Model transfer methods, computing devices, and machine-readable storage media

Abstract

This application discloses a model migration method, computing device, and machine-readable storage medium, belonging to the field of digital twin technology. The model migration method includes: responding to a component configuration request from construction machinery and determining a set of component configurations for the construction machinery; selecting from a preset proxy model library based on the component configuration set to obtain all corresponding target component proxy models; determining whether the accuracy of each target component proxy model is greater than a preset accuracy; and, if the accuracy of each target component proxy model is greater than the preset accuracy, assembling all target component proxy models according to the component connection relationships of the construction machinery to obtain a complete machine proxy model of the construction machinery. When the component configuration of the construction machinery changes, the complete machine proxy model can be assembled flexibly and quickly, thereby achieving rapid migration of the proxy model and improving the deployment efficiency of the proxy model for the construction machinery.

Description

Technical Field

[0001] This application relates to the field of digital twin technology, specifically to a model transfer method, computing device, and machine-readable storage medium. Background Technology

[0002] With the rapid development of control technology for construction machinery, digital twin technology is widely used in various types of construction machinery control scenarios. A digital twin refers to mapping construction machinery in a virtual space using real-time data and modeling techniques, thereby reflecting the operational status of the corresponding physical equipment. In the control scenarios of construction machinery such as tower cranes, a complete machine proxy model of the tower crane is typically built based on digital twin technology. This proxy model is then used to predict and optimize stress in the tower crane, improving its operational reliability.

[0003] However, tower cranes and other construction machinery are typically complex systems composed of multiple interchangeable components. During operation, these machines are often configured by assembling different interchangeable components to adapt to complex construction environments. When the tower crane's configuration changes, its proxy model needs to be rebuilt, retrained, and retested. This time-consuming redeployment of the proxy model leads to inefficient deployment of such models. Summary of the Invention

[0004] The purpose of this application is to provide a model migration method, computing device, and machine-readable storage medium to solve the problem of low deployment efficiency of proxy models for engineering machinery in the prior art.

[0005] To achieve the above objectives, the first aspect of this application provides a model transfer method, which includes:

[0006] Responding to the component configuration request of the construction machinery, determine the component configuration set of the construction machinery;

[0007] Based on the component configuration set, select from the preset proxy model library to obtain all corresponding target component proxy models. The preset proxy model library stores component proxy models of engineering machinery.

[0008] Determine whether the accuracy of the proxy model for each target component is greater than the preset accuracy;

[0009] When the accuracy of each target component proxy model is greater than the preset accuracy, all target component proxy models are assembled according to the component connection relationship of the construction machinery to obtain the whole machine proxy model of the construction machinery.

[0010] In the embodiments of this application, each target component proxy model includes at least one sub-network;

[0011] Model transfer methods also include:

[0012] If the accuracy of at least one target component proxy model is less than or equal to a preset accuracy, for each target component proxy model whose accuracy is less than or equal to the preset accuracy, a target component proxy model is trained based on the accuracy error of each sub-network in the target component proxy model until the accuracy of the target component proxy model is greater than the preset accuracy.

[0013] In the embodiments of this application, training the target component proxy model based on the accuracy error of each sub-network in the target component proxy model includes:

[0014] The accuracy error of each sub-network in the target component proxy model is obtained, and it is determined whether the accuracy error of each sub-network is greater than the preset error.

[0015] If the accuracy error of each sub-network is greater than the preset error, determine the deviation of the accuracy error of all sub-networks;

[0016] Based on the deviation, train the target component proxy model.

[0017] In the embodiments of this application, each target component proxy model further includes a summary network, and each sub-network is connected to the summary network;

[0018] Based on the deviation, train the target component surrogate model, including:

[0019] Training the summation network when the bias is linear;

[0020] If the bias is not linear, train all subnetworks and then train the sum network.

