Methods, devices, electronic equipment, and storage media for simulating the installation of large-size components
By acquiring measured data of components to generate high-precision point cloud data, and performing 3D model registration and pose parameter calculation, the problems of low assembly accuracy and safety hazards in the installation of large-size components are solved, and an efficient and safe installation process is achieved.
Patent Information
- Application Number
- CN202610435176.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-03
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2046-04-03
AI Technical Summary
Existing technologies cannot achieve priority alignment of key features in the installation of large-sized components, resulting in low assembly accuracy, potential safety hazards, and a lack of dynamic interference detection throughout the entire process, leading to low efficiency.
By acquiring measured data of components, high-precision point cloud data is generated, 3D model registration is performed, the optimal installation pose parameters are determined, and construction instructions are generated, enabling priority alignment of key features and dynamic simulation and prediction of installation risks throughout the entire process.
It improves the installation accuracy and one-time assembly success rate of large-sized components, reduces construction risks and reliance on manual experience, and improves construction efficiency and safety.
Smart Images

Figure CN122333591B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of information technology in building engineering, and in particular to a method, device, electronic equipment, and storage medium for simulating the installation of large-size components. Background Technology
[0002] During the installation of large-sized components, due to the limitations of the component's size and weight, the installation process is simulated before installation to generate corresponding installation control parameters, thereby ensuring the success rate of component installation.
[0003] Existing technical solutions utilize high-precision measuring equipment to acquire the actual geometric data of the components to be assembled, import it together with the preset three-dimensional design model into a virtual simulation platform, and use basic algorithms to align the models, simulate the hoisting and assembly process of the components, identify mismatches between theoretical design values and actual processing values in advance, and provide reference data for on-site construction.
[0004] However, existing technical solutions for pose adjustment mostly rely on simple geometric alignment or manual fine-tuning, only pursuing the overall surface alignment of the model, without considering the different impacts of various features on assembly accuracy. The core parts that directly determine the assembly quality, such as fastening bolt holes and mating sealing surfaces, have vastly different accuracy requirements from the non-mating surfaces of components. Existing solutions cannot achieve priority alignment of key features, which easily leads to the problem that the overall model appears aligned, but the core mating parts deviate beyond the standard. This results in a very low success rate for one-time assembly, requiring repeated on-site adjustments. Furthermore, it can only achieve static model comparison and lacks dynamic interference detection for the entire installation process. It cannot predict the collision risks during hoisting and closing, leading to frequent collisions between components and surrounding structures during on-site installation, and even structural damage caused by forced closing, posing significant safety and quality hazards. At the same time, for the machining allowance of the closing section, existing technologies can only achieve rough dimensional measurement and cannot quantify the spatial distribution of the allowance. It can only rely on manual grinding and trial and error on-site, which is extremely inefficient and cannot guarantee accuracy. Summary of the Invention
[0005] This invention provides a method, apparatus, electronic device, and storage medium for simulating the installation of large-sized components. By acquiring measured data corresponding to the large-sized components and determining the installation posture parameters and generating corresponding control commands based on the model dataset obtained by registering the three-dimensional model and the theoretical three-dimensional model, it achieves priority alignment of key features, dynamically simulates and predicts installation risks throughout the entire process, and realizes autonomous control of accuracy through closed-loop feedback. This improves the installation accuracy and one-time assembly success rate of large-sized components, reduces construction risks, and reduces reliance on manual experience.
[0006] According to one aspect of the present invention, a method for simulating the installation of large-size components is provided, comprising:
[0007] Acquire the actual measured data of the component corresponding to the component to be installed, and determine the high-precision point cloud data corresponding to the component to be installed based on the actual measured data. The actual measured data of the component includes the original point cloud data, measurement environment parameters and component temperature distribution data.
[0008] Based on the high-precision point cloud data component and the corresponding measured 3D model of the component to be installed, the measured 3D model and the theoretical 3D model are spatially registered to obtain the model dataset.
