Method and system for dynamically monitoring state of lithium ion battery

Abstract

The invention provides a method and system for dynamically monitoring the state of a lithium ion battery, and relates to the field of lithium battery dynamic monitoring, and the method comprises the steps: collecting the image data of the lithium battery at different time points, which is obtained through the reflection or attenuation after X-ray penetration; processing the acquired image data through a reconstruction algorithm to obtain three-dimensional images in the battery at different time points; key structure information in the battery is extracted through the three-dimensional image; the change information of the internal structure of the battery is obtained by extracting key structure information in the battery at different moments. According to the method, the microscopic CT and the thermal imaging technology are combined, and the rapid image reconstruction algorithm and the compressed sensing technology are combined, so that the efficiency of data acquisition and image reconstruction is greatly improved, the time interval of lithium ion battery state detection is remarkably shortened, and the detection accuracy of the lithium ion battery state is improved. And the dynamic real-time monitoring of the lithium ion battery is realized.

Description

Technical Field

[0001] The invention relates to the field of lithium ion battery testing and monitoring, and in particular to a method and system for dynamically monitoring the state of a lithium ion battery. Background Art

[0002] Lithium-ion batteries have been widely used in various fields such as portable electronic devices, electric vehicles and energy storage systems due to their high energy density, long cycle life and stability. However, lithium-ion batteries also have certain safety risks, especially under extreme conditions such as excessive charge and discharge, external impact or high temperature, which are prone to thermal runaway, fire or explosion and other serious accidents.

[0003] The failure of lithium-ion batteries is usually related to changes in their internal structure. Thermal runaway, internal short circuit or other structural changes are the main causes of safety accidents. Although some methods have been used to study the failure mechanism of lithium-ion batteries, most studies rely on static analysis and require a long wait to obtain meaningful results. As a result, each test of lithium batteries takes too long, making it impossible to respond to changes in the internal structure and performance of the battery in time during the interval between tests on both sides, and it is also impossible to monitor the dynamic changes of lithium batteries during use, resulting in the inability to obtain comprehensive monitoring data. Summary of the invention

[0004] Most existing lithium-ion battery testing methods rely on regular static testing or destructive experiments, which usually require a long testing cycle and cannot achieve dynamic real-time monitoring of the battery status. Existing non-destructive testing technologies, such as traditional CT scanning or thermal imaging technology, can provide information inside the battery, but due to the slow acquisition and processing speed, these methods cannot track the battery status changes in real time during the charging and discharging process.

[0005] In addition, existing detection technologies can often only monitor a single parameter (such as temperature, structure or electrochemical performance) and lack the ability to integrate and analyze multiple data types (such as three-dimensional images, temperature distribution and electrochemical characteristics), making it difficult to comprehensively evaluate the health and safety of the battery. In order to solve the shortcomings of existing technologies in dynamic real-time monitoring of lithium-ion battery status, this method uses a combination of micro-CT scanning and thermal imaging technology to achieve real-time monitoring of the internal structure and surface temperature distribution of the battery, and cooperates with iterative reconstruction and compressed sensing technology to accelerate data acquisition and processing. Through multimodal data fusion technology, this method can provide detailed battery status information in a very short time, and timely identify potential battery failures and thermal runaway risks, thereby providing more accurate monitoring and early warning for the safe operation of lithium-ion batteries. The specific contents are as follows.

[0006] A method for dynamically monitoring the state of a lithium-ion battery, the method comprising: Collect image data obtained by reflection or attenuation of lithium batteries after X-ray penetration at different time points. These images represent the attenuation of X-rays by different materials inside the battery. The image data can be obtained by scanning lithium-ion batteries using a micro-CT scanner. During the battery charging, discharging or working process, scans are performed regularly or at specific time points to capture changes inside the battery. During the scan, X-rays penetrate the battery and interact with different battery materials. Different materials attenuate X-rays to different degrees, generating projection images of different areas inside the battery, that is, image data obtained by reflection or attenuation of lithium batteries after X-ray penetration; The collected image data is processed through a reconstruction algorithm to obtain a three-dimensional image of the inside of the battery at different time points, and the key structural information inside the battery (such as electrode particles, cracks, pores, etc.) is extracted through the three-dimensional image; Analyze the three-dimensional image data at different time points to obtain the change information of the internal structure of the battery; By performing particle analysis on the three-dimensional images inside the battery at different times, the morphology and size changes of the electrode particles are tracked to obtain the microstructural changes of the electrode material; By regularly scanning and comparing 3D images at different time points, the crack evolution process inside the battery can be identified and tracked. In combination with image analysis algorithms, the crack location, length, width and other features can be extracted, and its expansion trend can be analyzed, so that the crack evolution inside the battery can be obtained regularly. Through three-dimensional images, the porosity and pore distribution inside the battery are measured, and the image segmentation technology is used to extract the pore area inside the battery, and the porosity and its distribution changes are quantitatively analyzed; the evolution of the pores inside the battery is tracked, the increase or change trend of the pores is identified, and then the reasons for the degradation of battery performance are analyzed.

[0007] Furthermore, the method also includes: Collect thermal imaging data of the battery surface at different time points, analyze the temperature distribution and thermal change trend of the battery surface through the thermal imaging data, and determine whether there is a hot spot area or abnormal temperature; By combining thermal imaging data with 3D image data, we analyze the relationship between the thermal response and internal structure of the battery during the charging and discharging process, identify whether there are overheating areas, and whether there is a potential risk of thermal runaway. By comparing the relationship between the battery surface temperature and the internal structure of the battery (such as pores, cracks, etc.), we can evaluate the safety of the battery.

[0008] Furthermore, the method also includes: Collect electrochemical data of lithium-ion batteries at different time points, including key parameters such as battery voltage, current, power, internal resistance, and charge and discharge efficiency. Obtain real-time battery status through the battery management system (BMS); Obtain 3D image data, thermal imaging data and electrochemical data, and use machine learning algorithms (such as deep learning, Kalman filtering, particle filtering, etc.) to analyze the multimodal information of the battery in real time to obtain the battery status information. Through multi-dimensional data fusion, obtain the real-time health status of the battery, including a comprehensive evaluation of the battery's internal structure, thermal response and electrochemical performance.

[0009] Furthermore, the method of processing the collected image data through a reconstruction algorithm to obtain a three-dimensional image of the inside of the battery at different time points is as follows; Using a filtering algorithm to remove noise from the image data, and performing smoothing processing on the image data to reduce the impact of the noise on subsequent reconstruction, so as to remove artifacts that may appear during rapid acquisition; The image data after noise removal, artifact correction and standardization is subjected to preliminary image reconstruction using an iterative reconstruction method (such as algebraic reconstruction technology ART), so that the accuracy of the image can be gradually improved. Through iteration, the reconstruction error can be continuously reduced until the reconstruction result converges.

