter are calculated, including signal integrity, consistency, and effectiveness, etc. Based on these indicators, the data is weighted processed. Thus, through this hierarchical data processing and feature extraction method, high-quality first multi-modal sensing data and second multi-modal sensing data are finally obtained.For example, during the charge and discharge process of a lithium-ion battery at a rate of 0.5C, the characteristic data obtained through the above processing shows that the voltage-current coupling characteristics reflect the change in the state of charge in the low-frequency band (<100 Hz), the polarization process is reflected in the mid-frequency band, and micro-electrochemical reactions are captured in the high-frequency band. At the same time, acoustic emission signals detect the formation of micro-cracks in the electrode material in the frequency band of 50 - 200 kHz, and the modal analysis of vibration signals reveals the stress distribution of the internal structure of the battery. These characteristics form a multi-modal mapping relationship with electrochemical parameters, providing multi-dimensional characteristic support for subsequent state assessment.

[0029] 102. Extract multi-level combined characteristics from the first multi-modal sensing data and the second multi-modal sensing data to obtain a multi-battery parameter combination feature set, and perform multi-dimensional feature fusion on each multi-battery parameter combination feature set to obtain a fusion feature vector of the intelligent battery state.

[0030] In this embodiment, hierarchical feature extraction is performed on the first multi-modal sensing data to obtain a voltage-current feature set and a temperature-internal resistance feature set, and hierarchical feature extraction is performed on the second multi-modal sensing data to obtain a vibration-acoustic emission feature set and an ultrasonic displacement feature set. Temporal pairing is performed on the voltage-current feature set and the vibration-acoustic emission feature set to obtain a multi-dimensional feature relationship matrix, and cross-correlation calculation and non-linear mapping are performed on the multi-dimensional feature relationship matrix to obtain a first battery parameter combination feature set. Spatial registration is performed on the temperature-internal resistance feature set and the ultrasonic displacement feature set to obtain a structural parameter relationship matrix, and time-frequency feature mapping is performed on the structural parameter relationship matrix to obtain a second battery parameter combination feature set. Wavelet packet multi-level decomposition is performed on the first battery parameter combination feature set to obtain multi-scale battery reaction stress characteristics, and empirical mode decomposition is performed on the second battery parameter combination feature set to obtain multi-scale thermal field deformation characteristics. Energy reconstruction and feature selection are performed on the multi-scale battery reaction stress characteristics to obtain electrochemical strain coupling characteristics, and modal screening and feature fusion are performed on the multi-scale thermal field deformation characteristics to obtain temperature deformation coupling characteristics. Hierarchical contribution degree calculation is performed on the electrochemical strain coupling characteristics and the temperature deformation coupling characteristics to obtain a multi-coupling weight matrix, and based on the multi-coupling weight matrix, weighted fusion is performed on the electrochemical strain coupling characteristics and the temperature deformation coupling characteristics to obtain a fusion feature vector of the intelligent battery state.

[0031] In practical applications, first, hierarchical analysis is performed on the first multimodal sensing data, that is, charge-discharge characteristics, polarization characteristics, and capacity characteristics are extracted through voltage-current curve analysis. At the same time, correlation analysis is carried out on the temperature field distribution and internal resistance change to extract thermal-resistance coupling characteristics, thereby obtaining a voltage-current feature set and a temperature-internal resistance feature set; meanwhile, feature extraction is performed on the second multimodal sensing data. Time-frequency analysis is respectively carried out on the vibration signal and the acoustic emission signal to extract modal characteristics and energy characteristics, and spatial correlation analysis is carried out on the ultrasonic propagation characteristics and the displacement field distribution to extract structural deformation characteristics, forming a vibration-acoustic emission feature set and an ultrasonic-displacement feature set; furthermore, the voltage-current feature set and the vibration-acoustic emission feature set are paired according to the time series. By calculating the correlation coefficient and time-delay parameter between features, a mapping relationship between electrochemical response and mechanical response is established, a multi-dimensional feature relationship matrix is constructed, and non-linear mapping and cross-correlation analysis are performed on this matrix to extract electrochemistry-mechanics coupling characteristics, obtaining a first battery parameter combination feature set; similarly, spatial registration is performed on the temperature-internal resistance feature set and the ultrasonic-displacement feature set to establish the corresponding relationship between the temperature field distribution and the structural deformation, forming a structural parameter relationship matrix, and then through time-frequency domain feature mapping, a second battery parameter combination feature set is obtained; furthermore, multi-level decomposition is performed on the first battery parameter combination feature set using wavelet packet transform to extract battery reaction-stress characteristics in different frequency bands and capture the coupling relationship between electrochemical reactions and mechanical stresses; at the same time, empirical mode decomposition is performed on the second battery parameter combination feature set to obtain multi-scale characteristics reflecting the temperature field distribution and structural deformation. Then, through energy reconstruction and feature selection, the battery reaction-stress characteristics are integrated into electrochemistry-strain coupling characteristics, and through mode screening and feature fusion, the thermal field-deformation characteristics are integrated into temperature-deformation coupling characteristics; furthermore, a multi-level feature importance evaluation method is adopted to calculate the contribution degrees of the electrochemistry-strain coupling characteristics and the temperature-deformation coupling characteristics to the battery state evolution respectively, that is, the SHAP values of the features are calculated through the XGBoost algorithm, combined with the feature screening results of Lasso regression and the feature importance scores of random forests to construct a multiple coupling weight matrix, and based on this weight matrix, weighted fusion is performed on the electrochemistry-strain coupling characteristics and the temperature-deformation coupling characteristics, and finally a fusion feature vector comprehensively reflecting the battery state is obtained. This multi-level feature extraction and fusion method not only realizes the deep coupling of the electrochemical characteristics and structural characteristics of the battery, but also can accurately capture the performance degradation and fault evolution characteristics of the battery. For example, through the correlation analysis of the voltage drop and vibration signal of a lithium-ion battery, it can be found that the sharp voltage drop at the end of discharge is often accompanied by a significant change in the characteristic frequency, and this correlation feature can be used to early warn of the battery failure risk.

