Thermoplastic thickness prediction method and device based on industrial internet and electronic equipment

Abstract

The embodiment of the invention provides a thermoplastic thickness prediction method and device based on the industrial internet and electronic equipment, and relates to the technical field of thermoplastic. A to-be-thermoplastic material is divided into a plurality of first grid units, a graph structure of the to-be-thermoplastic material is constructed according to the position relation among the first grid units, nodes of the graph structure are used for representing the corresponding first grid units, and edges of the graph structure are used for representing the position relation between the two corresponding first grid units; acquiring first temperature distribution of the thermoplastic equipment; determining second temperature distribution corresponding to each first grid unit based on the first temperature distribution; and on the basis of the second temperature distribution, node features of multiple nodes in the graph structure are updated to obtain an updated graph structure, and according to the updated graph structure, thermoplastic thicknesses corresponding to the multiple first grid units are predicted to obtain a thermoplastic thickness prediction result of the to-be-thermoplastic material. The high-precision prediction of the product thickness is realized, and the product quality and the production efficiency are improved.

Description

Technical Field

[0001] This application relates to the field of thermoplastic technology, and in particular to a thermoplastic thickness prediction method, device and electronic equipment based on the Industrial Internet. Background Technology

[0002] Thermoplastic processing is a common material molding process widely used in the processing of thermoplastic materials such as plastics and rubber. It primarily utilizes the property that thermoplastic materials soften when heated to a certain temperature and harden when cooled. By controlling the heating and cooling processes, materials can be thermoplasticized into desired shapes.

[0003] Currently, thermoplastic processing technology involves fixing a thermoplastic sheet above a mold. The sheet is then heated to a softened state using a heating device. Next, the mold is moved towards the sheet, forming a sealed cavity. Air is extracted from the cavity using a vacuum pump, and atmospheric pressure forces the sheet tightly against the mold surface, completing the molding process. After molding, the sheet solidifies during cooling. Finally, compressed air is used to easily demold the molded plastic part.

[0004] Although thermoforming offers advantages such as high efficiency and low cost, uneven temperature distribution on the heating plate or the inherent properties of the material during vacuum forming can lead to uneven thickness in the final product, affecting product quality. For new shapes, current technology relies on manual parameter adjustment, resulting in low production efficiency. Summary of the Invention

[0005] This application provides a thermoplastic thickness prediction method, device, and electronic device based on the Industrial Internet, which can solve the technical defects of existing thermoplastic technology, such as uneven thickness of thermoplastic products and reliance on manual parameter adjustment, resulting in low product quality and low production efficiency.

[0006] In a first aspect, this application provides a thermoplastic thickness prediction method based on the Industrial Internet, applied to thermoplastic equipment, including:

[0007] The thermoplastic material to be heated is divided into multiple first grid units, and a graph structure of the thermoplastic material is constructed according to the positional relationship between the multiple first grid units. The nodes of the graph structure are used to represent the corresponding first grid units, and the edges of the graph structure are used to represent the positional relationship between two corresponding first grid units.

[0008] A first temperature distribution is obtained in a thermoplastic equipment used to process the material to be thermoplasticized.

[0009] Determine the second temperature distribution corresponding to each of the first grid cells based on the first temperature distribution;

[0010] Based on the second temperature distribution, the node features of multiple nodes in the graph structure are updated to obtain an updated graph structure. Based on the updated graph structure, the thermoplastic thickness of multiple first grid cells is predicted, thereby obtaining the thermoplastic thickness prediction result of the material to be thermoplasticized.

[0011] Optionally, determining the second temperature distribution corresponding to each of the first grid cells based on the first temperature distribution includes:

[0012] The thermoplastic equipment is divided into temperature zones to obtain multiple second grid units corresponding to the thermoplastic equipment;

[0013] Determine the heat conduction mapping relationship between multiple second grid cells and the first grid cell, and determine the second temperature distribution corresponding to each first grid cell based on the mapping relationship and the first temperature distribution.

[0014] Optionally, constructing the graph structure of the thermoplastic material based on the positional relationship between the plurality of first grid cells includes:

[0015] Based on the positional relationship between the multiple first grid cells, determine the multiple adjacent first grid cells adjacent to any one first grid cell;

[0016] Determine the grid distance between the first grid cell and any of the plurality of adjacent first grid cells, wherein the grid distance is used to indicate the distance between the grid center point of the first grid cell and the grid center point of the adjacent first grid cell;

[0017] Based on the grid distance and the material parameters of the thermoplastic material to be heated, determine the edge weights of the first grid cell and the plurality of adjacent first grid cells;

[0018] By using any one of the first mesh cells as a node of the graph structure and the multiple edge weights as edges of the graph structure, the graph structure of the thermoplastic material to be obtained is obtained.

[0019] Optionally, after constructing the graphical structure of the thermoplastic material to be manufactured, the method further includes:

[0020] Obtain the process parameters of the thermoplastic material to be treated, including: adsorption pressure, adsorption time, and heating time during the thermoplastic treatment process;

[0021] Determine the material parameters of the thermoplastic material to be used;

[0022] The initial node characteristics of the multiple first grid cells are determined based on the material parameters and process parameters of the multiple first grid cells.

[0023] Optionally, the step of updating the node features of multiple nodes within the graph structure based on the second temperature distribution to obtain the updated graph structure includes:

[0024] Based on the graph structure, determine the heat flux value and attention coefficient of any first grid cell and its adjacent first grid cells in the graph structure;

[0025] Based on the heat flux value, the attention coefficient, the initial node features, and the second temperature distribution, the node features of multiple nodes within the graph structure are updated to obtain the updated graph structure.

[0026] Optionally, predicting the thermoplastic thickness corresponding to a plurality of the first grid cells based on the updated graph structure includes:

[0027] The updated graph structure is used as input data for the thermoplastic thickness prediction model, which is then used to control the model to perform iterative processing. The thermoplastic thickness prediction model is based on machine learning methods and is trained using historical thermoplastic production data and the experience of domain experts.

[0028] Determine the total loss function of the thermoplastic thickness prediction model, the total loss function including a first loss function of the thermoplastic thickness prediction model and a second loss function of physical constraints;

[0029] When the total loss function is determined to converge, the thermoplastic thickness prediction model is controlled to stop iterative processing and the predicted thickness corresponding to multiple first grid cells is output.

[0030] Optionally, determining the total loss function of the thermoplastic thickness prediction model includes:

[0031] Obtain the thermoplastic thickness prediction result output by the thermoplastic thickness prediction model and the actual thickness of the material to be thermoplasticized after thermoplastic treatment;

[0032] Determine the absolute value of the difference between the actual thickness and the predicted result;

[0033] Determine whether the absolute value of the difference is less than the error threshold;

[0034] If the absolute value of the difference is less than the error threshold, the first loss function is determined to be converged.

[0035] If the absolute value of the difference is not less than the error threshold, the first loss function is determined to be non-convergent.

[0036] If the first loss function indicates convergence, a second loss function is determined for the thermoplastic thickness prediction model, wherein the second loss function is used to indicate physical constraints.

[0037] Based on the first loss function and the second loss function, the total loss function of the thermoplastic thickness prediction model is determined.

[0038] Optionally, the method further includes:

[0039] If the actual thickness is greater than the target thickness, the adsorption pressure and / or adsorption time and / or heating time shall be adjusted upward.

