Lithium battery full life cycle quality tracing system based on digital twinning

Abstract

The invention discloses a lithium battery full life cycle quality tracing system based on digital twinning, and relates to the technical field of lithium battery quality tracing, and the system comprises a physical entity layer which integrates multiple types of sensors and detection equipment to collect full-link data; the digital twinning modeling layer is used for constructing a multi-scale model to map a battery state; the data acquisition and transmission layer is used for realizing data real-time transmission preprocessing through an edge gateway and an encryption protocol; the data storage and management layer is used for establishing a data warehouse by using a distributed database and performing backup; the intelligent analysis and decision-making layer integrates an algorithm to realize anomaly detection, performance prediction and tracing; the visualization and interaction layer is used for providing a multi-terminal visualization and interaction function; and the block chain evidence storage layer constructs an alliance chain to ensure that the data cannot be tampered. According to the invention, full-life-cycle data connection and credible tracing of the lithium battery are realized, the quality problem positioning efficiency is improved, intelligent early warning and optimization capabilities are provided, the operation threshold is reduced, and industrial quality control and green development are facilitated.

Description

Technical Field

[0001] This invention relates to the field of lithium battery quality traceability technology, and in particular to a lithium battery full life cycle quality traceability system based on digital twins. Background Technology

[0002] With the rapid development of new energy vehicles and energy storage industries, lithium batteries, as core energy components, have become a core focus of the manufacturing industry in terms of quality, safety, and full life-cycle traceability. Currently, the lithium battery production process encompasses multiple stages, including cell manufacturing, module assembly, and pack assembly, and involves warehousing, transportation, end-use applications, and decommissioning / recycling. Data from each stage is scattered across different systems, forming "data silos." Traditional quality traceability methods rely on manual recording and barcode scanning, which can only link information from some key nodes and cannot achieve seamless data flow from raw materials to decommissioning / recycling. When quality problems such as rapid capacity decay or thermal runaway occur, it is difficult to quickly pinpoint the root cause, resulting in low traceability efficiency and a tendency for blame-shifting.

[0003] Existing traceability systems have significant shortcomings in data integrity and analytical capabilities. On the one hand, data collection dimensions are limited, focusing primarily on basic parameters in the production process while neglecting parameters that significantly impact battery performance, such as storage environment temperature and humidity, transportation vibration intensity, and terminal charge / discharge rates. This results in an inability to fully reconstruct the scenarios in which quality issues arose during traceability analysis. On the other hand, there is a lack of deep data integration and intelligent analysis, enabling only post-event traceability and failing to predict performance degradation trends or provide early warnings of quality risks. For example, when a battery experiences a sudden capacity drop in end-use applications, traditional systems can only query factory inspection data and cannot combine data such as full-lifecycle charge / discharge records and environmental exposure history to determine whether the cause is a production defect or improper use, making it difficult to form a closed-loop quality improvement mechanism.

[0004] Furthermore, the application of digital twin technology in the lithium battery field is mostly limited to production process simulation, lacking deep integration with quality traceability. It cannot map the entire lifecycle state of physical batteries in real time through virtual models. Simultaneously, data reliability is lacking, and key quality parameters are at risk of tampering, resulting in insufficient authority of traceability results. With the large-scale expansion of the lithium battery industry and increasingly stringent quality standards, traditional traceability models can no longer meet the needs of full lifecycle quality control. There is an urgent need for a full lifecycle quality traceability system that integrates data from multiple stages, possesses intelligent analysis capabilities, and ensures data immutability. This system aims to improve the efficiency of locating quality problems, ensure safe battery operation, and promote the green and sustainable development of the industry. Summary of the Invention

[0005] The present invention proposes a lithium battery full life cycle quality traceability system based on digital twins to solve the problems mentioned in the prior art.

[0006] In order to achieve the above object, the present application adopts the following technical solutions:

[0007] A lithium battery full life cycle quality traceability system based on digital twinning includes:

[0008] A physical entity layer includes lithium battery cell production equipment, module assembly line, Pack packaging equipment, warehouse environment monitoring device, transportation state acquisition equipment and terminal application sensor;

[0009] A digital twinning modeling layer constructs a lithium battery full life cycle multi-scale digital twinning model, including a cell microstructure model, a module structure mechanics model, a Pack system thermal simulation model and a full life cycle performance attenuation model;

[0010] A data acquisition and transmission layer deploys an edge gateway and an industrial bus to realize real-time acquisition and preprocessing of physical entity layer data; the preprocessing includes outlier rejection, data interpolation and format conversion, the processed data is uploaded to the cloud through an encrypted transmission protocol, and key data is cached locally;

[0011] A data storage and management layer adopts a distributed database cluster to construct a lithium battery full life cycle data warehouse, the data warehouse divides data tables according to stages, configures a data backup mechanism and access permission management;

[0012] An intelligent analysis and decision layer integrates a machine learning algorithm library and a digital twinning simulation engine to realize quality anomaly detection, performance prediction and traceability analysis; the quality anomaly detection identifies anomalies by comparing real-time data with digital twinning model prediction values, the performance prediction predicts attenuation 30 cycles in advance based on an LSTM model, and the traceability analysis locates problems and generates reports by correlating the data warehouse;

[0013] A visualization and interaction layer develops a Web and mobile visual platform to provide digital twinning model three-dimensional display, full life cycle data trend chart and quality traceability path animation; a user interaction interface is configured for maintenance personnel to submit maintenance records and set permission verification mechanism;

[0014] A blockchain storage layer constructs a consortium chain network, stores key quality data on the chain, configures a smart contract to automatically execute data write permission verification and storage timestamp generation.

[0015] Further, it further includes a quality correlation degree analysis module, which identifies key influencing factors by calculating the correlation degree of each production parameter and quality index, and the correlation degree is calculated according to the formula , wherein R p is the correlation degree of the pth production parameter and the quality index, x pi is the pth production parameter value of the ith sample, is the sample average value of the pth production parameter, y i is the quality index value of the ith sample, is the sample average value of the quality index, n is the number of samples, through the correlation degree analysis, the key parameters with a correlation degree absolute value greater than or equal to 0.7 are preferentially monitored, and the quality anomaly detection efficiency is improved.

