A battery production whole-process data monitoring system and method

Abstract

The present application relates to the technical field of data monitoring, and particularly relates to a battery production whole-process data monitoring system and method.The method comprises the following steps: acquiring real-time multi-source data of a battery production line and assembly station images; analyzing current mutation amplitude and regional temperature difference standard deviation of the real-time multi-source data of the battery production line; identifying production stage boundary points of the assembly station images, and dividing battery production processes by using the current mutation amplitude and the regional temperature difference standard deviation, to generate battery production process data; extracting an electrode slurry preparation stage of the battery production process data, and deploying a quantum dot sensor array on a slurry agitator based on the electrode slurry preparation stage, to obtain sensor array deployment data.The present application improves the monitoring accuracy, abnormality recognition capability and intelligent collaborative response level of the whole battery production process by using multi-source real-time data fusion, quantum dot sensor array spectrum decoding, three-dimensional particle state reconstruction and digital twin technology.

Description

Technical Field

[0001] This invention relates to the field of data monitoring technology, and in particular to a data monitoring system and method for the entire battery production process. Background Technology

[0002] Early battery production relied primarily on manual recording and offline analysis, with limited data acquisition methods and poor timeliness, making it difficult to reflect production status promptly. Subsequently, automated production lines introduced sensor monitoring and PLC control, achieving real-time data acquisition at some key nodes, but still lacking global perception and dynamic response capabilities for the entire production process. In recent years, with the rise of technologies such as the Internet of Things, edge computing, and big data analytics, enterprises have gradually built digital monitoring systems covering multiple stages, including electrode preparation, slurry mixing, coating, rolling, slitting, assembly, liquid injection, and packaging. Through multi-source data fusion, image recognition, and AI anomaly detection, a shift from single-point monitoring to full-process collaboration has been gradually achieved. However, as electrode slurry preparation is a core stage, existing methods struggle to capture its microscopic particle evolution and aggregation behavior, and most rely on offline sampling inspection, resulting in significant lag. Furthermore, while traditional systems can monitor local data, they lack spatial mapping of the overall production process and anomaly coordination mechanisms, making it impossible to quickly respond to the impact of slurry anomalies on subsequent processes. Summary of the Invention

[0003] Therefore, it is necessary to provide a data monitoring system and method for the entire battery production process to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for monitoring data throughout the entire battery production process is provided, the method comprising the following steps:

[0005] Step S1: Acquire real-time multi-source data of the battery production line and images of the assembly station; analyze the current fluctuation amplitude and regional temperature difference standard deviation of the real-time multi-source data of the battery production line; identify the production stage boundary points of the assembly station images, and divide the battery production process using the current fluctuation amplitude and regional temperature difference standard deviation to generate battery production process data.

[0006] Step S2: Extract the electrode slurry preparation stage from the battery production process data, and deploy a quantum dot sensor array in the slurry stirring tank based on the electrode slurry preparation stage to obtain sensor array deployment data; decode the spectral response spectrum using the sensor array deployment data to generate slurry state spectral characteristic data;

[0007] Step S3: The microscopic particle state evolution of the electrode slurry during the preparation stage is analyzed using slurry state spectral characteristic data to generate slurry agglomeration behavior data; the slurry agglomeration behavior data is reconstructed in three dimensions, and the microscopic imbalance of slurry stirring is monitored based on the reconstruction results to generate slurry anomaly monitoring data;

[0008] Step S4: Construct a digital twin scenario based on battery production process data to generate three-dimensional monitoring data for the battery production process; coordinate battery production data with the three-dimensional monitoring data of the battery production process based on slurry anomaly monitoring data to execute full-process data monitoring of battery production.

[0009] This invention collects real-time multi-source data (such as current and temperature) from the battery production line and combines it with images from the assembly station. It uses the magnitude of current fluctuations and the standard deviation of regional temperature differences to divide the production stages, accurately identifying the boundaries of each process stage in battery production. This avoids the problems of traditional methods that rely on manual judgment or single parameters for production node division, improving the accuracy and automation level of process identification. A quantum dot sensor array is deployed during the electrode slurry preparation stage, and high-dimensional spectral characteristic data of the slurry state are obtained through spectral response spectrum decoding. This effectively enhances the monitoring capability of the slurry's microscopic physical and chemical state, overcoming the limitations of traditional sensors in terms of sensitivity and spectral resolution. Through a microscopic particle state evolution model driven by spectral characteristic data, the changing trend of slurry agglomeration behavior can be dynamically tracked. Spatial modeling of the slurry stirring state is achieved through three-dimensional distribution reconstruction, effectively identifying microscopic imbalances caused by uneven stirring efficiency or local stagnation during the stirring process, enabling early warning of slurry anomalies. By constructing a digital twin scenario based on battery production process data and coordinating it with slurry anomaly monitoring data, the system can visualize the status of each stage of production in real time in three-dimensional space. This enables real-time monitoring, intelligent decision-making, and closed-loop control of the entire battery production process, significantly improving the controllability and flexibility of production. Through a real-time monitoring and feedback control mechanism for micro-imbalances, anomalies can be detected early in electrode slurry preparation, allowing for timely intervention and effectively preventing electrode defects caused by slurry agglomeration or uneven mixing. This ensures product quality from the source, improves yield, and reduces rework rates and material waste. The method adopts a modular architecture design, which can flexibly adapt to different types of battery production lines and is easy to integrate with existing MES, SCADA, or quality traceability systems, demonstrating good engineering feasibility and industrial application prospects. Therefore, this invention improves the monitoring accuracy, anomaly identification capability, and intelligent collaborative response level of the entire battery production process by utilizing multi-source real-time data fusion, quantum dot sensor array spectral decoding, three-dimensional particle state reconstruction, and digital twin technology.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Acquire real-time multi-source data of the battery production line and images of the assembly station;

[0012] Step S12: Calculate the current fluctuation amplitude and regional temperature difference standard deviation based on real-time multi-source data from the battery production line to generate current-temperature difference joint characteristic data;

[0013] Step S13: Enhance the edge contour and extract the structural region of the assembly station image to generate structural feature data of the station image; identify the assembly action sequence of typical parts based on the structural feature data of the station image to generate boundary point data of the production stage;

[0014] Step S14: Use the current-temperature difference joint feature data to segment the production stage boundary point data into process behavior time windows to generate process feature slice data; perform key indicator clustering and dynamic stage alignment processing on the process feature slice data to generate battery production process data.

[0015] This invention achieves deep coupling analysis of process and image data by fusing real-time multi-source process data (current, temperature) with assembly station images, avoiding the uncertainty of traditional process identification relying on a single data source and improving the comprehensiveness and robustness of identification. In step S12, the constructed "current-temperature difference joint feature data" integrates two important dimensions of process change: current mutations reflect energy consumption fluctuations, and temperature difference standard deviations reveal thermal field anomalies. The combination of the two provides stronger process stage discrimination capabilities, helping to accurately capture production behavior change points and improve sensitivity to process nodes. Through image structure enhancement and recognition of typical assembly actions (such as welding, pressing, and inspection), production stage boundary point data is generated, effectively improving the temporal accuracy and spatial matching degree of production stage division and avoiding stage offset or misjudgment caused by relying solely on sensor data. The process behavior time window segmentation and key indicator clustering + dynamic stage alignment processing proposed in step S14 form stable and generalizable process feature slices, effectively solving problems such as inconsistent process duration and cycle time offset, and improving the time alignment and stage consistency of battery production process modeling.

[0016] Preferably, step S2, which involves deploying a quantum dot sensor array in the slurry stirring vessel based on the electrode slurry preparation stage, includes:

[0017] A quantum dot sensor array was deployed in the slurry stirring vessel during the electrode slurry preparation stage, and sensor array deployment data was obtained. The slurry temperature was set to be between 15 and 80°C, the viscosity range to be 500 to 5000 mPa·s, the particle size distribution to be 50 to 800 nm, the fluorescence response wavelength to be concentrated between 500 and 650 nm, and the fluorescence intensity sensitivity to be 10. 2 Up to 10 5 For relative units, the pH value of the slurry is set between 6.0 and 9.5, the conductivity range is 0.1 to 5.0 S / m, the vibration acceleration is set between 0.5 and 3.0 g, the sampling frequency of the sensor array is set between 10 and 100 Hz, and the array distribution density is 4 to 16 sensing points per square meter.

[0018] This invention achieves comprehensive sensing of the physicochemical state of slurry during mixing by setting monitoring ranges for key process parameters including temperature, viscosity, particle size distribution, pH value, conductivity, and vibration acceleration. This real-time sensing capability of multi-dimensional parameters far surpasses traditional single-physical-quantity monitoring methods, providing a data foundation for slurry quality control. The deployed quantum dot sensor array has fluorescence response wavelengths concentrated in the 500-650nm range and fluorescence intensity sensitivity reaching 10. 2 Up to 10 5 Compared to other sensors, this method can accurately respond to changes in the state of microscopic particles in slurry, such as aggregation, dispersion, or interfacial reactions, greatly enhancing the dynamic monitoring capability of microstructure evolution and effectively overcoming the limitations of traditional sensors in nanoscale sensing. By setting the sampling frequency of the sensor array to 10 to 100 Hz and the spatial distribution density to 4 to 16 sensing points per square meter, uniform spatial coverage and high-frequency dynamic data acquisition can be achieved. This ensures continuous tracking of physical disturbances, particle flow, and thermal changes throughout the entire slurry mixing cycle, improving the completeness and real-time performance of data acquisition. Setting reasonable slurry operating parameter ranges (such as temperature 1580℃, viscosity 500-5000 mPa·s, pH 6.0-9.5, conductivity 0.1-5.0 S / m, and particle size 50-800 nm) ensures that the sensor deployment environment is within the stable response range of quantum dot fluorescence, avoiding signal drift, fluorescence quenching, or data distortion due to extreme operating conditions, and guaranteeing the long-term stable operation of the sensor. The real-time output of high-frequency, high-density sensor array data and fluorescence spectral response provides precise input for subsequent slurry state visualization modeling, mixing uniformity analysis, and agglomeration behavior recognition. It also supports slurry preparation control algorithms driven by digital twins, enabling closed-loop control capabilities from "sensing → analysis → control".

[0019] Preferably, step S2, which involves deploying data via a sensor array to decode the spectral response spectrum, includes:

[0020] Extracting multi-channel response signals from sensor array deployment data;

[0021] Time-domain synchronization and spectrum normalization are performed on the raw response data of multiple channels to generate standard response spectrum data;

[0022] Perform spectral difference analysis and high-dimensional noise reduction on standard response spectrum data to generate effective spectral response interval data;

[0023] Based on the effective spectral response range data, component-sensitive spectral bands are decoded to generate a slurry component reflectance coefficient matrix;

[0024] Principal component features of the reflectance coefficient matrix of the slurry components are extracted, and chemical absorption peaks are calibrated for the multi-channel response signals based on the principal component features to generate spectral feature data of the slurry state.