[0040] If the actual thickness is not greater than the target thickness, the adsorption pressure and / or adsorption time and / or heating time shall be reduced.

[0041] Secondly, this application provides a thermoplastic thickness prediction device based on the Industrial Internet, comprising:

[0042] The processing module is used to divide the thermoplastic material to be heated into multiple first grid units, and to construct a graph structure of the thermoplastic material to be heated according to the positional relationship between the multiple first grid units. The nodes of the graph structure are used to represent the corresponding first grid units, and the edges of the graph structure are used to represent the positional relationship between two corresponding first grid units.

[0043] An acquisition module is used to acquire the first temperature distribution of the thermoplastic equipment, the thermoplastic equipment being used to process the material to be thermoplasticized;

[0044] The determination module is used to determine the second temperature distribution corresponding to each of the first grid cells based on the first temperature distribution;

[0045] The processing module is further configured to update the node features of multiple nodes within the graph structure based on the second temperature distribution to obtain an updated graph structure, and predict the thermoplastic thickness corresponding to multiple first grid cells based on the updated graph structure, thereby obtaining the thermoplastic thickness prediction result of the material to be thermoplasticized.

[0046] Optionally, the processing module is further configured to perform temperature zone division processing on the thermoplastic equipment to obtain multiple second grid units corresponding to the thermoplastic equipment;

[0047] The determining module is further configured to determine the heat conduction mapping relationship between the plurality of second grid cells and the first grid cell, and determine the second temperature distribution corresponding to each first grid cell based on the mapping relationship and the first temperature distribution.

[0048] Optionally, the determining module is further configured to determine, based on the positional relationship between the plurality of first grid cells, a plurality of adjacent first grid cells adjacent to any one first grid cell;

[0049] The determining module is further configured to determine the grid distance between the first grid cell and any one of the plurality of adjacent first grid cells, wherein the grid distance is used to indicate the distance between the grid center point of the first grid cell and the grid center point of the adjacent first grid cell;

[0050] The determining module is further configured to determine multiple edge weights of the first grid cell and the plurality of adjacent first grid cells based on the grid distance and the material parameters of the thermoplastic material to be heated;

[0051] The processing module is further configured to use any one of the first mesh cells as a node of the graph structure and the multiple edge weights as edges of the graph structure to obtain the graph structure of the thermoplastic material to be processed.

[0052] Optionally, the acquisition module is further configured to acquire the process parameters of the thermoplastic material to be thermoplasticized, the process parameters including: adsorption pressure, adsorption time and heating time during the thermoplasticization process;

[0053] The determining module is also used to determine the material parameters of the thermoplastic material to be thermoplasticized;

[0054] The determining module is further configured to determine the initial node characteristics of the plurality of first grid cells based on the material parameters and process parameters of the plurality of first grid cells.

[0055] Optionally, the determining module is further configured to determine the heat flux value and attention coefficient of any first grid cell and the adjacent first grid cells in the graph structure based on the graph structure;

[0056] The processing module is further configured to update the node features of multiple nodes within the graph structure based on the heat flux value, the attention coefficient, the initial node features, and the second temperature distribution, to obtain an updated graph structure.

[0057] Optionally, the processing module is further configured to use the updated graph structure as input data for the thermoplastic thickness prediction model, and control the thermoplastic thickness prediction model to perform iterative processing; the thermoplastic thickness prediction model is based on machine learning methods and is trained according to historical thermoplastic production data and domain expert experience.

[0058] The determining module is further configured to determine the total loss function of the thermoplastic thickness prediction model, the total loss function including a first loss function of the thermoplastic thickness prediction model and a second loss function of physical constraints;

[0059] The processing module is further configured to, when determining that the total loss function has converged, control the thermoplastic thickness prediction model to stop iterative processing and output the predicted thickness corresponding to multiple first grid cells.

[0060] Optionally, the device further includes: a determination module;

[0061] The acquisition module is also used to acquire the thermoplastic thickness prediction result output by the thermoplastic thickness prediction model and the actual thickness of the material to be thermoplasticized after thermoplastic treatment.

[0062] The determining module is further configured to determine the absolute value of the difference between the actual thickness and the predicted result;

[0063] The judgment module is used to determine whether the absolute value of the difference is less than the error threshold;

[0064] The determining module is further configured to determine the convergence of the first loss function when the absolute value of the difference is less than the error threshold;

[0065] The determining module is further configured to determine that the first loss function is non-convergent if the absolute value of the difference is not less than the error threshold.

[0066] The determining module is further configured to determine a second loss function of the thermoplastic thickness prediction model when the first loss function indicates convergence, wherein the second loss function is used to indicate physical law constraints;

[0067] The determining module is further configured to determine the total loss function of the thermoplastic thickness prediction model based on the first loss function and the second loss function.

[0068] Optionally, the processing module is further configured to increase the adsorption pressure and / or adsorption time and / or heating time when the actual thickness is greater than the target thickness.

[0069] The processing module is also used to reduce the adsorption pressure and / or adsorption time and / or heating time when the actual thickness is not greater than the target thickness.

[0070] Thirdly, this application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0071] The memory stores computer-executed instructions;

[0072] The processor executes computer execution instructions stored in the memory to implement the industrial internet-based thermoplastic thickness prediction method as described in the first aspect and various possible implementations of the first aspect above.

[0073] Fourthly, this application provides a computer-readable storage medium storing computer-executable instructions thereon, which, when executed by a processor, are used to implement the industrial internet-based thermoplastic thickness prediction method as described in the first aspect and various possible implementations of the first aspect.

[0074] Fifthly, this application provides a program product, including a computer program, which, when executed by a processor, implements the thermoplastic thickness prediction method based on the Industrial Internet as described above.

[0075] This application provides a method, apparatus, and electronic device for predicting thermoplastic thickness based on the Industrial Internet. The method divides the thermoplastic material to be processed into multiple first grid cells, constructs a graph structure based on the positional relationships between these cells, where each node represents a first grid cell and the edges represent the positional relationships between multiple first grid cells. It acquires a first temperature distribution of the thermoplastic equipment used to process the material; determines a possible second temperature distribution for each first grid cell during processing based on the first temperature distribution; updates the node information of each first grid cell in the graph structure according to the calculated second temperature distribution, thereby predicting the thermoplastic thickness of each first grid cell after processing; and integrates the predicted thickness information of all first grid cells to obtain the predicted thermoplastic thickness of the entire material after processing. This method considers both the uneven heating of different parts of the material and the spatial correlation between regions, avoiding prediction biases caused by neglecting local temperature differences or structural relationships in traditional methods. The prediction results are closer to the actual molding effect, achieving high-precision prediction of product thickness and improving product quality and production efficiency. Attached Figure Description

[0076] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0077] Figure 1 A flowchart illustrating a thermoplastic thickness prediction method based on the Industrial Internet provided in this application. Figure 1 ;

[0078] Figure 2 A flowchart illustrating a thermoplastic thickness prediction method based on the Industrial Internet provided in this application. Figure 2 ;

[0079] Figure 3 A flowchart illustrating a thermoplastic thickness prediction method based on the Industrial Internet provided in this application. Figure 3 ;

[0080] Figure 4A flowchart illustrating a thermoplastic thickness prediction method based on the Industrial Internet provided in this application. Figure 4 ;

[0081] Figure 5 A schematic diagram of a thermoplastic thickness prediction device based on the Industrial Internet provided in this application;

[0082] Figure 6 This is a schematic diagram of the structure of an electronic device provided in this application.

