Food cold chain warehouse digitized tracing system based on internet of things

Abstract

The application discloses a food cold chain storage digital tracing system based on an Internet of Things, and relates to the field of the Internet of Things and digital twin technology. The system performs heat diffusion calculation according to real-time data of a temperature sensor, generates three-dimensional heat diffusion boundary grid data evolving with time, generates a four-dimensional bounding box containing space and time boundaries in a three-dimensional digital twin scene, judges and filters affected tracing codes by intersection with tray coordinates and storage time in a tracing database, and highlights the affected tracing codes, supports dragging of a damaged batch of goods icon to a standby cold storage icon, calculates and displays temperature recovery curves and shelf life loss prediction results corresponding to multiple schemes in real time during the hovering process, and writes this event data into a blockchain account for storage after a user confirms the scheme, and the system sends transfer parameters to a physical scheduling system to drive an automated guided vehicle to execute the transfer.

Description

Technical Field

[0001] This invention relates to the fields of Internet of Things (IoT) and digital twin technology, specifically to a digital traceability system for cold chain food storage based on IoT. Background Technology

[0002] Food cold chain warehousing IoT traceability systems typically deploy IoT sensing devices such as temperature and humidity sensors and RFID tags on shelves, pallets, and outer packaging of goods. The collected data is uploaded to the back-end warehouse management system through a wireless gateway, and blockchain technology is used to store evidence of key operation nodes such as warehousing, outbound, and abnormal handling.

[0003] Existing advanced cold chain warehousing traceability systems generally include: using an IoT sensor array to collect temperature and humidity data from various areas of the warehouse in real time and uploading it to a backend server;

[0004] The backend server compares the collected data with a preset security threshold, and triggers an alarm event when the data exceeds the threshold range.

[0005] The front-end graphical user interface marks alarm markers on the corresponding shelf or pallet icons of the alarm sensors on the static three-dimensional warehouse model or two-dimensional plan view;

[0006] Managers can view batch traceability codes and historical data for a given shelf or pallet by clicking the alarm icon or filtering through the two-dimensional list.

[0007] Based on experience, managers judge the scope of the impact of the anomaly and manually create cargo transfer work orders through an independent warehouse management system, assigning manual labor or scheduling automated guided vehicles to complete the transfer.

[0008] The above-mentioned solutions typically include an IoT sensing layer, a network transmission layer, a data platform layer, and a human-computer interaction display layer. Data is transmitted sequentially and unidirectionally between each layer. The display layer only serves as the final information presentation outlet and does not have the ability to drive business scheduling downwards.

[0009] Its graphical user interface interaction method only supports clicking to select regular shelf or pallet icons. When the refrigeration failure causes the loss of cold air and presents an irregular three-dimensional thermal diffusion pattern, the user cannot directly select the spatial area that matches the actual thermal wave range on the interface. The user can only rely on manual investigation and comparison, which is time-consuming and has the risk of misjudgment.

[0010] Meanwhile, the presentation layer is independent of the scheduling system and the evidence storage system, lacking a direct linkage interface. The results of interface operations cannot be automatically converted into physical execution instructions, resulting in a broken response chain and poor timeliness. Summary of the Invention

[0011] The technical problem to be solved by the present invention is that the existing graphical user interface interaction method of the Internet of Things traceability system for cold chain food storage can only provide two-dimensional list query or static three-dimensional model superimposed regular alarm points. Users cannot directly perform selection operations in the irregular dynamic three-dimensional spatiotemporal range affected by heat waves, making it difficult to quickly and accurately locate the affected pallets and batch traceability codes.

[0012] Meanwhile, the existing system only has the function of post-event recording. Administrators cannot intuitively simulate and compare different emergency transfer plans through the interface when a failure occurs, and cannot foresee the consequences of temperature rise and shelf-life loss.

[0013] Furthermore, the virtual operations on the interface are disconnected from the actual scheduling, traceability, and evidence storage in the physical world, making it impossible to form a closed loop from human-computer interaction to physical execution.