[0025] This invention effectively eliminates signal distortion caused by sampling delay and sensor sensitivity drift between channels by performing time-domain synchronization and spectral normalization on the multi-channel raw response signals acquired by a quantum dot sensor array. This ensures the time-frequency consistency and feature comparability of the data used for decoding analysis, improving the accuracy and repeatability of the decoding results. By introducing spectral band difference analysis and high-dimensional noise reduction, effective bands with actual component response significance can be extracted from redundant and complex multi-channel spectra, reducing noise interference and information redundancy, improving the computational efficiency and stability of the decoding model in subsequent steps, and effectively ensuring the complete preservation of key information in spectral processing. By analyzing the effective spectral response range, component-sensitive spectral bands significantly related to key slurry components (such as binders, conductive agents, and active substances) are identified, and a reflection coefficient matrix is ​​generated through decoding. This enables the differentiation of different physical components in the multi-channel spectral space, providing physical support and spectroscopic basis for the dynamic identification of slurry components. By extracting principal component features from the reflection coefficient matrix, the main change patterns in the evolution of slurry state are captured. Combined with the absorption peak position, the chemical absorption peaks of the multi-channel response signals are calibrated. This allows for a precise characterization of the real-time changes in the micro-particle state, distribution uniformity, and chemical reactions of the slurry during stirring or dispersion. This significantly enhances the spectral perception capability of the slurry's physical-chemical state evolution. The generated slurry state spectral feature data, as a high-dimensional, multi-scale state characterization result, not only possesses good representativeness and discriminability but can also serve as input data for key aspects such as digital twin modeling, anomaly identification, and aggregation behavior prediction, forming an efficient bridge from "physical monitoring" to "data cognition."

[0026] Preferably, step S3, which involves analyzing the microscopic particle state evolution during the electrode slurry preparation stage using slurry state spectral characteristic data, includes:

[0027] Band reflectance fitting is performed on the spectral characteristic data of slurry state to generate multi-scale band reflectance curve data;

[0028] Particle size response mapping is performed on multi-scale band reflection curve data to generate particle size response distribution data;

[0029] Time window slicing is performed on particle size response distribution data to generate time-series particle size evolution fragment data;

[0030] Identify particle cluster aggregation trends in time-series particle size evolution data segments;

[0031] Analyze particle spacing data from particle aggregation trend data to generate particle spacing distribution data;

[0032] The particle spacing distribution data is compared with a preset particle spacing threshold. When the particle spacing distribution data is greater than or equal to the preset particle spacing threshold, the particle spacing distribution data is marked as clustered data.

[0033] Spatial density inversion is performed on the aggregation data to generate micro-aggregate concentration layer data;

[0034] The micro-agglomeration concentration layer data is spatiotemporally fused, and the stability discrete analysis of the electrode slurry preparation stage is performed based on the fused micro-agglomeration concentration layer data, ultimately generating slurry agglomeration behavior data.

[0035] This invention, by fitting band reflectance and mapping particle size response to spectral characteristic data of slurry state, can indirectly deduce particle size distribution changes without the need for complex physical particle imaging or particle size analyzer intervention. This significantly improves the real-time analysis capability and sensing efficiency of slurry micro-behavior, providing a new approach for non-invasive quality monitoring. By slicing the particle size response mapping results into time windows to construct time-series particle size evolution fragment data, it not only captures the dynamic trend of particle aggregation and dispersion during slurry mixing but also provides an evolutionary path and periodic analysis basis for subsequent particle agglomeration pattern recognition and aggregation intensity assessment. Based on particle spacing distribution data and compared with a preset agglomeration threshold, a scientific and objective agglomeration discrimination logic is constructed, effectively solving the problem that traditional particle size statistics cannot accurately describe the degree of "agglomeration" phenomenon, making slurry agglomeration identification more quantitative, regular, and engineered. By performing spatial density inversion on the agglomeration data, the generated micro-agglomeration concentration layer data reflects the spatial distribution density and intensity of agglomeration regions. Further spatiotemporal fusion and stability discretization analysis allows for the construction of a high-resolution three-dimensional agglomeration evolution map, significantly enhancing the visualization and traceability of the slurry preparation process. Based on the spatiotemporal analysis of agglomeration behavior, the resulting slurry agglomeration behavior data can be used to reflect the overall stability, discretization, and micro-mixing uniformity of the slurry system. This overcomes the bottleneck of relying solely on macroscopic viscosity or visual judgment to identify micro-agglomeration problems, greatly enhancing the scientific rigor and sensitivity of micro-state quality control.

[0036] Preferably, step S3, which involves reconstructing the three-dimensional distribution of the slurry agglomeration behavior data, includes:

[0037] Extract the spatial coordinate index of the slurry agglomeration behavior data to obtain the agglomeration spatial coordinate point set data;

[0038] Point cloud densification interpolation is performed on the clustered spatial coordinate point set data to generate particle clustered point cloud data;

[0039] Local voxelization encoding of particle agglomeration point cloud data, and global mesh splicing reconstruction of the encoded particle agglomeration point cloud data to generate slurry agglomeration 3D mesh skeleton data;

[0040] Particle size channel attribute mapping is performed on the 3D mesh skeleton data of slurry agglomeration to generate attribute-annotated mesh data;

[0041] Multi-view projection rendering is performed on attribute-annotated mesh data to generate 3D view data of particle aggregation;

[0042] Dynamic sequence frame encoding is performed based on particle aggregation 3D view data to generate adjustable-speed particle aggregation evolution sequence data.

[0043] Stability hot zone annotation is performed on particle aggregation evolution sequence data to obtain the reconstruction result of three-dimensional distribution reconstruction.

[0044] This invention reconstructs the dense distribution characteristics of agglomerated particles in the stirring space by extracting the spatial coordinate index of slurry agglomeration behavior data and performing point cloud densification interpolation. This significantly improves the spatial reconstruction accuracy and local aggregation recognition capability of micro-agglomeration behavior, laying the foundation for high-fidelity visualization in subsequent digital twin scenarios. Based on the 3D mesh skeleton data generated by local voxelization encoding and global mesh stitching reconstruction, a continuous and connected structural representation of slurry agglomeration morphology can be achieved. Further, the introduction of particle size channel attribute mapping and attribute annotation mesh construction endows the 3D mesh with particle-level semantic attributes, realizing a leap from "geometric reconstruction" to "physical semantic annotation," enhancing the depth of understanding of particle agglomeration behavior. Multi-view projection rendering of the attribute annotation mesh data generates realistic 3D view data of particle agglomeration, allowing the complex agglomeration evolution process to be visually presented intuitively. This facilitates morphological observation, local tracking, and comparative analysis for operators and system engineers, effectively improving the interpretability of data-driven judgment. By encoding 3D view data into adjustable-speed dynamic sequence frames, particle aggregation evolution sequence data can be generated. This allows for the dynamic presentation of the occurrence, development, and decay of particle aggregation behavior at different stirring stages, showcasing the evolutionary trajectory and mechanism trends of aggregation from non-existence to formation, from weak to strong, and then gradually dispersing. Based on the evolution sequence data, stability hotspot annotation can automatically identify unstable hotspots, persistent aggregation hotspots, or high cluster concentration areas during aggregation, and output visualized annotation information with risk warning functions. This provides key technical support for real-time control, parameter adjustment, and stirring strategy intervention.

[0045] Preferably, step S3, which involves monitoring the micro-imbalance of slurry mixing based on the reconstruction results, includes:

[0046] Based on the reconstruction results of the three-dimensional distribution reconstruction, the particle structure orientation information within the stirring area is extracted to generate local structure orientation data;

[0047] Spatial anisotropy tensor analysis is performed on local structural orientation data to generate orientation tensor deviation data;

[0048] Using orientation tensor deviation data, vector trajectory deduction of particle flow trend in electrode slurry preparation stage is performed to generate microscopic flow direction perturbation data.

[0049] The local shear stress gradient during the stirring process is calculated based on the microscopic flow disturbance data, and a shear imbalance intensity map is generated.

[0050] The stability domain boundary of the shear imbalance intensity map is identified, and particle aggregation time window statistics are performed on the stability domain boundary to generate particle swarm fluctuation interval data.

[0051] Microscopic disturbance frequency analysis was performed on the particle size fluctuation range data to generate slurry anomaly monitoring data.

[0052] This invention extracts particle structure orientation information from 3D reconstruction results and generates local structure orientation data, accurately depicting the spatial organization of particle arrangement and aggregation in the mixing zone. This solves the problem of insufficient perception of orientation changes and spatial inhomogeneities in traditional slurry condition monitoring, providing a precise basis for subsequent anisotropic modeling. Spatial anisotropic tensor analysis of the structure orientation data and extraction of orientation tensor deviation data quantitatively assesses the inconsistency of particle arrangement in different directions and local orientation disorder, forming a highly sensitive index system for microstructural inhomogeneity within the mixing zone, significantly improving anomaly detection resolution. Based on tensor deviation, vector trajectory deduction of particle flow trends can reconstruct the disturbance direction and bending trend of particle motion in the slurry, thereby obtaining microscopic flow disturbance data and revealing particle behavior deviations caused by asymmetric mixing, particle aggregation, or shear differences in local areas. The local shear stress gradient calculated based on the microscopic disturbance path forms a shear imbalance intensity map, clearly revealing the differences in shear action experienced by different mixing zones, accurately identifying typical problem areas such as "shear dead angles" and "mixing deviations" during mixing, providing a reference for mixing parameter calibration and equipment structure optimization. By identifying the stability domain boundaries in the shear spectrum and performing particle aggregation time window statistics, the generated particle swarm fluctuation range data can further reveal the behavioral patterns and concentrated occurrence times of agglomerated particles during the stirring cycle, supporting agglomeration early warning, disturbance prediction, and process optimization scheduling. Finally, micro-disturbance frequency analysis of the particle swarm fluctuation range data can effectively identify high-frequency oscillation regions and periodic abnormal aggregation phenomena, accurately generate slurry anomaly monitoring data, and achieve dynamic real-time feedback on the stability of electrode slurry preparation, exhibiting good process control foresight and responsiveness.

[0053] Preferably, the micro-perturbation frequency analysis of particle swarm fluctuation range data includes:

[0054] Micro-disturbance frequency analysis is performed on the particle size fluctuation range data. When any of the following conditions occur, it is determined to be an abnormal particle size concentration, and abnormal particle size concentration data is obtained: the characteristic particle size distribution range narrows by more than 20%, the D90 / D10 particle size ratio fluctuates by more than ±0.6, and the particle size distribution skewness index deviates from the average value by more than ±15% within three consecutive sampling periods.