[0009] Determine the optimal six-DOF installation pose parameters corresponding to the component to be installed based on the model dataset;
[0010] Based on the optimal six-degree-of-freedom installation pose parameters, installation instructions corresponding to the components to be installed are generated.
[0011] According to another aspect of the present invention, a large-size component installation simulation device is provided, comprising:
[0012] The data acquisition module is used to acquire the measured data of the component corresponding to the component to be installed, and to determine the high-precision point cloud data corresponding to the component to be installed based on the measured data. The measured data of the component includes the original point cloud data, measurement environment parameters and component temperature distribution data.
[0013] The 3D model building module is used to create a measured 3D model based on high-precision point cloud data components and the corresponding measured 3D model of the component to be installed, and to spatially register the measured 3D model with the theoretical 3D model to obtain the model dataset.
[0014] The pose parameter determination module is used to determine the optimal six-degree-of-freedom installation pose parameters corresponding to the component to be installed based on the model dataset.
[0015] The construction instruction generation module is used to generate installation construction instructions corresponding to the components to be installed based on the optimal six-degree-of-freedom installation pose parameters.
[0016] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:
[0017] At least one processor; and
[0018] A memory that is communicatively connected to at least one processor; wherein,
[0019] The memory stores a computer program that can be executed by at least one processor, such that the at least one processor is able to perform the large-size component installation simulation method of any embodiment of the present invention.
[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, which stores computer instructions for causing a processor to execute and implement the large-size component installation simulation method of any embodiment of the present invention.
[0021] The technical solution of this invention involves acquiring measured data of the component corresponding to the component to be installed, and determining high-precision point cloud data corresponding to the component based on the measured data. Based on the measured 3D model corresponding to the high-precision point cloud data component and the component to be installed, the measured 3D model and the theoretical 3D model are spatially registered to obtain a model dataset. Based on the model dataset, the optimal six-degree-of-freedom installation pose parameters corresponding to the component to be installed are determined. Installation instructions corresponding to the component to be installed are generated based on the optimal six-degree-of-freedom installation pose parameters. Based on the above technical solution, by acquiring measured data corresponding to large-size components and determining the installation pose parameters and generating corresponding control instructions based on the model dataset obtained by registering the 3D model and the theoretical 3D model, key features are prioritized for alignment. Furthermore, the entire process dynamically simulates and predicts installation risks, achieving autonomous control of accuracy through closed-loop feedback. This improves the installation accuracy and one-time assembly success rate of large-size components, reduces construction risks, and decreases reliance on manual experience.
[0022] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description
[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0024] Figure 1 This is a flowchart of a large-size component installation simulation method provided in an embodiment of the present invention;
[0025] Figure 2 This is a schematic diagram of the data flow in the large-size component installation simulation system provided in this embodiment of the invention;
[0026] Figure 3 This is a flowchart illustrating the construction of a digital twin model for a large-sized component provided in an embodiment of the present invention;
[0027] Figure 4 This is a flowchart of six-degree-of-freedom spatial pose calculation provided in an embodiment of the present invention;
[0028] Figure 5 This is a flowchart of dynamic interference and margin analysis provided in an embodiment of the present invention;
[0029] Figure 6 This is a flowchart of the closed-loop feedback control provided in an embodiment of the present invention;
[0030] Figure 7 This is a flowchart of a large-size component installation simulation method provided in an embodiment of the present invention;
[0031] Figure 8 A schematic diagram of a large-size component installation simulation device provided in an embodiment of the present invention;
[0032] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation
[0033] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.