[0032] 103. Separate the fused feature vector into the first sensing parameter channel and the second sensing parameter channel to obtain the electrochemical state feature stream and the structure evolution feature stream, and calculate the cross-attention weights of the two channels and predict the multi-objective time series for the electrochemical state feature stream and the structure evolution feature stream to obtain the multi-dimensional state evaluation index of the target intelligent battery;

[0033] In this embodiment, the attention weight corresponding to the first sensing parameter of the fusion feature vector is calculated, the initial electrochemical features are extracted, and the feature dimension of the initial electrochemical features is compressed and the electrochemical features are reconstructed to obtain an electrochemical feature subset of the target intelligent battery; the double-branch temporal convolutional features of the electrochemical feature subset are extracted to obtain an electrochemical response matrix, and the temporal reconstruction and performance feature enhancement of the electrochemical response matrix are performed to obtain an electrochemical state feature stream; the local similarity corresponding to the second sensing parameter of the fusion feature vector is calculated, the initial structural features of the corresponding feature components are extracted, and the reconstruction of the structural features and the screening of the key structural features of the initial structural features are performed to obtain a structural evolution feature subset of the target intelligent battery; the double-branch spatial convolutional features of the structural evolution feature subset are extracted to obtain a structural response matrix, and the spatial reconstruction and feature enhancement of the structural response matrix are performed to obtain a structural evolution feature stream; the self-attention weight calculation in the temporal dimension is performed on the corresponding charge-discharge feature branch and polarization feature branch in the electrochemical state feature stream to obtain an electrochemically enhanced feature, and the self-attention weight calculation in the spatial dimension is performed on the corresponding stress feature branch and deformation feature branch in the structural evolution feature stream to obtain a structurally enhanced feature; the cross-attention mapping calculation of the two channels is performed on the electrochemically enhanced feature and the structurally enhanced feature to obtain a feature weight mapping matrix, and the sliding temporal segmentation and the prediction of the battery performance trend of the feature weight mapping matrix are performed to obtain a state evolution sequence; based on a preset branch evaluation model, the multi-branch state of the state evolution sequence is evaluated to obtain an initial battery state evaluation value, and the confidence calibration and index optimization of the initial battery state evaluation value are performed to obtain the multi-dimensional state evaluation index of the target intelligent battery. The above-mentioned preset branch evaluation model refers to a dedicated model structure for evaluating three key states of the battery: namely, the state of charge (SOC) evaluation branch: a fusion model based on an improved Auger coulomb counting method and a Kalman filter, combined with voltage-current characteristics and temperature compensation, to evaluate the instant charging state of the battery; the state of health (SOH) evaluation branch: a comprehensive evaluation model based on the capacity attenuation rate, internal resistance growth rate, and coulomb efficiency, and continuously optimizing the parameters through online learning to evaluate the health of the battery; the remaining useful life (RUL) prediction branch: a life prediction model based on support vector regression, combined with historical degradation data and current state characteristics, to predict the remaining useful life of the battery; these three branch models are optimized using historical data in the training stage to form a preset evaluation benchmark for real-time evaluation of the battery state.

