Modeling Method, Device, Equipment and Medium for Finite Element Calculation Model of Transmission Tower

Abstract

The present invention discloses a method, device, equipment and medium for building a finite element calculation model of a transmission tower, which are used to solve the technical problems of complex process, high cost and large error in the existing modeling methods. The present invention includes: obtaining the point cloud data of the transmission tower; preprocessing the point cloud data to obtain target point cloud data; inputting the target point cloud data into a preset segmentation model to obtain the cross-sectional parameters of preset components of the transmission tower; calculating the angle steel cross-sectional size of the preset components by using the cross-sectional parameters; matching the limb width and limb thickness from a preset cross-sectional parameter library according to the angle steel cross-sectional size; and building a finite element calculation model of the transmission tower by using the limb width and limb thickness.

Description

Technical Field

[0001] The present invention relates to the technical field, and in particular to a method, device, equipment and medium for modeling a finite element calculation model of a transmission tower. Background Art

[0002] As an important facility of power engineering, transmission lines play a crucial role in the power supply of modern cities. To ensure the safe and reliable operation of the power grid, power operation units need to regularly inspect and maintain the jurisdiction lines and facilities. However, many old transmission towers have exceeded their designed service life, and these towers are still in use. Once they encounter natural disasters such as typhoons and low-temperature icing, it may cause tower damage and pose a safety hazard to the power grid operation. Therefore, it is urgent to establish calculation models for these old towers to evaluate their current disaster resistance capabilities. However, due to the long time of some old towers, their original design drawings have been lost. To re-obtain the structural parameters, relying solely on on-site manual measurement is not only costly but also very inefficient, especially for lines crossing mountains, which is even more difficult.

[0003] The existing technology mainly establishes a model through the following steps to evaluate the current disaster resistance ability of a transmission tower: collecting laser point cloud data of transmission line poles and towers, and establishing a pole and tower laser point cloud component model library; establishing a three-dimensional model of the pole and tower according to the pole and tower laser point cloud component model library, the pole and tower of the transmission line to be built and their position information; constructing the transmission line pole and tower according to the three-dimensional model of the pole and tower.

[0004] However, the above method requires accurate acquisition of the three-dimensional coordinate data of the transmission tower, which requires a large amount of labor costs; it is inefficient, the three-dimensional modeling process is complex, and a large amount of parameter extraction, mapping and calculation are required, and rapid and efficient modeling cannot be achieved; the process is complex, and multiple steps such as three-dimensional model construction, decomposition, and parameter mapping are required, and the process is complex; data dependence: it depends on accurate three-dimensional data, has extremely high requirements for data quality, but there are errors in manual surveying, which will lead to errors in the established model; the tower body needs to be divided into two parts, and the tower head and tower foot model libraries are established separately in advance. The model library is complex and has a large amount of data, and there are inaccuracies and errors in details such as cross-sectional dimensions, lengths and directions during construction. Summary of the Invention

[0005] The present invention provides a method, device, equipment and medium for modeling a finite element calculation model of a transmission tower, which are used to solve the technical problems of the existing modeling method being complex, costly and having large errors.

[0006] The present invention provides a method for modeling a finite element calculation model of a transmission tower, including:

[0007] Obtaining the point cloud data of the transmission tower;

[0008] Preprocess the point cloud data to obtain target point cloud data;

[0009] Input the target point cloud data into a preset segmentation model to obtain the cross-sectional parameters of the preset components of the transmission tower;

[0010] Calculate the angle steel cross-sectional dimensions of the preset components using the cross-sectional parameters;

[0011] Match the leg width and leg thickness from a preset cross-sectional parameter library according to the angle steel cross-sectional dimensions;

[0012] Construct a finite element calculation model of the transmission tower using the leg width and leg thickness.

[0013] Optionally, the step of obtaining the point cloud data of the transmission tower includes:

[0014] Obtain the position information of each node on the surface of the transmission tower through a laser sensor;

[0015] Obtain the color information of the surface of the transmission tower through a digital camera;

[0016] Generate the point cloud data of the transmission tower according to the position information and the color information.

[0017] Optionally, the step of preprocessing the point cloud data to obtain target point cloud data includes:

[0018] Divide the target point cloud data into multiple height intervals;

[0019] Calculate the existence probability of the point cloud data in each height interval;

[0020] Determine the height interval with the highest point cloud data existence probability as the ground height interval;

[0021] Divide the ground interval into multiple grids and calculate the ground height of each grid;

[0022] Obtain the actual floor area and actual height of the transmission tower;

[0023] Determine the ground point cloud according to the actual floor area and the ground height;

[0024] Determine the transmission tower point cloud according to the actual height, the ground point cloud and the ground height interval;

[0025] Perform compression and denoising processing on the transmission tower point cloud to obtain target point cloud data.

[0026] Optionally, the step of performing compression and denoising processing on the transmission tower point cloud to obtain target point cloud data includes:

[0027] Map the target point cloud data into a preset three-dimensional grid to obtain a point set for each three-dimensional grid;

[0028] Calculate the representative points of each three-dimensional grid according to the point set;

[0029] Generate a compressed point cloud using the representative points of all the three-dimensional grids;

[0030] Calculate the average distance of each point cloud in the compressed point cloud;

[0031] Determine candidate outlier points according to the average distance;

[0032] Perform density clustering on the candidate outlier points to obtain outlier points;

[0033] Remove the outlier points to obtain the target point cloud data.