[0055] When the following conditions occur simultaneously, it is determined to be a phenomenon of enhanced micro-agglomeration disturbance, and data on enhanced agglomeration disturbance are obtained: the ratio of agglomerates within the particle group increases by more than 25% within 10 minutes, the average agglomerate size exceeds the historical limit by more than 10%, and the density distribution inside the agglomerates shows a bimodal shift that lasts for more than 20 minutes.

[0056] When the following conditions are met simultaneously, it is determined to be an abnormal shear fluctuation state, and abnormal shear fluctuation data is obtained: the frequency of the stirring system rotation speed fluctuation exceeds 2Hz, the frequency of the particle group micro-perturbation response deviates from the reference spectrum range by more than ±10% during the stirring process, accompanied by an increase of more than 8% in the content of microbubbles in the slurry, and the duration of this state exceeds 30 minutes.

[0057] By integrating abnormal particle size concentration data, enhanced agglomeration disturbance data, and abnormal shear fluctuation data, and performing correlation weight evaluation and disturbance type fusion identification, slurry anomaly monitoring data is finally generated.

[0058] This invention achieves automatic anomaly identification of multi-scale particle group behavior based on data-driven principles by setting multiple specific and quantifiable judgment conditions (such as particle size distribution narrowing, D90 / D10 ratio, particle size skewness, agglomerate quantity and density characteristics, stirring frequency, and microbubble amplification). This strategy effectively overcomes the shortcomings of traditional anomaly identification, which relies on a single parameter and has a high false positive rate. By classifying abnormal behaviors in slurry into three categories: abnormal particle size concentration (reflecting an abnormally concentrated particle size distribution), enhanced agglomeration disturbance (reflecting a sharp increase in agglomeration trend), and abnormal shear fluctuation (reflecting an imbalance between the shear system and microscopic response), this classification system helps to clarify the source and behavioral patterns of different anomalies in slurry preparation, providing categorized strategy support for precise intervention and system response. By evaluating the correlation weights of the three types of abnormal data and fusing and identifying disturbance types, it not only achieves collaborative analysis and dynamic superposition judgment of different types of disturbances, but also integrates multiple indicator weights to form a global disturbance intensity judgment result, enhancing the monitoring system's comprehensive judgment capability for complex disturbance events. The method's coupled analysis of agglomeration behavior (such as particle size, density, and quantity) and agitation behavior (such as frequency fluctuations and bubble content) enables early warning when agglomeration is still in the "enhanced" rather than "severely out of control" stage. This allows for an early identification, adjustment, and intervention process control mechanism, effectively avoiding subsequent quality risks such as agglomerate deposition and uneven coating. By embedding time conditions such as "three consecutive sampling cycles," "increase within 10 minutes," and "lasting for more than 30 minutes" into anomaly detection, the method ensures that it is applicable not only to sudden anomalies but also to identifying slow-type and gradual process imbalances, forming a more complete time-series diagnostic capability for the dynamic behavior of the slurry system.

[0059] Preferably, step S4 includes the following steps:

[0060] Step S41: Extract the parameters of each assembly station, the flow of key materials and the timeline of the process from the battery production process data, perform spatial mapping, and generate structured digital scene data of the battery production process.

[0061] Step S42: Perform 3D visualization modeling on the structured digital scene data to generate 3D monitoring data for the battery production process; based on the slurry anomaly monitoring data, precisely locate the corresponding electrode slurry preparation stage in the 3D monitoring data to generate anomaly mapping segment data;

[0062] Step S43: Perform parameter cross-comparison between the abnormal mapping section data and the 3D monitoring data to generate visual abnormality marker layer data; overlay the visual abnormality marker layer data onto the 3D monitoring data of the battery production process to generate fused monitoring data with abnormality feedback information;

[0063] Step S44: Perform station-level data flow tracing on the fused monitoring data, identify the potential transmission path of slurry anomalies to downstream process nodes, and generate a process-level anomaly transmission map; use the process-level anomaly transmission map to perform data collaborative updates of the three-dimensional monitoring data of the battery production process, and finally realize the data monitoring operation of the entire battery production process.

[0064] This invention extracts assembly station parameters, material flow direction, and timeline information, performs spatial mapping and structural modeling, and generates structured digital scene data. This forms a virtual-real mapping digital foundation for the actual battery production process, providing a basic platform for multi-source data fusion, time-series process presentation, and state synchronization analysis. By using slurry anomaly monitoring data to map and precisely locate anomalies in corresponding time periods within the 3D monitoring scene, local imbalances originally hidden in microscopic data can be accurately projected into a visualized spatial model, enabling real-time dynamic feedback on the electrode slurry preparation stage. Through parameter cross-comparison and layer marking, abnormal events are visualized and overlaid on the 3D monitoring data using color coding, area highlighting, or icon prompts, constructing a visual feedback interface for anomalies. This enhances operators' intuitive perception of anomaly trends and improves their rapid response efficiency. By tracing the data flow at the station level from the fused monitoring data, the path of slurry anomalies from the preparation source to downstream key nodes such as coating, compaction, and stacking can be identified, forming a process-level anomaly transmission map. This expands anomaly monitoring beyond single-point identification to include full-process risk prediction and chain response. By utilizing anomaly propagation maps to perform local updates and full-process correlation adjustments of 3D monitoring data, a system with closed-loop logic of "anomaly identification → path location → impact feedback → data update" is constructed. This system has good self-evolution and dynamic adjustment capabilities and can adapt to changing working conditions in complex production environments.

[0065] This specification provides a battery production process data monitoring system for executing the aforementioned battery production process data monitoring method. The battery production process data monitoring system includes:

[0066] The process division module is used to acquire real-time multi-source data of the battery production line and images of the assembly station; analyze the current fluctuation amplitude and regional temperature difference standard deviation of the real-time multi-source data of the battery production line; identify the production stage boundary points of the assembly station images; and divide the battery production process using the current fluctuation amplitude and regional temperature difference standard deviation to generate battery production process data.

[0067] The slurry analysis module is used to extract data from the electrode slurry preparation stage of the battery production process, and to deploy a quantum dot sensor array in the slurry stirring vessel based on the electrode slurry preparation stage to obtain sensor array deployment data; the spectral response spectrum is decoded through the sensor array deployment data to generate slurry state spectral characteristic data;

[0068] The imbalance monitoring module is used to perform microscopic particle state evolution analysis on electrode slurry preparation stage using slurry state spectral characteristic data, generating slurry agglomeration behavior data; it performs three-dimensional distribution reconstruction on slurry agglomeration behavior data, and monitors microscopic imbalance of slurry stirring based on the reconstruction results, generating slurry anomaly monitoring data.

[0069] The data collaboration module is used to construct digital twin scenarios based on battery production process data and generate three-dimensional monitoring data for the battery production process; it also collaborates with the three-dimensional monitoring data of the battery production process based on slurry anomaly monitoring data to perform data monitoring operations for the entire battery production process.

[0070] The beneficial effects of this invention lie in the fact that, through the process segmentation module, multi-source real-time data such as the amplitude of current fluctuations and the standard deviation of regional temperature differences are linked with stage boundary points extracted from the assembly station image for analysis. This enables automatic segmentation of process time windows and precise process labeling for each stage of battery production, significantly improving segmentation accuracy and dynamic adaptability compared to traditional stage segmentation methods that rely on human experience. The slurry analysis module, by deploying a quantum dot sensor array and combining it with slurry physicochemical properties (temperature, pH, conductivity, particle size distribution, fluorescence response, etc.), and performing spectral decoding and principal component feature extraction on multi-channel response signals, can monitor slurry status with high sensitivity and high timeliness, solving the pain points of traditional slurry monitoring such as "slow response, limited information, and low accuracy." The imbalance monitoring module maps slurry state spectral data to the particle size evolution process and combines it with the three-dimensional distribution reconstruction of agglomeration behavior. Finally, through orientation tensor analysis, shear imbalance calculation, and perturbation frequency identification, it constructs a complete chain from microscale to macroscopic manifestation to imbalance identification. This effectively provides early warning of key risks such as agglomeration and abnormal stirring, improving slurry consistency control. Through the data collaboration module, the slurry anomaly monitoring results are dynamically linked with a three-dimensional visualized digital twin scene. A visualized anomaly layer is generated in the three-dimensional monitoring data, and process-level tracking and feedback updates are performed on anomaly sections, forming a closed-loop production monitoring system with anomaly feedback, autonomous updating, and dynamic collaboration capabilities. This significantly enhances the transparency and risk intervention capabilities of the entire process. The four modules in this solution work together from four levels: "process identification—state analysis—risk monitoring—data collaboration," connecting the entire battery production process from front-end identification, process analysis, risk identification to back-end monitoring and feedback. This has extremely high engineering practical value and effectively promotes the evolution of the battery manufacturing system towards intelligence, adaptability, and traceability. Therefore, this invention improves the monitoring accuracy, anomaly identification capability, and intelligent collaborative response level of the entire battery production process by utilizing multi-source real-time data fusion, quantum dot sensor array spectral decoding, three-dimensional particle state reconstruction, and digital twin technology. Attached Figure Description

[0071] Figure 1A flowchart illustrating the steps of a method for monitoring data throughout the entire battery production process;

[0072] Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S1.

[0073] Figure 3 for Figure 1 A detailed flowchart illustrating the implementation steps of step S4.

[0074] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0075] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0076] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0077] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0078] To achieve the above objectives, please refer to Figures 1 to 3 A method for monitoring data throughout the entire battery production process, the method comprising the following steps:

[0079] Step S1: Acquire real-time multi-source data of the battery production line and images of the assembly station; analyze the current fluctuation amplitude and regional temperature difference standard deviation of the real-time multi-source data of the battery production line; identify the production stage boundary points of the assembly station images, and divide the battery production process using the current fluctuation amplitude and regional temperature difference standard deviation to generate battery production process data.

[0080] Step S2: Extract the electrode slurry preparation stage from the battery production process data, and deploy a quantum dot sensor array in the slurry stirring tank based on the electrode slurry preparation stage to obtain sensor array deployment data; decode the spectral response spectrum using the sensor array deployment data to generate slurry state spectral characteristic data;

[0081] Step S3: The microscopic particle state evolution of the electrode slurry during the preparation stage is analyzed using slurry state spectral characteristic data to generate slurry agglomeration behavior data; the slurry agglomeration behavior data is reconstructed in three dimensions, and the microscopic imbalance of slurry stirring is monitored based on the reconstruction results to generate slurry anomaly monitoring data;

[0082] Step S4: Construct a digital twin scenario based on battery production process data to generate three-dimensional monitoring data for the battery production process; coordinate battery production data with the three-dimensional monitoring data of the battery production process based on slurry anomaly monitoring data to execute full-process data monitoring of battery production.

