Method, device, equipment, medium and product for finding force of cable-strut tension structure

Abstract

The application discloses a cable-strut tension structure force finding method, device, equipment, medium and product, and relates to the field of computer-aided design. The method comprises the following steps: analyzing geometric topology information of a cable-strut structure from a CAD file, and constructing graph network data according to the geometric topology information, wherein the graph network data comprises an association matrix; wherein the nodes in the graph network data comprise strut nodes and boundary nodes, and the edges comprise strut units and cable units; inputting the association matrix into a pre-trained force distribution prediction model to obtain unit force prediction values corresponding to the strut units and the cable units; wherein a loss function for training the force distribution prediction model is constructed based on physical constraints, and comprises a force distribution constraint loss and a force balance residual loss. The above scheme adopts the concept of topological and physical double driving, greatly improves the force finding accuracy of high-dimensional complex topological structures, enhances the universality and general applicability, and improves the automation degree.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer-aided design, in particular to a force-finding method, device, equipment, medium and product for a cable-strut tension structure. BACKGROUND

[0002] As the core form of large-span space structures, the stiffness of prestressed cable-strut structures (such as cable domes and tensegrity structures) is completely maintained by the self-balancing prestress between cables (tension) and struts (compression). Force-finding analysis is a key step in the design of such structures, aiming to determine a reasonable prestress distribution to make the structure meet the self-balancing condition under zero external load.

[0003] Traditional force-finding analysis methods, including nonlinear finite element method, dynamic relaxation method, singular value decomposition method, and heuristic algorithm, require manual grouping considering symmetry conditions, and need to preset prestress initial values, with complex iterative calculation and convergence conditions. Data-driven neural networks, such as BPNN (Back-Propagation Neural Network) and RBFNN (Radial Basis Function Neural Network), rely on a large amount of finite element simulation data for training, which is costly to obtain in practical engineering, and has weak generalization ability for new structural topological forms not trained. The new neural network architecture, Physics-Informed Neural Networks (PINNs), although can embed the balance equation into the loss function, reducing data dependence, but fails to explicitly express the topological relationship of the structure, as well as the characteristics of nodes and elements (node position, element length, etc.), and the establishment of constraints requires pre-processing for manual grouping, with low convergence accuracy when dealing with high-dimensional complex topological structures. SUMMARY

[0004] Therefore, it is necessary to provide a force-finding method, device, equipment, medium and product for a cable-strut tension structure to improve force-finding accuracy and automation, and enhance universality.

[0005] In a first aspect, the present application provides a force-finding method for a cable-strut tension structure. The method comprises:

[0006] parsing geometric topological information of the cable-strut structure from a CAD file, and constructing graph network data according to the geometric topological information, the graph network data comprising an incidence matrix; wherein the nodes in the graph network data comprise strut nodes and boundary nodes, and the edges comprise strut elements and cable elements;

[0007] input the association matrix into a pre-trained force distribution prediction model to obtain unit force prediction values corresponding to the strut unit and the cable unit;

[0008] The loss function for training the force distribution prediction model is constructed based on physical constraints, including force distribution constraint loss and force balance residual loss.

[0009] The force distribution constraint loss takes the variance of the unit force prediction values corresponding to the edges of the same category as the force distribution constraint loss. The edge category is divided according to the distance from the edge midpoint to the coordinate origin

[0010] The force balance residual loss obtains the node resultant force based on the association matrix and the unit force prediction value, extracts the resultant force residual of the strut node in the three coordinate axes direction of the three-dimensional coordinate system corresponding to the geometric topology information, and takes the L2 norm of the resultant force residual as the force balance residual loss in the three coordinate axes direction.

[0011] In one embodiment, the geometric topology information includes node coordinates, node categories, node connection relationships, and edge categories.

[0012] The node categories include strut nodes and boundary nodes, and the strut nodes include strut upper nodes and strut lower nodes corresponding to the upper and lower ends of the same strut unit, respectively.