[0034] Optionally, the generation process of the preset segmentation model includes:

[0035] Calculate the neighborhood space features of each single point of the sample point cloud data;

[0036] Aggregate the neighborhood space features to obtain aggregated features;

[0037] Decode the aggregated features to obtain point decoding features;

[0038] Generate segmentation labels for each point decoding feature through a feature learning model;

[0039] Optimize the feature learning model according to the difference between the segmentation label and the actual label to obtain the preset segmentation model.

[0040] Optionally, the step of matching the leg width and leg thickness from the preset cross-section parameter library according to the angle steel cross-section size includes:

[0041] Convert the data in the preset cross-section parameter library into library point cloud data;

[0042] Calculate the covariance matrix between the library point cloud data and the target point cloud data;

[0043] Perform singular value decomposition on the covariance matrix to obtain the eigenvector matrix;

[0044] Calculate the rotation transformation matrix and the translation vector according to the eigenvector matrix;

[0045] Align the library point cloud data and the target point cloud data according to the rotation transformation matrix and the translation vector, and match the leg width and leg thickness of the preset component from the library point cloud data according to the angle steel cross-section size.

[0046] The present invention also provides a device for modeling a finite element calculation model of a transmission tower, including:

[0047] A point cloud data acquisition module for acquiring the point cloud data of the transmission tower;

[0048] A preprocessing module for preprocessing the point cloud data to obtain target point cloud data;

[0049] A segmentation module for inputting the target point cloud data into a preset segmentation model to obtain the cross-sectional parameters of the preset components of the transmission tower;

[0050] An angle steel cross-sectional dimension calculation module for calculating the angle steel cross-sectional dimensions of the preset components by using the cross-sectional parameters;

[0051] A matching module for matching the limb width and limb thickness from a preset cross-sectional parameter library according to the angle steel cross-sectional dimensions;

[0052] A model construction module for constructing the finite element calculation model of the transmission tower by using the limb width and limb thickness.

[0053] Optionally, the point cloud data acquisition module includes:

[0054] A position information acquisition sub-module for acquiring the position information of each node on the surface of the transmission tower through a laser sensor;

[0055] A color information acquisition sub-module for acquiring the color information of the surface of the transmission tower through a digital camera;

[0056] A point cloud data generation sub-module for generating the point cloud data of the transmission tower according to the position information and the color information.

[0057] The present invention also provides an electronic device, which includes a processor and a memory:

[0058] The memory is used for storing program codes and transmitting the program codes to the processor;

[0059] The processor is used for executing the method for modeling a finite element calculation model of a transmission tower as described in any one of the above according to the instructions in the program codes.

[0060] The present invention also provides a computer-readable storage medium, which is used for storing program codes, and the program codes are used for executing the method for modeling a finite element calculation model of a transmission tower as described in any one of the above.

[0061] As can be seen from the above technical solutions, the present invention has the following advantages: The present invention provides a method for modeling a finite element calculation model of a transmission tower, specifically including: obtaining point cloud data of the transmission tower; preprocessing the point cloud data to obtain target point cloud data; inputting the target point cloud data into a preset segmentation model to obtain cross-section parameters of preset components of the transmission tower; calculating the angle steel cross-section size of the preset components using the cross-section parameters; matching the limb width and limb thickness from a preset cross-section parameter library according to the angle steel cross-section size; and constructing a finite element calculation model of the transmission tower using the limb width and limb thickness. By matching the point cloud data of the actual transmission tower components with the data of the standard tower, the present invention obtains the limb width and limb thickness of each rod of the transmission tower to construct a finite element calculation model, without the need to obtain accurate three-dimensional data of the transmission tower, thereby reducing the modeling complexity and data acquisition cost, and enabling targeted construction of the rods to improve the model construction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings without creative efforts based on these drawings.

[0063] Figure 1 It is a flowchart of the steps of a method for modeling a finite element calculation model of a transmission tower provided by an embodiment of the present invention;

[0064] Figure 2 It is a flowchart of the steps of a method for modeling a finite element calculation model of a transmission tower provided by another embodiment of the present invention;

[0065] Figure 3 It is a schematic structural diagram of a residual network structure provided by an embodiment of the present invention;

[0066] Figure 4 It is a block diagram of the structure of a device for modeling a finite element calculation model of a transmission tower provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0067] Embodiments of the present invention provide a method, device, equipment and medium for modeling a finite element calculation model of a transmission tower, which are used to solve the technical problems of complex process, high cost and large error in the existing modeling methods.

[0068] In order to make the objectives, features, and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the embodiments described below are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0069] Please refer to Figure 1 , Figure 1 , which is a step flow chart of a method for building a finite element calculation model of a transmission tower provided by an embodiment of the present invention.

[0070] A method for building a finite element calculation model of a transmission tower provided by the present invention specifically may include the following steps:

[0071] Step 101: Obtain the point cloud data of the transmission tower;

[0072] Transmission tower: The support point of the overhead line.