[0083] This invention collects real-time multi-source data (such as current and temperature) from the battery production line and combines it with images from the assembly station. It uses the magnitude of current fluctuations and the standard deviation of regional temperature differences to divide the production stages, accurately identifying the boundaries of each process stage in battery production. This avoids the problems of traditional methods that rely on manual judgment or single parameters for production node division, improving the accuracy and automation level of process identification. A quantum dot sensor array is deployed during the electrode slurry preparation stage, and high-dimensional spectral characteristic data of the slurry state are obtained through spectral response spectrum decoding. This effectively enhances the monitoring capability of the slurry's microscopic physical and chemical state, overcoming the limitations of traditional sensors in terms of sensitivity and spectral resolution. Through a microscopic particle state evolution model driven by spectral characteristic data, the changing trend of slurry agglomeration behavior can be dynamically tracked. Spatial modeling of the slurry stirring state is achieved through three-dimensional distribution reconstruction, effectively identifying microscopic imbalances caused by uneven stirring efficiency or local stagnation during the stirring process, enabling early warning of slurry anomalies. By constructing a digital twin scenario based on battery production process data and coordinating it with slurry anomaly monitoring data, the system can visualize the status of each stage of production in real time in three-dimensional space. This enables real-time monitoring, intelligent decision-making, and closed-loop control of the entire battery production process, significantly improving the controllability and flexibility of production. Through a real-time monitoring and feedback control mechanism for micro-imbalances, anomalies can be detected early in electrode slurry preparation, allowing for timely intervention and effectively preventing electrode defects caused by slurry agglomeration or uneven mixing. This ensures product quality from the source, improves yield, and reduces rework rates and material waste. The method adopts a modular architecture design, which can flexibly adapt to different types of battery production lines and is easy to integrate with existing MES, SCADA, or quality traceability systems, demonstrating good engineering feasibility and industrial application prospects. Therefore, this invention improves the monitoring accuracy, anomaly identification capability, and intelligent collaborative response level of the entire battery production process by utilizing multi-source real-time data fusion, quantum dot sensor array spectral decoding, three-dimensional particle state reconstruction, and digital twin technology.

[0084] In this embodiment of the invention, reference is made to Figure 1 The diagram shown illustrates the steps of a battery production process data monitoring method according to the present invention. In this example, the battery production process data monitoring method includes the following steps:

[0085] Step S1: Acquire real-time multi-source data of the battery production line and images of the assembly station; analyze the current fluctuation amplitude and regional temperature difference standard deviation of the real-time multi-source data of the battery production line; identify the production stage boundary points of the assembly station images, and divide the battery production process using the current fluctuation amplitude and regional temperature difference standard deviation to generate battery production process data.

[0086] Step S2: Extract the electrode slurry preparation stage from the battery production process data, and deploy a quantum dot sensor array in the slurry stirring tank based on the electrode slurry preparation stage to obtain sensor array deployment data; decode the spectral response spectrum using the sensor array deployment data to generate slurry state spectral characteristic data;

[0087] Step S3: The microscopic particle state evolution of the electrode slurry during the preparation stage is analyzed using slurry state spectral characteristic data to generate slurry agglomeration behavior data; the slurry agglomeration behavior data is reconstructed in three dimensions, and the microscopic imbalance of slurry stirring is monitored based on the reconstruction results to generate slurry anomaly monitoring data;

[0088] Step S4: Construct a digital twin scenario based on battery production process data to generate three-dimensional monitoring data for the battery production process; coordinate battery production data with the three-dimensional monitoring data of the battery production process based on slurry anomaly monitoring data to execute full-process data monitoring of battery production.

[0089] In this embodiment of the invention, multiple sensors installed at key nodes of the battery production line collect multi-source data in real time, specifically including current sensors, voltage sensors, temperature sensors, and humidity sensors. The sampling frequency is uniformly set to 1000 Hz to ensure the capture of minute changes. The data is transmitted to edge computing nodes for preprocessing using industrial protocols (such as OPC UA). Simultaneously, a high-definition industrial camera is fixed at the battery assembly station, with a resolution set to 1920×1080 pixels and a frame rate of 30 frames per second, continuously collecting image data of the assembly process. The images are displayed in both grayscale and RGB modes for subsequent multi-feature fusion and recognition. For real-time multi-source data, a sliding window mechanism (window length of 5 seconds, sliding step size of 1 second) is used to calculate the current fluctuation amplitude, calculated as the difference between the maximum and minimum current within the window. Simultaneously, the standard deviation of the temperature sensor area data is calculated, and the temperature dispersion at each point in the 3×3 sensor grid is statistically analyzed to reflect temperature fluctuations. The image data is processed by a pre-trained deep convolutional neural network (CNN) model for boundary point recognition. The model input is the pre-processed edge-enhanced image, and the output is the coordinates of key boundary points in the assembly stage. Boundary point accuracy is controlled within ±1 pixel. Utilizing the time-series characteristics of current surge amplitude and regional temperature difference standard deviation, combined with image boundary point information, data is synchronized according to timestamps. A process segmentation algorithm is executed to divide the battery production line data into several continuous production stage intervals. The process segmentation results are stored in a timestamp and process identifier format, generating complete battery production process data. Based on the electrode slurry preparation stage time period determined in step S1, the deployment of the quantum dot sensor array in the slurry mixing vessel is initiated. The sensor array adopts a grid layout with a spacing of 5 cm, covering all key mixing areas within the slurry vessel, ensuring a total number of sensors of no less than 64 to guarantee spatial resolution. The spectral response band of the quantum dot sensors covers 400 to 900 nm, with a spectral resolution of 1 nm. The sensors acquire the optical properties of suspended particles in the slurry in real time, and the output spectral signal is transmitted at high speed to the data processing unit via optical fiber, with a sampling frequency of 500 Hz. After time-series synchronous sampling, the spectral response signal undergoes denoising (wavelet transform denoising, threshold set to 0.02). Subsequently, a spectral decoding algorithm converts the spectral data into a multi-dimensional feature vector, including peak position, full width at half maximum (FWHM), and integral area. This output forms the spectral feature data of the slurry state. Using the spectral feature data of the slurry state from step S2, combined with particle dynamics analysis, the evolution of the agglomeration state of microparticles in the slurry is calculated. Particle size distribution and mean particle size changes are analyzed through time-series comparison to generate particle agglomeration behavior data. Based on the agglomeration behavior, a three-dimensional reconstruction algorithm (based on voxel mapping technology, voxel size set to 1 cubic millimeter) is used to reconstruct the spatial three-dimensional structure of the particle distribution in the slurry. The reconstruction accuracy is controlled within 1 millimeter.Further utilizing 3D reconstruction data, the uniformity and stability of particles inside the slurry mixing tank are analyzed. Microscopic particle imbalance indicators (such as local particle density fluctuation amplitude, with a threshold of 10%) are calculated during mixing. Combined with time-series monitoring data, slurry anomalies are identified. Slurry anomaly monitoring data is output, including anomaly type, anomaly time point, and location coordinates. Based on the battery production process data generated in step S1, a digital twin scene of the battery production line is constructed using a 3D digital twin modeling platform. By importing the CAD model of the production line and combining it with production process time nodes, dynamic 3D visualization of the production line is achieved, with the model spatial resolution set to millimeters. The slurry anomaly monitoring data from step S3 is mapped to the corresponding mixing tank location in the digital twin scene, enabling real-time marking and 3D display of anomaly states. The system synchronously updates the slurry physical state and production stage information, completing data fusion. This 3D monitoring data refreshes every 1 second, allowing production managers to view the entire process data and slurry status through the interface, and supporting automatic alarms and decision support.

[0090] As an example of the present invention, reference is made to... Figure 2 As shown, in this example, step S1 includes:

[0091] Step S11: Acquire real-time multi-source data of the battery production line and images of the assembly station;

[0092] Step S12: Calculate the current fluctuation amplitude and regional temperature difference standard deviation based on real-time multi-source data from the battery production line to generate current-temperature difference joint characteristic data;

[0093] Step S13: Enhance the edge contour and extract the structural region of the assembly station image to generate structural feature data of the station image; identify the assembly action sequence of typical parts based on the structural feature data of the station image to generate boundary point data of the production stage;

[0094] Step S14: Use the current-temperature difference joint feature data to segment the production stage boundary point data into process behavior time windows to generate process feature slice data; perform key indicator clustering and dynamic stage alignment processing on the process feature slice data to generate battery production process data.

[0095] In this embodiment of the invention, a multimodal sensor system installed at key nodes of the battery production line is used to collect physical quantity data such as current, temperature, and humidity in real time. The current sensor sampling frequency is set to 2000 Hz, with a resolution of 0.1 mA. The temperature sensors are arranged in a 3×3 matrix grid, covering the area surrounding the assembly station, with a temperature sampling frequency of 1 Hz and an accuracy of ±0.1 degrees Celsius. A high-resolution industrial camera with a resolution of 3840×2160 pixels and a frame rate of 30 frames per second is installed at the assembly station to continuously collect high-definition assembly images. The camera uses a global shutter mode to avoid motion blur, and the images are transmitted to the local processing unit in real time. All collected data have a unified timestamp, and a time synchronization module ensures the consistency of the timing of multi-source data, with the time synchronization accuracy controlled within 1 millisecond. The current signal collected from the battery production line is used to calculate the abrupt change amplitude using a sliding time window method, with a time window length set to 2 seconds and a step size of 0.5 seconds. The difference between the maximum and minimum values ​​of the current signal within the window is calculated to obtain the current abrupt change amplitude data sequence. Standard deviations were calculated for temperature sensor data in the assembly station area. Temperature values ​​within a 3×3 sensor matrix were used to statistically analyze the dispersion of temperature data within each time window, with the time window length consistent with the current fluctuation amplitude calculation. Current fluctuation amplitude and temperature difference standard deviation data were paired according to timestamps to generate a current-temperature difference joint feature data sequence, with the data format being timestamp and corresponding feature pairs. Assembly station images underwent preprocessing, including grayscale conversion, noise reduction (using Gaussian filtering with a standard deviation of 1.5), and edge feature enhancement. Subsequently, the Canny edge detection algorithm was applied, with thresholds set at a low threshold of 50 and a high threshold of 150, to extract edge contours from the images. Structural region features were extracted using region growing and morphological processing to form structural feature data for the station images. Extracted features included component edge curves, contact surface areas, and connection points. Combining image frame sequences, the assembly action sequence was identified based on the Dynamic Time Warping (DTW) algorithm, clarifying the start and end times of typical component assembly actions. Based on the time nodes of the assembly actions, corresponding production stage boundary point data were generated, stored in the form of timestamp and action identifier pairs. Using the current-temperature difference joint feature data generated in step S12, time windows are used to segment the production stage boundary points in step S13. The time window length is set to 3 seconds, and the sliding step is 1 second, segmenting the time intervals corresponding to the production stage boundary points to form process feature slice data. K-means clustering algorithm is applied to key indicators in the process feature slice data (such as peak current fluctuation amplitude and temperature difference standard deviation trend), with a preset cluster number of 5 to divide the process into different dynamic stages. Subsequently, dynamic time alignment is performed on the clustering results, adjusting the time axis of the process stages to ensure smooth transitions between stages across time periods and eliminate time misalignment at slice boundaries. Finally, battery production process data is output, with the data format including process stage number, corresponding time interval, cluster category, and key indicator values. The storage structure supports real-time updates and queries.