[0034] In practical applications, first, the self-attention weight of the features related to the first sensing parameter in the fusion feature vector is calculated. By setting the query matrix and the key-value matrix, the correlation score between the features is calculated using the multi-head self-attention mechanism. The calculation formula of the multi-head self-attention mechanism is as follows:

[0035] , , , ;

[0036] Among them, is the feature dimension, , , is a learnable weight matrix, F is the fused feature vector [n×d], n represents the number of data at time points, d contains all the fused features, Q, K, V are the query, key, and value matrices [n× , softmax is the normalization function (for example: for the voltage feature vector [1.2V, 3.6V, 4.2V], calculate the attention scores [0.2, 0.3, 0.5], indicating the importance of different voltage points), so as to extract the initial electrochemical features reflecting the battery's electrochemical performance; and perform dimensionality reduction and compression on these initial electrochemical features using a variational autoencoder, map the high-dimensional features to the latent space through the encoder, and then reconstruct the electrochemical features through the decoder to retain the key electrochemical information, thus obtaining a subset of the battery's electrochemical features (for example: compress 50-dimensional electrochemical features to a 10-dimensional latent space and then reconstruct to obtain the key electrochemical features); then input the subset of electrochemical features into a two-branch temporal convolutional network, where one branch extracts charge-discharge dynamic features through one-dimensional convolution, and the other branch extracts polarization dynamic features. Among them, the formula for the two-branch temporal convolutional network:

[0037] ;

[0038] Among them, \(W(\tau)\) is the convolution kernel, \(\tau\) is the time delay, \(X_t\) is the time series feature, \(W\) is the convolution kernel parameter, \(H_t\) is the convolution output, \(\tau\) is the time window (for example: using a convolution kernel with a length of 5 to extract features from the charge and discharge curve to capture the voltage change rate feature). Finally, the outputs of the two branches are combined to form an electrochemical response matrix. Furthermore, the response matrix is reconstructed and feature enhanced in the time series dimension, and the attention mechanism is used to highlight the key time series features, and finally a state feature stream representing the evolution of the battery's electrochemical performance is obtained. At the same time, local sensitive hashing is calculated for the second sensing parameter features in the fused feature vector to evaluate the local similarity between features, and the initial features reflecting the battery structure change are extracted. These initial structure features are reduced in dimension and reconstructed by a variational autoencoder, and the features most sensitive to the structure evolution are screened out to obtain a subset of the battery's structure evolution features. Then, the subset of the structure evolution features is input into a two-branch spatial convolution network. One branch is responsible for extracting the stress distribution features, and the other branch is responsible for extracting the deformation distribution features. The outputs of the two branches are fused to obtain a structure response matrix. Furthermore, the response matrix is reconstructed and feature enhanced in the spatial dimension, and the structure change features of the key regions are highlighted through the spatial attention mechanism to obtain a feature stream representing the evolution law of the battery structure. It not only realizes the effective decoupling of the electrochemical features and the structure features, but also maintains the integrity and representativeness of their respective features.

[0039] Secondly, perform time series processing on the charge and discharge feature branch and the polarization feature branch in the electrochemical state feature stream. By constructing a time series attention mechanism, calculate the correlation weights of the features at different time points, that is, use the charge and discharge curve features as the query sequence and the polarization features as the key-value sequence, and calculate their self-attention scores in the time series dimension to highlight the electrochemical features of the key time series nodes and obtain an enhanced electrochemical feature expression (for example, for the charge and discharge feature sequence [1.2V, 1.5V, 1.3V] and the polarization feature sequence [0.1Ω, 0.15Ω, 0.12Ω], the enhanced features are calculated through time series self-attention to highlight the electrochemical response at critical moments). At the same time, perform spatial dimension attention calculation on the stress feature branch and the deformation feature branch in the structure evolution feature stream. By establishing a spatial attention mapping relationship, calculate the importance weights of the features at different spatial positions to highlight the structure change features of the key regions and obtain an enhanced structure feature expression. Then, the enhanced electrochemical features and structure features are input into a two-channel cross-attention network, and its cross-attention network formula is:

[0040] ;

[0041] Where: , , , \(F_c\) is the electrochemical feature, \(F_s\) is the structure feature, \(W_q\), \(W_k\), \(W_v\) are the projection matrices, \(d\) is the feature dimension, and the dot product Represents the correlation between electrochemical characteristics and structural characteristics. Thus, by calculating the cross-correlation between the two types of characteristics, a mapping relationship between the electrochemical state and structural evolution is established, forming a characteristic weight mapping matrix. Furthermore, time series segmentation and trend prediction are performed on this matrix using the characteristic weight mapping matrix. An appropriate window length and sliding step are set to extract the time series characteristics of battery performance evolution, and a deep recurrent network is used to model these time series characteristics to predict the change trend of battery performance. The time series prediction model corresponding to the characteristic weight mapping matrix is as follows:

[0042] ;

[0043] Where: X(t) is the time series of electrochemical characteristics, and Z(t) is the time series of structural characteristics , is the basis function, , is the weight coefficient, τ is the prediction step, and N is the number of characteristics or the number of basis functions (for example: analyzing the capacity fade trend through a sliding window (window length 100, step 10) to predict the health state in the next 50 cycles), so as to obtain a complete state evolution sequence. Furthermore, based on a pre-trained three-branch evaluation model, the state of charge, health state, and remaining life in the state evolution sequence are evaluated respectively to obtain the initial evaluation value of the battery state. That is, by establishing a confidence interval based on historical data statistics, the reliability of the initial evaluation value is calibrated, and combined with the constraint conditions of the battery physical model, various state indicators are optimized, and finally a multi-dimensional evaluation index that accurately reflects the comprehensive state of the battery is obtained. It not only considers the interaction between electrochemical characteristics and structural characteristics but also realizes the accurate evaluation of the multi-dimensional state of the battery.

