A method and device for detecting micro-defects in battery manufacturing based on an ultrasound graph network

Abstract

The application belongs to the technical field of battery management, and provides a battery manufacturing micro defect detection method and device based on an ultrasonic graph network, which comprises the following steps: obtaining structural parameters of a battery electrode; applying a multi-frequency ultrasonic excitation to the battery electrode to obtain echo signals of different spatial positions of the electrode; analyzing and extracting characteristic parameters of the echo signals, constructing a plurality of analysis units and a characteristic parameter set thereof according to the spatial positions of the electrode; calculating physical intermediate variables of each analysis unit and propagation correlation parameters between different analysis units based on the characteristic parameter set of the plurality of analysis units, and constructing a fusion feature vector and a propagation correlation matrix; predicting the structural parameters of the battery electrode and the defect type of the battery electrode according to the fusion feature vector, the propagation correlation matrix and a graph neural network based on an acoustic propagation mechanism; constructing a structural dispersion index based on the prediction result of the structural parameters of the battery electrode, and combining the prediction result of the defect type of the battery electrode to perform hierarchical monitoring of the defect degree of the battery electrode, so that the defect detection of the battery electrode is completed.

Description

Technical Field

[0001] This invention belongs to the field of battery management technology, specifically relating to a method and device for detecting minute defects in battery manufacturing based on ultrasonic graph networks. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Uneven electrode thickness, abnormal porosity, and various physical defects that occur during battery operation can easily lead to serious consequences such as battery performance degradation, thermal runaway, and even fire and explosion, including shortened lifespan and safety hazards. Therefore, achieving precise monitoring of electrode defects is crucial to ensuring the safety and reliability of battery operation.

[0004] Current methods for detecting battery electrode defects include manual sampling, which involves cutting off sections of finished products and observing their microstructure under a microscope. However, this method suffers from problems such as destructive testing, low efficiency, and inability to cover full production. Alternatively, conventional non-destructive testing methods can be used, which involve full-scan imaging of the electrodes using X-rays or ultrasound. However, these methods suffer from low sensitivity in identifying localized micro-defects and insufficient detection accuracy. Finally, methods based on simple physical models can be used, which use a single physical parameter, such as velocity or attenuation, to estimate the electrode state. However, these methods struggle to handle complex acoustic propagation relationships and suffer from poor accuracy in multi-task joint prediction.

[0005] Based on the above-mentioned shortcomings, there is an urgent need for a battery electrode defect detection method that can deeply integrate acoustic physical characteristics and spatial correlation mechanisms to achieve high-sensitivity online joint prediction of multiple parameters, so as to ensure that the structural state and defect degree of each region of the electrode can be accurately and efficiently identified. Summary of the Invention

[0006] To address the aforementioned issues, this invention proposes a method and device for detecting minute defects in battery manufacturing based on an ultrasonic graph network. Through the deep fusion of multi-frequency ultrasonic excitation and an acoustic propagation mechanism graph neural network, it achieves refined monitoring of the degree of defects in battery electrodes.

[0007] According to some embodiments, the first aspect of the present invention provides a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks, employing the following technical solution: A method for detecting minute defects in battery manufacturing based on ultrasonic graph networks, comprising: Obtain the structural parameters of the battery electrodes; Multi-frequency ultrasonic excitation was applied to the battery electrodes to obtain echo signals from different spatial positions of the electrodes; The characteristic parameters of the acquired echo signal are analyzed and extracted, and several analysis units and their characteristic parameter sets are constructed according to the spatial position of the electrodes. Based on the feature parameter set of several constructed analysis units, calculate the physical intermediate variables of each analysis unit and the propagation correlation parameters between different analysis units, and construct the fused feature vector and propagation correlation matrix; Based on the fused feature vectors, propagation correlation matrix, and graph neural network based on acoustic propagation mechanism, the obtained battery electrode structural parameters and battery electrode defect types are predicted. Based on the predicted results of the battery electrode structural parameters, a structural dispersion index is constructed. Combined with the predicted results of battery electrode defect types, the degree of battery electrode defects is graded and monitored to complete the defect detection of the battery electrode.