[0013] In one embodiment, constructing the graph network data according to the geometric topology information includes:

[0014] Converting the node coordinates into a floating-point tensor;

[0015] Obtaining the edge lengths of each strut unit and each cable unit according to the floating-point tensor;

[0016] Obtaining the center point coordinates of each edge according to the node coordinates, obtaining the Euclidean distance from the center point coordinates to the coordinate origin, dividing the edges with the same Euclidean distance into the same edge category, and constructing a mapping relationship between the edge category and the edge;

[0017] Encoding the nodes according to the node categories to obtain node category labels including strut upper nodes, strut lower nodes, and boundary nodes;

[0018] Encoding the relationship between the edge and the node into a matrix form to obtain the association matrix;

[0019] Generating the graph network data according to the node coordinates, the node category labels, the node connection relationships, the edge categories, the edge lengths, and the association matrix.

[0020] In one embodiment, the force distribution prediction model sequentially includes:

[0021] An input layer for receiving the association matrix and flattening the association matrix into a one-dimensional vector;

[0022] a plurality of serial hidden layers for receiving a one-dimensional vector for processing; the hidden layers are connected through a nonlinear activation function;

[0023] an output layer for outputting a unit force prediction value, the output layer comprising a Softplus activation function.

[0024] In one embodiment, the Softplus activation function applies a negative Softplus activation to the hidden layer output corresponding to the strut unit, and applies a Softplus activation to the hidden layer output corresponding to the cable unit, followed by a linear transformation of 1.

[0025] In one embodiment, the unit force prediction value is visualized to obtain a three-dimensional structure topology diagram with a prestress distribution cloud diagram, and the graph network data is bound in the three-dimensional structure topology diagram.

[0026] In a second aspect, the present application also provides a cable-strut tension structure force finding device. The device comprises:

[0027] a data extraction module for parsing geometric topology information of a cable-strut structure from a CAD file, constructing graph network data according to the geometric topology information, and generating an associated matrix of the graph network data; wherein the nodes in the graph network data include strut nodes and boundary nodes, and the edges include strut units and cable units;

[0028] a prediction output module for inputting the associated matrix into a pre-trained force distribution prediction model to obtain unit force prediction values corresponding to the strut units and the cable units;

[0029] wherein the loss function for training the force distribution prediction model is constructed based on physical constraints, including a force distribution constraint loss and a force balance residual loss;

[0030] The force distribution constraint loss takes the variance of the unit force prediction values of edges of the same category as the force distribution constraint loss; the edge categories are divided according to the distance from the edge midpoint to the coordinate origin

[0031] The force balance residual loss obtains the node resultant force based on the associated matrix and the unit force prediction values, extracts the resultant force residual of the strut nodes in the three coordinate axis directions of the three-dimensional coordinate system corresponding to the geometric topology information, and takes the L2 norm of the resultant force residual as the force balance residual loss in the three coordinate axis directions.

[0032] In a third aspect, the present application also provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps in the above cable-strut tension structure force finding method when executing the computer program.

[0033] In a fourth aspect, the present application also provides a computer readable storage medium. The computer readable storage medium has a computer program stored thereon, and the computer program, when executed by a processor, implements the steps in the cable-strut tension structure force finding method.

[0034] In a fifth aspect, the present application also provides a computer program product. The computer program product comprises a computer program, and the computer program, when executed by a processor, implements the steps in the cable-strut tension structure force finding method.

[0035] The cable-strut tension structure force finding method, device, equipment, medium and product described above, the method comprises: parsing geometric topology information of a cable-strut structure from a CAD file, constructing graph network data according to the geometric topology information, and the graph network data comprises an association matrix; wherein the nodes in the graph network data comprise strut nodes and boundary nodes, and the edges comprise strut elements and cable elements; inputting the association matrix into a pre-trained force distribution prediction model to obtain element force prediction values corresponding to the strut elements and the cable elements; wherein a loss function for training the force distribution prediction model is constructed based on physical constraints, comprising a force distribution constraint loss and a force balance residual loss; the force distribution constraint loss takes the variance of the element force prediction values corresponding to edges of the same category as the force distribution constraint loss; the force balance residual loss obtains node resultant forces based on the association matrix and the element force prediction values, extracts force residual errors of the strut nodes in three coordinate axis directions of a three-dimensional coordinate system corresponding to the geometric topology information, and takes the L2 norm of the force residual errors as the force balance residual loss in the three coordinate axis directions. The above scheme adopts the concept of topological and physical double driving, proposes a GNN-PINN (graph neural network-physical information neural network) hybrid architecture applied to cable-strut tension structure force finding, greatly improves the force finding accuracy of high-dimensional complex topological structures, enhances the universality and general applicability, and because the force distribution prediction model can automatically process the input and generate the prestress distribution based on the parameters and algorithms learned in the pre-training after obtaining the input data, manual intervention is reduced and the automation degree is improved. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 Flowchart of a cable-strut tension structure force finding method in one embodiment Figure 1 ;