[0083] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0084] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0085] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with relevant laws, regulations, and standards, necessary confidentiality measures have been taken, they do not violate public order and good morals, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0086] Furthermore, the technical solution involved in this application, which involves big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.) and the use of artificial intelligence technology for automated decision-making, and makes decisions that have a significant impact on personal rights based on the results of automated decision-making, provides users with corresponding operation entry points for users to choose to agree to or reject the results of automated decision-making; if the user chooses to reject, the process will proceed to the expert decision-making process.

[0087] It should be noted that the thermoplastic thickness prediction method, device and electronic equipment based on the Industrial Internet provided in this application can be used in the field of thermoplastic technology, or in any field other than thermoplastic. The application field of the thermoplastic thickness prediction method, device and electronic equipment in this application is not limited.

[0088] Thermoplastic processing is a molding process widely used for thermoplastic materials such as plastics and rubber. Its basic principle is based on the property that thermoplastic materials soften and melt when heated, exhibiting good flowability and plasticity, and then harden again upon cooling, maintaining a specific shape. By precisely controlling the heating, deformation, and cooling processes, materials can be processed into products with complex structures and precise dimensions. This process offers advantages such as high production efficiency, low cost, and suitability for mass production, and is widely used in home appliances, automobiles, packaging, and daily necessities.

[0089] Currently, vacuum forming technology is widely used in thermoplastic processing. The specific process is as follows: First, the thermoplastic sheet is fixed above the mold; then, the sheet is uniformly heated to a softened state using a heating device; next, the mold is driven to rise or the sheet is driven to fall, forming a sealed cavity between them; then, the vacuum system is activated, and the air in the cavity is extracted by a vacuum pump, using external atmospheric pressure to press the softened sheet tightly against the mold surface, achieving precise forming; after forming, the sheet is cooled in the mold to maintain its geometric shape; finally, compressed air is introduced into the back of the mold to separate the product from the mold, completing the demolding operation.

[0090] Despite the advantages of thermoplastic processing, such as high production efficiency and low cost, key quality control challenges remain in vacuum forming. Uneven temperature distribution across the heating plate, or the influence of factors like material thermal conductivity and viscoelasticity, can lead to inconsistent stress and tension on the sheet material during forming. This can result in areas of excessively thin or thick walls in the final product, severely impacting quality. Furthermore, for trial molding of new structures or specifications, current technologies rely heavily on operator experience to manually adjust process parameters such as heating temperature, heating time, and vacuum duration. This adjustment process is cumbersome and time-consuming, reducing production efficiency and increasing raw material and energy consumption.

[0091] To address the aforementioned issues, this application proposes a thermoplastic thickness prediction method based on the Industrial Internet. This method couples the material's spatial structure with its temperature distribution in a model. First, the material is divided into multiple first-level grid cells, and a graph structure reflecting their adjacency relationships is constructed. Then, combined with the first temperature distribution of the thermoplastic equipment, a corresponding second temperature distribution is matched to each first-level grid cell. Based on the graph structure and updated node features, the thermoplastic thickness of each first-level grid cell is predicted. By no longer treating the material as a homogeneous whole, but rather from the perspective of "local heating + spatial correlation," the method accurately recreates the thickness changes caused by temperature unevenness and structural interactions during thermoplastic processing. This not only solves the problem of inaccurate predictions caused by neglecting local temperature differences and structural interactions in traditional methods, but also avoids the cost and time waste associated with relying on extensive trial-and-error experience. Thus, without increasing hardware modifications, high-precision prediction of product thickness is achieved, improving product quality and production efficiency.

[0092] Industrial Internet platforms, with their collaborative "cloud-edge-device" architecture, offer a new technological path for intelligent manufacturing. A typical platform is usually divided into a foundation layer, a capability layer, and an application layer: the foundation layer provides device access, data acquisition, and edge computing resources; the capability layer integrates core intelligent services such as data processing, model training, and inference; and the application layer implements business logic and human-machine interaction for specific industrial scenarios. In the embodiments of this application, the temperature control algorithm for the heating plate of a thermoplastic equipment can be deployed on an industrial Internet platform, fully leveraging the platform's advantages in data aggregation, computing power scheduling, and remote iteration. Specifically, on-site sensors collect key parameters such as the heating plate temperature in real time and upload them to the platform's edge computing nodes; operations including data preprocessing, model inference, and control strategy generation are completed at the capability layer; and instructions are issued through the PCL (Programmable Control Logic) in the industrial operating system to achieve precise adjustment of the heating device, completing one control iteration. This architecture not only improves the adaptability and robustness of the temperature control system but also lays the foundation for subsequent online model learning and cross-device knowledge transfer.

[0093] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0094] Figure 1 A flowchart illustrating a thermoplastic thickness prediction method based on the Industrial Internet provided in this application embodiment. Figure 1 .like Figure 1 As shown, the thermoplastic thickness prediction method provided in this embodiment includes:

[0095] S101. Divide the thermoplastic material to be heated into multiple first grid units, and construct a graph structure of the thermoplastic material according to the positional relationship between the multiple first grid units. The nodes of the graph structure are used to represent the corresponding first grid units, and the edges of the graph structure are used to represent the positional relationship between the corresponding two first grid units.

[0096] The "first grid cell" refers to the thermoplastic material obtained after spatial discretization. It can be a regular or irregular shape (such as a square, rectangle, triangle, or polygonal cell), and its size can be set according to the material shape and prediction accuracy requirements. Each first grid cell represents the physical state of a local region of the material.

[0097] A graph structure, composed of nodes and edges, is used to describe the relationships between objects. In this application, the graph structure is used to express the spatial topology of the material: nodes correspond to each first grid cell, and edges represent two first grid cells that are physically adjacent or have positional relationships such as thermal conduction or mechanical coupling.

[0098] One possible implementation involves obtaining a geometric model of the thermoplastic material to be thermoplasticized, and then using a mesh generation algorithm to divide the material into several first mesh cells. Specifically, during the partitioning process, adjacent cells are guaranteed to share boundaries or vertices to accurately reflect the continuity of the material. Each first mesh cell is traversed, and the adjacency relationships between them are identified. That is, if two first mesh cells are directly adjacent in space (i.e., share an edge), an edge is established between the corresponding two nodes, ultimately resulting in an undirected graph that fully represents the position and connectivity of the thermoplastic material.

[0099] S102, Obtain the first temperature distribution of the thermoplastic equipment.

[0100] Thermoplastic equipment refers to devices used to heat and soften thermoplastic materials and shape them, such as hot presses, infrared heating furnaces, hot air circulating ovens, or the heating chambers of injection molding machines.

[0101] The first temperature distribution refers to the temperature distribution of the heating area (such as the heating plate, the inside of the cavity, or the surface of the heating mold) of the thermoplastic equipment during operation. The first temperature distribution reflects the actual temperature levels at different locations of the equipment.

[0102] One possible approach is to pre-install multiple temperature sensors (such as thermocouples, infrared temperature measurement points, or thermal imaging arrays) at key locations in the thermoplastic equipment. When the equipment reaches a steady-state operation, real-time temperature data from each measurement point is collected synchronously, and a continuous spatial temperature distribution is generated through interpolation.