[0008] As a further technical limitation, time-domain, frequency-domain, and dispersion joint analysis is performed on the echo signals from different spatial locations of the acquired electrodes to extract characteristic parameters that characterize local propagation properties. Analysis units and their corresponding characteristic parameter sets are constructed according to spatial location. In the process of constructing the characteristic parameter set, a two-dimensional coordinate system is established on the horizontal plane according to the physical size of the electrode, and the electrode surface is divided into a regular grid array according to the preset spatial resolution. The electrode entity region corresponding to each grid is an independent analysis unit, and the ultrasonic characteristic parameter set corresponding to each unit is obtained, thus completing the construction of the characteristic parameter set.

[0009] As a further technical limitation, the comprehensive deviation of each analysis unit relative to the mean of the whole region is calculated based on the obtained prediction results of the battery electrode structure parameters. A structural dispersion index is constructed based on the comprehensive deviation of each analysis unit. A first judgment threshold and a second judgment threshold are determined based on the standard comprehensive deviation. The battery defect level is classified by combining the comprehensive deviation and the determined first and second judgment thresholds.

[0010] Furthermore, when the overall deviation does not exceed the first judgment threshold, the battery electrode at that location has an intact physical structure and no obvious defects; when the overall deviation is between the first and second judgment thresholds, the battery electrode at that location has a slight parameter deviation or defect risk, defined as a sub-healthy state, requiring enhanced quality attention; when the overall deviation exceeds the second judgment threshold, the battery electrode at that location has a serious physical defect that is the average value among standard qualified electrode samples, triggering a scrap warning and marking it as a defective area.

[0011] As a further technical limitation, the loss function of the graph neural network based on the acoustic propagation mechanism... for ;in, These are the weighting coefficients; and These are the comprehensive task loss function and the physical constraint loss function, respectively. ; ; , and These represent the actual values ​​for electrode thickness, porosity, and defect type, respectively; these are the labels. , They are respectively and , and The functional relationship between them is obtained by fitting the training data.

[0012] As a further technical limitation, the structural parameters of the obtained battery electrode include electrode thickness and porosity.

[0013] According to some embodiments, a second aspect of the present invention provides a battery manufacturing micro-defect detection system based on ultrasonic graph networks, employing the following technical solution: A battery manufacturing micro-defect detection system based on ultrasonic graph networks includes: The first acquisition module is configured to acquire the structural parameters of the battery electrodes; The second acquisition module is configured to apply multi-frequency ultrasonic excitation to the battery electrode to acquire echo signals at different spatial positions of the electrode. The first construction module is configured to analyze and extract the characteristic parameters of the acquired echo signal, and construct several analysis units and their characteristic parameter sets according to the spatial position of the electrodes; The second construction module is configured to calculate the physical intermediate variables of each analysis unit and the propagation correlation parameters between different analysis units based on the feature parameter set of the constructed analysis units, and construct the fused feature vector and the propagation correlation matrix. The electrode prediction module is configured to predict the acquired battery electrode structural parameters and battery electrode defect types based on the fused feature vector, the propagation correlation matrix, and the graph neural network based on the acoustic propagation mechanism. The defect detection module is configured to construct a structural dispersion index based on the obtained battery electrode structural parameter prediction results, and combine the battery electrode defect type prediction results to perform graded monitoring of the degree of battery electrode defects, thereby completing the defect detection of the battery electrode.

[0014] According to some embodiments, a third aspect of the present invention provides a computer-readable storage medium, employing the following technical solution: A computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of a method for detecting minute defects in battery manufacturing based on an ultrasonic graph network as described in the first aspect of the present invention.

[0015] According to some embodiments, the fourth aspect of the present invention provides an electronic device, which adopts the following technical solution: An electronic device includes a memory, a processor, and a program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of a method for detecting minute defects in battery manufacturing based on an ultrasonic graph network as described in the first aspect of the present invention.

[0016] According to some embodiments, the fifth aspect of the present invention provides a computer program product, which adopts the following technical solution: A computer program product includes software code, wherein the program in the software code performs the steps of a method for detecting minute defects in battery manufacturing based on an ultrasonic graph network as described in the first aspect of the present invention.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention is based on a graph neural network for acoustic propagation mechanisms. By applying multi-frequency ultrasonic excitation to the battery electrodes and combining time-frequency and dispersion analysis, a propagation correlation matrix and feature vector are constructed, integrating physical intermediate variables such as equivalent propagation speed and attenuation. Iterative training is performed using the multi-layer information propagation structure of the graph neural network to achieve multi-task joint prediction and refined classification of electrode thickness, porosity, and defect types. This overcomes the problems of traditional detection methods, which struggle to handle complex acoustic propagation correlations and have low accuracy in multi-parameter prediction. By constructing a propagation correlation matrix using propagation correlation parameters, it can quantify and analyze the interaction relationships between units and significantly improve the sensitivity to identify local minor defects. It also possesses high-precision structural characterization capabilities. By constructing a structural dispersion index and setting a threshold, it can directly perform refined classification and monitoring of electrode defect severity without relying on a single qualitative judgment. Attached Figure Description