[0073] Point cloud: In reverse engineering, the set of point data on the appearance surface of a product obtained by a measuring instrument is called a point cloud. Usually, the number of points obtained by a three-dimensional coordinate measuring machine is relatively small, and the distance between points is relatively large, which is called a sparse point cloud; while the point cloud obtained by using a three-dimensional laser scanner or a photogrammetric scanner has a relatively large and dense number of points, which is called a dense point cloud.

[0074] In the embodiment of the present invention, the point cloud data of the transmission tower can be collected by three-dimensional laser scanning technology.

[0075] Step 102: Preprocess the point cloud data to obtain target point cloud data;

[0076] To ensure accuracy, when collecting the point cloud data of the transmission tower with a three-dimensional laser scanner, a relatively high scanning density can be set. However, due to the complex and diverse scenes of the transmission tower, the existence of bird nests, and the fact that the original data contains ground, conductor, and ground wire data and noise points, these data are not helpful for building the model. To reduce the workload of identifying the cross-section parameters of components and improve the processing efficiency, it is necessary to preprocess the point cloud data to obtain target point cloud data that removes interference data such as the ground.

[0077] Step 103: Input the target point cloud data into a preset segmentation model to obtain the cross-section parameters of the preset components of the transmission tower;

[0078] In a specific implementation, the preprocessed target point cloud data contains rods with various cross-sections. Therefore, individual rods need to be extracted from the entire transmission tower to obtain independent rod point cloud data for subsequent model matching. In one example, through deep learning, using the point cloud data of the transmission tower as a sample and the corresponding rod labels as the output, a preset segmentation model can be trained. By inputting the target point cloud data into the preset segmentation model, the corresponding rod labels can be obtained, and further the corresponding cross-section parameters can be obtained.

[0079] Step 104, calculate the angle steel cross-section dimensions of the preset component using the cross-section parameters;

[0080] Component: includes the main materials, auxiliary materials, and diagonal materials of the transmission tower, etc.

[0081] Angle steel: that is, angle iron, which is a long strip of steel with two sides perpendicular to each other at an angle.

[0082] Angle steel cross-section dimensions: include information such as the cross-section shape, dimensions, and weight of the component.

[0083] In the embodiment of the present invention, after obtaining the cross-section parameters, the angle steel cross-section dimensions of the preset component can be calculated using the cross-section parameters.

[0084] Step 105, match the leg width and leg thickness from the preset cross-section parameter library according to the angle steel cross-section dimensions;

[0085] Step 106, construct a finite element calculation model of the transmission tower using the leg width and leg thickness.

[0086] Leg width: refers to the width from the vertex of the angle steel to the edge.

[0087] Leg thickness: refers to the thickness of one of the two long strips of steel that make up the angle steel.

[0088] After obtaining the angle steel cross-section dimensions of the transmission tower, the corresponding leg width and leg thickness can be matched from the preset cross-section parameter library, and thus a finite element calculation model of the transmission tower can be constructed according to the leg width and leg thickness.

[0089] The present invention matches the point cloud data of the actual transmission tower components with the data of the standard tower, thereby obtaining the leg width and leg thickness of each rod of the transmission tower to construct a finite element calculation model, without the need to obtain accurate three-dimensional data of the transmission tower, thereby reducing the modeling complexity and data acquisition cost, and can construct the rods targeted, improving the model construction accuracy.

[0090] Please refer to Figure 2 , Figure 2 , which is a step flowchart of a method for modeling a finite element calculation model of a transmission tower provided by another embodiment of the present invention. Specifically, it may include the following steps:

[0091] Step 201: Obtain the position information of each node on the surface of the transmission tower through a laser sensor;

[0092] Step 202: Obtain the color information of the surface of the transmission tower through a digital camera;

[0093] Step 203: Generate point cloud data of the transmission tower based on the position information and color information;

[0094] In the embodiment of the present invention, the three-dimensional laser scanning technology can collect the position information and color information of the surface of an object. The position information is combined with an inertial navigation system, and the laser sensor locates by obtaining the position of the UAV itself, the distance to the object, the reflection intensity, and the angle. The color information is obtained by a digital camera to obtain the color information of the surface of the object. Using a three-dimensional laser scanner to scan the transmission tower and conducting on-site survey scanning of the transmission tower through a UAV carrying this device, the point cloud data of the transmission tower can be obtained quickly.

[0095] Step 204: Preprocess the point cloud data to obtain target point cloud data;

[0096] To ensure accuracy, when using a three-dimensional laser scanner to collect the point cloud data of the transmission tower, a relatively high scanning density can be set. However, due to the complex and diverse scenes of the transmission tower, the existence of bird nests, and the fact that the original data contains ground, conductor, and ground wire data as well as noise points, these data are not helpful for model establishment. To reduce the workload of identifying the cross-sectional parameters of components and improve the processing efficiency, it is necessary to preprocess the point cloud data to obtain target point cloud data that removes interference data such as the ground.

[0097] In one example, step 204 may include the following sub-steps:

[0098] S41: Divide the target point cloud data into multiple height intervals;

[0099] S42: Calculate the existence probability of the point cloud data in each height interval;

[0100] S43: Determine the height interval with the highest existence probability of the point cloud data as the ground height interval;

[0101] Since the point cloud data contains a large amount of ground point cloud data that affects the subsequent processing speed of the point cloud data and has no effect on subsequent modeling, it is necessary to identify and segment the ground. In one example, the ground can be cut using a probability statistics-based method. This method calculates the probability of the existence of point cloud data in each height interval in the substation scene. Let the number of points in the interval be p, the total number of point cloud data be N, and the interval width be h. Then the probability statistics result f is:

[0102]

[0103] Among them, f is the probability of the point cloud data existing in each height interval; the height with the maximum probability is the ground height.