[0016] Further, it further includes a digital twin model dynamic updating module, which periodically updates the digital twin model parameters based on real-time acquisition data and physical entity state changes, and the model updating frequency is according to the formula is determined, wherein f u is the updating frequency, ΔM is the maximum deviation between the digital twin model prediction value and the physical entity measured value, M0 is the standard allowable deviation of the parameter, ΔP is the performance attenuation amount of the physical entity, P0 is the initial performance value, k1 and k2 are weight coefficients, and the updating frequency is dynamically adjusted.

[0017] Further, the physical entity layer further includes a retired battery recycling detection device, which integrates an internal resistance tester, a capacity tester and an appearance detection camera, collects the internal resistance, residual capacity and appearance damage degree data of the retired battery, and connects the recycling detection device with the data acquisition and transmission layer through an industrial Ethernet to upload the detection data in real time, associates the data with the unique ID of the retired battery, and provides data support for battery step utilization evaluation.

[0018] Further, the digital twin modeling layer further includes a supply chain digital twin sub-model, which integrates raw material supplier information, logistics transportation nodes and production plan scheduling, and constructs a full supply chain visualization model from raw materials to finished products; by inputting raw material procurement cycle and production rhythm, the influence of supply chain delay on production progress is simulated to provide simulation basis for supply chain optimization.

[0019] Further, the data acquisition and transmission layer further includes an edge computing node, which deploys a lightweight machine learning model to complete real-time data analysis locally, including vibration anomaly recognition, temperature sudden rise warning and equipment state evaluation; the edge computing node only uploads the analysis results and abnormal data to the cloud, while caching the original data for subsequent traceability analysis.

[0020] Further, the data storage and management layer is further configured with a data bloodline tracking module, which records the source, processing process and flow path of each data point to form a data bloodline map.

[0021] Further, the intelligent analysis and decision layer also contains a batch quality consistency evaluation module, which calculates the standard deviation and range of quality parameters of the same batch of lithium batteries to generate a consistency score; for batches with a consistency score < 70 points, the production process data is automatically associated, the key links leading to poor consistency are located through comparative analysis, and improvement suggestions are pushed to the production management terminal.

[0022] Further, the visualization and interaction layer also develops an AR assisted traceability function, which scans the battery shell two-dimensional code through the mobile terminal camera, triggers the AR interface to display the three-dimensional annotation of the key nodes in the full life cycle of the battery; dynamic arrow annotation is used in the module assembly stage to mark the cell stacking sequence, and the welding point temperature field distribution and insulation test results are superimposed and displayed; in the warehouse transportation stage, the storage environment temperature and humidity change curve and transportation vibration peak value are marked in the form of a time axis; in the terminal application stage, the current cycle number, capacity retention rate and historical fault record are displayed in real time, the AR interface supports double-finger zoom, single-finger rotation and sectioning operations, and is equipped with a voice interaction engine; for the annotated abnormal parameters, clicking automatically displays the abnormal diffusion path and associated batch influence range simulated by the digital twin model, and provides navigation to nearby maintenance points.

[0023] Further, it also includes a quality risk early warning module, which calculates the quality risk value of lithium batteries at each stage of the full life cycle to early warn potential quality problems, and the risk value is calculated according to the formula , wherein R s is the quality risk value of the s stage, a and β are weight coefficients, m is the number of quality parameters in the stage, w j is the weight of the jth parameter, v j is the measured value of the jth parameter, t j is the threshold value of the jth parameter, H s is the historical failure rate of the stage, when the risk value is greater than or equal to 80, the system automatically pushes the early warning information to the relevant person in charge, and the risk diffusion path simulated by the digital twin model is attached to assist in formulating preventive measures.

[0024] Compared with the existing technology, the beneficial effects of the present application are:

[0025] The system covers all links from cell production, module assembly, Pack assembly to warehouse transportation, terminal application and retirement recycling through multi-dimensional sensor deployment of the physical entity layer, breaks the "data island" through the edge gateway and encryption protocol of the data acquisition and transmission layer, realizes real-time acquisition and safe transmission of full life cycle data; the alliance chain network and smart contract design of the block chain storage layer ensure that the key quality data is tamper-proof and traceable, solve the problem of insufficient data credibility in traditional traceability, and when quality disputes occur, the responsible subject can be quickly located by retrieving the on-chain storage data.

[0026] In terms of intelligent analysis and quality control capabilities, the system performs outstandingly. The multi-scale model of the digital twin modeling layer can map the physical battery state in real time, combined with the machine learning algorithm of the intelligent analysis and decision layer, not only can realize real-time detection of quality abnormalities, but also can predict the battery performance degradation trend in advance, upgrade the traditional "after-tracing" to the whole-process quality control mode of "pre-warning + in-process control + after-tracing"; the addition of quality correlation analysis, batch consistency evaluation and other modules can accurately identify the key parameters and weak links affecting quality, provide data support for production process optimization, reduce the repeated occurrence of similar quality problems, and improve the overall production quality level.

[0027] The AR assisted tracing function of the visualization and interaction layer and the multi-terminal visualization platform greatly improve the convenience and intuitiveness of tracing operation. Field staff can quickly retrieve key information of the whole life cycle of the battery through mobile terminals, three-dimensional labeling and dynamic simulation functions can clearly restore the scene of quality problems, and voice interaction and navigation functions further reduce the operation threshold; at the same time, the system takes into account the data collection of retired battery recycling, provides basis for cascade utilization evaluation, promotes lithium battery resource recycling, and conforms to the concept of green manufacturing. Overall, the system realizes the intelligent, integrated and credible quality tracing of the whole life cycle of lithium batteries, which is of great significance to improve the quality control level of the industry, ensure the safety of the battery, and promote resource recycling. BRIEF DESCRIPTION OF DRAWINGS

[0028] Figure 1 The schematic diagram of the lithium battery whole life cycle quality tracing system based on digital twin proposed by the present application is shown in the figure;

[0029] Figure 2 The quality problem positioning time comparison column chart for different tracing systems is shown in the figure;

[0030] Figure 3 The data coverage rate comparison column chart for each link of the whole life cycle of lithium battery is shown in the figure;

[0031] Figure 4 The lithium battery field tracing operation time comparison line chart is shown in the figure. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0033] In the description of the present application, it should be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, which are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application.