[0014] To address the aforementioned technical problems, this invention provides a digital traceability method and corresponding system for food cold chain warehousing based on the Internet of Things. The technical solution includes the following steps: receiving storage environment temperature data collected by a temperature sensor array; determining the trigger point of a local refrigeration fault based on a preset safety threshold; performing heat diffusion calculations using the trigger point as the initial condition to obtain three-dimensional temperature distribution data that changes over time; and extracting the isosurface of the three-dimensional temperature distribution data according to a preset temperature threshold to generate three-dimensional thermal diffusion boundary grid data that evolves over time.

[0015] In a 3D digital twin scene, render the 3D thermal diffusion boundary mesh data, receive the time parameters specified by the user through the time axis slider and the operation parameters for dragging and trimming the mesh boundary, and generate 4D bounding box boundary data containing the 3D spatial boundary and the time boundary.

[0016] The boundary data of the four-dimensional bounding box is intersected with the three-dimensional spatial coordinates and entry timestamps of each pallet recorded in the traceability database. The set of traceability codes that fall within the range is then selected and the corresponding pallet icon is highlighted.

[0017] The system receives user drag-and-drop operations on the icons of damaged batches of goods. When the drag is hovered over the backup cold storage icon, the system obtains the corresponding ambient temperature parameters and the expected transfer time, calculates the temperature recovery prediction data and the shelf life loss prediction data and displays them visually. It supports parallel simulation and comparison of multiple backup cold storage icons during the same drag-and-drop process.

[0018] After receiving the user's confirmation of the plan, the transfer plan parameters are sent to the physical scheduling system to drive the automated guided vehicle to perform the transfer, and the relevant data of this abnormal event is written into the blockchain ledger for evidence storage.

[0019] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows:

[0020] 1. This invention provides an interactive mechanism that can dynamically generate irregular four-dimensional selection areas in a three-dimensional digital twin scene based on real-time thermal diffusion data, thereby achieving the technical effect of quickly and accurately locating the scope of abnormal impact;

[0021] 2. This invention provides an interactive control that supports drag-and-drop sand table simulation, enabling administrators to intuitively obtain predictive temperature rise curves and shelf-life loss before the plan is finalized, thereby achieving the technical effect of assisting decision-making and reducing the risk of cargo damage.

[0022] 3. This invention provides a closed-loop linkage mechanism that directly sends the results of graphical user interface operations to the physical scheduling system and simultaneously writes them into the blockchain for evidence storage, thereby achieving the technical effects of shortening the abnormal response link and ensuring the integrity and immutability of traceability data. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating the overall system architecture in an embodiment of the present invention.

[0024] Figure 2 This is a flowchart of the four-dimensional spatiotemporal thermolasso interaction method in an embodiment of the present invention;

[0025] Figure 3 This is a flowchart illustrating the closed-loop process of interface-driven physical traceability in an embodiment of the present invention. Detailed Implementation

[0026] The following is in conjunction with the appendix Figure 1-3 The specific embodiments of the present invention will be further described below. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0027] To make the technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the scope of protection of the present invention.

[0028] I. System Overall Architecture

[0029] The IoT-based digital traceability system for cold chain food storage provided in this embodiment of the invention consists of an IoT sensing layer, a network and edge computing layer, a thermal diffusion calculation engine and traceability database, a three-dimensional digital twin interaction layer, a prediction service, and a physical scheduling and evidence storage layer. There is both bottom-up data aggregation and top-down instruction issuance between the layers, forming a two-way closed loop.

[0030] The IoT sensing layer includes temperature and humidity sensors, RFID tags and reading devices distributed on shelves, pallets and outer packaging of goods, used to collect environmental and goods status data in real time. The network and edge computing layer includes wireless gateways and edge computing nodes, used to aggregate, pre-clean and upload the sensing layer data to the back-end server. The thermal diffusion calculation engine receives real-time temperature data and outputs irregular thermal field grid data.

[0031] The traceability database stores the traceability codes, three-dimensional spatial coordinates, and timestamps of warehousing and shelving for each pallet and batch. The three-dimensional digital twin interaction layer is the front-end graphical user interface, which includes a three-dimensional warehouse rendering module, a four-dimensional spatiotemporal lasso interaction module, a timeline slider control, and a drag-and-drop simulation interaction module. The simulation and prediction service includes a temperature rise prediction submodule and a shelf-life loss prediction submodule. The physical scheduling and evidence storage layer includes the standard scheduling interface between the warehouse management system and the warehouse control system, an automated guided vehicle scheduling server, and blockchain evidence storage nodes.