[0021] In the embodiments of this application, training the target component proxy model based on the accuracy error of each sub-network in the target component proxy model includes:

[0022] If the accuracy error of at least one subnetwork is greater than or equal to a preset error, and the number of subnetworks with accuracy errors less than the preset error is greater than the number corresponding to a preset percentage, then train the subnetworks with accuracy errors greater than the preset error, where the preset percentage is the percentage between the number of subnetworks with accuracy errors less than the preset error and the total number of subnetworks.

[0023] In the embodiments of this application, the preset proxy model library is obtained according to the following steps:

[0024] Identify all replaceable parts for the construction machinery;

[0025] Construct a proxy model for each replaceable component;

[0026] A proxy model library is built based on the proxy models of all replaceable parts.

[0027] In the embodiments of this application, constructing a proxy model for each replaceable component includes:

[0028] The load between the replaceable component and the first component is used as the input to the surrogate model, and the load and stress between the replaceable component and the second component are used as the output of the surrogate model to construct a surrogate model for the replaceable component, wherein the first component is connected to the second component through the replaceable component.

[0029] In the embodiments of this application, proxy models for all target components are selected from a preset proxy model library based on the component configuration set, including:

[0030] Obtain the initial component configuration set of the construction machinery, and the initial whole machine proxy model corresponding to the initial component configuration set;

[0031] Based on the component configuration set and the initial component configuration set, determine the components to be added and the components to be removed;

[0032] Select the proxy model of the component to be added from the proxy module library, and subtract the proxy model of the component to be subtracted from the initial whole machine proxy model to obtain all target component proxy models corresponding to the component configuration set.

[0033] A second aspect of this application provides a computing device, comprising:

[0034] The memory is configured to store instructions;

[0035] The processor is configured to retrieve instructions from memory and, when executing instructions, to implement the aforementioned model transfer method.

[0036] A third aspect of this application provides a machine-readable storage medium storing instructions that cause a machine to perform the model transfer method described above.

[0037] This application provides a model migration method, comprising: responding to a component configuration request from construction machinery and determining a set of component configurations for the construction machinery; selecting from a preset proxy model library based on the component configuration set to obtain all corresponding target component proxy models; determining whether the accuracy of each target component proxy model is greater than a preset accuracy; and, if the accuracy of each target component proxy model is greater than the preset accuracy, assembling all target component proxy models according to the component connection relationships of the construction machinery to obtain a complete machine proxy model of the construction machinery. This constructs a complete machine proxy model of construction machinery that is component-based and assemblable. When the component configuration of the construction machinery changes, the complete machine proxy model can be assembled flexibly and quickly, thereby achieving rapid migration of the proxy model. The migration efficiency of proxy models for construction machinery with different configurations is improved, eliminating the need for significant time to redeploy new complete machine proxy models, thus improving the deployment efficiency of proxy models for construction machinery.

[0038] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0039] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings:

[0040] Figure 1 The illustration shows a schematic flowchart of a model transfer method according to an embodiment of this application;

[0041] Figure 2 An example diagram illustrating the accuracy of a proxy model according to an embodiment of this application is shown schematically.

[0042] Figure 3 An example diagram illustrating a whole-machine agent model according to an embodiment of this application is shown schematically. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0044] It should be noted that if the embodiments of this application involve directional indication, the directional indication is only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0045] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0046] Figure 1 The illustration shows a schematic flowchart of a model transfer method according to an embodiment of this application. Figure 1 As shown in the figure, this application provides a model transfer method, which may include the following steps:

[0047] S110, responding to the component configuration request of the construction machinery, determines the component configuration set of the construction machinery.

[0048] The type of construction machinery is determined based on actual needs and is not limited here. For ease of understanding, the construction machinery in this embodiment is a tower crane. When the configuration of the tower crane changes, such as when it is necessary to add or remove components, the computing device responds to the component configuration request of the construction machinery and determines the new component configuration set corresponding to the construction machinery.