[0096] Preferably, step S2, which involves deploying a quantum dot sensor array in the slurry stirring vessel based on the electrode slurry preparation stage, includes:

[0097] A quantum dot sensor array was deployed in the slurry stirring vessel during the electrode slurry preparation stage, and sensor array deployment data was obtained. The slurry temperature was set to be between 15 and 80°C, the viscosity range to be 500 to 5000 mPa·s, the particle size distribution to be 50 to 800 nm, the fluorescence response wavelength to be concentrated between 500 and 650 nm, and the fluorescence intensity sensitivity to be 10. 2 Up to 10 5 For relative units, the pH value of the slurry is set between 6.0 and 9.5, the conductivity range is 0.1 to 5.0 S / m, the vibration acceleration is set between 0.5 and 3.0 g, the sampling frequency of the sensor array is set between 10 and 100 Hz, and the array distribution density is 4 to 16 sensing points per square meter.

[0098] In this embodiment of the invention, before arranging the quantum dot sensor array in the slurry mixing vessel, the following physicochemical parameters of the slurry are ensured to meet the following ranges: the slurry temperature is controlled between 15°C and 80°C to ensure the stability of the sensor material and signal; the slurry viscosity is controlled between 500 and 5000 mPa·s to ensure the slurry flow state is suitable for sensor signal acquisition; and the slurry particle size distribution is maintained between 50 nm and 800 nm to ensure that the quantum dots are sensitive to particle size changes and respond accurately. A quantum dot sensor with a fluorescence response wavelength covering 500 nm to 650 nm is selected to adapt to spectral band changes caused by variations in slurry composition. The fluorescence intensity sensitivity range is set to 100 to 100,000 relative units to meet the detection requirements of different concentration changes. The sensor array design considers controlling the slurry pH value between 6.0 and 9.5 to ensure stable quantum dot luminescence and avoid chemical corrosion. The conductivity range is controlled between 0.1 and 5.0 Siemens per meter to ensure that the sensor's electrical performance is not affected by abnormal electrolytes. Due to mechanical vibrations during slurry mixing, the vibration acceleration is set between 0.5 and 3.0 times the gravitational acceleration (g). The sensor structure is designed to withstand vibration, ensuring distortion-free signal acquisition. The sensor array is uniformly arranged with 4 to 16 sensing points per square meter, covering all key mixing areas within the slurry vessel to achieve high spatial resolution detection. The sampling frequency is set between 10 Hz and 100 Hz, balancing signal timeliness and data processing load, reflecting real-time changes in the slurry state. Fluorescence intensity signals collected by each sensor are transmitted to the local data processing module via fiber optic or wireless high-bandwidth channels. Data is time-synchronized to ensure unified time stamping across multiple data points, facilitating subsequent multi-dimensional analysis. After deployment, the sensor array is calibrated using standard fluorescent liquid samples to correct the sensitivity and baseline drift of each sensing point, ensuring consistent response across the entire array. Short-term continuous testing is performed after calibration to confirm data stability and repeatability.

[0099] Preferably, step S2, which involves deploying data via a sensor array to decode the spectral response spectrum, includes:

[0100] Extracting multi-channel response signals from sensor array deployment data;

[0101] Time-domain synchronization and spectrum normalization are performed on the raw response data of multiple channels to generate standard response spectrum data;

[0102] Perform spectral difference analysis and high-dimensional noise reduction on standard response spectrum data to generate effective spectral response interval data;

[0103] Based on the effective spectral response range data, component-sensitive spectral bands are decoded to generate a slurry component reflectance coefficient matrix;

[0104] Principal component features of the reflectance coefficient matrix of the slurry components are extracted, and chemical absorption peaks are calibrated for the multi-channel response signals based on the principal component features to generate spectral feature data of the slurry state.

[0105] In this embodiment of the invention, the multi-channel spectral response signals acquired by the quantum dot sensor array are read sequentially according to channel number. Each channel contains continuous sampling data of spectral intensity changing over time, with a sampling frequency set at 50 Hz and a sampling accuracy of 12-bit digital conversion. The raw data contains noise and baseline drift, which must be effectively removed in subsequent processing. Time alignment processing is performed on the multi-channel response data, using a sensor synchronization clock to achieve precise time-domain synchronization of the multi-channel signals, with the time error controlled within 1 millisecond. The spectral intensity of each channel is normalized by subtracting the baseline average value of that channel from the sampled value and then dividing by the maximum response value of the channel, ensuring that the spectral data of all channels are uniformly within the range of 0 to 1, facilitating subsequent comparison and fusion. The spectral response data is divided into multiple preset frequency band intervals, each with a width of approximately 5 nanometers, covering the sensor's response band range. The mean and variance of the spectral response are calculated within each frequency band to identify regions of significant response fluctuation as effective spectral response intervals. Principal component analysis (PCA) is used to denoise in ineffective regions, removing high-frequency random noise and environmental interference signals to improve the signal-to-noise ratio. Based on the characteristic absorption bands of the main chemical components of the slurry, effective spectral response ranges are matched, and a reflectance coefficient matrix is ​​constructed using multi-channel data. This matrix has sensor channels as rows and sensitive spectral bands as columns, with matrix elements representing the reflectance intensity coefficient of that channel in that spectral band. Matrix factorization is used to extract the spectral response intensity corresponding to each component, eliminating interference from non-target components and obtaining highly accurate component reflectance characteristics. Principal component extraction is performed on the component reflectance coefficient matrix to obtain principal component feature vectors representing changes in slurry state. Based on the principal component features and the known chemical absorption peak positions, the absorption peak positions and intensities of the multi-channel response signals are calibrated and corrected. Finally, the slurry state spectral characteristic data is output, including principal component values ​​and chemical absorption peak information for each sensitive spectral band. The data structure supports real-time updates and trend analysis, providing accurate data for slurry quality monitoring.

[0106] Preferably, step S3, which involves analyzing the microscopic particle state evolution during the electrode slurry preparation stage using slurry state spectral characteristic data, includes:

[0107] Band reflectance fitting is performed on the spectral characteristic data of slurry state to generate multi-scale band reflectance curve data;

[0108] Particle size response mapping is performed on multi-scale band reflection curve data to generate particle size response distribution data;

[0109] Time window slicing is performed on particle size response distribution data to generate time-series particle size evolution fragment data;

[0110] Identify particle cluster aggregation trends in time-series particle size evolution data segments;

[0111] Analyze particle spacing data from particle aggregation trend data to generate particle spacing distribution data;

[0112] The particle spacing distribution data is compared with a preset particle spacing threshold. When the particle spacing distribution data is greater than or equal to the preset particle spacing threshold, the particle spacing distribution data is marked as clustered data.

[0113] Spatial density inversion is performed on the aggregation data to generate micro-aggregate concentration layer data;

[0114] The micro-agglomeration concentration layer data is spatiotemporally fused, and the stability discrete analysis of the electrode slurry preparation stage is performed based on the fused micro-agglomeration concentration layer data, ultimately generating slurry agglomeration behavior data.

[0115] In this embodiment of the invention, multiple bands covering 400 nm to 900 nm are selected from the spectral characteristic data of the slurry state, with a band width of 10 nm. Curve fitting is used on the spectral reflectance data within each band, employing smooth spline curve fitting technology to generate continuous multi-scale band reflectance curve data. The fitting error is controlled within 0.5% to ensure the reflectance curve is smooth and accurately reflects the spectral characteristics. Based on the experimentally calibrated response relationship between slurry particle size and the reflectance of the corresponding spectral band, a particle size response mapping function is established. This function converts the multi-scale band reflectance curve into the response intensity of the corresponding particle size range, divided into 50 nm to 1000 nm, with a particle size step size of 25 nm. The mapping output generates particle size response distribution data, reflecting the relative distribution density of particles of different sizes in the slurry. The particle size response distribution data is divided into fixed-length time window segments according to the time series, with each time window length set to 5 minutes and no overlap between time windows. The sliced ​​data forms time-series particle size evolution segment data, reflecting the dynamic changes in the microscopic particle size state of the slurry over time. Based on time-series particle size evolution fragment data, spatial clustering algorithms (such as DBSCAN algorithm, with a neighborhood radius of 50 nm and a minimum number of cluster points of 10) are used to identify particle cluster aggregation trends in particle size distribution, determining particle aggregation blocks and isolated particle regions. For the identified particle aggregation trend data, Euclidean distances between particles are calculated, and particle spacing distribution data is statistically generated. The statistical range is set from 0 nm to 500 nm, and the spacing distribution is divided into 20 equidistant intervals to obtain the number and density of particles within different spacing intervals. A particle spacing threshold of 100 nm is set, and a comparison operation is performed on the particle spacing distribution data. Particle data within intervals with spacing greater than or equal to this threshold are marked as aggregation data to reflect the degree of particle aggregation and structural characteristics. An inversion algorithm is applied to the marked aggregation data to transform discrete particle aggregation information into a spatial density layer. The spatial resolution is set to 1 square millimeter per cell, and the kernel density estimation method is used to calculate local particle density, generating micro-aggregate concentration layer data. The micro-agglomeration concentration layers from multiple time windows were spatiotemporally fused, with fusion weights decreasing according to temporal proximity and equal weights for spatially adjacent units. After fusion, the micro-agglomeration concentration layers were discretized, and the standard deviation statistical index was used to evaluate the spatial stability and temporal fluctuations of agglomeration behavior. This ultimately generated slurry agglomeration behavior data, serving as an important characterization of the microstructure state during slurry preparation.

[0116] Of particular importance, identifying particle cluster aggregation trends in time-series particle size evolution fragment data also includes:

[0117] Multi-time-window moving average processing is performed on time-series particle size evolution fragment data to extract local particle size increment trend data;

[0118] Based on the particle size increment trend data, aggregated vector field analysis is performed on the particle boundary profile to generate data on the relative motion direction of particles.

[0119] By using data on the relative motion direction of particles, particle size overlap blocks can be identified, and particle contact density feature data can be extracted.

[0120] Cluster boundary identification is performed on particle contact density characteristic data to generate initial region data for potential particle clusters;

[0121] Based on the initial region data of particle clusters, continuous intra-frame consistency verification is performed to screen stable aggregation regions and generate particle cluster aggregation trend data.