[0010] Generate preliminary 3D reconstruction images that can reflect the distribution of the internal structure of the battery, including details such as electrode particles and cracks; Use 3D image processing and visualization software (such as ParaView, VTK, etc.) to load the 3D image data into a visualization environment, such as applying 3D rendering technology to generate interactive 3D models, so that users can intuitively view the internal structure of the battery to further analyze cracks, porosity, electrode particle morphology, etc.

[0011] Further, the method of obtaining the three-dimensional image data, thermal imaging data and electrochemical data, and analyzing the multimodal information of the battery in real time through a machine learning algorithm to obtain the state information of the battery is as follows; The acquired image data, thermal imaging data and electrochemical data are synchronized by precise matching of timestamps and sampling periods; to ensure that the temperature change, internal structure evolution and electrochemical parameters (such as voltage and current) at each time point can be associated together.

[0012] Align data from different sources to obtain multimodal data of the battery and ensure that they are compared within the same time series; for example, the battery's three-dimensional image data, thermal imaging data, and electrochemical data are collected and timestamped within the same time period to facilitate subsequent analysis.

[0013] Perform preliminary cleaning on multimodal data to remove noise; for example, for thermal imaging data, remove temperature fluctuations caused by environmental factors or equipment errors; for CT data, remove artifacts and errors in the image, and filter the electrochemical data to eliminate high-frequency noise to ensure that the data read by each sensor is accurate.

[0014] Extract the microstructural features of the battery from the 3D image data in the multimodal data; for example, identify and extract features such as electrode particle morphology, cracks, porosity, etc.

[0015] And calculate the particle distribution, crack length and porosity of each area of ​​the battery to provide basic data for subsequent health assessment; Temperature change characteristics are extracted from thermal imaging data to obtain the changing trend and abnormal areas of battery surface temperature. The thermal response characteristics of the battery can be evaluated by calculating the temperature rise rate and thermal peak position of the hot spot area, and the temporal and spatial characteristics related to temperature can be extracted, such as the delay time of thermal response and the rate of heating process. This information is helpful for analyzing the development of faults and thermal runaway risks inside the battery.

[0016] The working performance characteristics of the battery are extracted from the electrochemical data (battery voltage, current, power, internal resistance and other key data) and the key indicators such as battery charge and discharge efficiency, internal resistance change and cycle performance are calculated through the electrochemical analysis algorithm to obtain the battery health indicators.

[0017] Extract battery health indicators, such as SOC (State of Charge) and SOH (State of Health), and use them for subsequent data fusion analysis.

[0018] Multimodal data (CT image data, thermal imaging data and electrochemical data) from different sources are merged through data fusion technology; these data can be fused through methods such as Kalman filtering, particle filtering, and weighted averaging.

[0019] For thermal imaging and electrochemical data, they can be combined with CT image data through weighted average or correlation analysis methods, focusing on analyzing the relationship between internal structural changes of the battery (such as cracks and pores) and thermal response and electrochemical performance (such as temperature and voltage).

[0020] Through algorithmic models (such as deep learning networks), different data sets are integrated into a comprehensive data source to further analyze the health status of the battery.

[0021] Furthermore, the method also includes: Combined with spatiotemporal tracking algorithms (such as spatiotemporal convolutional neural networks ST-CNN or LSTM), the collected multimodal data are spatiotemporally tracked to analyze the spatiotemporal dynamic correlation between the structural changes, thermal response and electrochemical performance of the battery at different time points; this is used to evaluate the behavior patterns of the battery under different working conditions.

[0022] Through this spatiotemporal correlation analysis, the behavior patterns of the battery under different working conditions are evaluated and its potential sources of failure can be identified, especially whether the stress changes and temperature changes inside the battery during the charge and discharge process are consistent with the changes in electrochemical performance.

[0023] Furthermore, the method also includes: Build a battery health status assessment model through machine learning (such as support vector machine SVM, random forest RF, etc.) or deep learning (such as convolutional neural network CNN, long short-term memory network LSTM, etc.); The model input of the health status assessment model includes multimodal data features at different time points, such as three-dimensional image data, temperature change characteristics and electrochemical performance characteristics, and the output is a health status assessment value of the battery; It can use a large amount of historical data for model training and optimization to ensure that the evaluation model can accurately identify the health status and potential failures of the battery. During the training process, methods such as cross-validation can be used to improve the generalization ability of the model.

[0024] Optimize the evaluation model so that it can adapt to the health status assessment needs of different battery types and different working environments.

[0025] The trained model is used to evaluate the battery's health status in real time, predicting the battery's remaining life, fault sources, and possible failure mechanisms.

[0026] Based on the model's output, corresponding operational recommendations are given, such as whether safety measures need to be taken, whether charging restrictions should be imposed, whether the battery needs to be replaced, etc. Furthermore, based on the evaluation results of the health status evaluation model, a battery health evaluation report is generated, which includes a comprehensive evaluation of the battery's internal structural changes, thermal response mode, electrochemical performance, and health status.

[0027] A system for dynamically monitoring the state of a lithium-ion battery, the system comprising: Data acquisition module: used to collect multimodal data of lithium-ion batteries in real time, including image data, thermal imaging data, and electrochemical data; Data preprocessing and optimization module: used to perform preliminary processing and optimization on the collected raw data to ensure that the data quality is suitable for subsequent analysis and modeling; Image reconstruction and optimization module: used to reconstruct the image data obtained by reflection or attenuation of lithium batteries at different time points after X-ray penetration, generate high-quality three-dimensional images, and improve data acquisition efficiency through technologies such as compressed sensing; Structural change analysis module: Analyze the changes in the internal structure of the battery through the three-dimensional image of the battery, and monitor the microstructural changes that occur in the battery during the charge and discharge process, such as crack evolution and pore changes; Thermal response analysis module: used to monitor the thermal changes of the battery through thermal imaging data to identify the thermal response pattern of the battery and whether there is a risk of thermal runaway; Multimodal data fusion and analysis module: combines multiple data such as micro-CT, thermal imaging, and electrochemical data for comprehensive analysis to improve the accuracy and comprehensiveness of battery status monitoring; Health status assessment generation module: Analyzes multimodal data through machine learning or deep learning algorithms to generate a battery health status assessment model to obtain a battery health status assessment value, so that it can predict the remaining life and potential failure risk of the battery through this valuation; Visualization and report generation module: visualizes the battery's three-dimensional structure diagram, temperature distribution diagram, health status assessment diagram, etc., to help users intuitively understand the battery status.

[0028] Automatically generate a battery status report based on the analysis results, providing battery health assessment, failure prediction and maintenance recommendations.