[0032] Each of the above functional modules can be implemented by a computing device consisting of one or more processors and memory, which executes computer program instructions stored in the memory. The edge computing nodes and back-end servers can be implemented using general-purpose servers or industrial computers. The three-dimensional digital twin interaction layer can run on a terminal device equipped with a graphics processing unit and use a web-based three-dimensional graphics library for scene rendering. This invention does not limit the specific hardware form.

[0033] The data from the perception layer is uploaded to the thermal diffusion calculation engine and the traceability database via the network layer. The thermal field mesh data output by the thermal diffusion calculation engine and the batch data in the traceability database are uploaded together to the 3D digital twin interaction layer for rendering and collision detection.

[0034] The inference requests generated by the user's selection and drag-and-drop operations in the interaction layer are sent to the inference and prediction service for calculation, and the calculation results are sent back to the interaction layer for visualization.

[0035] The solution confirmed by the user at the interaction layer is sent to the automated guided vehicle scheduling server and blockchain nodes through the interface of the physical scheduling and evidence storage layer, forming a complete two-way closed loop.

[0036] II. Specific Implementation Methods of the Irregular Four-Dimensional Spatiotemporal Thermodynamic Lasso Interaction Mechanism

[0037] In this embodiment, the edge computing node periodically receives the measured temperature data reported by the temperature sensor array, constructs an initial temperature field distribution matrix, and marks the sensor locations that exceed the preset safety threshold as the initial conditions for heat source disturbance.

[0038] Using the marked initial conditions as boundary conditions, the temperature field is iteratively calculated over time using the three-dimensional finite difference method for heat conduction, yielding spatial temperature distribution data at different times. The specific calculation formula is as follows:

[0039] ;

[0040] In the formula, Indicates the first Each time step and spatial grid coordinate is Temperature value at;

[0041] This indicates the predicted temperature value at the same grid point in the next time step;

[0042] Let be the thermal diffusivity of the air in the cold storage, and be a known constant related to the properties of the medium.

[0043] For time step;

[0044] , , The grid division step size is defined for each of the three spatial coordinate directions.

[0045] For the three-dimensional temperature distribution data at each time point, isosurfaces are extracted according to a preset temperature threshold. An isosurface extraction algorithm is used to generate three-dimensional thermal diffusion boundary grid data at the corresponding time point. The grid data at each time point are arranged in chronological order to form a three-dimensional thermal diffusion data sequence with time dimension.

[0046] The above thermal diffusion data sequence is sent to the three-dimensional digital twin interaction layer. After the front end receives the time parameter specified by the user through the time axis slider, it extracts the grid data corresponding to that time and performs gradient color semi-transparent rendering.

[0047] Simultaneously, it receives the drag-and-trim operation parameters of the user on the mesh boundary, performs local vertex displacement adjustment on the original mesh boundary, and generates the four-dimensional bounding box boundary data after user confirmation, that is, the boundary data composed of three spatial dimensions and one time dimension.

[0048] The intersection of the above four-dimensional bounding box boundary data with the three-dimensional spatial coordinates and entry timestamps of each pallet recorded in the traceability database is performed. A two-stage detection method is adopted, which combines fast filtering with fine-grained verification of axially aligned bounding boxes: First, fast elimination is performed by the spatial projection of the axially aligned bounding boxes of each pallet with the four-dimensional bounding box boundary data. Then, the candidate pallets that pass the initial screening are subjected to intersection judgment using directional bounding boxes. The set of traceability codes obtained from the screening is output. The above intersection judgment condition can be expressed by the following expression:

[0049] ;

[0050] In the formula, This refers to the four-dimensional bounding box. This indicates the four-dimensional bounding box of the tray to be judged;

[0051] This represents the coordinate dimensions of the traversal, with values ​​ranging from the three axes of space. , , and timeline ;

[0052] , They represent bounding boxes. In dimensions Minimum and maximum coordinates on, , Corresponding meaning;

[0053] symbol This indicates that a logical AND operation is performed on all dimensions, meaning that two bounding boxes are considered to intersect only if the interval overlap condition is met in all four dimensions.