[0122] In this embodiment of the invention, particle size distribution data sequences at multiple time points are extracted from the electrode slurry preparation stage, with each set of data corresponding to a particle image of the same observation area. The time interval is set to once per second, and a total of 60 frames are collected to form a "time-series particle size evolution segment data". For this data segment, a sliding time window averaging method is used, with five frames as a sliding window and a window step size of one frame, to calculate the local mean of the particle diameter identified in each frame. This method can smooth out drastic fluctuations in the time series while preserving the detailed changes in particle size, thereby extracting local particle size increment trend data. The output data is a sequence of the mean particle size change of each particle over time, in micrometers. Based on the local particle size increment trend data, combined with the particle boundary extraction results from the original image data, the particle boundary contours are analyzed. The relative displacement of particle boundaries in adjacent time frames is calculated using image differencing to determine the offset path of each particle in consecutive time frames. This path data is used to construct a particle aggregation vector field. Each vector in this field represents the relative motion direction between the boundaries of two adjacent particles, pointing from the center of one particle to the center of the adjacent particle. The length represents the distance the two particles approach each other per unit time, measured in micrometers per second. This vector field can be viewed as a trend mapping of particle aggregation or separation, and the output is "particle relative motion direction data". Based on the particle relative motion direction data, regions where particle boundaries overlap or approach each other within consecutive time frames are identified. If the center distance between two particles is less than 1.2 times the average particle diameter, and the number of overlapping pixels exceeds 5 pixels, it is considered that a particle size overlap block exists. The number of adjacent particles of each particle in a unit region is counted, and the minimum distance between their boundaries is calculated. Combined with the degree of boundary overlap, contact tightness feature data is constructed. This feature is represented numerically, ranging from 0 (no contact) to 1 (complete fit), with each particle corresponding to a contact tightness value. Using contact density feature data, density-based spatial clustering methods (e.g., fixed threshold method) are employed to identify cluster boundaries. If the contact density of multiple consecutive particles within a region exceeds 0.6 and the distance between particles is less than twice the average particle size, the region is identified as a potential cluster. The clustering output is "Initial Region Data of Particle Clusters," formatted as multiple closed contour regions, each containing attributes such as particle ID, location coordinates, and local density value. Temporal consistency verification is performed on the initial regions of particle clusters. If the same cluster region remains stable across 10 consecutive frames (i.e., particle ID and location overlap exceeds 80%), the region is identified as a "Stable Aggregation Region." Finally, these time-verified cluster regions are output as "Particle Cluster Aggregation Trend Data," each trend data point containing information such as cluster location, duration, average particle size, and aggregation rate (particle spacing shortening rate). All information is recorded in structured tables for further analysis of slurry dispersion or agglomeration.

[0123] Preferably, step S3, which involves reconstructing the three-dimensional distribution of the slurry agglomeration behavior data, includes:

[0124] Extract the spatial coordinate index of the slurry agglomeration behavior data to obtain the agglomeration spatial coordinate point set data;

[0125] Point cloud densification interpolation is performed on the clustered spatial coordinate point set data to generate particle clustered point cloud data;

[0126] Local voxelization encoding of particle agglomeration point cloud data, and global mesh splicing reconstruction of the encoded particle agglomeration point cloud data to generate slurry agglomeration 3D mesh skeleton data;

[0127] Particle size channel attribute mapping is performed on the 3D mesh skeleton data of slurry agglomeration to generate attribute-annotated mesh data;

[0128] Multi-view projection rendering is performed on attribute-annotated mesh data to generate 3D view data of particle aggregation;

[0129] Dynamic sequence frame encoding is performed based on particle aggregation 3D view data to generate adjustable-speed particle aggregation evolution sequence data.

[0130] Stability hot zone annotation is performed on particle aggregation evolution sequence data to obtain the reconstruction result of three-dimensional distribution reconstruction.

[0131] In this embodiment of the invention, the three-dimensional spatial coordinates of all particle aggregation positions are extracted from slurry aggregation behavior data. The coordinates are accurate to the millimeter level, and the coordinate range corresponds to the internal space dimensions of the slurry mixing vessel, approximately 1000 mm × 1000 mm × 500 mm. The generated spatial coordinate point set is stored in a structured array for easy subsequent processing. The extracted spatial coordinate point set is densified using an interpolation algorithm based on the inverse distance weighting method. The interpolation radius is set to 20 mm to ensure a uniform distribution of the point cloud density after interpolation, filling the gaps in the original sampling and generating continuous and high-density particle aggregation point cloud data, with the average distance between points controlled within 5 mm. The particle aggregation point cloud data is divided into cubic voxel units with a side length of 10 mm. The number of points contained in each voxel unit is counted and encoded as a voxel density value. All local voxel units are spatially stitched together using an adjacency algorithm to form a global three-dimensional mesh skeleton data. The mesh side length is also 10 mm to ensure spatial continuity and structural integrity. Based on the particle size information corresponding to the original slurry agglomeration behavior data, particle size is mapped as an attribute to the corresponding grid cells. The particle size range is divided into five levels, corresponding to the following ranges: 50-150 nm, 151-300 nm, 301-500 nm, 501-800 nm, and 801-1000 nm. Each particle size level is assigned a different attribute value, generating attribute-annotated grid data that reflects the particle size distribution characteristics. The attribute-annotated grid data is rendered using two-dimensional projection from six fixed perspectives (front, back, left, right, top, and bottom) at a resolution of 1024×1024 pixels. The rendering employs volume rendering technology, supports semi-transparent effects, achieves color mapping of particle size attributes, and generates three-dimensional view data of particle agglomeration. The 3D view data of particle agglomeration collected at multiple time points is arranged chronologically at a frame rate of 30 frames per second. Inter-frame differential encoding is used to compress the data, preserving spatial location and particle size variation information. This generates a particle agglomeration evolution sequence data that can be played at adjustable speed, supporting dynamic analysis operations such as fast forward and slow motion. Based on the particle agglomeration evolution sequence data, the stability index of local regions is calculated by statistically analyzing the changes in grid density within consecutive time frames. For regions with stability indices higher than a threshold (set to a density fluctuation of less than 5% within 10 consecutive frames), hot zone annotations are performed to form 3D stability hot zone data, thus reconstructing the spatial distribution of slurry agglomeration.

[0132] Preferably, step S3, which involves monitoring the micro-imbalance of slurry mixing based on the reconstruction results, includes:

[0133] Based on the reconstruction results of the three-dimensional distribution reconstruction, the particle structure orientation information within the stirring area is extracted to generate local structure orientation data;

[0134] Spatial anisotropy tensor analysis is performed on local structural orientation data to generate orientation tensor deviation data;

[0135] Using orientation tensor deviation data, vector trajectory deduction of particle flow trend in electrode slurry preparation stage is performed to generate microscopic flow direction perturbation data.

[0136] The local shear stress gradient during the stirring process is calculated based on the microscopic flow disturbance data, and a shear imbalance intensity map is generated.

[0137] The stability domain boundary of the shear imbalance intensity map is identified, and particle aggregation time window statistics are performed on the stability domain boundary to generate particle swarm fluctuation interval data.

[0138] Microscopic disturbance frequency analysis was performed on the particle size fluctuation range data to generate slurry anomaly monitoring data.

[0139] In this embodiment of the invention, principal component analysis (PCA) is used to calculate the eigenvectors of particle point clouds within each voxel unit using slurry agglomerate voxel data reconstructed from a 3D distribution, obtaining the principal direction vectors of the local particle structure. The spatial resolution is set to a cube with a side length of 10 mm, and the extracted results form a local structure orientation dataset, with accuracy controlled within the range of nanoscale particle structure orientation variations. A 3D anisotropic tensor matrix is ​​constructed from the local structure orientation data, reflecting the uniformity and deviation of the orientation distribution. By calculating the eigenvalue distribution of the tensor, orientation tensor deviation data is generated. The deviation reflects the difference between the local structure and the ideal uniform distribution, with values ​​ranging from 0 to 1; a larger value indicates a more significant deviation. Using the orientation tensor deviation data, combined with a fluid dynamics simulation algorithm, the flow trend of particles inside the slurry under the influence of stirring is simulated. Particle flow direction vector trajectories are generated using the vector field integration method, with a time step of 0.1 seconds and a trajectory length covering the entire stirring cycle, generating a microscopic flow direction perturbation dataset. Based on the microscopic flow direction perturbation data, the local shear stress gradient distribution is calculated using a finite element shear stress model. The model parameters include slurry viscosity and shear rate. The spatial resolution of the shear stress gradient is 10 mm, generating a shear imbalance intensity map. The numerical range of the map corresponds to the magnitude of the shear stress gradient, reflecting the shear uniformity within the mixing area. An edge detection algorithm (such as the Canny algorithm) is applied to the shear imbalance intensity map to identify the stability boundaries between shear equilibrium and imbalance regions. For regions within these boundaries, a time window of 30 seconds is set to statistically analyze the number of particle aggregation changes and their temporal distribution, generating particle swarm fluctuation interval data to reflect the dynamic stability of the slurry's microstructure. Spectral analysis methods (such as Fast Fourier Transform) are used to calculate the perturbation frequency distribution of the particle swarm fluctuation interval data, identifying high-frequency perturbation regions and low-frequency stable regions. Through frequency threshold filtering, slurry anomaly monitoring data is generated, indicating microscopic imbalances and abnormal fluctuation points present during slurry mixing.

[0140] Of particular importance, the vector trajectory extrapolation of particle flow trends during the electrode slurry preparation stage using orientation tensor deviation data also includes:

[0141] Based on the orientation tensor deviation data, a particle flow direction vector field is constructed to generate a local particle flow vector distribution map.

[0142] Multi-scale grid interpolation is performed on the particle flow vector distribution map to generate continuous particle flow trajectory data;

[0143] By using particle flow continuous trajectory data to dynamically extrapolate vector trajectories, abnormal disturbance nodes are identified, and microscopic trajectory disturbance identification data is generated.

[0144] The micro-trajectory disturbance identification data is subjected to disturbance intensity assessment and direction decomposition, and finally micro-flow direction disturbance data is generated.