[0104] S44, divide the ground interval into multiple grids, and calculate the ground height of each grid;

[0105] S45, obtain the actual floor area and actual height of the transmission tower;

[0106] S46, determine the ground point cloud according to the actual floor area and the ground height;

[0107] S47, determine the transmission tower point cloud according to the actual height, the ground point cloud and the ground height interval;

[0108] Considering the ground undulation around the transmission tower, it is necessary to perform grid processing on the transmission tower. Divide it into square grids of equal size according to the actual floor area of the transmission tower, and count the ground height of each grid. After processing the grid according to this height and merging all the grids, the point cloud data after ground separation is obtained. After ground cutting, based on the ground height and referring to the height situation of the transmission tower, delete the point cloud exceeding the highest height to obtain the transmission tower point cloud with the ground and top point cloud data removed.

[0109] S48, perform compression and denoising processing on the transmission tower point cloud to obtain the target point cloud data.

[0110] Since the embodiments of the present invention only need to model the cross-sectional parameters and positions of components according to the point cloud data, the requirements for the detailed parts of the geometric appearance are relatively low. To speed up the subsequent processing of the point cloud data, the transmission tower point cloud can be compressed and denoised to obtain the target point cloud data.

[0111] In one example, the steps of performing compression and denoising processing on the transmission tower point cloud to obtain the target point cloud data may specifically include the following sub-steps:

[0112] S481, map the target point cloud data to a preset three-dimensional grid to obtain the point sets of each three-dimensional grid;

[0113] S482, calculate the representative points of each three-dimensional grid according to the point sets;

[0114] S483, generate a compressed point cloud using the representative points of all three-dimensional grids;

[0115] S484, calculate the average distance of each point cloud in the compressed point cloud;

[0116] S485, determine the candidate outlier points according to the average distance;

[0117] S486, perform density clustering on the candidate outlier points to obtain the outlier points;

[0118] S487. Remove outliers to obtain the target point cloud data.

[0119] In specific implementation, first, it is necessary to determine the size of the spatial grid according to the range and level of detail requirements of the point cloud scene. The grid size directly affects the result of downsampling. Generally, a relatively large grid is initially set, and it is adjusted iteratively multiple times to obtain the optimal size. Map all the points in the original point cloud to a preset three-dimensional grid according to their three-dimensional coordinates, so that each grid cell contains a set of point clouds, and an octree is used to optimize the search speed. Calculate the geometric center of the point set in each grid cell using the centroid method as the representative point of the three-dimensional grid cell. Retain the representative points of each grid cell and delete the remaining points in the cell to obtain the compressed point cloud.

[0120] Next, perform outlier removal on the compressed point cloud. First, construct the neighbor relationship of the point cloud, calculate the average distance from each point cloud to its K nearest neighbor points, and determine the distance threshold for determining outliers according to the distribution of the average distances of all point clouds. Then compare the average distance of each point with the distance threshold, and the points greater than the threshold are marked as candidate outliers. Then perform density clustering analysis on the candidate outliers. If the class where the point cloud is located is relatively sparse, it is confirmed as an outlier. Delete the confirmed outliers from the original point cloud data to generate the denoised target point cloud data.

[0121] Step 205. Input the target point cloud data into a preset segmentation model to obtain the cross-section parameters of the preset components of the transmission tower.

[0122] In specific implementation, the target point cloud data after preprocessing contains various cross-section rods. Therefore, it is necessary to extract a single rod from the entire transmission tower to obtain independent rod point cloud data for subsequent model matching.

[0123] In one example, the training process of the preset segmentation model is as follows:

[0124] S51. Calculate the neighborhood space features of each single point of the sample point cloud data.

[0125] S52. Aggregate the neighborhood space features to obtain the aggregated features.

[0126] S53. Decode the aggregated features to obtain the point decoding features.

[0127] S54. Generate the segmentation labels of each point decoding feature through the feature learning model.

[0128] S55. Optimize the feature learning model according to the difference between the segmentation labels and the actual labels to obtain the preset segmentation model.

[0129] In the embodiment of the present invention, a coding structure based on neighborhood feature aggregation is constructed on the basis of the encoding-decoding structure of the deep learning semantic segmentation algorithm. The method divides the segmentation extraction process into three modules: feature encoding, feature decoding, and feature learning.

[0130] First, the neighborhood spatial features of the rod members are extracted by using the feature encoding module. Specifically, the core of the feature encoding module is to analyze and extract the three-dimensional coordinates of each point in the processed entire point cloud data. In order to expand the receptive field of each point cloud and enhance the local feature learning ability, a neighborhood feature aggregation module is constructed to highlight the local spatial features of the rod members. This module needs to complete two tasks. One is to represent the neighborhood spatial features of the three-dimensional coordinates of a single point, and the other is to aggregate the neighborhood features of each single point. For the task of representing neighborhood spatial features, the k-nearest neighbor algorithm (k-NN) is used to incorporate the information of the k nearest points around a single point into the spatial features of this point, and link the coordinates of this point with the relationship with the neighborhood points to form a kind of encoding to represent the neighborhood spatial features of this single point. Taking the K nearest neighbor points of the input point cloud data p i as an example, the encoding is represented as follows: For example, the encoding is represented as follows:

[0131]

[0132] where r i k is the relative coordinate, is the Euclidean distance.