[0034] In addition, the terms "first", "second" are only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited. In addition, the terms "mounting", "connecting", "connecting" should be broadly understood, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium, or it can be the communication between two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances, and the present application will be further described in detail below with reference to the drawings.

[0035] Referring to Figures 1 to 4 A lithium battery full life cycle quality traceability system based on digital twinning, comprising:

[0036] A physical entity layer includes lithium battery cell production equipment, module assembly line, Pack packaging equipment, warehouse environment monitoring device, transportation state acquisition equipment and terminal application sensor. The cell production equipment integrates a laser thickness gauge, an X-ray flaw detector and a pressure sensor, the laser thickness gauge has an accuracy of ±0.5 μm, the X-ray flaw detector has a resolution of 50 μm, and the pressure sensor has a range of 0-50 kN. Real-time acquisition of tab thickness, internal defects and roll pressure data. The module assembly line is configured with a visual positioning system and a torque wrench, the visual positioning system has a positioning error of ≤0.1 mm, and the torque wrench has an accuracy of ±1% FS. Record the cell stacking position and bolt tightening torque. The Pack packaging equipment is equipped with a barcode scanner and a sealing detector, the barcode scanner has a decoding speed of ≥300 times / s, and the sealing detector has a lower limit of leakage detection of 1×10 -6 Pa·m 3 / s, associate the package barcode with the sealing performance parameters. The warehouse environment monitoring device includes a temperature and humidity sensor and a gas detector. The temperature and humidity sensor measures a range of -40 to 85°C and 0 to 100% RH. The gas detector has a detection accuracy of ±5% LEL. The environment data is collected and stored. The transportation state acquisition equipment integrates a three-axis acceleration sensor and a GPS positioning module. The three-axis acceleration sensor has a range of ±16g. The GPS positioning module has a positioning accuracy of ≤5m. The transportation vibration and position information are recorded. The terminal application monitors the battery operating voltage, current, and surface temperature in real time. The voltage measurement range is 0 to 5V with an accuracy of ±0.5% FS. The current is 0 to 100A with an accuracy of ±1% FS. The surface temperature is -20 to 100°C with an accuracy of ±0.5°C.

[0037] The digital twin modeling layer constructs a multi-scale digital twin model of the lithium battery full life cycle, including a cell microstructure model, a module structure mechanics model, a Pack system thermal simulation model, and a full life cycle performance attenuation model. The cell microstructure model is based on a scanning electron microscope image with a resolution of 10 nm. The electrode material particle distribution with a particle diameter of 2 to 20 μm and the pore structure with a porosity of 20% to 40% are reconstructed. The lithium ion diffusion path is calculated by finite element analysis. The module structure mechanics model uses a tetrahedral mesh division with a unit size of 0.5 to 2 mm to simulate the stress deformation of the cell stack with a maximum deformation of ≤0.2 mm. The Pack system thermal simulation model includes fluid dynamics analysis of the heat dissipation channel with a flow rate of 0.5 to 2 m / s, and temperature field distribution calculation with a temperature difference control of ≤5°C. The full life cycle performance attenuation model integrates parameters such as cycle number 0 to 5000 times, charge and discharge rate 0.1C to 5C, and environmental temperature -20 to 60°C to predict the capacity retention rate, and takes 80% of the initial capacity as the retirement threshold.

[0038] The data acquisition and transmission layer deploys edge gateways and industrial buses. The edge gateway supports 5G / Ethernet dual-mode communication and has a data transmission rate of ≥100Mbps. The industrial bus PROFINET has a real-time performance of ≤1ms, realizing real-time data acquisition and preprocessing of the physical entity layer. The preprocessing includes outlier rejection based on the 3σ criterion (σ is the standard deviation of the data), cubic spline interpolation data with a time interval of 0.1 to 1s, and conversion to JSON format with information such as time stamp. The data is transmitted to the cloud through a protocol encrypted by AES-256 and with a key updated every hour. At the same time, the key data is cached locally, with a storage capacity of ≥1TB and a storage period of 3 months.

[0039] The data storage and management layer adopts a distributed database cluster (including a relational database MySQL to store structured data, a time series database InfluxDB to store time series data, and a non-relational database MongoDB to store unstructured data) to build a lithium battery full life cycle data warehouse. The data warehouse divides data tables by stages: a cell production table (including tab information, winding parameters, and packaging data, each record being associated with a unique cell ID), a module assembly table (recording cell combination methods, welding parameters, and insulation test results), a Pack grouping table (storing BMS parameter configurations, shell assembly data, and factory detection reports), a warehouse record table (associated with storage locations, environmental parameters, and warehouse entry and exit times), a transportation trajectory table (including GPS coordinates, vibration data, and transportation tool information), and a terminal application table (recording charge and discharge cycles, fault alarms, and maintenance records). A data backup mechanism (real-time incremental backup + daily full backup, with backup data saved for 10 years) and access permission management (based on an RBAC model, with different permissions for administrators, quality inspectors, and operation and maintenance personnel) are configured.

[0040] The intelligent analysis and decision-making layer integrates a machine learning algorithm library (including random forest, LSTM neural network, and reinforcement learning model) and a digital twin simulation engine to realize quality anomaly detection, performance prediction, and traceability analysis. The quality anomaly detection identifies anomalies such as tab thickness out-of-tolerance and virtual welding by comparing real-time collected data with predicted values of the digital twin model (with a deviation threshold of 0.5% to 5%, dynamically adjusted according to parameter importance). The performance prediction based on an LSTM model (with input features including cycle number, average charge and discharge rate, and environmental temperature, and output features including capacity retention rate) provides early warning of performance degradation trends 30 cycles in advance. The traceability analysis locates the stage (cell production / module assembly / Pack grouping / warehouse transportation / terminal application) and specific equipment where quality problems occur, and generates a traceability report (including abnormal parameters, associated batches, and impact scope).