[0037] Figure 2 Flowchart of a cable-strut tension structure force finding method in one embodiment Figure 2 ;

[0038] Figure 3 Three-dimensional topological graph of a sunflower-type spoke-type cable truss structure in one embodiment

[0039] Figure 4 Three-dimensional topological graph of a Kiewitt cable dome structure in one embodiment

[0040] Figure 5 Structure diagram of force finding device for cable-strut tension structure in one embodiment;

[0041] Figure 6 Internal structure diagram of computer device in one embodiment. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0043] The embodiment of the present application provides a cable-strut tension structure force finding method, as shown in Figure 1 and 2 , comprising the following steps:

[0044] Step 102, parse the geometric topology information of the cable-strut structure from the CAD file, and construct graph network data according to the geometric topology information, wherein the graph network data includes an association matrix; wherein the nodes in the graph network data include strut nodes and boundary nodes, and the edges include strut elements and cable elements.

[0045] Import the three-dimensional cable-strut structure CAD file, perform CAD geometric analysis and topology extraction. Read the line segment, i.e. edge, from the CAD file, extract the start point coordinates and end point coordinates of the line segment, i.e. node coordinates. Determine the geometric center of the three-dimensional cable-strut structure according to the extracted node coordinates, if the coordinates (x i , y i , z i ) of each node are known, the coordinates (Cx, Cy, Cz) of the geometric center can be calculated by the following formula:

[0046]

[0047]

[0048]

[0049] Wherein, N represents the number of nodes. Through coordinate translation transformation, transform the coordinate system of the original node coordinates to a new coordinate system with the geometric center coordinates as the reference, and the specific operation is to subtract the center point coordinates from all node coordinates to update the node coordinates. The subsequent operations are all based on the updated node coordinates.

[0050] In the extraction of nodes and coordinates, key elements need to be identified, including strut elements, strut nodes and boundary nodes. First, identify the line segment that satisfies |x1-x2|<0.001 and |y1-y2|<0.001, while the z coordinates are different, which can be regarded as a strut element approximately vertically arranged, and the upper and lower ends of the strut element are the upper node and the lower node respectively. Through the two end nodes of the strut element, the node with the larger z coordinate value is determined as the upper node of the strut, and the node with the smaller z coordinate value is determined as the lower node of the strut. In addition to the strut nodes, the remaining nodes are boundary nodes. Boundary nodes are subject to constraints and do not need to calculate node balance. In addition to strut elements, the remaining line segments are cable elements.

[0051] After the above operation, the geometric topological information of the cable-strut structure is obtained, including node categories (boundary nodes, strut upper nodes, strut lower nodes), node coordinates, and node connection relationships.

[0052] Next, the geometric topological information is used to construct a graph network data. The specific operation includes:

[0053] 1. Coordinate tensorization: convert node coordinates to floating point tensors, mathematically represented as: , represents the coordinate vector of the node in the x, y and z axis directions, and N is the number of nodes;

[0054] 2. Element attribute identification: edge length calculation, the calculation method uses the Euclidean distance formula: , where is the edge length, and are the i and j node coordinates of the two ends of the edge and ;

[0055] 3. Edge category coding: take the center point of each edge, calculate the L2 norm of the center point to the coordinate origin, that is, get the Euclidean distance from the center point of each edge to the coordinate origin, and classify edges with the same Euclidean distance into the same category, and establish a mapping dictionary from each edge category set to the edge category name, such as {"k1":[0,3,5], "k2":[1,4,8],...}, where k1 and k2 are the edge category names corresponding to the edge category, and [0,3,5] and [1,4,8] are the edge number sets in the corresponding k1 type and k2 type; the edge category name is used as the edge type identifier to define the edge category; edge_stype: put the indices of the same type of edge in a set;

[0056] After edge category coding, struts and cables are generally classified into different edge categories. Based on the overall structure of the cable-stayed structure, edges with different positional characteristics are classified into different types. The accompanying effect of this process is to automatically distinguish strut units and cable units. Edges of the same category exhibit mechanical similarity. This coding method is applicable to both symmetrical and asymmetrical cable-stayed structures. The difference is that asymmetrical cable-stayed structures have a larger number of types.