[0034] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0035] Figure 1 This is a flowchart illustrating a large-size component installation simulation method provided by an embodiment of the present invention. This embodiment is applicable to situations where, during the installation of large-size components, a three-dimensional model is aligned based on collected measured data, and installation instructions are determined based on the model data. This method can be executed by a large-size component installation simulation device, which can be implemented in hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method specifically includes the following steps:
[0036] It should be noted that, as Figure 2 As shown, the technical solution of this invention constructs a full-process installation simulation system encompassing data acquisition, environmental compensation, model reconstruction, model registration, pose calculation, interference margin analysis, installation strategy optimization, command generation, and precision closed-loop control. The system acquires the original geometric data of the component to be installed using multi-source high-precision measuring equipment, eliminates environmental interference through real-time environmental parameter compensation, and generates high-precision point cloud data. Based on the high-precision point cloud, a measured 3D model representing the actual shape of the component is reconstructed and highly registered with the theoretical design model. The optimal installation pose of the component is calculated using a weighted least squares algorithm, and the complete installation process is simulated in virtual space, completing collision detection and margin analysis. Finally, directly executable construction commands are generated, and a closed-loop control system is formed through precision evaluation and backtracking mechanisms to ensure that the installation accuracy meets the requirements.
[0037] S110. Obtain the measured data of the component corresponding to the component to be installed, and determine the high-precision point cloud data corresponding to the component to be installed based on the measured data of the component.
[0038] The measured data for the components includes raw point cloud data, measurement environment parameters, and component temperature distribution data. The component to be installed can be understood as a large-sized component requiring precision installation in scenarios such as heavy equipment manufacturing and large ship assembly. The measured data can be a dataset characterizing the component's geometry and the measurement environment. The raw point cloud data can be discrete point coordinate data of the component's surface collected by multi-source measurement equipment. The measurement environment parameters can be temperature, humidity, and air pressure data from the measurement site, used to eliminate environmental interference with measurement accuracy. The component temperature distribution data can be understood as multi-point temperature data on the component's surface, used to correct for the effects of thermal deformation. The high-precision point cloud data can be point cloud data reflecting the actual shape of the component after multi-source stitching and environmental compensation.
[0039] Specifically, the measured data of the component corresponding to the component to be installed is obtained, and high-precision point cloud data corresponding to the component to be installed is determined based on the measured data. For example, a measurement cluster consisting of a laser scanning device, a visual measurement device, and a total station can be used to collect the original point cloud data of the component and the coordinates of the coded marker points; the measurement environment parameters and component temperature distribution data are collected simultaneously, and the original point cloud data is obtained by stitching together the multi-source point clouds; then, the original point cloud data is corrected for errors through atmospheric refractive index correction and thermal deformation compensation to obtain high-precision point cloud data.
[0040] For example, such as Figure 3As shown, to acquire complete and high-precision original point cloud data of the component to be installed, a laser scanning device, a vision measurement device, and a total station are integrated to overcome the limitations of a single measurement method. These three devices work together to form a three-dimensional spatial coordinate measurement cluster. The laser scanning device, as the main equipment for dense point cloud acquisition, is deployed at multiple scanning stations around the component. Using the time-of-flight method or phase method, it emits a laser beam to scan the component surface, records the round-trip time and scanning angle, and calculates the three-dimensional coordinates of each sampling point in the local coordinate system of the scanning device, quickly acquiring massive amounts of original point cloud data of the component surface. The vision measurement device provides a spatial reference for multi-site point cloud stitching. Encoded markers with unique identification information are pre-attached to key locations such as corners, mating surfaces, and bolt holes of the component. The vision measurement device captures these markers using a high-resolution image sensor and, through a sub-pixel edge extraction algorithm, accurately locates the center coordinates of the markers, obtaining their two-dimensional coordinates in the image coordinate system. Total Station: As the control core of the global coordinate system, it deploys reflecting prisms corresponding one-to-one with the coded marker points at the measurement site. Through automatic aiming and precise distance and angle measurement, it determines the three-dimensional coordinates of all prisms in the global coordinate system (factory geodetic coordinate system / assembly reference coordinate system), establishing a high-precision three-dimensional reference frame. Multi-source Point Cloud Stitching: Based on the global coordinates of the coded marker points measured by the total station, a mapping relationship is established with the image coordinates of the corresponding marker points identified by the visual measurement device; the coded marker points in the laser point cloud are identified and matched with the marker points with known global coordinates, calculating the pose of each scanning station in the global coordinate system. The point cloud data of all stations are uniformly converted to the global coordinate system, completing the precise stitching of the multi-source point clouds, and finally obtaining the original point cloud data covering the entire surface of the component.