[0103] Another possible approach is to call the real-time temperature distribution data recorded by the temperature monitoring system (such as industrial control software) built into the thermoplastic equipment to obtain the first temperature distribution of the thermoplastic equipment.

[0104] Regardless of the method used, the obtained first temperature distribution should reflect the temperature distribution of the thermoplastic material as accurately as possible during the actual processing.

[0105] S103. Determine the second temperature distribution corresponding to each first grid cell based on the first temperature distribution.

[0106] The second temperature distribution refers to the actual heating state received by each first grid cell from the thermoplastic equipment at its spatial location; it can be understood as the effective temperature distribution experienced by that grid cell during processing. It represents the local temperature distribution "sensed" or "endured" by each discrete first cell on the thermoplastic material.

[0107] The second temperature distribution corresponding to each first grid cell is determined based on the first temperature distribution. This is essentially a spatial matching or mapping process, which associates the temperature information of the thermoplastic equipment (first temperature distribution) with the spatial coordinates of the first grid cells of the material to be thermoplasticized, thereby assigning a second temperature distribution corresponding to the location of each first cell.

[0108] One possible implementation involves obtaining the actual placement position of the thermoplastic material on the heating device of the thermoplastic equipment, ensuring that the grid cells of the material and the second grid cells of the thermoplastic equipment are in the same spatial coordinate system. For each first grid cell, the spatial coordinates of its geometric center point are extracted, and then the distribution at the corresponding position is obtained by searching or interpolation in the first temperature distribution based on these coordinates. Finally, each first grid cell is assigned a specific temperature value or short-time temperature sequence, i.e., the "second temperature distribution".

[0109] S104. Based on the second temperature distribution, update the node features of multiple nodes in the graph structure to obtain the updated graph structure. Based on the updated graph structure, predict the thermoplastic thickness of multiple first grid cells to obtain the thermoplastic thickness prediction result of the material to be thermoplasticized.

[0110] In this context, node features refer to the set of information carried by each node in the graph structure, which characterizes the state of the corresponding first grid cell. Initially, node features may include location coordinates, area, material properties, etc.; after updating the node features, the node features will incorporate thermal state information provided by the second temperature distribution.

[0111] Understandably, the update process refers to supplementing the original node features based on the second temperature distribution, so that the node features can reflect their temperature distribution. Therefore, updating the graph structure does not change the topological connections of the graph structure (i.e., the edge relationships remain unchanged), but only assigns information containing the temperature distribution to the feature vectors of each node to support subsequent thermoplastic thickness prediction.

[0112] One possible implementation is to add the second temperature distribution corresponding to each first grid cell (e.g., the temperature change trend at that location during the heating process) as new information to the features of the corresponding node in the graph structure. For example, a node that originally only recorded "I am in the upper left corner of the material and the size of my area" would also record "the temperature at my location is 180℃" after the update.

[0113] The thickness of each first grid cell after thermoforming is predicted based on the updated node characteristics. In some embodiments, the thickness can be directly determined by looking up tables or applying empirical rules based on the node's temperature and material properties. For example, the higher the temperature, the softer the material to be thermoformed, and the thinner the material will be. In other embodiments, a comprehensive judgment can be made by combining the temperature and state of the node with its neighboring nodes. For example, if a first grid cell itself is not hot, but the other connected first grid cells are very hot, then the first grid cell may be "thinned"; conversely, if the surrounding area is cool and the support is strong, the thermoformed thickness may be thicker.

[0114] By injecting temperature distribution into the nodes of the graph structure and preserving spatial relationships, the thermoplastic thickness is predicted. This method reflects both the heating degree of each first grid region and the mutual influence between adjacent first grid cells. This makes the prediction of thermoplastic thickness closer to the actual physical process, significantly improving the precision and accuracy of the prediction compared to traditional methods that rely solely on average temperature or estimation.

[0115] This embodiment provides a thermoplastic thickness prediction method based on the Industrial Internet. The method divides the material to be thermoplasticized into multiple first grid cells and constructs a graph structure of the material based on the positional relationships between these cells. Nodes in the graph structure represent the corresponding first grid cells, and edges represent the positional relationships between two corresponding first grid cells. The method acquires a first temperature distribution of the thermoplasticizing equipment; determines a second temperature distribution for each first grid cell based on the first temperature distribution; updates the node features of multiple nodes within the graph structure based on the second temperature distribution to obtain an updated graph structure; and predicts the thermoplastic thickness corresponding to multiple first grid cells based on the updated graph structure, thereby obtaining the predicted thermoplastic thickness of the material. This method achieves high-precision prediction of product thickness, improving product quality and production efficiency.

[0116] Figure 2 A flowchart illustrating a thermoplastic thickness prediction method based on the Industrial Internet provided in this application embodiment. Figure 2 .like Figure 2 As shown, in Figure 1 Based on the examples, the method for predicting thermoplastic thickness based on the Industrial Internet is described in detail, including:

[0117] S201. Divide the thermoplastic material to be heated into multiple first grid units, and construct a graph structure of the thermoplastic material according to the positional relationship between the multiple first grid units. The nodes of the graph structure are used to represent the corresponding first grid units, and the edges of the graph structure are used to represent the positional relationship between the corresponding two first grid units.

[0118] Step S201 is the same as step S101 above, and will not be repeated here.

[0119] S202, Obtain the first temperature distribution of the thermoplastic equipment.

[0120] Step S202 is the same as step S102 above, and will not be repeated here.

[0121] S203. Perform temperature zone division processing on the thermoplastic equipment to obtain multiple second grid units corresponding to the thermoplastic equipment.

[0122] The thermoplasticizing equipment includes upper and lower heating plates and can be applied to manufacturing industries such as home appliances, automobiles, and packaging. Specifically, it utilizes the property that the material to be thermoplasticized softens when heated to a certain temperature. By vacuuming, the material is forced to adhere tightly to the heating plates of the thermoplasticizing equipment, thereby forming the desired shape and completing the thermoplasticizing process.

[0123] Temperature zone division refers to dividing the heating device of thermoplastic equipment, such as upper and lower heating plates, into several temperature zones in physical space. Each temperature zone is called a "temperature zone". Each temperature zone is further subdivided into smaller units to obtain a second grid unit.

[0124] One possible implementation involves a 1.2m x 0.8m rectangular metal plate as the upper heating plate of a refrigerator liner thermoforming device. The heating plate is divided into four temperature zones: upper left, upper right, lower left, and lower right. Each temperature zone is not the smallest unit; it needs to be further subdivided into smaller second grid units to construct the graph structure. For example, each of the above temperature zones can be further divided into 5x5 small grids, i.e., the second grid unit. It is important to note that the second grid unit is not a physically independent heating element, but rather the logical smallest temperature monitoring unit.

[0125] S204. Determine the heat conduction mapping relationship between multiple second grid cells and the first grid cell, and determine the second temperature distribution corresponding to each first grid cell based on the mapping relationship and the first temperature distribution.

[0126] The first temperature value indicates the actual temperature of the heating element, i.e., the heating plate, in the thermoplastic equipment. The second grid cell indicates the smallest logical temperature detection unit after the heating plate of the thermoplastic equipment is divided into temperature zones. Mapping the first temperature value to each second grid cell provides input data for subsequent heat conduction simulation, thickness prediction, etc.