[0057] 4. Node type coding: Definition 0: Boundary node (fixed constraint); 1: Node on the strut (free node); 2: Node below the strut (free node).

[0058] 5. Construction of the association matrix: Encode the node-edge connection relationships into matrix form and build the association matrix. Correlation matrix The row and number The elements of a column are denoted as , representing a node and edge The connection relationship between nodes. If nodes It is the edge The starting point, then =1, if node It is the edge The endpoint, then =-1, if node With edge If not connected, then =0.

[0059]

[0060] The correlation matrix is ​​used for subsequent topological feature identification and node force balance conditions.

[0061] The above data is automatically encapsulated to output a graph network data object that conforms to the PyG (PyTorch Geometric) standard, which includes: geometric features (node ​​coordinates), topological features (node ​​connection relationships), physical features (edge ​​lengths), conditional features (edge / node types), and mechanical features (incidence matrix).

[0062] Step 104: Input the correlation matrix into the pre-trained force distribution prediction model to obtain the predicted values ​​of the unit forces corresponding to the strut unit and the cable unit.

[0063] The force distribution prediction model includes, in order:

[0064] The input layer receives the correlation matrix and flattens it into a one-dimensional vector.

[0065] Several cascaded hidden layers are used to receive one-dimensional vectors for processing; the hidden layers are connected by non-linear activation functions; the function of the hidden layers is to extract and transform features from the input data; the neurons in the hidden layers are interconnected with all neurons in the previous and next layers; each hidden layer can learn features at different levels and abstractions.

[0066] The output layer is used to output the predicted unit force values. The output layer includes the Softplus activation function.

[0067] When constructing the force distribution prediction model, a multi-layer fully connected neural network is first established. The flattened vector of the correlation matrix (dimension: number of nodes N × number of edges M) is used as the input layer feature of the cable-stayed structure topology. Each hidden layer in the fully connected neural network has the same dimension, and a non-linear activation function is used between adjacent hidden layers. Connection, output layer output element force prediction value .because It contains tension cable forces and compression strut forces; therefore, the linear transformation of the nonlinear activation function of the network output using the Softplus activation function with classification labels is as follows:

[0068]

[0069] In the formula, It is the first Predicted element force values ​​for the strip edge. It is a set of strut element indices.

[0070] A negative Softplus activation is applied to the strut elements to ensure that the internal force is negative (i.e., compression); a linear transformation of Softplus activation plus 1 is applied to the non-strut elements to ensure that the minimum cable force is not less than 1. This set of linear transformations of the Softplus activation function is a key innovation that strictly follows the mechanical principle of "struts are under compression, cable elements are under tension," enabling the output of the force distribution prediction model to distinguish between cable tension and strut compression in terms of data representation.

[0071] During training, the loss function of the force distribution prediction model is constructed based on physical constraints. The loss function includes force distribution constraint loss and force balance residual loss. The force distribution constraint loss uses the variance of the predicted force values ​​of the corresponding units of the same type of edge as the force distribution constraint loss. The force balance residual loss obtains the resultant force of the nodes based on the correlation matrix and the predicted force values ​​of the units, extracts the resultant force residual of the strut nodes in the three coordinate axis directions of the three-dimensional coordinate system corresponding to the geometric topology information, and uses the L2 norm of the resultant force residual as the force balance residual loss in the three coordinate axis directions.

[0072] The force distribution constraint loss is based on the physical law that the internal forces of the same type of edge should be as close as possible. The variance of the internal force of the same type of edge is calculated as the force distribution constraint loss to punish the mechanical discretization. The specific formula is as follows:

[0073] The average of the category edge is calculated as:

[0074]

[0075] The variance of the category edge is calculated as:

[0076]

[0077] The force distribution constraint loss is calculated as:

[0078]

[0079] In the formula, is the number of edges in the kth category, is the number of edge categories.