[0044] 104. Perform time series anomaly detection and determination of abnormal spatial distribution on the multi-dimensional state evaluation index to obtain a multi-dimensional fault feature vector, and calculate the propagation path and divide the risk level of the multi-dimensional fault feature vector to obtain the fault risk level of the target intelligent battery;

[0045] In this embodiment, the deviation calculation of preset multi-monitoring parameters for the multi-dimensional state evaluation index is performed to obtain a state change sequence, and the state change sequence is classified and abnormally located according to multi-level battery abnormal working thresholds to obtain an abnormal state sequence; the abnormal state sequence is subjected to spatial distribution mapping and correlation feature extraction to obtain a multi-dimensional fault feature vector, and a fault correlation network is constructed for the multi-dimensional fault feature vector to obtain an initial propagation path matrix; the diffusion rate and propagation direction of the initial propagation path matrix are calculated to obtain fault dynamic evolution parameters, and the fault dynamic evolution parameters are subjected to spatio-temporal sequence reconstruction to obtain a battery fault propagation matrix; the multi-dimensional risk index of the fault propagation matrix is calculated to obtain a fault risk score, and the fault risk score is classified and prioritized to obtain the fault risk level of the target intelligent battery. The above-mentioned preset multi-monitoring parameters refer to the voltage change rate of the state of charge, the capacity attenuation rate of the state of health, and the performance degradation rate of the remaining life.

[0046] In practical applications, reference values for the normal operating range of the battery are set for the multi-dimensional state evaluation index, including the voltage range of the state of charge, the capacity range of the state of health, and the degradation threshold of life prediction. By calculating the deviation between the actual state index and the reference value, the deviation calculation formula is:

[0047] ;

[0048] where ΔS(t) is the comprehensive state deviation, is the weight coefficient of each state index, is the current state value, is the reference state value, is the time decay coefficient, is the initial moment, α is the spatial diffusion coefficient, and ▽² is the Laplace operator, representing the spatial diffusion effect, so as to obtain a state sequence reflecting the battery performance change (for example, when the state of health SOH of the battery drops from 90% to 85%, calculate its state deviation: ); Furthermore, based on preset multi-level warning thresholds, anomaly detection is performed on the state change sequence, including threshold judgments for three levels: slight anomaly, moderate anomaly, and severe anomaly. At the same time, combined with the constraints of the battery physical model, spatial positioning is performed on the detected abnormal states to determine the specific location and time point where the anomaly occurs, thereby obtaining a complete abnormal state sequence; furthermore, the abnormal state sequence is mapped into the spatial structure model of the battery, the distribution law and propagation characteristics of the abnormal state inside the battery are analyzed, the correlation characteristics between abnormal states are extracted, a multi-dimensional fault feature vector is constructed, and a fault correlation network is established based on this feature vector. By calculating the mutual influence degree and propagation path between fault points, an initial propagation path matrix is formed; furthermore, quantitative analysis is performed on this matrix, that is, calculating the diffusion rate and propagation direction of the fault inside the battery, obtaining key parameters describing the dynamic evolution of the fault, and through the spatio-temporal sequence reconstruction method, these parameters are integrated into a complete fault propagation matrix, which comprehensively reflects the development law and influence range of the fault; furthermore, multi-dimensional risk indicators are calculated based on the fault propagation matrix, including the diffusion range, propagation speed, and influence degree of the fault, and the fault risk score is comprehensively evaluated. According to the distribution characteristics and clustering results of the score, the fault risk is divided into different levels, and considering the urgency and harmfulness of the fault, the priority of various faults is sorted, and finally the fault risk level of the battery pack is obtained. Thus, a full-process analysis from anomaly detection to risk assessment is achieved.

[0049] 105. Perform safety threshold grading calculation and conversion of management instructions on the fault risk level and multi-dimensional state evaluation indicators to generate the battery management result of the target intelligent battery.

[0050] In this embodiment, safety factor mapping and working interval conversion are performed on the fault risk level and the multi-dimensional state evaluation indicators to obtain battery working constraint parameters; charge and discharge parameter conversion of corresponding power limit coefficients is performed on the battery working constraint parameters to obtain power management instructions, and cooling channel flow rate distribution and heat dissipation power distribution calculations of corresponding temperature thresholds are performed on the battery working constraint parameters to obtain temperature management instructions, and active balancing circuit parameter calculations of corresponding single-cell battery state differences are performed on the battery working constraint parameters to obtain balancing management instructions; priority sorting and battery state regulation are performed on the power management instructions, the temperature management instructions, and the balancing management instructions to generate the battery management result of the target intelligent battery.