[0133] Finally, the features of all input point cloud data are cascaded encoded and screened to complete the aggregation of all point features and obtain the aggregated features. The feature encoding module performs 4 times of downsampling and neighborhood feature aggregation on the input sample point cloud data.

[0134] In addition to the above two tasks, in order to avoid the loss of shallow features after continuous feature aggregation, the embodiment of the present invention introduces a residual network structure, summarizes the features before encoding and the features after neighborhood feature aggregation, and enhances the nonlinearity of the weighted obtained features by using the ReLU activation function. The residual network structure is as Figure 3 shown.

[0135] The core of feature decoding is to restore the number of points after downsampling by the encoding module to the original number of points. The restoration method can adopt the nearest neighbor interpolation method, and at the same time use the MLP to learn the result of each upsampling to ensure that the features will not be damaged by interpolation. The nearest neighbor interpolation method is to fill the surrounding of each point with a fixed ratio a reduction value of this point. The specific formula is as follows:

[0136]

[0137] Then, the extracted features are screened and the feature decoding is completed. After decoding, the important feature description of the point cloud data is obtained as the point decoding feature.

[0138] Finally, the point decoding features are learned to obtain the segmentation label of each point, and the segmentation and extraction of the rod members are completed.

[0139] The feature learning module consists of three fully connected layers. The function of the fully connected layer is similar to polynomial fitting. Let a function Y = f(X) = wX + b. To find the relationship f(X) between Y and X, where w is called the weight and b is called the bias value. Usually, X and Y are multi-dimensional vectors. The fully connected layer continuously updates these two parameter values and returns the loss value to continuously fit the functional relationship between X and Y. For point cloud data, the input X not only contains the coordinates of the points but also the features of the points, and there are many objects to be learned. Therefore, the learning effect of a single fully connected layer is usually not good. Therefore, the embodiment of the present invention selects three fully connected layers as the basic structure. At the same time, considering that increasing the number of fully connected layers will bring the problem of too many parameters to be learned, a Dropout layer is added after the second fully connected layer to randomly discard some parameters and reduce the computational amount.

[0140] The training parameters are shown in Table 1 below, and the specific settings can be determined according to the actual situation.

[0141]

[0142] Table 1

[0143] After the preset segmentation model is trained, the preprocessed target point cloud data can be input into the preset segmentation model to segment and extract individual components, and the parameter settings can be iteratively modified according to the obtained results until the expected efficiency and accuracy are achieved.

[0144] Step 206, calculate the angle steel cross-section size of the preset component using the cross-section parameters;

[0145] In the embodiment of the present invention, after obtaining the cross-section parameters, the angle steel cross-section size of the preset component can be calculated using the cross-section parameters.

[0146] Step 207, match the limb width and limb thickness from the preset cross-section parameter library according to the angle steel cross-section size;

[0147] After obtaining the angle steel cross-section size, the limb width and limb thickness can be matched from the preset cross-section parameter library according to the angle steel cross-section size.

[0148] In specific implementation, the construction of the cross-section parameter library of commonly used angle steels for transmission towers is the key to component-based modeling of tower poles. The components of transmission towers are generally made of angle steels, steel plates, and steel pipe components, and are combined by bolt connection or welding connection. Based on comprehensively considering the differences and similarities of different three-dimensional models of tower poles, high-precision three-dimensional modeling of different components is carried out through CAD software, and a cross-section size library of commonly used components of the tower is established. The specific steps are as follows:

[0149] 1. Determine the classification of the cross-section library: Classify the cross-section library into main materials, auxiliary materials, and diagonal materials. Main materials refer to the main components of transmission towers, such as angle steels, I-beams, etc., and their attributes are set as beam elements; auxiliary materials refer to the secondary components of transmission towers, such as bolts, gaskets, etc.; diagonal materials refer to the diagonal components of transmission towers, such as diagonal braces, etc., and their attributes are set as rod elements.

[0150] 2. Collect cross-section size data: Collect the cross-section size data of commonly used components of transmission towers, including information such as the cross-section shape, size, and weight of main materials, auxiliary materials, and diagonal materials.

[0151] 3. Organize the cross-section size data: Organize the collected cross-section size data into the form of a table or database for subsequent use and management.

[0152] In order to match the leg width and leg thickness from the preset cross-section parameter library, it is first necessary to realize the matching of the target point cloud data and the preset cross-section parameter library. Specifically, the coordinate system of the three-dimensional CAD model of the preset cross-section parameter library can be adjusted through translation, rotation, and scale transformation to make it basically consistent with the coordinate system of the corresponding target point cloud data, so as to determine the initial pose.