[0080] The force balance residual loss is constructed from the perspective of node static balance. The force balance residual loss is based on the incidence matrix and the force vector The combined force of the node is calculated, and the combined force residual of the free node (the node on the strut and the node below the strut) in the x, y and z axis directions is extracted. The L2 norm of the combined force residual is taken as the loss term of the force balance residual in the x, y and z directions.

[0081] The calculation formula of the force balance residual loss is as follows:

[0082] The force vector of the th edge is calculated as:

[0083]

[0084] In the formula, and are the starting coordinate vector and the ending coordinate vector of the th edge, respectively. The combined force matrix of all nodes is calculated as:

[0085]

[0086] In the formula, is the N×M incidence matrix, is the vector set of the force vector of all edges, ​​, M represents the total number of edges. The force residual of the free nodes (strut upper nodes and strut lower nodes) in the x, y and z axis directions is as follows:

[0087]

[0088]

[0089]

[0090] wherein, is the set of index numbers of the free nodes (strut upper nodes and strut lower nodes), represents the force matrix of the free nodes, represents the force matrix in the x axis direction, represents the force matrix in the y axis direction, represents the force matrix in the z axis direction. The force balance residual loss function is established as follows:

[0091]

[0092] The total loss function is:

[0093] T_loss=

[0094] wherein, is the force balance residual loss weight, is the force distribution constraint loss weight. The composite loss function makes the strut structure satisfy the node static force balance condition and the uniformity of the internal force distribution in the strut element at the same time.

[0095] Before training the force distribution prediction model, initialize the call graph network data, the total loss function and the optimizer, create an Adam optimizer with adaptive learning rate, and set the initial learning rate to 0.001. Start to perform iterative training, first input the incidence matrix as the input feature into the neural network, pass through an input layer, multiple hidden layers and an output layer, and output the unit force prediction value ; then input into the total loss function T_loss to calculate the total loss function value; then perform back propagation of the neural network to reduce the total loss function T_loss. Stop training when the total loss is lower than the set tolerance limit or reaches the preset number of iterations, and output the final unit force prediction value .

[0096] In one embodiment, the unit force prediction value is visualized to obtain a three-dimensional structure topology with a prestress distribution cloud chart, and a graph network data is bound in the three-dimensional structure topology.

[0097] As shown in the sunflower type spoke cable truss structure Figure 3 The specific process of finding force of the cable-strut tension structure is illustrated by taking the Kiewitt dome structure as an example, and the operation includes:

[0098] The three-dimensional graph geometry topology is imported from CAD, and there are 168 nodes and 408 edges. The nodes are divided into three categories, including 24 boundary nodes, 72 upper strut nodes and 72 lower strut nodes. The graph network structure data is established by using the torch_geometric class in the PyTorch library, including node coordinates, node connection relationship, strut unit index (special unit identifier to distinguish strut units and edge units), node category and association matrix B.

[0099] A four-layer neural network is established, including an input layer, two hidden layers and an output layer after linear transformation; the Tanh activation function is used between the input layer and the hidden layer, and the linearly transformed activation functions Softplus and Softplus+1 are used between the hidden layer and the output layer for struts and cables respectively. The total loss function (force distribution constraint loss and force balance residual loss) is calculated by using the output force distribution, then the total loss function is back propagated and minimized by using the Adam optimizer, and after 3000 iterations, the force balance residual loss is reduced to 8.9 e-4 , and the force distribution constraint loss is reduced to 7.0 e-5 . Finally, a feasible force distribution is obtained.

[0100] As shown in the Kiewitt dome structure Figure 4 The specific process of finding force of the cable-strut tension structure is illustrated by taking the Kiewitt dome structure as an example, and the operation includes:

[0101] The three-dimensional graph geometry topology is imported from CAD, and there are 56 nodes and 145 edges. The nodes are divided into three categories, including 18 boundary nodes, 19 upper strut nodes and 19 lower strut nodes. The graph network structure data is established by using the torch_geometric class in the PyTorch library, including node coordinates, edge connection relationship, strut unit index (special unit identifier to distinguish strut units and edge units), node position category and association matrix B.