[0041] Based on the above technical solution, high-precision point cloud data corresponding to the component to be installed is determined according to the measured data of the component, including: completing multi-source point cloud stitching based on the measured data of the component to obtain the original point cloud data of the component to be installed; and performing environmental compensation on the original point cloud data based on the measured environmental parameters and component temperature distribution data to obtain high-precision point cloud data.
[0042] Specifically, a mapping relationship can be established between the global coordinates of the coded marker points measured by the total station and the image coordinates of the marker points identified by the visual measurement device. The coded marker points in the laser point cloud are matched, the pose of each scanning station is calculated, and all point clouds are unified to the global coordinate system. Multi-source point cloud stitching is then completed to obtain the original point cloud data. The real-time atmospheric refractive index is calculated based on the measurement environment parameters to correct the laser ranging error. Combined with the component temperature distribution data, thermal deformation is corrected through scaling formulas or finite element methods. Environmental compensation is then performed on the original point cloud data to obtain high-precision point cloud data.
[0043] For example, to eliminate the interference of environmental factors on measurement accuracy, real-time compensation is performed on the original point cloud data to generate high-precision point cloud data, which includes two dimensions: atmospheric refractive index correction and thermal deformation compensation. First, the temperature, humidity, and air pressure data of the measurement area, as well as the multi-point temperature distribution data of the component, are acquired in real time through a sensor network deployed at the measurement site.
[0044] Atmospheric refractive index compensation: Based on the collected temperature, humidity, and air pressure data, the real-time atmospheric refractive index is calculated to correct the laser ranging value. The specific formula and calculation process are as follows: First, calculate the real-time air refractive index n using an empirical formula: Where, n: corrected air refractive index (dimensionless); P: atmospheric pressure, in hectopascals (hPa), directly measured by a barometer; T: thermodynamic temperature, in Kelvin (K), converted from Celsius temperature measured by a temperature sensor; e: water vapor partial pressure, in hectopascals (hPa), calculated based on relative humidity measured by a humidity sensor and saturated water vapor pressure at the current temperature; Q: average humidity distribution weight along the measurement path, obtained by weighted averaging of data from multiple humidity sensors arranged along the measurement path, or approximately by taking the arithmetic mean of water vapor partial pressures along the path.
[0045] The second step is to compensate for the original laser ranging value based on the inverse relationship between the speed of light and the refractive index. The compensation formula is as follows: ;in, : The distance measurement value after compensation; Original measured distance value; : Refractive index under standard atmospheric conditions, fixed at 1.000292; n: Real-time atmospheric refractive index calculated above. This compensation is applied to the three-dimensional coordinate calculation of each measuring point to eliminate the ranging error introduced by changes in atmospheric conditions.
[0046] Thermal deformation compensation: Based on the component material properties and the measured temperature field, thermal deformation correction is performed on the point cloud coordinates, divided into two scenarios: when the component temperature distribution is uniform, the temperature difference is calculated based on the linear expansion coefficient α of the component material, with 20℃ as the standard reference temperature. ( (Based on the average temperature of the component), the point cloud coordinates are corrected using a scaling formula: ;in, : Final point cloud coordinates after thermal deformation correction; : Point cloud coordinates after refractive index compensation; α: Linear expansion coefficient of the component material; ΔT: Difference between the measured average temperature of the component and the standard reference temperature.