[0127] One possible implementation involves dividing the heating plate of the thermoplastic equipment into four temperature zones: upper left, upper right, lower left, and lower right. Each temperature zone is further subdivided into 10x10 second grid units. For example, if the temperature of the upper left temperature zone is 200℃, then the temperature value of all corresponding second grid units in the upper left temperature zone will also be 200℃. The other temperature zones are similar and will not be described in detail here.

[0128] Align the first and second grid cells spatially to ensure they share the same reference frame, establishing a positional correspondence between them. Use the heat conduction equation to determine the heat conduction mapping relationship. Based on this mapping relationship and the first temperature distribution, determine the second temperature distribution for each arbitrary first grid cell.

[0129] That is, from the known temperature of the thermoplastic equipment, through physical modeling and spatial mapping, the actual temperature of the material to be thermoplasticized is deduced, that is, the second temperature distribution of each first grid cell.

[0130] The process takes into account heat loss during heat conduction, avoiding prediction errors caused by simply approximating the actual temperature of the material to be thermoplasticized as the heating temperature of the thermoplastic equipment.

[0131] S205. Based on the graph structure, determine the heat flux value and attention coefficient of any first grid cell and its adjacent first grid cells in the graph structure.

[0132] The formula for calculating the heat flux value is, for example:

[0133]

[0134] in, For the heat flow from node i to j, The thermal conductivity coefficient, Let i be the temperature value of node i. Let j be the temperature value of node j. Let i be the spatial coordinates of node i. Let j be the spatial coordinates of node j. Let be the grid distance between node i and node j.

[0135] The formula for calculating the attention coefficient is as follows:

[0136]

[0137] in, Let be the attention coefficient between node i and node j. For learnable weight matrix, For attention vectors, Let i be the set of adjacent grids of node i. Let i be the node characteristics. Let j be the node characteristics of node j.

[0138] S206. Based on the heat flux value, attention coefficient, initial node characteristics, and second temperature distribution, update the node characteristics of multiple nodes in the graph structure to obtain the updated graph structure.

[0139] Specifically, for each node i, information about all its neighboring grids j is collected, and the node features of multiple nodes within the graph structure are updated by weighting the heat flux value and attention coefficient, resulting in the updated graph structure.

[0140] S207. Use the updated graph structure as input data for the thermoplastic thickness prediction model and control the thermoplastic thickness prediction model to perform iterative processing; the thermoplastic thickness prediction model is based on machine learning methods and is trained according to historical thermoplastic production data and domain expert experience.

[0141] S208. Determine the total loss function of the thermoplastic thickness prediction model. The total loss function includes the first loss function of the thermoplastic thickness prediction model and the second loss function of the physical constraints.

[0142] Optionally, the total loss function of the thermoplastic thickness prediction model is determined, including:

[0143] Obtain the thermoplastic thickness prediction results output by the thermoplastic thickness prediction model and the actual thickness of the material to be thermoplasticized after thermoplastic treatment.

[0144] Determine the absolute value of the difference between the actual thickness and the predicted result;

[0145] Determine whether the absolute value of the difference is less than the error threshold;

[0146] If the absolute value of the difference is less than the error threshold, the first loss function is determined to be convergent.

[0147] If the absolute value of the difference is not less than the error threshold, the first loss function is determined to be non-convergent.

[0148] If the first loss function indicates convergence, a second loss function is determined for the thermoplastic thickness prediction model. The second loss function is used to indicate physical constraints.

[0149] Based on the first loss function and the second loss function, the total loss function of the thermoplastic thickness prediction model is determined.

[0150] In this process, after the thermoplastic thickness prediction model completes one forward inference, it outputs the predicted thickness value corresponding to each first grid cell. The predicted thickness value is calculated based on the updated graph structure. After the thermoplastic sheet is thermoformed in a real thermoplastic equipment, the actual thickness of the thermoplastic sheet is measured using appropriate measuring tools.

[0151] For any first grid cell, calculate the absolute value of the difference between the actual thickness and the predicted thickness. Based on a pre-set error threshold, compare the threshold with the absolute value of the difference to determine if the first loss function has converged. Assuming the first loss function has converged, a second loss function is introduced. Even if the prediction is accurate, it cannot be guaranteed that the prediction results of the thermoplastic thickness prediction model conform to physical laws; therefore, a physical law constraint loss is introduced as the second loss function.

[0152] The second loss function is used to measure whether the output of the thermoplastic thickness prediction model satisfies known physical laws. For example, while driving learning with data, it enforces the principle of energy conservation to avoid predicting non-physical phenomena such as "heat flowing from low temperature to high temperature".

[0153] Based on the first and second loss functions, the total loss function is determined, which is a weighted sum of the first and second loss functions.

[0154] S209. When the total loss function is determined to converge, control the thermoplastic thickness prediction model to stop iterative processing and output the predicted thickness corresponding to multiple first grid cells.

[0155] The updated graph structure ensures that each node contains the latest feature information. This feature information can include, but is not limited to, temperature, material properties, and process parameters, all of which are key factors affecting the final thermoplastic thickness. The updated graph structure is used as input data for the thermoplastic thickness prediction model. The model processes the input graph structure to obtain the preliminary predicted thickness for each first grid cell. The difference between the model output and the actual observations is calculated, and one or more loss functions are used to quantify this difference.

[0156] Determine the total loss function of the thermoplastic thickness prediction model. If the total loss function indicates convergence, it means that the thermoplastic thickness prediction model has learned an effective mapping relationship from input to output. At this point, the thermoplastic thickness prediction model can be used to predict the thickness of all first grid cells, outputting the predicted thickness for multiple first grid cells.

[0157] Optionally, after obtaining the predicted thickness corresponding to multiple first grid cells, in some embodiments, the actual thickness of the thermoplastic material after thermoplasticization is greater than the target thickness. In this case, the adsorption pressure and / or adsorption time and / or heating time can be adjusted upward.

[0158] In other embodiments, the actual thickness of the thermoplastic material after thermoplasticization is not greater than the target thickness. In this case, the adsorption pressure and / or adsorption time and / or heating time can be reduced.

[0159] Based on the actual thermoplastic results, i.e. the deviation between the actual thickness and the design requirements, i.e. the target thickness, process parameters such as adsorption pressure, adsorption time, and heating time are adjusted to gradually approach the ideal molding quality.

[0160] Specifically, if the actual thickness is greater than the target thickness, it indicates that there is excessive accumulation of the thermoplastic material, resulting in localized overthickness. This may be due to insufficient adsorption pressure and / or insufficient adsorption time and / or insufficient heating time of the thermoplastic material. Therefore, the adsorption pressure and / or adsorption time and / or heating time are adjusted upwards to gradually bring the actual thickness closer to the target thickness.

[0161] If the actual thickness is not greater than the target thickness, it indicates that the thermoplastic material is overstretched and there is a risk of breakage. This may be caused by excessive adsorption pressure and / or excessive adsorption time and / or excessive heating time. Therefore, the adsorption pressure and / or adsorption time and / or heating time should be reduced to gradually bring the actual thickness closer to the target thickness.