[0054] The 3D digital twin interactive layer receives the set of traceability codes obtained from the above filtering, highlights the corresponding pallet icon in the 3D window, and simultaneously displays detailed traceability information such as batch number, production date, and original shelf life associated with each traceability code in the side information panel for managers to view.

[0055] III. Specific Implementation Methods of Context-Aware Parallel Inference Drag-and-Drop Interaction Mechanism

[0056] The 3D digital twin interaction layer listens to the drag events of users on the icons of damaged batches of goods, and obtains in real time the identifier of the target backup cold storage area pointed to by the current drag hover and its corresponding environmental temperature parameters, as well as the estimated transfer time obtained based on the warehouse plan path.

[0057] The prediction service receives the target cold storage ambient temperature uploaded from the front end, the initial temperature of the goods at the start of the transfer collected by the temperature sensor on the outer packaging, and the estimated transfer duration. It then uses a temperature recovery prediction model based on Newton's law of cooling to calculate the temperature change curve of the goods over time during the transfer process. The calculation formula is as follows:

[0058] ;

[0059] In the formula, Indicates the elapsed time since the start of the transfer. Predicted cargo temperature at any given time;

[0060] This indicates the ambient temperature of the target backup cold storage area;

[0061] Indicates the initial temperature of the goods at the moment the transfer begins;

[0062] The coefficient representing the temperature recovery rate, which is related to the insulation conditions of the transfer path and the packaging method, is a pre-calibrated known parameter;

[0063] This indicates the time elapsed since the start of the transfer.

[0064] Based on the temperature time series obtained above Based on the biochemical characteristics of this batch of food, such as activation energy and frequency factor, the shelf-life loss rate is determined using the following formula:

[0065] ;

[0066] In the formula, Indicates temperature Shelf life loss rate constant under certain conditions;

[0067] The frequency factor representing the food type is a known constant calibrated experimentally;

[0068] This indicates the activation energy of the reaction that causes the food to deteriorate in quality;

[0069] Represents the gas constant;

[0070] It represents absolute temperature in Kelvin.

[0071] The above loss rate Numerical integration over the transfer time interval yields the cumulative shelf-life loss corresponding to this transfer plan, calculated as follows:

[0072] ;

[0073] In the formula, Indicates the start time of the transition Until the end of the transfer Cumulative shelf-life loss;

[0074] A function representing the rate of loss as it changes over time;

[0075] Integral upper and lower limits , These correspond to the start and end times of the transfer, respectively.

[0076] The simulation and prediction service will return the temperature recovery curve data and cumulative shelf-life loss results obtained from the above calculations to the corresponding floating information panel in the front-end interaction layer for visualization in real time.

[0077] When a user hovers over different target icons in sequence during the same drag-and-drop process, the above calculation process is repeated for each hovered target, thereby enabling parallel simulation and comparison of multiple solutions. In other words, multiple candidate transfer solutions under the same abnormal event can simultaneously present their respective prediction results on the interface for the administrator to compare and select.

[0078] Once the user completes the drag-and-drop and clicks "confirm" in the pop-up confirmation dialog box, the front-end interaction layer will package the selected target cold storage identifier, the predicted temperature curve, and the shelf-life loss results as the simulation results and pass them to the subsequent physical scheduling and evidence storage process.

[0079] IV. Specific Implementation Methods for Using Graphical User Interface Parameter Manipulation to Drive a Closed-Loop Tracking System for the Physical World

[0080] After the user clicks the confirmation button, the front-end interaction layer serializes the solution data generated during the drag-and-drop simulation stage (including the set of affected traceability codes, target cold storage identifier, key points of the predicted temperature curve, and predicted shelf-life loss) into a standardized data structure and sends it to the back-end scheduling service through a secure interface protocol.

[0081] The backend scheduling service verifies the received data structure and, after confirming that constraints such as the current remaining capacity of the target cold storage and path reachability are met, generates a scheduling instruction that includes the coordinates of the starting point and the ending point, a list of batch traceability codes, and the priority level.

[0082] The backend scheduling service sends the above scheduling instructions to the automated guided vehicle (AGV) scheduling server through the standard interface between the warehouse management system and the warehouse control system. The AGV scheduling server assigns tasks to idle AGVs based on its own path planning algorithm and directs them to complete the actual cargo handling.