[0029] The beneficial effects of this application are as follows: This method greatly improves the efficiency of data acquisition and image reconstruction by combining micro-CT with thermal imaging technology, fast image reconstruction algorithm and compressed sensing technology, significantly shortens the time interval for lithium-ion battery status detection, and realizes dynamic real-time monitoring of lithium-ion batteries, so that the status of lithium-ion batteries can be dynamically monitored in a short time. Through multimodal data fusion, multi-dimensional data such as structural changes, temperature distribution, and charge and discharge performance inside the battery can be obtained in real time, and in-depth analysis can be performed, thereby providing more accurate data support for the study of battery failure mechanisms.

[0030] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A schematic flow chart of a method for dynamically monitoring the status of a lithium-ion battery provided in an embodiment of the present application.

[0032] Figure 2 A schematic diagram of the structure of a system for dynamically monitoring the status of a lithium-ion battery provided in an embodiment of the present application.

[0033] Explanation of the accompanying drawings: 10. Data acquisition module, 20. Data preprocessing and optimization module, 30. Image reconstruction and optimization module, 40. Structural change analysis module, 50. Thermal response analysis module, 60. Multimodal data fusion and analysis module, 70. Health status assessment generation module, 80. Visualization and report generation module. DETAILED DESCRIPTION

[0034] The present application provides a method for dynamically monitoring the status of lithium-ion batteries to solve the problem that most of the lithium-ion battery detection methods in the prior art rely on regular static detection or destructive experiments. These methods usually require a long detection cycle and cannot achieve dynamic real-time monitoring of the battery status. Existing non-destructive testing technologies, such as traditional CT scanning or thermal imaging technology, can provide information inside the battery, but due to the slow acquisition and processing speed, these methods cannot track the state changes of the battery in real time during the charging and discharging process.

[0035] In addition, existing detection technologies can often only monitor a single parameter (such as temperature, structure or electrochemical performance) and lack the ability to integrate and analyze multiple data types (such as three-dimensional images, temperature distribution and electrochemical characteristics), making it difficult to comprehensively evaluate the health and safety of the battery. In order to solve the shortcomings of existing technologies in dynamic real-time monitoring of lithium-ion battery status, this method uses a combination of micro-CT scanning and thermal imaging technology to achieve real-time monitoring of the internal structure and surface temperature distribution of the battery, and cooperates with iterative reconstruction and compressed sensing technology to accelerate data acquisition and processing. Through multimodal data fusion technology, this method can provide detailed battery status information in a very short time, and promptly identify potential battery failures and thermal runaway risks, thereby providing more accurate monitoring and early warning for the safe operation of lithium-ion batteries.

[0036] Specifically, this method includes the following technical links: Data acquisition and optimization: Use micro-CT scanning and thermal imaging technology to obtain the three-dimensional structure image and surface temperature data of the battery in real time, and use compressed sensing technology to optimize data acquisition efficiency, reduce redundant data, and increase acquisition speed; Image reconstruction and optimization: Iterative reconstruction method is used to accelerate image processing, and GPU acceleration technology is combined to achieve fast and high-quality 3D image reconstruction, ensuring that accurate structural data inside the battery is obtained within a limited time; Dynamic monitoring and analysis: Using deep learning algorithms and spatiotemporal data analysis technology, CT images, thermal imaging and electrochemical data are integrated to track battery status changes in real time to assess the battery's health and potential risks.

[0037] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically introduced below in conjunction with the accompanying drawings of the specification. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0038] Example 1: Figure 1 As shown, A method for dynamically monitoring the state of a lithium-ion battery, the method comprising: S100 collects image data obtained by reflection or attenuation of lithium batteries after X-ray penetration at different time points. These images represent the attenuation of X-rays by different materials inside the battery. The image data can be obtained by scanning lithium-ion batteries using a micro CT scanner. During the battery charging, discharging or working process, scans are performed regularly or at specific time points to capture changes inside the battery. During the scan, X-rays penetrate the battery and interact with different battery materials. Different materials attenuate X-rays to different degrees, generating projection images of different areas inside the battery, that is, image data obtained by reflection or attenuation of the lithium battery after X-ray penetration; Specifically: It can use a high-resolution micro-CT scanner to scan lithium-ion batteries. The micro-CT scanner penetrates the battery with X-rays and collects the X-ray data after penetration. These data reflect the attenuation degree of X-rays by different materials inside the battery (such as electrodes, diaphragms, etc.), thereby providing basic data for subsequent three-dimensional image reconstruction. The scanning process of micro-CT is to scan the battery at multiple angles to generate projection image data at different angles; During the scanning process, the battery can be scanned regularly during the charging and discharging process. Or according to the working state of the battery, the scanning time point can be set (for example, at regular intervals or at the key moments of charging and discharging) to obtain the dynamic change information inside the battery. Through this regular scanning, the microscopic changes of the electrode materials inside the battery (such as crack expansion of electrode particles, changes in porosity, etc.) can be captured.

[0039] The data collected by micro-CT are two-dimensional projection images, which record the attenuation intensity of X-rays after penetrating different materials. These two-dimensional images are usually presented in the form of grayscale images, and the high and low grayscale values ​​represent the density differences of different materials. Through subsequent reconstruction algorithms, these two-dimensional images can be converted into three-dimensional internal structures of batteries.

[0040] S101: Process the three-dimensional image inside the battery through compressed sensing to reduce unnecessary sampling points, so as to further shorten the interval time of each sampling: Specifically: The detailed steps are as follows: Acquire the three-dimensional image information of the inside of the battery obtained in step S200; Perform edge detection and feature segmentation on the three-dimensional image information inside the battery to extract key areas, such as electrode material distribution, crack starting point, and pore area; Using machine learning methods (such as U-Net, ResNet, etc.) to mark key areas, identify high-risk areas, and use sparse transforms (such as wavelet transform, Fourier transform) to represent data, so that the data presents sparse characteristics. This step is because the inventors found that the internal structure of the battery has spatial sparsity (that is, the structure of most areas changes little, and only local areas change), so they choose to use sparse transforms (such as wavelet transform, Fourier transform) to represent data, so that the data presents sparse characteristics; By using a random projection matrix (such as a Gaussian random matrix or a Toeplitz matrix) to perform a compressed sensing transformation on the three-dimensional image inside the battery, an optimized scanning path is generated, so that the optimized measurement matrix can be used to preferentially collect data in areas with large changes (such as cracks, pores, and hot spots), while data compression is performed on areas with small structural changes. Use compressed sensing optimized sampling point selection algorithms (such as greedy algorithms, OMP) to generate new scanning paths, reduce repeated and unnecessary scanning points, and focus on the structural change area; For areas with obvious structural changes (such as crack extension areas and areas with severe polarization), increase the scanning density and improve the resolution; For areas with less variation (such as areas of uniform electrode material), the scanning frequency is reduced to reduce the amount of data.