[0018] The accompanying drawings, which form part of this embodiment, are used to provide a further understanding of this embodiment. The illustrative embodiments and their descriptions are used to explain this embodiment and do not constitute an improper limitation of this embodiment.

[0019] Figure 1 This is a flowchart of a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks, according to Embodiment 1 of the present invention. Figure 2 This is a structural block diagram of a battery manufacturing micro-defect detection system based on ultrasonic graph networks, according to Embodiment 2 of the present invention. Detailed Implementation

[0020] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0021] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0022] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0023] In this invention, terms such as "upper," "lower," "left," "right," "front," "back," "vertical," "horizontal," "side," and "bottom" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used only to facilitate the description of the structural relationships of the various components or elements of this invention and do not specifically refer to any component or element in this invention. They should not be construed as limiting the invention.

[0024] In this invention, terms such as "fixed connection," "connected," and "linked" should be interpreted broadly, indicating a fixed connection, an integral connection, or a detachable connection; a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can determine the specific meaning of these terms in this invention based on the specific circumstances, and they should not be construed as limitations on the invention.

[0025] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0026] Example 1 Embodiment 1 of this invention introduces a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks.

[0027] like Figure 1 The method for detecting minute defects in battery manufacturing based on ultrasonic graph networks, as shown, includes: Obtain the structural parameters of the battery electrodes; Multi-frequency ultrasonic excitation was applied to the battery electrodes to obtain echo signals from different spatial positions of the electrodes; The characteristic parameters of the acquired echo signal are analyzed and extracted, and several analysis units and their characteristic parameter sets are constructed according to the spatial position of the electrodes. Based on the feature parameter set of several constructed analysis units, calculate the physical intermediate variables of each analysis unit and the propagation correlation parameters between different analysis units, and construct the fused feature vector and propagation correlation matrix; Based on the fused feature vectors, propagation correlation matrix, and graph neural network based on acoustic propagation mechanism, the obtained battery electrode structural parameters and battery electrode defect types are predicted. Based on the predicted results of the battery electrode structural parameters, a structural dispersion index is constructed. Combined with the predicted results of battery electrode defect types, the degree of battery electrode defects is graded and monitored to complete the defect detection of the battery electrode.

[0028] As one or more implementation methods, this embodiment collects the structural parameters of the battery electrode, applies multi-frequency ultrasonic excitation to the battery electrode, performs scanning detection on different spatial positions of the electrode, and collects the echo signals corresponding to each position.

[0029] The structural parameters in this embodiment include at least electrode thickness and porosity. The multi-frequency ultrasonic excitation is a broadband or multi-frequency combined ultrasonic signal based on a preset frequency band range. The preset frequency band range is determined according to the characteristics of the specific electrode material. The scanning detection process sets up the electrode to be partitioned or scanned point by point according to the electrode structure size and spatial resolution requirements, and simultaneously acquires the ultrasonic echo signal and its propagation response data corresponding to each detection position.

[0030] As one or more implementation methods, this embodiment performs time-domain, frequency-domain, and dispersion joint analysis on the echo signals collected at various locations, extracts characteristic parameters that characterize local propagation characteristics, and constructs analysis units and their corresponding characteristic parameter sets according to spatial location.

[0031] In this embodiment, the echo signals acquired at various spatial locations are subjected to joint analysis in the time domain, frequency domain, and dispersion domain to extract characteristic parameters for characterizing local ultrasonic propagation properties, including but not limited to time of flight, attenuation coefficient, and spectral characteristics. To suppress the global averaging effect of large-size electrode detection signals and improve the positioning accuracy and identification sensitivity of local micro-defects, a two-dimensional coordinate system is established on the horizontal plane according to the physical size of the electrode. The electrode surface is divided into a regular grid array according to a preset spatial resolution. The electrode entity region corresponding to each grid is an independent analysis unit, and a set of characteristic parameters corresponding to each unit is constructed.