[0041] The visualization and interaction layer develops a web-based and mobile-based visualization platform that supports responsive design and adapts to resolutions from 1366x768 to 3840x2160, provides three-dimensional display of digital twin models (supporting zooming, rotating, and slicing operations with a frame rate of ≥30 fps), full life cycle data trend charts (including line charts, column charts, and heat maps), and quality traceability path animations. The platform also configures user interaction interfaces that support quality inspectors to initiate anomaly labeling containing abnormal types, operation and maintenance personnel to submit maintenance records containing maintenance measures, and management personnel to view quality statistical reports containing batch qualification rates, and sets a permission verification mechanism (username and password + verification code login, with a session timeout time of 30 minutes).

[0042] The blockchain storage layer constructs a consortium chain network, the nodes of the consortium chain network include the battery cell factory, the module factory, the pack factory, the logistics company and the vehicle enterprise, the consensus mechanism adopts PBFT, the block time is less than or equal to 3 seconds, and the key quality data is stored in the chain through the network.

[0043] The key data specifically includes a battery cell factory detection report, a module welding parameter, a pack factory qualified certificate, a transportation handover record and terminal fault information: the battery cell factory detection report is stored by using a hash value, the original data address is recorded on the chain; the module welding parameter is stored by combining a time stamp, an equipment ID and a parameter value; the pack factory qualified certificate is verified by a digital signature; the transportation handover record needs a digital signature of both parties; and the terminal fault information is associated with a battery unique ID.

[0044] Meanwhile, a smart contract is configured, the smart contract is written by using a Solidity language, can automatically execute data write permission verification and storage timestamp generation, the data write permission verification verifies the validity of a node certificate, the storage timestamp generation is accurate to milliseconds, so that the data is ensured to be non-tamperable and traceable.

[0045] In the application, a quality correlation degree analysis module is further included, the correlation degree of each production parameter and a quality index is calculated to identify key influencing factors. p = is calculated, wherein R p is the correlation degree of the pth production parameter and the quality index (the value range is -1-1, and the greater the absolute value is, the stronger the correlation is), x pi is the pth production parameter value (such as a pole piece rolling pressure and a welding current) of the ith sample, is the sample average value of the pth production parameter, y i is the quality index value (such as a capacity deviation rate and a cycle life) of the ith sample, is the sample average value of the quality index, and n is the sample quantity (n is greater than or equal to 30). Through the correlation degree analysis, the key parameters with the absolute value of the correlation degree greater than or equal to 0.7 can be preferentially monitored, and the pertinence and efficiency of quality anomaly detection are improved.

[0046] In the application, a digital twin model dynamic updating module is further included, the digital twin model parameters are periodically updated based on real-time acquisition data and physical entity state changes. The model updating frequency is determined according to the formula , wherein f uFor the update frequency (unit: times / hour, value range 0.1-10), ΔM is the maximum deviation of the predicted value of the digital twin model and the measured value of the physical entity, M0 is the standard allowable deviation of the parameter, ΔP is the performance attenuation of the physical entity, P0 is the initial performance value, k1, k2 are weight coefficients (k1+k2=1, default k1=0.6, k2=0.4). By dynamically adjusting the update frequency, the model accuracy (prediction deviation ≤3%) is ensured while reducing the calculation resource consumption (40%-60% less computing power than fixed frequency update).

[0047] In the present application, the physical entity layer further comprises a retired battery recycling detection device, which integrates an internal resistance tester, a capacity tester and an appearance detection camera. The internal resistance tester has a measurement range of 0-200 mΩ and an accuracy of ±1%, the capacity tester has a charge and discharge accuracy of ±0.5% FS, and the appearance detection camera has a resolution of 5 million pixels. The internal resistance, residual capacity and appearance damage degree data of the retired battery are collected through these devices. The recycling detection device is connected with the data acquisition and transmission layer through industrial Ethernet, the communication delay is ≤50 ms, the detection data can be uploaded in real time, and the data is associated with the unique ID of the retired battery. The ID is bound with the original ID in the whole life cycle data warehouse, which provides data support for battery step utilization evaluation, such as the capacity threshold of energy storage power station reuse needs to be ≥60% of the initial capacity.

[0048] In the present application, the digital twin modeling layer further comprises a supply chain digital twin sub-model, which integrates raw material supplier information, logistics transportation nodes and production plan scheduling to build a full supply chain visualization model from raw materials to finished products. The raw material supplier information includes positive material purity ≥99.5%, separator thickness 12-25 μm; the logistics transportation nodes include warehouse center coordinates and transportation route; the production plan scheduling includes daily production capacity and equipment utilization rate ≥85%. By inputting the raw material procurement period of 3-7 days and the production rhythm of 10-30 seconds / cell, the influence of supply chain delay on production progress (such as the production capacity loss rate caused by 1-day delay) is simulated, which provides simulation basis for supply chain optimization such as safety stock setting and logistics route adjustment.

[0049] In the application, the data acquisition and transmission layer also contains an edge computing node, the node computing power is greater than or equal to 10 TOPS, supports GPU acceleration, deploys a lightweight machine learning model, the model size is less than or equal to 500 MB, and the real-time data can be quickly analyzed locally, and the response time is less than or equal to 100 ms. The analysis content includes vibration anomaly identification, temperature sudden rise warning and equipment state evaluation: the vibration anomaly identification is based on the fast Fourier transform of the acceleration signal, the characteristic frequency range is 1-1000 Hz; the temperature sudden rise warning is triggered when the temperature change rate is greater than or equal to 5 DEG C / min; and the equipment state evaluation judges the bearing wear degree through the vibration spectrum characteristics. The edge computing node only uploads the analysis results and abnormal data to the cloud, and the data transmission amount is reduced by 60%-80% compared with full amount uploading, and the original data is cached, and the saving period is 24 hours, for subsequent traceability analysis and calling.

[0050] In the application, the data storage and management layer is also configured with a data bloodline tracking module, which records the source, processing process and flow path of each data point to form a data bloodline map. The source includes equipment number and sensor ID; the processing process includes interpolation algorithm and outlier elimination mark; the flow path includes transmission time from the edge node to the cloud database; and the data bloodline map is stored in the form of a directed graph, the node is a data processing link, and the edge is a data flow direction. By querying the data bloodline map, the original collection equipment and processing method (such as whether smoothing processing is performed) of any quality data can be traced, the traceability distortion problem caused by data tampering or misprocessing is solved, and the accuracy is greater than or equal to 99.9%.