[0041] Use sparse reconstruction algorithms (such as L1 norm minimization and Bayesian compressed sensing) to restore the collected compressed data and calculate the reconstruction error (such as mean square error MSE and structural similarity SSIM) to ensure that the error between the optimized scan data and the complete scan data is within an acceptable range so that the data quality is not affected; Combined with deep learning models (such as LSTM or Transformer), adaptive learning is performed on the scan data to continuously optimize the scan path and improve efficiency. By comparing the data collection results before and after optimization, the compressed sensing parameters can be adjusted to further improve the scanning accuracy while further reducing redundant calculations and improving the efficiency of the entire battery monitoring system. S200 processes the collected image data through a reconstruction algorithm to obtain three-dimensional images of the battery at different time points, and extracts key structural information inside the battery through the three-dimensional images (such as electrode particles, cracks, pores, etc.); Specifically: The goal of this step S200 is to convert the collected raw image data into an accurate three-dimensional image through a series of data processing and reconstruction algorithms, and extract the key structural information inside the battery. Through noise removal, smoothing, artifact correction and other technologies, ensure that the image quality reaches a high level of accuracy; use iterative reconstruction methods to continuously optimize the accuracy of the image and generate a three-dimensional image; finally, through three-dimensional image processing and visualization technology, generate an interactive three-dimensional model to help users analyze the internal structural changes of the battery. These processing steps provide a solid data foundation for the subsequent battery health status assessment.

[0042] The specific implementation steps of step S200 of this application are as follows: S210: removing noise from the image using a filtering algorithm and performing smoothing on the image data to reduce the impact of the noise on subsequent reconstruction and remove artifacts that may occur during rapid acquisition; Specifically, the inventors found that the collected image data usually contains some noise, which may be caused by factors such as equipment errors and interference from the scanning environment. Therefore, in this step, the original image data is de-noised by using a filtering algorithm. The filtering algorithm identifies and removes random noise in the image based on the grayscale changes and spatial characteristics of the image.

[0043] The original image processed by the filtering algorithm is reprocessed through a smoothing algorithm to further reduce the excessive noise in the details of the image, thereby avoiding affecting the accuracy of the image in the subsequent reconstruction process. The smoothed image is more conducive to clearly presenting the main features of the internal structure of the battery and avoiding the negative impact of noise on subsequent processing; The inventors have found that during the rapid scanning process, artifacts (artifacts refer to false or inaccurate image information) may be generated due to incomplete sampling or equipment errors. Therefore, the smoothed image needs to be corrected by a specific correction algorithm (such as local adjustment and model correction) to eliminate these artifacts, ensure the accuracy and authenticity of the image data, and further improve its imaging quality.

[0044] S220 uses an iterative reconstruction method (such as algebraic reconstruction technology ART) to perform preliminary image reconstruction on the image data after noise removal, artifact correction and standardization, so that the accuracy of the image can be gradually improved. Through iteration, the reconstruction error can be continuously reduced until the reconstruction result converges to generate preliminary 3D reconstructed images. These images can reflect the distribution of the internal structure of the battery, including details such as electrode particles and cracks. Specifically, the image data after noise removal, artifact correction and standardization is used as input, and the reconstruction algorithm is used to perform preliminary image reconstruction. This process uses iterative reconstruction methods (such as algebraic reconstruction technology ART) to construct a preliminary three-dimensional image. Algebraic reconstruction technology ART gradually approaches the real three-dimensional structure by continuously adjusting and optimizing image data in stages; During the initial reconstruction process, the image may have certain errors and inaccuracies. Therefore, the reconstruction process is an iterative process. In each round of iteration, the algorithm will adjust the error according to the reconstruction result, gradually reduce the error, until the reconstruction result converges and generates an accurate three-dimensional image. Therefore, each iterative optimization process is an improvement in image accuracy. The final three-dimensional image will be able to accurately reflect the structural details inside the battery, such as the morphology of electrode particles, the distribution of cracks, and the structure of pores.

[0045] S230 uses 3D image processing and visualization software (such as ParaView, VTK, etc.) to load 3D image data into a visualization environment, such as applying 3D rendering technology to generate interactive 3D models, allowing its users to intuitively view the internal structure of the battery to further analyze cracks, porosity, electrode particle morphology, etc.

[0046] Specifically: the iteratively optimized 3D image is loaded into a visualization environment through 3D image processing and visualization software (such as ParaView, VTK, etc.). In this environment, the image data is converted into an operable 3D model, allowing users to view the internal structure of the battery through interactive operations; It can also further visualize the three-dimensional image of the battery through three-dimensional rendering technology, vividly presenting the internal structure of the battery. Three-dimensional rendering technology can simulate real image details, such as density differences in different areas, the direction of cracks, etc., to help users understand the changes inside the battery more intuitively; By visualizing the final generated 3D image model, users can interact with the model. Users can view the internal structure of the battery from different angles and zoom in on certain areas for detailed analysis. This step helps to deeply analyze key factors such as cracks, porosity, and electrode particle morphology inside the battery, so as to further evaluate the health of the battery.

[0047] S300 obtains information about changes in the internal structure of the battery by performing detailed analysis of the three-dimensional image data at different time points. It also analyzes and monitors changes in the electrode particles, crack evolution, and porosity inside the battery to further understand the health status of the battery and potential failure risks. The following is a detailed operation flow of this step; S310 performs particle analysis on the three-dimensional images inside the battery at different times, tracking the changes in the morphology and size of the electrode particles to obtain the microstructural changes of the electrode material; Specifically: During the charge and discharge cycle of the battery, the electrode particles will undergo certain changes. Therefore, 3D image data is used to accurately observe the morphology and size changes of the electrode particles inside the battery. By performing particle analysis on 3D images at different times, the evolution of the electrode particles during the working process can be effectively captured; It tracks the electrode particles at each time point and analyzes their morphological changes (such as particle enlargement, rupture or rearrangement) and size changes (such as particle shrinkage or expansion). This helps to deeply understand the microstructural changes of electrode materials and then analyze the battery's charge and discharge performance, stability and decay; Through the particle analysis algorithm, the specific morphological characteristics and size change data of electrode particles can be extracted. This information helps to evaluate the degradation process of battery performance and identify material damage and its possible failure mode.

[0048] S320 regularly scans and compares 3D images at different time points to identify and track the evolution of cracks inside the battery. It also uses image analysis algorithms to extract the location, length, width and other features of the cracks and analyze their expansion trends, so that it can regularly obtain the evolution of cracks inside the battery. Specifically: It compares 3D images at different time points to analyze whether cracks are generated and expanded inside the battery. The image analysis algorithm can identify the location, shape and expansion path of the cracks and accurately track the evolution of the cracks inside the battery. In addition, this community can further extract key features such as crack length, width, and depth based on crack tracking. Through the analysis of these features, it can be used as important information to evaluate the crack expansion trend and its impact on battery performance; it can be combined with image analysis results to analyze the crack expansion trend and predict crack problems that may occur in the battery under future working conditions. This is of great significance for predicting the time and cause of battery failure.