[0049] The method in the embodiments of this application selects from a preset proxy model library based on a component configuration set to obtain all corresponding target component proxy models, including:

[0050] Obtain the initial component configuration set of the construction machinery, and the initial whole machine proxy model corresponding to the initial component configuration set;

[0051] Based on the component configuration set and the initial component configuration set, determine the components to be added and the components to be removed;

[0052] Select the proxy model of the component to be added from the proxy module library, and subtract the proxy model of the component to be subtracted from the initial whole machine proxy model to obtain all target component proxy models corresponding to the component configuration set.

[0053] When the component configuration of construction machinery changes, it may be necessary to add or remove components. The process involves obtaining the initial component configuration set of the construction machinery, and the corresponding initial whole-machine proxy model. Based on the component configuration set and the initial component configuration set, the components to be added and removed are determined. Proxy models for the components to be added are selected from the proxy module library, and the proxy models for the components to be removed are subtracted from the initial whole-machine proxy model. Based on the component configuration set, proxy models for the new components are added, and proxy models for redundant components in the initial component configuration are subtracted, quickly obtaining all target component proxy models corresponding to the component configuration set.

[0054] S120: Select from the preset proxy model library according to the component configuration set to obtain all corresponding target component proxy models. The preset proxy model library stores component proxy models of engineering machinery.

[0055] Using tower crane components as units, pre-constructed proxy models for various common tower crane components are created, forming a pre-defined proxy model database. Based on the component configuration set, proxy models for all corresponding target components are selected from this database. Specifically, the target components that need to be added and / or removed from the tower crane are determined based on the component configuration set. When a target component needs to be removed, its corresponding proxy model is directly removed. When a target component needs to be added, it is selected from the pre-defined proxy model database to obtain all corresponding target component proxy models.

[0056] S130, determine whether the accuracy of the proxy model for each target component is greater than the preset accuracy.

[0057] Determine whether the accuracy of each target component proxy model is greater than the preset accuracy. If the accuracy of the target component proxy model is less than or equal to the preset accuracy, then the target component proxy model needs to be recalibrated until the accuracy of each target component proxy model is greater than the preset accuracy.

[0058] S140: When the accuracy of each target component proxy model is greater than the preset accuracy, all target component proxy models are assembled according to the component connection relationship of the construction machinery to obtain the whole machine proxy model of the construction machinery.

[0059] With the accuracy of each target component proxy model exceeding the preset accuracy, all target component proxy models are assembled based on the component connection relationships of the construction machinery, i.e., based on the actual topological connection relationships of the tower crane, to obtain the overall proxy model of the construction machinery. Using the overall proxy model, parameters such as stress in each key component of the construction machinery can be predicted. It is important to understand that key components are those with high load-bearing capacity and relatively weak fatigue life.

[0060] A complete proxy model for construction machinery is constructed, using components as units and capable of assembly. When the component configuration of the construction machinery changes, the complete proxy model can be assembled flexibly and quickly, thereby achieving rapid migration of the proxy model. The migration efficiency of proxy models for construction machinery with different configurations is improved, eliminating the need for spending a lot of time redeploying new complete proxy models, thus improving the deployment efficiency of proxy models for construction machinery.

[0061] In the embodiments of this application, each target component proxy model includes at least one sub-network;

[0062] Model transfer methods also include:

[0063] If the accuracy of at least one target component proxy model is less than or equal to a preset accuracy, for each target component proxy model whose accuracy is less than or equal to the preset accuracy, a target component proxy model is trained based on the accuracy error of each sub-network in the target component proxy model until the accuracy of the target component proxy model is greater than the preset accuracy.

[0064] In this embodiment, each target component proxy model includes at least one network, and each sub-network corresponds to a sub-model. For components that require load output, the target component proxy model includes at least two networks, one of which is used to output the load, and the remaining at least one network is used to output stress and / or displacement data. For example, the foundation section of a tower crane is the lowest component. Since the foundation section does not have any next-level connecting components, the proxy model for the foundation section does not need to output the load. For components that do not require load output, the target component proxy model includes at least one network, excluding the network used to output the load. The proxy model includes at least one network used to output stress and / or displacement data. When the accuracy error of each sub-network in the target proxy model is small, the target proxy model can be used directly. When the accuracy of at least one target component proxy model is less than or equal to a preset accuracy, the target component proxy model with an accuracy less than or equal to the preset accuracy is a proxy model that needs to be recalibrated.