[0162] This embodiment provides a thermoplastic thickness prediction method based on the Industrial Internet. The method discretizes the material to be thermoplasticized into multiple first grid cells and constructs a graph structure that reflects their spatial adjacency relationship. At the same time, it obtains the first temperature distribution of the thermoplastic equipment and divides the thermoplastic equipment into temperature zones to obtain multiple second grid cells. By establishing a heat conduction mapping relationship between the second grid cell and the first grid cell, the temperature distribution of the thermoplastic equipment is accurately mapped to the material to be thermoplasticized, obtaining the second temperature distribution corresponding to each first grid cell. Based on this, the heat flow transfer characteristics between any first grid cell and its adjacent first cells are analyzed using graph structure analysis, and an attention mechanism is introduced to quantify the degree of influence of different adjacent first cells on the first cell. By combining the initial node features, the second temperature distribution, the heat flow value, and the attention coefficient, the features of each node in the graph structure are dynamically updated to form an updated graph structure containing temperature distribution and structural coupling information. The updated graph structure is input into a pre-trained thermoplastic thickness prediction model. This model integrates historical production data and domain expert experience and is constructed using machine learning methods. During the model training process, not only is the error between the predicted thickness and the measured value considered (first loss function), but also physical laws such as heat conduction and mass conservation are embedded as constraints (second loss function), forming the total loss function. When the total loss function converges, the model stops iterating and outputs the predicted thickness of each first grid cell, finally obtaining the predicted thermoplastic thickness of the material to be thermoplasticized. By acquiring the actual temperature distribution of the thermoplastic equipment and combining temperature zone division with heat conduction mapping, the non-uniform temperature distribution of the equipment is accurately mapped to each thermoplastic material grid cell, improving the realism of the thermoplastic material temperature distribution. Based on this, heat flux and attention coefficients are introduced to dynamically quantify the interaction between adjacent grid cells, and a second temperature distribution is integrated to update the graph node features, enabling the prediction model to accurately reflect the effect of local heating and structural coupling on thermoplastic thickness. The updated graph structure is input into a machine learning prediction model that integrates historical data, expert experience, and physical constraints (such as heat conduction laws). By balancing data fitting and physical rationality, the accuracy and generalization ability of thickness prediction are significantly improved. Therefore, high-precision prediction of product thickness distribution can be achieved, effectively reducing scrap rates and improving product quality and production efficiency.

[0163] Figure 3 A flowchart illustrating a thermoplastic thickness prediction method based on the Industrial Internet provided in this application embodiment. Figure 3 .like Figure 3 As shown, in Figure 1 Based on the embodiments, a possible implementation of constructing a graph structure of the thermoplastic material according to the positional relationship between multiple first grid cells is described in detail, including:

[0164] S301. Based on the positional relationship between multiple first grid cells, determine multiple adjacent first grid cells adjacent to any one first grid cell.

[0165] Among them, the adjacent first grid cell refers to other grid cells that are in direct contact or adjacent to a certain first grid cell in space and have interactions such as heat exchange and force transmission.

[0166] Because heat conduction primarily occurs on the forward contact surface, industrial applications commonly use four-neighbor domains, i.e., adjacent first grid cells in the top, bottom, left, and right directions. Based on the positional relationship between the first grid cell and its adjacent first grid cells, the adjacent grid cells in the top, bottom, left, and right directions corresponding to the first grid cell are determined. However, for first grid cells located at the edges or corners of the thermoplastic material, the number of adjacent first grid cells will decrease, requiring boundary checks to prevent them from exceeding the boundaries.

[0167] One possible implementation is that the first grid cell G(5,5) has the following adjacent grid cells in the four directions of up, down, left and right: G(4,5), G(6,5), G(5,4), G(5,6).

[0168] Determining the multiple adjacent first grid cells corresponding to any first grid cell essentially involves establishing a "local connection network" for each first grid cell.

[0169] S302. Determine the grid distance between the first grid cell and any of the multiple adjacent first grid cells.

[0170] This process involves determining the spatial location information of each first grid cell, typically represented by the coordinates of its center point. For any given first grid cell, one adjacent first grid cell is randomly selected from its multiple neighboring cells, and the grid distance is determined based on the coordinates of the two cells. The grid distance indicates the distance between the center point of the first grid cell and the center points of its adjacent first grid cells.

[0171] One possible implementation is to use G(5,5) as the first grid cell and its right neighbor G(5,6) as the adjacent first grid cell. The grid distance between G(5,5) and its right neighbor G(5,6) is determined based on their coordinates.

[0172] S303. Based on the grid distance and the material parameters of the thermoplastic material to be heated, determine the edge weights of the first grid cell and multiple adjacent first grid cells.

[0173] Material parameters are physical quantities used to indicate the thermal or mechanical properties of a material, such as thermal conductivity. The values ​​carried by the "edges" connecting two first grid cells in the graph structure represent the strength of their interaction, such as heat transfer capacity.

[0174] For any first grid cell G(i,j), all its adjacent grid cells, such as G(i−1,j), G(i,j+1), etc., are known. Each pair constitutes a "connection relationship", which corresponds to an "edge" in the graph structure.

[0175] Specifically, for example, if G(5,5) has four neighbors, the following four edges will be generated:

[0176] G(5,5)–G(4,5);

[0177] G(5,5)–G(6,5);

[0178] G(5,5)–G(5,4);

[0179] G(5,5)–G(5,6).

[0180] The edge weights of the four edges mentioned above are determined based on the mesh distance and material parameters. Specifically, they are calculated using the following formula:

[0181]

[0182] in, Let the edge weights be those between the first grid node i and the first grid node j. This refers to the material parameters of the thermoplastic material to be tested, namely, the thermal conductivity coefficient. Let i be the spatial coordinates of the first grid node. Let j be the spatial coordinates of the first grid node. Let be the grid distance between the first grid node i and the first grid node j.

[0183] S304. Take any first grid cell as a node of the graph structure and multiple edge weights as edges of the graph structure to obtain the graph structure of the thermoplastic material to be tested.

[0184] In this process, each first grid cell is mapped to a graph node. For each node, i.e., the first grid cell, its multiple adjacent first grid cells are known. An edge is added between this node and each of its adjacent nodes in the graph. The weight of each edge reflects the thermal conductivity relationship between the two grids. By integrating all nodes and weighted edges, a complete graph is obtained, which represents the graph structure of the thermoplastic material.

[0185] This embodiment provides a method for predicting the thickness of thermoplastic materials based on the Industrial Internet. This method determines the adjacent grid cells of each first grid cell according to the positional relationships between multiple first grid cells; calculates the grid distance between any two adjacent first grid cells, i.e., the distance between their center points; determines the edge weights between a first grid cell and its adjacent first grid cells based on the grid distances and the specific material parameters of the material to be thermoplasticized, thereby reflecting the degree of mutual influence between them; and constructs a graph structure representing the entire thermoplastic material by treating each first grid cell as a node in a graph structure and using the calculated edge weights as the edges connecting these nodes. This method improves the simulation accuracy of the thermoplastic process by precisely quantifying the influence between adjacent first grid cells.

[0186] Figure 4 A flowchart illustrating a thermoplastic thickness prediction method based on the Industrial Internet provided in this application embodiment. Figure 4 .like Figure 4 As shown, in Figure 1 Based on the examples, the method for predicting thermoplastic thickness based on the Industrial Internet is described in detail, including:

[0187] S401. Obtain the process parameters of the thermoplastic material to be treated. The process parameters include: adsorption pressure, adsorption time and heating time during the thermoplastic treatment process.

[0188] Process parameters include, for example, the adsorption pressure, adsorption time, and heating time during the thermoplastic treatment of the material to be thermoplasticized.