[0047] When the temperature distribution of a component is uneven or it is composed of multiple materials, a finite element-based thermal deformation analysis method is adopted. The specific steps are as follows: ① Based on the theoretical three-dimensional model of the component, a finite element mesh model is established; ② The temperature values measured by multiple temperature sensors are mapped to the nodes of the finite element mesh through interpolation to form the temperature field boundary conditions; ③ The thermoelastic equations are solved to obtain the displacement vector of each mesh node; ④ Based on the node displacement, the point cloud coordinates are interpolated and corrected to accurately restore the actual geometric shape of the component under non-uniform temperature field.
[0048] The technical solution of this invention achieves full surface data coverage of components through multi-source point cloud stitching, and eliminates systematic errors caused by temperature, humidity, air pressure and thermal deformation through dual-dimensional environmental compensation, which significantly improves the accuracy and reliability of point cloud data and provides solid data support for subsequent high-precision installation simulation.
[0049] S120. Based on the high-precision point cloud data component and the corresponding measured 3D model of the component to be installed, the measured 3D model and the theoretical 3D model are spatially registered to obtain the model dataset.
[0050] The measured 3D model can be a continuous triangular mesh surface model representing the actual geometric shape of the component to be installed. The theoretical 3D model can be an ideal shape model of the component built based on the design drawings. Spatial registration is the process of accurately aligning the measured and theoretical models in the same coordinate system. The model dataset can be understood as a collection of data containing the registered model and geometric deviation information, used for subsequent pose calculation and accuracy analysis.
[0051] Specifically, high-precision point cloud data can be denoised and simplified, and a measured 3D model can be generated using the Poisson reconstruction algorithm. First, the two models are coarsely registered by aligning the principal inertial axes, and then finely registered by using the iterative nearest point algorithm. The geometric feature deviation between the two models is calculated, and the registered model and deviation information are integrated into a model dataset.
[0052] For example, such as Figure 3 As shown, the high-precision point cloud data after environmental compensation is transformed into a continuous, high-fidelity measured 3D model. Specifically, it is divided into three processing steps: (1) Point cloud denoising: The statistical outlier removal algorithm is used to remove noise points. Specifically, it is implemented as follows: ① For each sampling point in the point cloud, the nearest 30 points in its neighborhood are searched through spatial index, and the average Euclidean distance between the point and all neighborhood points is calculated; ② The average neighborhood distance of all sampling points is calculated, and the global mean and standard deviation are obtained; ③ The threshold is set as "global mean + 2 times standard deviation". If the average neighborhood distance of a point exceeds the threshold, it is judged as flying point noise and removed from the point cloud.
[0053] (2) Point cloud simplification: Voxel mesh downsampling technology is used to compress the amount of point cloud data while retaining key geometric features. Specifically, the following steps are taken: ① Determine the three-dimensional bounding box based on the overall spatial range of the point cloud and divide the space into regular cubic meshes (voxels) with a side length of 5 mm; ② For each non-empty voxel mesh, calculate the average coordinate of all points in the mesh and use this average value as the representative point to replace all the original point clouds in the mesh; ③ Finally, the amount of point cloud data is compressed to about 15% of the original size, balancing the model accuracy and the computational efficiency of subsequent simulations.
[0054] (3) Surface reconstruction: The Poisson reconstruction algorithm is used to transform the discrete point cloud into a continuous triangular mesh surface to generate a measured three-dimensional model. Specifically, the following steps are taken: ① Normal vector estimation: For each point, its neighboring points are searched to form a local point set. The covariance matrix of the local point set is calculated, and the eigenvector corresponding to the smallest eigenvalue is taken as the initial estimate of the normal vector of the point. The normal direction of all points is unified through the normal propagation algorithm. ② An adaptive octree is constructed, and the Poisson equation is solved based on the position and normal vector of the point to obtain the indicator function. ③ The isosurface of the indicator function is extracted to generate a smooth triangular mesh surface, which is the measured three-dimensional model that represents the actual geometric shape of the component. The surface accuracy of the model can reach 0.1 mm.