[0032] As one or more implementation methods, this embodiment calculates physical intermediate variables such as equivalent propagation speed, attenuation, and structural uniformity based on the characteristic parameter set of each analysis unit, and constructs a fused feature vector by combining the propagation correlation between units; specifically: In this embodiment, each analysis unit is equivalent to a uniform acoustic medium, and based on this assumption, the physical intermediate variables corresponding to each analysis unit are calculated, including the equivalent propagation velocity parameter. Equivalent attenuation characterization parameters and structural uniformity characterization parameters Used to characterize local acoustic propagation characteristics, specifically: ; ; ; in, i Indicates the sequence number of the analysis unit. This indicates the length of the ultrasound propagation path within the analysis unit. Indicates the corresponding propagation time. and These represent the amplitudes of the ultrasonic signal at the input and output terminals of the analysis unit, respectively. Indicates the first [unit] within this analysis unit Ultrasonic feature values ​​of each sampling point This represents the mean value of the ultrasonic features within the analysis unit. Indicates the number of sampling points.

[0033] This embodiment calculates the propagation correlation parameters between different analysis units. Specifically: ; in, i , j Each represents a different analysis unit. Indicates the spatial distance between two analysis units. express and The arithmetic mean, express and The harmonic mean, This represents the reference propagation speed, which is determined based on actual conditions or experience. This embodiment concatenates physical intermediate variables with the feature parameter set to construct a fused feature vector. .

[0034] As one or more implementation methods, this embodiment constructs a propagation correlation matrix, inputs it and a fused feature vector into a graph neural network based on the acoustic propagation mechanism, and achieves joint prediction of electrode thickness, porosity, and defect type through iterative training; specifically: according to Constructing a propagation correlation matrix ,in ;Will Arrange the nodes in order of their numbers and construct a node feature matrix. ,in, The number of nodes; graph neural networks use and As input, node features are updated through a multi-layered information propagation structure.

[0035] This embodiment introduces a propagation mechanism based on propagation correlation parameters into the graph neural network to iteratively update node features. The update formula for a single node is as follows: ; in, Represents nodes A set of neighboring nodes that have a propagation relationship. For the first l The layer's trainable weight matrix is ​​obtained by training by minimizing the loss function. It is a non-linear activation function.

[0036] Based on the feature matrix of the final layer nodes By output mapping function The prediction result is as follows: ; in, The model outputs a matrix, where each row corresponds to the predicted value of a node, specifically including the electrode thickness. Porosity and defect types The multi-task outputs are all based on supervised learning using labeled training data, which includes the actual structural parameters and defect type annotation information of the corresponding analysis unit, and are paired with the ultrasound data at the corresponding location to construct training samples.

[0037] A loss function is constructed based on acoustic propagation laws and physical constraints. The specific formula is as follows: ; in, These are weighting coefficients, and their values ​​are determined through parameter tuning during model training. and These are the comprehensive task loss function and the physical constraint loss function, respectively, as follows: ; ; in, , and These represent the actual values ​​for electrode thickness, porosity, and defect type, respectively; these are the labels. , They are respectively and , and The functional relationship between them is obtained by fitting the training data.

[0038] In this embodiment, the graph neural network is iteratively trained by minimizing the loss function to obtain a converged model. Based on the trained model, inference calculations are performed on each node to output the electrode thickness, porosity, and defect type.

[0039] As one or more implementation methods, this embodiment constructs a structural dispersion index based on the predicted electrode thickness and porosity, and sets a threshold to classify the degree of electrode defects, thereby achieving refined monitoring of electrode defects; specifically: Calculate the overall deviation of each unit relative to the mean of the entire region, i.e. Determine the first judgment threshold. Second determination threshold ,Right now ; ;in, This represents the overall deviation of all units in the standard qualified electrode sample.

[0040] This embodiment is based on The relationship with the threshold classifies the state of the electrode unit into three defect severity levels: if The physical structure at that location is determined to be intact and without obvious defects; if If a slight parameter deviation or defect risk is detected at this location, it is defined as a sub-healthy state, requiring increased quality attention; if If the location is determined to have a serious physical defect, which is the average value among the standard qualified electrode samples, a scrap warning will be triggered and the area will be marked as a defective area.