[0051] In practical applications, a safety margin mapping relationship is established based on the fault risk level and multi-dimensional state evaluation indicators. The fault risk level is converted into a power limit coefficient, and the state evaluation indicators are converted into the boundaries of the working range. The battery operating constraint parameters are obtained through comprehensive calculation (where these constraint parameters not only consider the safety boundaries of the battery but also include the target requirements for performance optimization, providing a basic basis for the subsequent generation of management instructions). Furthermore, during the generation of management instructions, first, the power limit coefficient in the operating constraint parameters is converted into specific charge and discharge parameters, including the upper limit value of the charging current, the discharge power threshold, and the cut-off voltage, etc., to form power management instructions. At the same time, based on the temperature threshold parameters, the flow distribution ratio of each cooling channel and the power distribution scheme of the radiator are calculated, considering the uniformity of the temperature field distribution and the key control of the hot spot area, to generate temperature management instructions. In addition, according to the state differences between individual batteries, the operating parameters of the equalization circuit are calculated, including the magnitude of the equalization current, the equalization time, and the selection of the equalization path, and energy transfer is achieved through the DC-DC conversion circuit to form equalization management instructions. Then, the three types of management instructions are uniformly coordinated and prioritized. When there are multiple management requirements, the instructions related to safety are preferentially ensured to be executed, followed by the instructions related to performance optimization, and finally, the maintenance management instructions are executed. Furthermore, through the reasonable arrangement of execution priorities and the coordination and cooperation between instructions, the safe and controllable operation of the battery pack is realized, and finally, a complete battery management result is generated. This hierarchical management strategy not only ensures the safe operation of the battery but also realizes the optimization and improvement of performance.

[0052] In the embodiments of the present invention, through multi-source data collection and feature separation of the traditional electrochemical parameters and non-traditional physical parameters of the target intelligent battery, the basic state features are obtained. Then, these features are subjected to frequency-domain decomposition and spatio-temporal reconstruction to obtain the multi-modal data of the battery. Furthermore, through multi-level combined feature extraction and fusion, the deep coupling of electrochemical features and structural features is realized, and a fusion feature vector is obtained. Then, based on the dual-channel cross-attention mechanism and multi-objective time-series prediction, the comprehensive state evaluation indicators of the battery are obtained, so as to generate the battery management result of the target intelligent battery based on the comprehensive state evaluation indicators. Through hierarchical feature extraction and state evaluation, the problem of accurate identification of battery state monitoring is solved. Especially in aspects such as performance degradation, structural evolution, and fault diagnosis, the dynamic characteristics and degradation mechanism of the battery are fully considered, effectively improving the evaluation accuracy. And a multi-level feature fusion and risk warning strategy is adopted, which not only realizes the collaborative analysis of multi-source parameters but also enhances the reliability of fault diagnosis. In addition, through fault propagation analysis and management instruction optimization, the intelligent management and safety control of the battery pack are realized.

[0053] The intelligent battery management method in the embodiments of the present invention has been described above. Next, the intelligent battery management device in the embodiments of the present invention will be described. Please refer to Figure 2, an embodiment of the intelligent battery management device in the embodiments of the present invention includes:

[0054] A preprocessing module 201, configured to collect first multi-source original sensing parameter data corresponding to a first sensing parameter of a target intelligent battery and second multi-source original sensing parameter data corresponding to a second sensing parameter, and perform spatio-temporal association and multiple preprocessing on the first multi-source original sensing parameter data and the second multi-source original sensing parameter data respectively to obtain first multi-modal sensing data and second multi-modal sensing data;

[0055] A feature fusion module 202, configured to extract multi-level combined features from the first multi-modal sensing data and the second multi-modal sensing data to obtain a multi-battery parameter combined feature set, and perform multi-dimensional feature fusion on each of the battery parameter combined feature sets to obtain a fusion feature vector of the intelligent battery state;

[0056] An index evaluation module 203, configured to separate the fusion feature vector into an electro-chemical state feature stream and a structural evolution feature stream for a first sensing parameter channel and a second sensing parameter channel, and calculate cross-attention weights for the two channels and predict multi-object time series to obtain multi-dimensional state evaluation indexes of the target intelligent battery;

[0057] A risk division module 204, configured to perform time series anomaly detection and determination of abnormal spatial distribution on the multi-dimensional state evaluation indexes to obtain a multi-dimensional fault feature vector, and calculate a propagation path and divide a risk level for the multi-dimensional fault feature vector to obtain a fault risk level of the target intelligent battery;

[0058] An instruction conversion module 205, configured to perform safety threshold classification calculation and conversion of management instructions on the fault risk level and the multi-dimensional state evaluation indexes to generate a battery management result of the target intelligent battery.