[0102] A 4-layer neural network is established, including an input layer, two hidden layers and a linear transformed output layer; a Tanh activation function is used between the input layer and the hidden layer, and a linear transformed activation function-Softplus and Softplus+1 is used between the hidden layer and the output layer for the strut and the cable respectively. The total loss function (force distribution constraint loss and force balance residual loss) is calculated using the output force distribution, then back propagation is performed, and the total loss function is minimized using the Adam optimizer, which undergoes 3000 iterations, and the force balance residual loss decreases to 6.3 e-5 , and the force distribution constraint loss decreases to 1.9 e-6 . A feasible force distribution is finally obtained.

[0103] The cable-strut structure is converted into graph network data, a graph neural network is created, the force balance residual condition and the force distribution constraint condition can be automatically created based on the digitized topological relationship, the force transmission path under the topological constraint is explicitly learned, the balance equation is constructed in the loss function using the correlation matrix B =0, which guarantees that the solution satisfies the basic laws of mechanics, thereby constructing a GNN-PINN hybrid architecture that can simultaneously enhance the topological adaptability of the prestressed cable-strut structure and the compliance with physical laws, greatly improving the force finding accuracy of high-dimensional complex topological structures, enhancing the universality and general applicability. In addition, this method is applicable to all prestressed cable-strut structures with node balance topological conditions, and has very high convergence accuracy for high-dimensional problems. Finally, the prestress distribution can be automatically generated by importing the structure topology, improving the degree of automation and making it easier to implement.

[0104] It should be understood that, although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowcharts involved in the above-described embodiments can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or stages or steps or stages in other steps.

[0105] Based on the same inventive concept, the embodiments of the present application also provide a device for implementing the above-mentioned cable-strut tension structure force finding method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more cable-strut tension structure force finding device embodiments provided below can refer to the limitations of the cable-strut tension structure force finding method described above, which will not be repeated here.

[0106] In one embodiment, as shown in Figure 5 A cable-strut tension structure force finding device is provided, comprising:

[0107] A data extraction module is configured to parse geometric topology information of the cable-strut structure from a CAD file, construct graph network data according to the geometric topology information, and generate an incidence matrix of the graph network data; wherein nodes in the graph network data include strut nodes and boundary nodes, and edges include strut elements and cable elements;

[0108] A prediction output module is configured to input the incidence matrix into a pre-trained force distribution prediction model to obtain element force prediction values corresponding to the strut elements and the cable elements;

[0109] The loss function for training the force distribution prediction model is constructed based on physical constraints, including a force distribution constraint loss and a force balance residual loss;

[0110] The force distribution constraint loss takes the variance of the element force prediction values corresponding to edges of the same category as the force distribution constraint loss; the edge category is divided according to the distance from the midpoint of the edge to the coordinate origin

[0111] The force balance residual loss obtains the node resultant force based on the incidence matrix and the element force prediction values, extracts the resultant force residual of the strut nodes in the three coordinate axis directions of the three-dimensional coordinate system corresponding to the geometric topology information, and takes the L2 norm as the force balance residual loss in the three coordinate axis directions.

[0112] The above-mentioned modules in the cable-strut tension structure force finding device can be realized by software, hardware and their combinations in whole or in part. The above-mentioned modules can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above-mentioned modules.

[0113] In one exemplary embodiment, a computer device is provided, which can be a terminal, and its internal structure diagram can be as shown in Figure 6As shown in the figure. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. Among them, the processor, the memory and the input / output interface are connected through the system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capability. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner. Wireless mode can be achieved through WIFI, mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by the processor to implement a force finding method of a cable-strut tensile structure. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device. It can also be an external keyboard, touchpad or mouse, etc.

[0114] Those skilled in the art can understand that, Figure 6 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device.

[0115] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps in all method embodiments described above.

[0116] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the steps in all method embodiments described above.

[0117] In one embodiment, a computer program product is provided, including a computer program, and the computer program is executed by a processor to implement the steps in all method embodiments described above.

[0118] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of the country and region.

[0119] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of each method can be included. In the embodiments provided in the present application, any reference to memory, database or other medium can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0120] The technical features of the above embodiments can be combined in any way. To make the description concise, not all possible combinations of the technical features in the above embodiments are described, but as long as the combinations of the technical features do not exist contradictions, they should be considered as the scope of the present application.

[0121] The above-described embodiments are merely illustrative of several embodiments of the present application, and the description is relatively specific and detailed, but should not be understood as a limitation on the scope of the patent. It should be noted that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of the present application. Therefore, the scope of protection of the present application should be subject to the appended claims.