[0041] This embodiment is based on a graph neural network for acoustic propagation mechanisms. By applying multi-frequency ultrasonic excitation to the battery electrodes and combining time-frequency and dispersion analysis, a propagation correlation matrix and feature vector are constructed, which integrate physical intermediate variables such as equivalent propagation speed and attenuation. Iterative training is performed using the multi-layer information propagation structure of the graph neural network to achieve multi-task joint prediction and refined classification of electrode thickness, porosity, and defect types. This overcomes the problems of traditional detection methods, which struggle to handle complex acoustic propagation correlations and have low accuracy in multi-parameter prediction. By constructing a propagation correlation matrix using propagation correlation parameters, it can quantify and analyze the interaction relationships between units and significantly improve the sensitivity to identify local minor defects. It also has high-precision structural characterization capabilities. By constructing a structural dispersion index and setting a threshold, it can directly perform refined classification and monitoring of electrode defect severity without relying on a single qualitative judgment.

[0042] Example 2 Embodiment 2 of the present invention introduces a battery manufacturing micro-defect detection system based on ultrasonic graph networks.

[0043] like Figure 2 The battery manufacturing micro-defect detection system based on ultrasonic graph networks, as shown, includes: The first acquisition module is configured to acquire the structural parameters of the battery electrodes; The second acquisition module is configured to apply multi-frequency ultrasonic excitation to the battery electrode to acquire echo signals at different spatial positions of the electrode. The first construction module is configured to analyze and extract the characteristic parameters of the acquired echo signal, and construct several analysis units and their characteristic parameter sets according to the spatial position of the electrodes; The second construction module is configured to calculate the physical intermediate variables of each analysis unit and the propagation correlation parameters between different analysis units based on the feature parameter set of the constructed analysis units, and construct the fused feature vector and the propagation correlation matrix. The electrode prediction module is configured to predict the acquired battery electrode structural parameters and battery electrode defect types based on the fused feature vector, the propagation correlation matrix, and the graph neural network based on the acoustic propagation mechanism. The defect detection module is configured to construct a structural dispersion index based on the obtained battery electrode structural parameter prediction results, and combine the battery electrode defect type prediction results to perform graded monitoring of the degree of battery electrode defects, thereby completing the defect detection of the battery electrode.

[0044] The detailed steps are the same as those provided in Example 1, which is a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks, and will not be repeated here.

[0045] Example 3 Embodiment 3 of the present invention provides a computer-readable storage medium.

[0046] A computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of a method for detecting minute defects in battery manufacturing based on an ultrasonic graph network as described in Embodiment 1 of the present invention.

[0047] The detailed steps are the same as those provided in Example 1, which is a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks, and will not be repeated here.

[0048] Example 4 Embodiment 4 of the present invention provides an electronic device.

[0049] An electronic device includes a memory, a processor, and a program stored in the memory and running on the processor. When the processor executes the program, it implements the steps in a method for detecting minute defects in battery manufacturing based on an ultrasonic graph network as described in Embodiment 1 of the present invention.

[0050] The detailed steps are the same as those provided in Example 1, which is a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks, and will not be repeated here.

[0051] Example 5 Embodiment 5 of the present invention provides a computer program product.

[0052] A computer program product includes software code, wherein the program in the software code performs the steps of a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks as described in Embodiment 1 of the present invention.

[0053] The detailed steps are the same as those provided in Example 1, which is a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks, and will not be repeated here.

[0054] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0055] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0056] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0057] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0058] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0059] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0060] The above description is merely a preferred embodiment of this practice and is not intended to limit the scope of this practice. Various modifications and variations can be made to this practice by those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of this practice should be included within the protection scope of this practice.

Claims

1. A method for detecting minute defects in battery manufacturing based on ultrasonic graph networks, characterized in that, include: Obtain the structural parameters of the battery electrodes; Multi-frequency ultrasonic excitation was applied to the battery electrodes to obtain echo signals from different spatial positions of the electrodes; The characteristic parameters of the acquired echo signal are analyzed and extracted, and several analysis units and their characteristic parameter sets are constructed according to the spatial position of the electrodes. Based on the feature parameter set of several constructed analysis units, calculate the physical intermediate variables of each analysis unit and the propagation correlation parameters between different analysis units, and construct the fused feature vector and propagation correlation matrix; Based on the fused feature vectors, propagation correlation matrix, and graph neural network based on acoustic propagation mechanism, the obtained battery electrode structural parameters and battery electrode defect types are predicted. Based on the predicted results of the battery electrode structural parameters, a structural dispersion index is constructed. Combined with the predicted results of battery electrode defect types, the degree of battery electrode defects is graded and monitored to complete the defect detection of the battery electrode.