[0145] In this embodiment of the invention, particle orientation tensor deviation data in different observation areas are calculated using multidimensional spectral response maps acquired based on a quantum dot sensor array. The deviation is calculated based on the angular deviation between the principal tensor direction and the theoretical uniform flow direction, with a numerical range of 0 to 90 degrees. Using the spatial unit where each sensor group is located (each unit has a volume of approximately 1 cubic centimeter) as the basic unit, the distribution of the main flow direction of particles in three-dimensional space is summarized. The principal tensor direction is converted into a standard three-dimensional vector form, and combined with the deviation value, a local particle flow direction vector field is constructed. In the actual construction process, each vector has specific position coordinates, direction components (X, Y, Z), and a deviation label. The output is a particle flow vector distribution map, where each vector represents the main motion direction of the particles with an arrow, and the degree of deviation is represented by a color gradient (e.g., blue to red), with red indicating a larger deviation and blue indicating a closer approximation of uniform flow. Based on the particle flow direction vector field data, a multi-layered grid system is established according to different spatial scales. The basic grid cell size is set to 2 cubic millimeters, the medium-scale size to 5 cubic millimeters, and the large-scale size to 1 cubic centimeter, with three levels operating simultaneously to cover particle motion characteristics at different scales. Three-dimensional spline interpolation is used to estimate the continuity of vector directions within each grid layer, eliminating local data gaps caused by uneven sensor distribution. After interpolation, continuous particle flow trajectory data is generated, recorded as a set of path points. Each trajectory consists of continuous spatial points, with velocity direction and gradient change values ​​recorded at each point. This trajectory data is used for subsequent dynamic evolution analysis. Dynamic trajectory extrapolation is performed using the continuous particle flow trajectory data. The extrapolation is based on two key indicators: the velocity field change rate (i.e., trajectory curvature change) and the trajectory angle offset rate. If the angle change of a trajectory segment between three consecutive points exceeds 45 degrees, or the radius of curvature is less than 2 millimeters, it is marked as a "potentially disturbed segment"; if this change overlaps in more than three trajectories, it is identified as a disturbed node. These abnormal deformation segments are numbered, located, and mapped one-to-one with spatial coordinates, ultimately generating microscopic trajectory disturbance identification data. This data includes the location (XYZ coordinates) of the disturbance nodes, the corresponding time point, the disturbance type (sharp turn, sudden velocity change, direction reversal, etc.), and a visual identifier. The disturbance intensity of each micro-disturbance node is quantitatively assessed. The assessment uses three dimensions: directional offset angle: the angle with the average flow direction exceeds 30 degrees; local velocity fluctuation: the standard deviation of the velocity exceeds 20% of the overall average velocity; and curvature change rate: the rate of curvature change per unit time is greater than 10 degrees / millisecond. Each disturbance node is scored based on the above three indicators, and the scores are accumulated to obtain the disturbance intensity level, categorized as low disturbance (below 1 point), medium disturbance (1–2 points), and high disturbance (above 2 points). Simultaneously, the disturbance direction is vector-decomposed into XYZ components to analyze whether there is a concentrated trend in directional offset.The final output micro-flow disturbance data is a structured data table, including disturbance location, disturbance type, disturbance intensity level, directional component information, fluctuation duration, associated trajectory number, etc.

[0146] Preferably, the micro-perturbation frequency analysis of particle swarm fluctuation range data includes:

[0147] Micro-disturbance frequency analysis is performed on the particle size fluctuation range data. When any of the following conditions occur, it is determined to be an abnormal particle size concentration, and abnormal particle size concentration data is obtained: the characteristic particle size distribution range narrows by more than 20%, the D90 / D10 particle size ratio fluctuates by more than ±0.6, and the particle size distribution skewness index deviates from the average value by more than ±15% within three consecutive sampling periods.

[0148] When the following conditions occur simultaneously, it is determined to be a phenomenon of enhanced micro-agglomeration disturbance, and data on enhanced agglomeration disturbance are obtained: the ratio of agglomerates within the particle group increases by more than 25% within 10 minutes, the average agglomerate size exceeds the historical limit by more than 10%, and the density distribution inside the agglomerates shows a bimodal shift that lasts for more than 20 minutes.

[0149] When the following conditions are met simultaneously, it is determined to be an abnormal shear fluctuation state, and abnormal shear fluctuation data is obtained: the frequency of the stirring system rotation speed fluctuation exceeds 2Hz, the frequency of the particle group micro-perturbation response deviates from the reference spectrum range by more than ±10% during the stirring process, accompanied by an increase of more than 8% in the content of microbubbles in the slurry, and the duration of this state exceeds 30 minutes.

[0150] By integrating abnormal particle size concentration data, enhanced agglomeration disturbance data, and abnormal shear fluctuation data, and performing correlation weight evaluation and disturbance type fusion identification, slurry anomaly monitoring data is finally generated.

[0151] In this embodiment of the invention, dynamic analysis of particle size distribution data within the particle size fluctuation range is performed using a sliding time window method. The time window length is set to three consecutive sampling periods (typically 5 minutes per period, totaling 15 minutes): The system monitors changes in the characteristic particle size distribution range. If the range narrows by more than 20%, i.e., the difference between the maximum and minimum particle size values ​​within the current time window decreases by more than one-fifth compared to the historical average, an anomaly is identified. The ratio of D90 to D10 particle sizes is calculated; if the fluctuation of this ratio exceeds ±0.6, it is considered an abnormal fluctuation. The particle size distribution skewness index is statistically analyzed; if it deviates from the historical average by more than 15% within three consecutive sampling periods, it is considered an abnormal skewness. When any of the above conditions are met, the system automatically records the current abnormal period and corresponding data, generating abnormal particle size concentration data. Within a 10-minute statistical window, the quantity and properties of particle agglomerates are analyzed: if the agglomerate quantity ratio increases by more than 25% compared to the previous baseline, it is considered an abnormal increase in quantity; the average particle size of agglomerates is monitored, and if it exceeds the historical maximum value by more than 10%, it is determined to be an agglomerate size anomaly; the density distribution characteristics inside the agglomerates are analyzed, and if a bimodal shift occurs (i.e., the density distribution map shows two obvious peaks with offset positions) and this state lasts for more than 20 minutes, it is confirmed that the agglomeration disturbance is enhanced. When the above three conditions are met simultaneously, data on enhanced agglomeration disturbance is generated. The status of the slurry mixing system and the microscopic response of the slurry are monitored, and the following judgment conditions are set: the fluctuation frequency of the mixing system speed exceeds 2 times per second; the particle group micro-disturbance response frequency (measured by vibration sensor or particle size change frequency) deviates from the reference spectrum by more than ±10%; at the same time, the microbubble content in the slurry increases by more than 8% compared to the normal level; and the above state lasts for no less than 30 minutes. When all conditions are met, it is determined to be an abnormal shear fluctuation, and corresponding abnormal data is generated. The data on abnormal particle size concentration, enhanced agglomeration disturbance, and abnormal shear fluctuation are compared and correlated with weights in time and space. The weight index is calculated based on the frequency, duration, and intensity of abnormal data. Multi-factor fusion algorithms (such as weighted average fusion and fuzzy logic reasoning) are used to identify the disturbance type of different types of abnormal data. Finally, structured slurry anomaly monitoring data is formed, which includes anomaly type classification, severity level, and corresponding time period identifier.

[0152] As an example of the present invention, reference is made to... Figure 3 As shown, step S4 in this example includes:

[0153] Step S41: Extract the parameters of each assembly station, the flow of key materials and the timeline of the process from the battery production process data, perform spatial mapping, and generate structured digital scene data of the battery production process.

[0154] Step S42: Perform 3D visualization modeling on the structured digital scene data to generate 3D monitoring data for the battery production process; based on the slurry anomaly monitoring data, precisely locate the corresponding electrode slurry preparation stage in the 3D monitoring data to generate anomaly mapping segment data;

[0155] Step S43: Perform parameter cross-comparison between the abnormal mapping section data and the 3D monitoring data to generate visual abnormality marker layer data; overlay the visual abnormality marker layer data onto the 3D monitoring data of the battery production process to generate fused monitoring data with abnormality feedback information;

[0156] Step S44: Perform station-level data flow tracing on the fused monitoring data, identify the potential transmission path of slurry anomalies to downstream process nodes, and generate a process-level anomaly transmission map; use the process-level anomaly transmission map to perform data collaborative updates of the three-dimensional monitoring data of the battery production process, and finally realize the data monitoring operation of the entire battery production process.

[0157] In this embodiment of the invention, process parameter data of each assembly station in the battery production process is collected by calling the real-time database interface of the production line. This includes, but is not limited to, current, voltage, temperature, and pressure. The sampling frequency is set to once per second to ensure data integrity. Simultaneously, the material tracking system of the assembly station is used to obtain the flow path and assembly sequence information of key materials (such as slurry and electrode sheets), with timestamps accurate to the millisecond level. The process timeline is extracted through the process management system to clarify the start and end times and duration of each process, with time errors controlled within 100 milliseconds. After spatiotemporal synchronization processing, the above data is mapped to a unified spatial coordinate system. The XYZ three-dimensional coordinates are used to represent the assembly station layout, completing the spatial mapping of station parameters, material flow, and timelines. This generates structured digital scene data, using a three-dimensional vector model with time-series tags, and the file size is controlled to within approximately 500MB to ensure real-time processing capabilities. Import the structured digital scene data generated in step S41 using professional 3D modeling software (such as Unity 3D or Unreal Engine) for high-precision modeling. The model's spatial accuracy is set to 1 mm to ensure that the spatial position and size of the assembly station accurately reflect the real production environment. Combined with slurry anomaly monitoring data, anomaly events are precisely mapped to the corresponding assembly station and time period in the 3D model according to timestamps, with the positioning time period error controlled within ±0.5 seconds, generating anomaly mapping segment data. This step includes time synchronization and spatial comparison of anomaly events, using a time window mechanism to expand the segment by 5 seconds before and after the anomaly time point to cover the impact range of the anomaly. For the anomaly mapping segment data, extract the change trends of key parameters (current, voltage, temperature, etc.) of the corresponding assembly station, and identify anomaly feature points through parameter threshold comparison and change rate calculation. Set anomaly judgment thresholds, such as current fluctuations exceeding 20 amperes or temperature differences exceeding 3 degrees Celsius, to automatically mark anomaly nodes. Generate a visual anomaly marker layer as a semi-transparent red highlighted layer, with the layer's spatial resolution consistent with the 3D model. By utilizing layer overlay technology, anomaly marker layers are accurately superimposed onto the 3D monitoring data of the battery production process, ensuring real-time display and spatial accuracy, and generating fused monitoring data with anomaly feedback information. Based on the fused monitoring data, the production line material flow and process flow database is invoked to track downstream process nodes of the abnormal workstations along the time axis, with time tracking accuracy controlled within 1 second to ensure accurate tracking paths. Topology analysis is used to identify anomaly propagation paths, including material flow direction propagation and process parameter impact transmission. The tracking results are structured into a process-level anomaly propagation map. Map nodes include workstation identification, anomaly level, and propagation time, while edges represent propagation relationships and impact strength, with impact strength represented as decimals between 0 and 1. The map drives the dynamic updating of 3D monitoring data, reflecting the scope and propagation trend of anomaly impacts in real time, ensuring the integrity and real-time nature of data collaboration, and ultimately achieving full-process data monitoring of battery production.