[0049] S330 measures the porosity and pore distribution inside the battery through three-dimensional images, uses image segmentation technology to extract the pore area inside the battery, and quantitatively analyzes the porosity and its distribution changes; it tracks the evolution of the pores inside the battery, identifies the increase or change trend of the pores, and then analyzes the reasons for the degradation of battery performance.

[0050] Specifically: It uses 3D image data to accurately measure the porosity and pore distribution inside the battery. Through image segmentation technology, the pore area inside the battery can be extracted and the porosity can be calculated. This helps identify uneven phenomena such as voids and tiny cracks that may appear inside the battery. And by scanning the lithium battery regularly, we can obtain the pore data at different time points. By comparing the images at different time points, we can track the increase or distribution of pores, which can serve as an important reference condition for analyzing the degradation process of battery performance; This allows the quantitative analysis of porosity and its distribution changes to further identify trends in pore growth or other structural changes. Porosity growth usually indicates damage or degradation of battery materials, so analyzing pore changes helps determine whether the battery has performance degradation or potential failure risks.

[0051] S340: Comprehensively analyze the morphological changes, crack evolution and porosity changes of electrode particles. Use data fusion technology to combine these structural change data with the temperature, pressure and electrochemical data of the battery to obtain a comprehensive assessment of the internal structural changes of the battery.

[0052] By combining the data from various analysis modules and using machine learning algorithms or data-driven models to assess the structural health status, it is possible to determine whether the battery is in a healthy state and whether there is a potential risk of failure or malfunction.

[0053] Finally, a comprehensive structural change report is generated, which contains information such as electrode particles, cracks and porosity, and provides health status assessment results. This report is helpful for battery health monitoring, life prediction and failure mechanism analysis.

[0054] Specifically: This step uses data fusion technology based on the analysis of changes in the internal structure of the battery to combine information such as changes in electrode particle morphology, crack evolution, and porosity with other battery status data (such as temperature, pressure, electrochemical data, etc.) for comprehensive evaluation. Data fusion technology can integrate multi-source data into a more complete analysis framework, which helps to comprehensively evaluate the health status of the battery; By combining the data from each analysis module, the structural health status is evaluated using machine learning algorithms or data-driven models. Machine learning algorithms can identify the health level of the battery by learning historical data and current status, and determine whether it is in normal working condition and whether there is a risk of potential failure or failure. Based on the above comprehensive analysis results, it can generate a detailed structural change report. The report will cover important information such as electrode particles, cracks, porosity, etc. inside the battery, and provide battery health status assessment results. The health status assessment report can be used for battery health monitoring, life prediction and failure mechanism analysis to help users better understand the battery status.

[0055] S400 collects temperature distribution data on the battery surface through thermal imaging technology, and combines it with the internal structure change data to analyze the thermal response behavior of the battery. Thermal response analysis can reveal the temperature changes caused by internal structural changes (such as cracks, pores, overheating, etc.) during the battery charging and discharging process, and provide a basis for identifying potential thermal runaway risks. The specific steps are as follows: S410 collects thermal imaging data of the battery surface at different time points, analyzes the temperature distribution and thermal change trend of the battery surface through the thermal imaging data, and determines whether there is a hot spot area or abnormal temperature; Specifically: This step uses thermal imaging technology to regularly collect temperature distribution data on the battery surface under different working conditions. The thermal imaging camera can detect and record temperature changes on the battery surface in real time, thereby obtaining hot spots and other temperature anomalies that may occur during the battery's charge and discharge process; The inventors found that during the charging and discharging process, the battery may have local overheating problems due to certain local faults (such as electrode short circuit, diaphragm damage, etc.), and the collected thermal imaging data can reflect the temperature distribution on the battery surface. By analyzing these data, it is possible to identify whether there are areas of abnormal temperature rise (such as hot spots) on the battery surface. At the same time, the trend of temperature changes (such as heating rate, temperature fluctuations, etc.) can also reflect the hot spots on the surface and whether the battery has overheating problems during operation; S420 combines thermal imaging data with 3D image data to analyze the relationship between the thermal response and internal structure of the battery during charging and discharging, identify whether there are overheated areas, and whether there is a potential risk of thermal runaway. By comparing the relationship between the battery surface temperature and the internal structure of the battery (such as pores, cracks, etc.), the safety of the battery is evaluated.

[0056] Specifically, this step combines thermal imaging data with the three-dimensional structural image data inside the battery to further analyze the relationship between the structural changes inside the battery and the surface temperature changes. The three-dimensional image data provides detailed information about the internal structure of the battery (such as cracks, pores, etc.), while the thermal imaging data provides the distribution of the battery surface temperature. By combining these two types of data, the risk of thermal runaway inside the battery can be more accurately identified; The inventors found that during the battery charging and discharging process, changes in the internal structure of the battery (such as crack expansion, increase in pores, etc.) may cause abnormal accumulation or dissipation of heat. By analyzing the relationship between thermal imaging data and 3D image data, it is possible to reveal which internal structural changes (such as crack expansion or pore increase) are related to the increase in surface temperature, and then evaluate whether these structural changes will threaten the thermal stability of the battery; Therefore, this step combines thermal imaging data with internal structure information to determine whether there are overheated areas in the battery. These overheated areas may be a sign of local battery failure, especially when there are internal damages such as cracks or pores. This analysis can help assess whether the battery is at potential risk of thermal runaway, thereby identifying potential hazards that may lead to battery safety issues in advance.

[0057] S500 collects electrochemical data of lithium-ion batteries at different time points, that is, extracts the working performance characteristics of the battery from the electrochemical data (battery voltage, current, power, internal resistance and other key data) and calculates key indicators such as battery charge and discharge efficiency, internal resistance change and cycle performance through electrochemical analysis algorithms, so that it can obtain real-time battery status and battery health indicators through the battery management system (BMS); Specifically: This step collects electrochemical data of lithium-ion batteries periodically or according to the working cycle of the battery (such as charging and discharging process, etc.). Electrochemical data includes but is not limited to key parameters such as battery voltage, current, power, and internal resistance. The battery voltage reflects the potential state of the battery, the current and power reveal the energy output or input of the battery, and the internal resistance provides the internal impedance information of the battery, which are all important indicators of battery performance; When collecting these electrochemical data, synchronize them with other types of data (such as image data and thermal imaging data) in time to fully evaluate the comprehensive performance of the battery under various working conditions, and collect and analyze the electrochemical data of the battery in real time through the battery management system (BMS). These real-time data will be used to track the current working status of the battery, including the battery's power level, health status, etc., to ensure that the battery operates within the optimal performance range.

[0058] The S600 obtains 3D image data, thermal imaging data, and electrochemical data, and uses machine learning algorithms (such as deep learning, Kalman filtering, particle filtering, etc.) to analyze the multimodal information of the battery in real time to obtain the battery status information. Through multi-dimensional data fusion, the real-time health status of the battery is obtained, including a comprehensive evaluation of the battery's internal structure, thermal response, and electrochemical performance.