[0059] In the embodiments of the present invention, multi-source data collection and feature separation are performed on the traditional electrochemical parameters and non-traditional physical parameters of the target intelligent battery to obtain the basic state features. Then, these features are decomposed in the frequency domain and reconstructed in space-time to obtain the battery multi-modal data. Furthermore, through multi-level combined feature extraction and fusion, the deep coupling of electrochemical features and structural features is realized, and a fused feature vector is obtained. Furthermore, based on the dual-channel cross-attention mechanism and multi-objective time series prediction, a comprehensive state evaluation index of the battery is obtained, so as to generate the battery management result of the target intelligent battery based on the comprehensive state evaluation index. Through hierarchical feature extraction and state evaluation, the problem of accurate identification of battery state monitoring is solved. Especially in aspects such as performance degradation, structural evolution, and fault diagnosis, the dynamic characteristics and degradation mechanism of the battery are fully considered, and the evaluation accuracy is effectively improved. And a multi-level feature fusion and risk warning strategy is adopted, which not only realizes the collaborative analysis of multi-source parameters but also enhances the reliability of fault diagnosis. In addition, through fault propagation analysis and management instruction optimization, the intelligent management and safety control of the battery pack are realized.

[0060] above Figure 2 The intelligent battery management device in the embodiments of the present invention is described in detail from the perspective of modular functional entities. Next, the intelligent battery management device in the embodiments of the present invention is described in detail from the perspective of hardware processing.

[0061] Figure 3 FIG. is a schematic structural diagram of an intelligent battery management device provided by an embodiment of the present invention. The intelligent battery management device 300 may vary greatly due to configuration or performance differences, and may include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 (for example, one or more mass storage devices) for storing application programs 333 or data 332. Among them, the memory 320 and the storage media 330 may be transient storage or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the intelligent battery management device 300. Further, the processor 310 may be configured to communicate with the storage media 330 and execute a series of instruction operations in the storage media 330 on the intelligent battery management device 300.

[0062] The intelligent battery management device 300 may further include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, and so on. Those skilled in the art can understand that Figure 3 The structure of the illustrated intelligent battery management device does not constitute a limitation on the intelligent battery management device, and it may include more or fewer components than those shown, or combine certain components, or have different component arrangements.

[0063] The present invention also provides an intelligent battery management device. The computer device includes a memory and a processor. When the computer-readable instructions stored in the memory are executed by the processor, the processor is caused to execute each step of the intelligent battery management method in the above-mentioned embodiments.

[0064] The present invention also provides a computer-readable storage medium. The computer-readable storage medium may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions run on a computer, the computer is caused to execute each step of the intelligent battery management method.

[0065] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0066] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0067] This application can be used in numerous general-purpose or special-purpose computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0068] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent battery management method, characterized in that The described intelligent battery management method includes: Collecting the first multi-source original sensing parameter data corresponding to the first sensing parameter of the target intelligent battery and the second multi-source original sensing parameter data corresponding to the second sensing parameter, and respectively performing spatio-temporal correlation and multiple preprocessings on the first multi-source original sensing parameter data and the second multi-source original sensing parameter data to obtain the first multi-modal sensing data and the second multi-modal sensing data; Extracting multi-level combined features from the first multi-modal sensing data and the second multi-modal sensing data to obtain a multi-battery parameter combination feature set, and performing multi-dimensional feature fusion on each of the battery parameter combination feature sets to obtain a fusion feature vector of the intelligent battery state; Calculating the attention weight corresponding to the first sensing parameter of the fusion feature vector, extracting the initial electrochemical feature, and performing feature dimension compression and electrochemical feature reconstruction on the initial electrochemical feature to obtain an electrochemical feature subset of the target intelligent battery; extracting double-branch temporal convolutional features from the electrochemical feature subset to obtain an electrochemical response matrix, and performing temporal reconstruction and performance feature enhancement on the electrochemical response matrix to obtain an electrochemical state feature stream; calculating the local similarity corresponding to the second sensing parameter of the fusion feature vector, extracting the initial structural feature of the corresponding feature component, and performing structural feature reconstruction and screening of key structural features on the initial structural feature to obtain a structural evolution feature subset of the target intelligent battery; extracting double-branch spatial convolutional features from the structural evolution feature subset to obtain a structural response matrix, and performing spatial reconstruction and feature enhancement on the structural response matrix to obtain a structural evolution feature stream, and calculating the cross-attention weight of the two channels and predicting multi-objective time series for the electrochemical state feature stream and the structural evolution feature stream to obtain the multi-dimensional state evaluation index of the target intelligent battery; Calculating the deviation of the preset multi-monitoring parameters for the multi-dimensional state evaluation index to obtain a state change sequence, and classifying and abnormally locating the multi-level battery abnormal working threshold for the state change sequence to obtain an abnormal state sequence; performing spatial distribution mapping and correlation feature extraction on the abnormal state sequence to obtain a multi-dimensional fault feature vector, and constructing a fault correlation network for the multi-dimensional fault feature vector to obtain an initial propagation path matrix; calculating the diffusion rate and propagation direction for the initial propagation path matrix to obtain a fault dynamic evolution parameter, and performing spatio-temporal sequence reconstruction on the fault dynamic evolution parameter to obtain a battery fault propagation matrix; calculating a multi-dimensional risk index for the fault propagation matrix to obtain a fault risk score, and performing grade division and priority sorting on the fault risk score to obtain the fault risk level of the target intelligent battery; Performing safety threshold classification calculation and conversion of management instructions on the fault risk level and the multi-dimensional state evaluation index to generate the battery management result of the target intelligent battery.