2. The method for detecting minute defects in battery manufacturing based on ultrasonic graph networks as described in claim 1, characterized in that, Echo signals from different spatial locations of the acquired electrodes are subjected to time-domain, frequency-domain, and dispersion joint analysis to extract characteristic parameters representing local propagation characteristics. Analysis units and their corresponding characteristic parameter sets are constructed according to spatial location. In the process of constructing the characteristic parameter set, a two-dimensional coordinate system is established on the horizontal plane according to the physical size of the electrode. The electrode surface is divided into a regular grid array according to the preset spatial resolution. The electrode entity region corresponding to each grid is an independent analysis unit, and the ultrasonic characteristic parameter set corresponding to each unit is obtained, thus completing the construction of the characteristic parameter set.

3. The method for detecting minute defects in battery manufacturing based on ultrasonic graph networks as described in claim 1, characterized in that, Based on the obtained prediction results of battery electrode structure parameters, the comprehensive deviation of each analysis unit relative to the mean of the whole region is calculated, and the structural dispersion index is constructed according to the comprehensive deviation of each analysis unit. The first and second judgment thresholds are determined based on the overall deviation of the standard. The battery defect level is classified by combining the overall deviation and the determined first and second judgment thresholds.

4. The method for detecting minute defects in battery manufacturing based on ultrasonic graph networks as described in claim 3, characterized in that, When the overall deviation does not exceed the first judgment threshold, the battery electrode at that location has an intact physical structure and no obvious defects. When the overall deviation is between the first and second judgment thresholds, the battery electrode at that location has a slight parameter deviation or defect risk, which is defined as a sub-healthy state and requires enhanced quality attention. When the overall deviation exceeds the second judgment threshold, the battery electrode at that location has a serious physical defect that is the average value of the standard qualified electrode samples, triggering a scrap warning and marking it as a defective area.

5. The method for detecting minute defects in battery manufacturing based on ultrasonic graph networks as described in claim 1, characterized in that, The loss function of the graph neural network based on the acoustic propagation mechanism for ;in, These are the weighting coefficients; and These are the comprehensive task loss function and the physical constraint loss function, respectively. ; ; , and These represent the actual values ​​for electrode thickness, porosity, and defect type, respectively; these are the labels. , They are respectively and , and The functional relationship between them is obtained by fitting the training data.

6. The method for detecting minute defects in battery manufacturing based on ultrasonic graph networks as described in claim 1, characterized in that, The structural parameters of the battery electrode obtained include electrode thickness and porosity.

7. A battery manufacturing micro-defect detection system based on ultrasonic graph networks, characterized in that, include: The first acquisition module is configured to acquire the structural parameters of the battery electrodes; The second acquisition module is configured to apply multi-frequency ultrasonic excitation to the battery electrode to acquire echo signals at different spatial positions of the electrode. The first construction module is configured to analyze and extract the characteristic parameters of the acquired echo signal, and construct several analysis units and their characteristic parameter sets according to the spatial position of the electrodes; The second construction module is configured to calculate the physical intermediate variables of each analysis unit and the propagation correlation parameters between different analysis units based on the feature parameter set of the constructed analysis units, and construct the fused feature vector and the propagation correlation matrix. The electrode prediction module is configured to predict the acquired battery electrode structural parameters and battery electrode defect types based on the fused feature vector, the propagation correlation matrix, and the graph neural network based on the acoustic propagation mechanism. The defect detection module is configured to construct a structural dispersion index based on the obtained battery electrode structural parameter prediction results, and combine the battery electrode defect type prediction results to perform graded monitoring of the degree of battery electrode defects, thereby completing the defect detection of the battery electrode.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks as described in any one of claims 1-6.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the program, it implements the steps of a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks as described in any one of claims 1-6.

10. A computer program product, comprising software code, characterized in that, The program in the software code executes the steps of a method for detecting minute defects in battery manufacturing based on ultrasonic graph networks as described in any one of claims 1-6.

Images