[0059] The specific steps are as follows: S610 obtains three-dimensional image data, thermal imaging data and electrochemical data, and uses machine learning algorithms to analyze multi-modal information of the battery in real time. The method for obtaining battery status information is as follows; Specifically: This step collects data sources from different sensors, including three-dimensional image data, thermal imaging data, and electrochemical data of the battery. These data provide internal structure information, thermal response information, and electrochemical performance data of the battery, respectively, and use machine learning algorithms (such as deep learning, Kalman filtering, particle filtering, etc.) to analyze these multimodal data in real time to extract battery status information. Machine learning algorithms can identify complex relationships between data and help analyze the health status of the battery; S620 synchronizes the acquired image data, thermal imaging data and electrochemical data by accurately matching the timestamp and sampling period; to ensure that the temperature change, internal structure evolution and electrochemical parameters (such as voltage and current) at each time point can be associated together.

[0060] Specifically, in order to ensure consistency and accuracy between data, it is first necessary to ensure that different data sources (such as three-dimensional images, thermal imaging, and electrochemical data) have consistent time synchronization. By accurately matching the timestamp and sampling period of the data, ensure that the temperature changes, internal structure evolution, and electrochemical parameters (such as voltage, current, etc.) at each moment can correspond, and ensure that all data are collected in the same time period and marked with the same timestamp to facilitate subsequent correlation analysis. Through synchronized data, the different performance of batteries can be accurately compared and provide a reliable basis for health assessment.

[0061] S630 aligns data from different sources to obtain multimodal data of the battery, ensuring that they are compared within the same time series; for example, the battery's three-dimensional image data, thermal imaging data, and electrochemical data are collected and timestamped within the same time period to facilitate subsequent analysis.

[0062] Specifically: align multimodal data from different sensors to ensure that they can be compared and analyzed in the same time series. For example, organize the battery's three-dimensional image data, thermal imaging data, and electrochemical data in chronological order to make them consistent in subsequent analysis; and annotate each data source (such as image, temperature, chemical properties, etc.) to ensure that the data can accurately reflect the performance of the battery under different working conditions and integrate multimodal data.

[0063] The S640 performs preliminary cleaning on multimodal data to remove noise. For example, for thermal imaging data, it removes temperature fluctuations caused by environmental factors or equipment errors. For CT data, it removes artifacts and errors in the image. It also filters electrochemical data to eliminate high-frequency noise, ensuring that the data read by each sensor is accurate.

[0064] Extract the microstructural features of the battery from the 3D image data in the multimodal data; for example, identify and extract features such as electrode particle morphology, cracks, porosity, etc.

[0065] And calculate the particle distribution, crack length and porosity of each area of ​​the battery to provide basic data for subsequent health assessment; Temperature change characteristics are extracted from thermal imaging data to obtain the changing trend and abnormal areas of battery surface temperature. The thermal response characteristics of the battery can be evaluated by calculating the temperature rise rate and thermal peak position of the hot spot area, and the temporal and spatial characteristics related to temperature can be extracted, such as the delay time of thermal response and the rate of heating process. This information is helpful for analyzing the development of faults and thermal runaway risks inside the battery.

[0066] The battery's operating performance characteristics are extracted from electrochemical data (battery voltage, current, power, internal resistance and other key data) and key indicators such as battery charge and discharge efficiency, internal resistance change and cycle performance are calculated through electrochemical analysis algorithms, so that the real-time battery status can be obtained through the battery management system (BMS) and the battery health indicators can be obtained.

[0067] Extract battery health indicators, such as SOC (State of Charge) and SOH (State of Health), and use them for subsequent data fusion analysis.

[0068] Specifically, the present application removes noise by cleaning the collected multimodal data. For thermal imaging data, remove temperature fluctuations caused by environmental factors or equipment errors; for three-dimensional image data, remove artifacts and errors in the image; for electrochemical data, use filtering algorithms to remove high-frequency noise to ensure that the data read by each sensor is true and accurate, so that it can ensure the data quality in subsequent analysis through the data cleaning process in this step, thereby improving the accuracy of subsequent modeling and health assessment; S650 combines multimodal data (CT image data, thermal imaging data and electrochemical data) from different sources through data fusion technology; these data can be fused through methods such as Kalman filtering, particle filtering, and weighted averaging.

[0069] For thermal imaging and electrochemical data, they can be combined with CT image data through weighted average or correlation analysis methods, focusing on analyzing the relationship between internal structural changes of the battery (such as cracks and pores) and thermal response and electrochemical performance (such as temperature and voltage).

[0070] Through algorithmic models (such as deep learning networks), different data sets are integrated into a comprehensive data source to further analyze the health status of the battery.

[0071] Specifically: This step can merge multimodal data from different sources through data fusion techniques (such as Kalman filtering, particle filtering, weighted averaging, etc.). Combining the three-dimensional image of the battery, thermal imaging data and electrochemical data, the fusion of these data helps to analyze the relationship between the internal structural changes of the battery (such as cracks, pores, etc.) and the thermal response and electrochemical performance (such as temperature, voltage, etc.).

[0072] Then, various types of data are integrated through weighted average or correlation analysis methods, and in-depth analysis is performed using the comprehensive data source of the battery to generate a battery health status assessment report. This process uses machine learning algorithms (such as deep learning networks) to comprehensively analyze the performance of the battery and assess the battery's health status, potential failure risks, etc.

[0073] S660 combines spatiotemporal tracking algorithms (such as spatiotemporal convolutional neural networks ST-CNN or LSTM) to perform spatiotemporal tracking of the collected multimodal data, analyze the spatiotemporal dynamic correlation between the battery's structural changes, thermal response, and electrochemical performance at different time points, and evaluate the battery's behavior patterns under different working conditions.

[0074] Through this spatiotemporal correlation analysis, the behavior patterns of the battery under different working conditions are evaluated and its potential sources of failure can be identified, especially whether the stress changes and temperature changes inside the battery during the charge and discharge process are consistent with the changes in electrochemical performance.

[0075] Specifically: combined with spatiotemporal tracking algorithms (such as spatiotemporal convolutional neural networks (ST-CNN) or long short-term memory networks (LSTM), the collected multimodal data can be spatiotemporally tracked and analyzed. Spatiotemporal tracking can analyze the dynamic relationship between the structural changes, thermal responses, and electrochemical properties of the battery at different time points; it can combine three-dimensional image data, thermal imaging data, and electrochemical data through spatiotemporal analysis to evaluate their dynamic changes in time and space. For example, as the battery charge and discharge process proceeds, there may be time delays or cross-effects in the stress, temperature, and electrochemical performance changes inside the battery. Spatiotemporal tracking analysis can reveal the spatiotemporal relationship of these changes, so that it can evaluate the behavior patterns of the battery under different working conditions through spatiotemporal correlation analysis. This includes whether the stress changes, temperature changes, and electrochemical performance changes inside the battery during the charge and discharge process are consistent. This analysis helps to identify potential sources of battery failure or thermal runaway risks and ensure the safety of the battery.