2. The intelligent battery management method according to claim 1, wherein Separately performing spatio-temporal correlation and multiple preprocessing on the first multi-source original sensing parameter data and the second multi-source original sensing parameter data to obtain first multi-modal sensing data and second multi-modal sensing data, including: Performing frequency segmentation on the first multi-source original sensing parameter data to obtain an electrochemical low-frequency data segment, an electrochemical intermediate-frequency data segment, and an electrochemical high-frequency data segment, and performing frequency segmentation on the second multi-source original sensing parameter data to obtain a mechanical low-frequency data segment, a mechanical intermediate-frequency data segment, and a mechanical high-frequency data segment; Performing Kalman filtering on the electrochemical low-frequency data segment and the mechanical low-frequency data segment to obtain state basic features, performing wavelet multi-scale decomposition on the electrochemical intermediate-frequency data segment and the mechanical intermediate-frequency data segment to obtain dynamic response features, and performing empirical mode decomposition on the electrochemical high-frequency data segment and the mechanical high-frequency data segment to obtain transient evolution features; Performing spatio-temporal synchronous matching on the state basic features, the dynamic response features, and the transient evolution features to obtain initial multi-modal sensing data, and performing signal-to-noise ratio evaluation and weighted calculation of corresponding original sensing parameters on the initial multi-modal sensing data to obtain first multi-modal sensing data and second multi-modal sensing data.

3. The intelligent battery management method according to claim 1, wherein Performing extraction of multi-level combined features on the first multi-modal sensing data and the second multi-modal sensing data to obtain a multi-battery parameter combination feature set, and performing multi-dimensional feature fusion on each of the battery parameter combination feature sets to obtain a fusion feature vector of the intelligent battery state, including: Performing hierarchical feature extraction on the first multi-modal sensing data to obtain a voltage-current feature set and a temperature-internal resistance feature set, and performing hierarchical feature extraction on the second multi-modal sensing data to obtain a vibration-acoustic emission feature set and an ultrasonic displacement feature set; Performing time-series pairing on the voltage-current feature set and the vibration-acoustic emission feature set to obtain a multi-dimensional feature relationship matrix, and performing cross-correlation calculation and non-linear mapping on the multi-dimensional feature relationship matrix to obtain a first battery parameter combination feature set; Performing spatial registration on the temperature-internal resistance feature set and the ultrasonic displacement feature set to obtain a structural parameter relationship matrix, and performing time-frequency feature mapping on the structural parameter relationship matrix to obtain a second battery parameter combination feature set; Performing wavelet packet multi-level decomposition on the first battery parameter combination feature set to obtain multi-scale battery reaction stress features, performing empirical mode decomposition on the second battery parameter combination feature set to obtain multi-scale thermal field deformation features, performing energy reconstruction and feature selection on the multi-scale battery reaction stress features to obtain electrochemical strain coupling features, and performing mode screening and feature fusion on the multi-scale thermal field deformation features to obtain temperature deformation coupling features; Performing hierarchical contribution degree calculation on the electrochemical strain coupling features and the temperature deformation coupling features to obtain a multiple coupling weight matrix, and based on the multiple coupling weight matrix, performing weighted fusion on the electrochemical strain coupling features and the temperature deformation coupling features to obtain a fusion feature vector of the intelligent battery state.

4. The intelligent battery management method according to claim 1, wherein, Calculating the cross-attention weights of the dual channels for the electrochemical state feature stream and the structural evolution feature stream and predicting the multi-objective time series to obtain the multi-dimensional state evaluation index of the target intelligent battery, including: Calculating the self-attention weights in the time series dimension for the corresponding charge-discharge feature branch and polarization feature branch in the electrochemical state feature stream to obtain the electrochemically enhanced feature, and calculating the self-attention weights in the spatial dimension for the corresponding stress feature branch and deformation feature branch in the structural evolution feature stream to obtain the structurally enhanced feature; Performing cross-attention mapping calculation on the electrochemically enhanced feature and the structurally enhanced feature for the dual channels to obtain the feature weight mapping matrix, and performing sliding time series segmentation and prediction of the battery performance trend on the feature weight mapping matrix to obtain the state evolution sequence; Based on a preset branch evaluation model, evaluating the multi-branch state of the state evolution sequence to obtain the initial battery state evaluation value, and performing confidence calibration and index optimization on the initial battery state evaluation value to obtain the multi-dimensional state evaluation index of the target intelligent battery.