Claims

1. A method for finding forces in a cable-strut tension structure, characterized by, The method comprises: parsing geometric topology information of a cable-strut structure from a CAD file, and constructing graph network data according to the geometric topology information, wherein the graph network data comprises an association matrix; wherein nodes in the graph network data comprise strut nodes and boundary nodes, and edges comprise strut elements and cable elements; inputting the association matrix into a pre-trained force distribution prediction model to obtain element force prediction values corresponding to the strut elements and the cable elements; wherein a loss function for training the force distribution prediction model is constructed based on physical constraints, and comprises a force distribution constraint loss and a force balance residual loss; the force distribution constraint loss takes the variance of the element force prediction values corresponding to edges of the same category as the force distribution constraint loss; edge categories are divided according to the distance from the edge midpoint to the coordinate origin; the force balance residual loss obtains node resultant forces based on the association matrix and the element force prediction values, extracts the resultant force residual of the strut nodes in three coordinate axis directions of a three-dimensional coordinate system corresponding to the geometric topology information, and takes the L2 norm of the resultant force residual as the force balance residual loss in the three coordinate axis directions.

2. The method of claim 1, wherein: the geometric topology information comprises node coordinates, node categories, and node connection relationships; wherein the node categories comprise the strut nodes and the boundary nodes, and the strut nodes comprise strut upper nodes and strut lower nodes corresponding to the upper and lower ends of the same strut element, respectively.

3. The method of claim 2, wherein, constructing the graph network data according to the geometric topology information comprises: converting the node coordinates into a floating-point tensor; obtaining edge lengths of each strut element and each cable element based on the floating-point tensor; obtaining center point coordinates of each edge based on the node coordinates, obtaining the Euclidean distance from the center point coordinates to the coordinate origin, dividing edges with the same Euclidean distance into the same edge category, and constructing a mapping relationship between edge categories and edges; encoding nodes according to the node categories to obtain node category labels including strut upper nodes, strut lower nodes, and boundary nodes; encoding the connection relationship between edges and nodes into a matrix form to obtain the association matrix; generating graph network data based on the node coordinates, the node category labels, the node connection relationships, the edge categories, the edge lengths, and the association matrix.

4. The method of claim 1, wherein, the force distribution prediction model comprises, in sequence: an input layer for receiving the association matrix and flattening the association matrix into a one-dimensional vector; a plurality of serial hidden layers for receiving and processing the one-dimensional vector; the hidden layers are connected through a nonlinear activation function; an output layer for outputting the element force prediction values, the output layer comprising a Softplus activation function.

5. The method of claim 4, wherein, the Softplus activation function applies negative Softplus activation to the hidden layer output corresponding to the strut elements, and applies Softplus activation to the hidden layer output corresponding to the cable elements and then performs a linear transformation by adding 1.

6. The method of claim 1, wherein, The method further comprises: performing visual processing on the element force prediction values to obtain a three-dimensional structure topology graph with a prestress distribution cloud chart, and binding the graph network data in the three-dimensional structure topology graph.

7. A force finding device for cable-strut tension structures, characterized by The device comprises: The data extraction module is configured to parse geometric topology information of the cable-strut structure from the CAD file, construct graph network data according to the geometric topology information, and generate an incidence matrix of the graph network data; wherein nodes in the graph network data include strut nodes and boundary nodes, and edges include strut elements and cable elements; The prediction output module is configured to input the incidence matrix into a pre-trained force distribution prediction model to obtain element force prediction values corresponding to the strut elements and the cable elements; The loss function for training the force distribution prediction model is constructed based on physical constraints, including a force distribution constraint loss and a force balance residual loss; The force distribution constraint loss takes a variance of the element force prediction values corresponding to edges of the same category as the force distribution constraint loss; edge categories are divided according to distances from edge midpoints to a coordinate origin The force balance residual loss obtains node resultant forces based on the incidence matrix and the element force prediction values, extracts force balance residuals of the strut nodes in three coordinate axis directions of a three-dimensional coordinate system corresponding to the geometric topology information, and takes L2 norms of the force balance residuals as the force balance residual losses in the three coordinate axis directions, respectively.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method of any one of claims 1 to 6.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 6.

Images