[0055] Based on the above technical solution, the measured three-dimensional model corresponding to the component to be installed is based on the high-precision point cloud data component, including: performing statistical outlier removal and denoising on the high-precision point cloud data to obtain the first point cloud data, performing voxel mesh downsampling on the first point cloud data to determine the second point cloud data, and performing Poisson surface reconstruction based on the second point cloud data to obtain the measured three-dimensional model corresponding to the component to be installed.
[0056] Statistical outlier removal denoising can be a method based on neighborhood distance statistics to remove noisy points from a point cloud. Voxel mesh downsampling can be understood as a technique for compressing point cloud data while preserving key geometric features. Poisson surface reconstruction can be an algorithm that transforms discrete point clouds into continuous smooth triangular mesh surfaces.
[0057] Specifically, a statistical outlier removal algorithm can be used to remove flying point noise from high-precision point cloud data to obtain the first point cloud data; the data volume can be compressed by voxel grid downsampling technology to obtain the second point cloud data that balances accuracy and efficiency; the normal vector of the second point cloud data can be estimated and unified, an adaptive octree can be constructed to solve the Poisson equation, and the isosurface can be extracted to generate the measured 3D model.
[0058] The technical solution of this invention improves the purity of point cloud data by denoising, balances data volume and computational efficiency by downsampling, and generates a high-fidelity continuous surface model by Poisson reconstruction, ensuring that the measured 3D model can accurately restore the actual shape of the component, and providing high-quality model support for subsequent registration and pose calculation.
[0059] Based on the above technical solution, the measured 3D model and the theoretical 3D model are spatially registered to obtain a model dataset, including: performing coarse registration processing on the measured 3D model and the theoretical 3D model using the principal inertial axis alignment algorithm to obtain the first model data; performing fine registration on the first model data using the iterative nearest point algorithm to obtain the second model data; determining the geometric feature deviation between the measured 3D model and the theoretical 3D model based on the second model data; and using the geometric feature deviation and the second model data as the model dataset.
[0060] Among them, the principal inertial axis alignment algorithm can be a coarse registration method that achieves initial alignment by matching the principal inertial axes of the model. The Iterative Closest Point (ICP) algorithm can be understood as a fine registration method that achieves high-precision alignment by iteratively optimizing the distance between corresponding points. Geometric feature deviation can be used to quantify the spatial differences between the measured model and the theoretical model in key features and surfaces, and is used to quantify manufacturing and deformation errors.
[0061] Specifically, the covariance matrix and eigenvectors of the measured and theoretical models can be calculated separately. The principal inertial axes of the two models can be aligned by rotation and translation to obtain the first model data. Based on the first model data, the corresponding points are iteratively searched and the optimal transformation matrix is solved to complete the fine registration and obtain the second model data. The spatial distance between the key features and surface vertices of the two models is calculated to determine the geometric feature deviation, and then the deviation is integrated with the second model data to form a model dataset.
[0062] For example, the reconstructed measured 3D model is spatially registered with a pre-stored theoretical design 3D model with high precision to establish a spatial mapping relationship between the two. This process involves three steps:
[0063] (1) Coarse registration: By aligning the principal inertial axes, a good initial pose is provided for fine registration. Specifically, the following steps are taken: ① Read the set of mesh vertices of the measured model and the theoretical model respectively, calculate their respective covariance matrices, and solve for the eigenvectors; ② The eigenvector corresponding to the largest eigenvalue is the principal inertial axis of the model, and the second largest eigenvalue corresponds to the second principal axis; ③ Through rotation and translation transformation, make the principal axis direction of the measured model consistent with the principal axis of the theoretical model, and make their centroids coincide, thus completing the coarse registration.