[0076] S700 builds a battery health status assessment model through machine learning (such as support vector machine SVM, random forest RF, etc.) or deep learning (such as convolutional neural network CNN, long short-term memory network LSTM, etc.); Specifically: The model input of the health status assessment model includes multimodal data features at different time points, such as three-dimensional image data, temperature change characteristics and electrochemical performance characteristics, and the output is a health status assessment value of the battery; It can use a large amount of historical data for model training and optimization to ensure that the evaluation model can accurately identify the health status and potential failures of the battery. During the training process, methods such as cross-validation can be used to improve the generalization ability of the model.

[0077] Optimize the evaluation model so that it can adapt to the health status assessment needs of different battery types and different working environments.

[0078] The trained model is used to evaluate the battery's health status in real time, predicting the battery's remaining life, fault sources, and possible failure mechanisms.

[0079] Based on the model's output, corresponding operational recommendations are given, such as whether safety measures need to be taken, whether charging restrictions should be imposed, whether the battery needs to be replaced, etc. The S701 combines and analyzes multiple data from different sources (such as micro-CT data, thermal imaging data, and electrochemical data) to provide a comprehensive battery health status assessment; Specifically, it collects the electrochemical data of the lithium-ion battery, such as voltage, current, power, internal resistance, etc. at different time points, and calculates the battery's charging and discharging efficiency, internal resistance and other parameters through electrochemical analysis algorithms; Obtain 3D image data, thermal imaging data and electrochemical data, and use machine learning algorithms (such as deep learning, Kalman filtering, particle filtering, etc.) to analyze the multimodal information of the battery in real time to obtain the battery status information. Through multi-dimensional data fusion, obtain the real-time health status of the battery, including a comprehensive evaluation of the battery's internal structure, thermal response and electrochemical performance.

[0080] S702 generates a battery health assessment report based on the assessment results of the health status assessment model, including a comprehensive assessment of the battery's internal structural changes, thermal response mode, electrochemical performance, and health status.

[0081] Specifically, this step generates a battery health assessment report based on the assessment results of the health status assessment model. The report contains comprehensive assessment results of the battery's internal structural changes (such as cracks, porosity, etc.), thermal response patterns (such as temperature changes, hot spots, etc.), and electrochemical performance (such as charge and discharge efficiency, internal resistance, etc.).

[0082] A system for dynamically monitoring the state of a lithium-ion battery, the system comprising: Data acquisition module 10: used to collect multimodal data of lithium-ion batteries in real time, including image data, thermal imaging data, and electrochemical data.

[0083] This module includes the following units: Micro-CT scanning unit: responsible for collecting three-dimensional CT image data inside the battery, providing detailed information on the internal structure of the battery, such as electrode particle morphology, cracks, porosity, etc.

[0084] Thermal imaging unit: acquires temperature change data on the battery surface in real time, provides thermal response information of the battery, and helps identify whether the battery is at risk of thermal runaway.

[0085] Electrochemical data acquisition unit: collects the battery's voltage, current, internal resistance and other electrochemical parameters, and provides information such as the battery's charge and discharge efficiency and internal resistance.

[0086] Data preprocessing and optimization module 20: used to preliminarily process and optimize the collected raw data to ensure that the data quality is suitable for subsequent analysis and modeling; This module includes the following units: Data denoising unit: denoises the noise in thermal imaging, CT images and electrochemical data to improve data quality and remove irrelevant interference; Data standardization unit: standardizes data from different sources to ensure comparability and consistency between modal data, facilitating subsequent fusion analysis; Image reconstruction and optimization module 30: used to reconstruct image data obtained by reflection or attenuation of lithium batteries after X-ray penetration at different time points, generate high-quality three-dimensional images, and improve data acquisition efficiency through technologies such as compressed sensing.

[0087] This module includes the following units: CT data reconstruction unit: through iterative reconstruction algorithm, the CT projection data is converted into a three-dimensional reconstructed image to present the structural details inside the battery; Image optimization unit: performs denoising and sharpening on the reconstructed image to enhance the clarity and accuracy of the image; Multi-resolution reconstruction unit: adjusts the reconstruction resolution according to the requirements of different areas of the battery to ensure high-precision reconstruction of key areas; Structural change analysis module 40: Analyzes changes in the internal structure of the battery through a three-dimensional image of the battery, and monitors microstructural changes that occur in the battery during the charge and discharge process, such as crack evolution, pore changes, etc.

[0088] This module includes the following units: Particle morphology analysis unit: Analyze the morphology changes of battery electrode particles through three-dimensional image data and evaluate their impact on battery performance; Crack evolution monitoring unit: monitors the generation and expansion of cracks inside the battery and evaluates the impact of cracks on battery life and safety; Porosity analysis unit: extract the battery pore area through image segmentation technology, and quantitatively analyze the porosity and its change trend; Thermal response analysis module 50: used to monitor the thermal changes of the battery through thermal imaging data to identify the thermal response pattern of the battery and whether there is a risk of thermal runaway.

[0089] This module includes the following units: Thermal map analysis unit: Generates battery surface temperature distribution map based on thermal imaging data to identify possible thermal hot spots in the battery; Thermal runaway risk assessment unit: assesses whether the battery has a potential risk of thermal runaway based on thermal response data and battery internal structure information; Multimodal data fusion and analysis module 60: combines multiple data such as micro-CT, thermal imaging, and electrochemical data for comprehensive analysis to improve the accuracy and comprehensiveness of battery status monitoring.

[0090] This module includes the following units: Data fusion unit: Through Kalman filtering, particle filtering and other algorithms, the information of various data sources is integrated to ensure the mutual complementation of data of various modalities; Spatiotemporal tracking and analysis unit: Analyzes multimodal data at different time points based on deep learning models (such as ST-CNN, LSTM, etc.) to track the dynamic changes of battery status; Health status assessment generation module 70: Analyze multimodal data through machine learning or deep learning algorithms to generate a battery health status assessment model, so as to obtain the battery health status assessment value through the health status assessment model, so that the remaining life and potential failure risk of the battery can be predicted through this valuation.

[0091] This module includes the following units: Health status assessment unit: Analyzes multimodal data based on machine learning or deep learning algorithms to assess the health status of the battery and determine whether there are signs of failure or aging; Fault prediction unit: predicts the possible fault type and failure time of the battery by learning historical data and analyzing real-time data; Status feedback and instruction generation unit: Generates a battery health status report based on the evaluation results and predictive analysis, and provides targeted operational suggestions, such as whether the battery needs to be replaced or whether the usage strategy needs to be adjusted; Visualization and report generation module 80: visualizes the three-dimensional structure diagram, temperature distribution diagram, health status assessment diagram, etc. of the battery, so that users can intuitively understand the battery status; Automatically generate a battery status report based on the analysis results, providing battery health assessment, failure prediction and maintenance recommendations.