[0153] In specific implementation, step 207 may include the following sub-steps:

[0154] S71. Convert the data in the preset cross-section parameter library into library point cloud data;

[0155] S72. Calculate the covariance matrix between the library point cloud data and the target point cloud data;

[0156] S73. Perform singular value decomposition on the covariance matrix to obtain the eigenvector matrix;

[0157] S74. Calculate the rotation transformation matrix and translation vector according to the eigenvector matrix;

[0158] S75. Align the library point cloud data and the target point cloud data according to the rotation transformation matrix and translation vector, and match the leg width and leg thickness of the preset structure from the library point cloud data.

[0159] In a specific implementation, first, the CAD model in the preset cross-section parameter library can be converted into library point cloud data. In the embodiments of the present invention, the STEP format is used to record the CAD model. STEP is a format for recording CAD model data, and the described object information includes not only geometric information of the model such as points, lines, surfaces, and solids, but also annotation information and structural information, etc. The CAD model in the STEP format is discretized through the OpenCASCADE discretization function to obtain the corresponding triangular mesh model.

[0160] Next, through the principal component analysis method, the three principal axis directions of the point cloud and the discrete model are obtained, the center of gravity of the model is set as the coordinate origin, a PCA coordinate system for the data is established, and the initial pose of the point cloud data and the discrete model can be determined by adjusting the coordinate system. The specific solution process of the algorithm is as follows:

[0161] Let represent two point cloud data respectively, P D represents the original measured point cloud data, P M is the point set corresponding to the CAD model, where N d , N m are the number of points of the point cloud respectively. Construct the covariance matrix of the point cloud:

[0162]

[0163]

[0164] where corresponds to the center of gravity of the point cloud:

[0165]

[0166]

[0167] Perform singular value decomposition on the covariance matrix. The principal component axes of the point cloud are the eigenvectors of the matrix respectively, and the eigenvector matrix is obtained by solving:

[0168]

[0169]

[0170] L D , L M is the eigenvalue diagonal matrix, U D , U M is the eigenvector matrix of the covariance matrix. Thus, the rotation transformation matrix R and the translation vector T for determining the initial pose can be calculated:

[0171]

[0172]

[0173] Furthermore, after determining the initial pose, in order to reduce the pose deviation of the model, the registration result can be further adjusted to achieve the precise registration of the discrete model and the point cloud data, which mainly includes the following three steps:

[0174] 1. Obtain the nearest corresponding point set;

[0175] 2. Minimize the error;

[0176] 3. Repeat these two steps until the registration error is less than the threshold to achieve the correction of the registration error.

[0177] The specific process is as follows:

[0178] After calculating the corresponding point set of the discrete model and the point cloud data, it is necessary to establish an error equation for solving the transformation matrix:

[0179] f(R,T;P) = RP + T

[0180]

[0181] f(R,T;P) is the coordinate transformation equation of the corresponding point set, Error(P) is the error equation, R represents the rotation matrix, T represents the translation vector, P represents the measurement data set, N represents the number of points, m i represents the coordinate of a point in the corresponding point set of the CAD model, p i is the coordinate of a point in the point cloud data. In this paper, the problem of registration error correction is transformed into the problem of minimizing the error equation, that is:

[0182]

[0183] By solving the equation (R * ,T * ), the optimal rotation transformation matrix and translation matrix can be obtained.

[0184] Then, according to the optimal rotation transformation matrix and translation matrix, following the principle of preferentially matching the limb width, the limb width and limb thickness parameters of the component are obtained.

[0185] Furthermore, after completing the matching of the limb width and limb thickness, the main material, auxiliary material and diagonal material can also be matched to obtain the corresponding parameters.

[0186] Step 208, construct a finite element calculation model of the transmission tower by using the limb width and limb thickness.

[0187] In the specific implementation, according to the identified limb width parameters, the finite element calculation model of the transmission tower can be constructed under the same limb width using the principle of the thinnest limb thickness. The risk calculation is performed using the finite element calculation model. If the calculation result is safe, it proves that the tower's wind and ice resistance meet safety requirements. If the calculation result is dangerous, the limb thickness will be obtained through manual measurement or higher-precision point cloud data, and remodeling and calculation will be performed to analyze its safety.

[0188] The present invention matches the point cloud data of actual transmission tower components with the data of standard towers to obtain the limb width and limb thickness of each rod of the transmission tower to construct a finite element calculation model. There is no need to obtain accurate three-dimensional data of the transmission tower, thereby reducing modeling complexity and data acquisition costs, and can carry out targeted construction of the rods to improve the accuracy of model construction.

[0189] See also Figure 4 , Figure 4 A structural block diagram of a modeling device for a finite element calculation model of a transmission tower provided in an embodiment of the present invention.

[0190] The embodiment of the present invention provides a device for modeling a finite element calculation model of a transmission tower, comprising:

[0191] Point cloud data acquisition module 401, used to acquire point cloud data of the transmission tower;

[0192] A preprocessing module 402 is used to preprocess the point cloud data to obtain target point cloud data;

[0193] The segmentation module 403 is used to input the target point cloud data into a preset segmentation model to obtain the cross-sectional parameters of the preset components of the transmission tower;

[0194] Angle steel cross-sectional dimension calculation module 404, used to calculate the angle steel cross-sectional dimension of a preset component using cross-sectional parameters;

[0195] A matching module 405 is used to match the limb width and limb thickness from a preset cross-section parameter library according to the cross-section size of the angle steel;

[0196] The model building module 406 is used to build a finite element calculation model of the transmission tower using the limb width and limb thickness.