[0158] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0159] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for monitoring data throughout the entire battery production process, characterized in that, Includes the following steps: Step S1: Acquire real-time multi-source data of the battery production line and images of the assembly station; analyze the current fluctuation amplitude and regional temperature difference standard deviation of the real-time multi-source data of the battery production line; identify the production stage boundary points of the assembly station images, and divide the battery production process using the current fluctuation amplitude and regional temperature difference standard deviation to generate battery production process data. Step S2: Extract the electrode slurry preparation stage from the battery production process data, and deploy a quantum dot sensor array in the slurry stirring tank based on the electrode slurry preparation stage to obtain sensor array deployment data; decode the spectral response spectrum using the sensor array deployment data to generate slurry state spectral characteristic data; Step S3: The microscopic particle state evolution during the electrode slurry preparation stage is analyzed using slurry state spectral characteristic data to generate slurry agglomeration behavior data; The slurry agglomeration behavior data is reconstructed into a three-dimensional distribution, and the micro-imbalance of slurry mixing is monitored based on the reconstruction results to generate slurry anomaly monitoring data; wherein, the three-dimensional distribution reconstruction of slurry agglomeration behavior data in step S3 includes: Extract the spatial coordinate index of the slurry agglomeration behavior data to obtain the agglomeration spatial coordinate point set data; Point cloud densification interpolation is performed on the clustered spatial coordinate point set data to generate particle clustered point cloud data; Local voxelization encoding of particle agglomeration point cloud data, and global mesh splicing reconstruction of the encoded particle agglomeration point cloud data to generate slurry agglomeration 3D mesh skeleton data; Particle size channel attribute mapping is performed on the 3D mesh skeleton data of slurry agglomeration to generate attribute-annotated mesh data; Multi-view projection rendering is performed on attribute-annotated mesh data to generate 3D view data of particle aggregation; Dynamic sequence frame encoding is performed based on particle aggregation 3D view data to generate adjustable-speed particle aggregation evolution sequence data. Stability hot zone annotation is performed on particle aggregation evolution sequence data to obtain the reconstruction result of three-dimensional distribution reconstruction; Step S4: Construct a digital twin scenario based on battery production process data to generate three-dimensional monitoring data for the battery production process; coordinate battery production data with the three-dimensional monitoring data of the battery production process based on slurry anomaly monitoring data to execute full-process data monitoring of battery production.

2. The battery production process data monitoring method according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Acquire real-time multi-source data of the battery production line and images of the assembly station; Step S12: Calculate the current fluctuation amplitude and regional temperature difference standard deviation based on real-time multi-source data from the battery production line to generate current-temperature difference joint characteristic data; Step S13: Enhance the edge contour and extract the structural region of the assembly station image to generate structural feature data of the station image; identify the assembly action sequence of typical parts based on the structural feature data of the station image to generate boundary point data of the production stage; Step S14: Use the current-temperature difference joint feature data to segment the production stage boundary point data into process behavior time windows to generate process feature slice data; perform key indicator clustering and dynamic stage alignment processing on the process feature slice data to generate battery production process data.

3. The battery production process data monitoring method according to claim 1, characterized in that, Step S2, which involves deploying a quantum dot sensor array in the slurry mixing vessel based on the electrode slurry preparation stage, includes: A quantum dot sensor array was deployed in the slurry stirring vessel during the electrode slurry preparation stage, and sensor array deployment data was obtained. The slurry temperature was set to be between 15 and 80°C, the viscosity range to be 500 to 5000 mPa·s, the particle size distribution to be 50 to 800 nm, the fluorescence response wavelength to be concentrated between 500 and 650 nm, and the fluorescence intensity sensitivity to be between 10² and 10⁻⁶. 5 For relative units, the pH value of the slurry is set between 6.0 and 9.5, the conductivity range is 0.1 to 5.0 S / m, the vibration acceleration is set between 0.5 and 3.0 g, the sampling frequency of the sensor array is set between 10 and 100 Hz, and the array distribution density is 4 to 16 sensing points per square meter.

4. The battery production process data monitoring method according to claim 1, characterized in that, Step S2, which involves deploying data via a sensor array to decode the spectral response spectrum, includes: Extracting multi-channel response signals from sensor array deployment data; Time-domain synchronization and spectrum normalization are performed on the raw response data of multiple channels to generate standard response spectrum data; Perform spectral difference analysis and high-dimensional noise reduction on standard response spectrum data to generate effective spectral response interval data; Based on the effective spectral response range data, component-sensitive spectral bands are decoded to generate a slurry component reflectance coefficient matrix; Principal component features of the reflectance coefficient matrix of the slurry components are extracted, and chemical absorption peaks are calibrated for the multi-channel response signals based on the principal component features to generate spectral feature data of the slurry state.

5. The battery production process data monitoring method according to claim 1, characterized in that, Step S3 involves analyzing the microscopic particle state evolution during the electrode slurry preparation stage using slurry state spectral characteristic data, including: Band reflectance fitting is performed on the spectral characteristic data of slurry state to generate multi-scale band reflectance curve data; Particle size response mapping is performed on multi-scale band reflection curve data to generate particle size response distribution data; Time window slicing is performed on particle size response distribution data to generate time-series particle size evolution fragment data; Identify particle cluster aggregation trends in time-series particle size evolution data segments; Analyze particle spacing data from particle aggregation trend data to generate particle spacing distribution data; The particle spacing distribution data is compared with a preset particle spacing threshold. When the particle spacing distribution data is greater than or equal to the preset particle spacing threshold, the particle spacing distribution data is marked as clustered data. Spatial density inversion is performed on the aggregation data to generate micro-aggregate concentration layer data; The micro-agglomeration concentration layer data is spatiotemporally fused, and the stability discrete analysis of the electrode slurry preparation stage is performed based on the fused micro-agglomeration concentration layer data, ultimately generating slurry agglomeration behavior data.

6. The battery production process data monitoring method according to claim 1, characterized in that, Step S3, which involves monitoring the micro-imbalance of slurry mixing based on the reconstruction results, includes: Based on the reconstruction results of the three-dimensional distribution reconstruction, the particle structure orientation information within the stirring area is extracted to generate local structure orientation data; Spatial anisotropy tensor analysis is performed on local structural orientation data to generate orientation tensor deviation data; Using orientation tensor deviation data, vector trajectory deduction of particle flow trend in electrode slurry preparation stage is performed to generate microscopic flow direction perturbation data. The local shear stress gradient during the stirring process is calculated based on the microscopic flow disturbance data, and a shear imbalance intensity map is generated. The stability domain boundary of the shear imbalance intensity map is identified, and particle aggregation time window statistics are performed on the stability domain boundary to generate particle swarm fluctuation interval data. Microscopic disturbance frequency analysis was performed on the particle size fluctuation range data to generate slurry anomaly monitoring data.

7. The battery production process data monitoring method according to claim 6, characterized in that, Micro-perturbation frequency analysis of particle swarm fluctuation range data includes: Micro-disturbance frequency analysis is performed on the particle size fluctuation range data. When any of the following conditions occur, it is determined to be an abnormal particle size concentration, and abnormal particle size concentration data is obtained: the characteristic particle size distribution range narrows by more than 20%, the D90 / D10 particle size ratio fluctuates by more than ±0.6, and the particle size distribution skewness index deviates from the average value by more than ±15% within three consecutive sampling periods. When the following conditions occur simultaneously, it is determined to be a phenomenon of enhanced micro-agglomeration disturbance, and data on enhanced agglomeration disturbance are obtained: the ratio of agglomerates within the particle group increases by more than 25% within 10 minutes, the average particle size of agglomerates exceeds the historical limit by more than 10%, and the density distribution inside the agglomerates shows a bimodal shift that lasts for more than 20 minutes. When the following conditions are met simultaneously, it is determined to be an abnormal shear fluctuation state, and abnormal shear fluctuation data is obtained: the frequency of the stirring system rotation speed fluctuation exceeds 2Hz, the frequency of the particle group micro-perturbation response deviates from the reference spectrum range by more than ±10% during the stirring process, accompanied by an increase of more than 8% in the content of microbubbles in the slurry, and the duration of this state exceeds 30 minutes. By integrating abnormal particle size concentration data, enhanced agglomeration disturbance data, and abnormal shear fluctuation data, and performing correlation weight evaluation and disturbance type fusion identification, slurry anomaly monitoring data is finally generated.

8. The battery production process data monitoring method according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Extract the parameters of each assembly station, the flow of key materials and the timeline of the process from the battery production process data, perform spatial mapping, and generate structured digital scene data of the battery production process. Step S42: Perform 3D visualization modeling on the structured digital scene data to generate 3D monitoring data for the battery production process; based on the slurry anomaly monitoring data, precisely locate the corresponding electrode slurry preparation stage in the 3D monitoring data to generate anomaly mapping segment data; Step S43: Perform parameter cross-comparison between the abnormal mapping section data and the 3D monitoring data to generate visual abnormality marker layer data; overlay the visual abnormality marker layer data onto the 3D monitoring data of the battery production process to generate fused monitoring data with abnormality feedback information; Step S44: Perform station-level data flow tracing on the fused monitoring data, identify the potential transmission path of slurry anomalies to downstream process nodes, and generate a process-level anomaly transmission map; use the process-level anomaly transmission map to perform data collaborative updates of the three-dimensional monitoring data of the battery production process, and finally realize the data monitoring operation of the entire battery production process.

9. A data monitoring system for the entire battery production process, characterized in that, For executing the battery production process data monitoring method as described in claim 1, the battery production process data monitoring system includes: The process division module is used to acquire real-time multi-source data of the battery production line and images of the assembly station; analyze the current fluctuation amplitude and regional temperature difference standard deviation of the real-time multi-source data of the battery production line; identify the production stage boundary points of the assembly station images; and divide the battery production process using the current fluctuation amplitude and regional temperature difference standard deviation to generate battery production process data. The slurry analysis module is used to extract data from the electrode slurry preparation stage of the battery production process, and to deploy a quantum dot sensor array in the slurry stirring vessel based on the electrode slurry preparation stage to obtain sensor array deployment data; the spectral response spectrum is decoded through the sensor array deployment data to generate slurry state spectral characteristic data; The imbalance monitoring module is used to perform microscopic particle state evolution analysis on electrode slurry preparation stage using slurry state spectral characteristic data, generating slurry agglomeration behavior data; it performs three-dimensional distribution reconstruction on slurry agglomeration behavior data, and monitors microscopic imbalance of slurry stirring based on the reconstruction results, generating slurry anomaly monitoring data. The data collaboration module is used to construct digital twin scenarios based on battery production process data and generate three-dimensional monitoring data for the battery production process; it also collaborates with the three-dimensional monitoring data of the battery production process based on slurry anomaly monitoring data to perform data monitoring operations for the entire battery production process.

Images