[0051] In the application, the intelligent analysis and decision layer also contains a batch quality consistency evaluation module, which calculates the standard deviation and range of quality parameters of the same batch of lithium batteries to generate a consistency score. In the standard deviation of the quality parameters, the capacity standard deviation is less than or equal to 50 mAh, and the resistance standard deviation is less than or equal to 5 m omega; the range is the difference between the maximum value and the minimum value; and the consistency score is full score 100 points, of which greater than or equal to 85 points is a high-quality batch. For the batch with a consistency score less than 70 points, the module automatically associates the production process data, such as the equipment parameter fluctuation in the same time period, locates the key link causing poor consistency, such as uneven electrode coating and unstable welding parameters, and pushes the improvement suggestions, such as adjusting the coating speed and calibrating the welding current, to the production management terminal.

[0052] In the application, the visualization and interaction layer also develops an AR assisted traceability function, a mobile terminal camera scans a battery shell two-dimensional code (recognition distance 0.5-2 m, recognition response time less than or equal to 300 ms), triggers an AR interface to display three-dimensional labels of key nodes in the whole life cycle of the battery; and the label information is displayed in layers according to stages: the electrode thickness average (such as 85 mu m plus or minus 2 mu m), the winding tension (30 N plus or minus 2 N) and the packaging air tightness (less than or equal to 1*10 -7 Pa·m3 The AR interface supports double-finger zooming (zooming range 1:0.5-1:5), single-finger rotation (360° no dead angle viewing), and sectioning operation (internal cell arrangement can be viewed), and is equipped with a voice interaction engine (supports more than 20 instructions, such as "display the last 3 times of charging and discharging data" and "locate the welding abnormal position", and the recognition accuracy is ≥95%). For the marked abnormal parameters (such as the resistance of a certain cell exceeding the standard), clicking automatically displays the abnormal diffusion path simulated by the digital twin model (such as the time estimate of 2 hours from cell A3 to the module) and the associated batch influence range (such as 15 of the 120 batteries in the same batch), and provides nearby maintenance point navigation (based on GPS positioning, error ≤10m), greatly improving the intuitiveness and operation efficiency of on-site tracing.

[0053] In the present application, a quality risk early warning module is also included, which calculates the quality risk value of lithium batteries at each stage of the whole life cycle to early warn potential quality problems. The risk value is calculated according to the formula wherein R s is the quality risk value of the s stage (≥0, ≥80 is high risk), α and β are weight coefficients (α+β=1, default α=0.7, β=0.3), m is the number of quality parameters in the stage, w j is the weight of the jth parameter (∑w j =1, the weight of a key parameter such as the separator thickness is ≥0.2), v j is the measured value of the jth parameter, t j is the threshold value of the jth parameter (such as the lower limit of the separator thickness 12μm), and H s is the historical failure rate of the stage (the number of failures in the same stage in the last 3 months / total sample number). When the risk value is ≥80, the system automatically pushes the early warning information to the relevant person in charge (through SMS, APP push, response delay ≤5 minutes), and attaches the risk diffusion path simulated by the digital twin model to assist in formulating preventive measures.

[0054] The specific implementation of the system is further illustrated by two embodiments as follows:

[0055] Embodiment 1: New energy vehicle power battery full life cycle quality traceability (ternary lithium battery for passenger cars)

[0056] This embodiment is aimed at ternary lithium battery for passenger cars (capacity 75 kWh, number of cells 288, number of modules 12, Pack group) of a vehicle enterprise, and a full life cycle quality traceability system based on digital twinning is constructed, covering the complete process from cell production to terminal retirement.

[0057] Physical entity layer configuration: The cell production equipment selects the pole piece laser thickness gauge (model LS-300, precision ±0.5μm, sampling frequency 10Hz), X-ray flaw detector (model XTH5000, resolution 50μm, detection speed 2 pieces / min), real-time acquisition of positive pole piece thickness (standard 85μm±2μm) and internal pinhole defects (allowance ≤1 / ㎡); The roller pressing equipment integrates the pressure sensor (model CY-303, range 0-50kN, precision ±0.1kN), records the roller pressing pressure (standard 25kN±1kN). The module assembly line uses a visual positioning system (model VS-600, positioning error ≤0.1mm) to guide the mechanical arm to stack the cells, and a torque wrench (model TW-800, precision ±1%FS) to record the bolt fastening torque (standard 8N·m±0.2N·m). The Pack packaging equipment is equipped with a bar code scanner (model SC-900, decoding speed 350 times / s), which is associated with the Pack unique code (18-bit code, including production date, batch information) and the sealing detector (model LT-500, leakage detection lower limit 1×10 -6 Pa·m 3 / s) data (standard leakage rate ≤5×10 - 7 Pa·m 3The warehouse environment monitoring device selects a temperature and humidity sensor (model TH-400, -40~85°C, 0~100% RH, accuracy ±0.5°C / ±3% RH) and a hydrogen detector (model GD-200, detection accuracy ±5% LEL), and collects data every 10 seconds (storage standard 25°C±2°C, RH 40%~60%, hydrogen concentration≤1% LEL). The transportation state acquisition equipment integrates a three-axis acceleration sensor (model AD-600, range ±16g, sampling frequency 50Hz) and a GPS module (model GP-800, positioning accuracy≤5m), records the transportation vibration (standard≤3g) and the position information. The terminal application sensor (model BT-700) monitors the battery voltage (0~5V, accuracy ±0.5% FS), current (0~100A, accuracy ±1% FS) and surface temperature (-20~100°C, accuracy ±0.5°C) in real time, and uploads data every 1 second.

[0058] The digital twin modeling layer constructs a multi-scale model: the microstructure model of the battery cell is based on scanning electron microscope (SEM, model SU8020, resolution 10nm) images, reconstructs the distribution of NCM811 positive material particles (diameter 5~15μm) and graphite negative particles (diameter 2~8μm), porosity 30%±5%, and calculates the lithium ion diffusion coefficient (standard 1×10 -10 ~5×10 -10 m 2 / s) by COMSOL Multiphysics; the module structure mechanics model uses ANSYS to divide tetrahedral mesh (element size 1mm), simulates the stress of battery cell stacking (maximum deformation≤0.2mm); the Pack system thermal simulation model analyzes the flow velocity (1.2m / s±0.2m / s) of the heat dissipation channel by Fluent, and controls the temperature field temperature difference within 3°C; the full life cycle performance attenuation model integrates the cycle number (0~3000 times), the charge and discharge rate (0.5C~2C) and the environmental temperature (-10~45°C), and predicts the capacity retention rate (80% of the initial capacity is the retirement threshold).