[0197] In the embodiment of the present invention, the point cloud data acquisition module 401 includes:

[0198] The location information acquisition submodule is used to obtain the location information of each node on the surface of the transmission tower through a laser sensor;

[0199] The color information acquisition submodule is used to acquire the color information of the surface of the transmission tower through a digital camera;

[0200] A point cloud data generation sub-module, configured to generate point cloud data of a transmission tower according to position information and color information.

[0201] In an embodiment of the present invention, the preprocessing module 402 includes:

[0202] A height interval division sub-module, configured to divide the target point cloud data into multiple height intervals;

[0203] A point cloud data existence probability calculation sub-module, configured to calculate the existence probability of the point cloud data in each height interval;

[0204] A ground height interval determination sub-module, configured to determine the height interval with the highest point cloud data existence probability as the ground height interval;

[0205] A ground height calculation sub-module, configured to divide the ground interval into multiple grids and calculate the ground height of each grid;

[0206] An actual floor area and actual height acquisition sub-module, configured to acquire the actual floor area and actual height of the transmission tower;

[0207] A ground point cloud determination sub-module, configured to determine the ground point cloud according to the actual floor area and the ground height;

[0208] A transmission tower point cloud determination sub-module, configured to determine the transmission tower point cloud according to the actual height, the ground point cloud, and the ground height interval;

[0209] A target point cloud data acquisition sub-module, configured to perform compression and denoising processing on the transmission tower point cloud to obtain the target point cloud data.

[0210] In an embodiment of the present invention, the target point cloud data acquisition sub-module includes:

[0211] A mapping unit, configured to map the target point cloud data into a preset three-dimensional grid to obtain a point set of each three-dimensional grid;

[0212] A representative point calculation unit, configured to calculate the representative point of each three-dimensional grid according to the point set;

[0213] A compressed point cloud generation unit, configured to generate a compressed point cloud by using the representative points of all three-dimensional grids;

[0214] An average distance calculation unit, configured to calculate the average distance of each point cloud in the compressed point cloud;

[0215] A candidate outlier determination unit, configured to determine candidate outliers according to the average distance;

[0216] A density clustering unit, configured to perform density clustering on the candidate outliers to obtain outliers;

[0217] A target point cloud data determination unit, configured to remove outliers to obtain target point cloud data.

[0218] In an embodiment of the present invention, a preset segmentation model generation module includes:

[0219] A neighborhood space feature calculation sub-module, configured to calculate the neighborhood space features of each single point of the sample point cloud data;

[0220] An aggregation sub-module, configured to aggregate the neighborhood space features to obtain aggregated features;

[0221] A decoding sub-module, configured to decode the aggregated features to obtain point decoding features;

[0222] A feature learning sub-module, configured to generate segmentation labels for each point decoding feature through a feature learning model;

[0223] An optimization sub-module, configured to optimize the feature learning model according to the difference between the segmentation labels and the actual labels to obtain a preset segmentation model.

[0224] In an embodiment of the present invention, a matching module 405 includes:

[0225] A library point cloud data conversion sub-module, configured to convert the data in a preset cross-section parameter library into library point cloud data;

[0226] A covariance matrix calculation sub-module, configured to calculate the covariance matrix between the library point cloud data and the target point cloud data;

[0227] A singular value decomposition sub-module, configured to perform singular value decomposition on the covariance matrix to obtain an eigenvector matrix;

[0228] A rotation transformation matrix and translation vector calculation sub-module, configured to calculate a rotation transformation matrix and a translation vector according to the eigenvector matrix;

[0229] A matching sub-module, configured to align the library point cloud data and the target point cloud data according to the rotation transformation matrix and the translation vector, and match the preset built limb width and limb thickness from the library point cloud data according to the angle steel cross-section size.

[0230] An embodiment of the present invention further provides an electronic device, which includes a processor and a memory:

[0231] The memory is used to store program codes and transmit the program codes to the processor;

[0232] The processor is configured to execute the power transmission tower finite element calculation model modeling method according to the instructions in the program codes.

[0233] An embodiment of the present invention further provides a computer-readable storage medium, which is used to store program codes for executing the modeling method of the finite element calculation model of the transmission tower according to the embodiment of the present invention.

[0234] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0235] Each embodiment in this specification is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other.

[0236] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a device, or a computer program product. Therefore, the embodiments of the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0237] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing terminal devices to generate a machine, so that the instructions executed by the processors of the computer or other programmable data processing terminal devices generate a device for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0238] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured product including an instruction device, and the instruction device implements the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0239] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable terminal device provide steps for implementing the functions specified in one process or multiple processes and / or blocks Figure 1 one process or multiple processes and / or blocks Figure 1 steps for implementing the functions specified in one block or multiple blocks.

[0240] Although the preferred embodiments of the embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they know the basic creative concept. Therefore, the appended claims are intended to be construed as including the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

[0241] Finally, it should also be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or terminal device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or elements inherent to such process, method, article or terminal device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or terminal device comprising the element.