[0189] Adsorption pressure refers to the pressure value used to adsorb the heated and softened sheet material onto the mold surface during the thermoplastic molding process.

[0190] Adsorption time refers to the time required from the start of adsorption to complete formation.

[0191] Heating time refers to the time required for the sheet material to be heated to a sufficient softening temperature on the heating plate of the thermoplastic equipment.

[0192] S402. Determine the material parameters of the thermoplastic material to be used.

[0193] The material parameters of the thermoplastic material to be used include, but are not limited to, thermal conductivity, specific heat capacity, density, and thickness.

[0194] S403. Determine the initial node characteristics of multiple first grid cells based on their material and process parameters.

[0195] Each first grid cell corresponds to a node in the graph, and its feature vector may include, but is not limited to, the following:

[0196] Location information: Coordinates (x, y) of the grid center point;

[0197] Material parameters: thermal conductivity, specific heat capacity, density, thickness;

[0198] Process parameters: adsorption pressure, adsorption time, heating time.

[0199] Based on the aforementioned process and material parameters, the nodes of each first mesh element are initialized to determine the initial node features of multiple first mesh elements. For example, the initial node features of node i are:

[0200]

[0201] in, Let i be the temperature value of node i. Let i be the thermal conductivity of the material at node i. Let i be the viscoelastic parameters of the material. Let be the adsorption pressure at node i. Let be the adsorption time of node i.

[0202] By determining the initial node features of several points in the first grid, the features of each node are converted into a vector form, which is convenient for subsequent input into the machine learning model.

[0203] This embodiment provides a method for predicting thermoplastic thickness based on the Industrial Internet. This method acquires key process parameters (including adsorption pressure, adsorption time, and heating time) and physical property parameters of the material itself (such as thermal conductivity, specific heat capacity, and glass transition temperature) during the thermoplastic processing. These two types of parameters are used together as basic information to assign initial node features to each first grid cell in the graph structure. Specifically, the initial node features of each grid cell are composed of its corresponding material parameters and globally or locally applicable process parameters. By integrating real process conditions and material properties into the initial state of each first grid cell, subsequent thermal response analysis and thickness prediction can better reflect actual processing scenarios, avoiding prediction biases caused by ignoring process differences or using uniform default parameters in traditional methods.

[0204] Figure 5 This application provides a structural schematic diagram of a thermoplastic thickness prediction device based on the Industrial Internet. Figure 5 As shown, this application provides a thermoplastic thickness prediction device based on the Industrial Internet, the thermoplastic thickness prediction device 500 based on the Industrial Internet includes:

[0205] The processing module 501 is used to divide the thermoplastic material to be heated into multiple first grid units, and to construct a graph structure of the thermoplastic material according to the positional relationship between the multiple first grid units. The nodes of the graph structure are used to represent the corresponding first grid units, and the edges of the graph structure are used to represent the positional relationship between the corresponding two first grid units.

[0206] The acquisition module 502 is used to acquire the first temperature distribution of the thermoplastic equipment, which is used to process the thermoplastic material.

[0207] The determination module 503 is used to determine the second temperature distribution corresponding to each first grid cell based on the first temperature distribution;

[0208] The processing module 501 is also used to update the node features of multiple nodes in the graph structure based on the second temperature distribution to obtain the updated graph structure, and predict the thermoplastic thickness corresponding to multiple first grid cells based on the updated graph structure, thereby obtaining the thermoplastic thickness prediction result of the material to be thermoplasticized.

[0209] Optionally, the processing module 501 is also used to perform temperature zone division processing on the thermoplastic equipment to obtain multiple second grid units corresponding to the thermoplastic equipment;

[0210] The determination module 503 is also used to determine the heat conduction mapping relationship between multiple second grid cells and the first grid cell, and to determine the second temperature distribution corresponding to each first grid cell based on the mapping relationship and the first temperature distribution.

[0211] Optionally, the determining module 503 is further configured to determine, based on the positional relationship between the multiple first grid cells, multiple adjacent first grid cells adjacent to any one first grid cell;

[0212] The determining module 503 is further configured to determine the grid distance between the first grid cell and any one of the plurality of adjacent first grid cells, wherein the grid distance is used to indicate the distance between the grid center point of the first grid cell and the grid center point of the adjacent first grid cell;

[0213] The determination module 503 is also used to determine the edge weights of the first grid cell and multiple adjacent first grid cells based on the grid distance and the material parameters of the thermoplastic material to be heated;

[0214] The processing module 501 is also used to take any first grid cell as a node of the graph structure and multiple edge weights as edges of the graph structure to obtain the graph structure of the thermoplastic material to be processed.

[0215] Optionally, the acquisition module 502 is also used to acquire the process parameters of the thermoplastic material to be thermoplasticized, including: adsorption pressure, adsorption time and heating time during the thermoplasticization process;

[0216] The determination module 503 is also used to determine the material parameters of the thermoplastic material to be used;

[0217] The determination module 503 is also used to determine the initial node characteristics of the multiple first grid cells based on the material parameters and process parameters of the multiple first grid cells.

[0218] Optionally, the determining module 503 is also used to determine the heat flux value and attention coefficient of any first grid cell and its adjacent first grid cells in the graph structure based on the graph structure;

[0219] The processing module 501 is also used to update the node features of multiple nodes in the graph structure based on the heat flux value, attention coefficient, initial node features and second temperature distribution, so as to obtain the updated graph structure.

[0220] Optionally, the processing module 501 is also used to use the updated graph structure as input data for the thermoplastic thickness prediction model and control the thermoplastic thickness prediction model to perform iterative processing; the thermoplastic thickness prediction model is based on machine learning methods and is trained according to historical thermoplastic production data and domain expert experience.

[0221] The determination module 503 is also used to determine the total loss function of the thermoplastic thickness prediction model, which includes the first loss function of the thermoplastic thickness prediction model and the second loss function of the physical constraints;

[0222] The processing module 501 is also used to control the thermoplastic thickness prediction model to stop iterative processing when the total loss function is determined to converge, and to output the predicted thickness corresponding to multiple first grid cells.

[0223] Optionally, the device may also include: a judgment module 504;

[0224] The acquisition module 502 is also used to acquire the thermoplastic thickness prediction result output by the thermoplastic thickness prediction model and the actual thickness of the thermoplastic material after thermoplastic treatment.

[0225] The determination module 503 is also used to determine the absolute value of the difference between the actual thickness and the predicted result;

[0226] The judgment module 504 is used to determine whether the absolute value of the difference is less than the error threshold.

[0227] The determination module 503 is also used to determine the convergence of the first loss function when the absolute value of the difference is less than the error threshold;

[0228] The determination module 503 is also used to determine whether the first loss function is non-convergent when the absolute value of the difference is not less than the error threshold.

[0229] The determination module 503 is also used to determine a second loss function for the thermoplastic thickness prediction model when the first loss function indicates convergence. The second loss function is used to indicate physical law constraints.

[0230] The determination module 503 is also used to determine the total loss function of the thermoplastic thickness prediction model based on the first loss function and the second loss function.

[0231] Optionally, the processing module 501 is also used to increase the adsorption pressure and / or adsorption time and / or heating time when the actual thickness is greater than the target thickness.

[0232] The processing module 501 is also used to reduce the adsorption pressure and / or adsorption time and / or heating time when the actual thickness is not greater than the target thickness.