[0059] The data acquisition and transmission layer deploys an edge gateway (model EG-500, supports 5G / Ethernet, transmission rate 150Mbps) and a PROFINET industrial bus (real-time≤1ms), the preprocessing uses 3σ criterion to eliminate outliers (such as the thickness of the pole piece exceeding 83~87μm is marked as abnormal), cubic spline interpolation (time interval 0.5s) and JSON format conversion (including timestamp, device ID, parameter name and value), and uploads to the cloud through AES-256 encryption, and local cache (1TB SSD, saves for 3 months).

[0060] The data storage and management layer uses MySQL (stores the production table of the battery cell, including the thickness of the pole piece and the winding tension of 30N±2N), InfluxDB (stores time series data such as charging and discharging current), and MongoDB (stores SEM images and detection report PDF). The data backup is real-time incremental + daily full (saved for 10 years), and the RBAC permission management distinguishes between administrators (full permission), quality inspection personnel (data query / annotation), and maintenance personnel (maintenance record).

[0061] The intelligent analysis and decision-making layer integrates a random forest model and an LSTM model. The random forest model is used for anomaly detection with a deviation threshold of 2%. The LSTM model is used for performance prediction, with input cycle times, rate, and temperature, and output capacity retention rate. Traceability analysis can locate the problem stage, such as rapid capacity attenuation of a certain battery, which can be traced back to the production stage of the pole piece thickness of the battery cell. The visualization and interaction layer supports 3840x2160 resolution on the Web, displays a three-dimensional model with a frame rate of 35fps, and supports anomaly annotation on mobile devices. The blockchain storage layer constructs a consortium chain with nodes including battery cell factories, vehicle manufacturers, and logistics companies, adopts a PBFT consensus mechanism, and has a block time of 2.5 seconds. The chained data includes the hash value of the battery detection report and the digital signature of the Pack qualification certificate.

[0062] The quality correlation analysis module calculates the parameter correlation, such as the correlation between the thickness of the pole piece (x p ) and the capacity deviation rate (y), n = 50 samples, Substitute the formula R p = 0.82 (absolute value ≥ 0.7), which is listed as a key monitoring parameter. The dynamic update frequency of the digital twin model is calculated as ΔM = 0.03V (voltage deviation), M0 = 0.05V, ΔP = 2kWh (capacity attenuation), P0 = 75kWh, f u = 0.6x(0.03 / 0.05) + 0.4x(2 / 75) ≈ 0.38 times / hour (approximately 2.6 hours of update). The quality risk warning module is calculated as m = 4 (thickness of the pole piece, roll pressure, sealing rate, vibration value), w j = 0.3, 0.25, 0.25, 0.2, respectively, v j is the measured value, t j is the threshold value, H s = 0.02 (historical failure rate), and R s = 75 (<80, medium risk) for a certain batch, which pushes the warning to the production department.

[0063] Example 2: Quality traceability of lithium iron phosphate batteries in the whole life cycle of energy storage power stations (1MWh energy storage cabinet)

[0064] This embodiment is aimed at a certain energy storage power station 1MWh lithium iron phosphate battery energy storage cabinet (single cell capacity 280Ah, cell number 3600, module 300, Pack 10), deploying a full life cycle quality traceability system, focusing on performance degradation and safety traceability in long-term energy storage applications.

[0065] The physical entity layer supplements the retired battery recycling detection equipment: internal resistance tester (model IR-300, 0-200mΩ, accuracy ±1%), capacity tester (model CA-600, charge and discharge accuracy ±0.5%FS), appearance detection camera (model VC-800, 5 million pixels), collects the internal resistance (standard ≤100mΩ), residual capacity (step utilization threshold ≥60% of initial capacity) and appearance damage (such as shell rupture, pole corrosion) of the retired battery, uploads data through industrial Ethernet, and associates the retired battery ID (bound with the original production ID).

[0066] The digital twin modeling layer adds a supply chain digital twin sub-model: integrates positive material (lithium iron phosphate purity 99.8%) and diaphragm (thickness 16μm±1μm) supplier information, logistics transportation nodes (warehouse center coordinates N30°, E120°, transportation route highway priority), production planning scheduling (daily capacity 500 cells, equipment utilization rate 88%), through AnyLogic simulation raw material procurement cycle (5 days) and production rhythm (20 seconds / cell), simulates that the delay of 1 day causes the capacity loss rate of 8%, optimizes the safety stock (increases the usage of 3 days).

[0067] The data acquisition and transmission layer deploys edge computing nodes (model EC-800, computing power 15TOPS, GPU acceleration), deploys lightweight CNN models (size 450MB), analyzes vibration abnormalities (FFT transform, characteristic frequency 10-500Hz, identifies bearing wear), temperature surge (change rate ≥5℃ / min triggers early warning) locally, uploads only abnormal data (reduces 70% transmission volume compared to full volume upload), and locally caches original data for 24 hours.

[0068] Data storage and management layer configuration data bloodline tracking module: records data source (such as cell capacity data from device CA-600, ID = C508), processing process (cubic spline interpolation, outlier rejection mark), flow path (edge node → cloud time-consuming 0.8 seconds), form directed graph bloodline atlas, query accuracy 99.95%. Intelligent analysis and decision layer batch quality consistency evaluation module calculates the capacity standard deviation of 300 modules (≤30 mAh), resistance standard deviation (≤3 mΩ), consistency score 92 points (high-quality batch), for another score 68 points batch, associated with the discovery of the coating speed fluctuation of the pole piece (2.5-3.5 m / min), push adjustment suggestion (stabilize at 3 m / min ± 0.1 m / min).