[0242] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A modeling method for the finite element calculation model of a transmission tower, characterized in that Including: Obtain the point cloud data of the transmission tower; Preprocess the point cloud data to obtain target point cloud data; Input the target point cloud data into a preset segmentation model to obtain the cross-section parameters of the preset components of the transmission tower; Calculate the angle steel cross-section size of the preset component using the cross-section parameters; Match the leg width and leg thickness from a preset cross-section parameter library according to the angle steel cross-section size; Construct a finite element calculation model of the transmission tower using the leg width and leg thickness; Among them, the step of preprocessing the point cloud data to obtain target point cloud data includes: Divide the target point cloud data into multiple height intervals; Calculate the existence probability of the point cloud data in each height interval; Determine the height interval with the highest point cloud data existence probability as the ground height interval; Divide the ground interval into multiple grids and calculate the ground height of each grid; Obtain the actual floor area and actual height of the transmission tower; Determine the ground point cloud according to the actual floor area and the ground height; Determine the point cloud of the transmission tower according to the actual height, the ground point cloud, and the ground height interval; Perform compression and denoising processing on the point cloud of the transmission tower to obtain target point cloud data.

2. The method according to claim 1, characterized in that, The step of obtaining the point cloud data of the transmission tower includes: Obtain the position information of each node on the surface of the transmission tower through a laser sensor; Obtain the color information on the surface of the transmission tower through a digital camera; Generate the point cloud data of the transmission tower according to the position information and the color information.

3. The method according to claim 1, characterized in that, The step of performing compression and denoising processing on the point cloud of the transmission tower to obtain target point cloud data includes: Map the target point cloud data into a preset three-dimensional grid to obtain a point set of each three-dimensional grid; Calculate the representative point of each three-dimensional grid according to the point set; Generate a compressed point cloud using the representative points of all the three-dimensional grids; Calculate the average distance of each point cloud in the compressed point cloud; Determine candidate outlier points according to the average distance; Perform density clustering on the candidate outlier points to obtain outlier points; Remove the outlier points to obtain target point cloud data.

4. The method according to claim 1, wherein The generation process of the preset segmentation model includes: Calculate the neighborhood space feature of each single point of the sample point cloud data; Aggregate the neighborhood space features to obtain an aggregated feature; Decode the aggregated feature to obtain a point decoding feature; Generate a segmentation label for each point decoding feature through a feature learning model; Optimize the feature learning model according to the difference between the segmentation label and the actual label to obtain a preset segmentation model.

5. The method according to claim 1, wherein The step of matching the leg width and leg thickness from a preset cross-section parameter library according to the angle steel cross-section size includes: Convert the data in the preset cross-section parameter library into library point cloud data; Calculate the covariance matrix between the library point cloud data and the target point cloud data; Perform singular value decomposition on the covariance matrix to obtain an eigenvector matrix; Calculate the rotation transformation matrix and translation vector according to the eigenvector matrix; Align the library point cloud data and the target point cloud data according to the rotation transformation matrix and the translation vector, and match the limb width and limb thickness of the preset component from the library point cloud data according to the angle steel section size.

6. A modeling device for a finite element calculation model of a transmission tower, characterized in that, Including: A point cloud data acquisition module, configured to acquire the point cloud data of the transmission tower; A preprocessing module, configured to preprocess the point cloud data to obtain target point cloud data; A segmentation module, configured to input the target point cloud data into a preset segmentation model to obtain the cross-section parameters of the preset components of the transmission tower; An angle steel section size calculation module, configured to calculate the angle steel section size of the preset component by using the cross-section parameters; A matching module, configured to match the limb width and limb thickness from a preset cross-section parameter library according to the angle steel section size; A model construction module, configured to construct a finite element calculation model of the transmission tower by using the limb width and limb thickness; Among them, the preprocessing module includes: A height interval division sub-module, configured to divide the target point cloud data into multiple height intervals; A point cloud data existence probability calculation sub-module, configured to calculate the existence probability of the point cloud data in each height interval; A ground height interval determination sub-module, configured to determine the height interval with the maximum point cloud data existence probability as the ground height interval; A ground height calculation sub-module, configured to divide the ground interval into multiple grids and calculate the ground height of each grid; An actual floor area and actual height acquisition sub-module, configured to acquire the actual floor area and actual height of the transmission tower; A ground point cloud determination sub-module, configured to determine the ground point cloud according to the actual floor area and the ground height; A transmission tower point cloud determination sub-module, configured to determine the transmission tower point cloud according to the actual height, the ground point cloud and the ground height interval; A target point cloud data acquisition sub-module, configured to perform compression and denoising processing on the transmission tower point cloud to obtain target point cloud data.

7. The device according to claim 6, characterized in that The point cloud data acquisition module includes: A position information acquisition sub-module, configured to acquire the position information of each node on the surface of the transmission tower through a laser sensor; A color information acquisition sub-module, configured to acquire the color information of the surface of the transmission tower through a digital camera; A point cloud data generation sub-module, configured to generate the point cloud data of the transmission tower according to the position information and the color information.

8. An electronic device, characterized in that, The device includes a processor and a memory: The memory is used to store program codes and transmit the program codes to the processor; The processor is configured to execute the method for modeling a finite element calculation model of a transmission tower according to any one of claims 1-5 according to the instructions in the program codes.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store program codes, and the program codes are used to execute the method for modeling a finite element calculation model of a transmission tower according to any one of claims 1-5.

Images