[0083] The backend scheduling service synchronously calls the write interface of the blockchain smart contract to encapsulate the four-dimensional bounding box boundary data of this event, the set of affected traceability codes, the confirmed selected transfer plan and prediction data into a structured record, and submits it to the blockchain network as an emergency transfer prospective record for consensus and evidence storage.

[0084] After the blockchain network completes consensus, it returns a unique hash identifier for this notarization transaction. The backend scheduling service writes this hash identifier back to the traceability database and establishes an association index with the traceability codes of the affected batches.

[0085] After receiving the scheduling acceptance result and evidence storage success mark returned by the backend, the front-end interaction layer updates the display status of the corresponding tray icon in the 3D window from pending disposal to transfer, and displays the newly added emergency disposal evidence storage record in the traceability details panel for verification by managers and subsequent traceability query personnel.

[0086] V. Other Optional Implementation Methods

[0087] In one alternative implementation, the four-dimensional spatiotemporal thermal lasso does not employ a two-level collision detection method combining isosurface grids and axis-aligned bounding boxes with directed bounding boxes. Instead, the storage space is pre-divided into regular voxel grids, and the thermal diffusion calculation results are discretized using the temperature values ​​of voxel units. User selection operations directly add, delete, and trim voxel sets. The traceability code collision judgment is simplified to determining whether the voxel at the center point of the pallet is covered by the selected voxel set. This implementation has relatively lower computational complexity and is suitable for application scenarios with high real-time requirements and relatively relaxed requirements for boundary accuracy. It can also achieve the functions of selecting irregular dynamic areas and filtering traceability codes.

[0088] In another alternative implementation, the drag-and-drop parallel simulation interaction does not rely on mouse drag-and-drop operations. Instead, it uses gesture recognition from virtual reality or augmented reality head-mounted displays. Users use gestures to grab virtual objects of damaged batches of goods and place them in the target cold storage virtual area. The system also triggers real-time prediction rendering of temperature recovery curves and shelf-life losses. This alternative interaction method is suitable for warehouse management scenarios equipped with virtual reality or augmented reality terminal devices and can achieve the same parallel simulation and solution confirmation effect as mouse drag-and-drop operations.

[0089] VI. Exemplary Application Scenarios

[0090] To facilitate understanding, the implementation process of this embodiment of the invention is illustrated by an example: When a local refrigeration unit in a cold storage fails and the temperature sensor reading at the corresponding location exceeds a preset safety threshold, the thermal diffusion calculation engine performs iterative calculations with the trigger point as the initial condition, and renders an irregular semi-transparent thermal diffusion area centered on the fault point and gradually expanding over time in the three-dimensional digital twin scene.

[0091] Managers select the diffusion boundary at a certain moment using the timeline slider and drag and trim the boundary locally to fit the actual shelf layout. The system then uses this four-dimensional bounding box to perform collision filtering on the traceability database and highlights several pallet icons that have fallen into the area in the three-dimensional view.

[0092] The manager selects one of the damaged batches of goods icons and drags it to the multiple backup cold storage icons on the interface. A floating panel pops up in real time, showing the predicted temperature recovery curves and shelf-life loss percentages after the goods are transferred to different backup cold storages. The manager then selects one of the backup cold storages and releases it for confirmation.

[0093] The system then generates a dispatch instruction and sends it to the automated guided vehicle to execute the transfer. At the same time, it writes the complete data of this abnormal handling into the blockchain ledger for evidence storage and adds an immutable emergency handling node to the historical record of the corresponding batch in the traceability database.

[0094] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of the present invention. Therefore, any modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope defined by the claims of the present invention.