[0064] (2) Fine registration: The Iterative Closest Point (ICP) algorithm is used to achieve high-precision spatial alignment between the two models. Specifically, the following steps are taken: ① Using the vertex of the measured model after coarse registration as the reference, the nearest point is searched on the surface of the theoretical model to construct a set of corresponding point pairs; ② Based on the corresponding point pairs, the optimal rigid body transformation matrix (rotation matrix + translation vector) is solved to minimize the sum of squared Euclidean distances of all corresponding point pairs; ③ The measured model is updated according to the solved transformation matrix, and the iterative process of "searching for corresponding points - solving the transformation matrix" is repeated until the transformation difference between the two iterations is less than the preset threshold or the maximum number of iterations is reached to complete the fine registration.
[0065] (3) Difference analysis: Quantitative analysis of the geometric deviation between the two models. Specifically, the following is implemented: ① Identify the key features common to the two models through a geometric feature detection algorithm, including the central axis, bolt hole center point, flange face, corner point, etc.; ② Automatically calculate the spatial position deviation after registration of these features, and calculate the distance between each vertex on the model surface and the corresponding point on the theoretical model; ③ After normalizing the distance value, map it to the color space and present the manufacturing error, welding deformation or structural distortion of the model surface in the form of a color distribution map.
[0066] The technical solution of this invention achieves high-precision spatial alignment between the measured and theoretical models by combining coarse and fine registration. At the same time, it quantifies geometric feature deviations, providing accurate deviation references for subsequent pose calculations and ensuring the targeted nature of installation pose adjustments.
[0067] S130. Determine the optimal six-degree-of-freedom installation pose parameters corresponding to the component to be installed based on the model dataset.
[0068] The optimal six-DOF installation pose parameters can be the three-dimensional translation and three-dimensional rotation angles (Eulerian angles) that enable precise component installation. The key feature point set can be understood as the set of feature points that affect assembly accuracy. The weighting coefficients can be considered as differential coefficients assigned based on the importance of feature points, used to prioritize the alignment of key features.
[0069] Specifically, a set of key feature points can be formed by extracting sampling points of the docking surface, center points of fastening bolt holes, and centroid points of components from the model dataset, and weight coefficients can be assigned according to their importance; a weighted least squares objective function can be constructed, and the optimal rotation matrix and translation vector can be solved through singular value decomposition, which can then be converted into six-degree-of-freedom installation pose parameters.
[0070] For example, such as Figure 4 As shown, the optimal installation pose parameters of the component to be installed (3 translations + 3 rotation angles in 3D space, i.e., six degrees of freedom) are calculated. The specific implementation process is as follows:
[0071] (1) Feature point extraction: Extract a set of key feature points from the registered measured 3D model and theoretical 3D model, including at least: ① sampling points of the docking surface: generate grid points at fixed intervals in the docking surface area of the theoretical model, and map them to the measured model through nearest point search to obtain the corresponding measured point set; ② center point of fastening bolt hole: extract the hole edge points in the point cloud through edge detection, fit the spatial circle using the least squares method, and calculate the center coordinates of the circle; ③ centroid of component: calculate the centroid position of the component as a whole based on the weighted calculation of the model grid surface.
[0072] (2) Weighting: Based on the influence of each feature point on the assembly accuracy and the measurement accuracy, a differentiated weight coefficient is assigned to each feature point. All weight values are normalized and summed to 1. Among them: the center point of the fastening bolt hole directly determines the connection accuracy of the component and is given the highest weight; the sampling point of the mating surface affects the fitting accuracy and is given a medium weight; the centroid of the component is only used for overall pose reference and is given a low weight.
[0073] (3) Construction of objective function: Construct a weighted least squares objective function to measure the degree of deviation between the measured model feature points after rigid body transformation and the corresponding reference points of the theoretical model. The function expression is as follows: Where R is a 3×3 rotation matrix describing the orientation adjustment of the component; t is a 3×1 translation vector describing the positional movement of the component. : The coordinate vector of the i-th feature point on the measured model; : The coordinate vector of the i-th corresponding reference point in the theoretical model; The weight coefficients of the i-th feature point satisfy the following conditions: M: Total number of feature points; : Euclidean distance norm. The smaller the function value, the closer the transformed pose is to the ideal installation state.