[0092] This module includes the following units: Data visualization unit: visualizes the battery's three-dimensional structure diagram, temperature distribution diagram, health status assessment diagram, etc., to facilitate users to intuitively understand the battery status; Report generation unit: Automatically generates battery status reports based on analysis results, providing battery health assessment, fault prediction and maintenance recommendations.

[0093] Through the above detailed description of a method and system for dynamically monitoring the status of a lithium-ion battery, those skilled in the art can clearly understand the system for dynamically monitoring the status of a lithium-ion battery in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0094] Through the above description of the disclosed embodiments, it is believed that the professional and technical personnel in the field can implement or use the present application. The various modifications to these embodiments will be apparent to the professional and technical personnel in the field, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will conform to the widest range consistent with the principles and novel features disclosed herein.

Claims

1. A method for dynamically monitoring the state of a lithium-ion battery, characterized in that: The method comprises: Collect image data obtained by reflection or attenuation of lithium batteries after X-ray penetration at different time points; The collected image data is processed by a reconstruction algorithm to obtain a three-dimensional image of the inside of the battery at different time points, and key structural information inside the battery is extracted through the three-dimensional image; Analyze the three-dimensional image data at different time points to obtain information on changes in the internal structure of the battery.

2. A method for dynamically monitoring the state of a lithium-ion battery according to claim 1, characterized in that: The method of analyzing the three-dimensional image data at different time points to obtain the change information of the internal structure of the battery is as follows; By performing particle analysis on the three-dimensional images inside the battery at different times, the morphology and size changes of the electrode particles are tracked to obtain the microstructural changes of the electrode material; By regularly scanning and comparing 3D images at different time points, the crack evolution process inside the battery can be identified and tracked. In combination with image analysis algorithms, the crack location, length, width and other features can be extracted, and its expansion trend can be analyzed to regularly obtain the crack evolution inside the battery. The porosity and pore distribution inside the battery are measured through three-dimensional images, the pore area inside the battery is extracted using image segmentation technology, and the porosity and its distribution changes are quantitatively analyzed.

3. The method for dynamically monitoring the state of a lithium-ion battery according to claim 1, characterized in that: The method further comprises: Collect thermal imaging data of the battery surface at different time points, analyze the temperature distribution and thermal change trend of the battery surface through the thermal imaging data, and determine whether there is a hot spot area or abnormal temperature; By combining thermal imaging data with three-dimensional image data, the relationship between the thermal response and internal structure of the battery during the charging and discharging process is analyzed to identify whether there are overheating areas and potential risks of thermal runaway.

4. The method for dynamically monitoring the status of a lithium-ion battery according to claim 1, characterized in that: The method of processing the collected image data through a reconstruction algorithm to obtain a three-dimensional image of the inside of the battery at different time points is as follows; Using a filtering algorithm to remove noise from the image data, so as to remove artifacts that may appear during rapid acquisition; The image data after noise removal, artifact correction and standardization are used for preliminary image reconstruction using an iterative reconstruction method to generate a preliminary three-dimensional reconstructed image.

5. The method for dynamically monitoring the status of a lithium-ion battery according to claim 1, characterized in that: The method further comprises: Collect electrochemical data of lithium-ion batteries at different time points; Obtain three-dimensional image data, thermal imaging data and electrochemical data, and use machine learning algorithms to analyze the multimodal information of the battery in real time to obtain the battery status information.

6. A method for dynamically monitoring the status of a lithium-ion battery according to claim 5, characterized in that: The method of obtaining three-dimensional image data, thermal imaging data and electrochemical data, and analyzing the multimodal information of the battery in real time through a machine learning algorithm to obtain the battery status information is as follows; The acquired image data, thermal imaging data and electrochemical data are synchronized in time by precise matching of time stamps and sampling periods; Aligning data from different sources to obtain multimodal data of the battery, and preliminarily cleaning the multimodal data to remove noise therein; Extract the microstructural features of the battery from the 3D image data in the multimodal data, and calculate the particle distribution, crack length, porosity and other features of each area of ​​the battery; Extract temperature variation features from thermal imaging data to obtain the temperature variation trend and abnormal areas on the battery surface, and extract the temporal and spatial features related to temperature; Extract the working performance characteristics of the battery from the electrochemical data, and calculate the key indicators such as the battery's charge and discharge efficiency, internal resistance change, and cycle performance through the electrochemical analysis algorithm to obtain the battery's health indicators; Merge multimodal data from different sources through data fusion technology; Different data sets are integrated into a comprehensive data source through algorithmic models.

7. A method for dynamically monitoring the state of a lithium-ion battery according to claim 6, characterized in that: The method further comprises: Combined with the space-time tracking algorithm, the collected multimodal data are tracked in space and time to analyze the space-time dynamic correlation between the battery's structural changes, thermal response and electrochemical performance at different time points.

8. A method for dynamically monitoring the state of a lithium-ion battery according to any one of claims 2 to 7, characterized in that: The method further comprises: Build a battery health status assessment model through machine learning or deep learning; The model input of the health status assessment model includes multimodal data features at different time points, such as three-dimensional image data, temperature change characteristics and electrochemical performance characteristics, and the output is the health status assessment value of the battery.

9. A method for dynamically monitoring the status of a lithium-ion battery according to claim 8, characterized in that: The method also includes: generating a battery health assessment report based on an assessment result of the health status assessment model.

10. A system for dynamically monitoring the status of a lithium-ion battery, characterized in that: The system is used to implement a method for dynamically monitoring the state of a lithium-ion battery according to any one of claims 1 to 8, and the system comprises: Data acquisition module: used to collect multimodal data of lithium-ion batteries in real time, including image data, thermal imaging data, and electrochemical data; Data preprocessing and optimization module: used to perform preliminary processing and optimization on the collected raw data; Image reconstruction and optimization module: used to reconstruct the image data obtained by reflection or attenuation of lithium batteries through X-ray penetration at different time points to generate high-quality three-dimensional images; Structural change analysis module: Analyze the changes in the internal structure of the battery through the three-dimensional image of the battery, and monitor the microscopic structural changes that occur during the charge and discharge process of the battery; Thermal response analysis module: used to monitor the thermal changes of the battery through thermal imaging data to identify the thermal response pattern of the battery and whether there is a risk of thermal runaway; Multimodal data fusion and analysis module: combines multiple data such as micro-CT, thermal imaging, and electrochemical data for comprehensive analysis; Health status assessment generation module: Analyze multimodal data through machine learning or deep learning algorithms to generate a battery health status assessment model to obtain the battery health status assessment value.

Images