Claims

1. A digital traceability method for food cold chain warehousing based on the Internet of Things, characterized in that, include: The system receives storage environment temperature data collected by a temperature sensor array, determines the trigger point of local refrigeration failure based on a preset safety threshold, performs heat diffusion calculation with the trigger point as the initial condition, obtains three-dimensional temperature distribution data that changes over time, and extracts the isosurface of the three-dimensional temperature distribution data according to a preset temperature threshold to generate three-dimensional thermal diffusion boundary grid data that evolves over time. In a 3D digital twin scene, the 3D thermal diffusion boundary mesh data is rendered. The user receives time parameters specified by the time axis slider and operation parameters for dragging and trimming the boundary of the 3D thermal diffusion boundary mesh to generate 4D bounding box boundary data, which includes 3D spatial boundary and time boundary. The intersection of the four-dimensional bounding box boundary data with the three-dimensional spatial coordinates and entry timestamps of each pallet recorded in the traceability database is used to filter out the set of traceability codes that fall within the range of the four-dimensional bounding box boundary data, and the pallet icon corresponding to the set of traceability codes is highlighted in the three-dimensional digital twin scene. The system receives user drag-and-drop operations on the icons of damaged batches of goods corresponding to the traceability code set. When the drag is hovered over the backup cold storage icon, the system obtains the ambient temperature parameters and estimated transfer time corresponding to the backup cold storage icon. Based on the ambient temperature parameters, the initial temperature of the goods at the start of the transfer obtained by temperature sensors deployed on the outer packaging of the goods, and the estimated transfer time, the system calculates temperature recovery prediction data of the goods during the transfer process over time. Based on the temperature recovery prediction data, the system calculates shelf-life loss prediction data and visualizes the temperature recovery prediction data and the shelf-life loss prediction data in a floating information panel next to the backup cold storage icon. When the drag is hovered over multiple backup cold storage icons, the system calculates and displays the corresponding temperature recovery prediction data and shelf-life loss prediction data respectively, so as to realize the parallel simulation and display of multiple transfer schemes. The system receives confirmation from the user after dragging and dropping the icon to the target backup cold storage and releasing it. The confirmed transfer plan parameters are then sent to the physical scheduling system, generating a scheduling instruction which is sent to the automated guided vehicle (AGV) scheduling server to drive the AGV to perform the cargo transfer. Simultaneously, the system writes the four-dimensional bounding box boundary data, the traceability code set, the confirmed transfer plan parameters, the temperature recovery prediction data, and the shelf-life loss prediction data corresponding to this abnormal event into the blockchain ledger for evidence storage. The blockchain-returned evidence transaction identifier is then associated with and stored in the corresponding traceability code in the traceability database.

2. The method according to claim 1, characterized in that, In the step of generating the three-dimensional thermal diffusion boundary mesh data, the temperature field is iteratively calculated over time using the three-dimensional thermal conduction finite difference method to obtain spatial temperature distribution data at different times, and the isosurface extraction algorithm is used to extract the isosurface of the spatial temperature distribution data to generate the three-dimensional thermal diffusion boundary mesh data.

3. The method according to claim 1, characterized in that, The steps for selecting the set of traceability codes that fall within the boundary data range of the four-dimensional bounding box include: firstly, aligning the spatial projection of the bounding box and the boundary data of the four-dimensional bounding box with the axis of each pallet for preliminary screening and elimination; then, using a directed bounding box to perform intersection judgment on the candidate pallets that have passed the preliminary screening, and outputting the set of traceability codes obtained from the screening.

4. The method according to claim 1, characterized in that, In the step of calculating the predicted temperature rise data, based on Newton's law of cooling, and considering the ambient temperature of the target backup cold storage area, the initial temperature of the goods at the start of the transfer, and the temperature rise rate coefficient related to the insulation conditions and packaging method of the transfer path, the predicted temperature of the goods during the transfer process is calculated as the temperature changes with the transfer time. In the step of calculating the predicted shelf-life loss data, based on the predicted temperature, and combined with the activation energy and frequency factor of the batch of food, the shelf-life loss rate is calculated, and the shelf-life loss rate is integrated over the transfer time interval to obtain the cumulative shelf-life loss.

5. The method according to claim 1, characterized in that, The records written into the blockchain ledger for evidence storage include the trigger time of this abnormal event, the boundary data of the four-dimensional bounding box, the traceability code set, the identifier of the selected target backup cold storage, the key points of the temperature recovery prediction data, and the shelf-life loss prediction data.

6. The method according to claim 1, characterized in that, Also includes: After receiving the scheduling acceptance result and blockchain evidence success identifier returned by the physical scheduling system, the display status of the corresponding tray icon in the three-dimensional digital twin scene is updated from pending processing to transfer.