[0069] Visual and interactive layer AR assisted traceability function: mobile terminal (mobile phone / tablet) scans Pack two-dimensional code (recognition distance 0.5-2 m, response time 250 ms), AR interface layered display: cell production stage marks the thickness of the pole piece 88 μm ± 1 μm, winding tension 28 N ± 1 N; module assembly stage dynamic arrow shows the cell stacking order (A1-A12 clockwise arrangement), temperature field red area marks the welding over-temperature 230℃±3℃(qualified); Pack stage displays BMS overcharge protection voltage 3.65V±0.02V; warehouse transportation stage time axis marks the temperature and humidity curve (22℃±1℃, 55%RH±3%), vibration peak 2.5g; terminal application stage displays the current cycle number 2500 times, capacity retention rate 82%, historical failure (overcurrent alarm on March 10, 2024, processing measures: restart BMS). AR supports double-finger zoom (1:0.5-1:5), 360° rotation, voice command "view recent charge and discharge data" recognition accuracy 96%, abnormal parameter click to display simulation diffusion path and nearby maintenance point navigation.

[0070] Performance comparison table 1: traceability efficiency and data integrity

[0071]

[0072] Table 1

[0073] Table 1 data shows that the system has significant advantages in traceability efficiency and data reliability. Traditional traceability relies on manual query of scattered systems, and it takes 8-12 hours to locate the problem. The data only covers the production link (60-70%), and there is a 25-30% risk of tampering. The system uses digital twin modeling and blockchain storage to shorten the positioning time to 1-1.5 hours. The data covers the whole link from production to retirement (more than 98%), and the on-chain data tampering risk is 0. This is due to real-time data collection and encrypted transmission in multiple links, as well as the tamper-proof nature of blockchain, ensuring quick problem location and authoritative and reliable traceability results, especially for passenger cars and energy storage scenarios that require high safety.

[0074] Performance comparison table 2: intelligent analysis and operational convenience

[0075]

[0076] Table 2

[0077] Table 2 data reflects the value of system intelligent analysis and interaction design. The traditional system has no prediction and early warning function, and it takes 15-20 minutes per site to trace. The system uses LSTM model to achieve more than 92% performance prediction accuracy, and the abnormal early warning response time is 3-5 minutes. AR assisted traceability reduces the on-site operation time to 1.5-3 minutes. This is due to the real-time simulation capability of digital twin model and the intuitive interaction of AR, which can avoid quality risks in advance, reduce the difficulty of on-site operator operation, improve the efficiency of energy storage power station long-term operation and passenger car after-sales traceability, and reduce economic losses caused by failures.

[0078] Referring to Figure 2 The figure clearly shows the significant advantage of the system in the efficiency of locating quality problems. Manual record tracing relies on paper documents and manual checking, and the information is fragmented, resulting in a positioning time of up to 11.5 hours. Traditional bar code tracing can associate some data, but it still takes 8.2 hours to switch between multiple systems for query. The industry average intelligent traceability cannot cover the whole life cycle due to incomplete data integration, and it takes 4.8 hours. The system uses digital twin model to map the whole life cycle state of lithium battery in real time, and combines with the trusted data of blockchain storage to quickly associate abnormal parameters and corresponding links. For example, in Example 1, the thickness of the pole piece is directly located to the battery production equipment, and the positioning time is only 1.2-1.4 hours, which is more than 80% shorter than the traditional method, greatly reducing the time cost of quality problem troubleshooting, and saving valuable time for after-sales maintenance and production improvement.

[0079] Referring to Figure 3The figure verifies the effectiveness of the "full-link data penetration" design of the system. Due to the lack of special sensors for the storage and transportation links and the detection equipment for the retired batteries, the data coverage of the two links is only 45% and 10%, and the terminal application link only records fault information, with a data coverage of 60%, which cannot completely restore the full scene of the quality problem. Through the multi-dimensional sensor deployment of the physical entity layer, such as the vibration sensor in the transportation link and the internal resistance tester in the retirement link in embodiment 2, combined with the edge gateway of the data acquisition layer, the system realizes real-time acquisition and transmission of data in each link, and the data coverage of each link reaches more than 95%, among which the data coverage of the battery cell production, module assembly and terminal application links reaches 99%. Complete data coverage ensures that all parameters in each stage can be fully retrieved during traceability analysis, for example, when the capacity of a certain energy storage battery decays, the porosity of the pole piece during production, the vibration intensity during transportation and the charge-discharge rate during application can be analyzed at the same time, avoiding the broken chain of traceability caused by data loss.

[0080] Referring to Figure 4 The figure directly reflects the improvement of the AR assisted traceability function of the system on the efficiency of the on-site operation. The traditional on-site traceability requires the staff to carry a tablet computer and switch between multiple APPs to query, such as logging into the MES system to view production parameters and logging into the after-sales system to query fault records, which is cumbersome and time-consuming, with an average time consumption of 9.8 minutes per battery. The AR assisted traceability function of the system directly superimposes three-dimensional labels on the physical battery by scanning the two-dimensional code on the battery shell, such as the welding point temperature field and the cell stacking sequence of the energy storage battery in embodiment 2, and supports voice instructions to quickly retrieve data, such as directly displaying the performance trend through the voice instruction "view the latest charge-discharge curve". The time consumption of each operation is reduced to 1.6-2.5 minutes, and the average time consumption is only 2.0 minutes, which is 80% shorter than the traditional way. This advantage greatly reduces the operation threshold of on-site staff, especially suitable for the rapid traceability and troubleshooting of a large number of batteries in energy storage power stations and other scenarios, significantly improving the operation and maintenance efficiency.

[0081] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can make equivalent replacements or changes within the technical scope disclosed by the present application according to the technical solution and inventive concept of the present application, which should be covered within the protection scope of the present application.