[0065] Obtain the accuracy error of each sub-network in the target component proxy model whose accuracy is less than or equal to the preset accuracy. For each target component proxy model whose accuracy is less than or equal to the preset accuracy, train the target component proxy model based on the accuracy error of each sub-network in the target component proxy model until the accuracy of the target component proxy model is greater than the preset accuracy.

[0066] In the embodiments of this application, the target component proxy model is trained based on the accuracy error of each sub-network in the target component proxy model, including:

[0067] The accuracy error of each sub-network in the target component proxy model is obtained, and it is determined whether the accuracy error of each sub-network is greater than the preset error.

[0068] If the accuracy error of each sub-network is greater than the preset error, determine the deviation of the accuracy error of all sub-networks;

[0069] Based on the deviation, train the target component proxy model.

[0070] For each target component proxy model with an accuracy less than or equal to a preset accuracy, the accuracy error of each node in each sub-network of the target component proxy model is obtained, and it is determined whether the accuracy error of each sub-network is greater than the preset error, so as to determine the error category of different proxy models.

[0071] The surrogate model is recalibrated based on its error category to ensure that the accuracy of all target component surrogate models exceeds a preset accuracy. Specifically, when the accuracy error of each sub-network is greater than the preset error, the deviation of the accuracy error of all sub-networks is determined. Based on the deviation, it is determined whether the overall prediction accuracy of each network output node maintains approximately the same error level, and then the target component surrogate model is trained to recalibrate the surrogate model.

[0072] In the embodiments of this application, each target component proxy model further includes a summary network, and each sub-network is connected to the summary network;

[0073] Based on the deviation, train the target component surrogate model, including:

[0074] Training the summation network when the bias is linear;

[0075] If the bias is not linear, train all subnetworks and then train the sum network.

[0076] In this embodiment, each target component proxy model also includes a summary network. The output layer of each sub-network is connected to the input layer of the summary network. The input of each sub-network is the load at the connection point of the previous component. The output of the summary network is the key node parameters of the target component. The key nodes are the nodes with relatively large load-bearing capacity and the nodes with relatively weak fatigue life in the component.

[0077] When the bias is linear, it is determined that the overall accuracy error of each network output node remains at approximately the same level, and the aggregate network is trained without retraining the sub-networks. When the bias is non-linear, it is determined that the accuracy error of each network output node is relatively poor, and the accuracy errors vary in size. All sub-networks are trained, and the aggregate network is trained as well.

[0078] In the embodiments of this application, training the target component proxy model based on the accuracy error of each sub-network in the target component proxy model includes:

[0079] If the accuracy error of at least one subnetwork is greater than or equal to a preset error, and the number of subnetworks with accuracy errors less than the preset error is greater than the number corresponding to a preset percentage, then train the subnetworks with accuracy errors greater than the preset error, where the preset percentage is the percentage between the number of subnetworks with accuracy errors less than the preset error and the total number of subnetworks.

[0080] The preset percentage is used to determine whether a large number of subnetworks have low accuracy errors. If the accuracy error of at least one subnetwork is greater than or equal to the preset error, and the number of subnetworks with accuracy errors less than the preset error is greater than the number corresponding to the preset percentage, it is determined that only some website nodes have high accuracy errors. In this case, only the subnetworks with accuracy errors greater than the preset error are trained, and it is not necessary to train all subnetworks.

[0081] Please see Figure 2 , Figure 2 An example diagram illustrating the accuracy of a proxy model according to an embodiment of this application is shown schematically.