7. A digital traceability system for food cold chain warehousing based on the Internet of Things, characterized in that, include: The Internet of Things (IoT) sensing layer includes temperature and humidity sensors distributed on shelves, pallets, and outer packaging of goods to collect real-time temperature data of the warehouse environment. The thermal diffusion calculation engine is connected to the IoT sensing layer and is used to determine the local cooling fault trigger point based on a preset safety threshold. The trigger point is used as the initial condition to perform thermal diffusion calculation, obtain three-dimensional temperature distribution data that changes over time, and extract the isosurface of the three-dimensional temperature distribution data according to the preset temperature threshold to generate three-dimensional thermal diffusion boundary mesh data that evolves over time. The traceability database stores the traceability code, three-dimensional spatial coordinates, and entry timestamp for each pallet. A three-dimensional digital twin interaction layer, communicatively connected to the thermal diffusion calculation engine and the traceability database, includes a three-dimensional warehouse rendering module, a four-dimensional spatiotemporal lasso interaction module, and a drag-and-drop simulation interaction module. The four-dimensional spatiotemporal lasso interaction module renders the three-dimensional thermal diffusion boundary mesh data, receives time parameters specified by the user via the time axis slider, and operation parameters for dragging and trimming the boundaries of the three-dimensional thermal diffusion boundary mesh to generate four-dimensional bounding box boundary data. It then performs intersection analysis between the four-dimensional bounding box boundary data and the three-dimensional spatial coordinates and entry timestamps of each pallet recorded in the traceability database to filter out a set of traceability codes, and highlights the pallet icons corresponding to the traceability code set. The drag-and-drop simulation interaction module receives user drag-and-drop operations on the icons of damaged batches of goods corresponding to the traceability code set, and obtains the corresponding ambient temperature parameters and estimated transfer time when the drag is hovered over the backup cold storage icon. The simulation and prediction service, which is connected to the three-dimensional digital twin interactive layer, includes a temperature rise prediction submodule and a shelf-life loss prediction submodule. It is used to calculate temperature rise prediction data based on the ambient temperature parameters, the initial temperature of the goods and the expected transfer time, and to calculate shelf-life loss prediction data based on the temperature rise prediction data. The temperature rise prediction data and the shelf-life loss prediction data are returned to the drag-and-drop simulation and interaction module for visualization in the floating information panel, and support the parallel simulation and display of multiple transfer schemes corresponding to multiple backup cold storage icons. The physical scheduling and evidence storage layer is communicatively connected to the three-dimensional digital twin interaction layer, including a scheduling interface, an automated guided vehicle scheduling server, and a blockchain evidence storage node. The scheduling interface is used to receive the transfer scheme parameters confirmed by the three-dimensional digital twin interaction layer, generate scheduling instructions and send them to the automated guided vehicle scheduling server to drive the automated guided vehicle to perform cargo transfer. The blockchain evidence storage node is used to write the four-dimensional bounding box boundary data, the traceability code set, the confirmed transfer scheme parameters, the temperature recovery prediction data, and the shelf-life loss prediction data corresponding to this abnormal event into the blockchain ledger for evidence storage, and to associate and store the evidence storage transaction identifier in the traceability database.

8. The system according to claim 7, characterized in that, When the four-dimensional spatiotemporal lasso interaction module performs intersection judgment, it first performs preliminary screening and elimination by aligning the bounding box of each pallet with the spatial projection of the boundary data of the four-dimensional bounding box. Then, it performs intersection judgment on the candidate pallets that pass the preliminary screening using a directed bounding box and outputs the traceability code set.

9. The system according to claim 7, characterized in that, The temperature recovery prediction submodule is used to calculate the predicted temperature of the goods during the transfer process based on Newton's law of cooling, the ambient temperature of the target backup cold storage area, the initial temperature of the goods at the start of the transfer, and the temperature recovery rate coefficient related to the insulation conditions and packaging methods of the transfer path. The shelf-life loss prediction submodule is used to calculate the shelf-life loss rate based on the predicted temperature, combined with the activation energy and frequency factor of the batch of food, and integrate the shelf-life loss rate over the transfer time interval to obtain the cumulative shelf-life loss.

10. The system according to claim 7, characterized in that, The physical scheduling and evidence storage layer is also used to drive the three-dimensional digital twin interaction layer to update the display status of the corresponding tray icon from the pending status to the transferring status after receiving the scheduling acceptance result and the blockchain evidence storage success identifier.

Images