5. The intelligent battery management method according to claim 1, wherein Converting the safety threshold grading calculation and management instructions for the failure risk level and the multi-dimensional state evaluation index to generate the battery management result of the target intelligent battery, including: Performing safety factor mapping and working interval conversion on the failure risk level and the multi-dimensional state evaluation index to obtain the battery working constraint parameters; Performing charge-discharge parameter conversion of the corresponding power limit coefficient on the battery working constraint parameters to obtain the power management instruction, performing calculation of the cooling channel flow rate allocation and heat dissipation power allocation corresponding to the temperature threshold on the battery working constraint parameters to obtain the temperature management instruction, and performing calculation of the active equalization circuit parameters corresponding to the difference in the state of each single battery on the battery working constraint parameters to obtain the equalization management instruction; Performing priority sorting and battery state regulation on the power management instruction, the temperature management instruction, and the equalization management instruction to generate the battery management result of the target intelligent battery.

6. An intelligent battery management device, characterized in that, The intelligent battery management device includes: A preprocessing module, configured to collect the first multi-source original sensing parameter data corresponding to the first sensing parameter of the target intelligent battery and the second multi-source original sensing parameter data corresponding to the second sensing parameter, and perform spatio-temporal association and multiple preprocessing on the first multi-source original sensing parameter data and the second multi-source original sensing parameter data respectively to obtain the first multi-modal sensing data and the second multi-modal sensing data; A feature fusion module, configured to extract the multi-level combined features of the first multi-modal sensing data and the second multi-modal sensing data to obtain the multi-battery parameter combination feature set, and perform multi-dimensional feature fusion on each of the battery parameter combination feature sets to obtain the fusion feature vector of the intelligent battery state; The index evaluation module is used to calculate the attention weight corresponding to the first sensing parameter of the fused feature vector, extract the initial electrochemical features, compress the feature dimension of the initial electrochemical features and reconstruct the electrochemical features to obtain an electrochemical feature subset of the target intelligent battery; extract the dual-branch temporal convolutional features of the electrochemical feature subset to obtain an electrochemical response matrix, and perform temporal reconstruction and enhancement of performance features on the electrochemical response matrix to obtain an electrochemical state feature stream; calculate the local similarity corresponding to the second sensing parameter of the fused feature vector, extract the initial structural features of the corresponding feature components, and reconstruct the structural features and screen the key structural features of the initial structural features to obtain a structural evolution feature subset of the target intelligent battery; extract the dual-branch spatial convolutional features of the structural evolution feature subset to obtain a structural response matrix, and perform spatial reconstruction and feature enhancement on the structural response matrix to obtain a structural evolution feature stream, and calculate the cross-attention weight of the two channels and predict the multi-objective time series of the electrochemical state feature stream and the structural evolution feature stream to obtain the multi-dimensional state evaluation index of the target intelligent battery; The risk division module is used to calculate the deviation of the preset multi-monitoring parameters from the multi-dimensional state evaluation index to obtain a state change sequence, classify and locate the anomalies of the state change sequence according to the multi-level battery abnormal working thresholds to obtain an abnormal state sequence; perform spatial distribution mapping and correlation feature extraction on the abnormal state sequence to obtain a multi-dimensional fault feature vector, and construct a fault correlation network for the multi-dimensional fault feature vector to obtain an initial propagation path matrix; calculate the diffusion rate and propagation direction of the initial propagation path matrix to obtain fault dynamic evolution parameters, and perform spatio-temporal sequence reconstruction on the fault dynamic evolution parameters to obtain a battery fault propagation matrix; calculate the multi-dimensional risk index of the fault propagation matrix to obtain a fault risk score, and perform grade division and priority sorting on the fault risk score to obtain the fault risk level of the target intelligent battery; The instruction conversion module is used to perform safety threshold classification calculation and conversion of management instructions on the fault risk level and the multi-dimensional state evaluation index to generate the battery management result of the target intelligent battery.

7. An intelligent battery management device, characterized in that, The intelligent battery management device includes: a memory and at least one processor, and instructions are stored in the memory; The at least one processor calls the instructions in the memory so that the intelligent battery management device executes each step of the intelligent battery management method as described in any one of claims 1-5.

8. A computer-readable storage medium having instructions stored thereon, characterized in that, When the instructions are executed by the processor, each step of the intelligent battery management method as described in any one of claims 1-5 is implemented.

Images