[0082] As shown in the figure, 15 target component proxy models are obtained in the embodiments of this application, and the names of the 15 proxy models are: QZB1, QZB2, QZB3, QZB4, QZB4, QZB5, QZB5, QZB6, QZB7, PHB1, PHB2, PHB3, TT, SZZ, XZZ, BZJ, and GDJ. The accuracy of the proxy model is inversely proportional to its error; that is, the higher the accuracy of the proxy model, the lower its error.

[0083] The surrogate models named QZB6, QZB3, PHB3, PHB2, and SZZ have relatively concentrated errors, while the errors of the remaining surrogate models are more consistent, requiring retraining of the aggregation network. Some subnetworks of the surrogate models named QZB4, QZB5, QZB1, XZZ, and BZJ need to be retrained. The surrogate model named QZB7 has an accuracy greater than, less than, or equal to the preset accuracy and does not require retraining. The surrogate models named QZB6, QZB3, PHB3, PHB2, and SZZ require retraining of the aggregation model.

[0084] In the embodiments of this application, the preset proxy model library is obtained according to the following steps:

[0085] Identify all replaceable parts for the construction machinery;

[0086] Construct a proxy model for each replaceable component;

[0087] A proxy model library is built based on the proxy models of all replaceable parts.

[0088] Construction machinery components include replaceable and non-replaceable components. This document identifies all replaceable components for the construction machinery. In this embodiment, the tower crane's replaceable components include 7 booms, 3 counterweight booms, the tower head, upper support, lower support, standard sections, and transition sections. A proxy model is constructed for each replaceable component, using the component as a unit. Based on the proxy models of all replaceable components, a proxy model library is built.

[0089] In the embodiments of this application, constructing a proxy model for each replaceable component includes:

[0090] The load between the replaceable component and the first component is used as the input to the surrogate model, and the load and stress between the replaceable component and the second component are used as the output of the surrogate model to construct a surrogate model for the replaceable component, wherein the first component is connected to the second component through the replaceable component.

[0091] The first component is connected to the second component via a replaceable component; that is, the first component is the previous component connected to the replaceable component, and the second component is the next component connected to the replaceable component. In this embodiment, the proxy model includes a load prediction proxy model and a stress prediction proxy model. The accuracy of both the load prediction proxy model and the stress prediction proxy model is set according to actual needs and is not limited here. For ease of understanding, in the embodiments of this application, the accuracy of the load prediction proxy model is within 5%, and the accuracy of the stress prediction proxy model is within 10%.

[0092] Please see Figure 3 , Figure 3 An example diagram illustrating a whole-machine agent model according to an embodiment of this application is shown schematically.

[0093] The load between the replaceable component and the first component is used as the input to the surrogate model, and the load and stress between the replaceable component and the second component are used as the output to construct the surrogate model for the replaceable component. When assembling the surrogate models for all target components, the input to the load prediction surrogate model is the load at the connection point of the previous target component, and the output of the load prediction surrogate model is the load at the connection point with the next target component, thus obtaining the overall surrogate model of the engineering machinery. As shown in the figure, the overall surrogate model is obtained by assembling the surrogate models of boom 2, boom 1, counterweight boom 1, tower head, upper support, and standard section.

[0094] This application provides a model migration method, comprising: responding to a component configuration request from construction machinery and determining a set of component configurations for the construction machinery; selecting from a preset proxy model library based on the component configuration set to obtain all corresponding target component proxy models; determining whether the accuracy of each target component proxy model is greater than a preset accuracy; and, if the accuracy of each target component proxy model is greater than the preset accuracy, assembling all target component proxy models according to the component connection relationships of the construction machinery to obtain a complete machine proxy model of the construction machinery. This constructs a complete machine proxy model of construction machinery that is component-based and assemblable. When the component configuration of the construction machinery changes, the complete machine proxy model can be assembled flexibly and quickly, thereby achieving rapid migration of the proxy model. The migration efficiency of proxy models for construction machinery with different configurations is improved, eliminating the need for significant time to redeploy new complete machine proxy models, thus improving the deployment efficiency of proxy models for construction machinery.