[0233] The thermoplastic thickness prediction device based on the Industrial Internet provided in this application has a similar implementation principle and technical effect to the implementation of each part of the aforementioned thermoplastic thickness prediction method based on the Industrial Internet, and will not be described again here.

[0234] like Figure 6 As shown, this application provides an electronic device 600, which includes: a receiver 601, a transmitter 602, a processor 603, and a memory 604.

[0235] Receiver 601 is used to receive instructions and data;

[0236] Transmitter 602 is used to send commands and data;

[0237] Memory 604 is used to store instructions executed by the computer;

[0238] Processor 603 is used to execute computer execution instructions stored in memory 604 to implement the various steps of the thermoplastic thickness prediction method based on the Industrial Internet in the above embodiments. For details, please refer to the relevant descriptions in the foregoing embodiments of the thermoplastic thickness prediction method based on the Industrial Internet.

[0239] Optionally, the memory 604 can be either standalone or integrated with the processor 603.

[0240] When the memory 604 is set up independently, the electronic device also includes a bus for connecting the memory 604 and the processor 603.

[0241] The implementation principle and technical effects of the electronic device provided in this embodiment can be found in the foregoing embodiments, and will not be repeated here.

[0242] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the method of any of the foregoing embodiments.

[0243] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the method of any of the foregoing embodiments.

[0244] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0245] It should be further noted that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0246] It should be understood that the above-described device embodiments are merely illustrative, and the device of this application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units, modules, or components may be combined, or integrated into another system, or some features may be ignored or not executed.

[0247] Furthermore, unless otherwise specified, the functional units / modules in the various embodiments of this application can be integrated into one unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated together. The integrated units / modules described above can be implemented in hardware or as software program modules.

[0248] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0249] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0250] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for predicting thermoplastic thickness based on the Industrial Internet, characterized in that, The method includes: The thermoplastic material to be heated is divided into multiple first grid units, and a graph structure of the thermoplastic material is constructed according to the positional relationship between the multiple first grid units. The nodes of the graph structure are used to represent the corresponding first grid units, and the edges of the graph structure are used to represent the positional relationship between two corresponding first grid units. A first temperature distribution is obtained in a thermoplastic equipment used to process the material to be thermoplasticized. Determine the second temperature distribution corresponding to each of the first grid cells based on the first temperature distribution; Based on the second temperature distribution, the node features of multiple nodes in the graph structure are updated to obtain an updated graph structure. Based on the updated graph structure, the thermoplastic thickness of multiple first grid cells is predicted, thereby obtaining the thermoplastic thickness prediction result of the material to be thermoplasticized.

2. The method according to claim 1, characterized in that, Determining the second temperature distribution corresponding to each of the first grid cells based on the first temperature distribution includes: The thermoplastic equipment is divided into temperature zones to obtain multiple second grid units corresponding to the thermoplastic equipment; Determine the heat conduction mapping relationship between multiple second grid cells and the first grid cell, and determine the second temperature distribution corresponding to each first grid cell based on the mapping relationship and the first temperature distribution.

3. The method according to claim 1, characterized in that, The step of constructing the graph structure of the thermoplastic material based on the positional relationship between multiple first grid cells includes: Based on the positional relationship between the multiple first grid cells, determine the multiple adjacent first grid cells adjacent to any one first grid cell; Determine the grid distance between the first grid cell and any of the plurality of adjacent first grid cells, wherein the grid distance is used to indicate the distance between the grid center point of the first grid cell and the grid center point of the adjacent first grid cell; Based on the grid distance and the material parameters of the thermoplastic material to be heated, determine the edge weights of the first grid cell and the plurality of adjacent first grid cells; By using any one of the first mesh cells as a node of the graph structure and the multiple edge weights as edges of the graph structure, the graph structure of the thermoplastic material to be obtained is obtained.

4. The method according to claim 1, characterized in that, After constructing the graphical structure of the thermoplastic material to be manufactured, the method further includes: Obtain the process parameters of the thermoplastic material to be treated, including: adsorption pressure, adsorption time, and heating time during the thermoplastic treatment process; Determine the material parameters of the thermoplastic material to be used; The initial node characteristics of the multiple first grid cells are determined based on the material parameters and process parameters of the multiple first grid cells.

5. The method according to claim 4, characterized in that, The step of updating the node features of multiple nodes within the graph structure based on the second temperature distribution to obtain the updated graph structure includes: Based on the graph structure, determine the heat flux value and attention coefficient of any first grid cell and its adjacent first grid cells in the graph structure; Based on the heat flux value, the attention coefficient, the initial node features, and the second temperature distribution, the node features of multiple nodes within the graph structure are updated to obtain the updated graph structure.

6. The method according to claim 1, characterized in that, The step of predicting the thermoplastic thickness corresponding to multiple first grid cells based on the updated graph structure includes: The updated graph structure is used as input data for the thermoplastic thickness prediction model, which is then used to control the model to perform iterative processing. The thermoplastic thickness prediction model is based on machine learning methods and is trained using historical thermoplastic production data and the experience of domain experts. Determine the total loss function of the thermoplastic thickness prediction model, the total loss function including a first loss function of the thermoplastic thickness prediction model and a second loss function of physical constraints; When the total loss function is determined to converge, the thermoplastic thickness prediction model is controlled to stop iterative processing and the predicted thickness corresponding to multiple first grid cells is output.

7. The method according to claim 6, characterized in that, Determining the total loss function of the thermoplastic thickness prediction model includes: Obtain the thermoplastic thickness prediction result output by the thermoplastic thickness prediction model and the actual thickness of the material to be thermoplasticized after thermoplastic treatment; Determine the absolute value of the difference between the actual thickness and the predicted result; Determine whether the absolute value of the difference is less than the error threshold; If the absolute value of the difference is less than the error threshold, the first loss function is determined to be converged. If the absolute value of the difference is not less than the error threshold, the first loss function is determined to be non-convergent. If the first loss function indicates convergence, a second loss function is determined for the thermoplastic thickness prediction model, wherein the second loss function is used to indicate physical constraints. Based on the first loss function and the second loss function, the total loss function of the thermoplastic thickness prediction model is determined.

8. The method according to claim 7, characterized in that, The method further includes: If the actual thickness is greater than the target thickness, the adsorption pressure and / or adsorption time and / or heating time shall be adjusted upward. If the actual thickness is not greater than the target thickness, the adsorption pressure and / or adsorption time and / or heating time shall be reduced.

9. A thermoplastic thickness prediction device based on the Industrial Internet, characterized in that, The device includes: The processing module is used to divide the thermoplastic material to be heated into multiple first grid units, and to construct a graph structure of the thermoplastic material to be heated according to the positional relationship between the multiple first grid units. The nodes of the graph structure are used to represent the corresponding first grid units, and the edges of the graph structure are used to represent the positional relationship between two corresponding first grid units. An acquisition module is used to acquire the first temperature distribution of the thermoplastic equipment, the thermoplastic equipment being used to process the material to be thermoplasticized; The determination module is used to determine the second temperature distribution corresponding to each of the first grid cells based on the first temperature distribution; The processing module is further configured to update the node features of multiple nodes within the graph structure based on the second temperature distribution to obtain an updated graph structure, and predict the thermoplastic thickness corresponding to multiple first grid cells based on the updated graph structure, thereby obtaining the thermoplastic thickness prediction result of the material to be thermoplasticized.

10. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1 to 8.

12. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 8.

Images