Claims

1. A lithium battery full life cycle quality traceability system based on digital twins, characterized in that, include: The physical entity layer includes lithium battery cell production equipment, module assembly lines, pack packaging equipment, warehouse environment monitoring devices, transportation status acquisition equipment, and terminal application sensors. The digital twin modeling layer constructs a multi-scale digital twin model of the entire life cycle of lithium batteries, including cell microstructure model, module structural mechanics model, pack system thermal simulation model, and full life cycle performance degradation model. The data acquisition and transmission layer deploys edge gateways and industrial buses to achieve real-time acquisition and preprocessing of physical entity layer data. The preprocessing includes outlier removal, data interpolation, and format conversion. The processed data is uploaded to the cloud through an encrypted transmission protocol, while key data is cached locally. The data storage and management layer adopts a distributed database cluster to build a lithium battery full life cycle data warehouse. The data warehouse is divided into data tables according to stages and configured with data backup mechanism and access permission management. The intelligent analysis and decision-making layer integrates a machine learning algorithm library and a digital twin simulation engine to achieve quality anomaly detection, performance prediction, and traceability analysis. The quality anomaly detection identifies anomalies by comparing real-time data with the predicted values ​​of the digital twin model. The performance prediction is based on an LSTM model that provides an early warning of attenuation 30 cycles in advance. The traceability analysis is linked to the data warehouse to locate problems and generate reports. The visualization and interaction layer develops web and mobile visualization platforms, providing 3D displays of digital twin models, full lifecycle data trend charts, and quality traceability path animations; it also configures user interaction interfaces, allowing maintenance personnel to submit maintenance records and setting up permission verification mechanisms. The blockchain evidence storage layer constructs a consortium blockchain network to store key quality data on the blockchain, configures smart contracts, and automatically executes data write permission verification and evidence storage timestamp generation.

2. The lithium battery full life cycle quality traceability system based on digital twins according to claim 1, characterized in that, It also includes a quality correlation analysis module, which identifies key influencing factors by calculating the correlation between various production parameters and quality indicators. The correlation is calculated using the formula... Calculate, where R p Let x be the correlation degree between the p-th production parameter and the quality index. pi Let p be the production parameter value of the i-th sample. Let y be the sample average of the p-th production parameter. i Let be the quality index value of the i-th sample. Here, n is the sample average of the quality indicators and n is the number of samples. Through this correlation analysis, key parameters with an absolute correlation value ≥ 0.7 are monitored first to improve the efficiency of quality anomaly detection.

3. The lithium battery full life cycle quality traceability system based on digital twins according to claim 1, characterized in that, It also includes a digital twin model dynamic update module, which periodically updates the digital twin model parameters based on real-time collected data and changes in the physical entity's state. The model update frequency is calculated according to the formula. Determined, where f u The update frequency is defined as follows: ΔM is the maximum deviation between the predicted value of the digital twin model and the measured value of the physical entity; M0 is the standard allowable deviation of this parameter; ΔP is the performance degradation of the physical entity; P0 is the initial performance value; and k1 and k2 are weighting coefficients. The update frequency is dynamically adjusted.

4. The lithium battery full life cycle quality traceability system based on digital twin as described in claim 1, characterized in that, The physical entity layer also includes a retired battery recycling and testing device. This device integrates an internal resistance tester, a capacity tester, and an appearance inspection camera to collect data on the internal resistance, remaining capacity, and degree of appearance damage of retired batteries. The recycling and testing device is connected to the data acquisition and transmission layer via an industrial Ethernet to upload the test data in real time. The data is associated with the unique ID of the retired battery, providing data support for the evaluation of battery cascade utilization.

5. The lithium battery full life cycle quality traceability system based on digital twins according to claim 1, characterized in that, The digital twin modeling layer also includes a supply chain digital twin sub-model, which integrates raw material supplier information, logistics and transportation nodes, and production planning and scheduling to construct a full supply chain visualization model from raw materials to finished products. By inputting the raw material procurement cycle and production takt time, it simulates the impact of supply chain delays on production progress and provides simulation basis for supply chain optimization.

6. The lithium battery full life cycle quality traceability system based on digital twin as described in claim 1, characterized in that, The data acquisition and transmission layer also includes edge computing nodes, which deploy lightweight machine learning models to perform real-time data analysis locally, including vibration anomaly identification, temperature surge warning, and equipment status assessment. The edge computing nodes only upload the analysis results and abnormal data to the cloud, while caching the original data for subsequent traceability analysis.

7. The lithium battery full life cycle quality traceability system based on digital twin as described in claim 1, characterized in that, The data storage and management layer is also equipped with a data lineage tracking module, which records the source of each data point, the processing process, and the flow path to form a data lineage map.

8. The lithium battery full life cycle quality traceability system based on digital twins according to claim 1, characterized in that, The intelligent analysis and decision-making layer also includes a batch quality consistency assessment module, which calculates the standard deviation and range of quality parameters of the same batch of lithium batteries and generates a consistency score. For batches with a consistency score of <70, the system automatically links the production process data, identifies the key links causing poor consistency through comparative analysis, and pushes improvement suggestions to the production management terminal.

9. The lithium battery full life cycle quality traceability system based on digital twin as described in claim 1, characterized in that, The visualization and interaction layer also develops an AR-assisted traceability function, which uses a mobile terminal camera to scan the QR code on the battery casing to trigger an AR interface that displays 3D annotations of key nodes throughout the battery's life cycle. During the module assembly stage, dynamic arrows are used to mark the cell stacking order, and the temperature field distribution and insulation test results at the welding points are superimposed and displayed. During the warehousing and transportation phase, the temperature and humidity change curves of the storage environment and the peak value of transportation vibration are marked in the form of a time axis; during the terminal application phase, the current number of cycles, capacity retention rate and historical fault records are displayed in real time. The AR interface supports two-finger zoom, single-finger rotation and segmentation operations, and is equipped with a voice interaction engine. Clicking on the marked abnormal parameters will automatically display the abnormal propagation path of the digital twin model simulation and the impact range of related batches, and provide navigation to nearby maintenance points.

10. The lithium battery full life cycle quality traceability system based on digital twin as described in claim 1, characterized in that, It also includes a quality risk early warning module, which calculates the quality risk value of lithium batteries at each stage of their entire life cycle and provides early warnings of potential quality problems. The risk value is calculated using a formula. Calculate, where R s Let w be the quality risk value for stage s, α and β be weighting coefficients, m be the number of quality parameters for that stage, and w be the quality risk value for stage s. j Let v be the weight of the j-th parameter. j Let t be the measured value of the j-th parameter. j H is the threshold for the j-th parameter. s The system calculates the historical failure rate for this phase. When the risk value is ≥80, the system automatically pushes an early warning message to the relevant responsible persons, along with a risk diffusion path simulated by a digital twin model, to assist in the formulation of preventive measures.

Images