[0095] This application embodiment also provides a computing device, including:

[0096] The memory is configured to store instructions;

[0097] The processor is configured to retrieve instructions from memory and, when executing instructions, to implement the aforementioned model transfer method.

[0098] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured; adjusting kernel parameters can address the problem of low deployment efficiency of proxy models in engineering machinery.

[0099] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0100] This application also provides a machine-readable storage medium storing instructions that cause a machine to perform the model transfer method described above.

[0101] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0102] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0103] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0104] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0105] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0106] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0107] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0108] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0109] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A model transfer method, characterized in that, The model transfer method includes: In response to a component configuration request from the construction machinery, determine the component configuration set of the construction machinery; Based on the component configuration set, a selection is made from the preset proxy model library to obtain all corresponding target component proxy models. The preset proxy model library stores component proxy models of the engineering machinery. Each target component proxy model includes a summary network and at least one sub-network, and each sub-network is connected to the summary network. Determine whether the accuracy of each target component proxy model is greater than the preset accuracy; If the accuracy of each target component proxy model is greater than the preset accuracy, all the target component proxy models are assembled according to the component connection relationship of the engineering machinery to obtain the whole machine proxy model of the engineering machinery. The model transfer method also includes: If the accuracy of at least one of the target component proxy models is less than or equal to a preset accuracy, for each target component proxy model whose accuracy is less than or equal to the preset accuracy, the target component proxy model is trained based on the accuracy error of each sub-network in the target component proxy model until the accuracy of the target component proxy model is greater than the preset accuracy. Training the target component proxy model based on the accuracy error of each sub-network in the target component proxy model includes: The accuracy error of each sub-network in the target component proxy model is obtained respectively, and it is determined whether the accuracy error of each sub-network is greater than a preset error. If the accuracy error of each sub-network is greater than the preset error, the deviation of the accuracy error of all sub-networks is determined. Based on the deviation, train the target component proxy model; The step of training the target component proxy model based on the deviation includes: The summarization network is trained when the deviation is linear. If the deviation is not linear, train all the sub-networks and train the summation network.

2. The method according to claim 1, characterized in that, Training the target component proxy model based on the accuracy error of each sub-network in the target component proxy model includes: If the accuracy error of at least one of the sub-networks is greater than or equal to a preset error, and the number of sub-networks with accuracy errors less than the preset error is greater than the number corresponding to a preset percentage, then the sub-networks with accuracy errors greater than the preset error are trained, wherein the preset percentage is the percentage between the number of sub-networks with accuracy errors less than the preset error and the total number of sub-networks.

3. The method according to claim 1, characterized in that, The preset proxy model library is obtained according to the following steps: Identify all replaceable parts for the construction machinery; Construct a proxy model for each of the replaceable components; A proxy model library is constructed based on the proxy models of all the aforementioned replaceable parts.

4. The method according to claim 3, characterized in that, The process of constructing a proxy model for each of the replaceable components includes: The load between the replaceable component and the first component is used as the input to the surrogate model, and the load and stress between the replaceable component and the second component are used as the output of the surrogate model to construct a surrogate model for the replaceable component, wherein the first component is connected to the second component through the replaceable component.

5. The method according to claim 1, characterized in that, The step of selecting from a preset proxy model library based on the component configuration set to obtain all corresponding target component proxy models includes: Obtain the initial component configuration set of the construction machinery, and the initial whole machine proxy model corresponding to the initial component configuration set; Based on the component configuration set and the initial component configuration set, determine the components to be added and the components to be removed; Select the proxy model of the component to be added from the proxy module library, and subtract the proxy model of the component to be subtracted from the initial whole machine proxy model to obtain all target component proxy models corresponding to the component configuration set.

6. A computing device, characterized in that, include: The memory is configured to store instructions; A processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the model migration method according to any one of claims 1 to 5.

7. A machine-readable storage medium, characterized in that, The machine-readable storage medium stores instructions for causing the machine to perform the model transfer method according to any one of claims 1 to 5.

Images