[0074] (4) Pose calculation: The singular value decomposition (SVD) method is used to solve the above least squares problem to obtain the optimal pose parameters. The specific calculation process is as follows: ① Calculate the weighted centroids of the two sets of points: Weighted centroids of the measured point set: Theoretical point set weighted centroid: ② Center the two sets of points to eliminate the effects of translation: ③ Construct the weighted covariance matrix: ④ Perform singular value decomposition on the covariance matrix H: Where U and V are 3×3 orthogonal matrices, and Σ is a 3×3 diagonal matrix. ⑤ Calculate the optimal rotation matrix: If the determinant of R obtained from the decomposition is negative, the last column of V needs to be inverted to ensure that R belongs to the rotation group (determinant is 1). ⑥ Calculate the optimal translation vector: ⑦ The rotation matrix R is decomposed into Euler angles of rotation around the spatial coordinate axes, and the translation vector t directly corresponds to the movement in three directions, which is finally converted into pose parameters that can guide on-site construction.
[0075] Based on the above technical solution, the optimal six-degree-of-freedom installation pose parameters corresponding to the component to be installed are determined based on the model dataset, including: extracting a set of key feature points from the model dataset and assigning weight coefficients to each feature point in the set of key feature points; constructing a weighted least squares objective function based on the set of key feature points and the weight coefficients, and solving the weighted least squares objective function to obtain the optimal six-degree-of-freedom installation pose parameters.
[0076] The key feature point set includes sampling points on the mating surface, center points of fastening bolt holes, and centroid points of components.
[0077] Specifically, a set of key feature points can be extracted from the registered measured and theoretical models, including sampling points on the mating surface, center points of fastening bolt holes, and centroid points of components. The center points of bolt holes are assigned the highest weight, the sampling points on the mating surface are assigned a medium weight, and the centroid points are assigned a low weight. A function is constructed with the objective of minimizing the sum of squared weighted distances of the feature points. The optimal rotation matrix and translation vector are solved through singular value decomposition, and then converted into three-dimensional translations and Euler angles to obtain the optimal six-degree-of-freedom installation pose parameters.
[0078] The technical solution of this invention highlights the importance of core features through differentiated weight allocation, and the weighted least squares algorithm ensures the accuracy of pose calculation, making the calculation results more in line with actual assembly requirements, and effectively improving the accuracy and reliability of component installation.
[0079] S140. Generate installation instructions corresponding to the component to be installed based on the optimal six-degree-of-freedom installation pose parameters.
[0080] The installation instructions can be specific commands guiding on-site equipment to perform installation operations, including hydraulic pump station pressure, jack extension / retraction stroke, CNC cutting parameters, etc. Dynamic collision detection can be understood as simulating the real-time detection of whether components interfere with surrounding obstacles during installation. Machining allowance distribution data can be visualized data used to quantify the spatial distribution of the closure section's allowance. The optimal collision-free installation path is an installation path that balances safety, accuracy, and efficiency.
[0081] Specifically, based on the optimal pose parameters and preset path, the installation process can be simulated using a bounding box tree collision detection algorithm to obtain collision detection results; the machining allowance of the closure segment can be quantitatively analyzed based on the model dataset to generate allowance distribution data; combining the collision detection results and allowance data, the optimal collision-free installation path can be planned using the artificial potential field method; and based on the path and pose parameters, directly executable installation and construction commands such as hydraulic pump station pressure and jack extension stroke can be calculated and generated.
[0082] The technical solution of this invention avoids installation risks in advance through dynamic collision detection, replaces manual grinding and trial and error with quantitative analysis of margin, and automatically generates construction instructions to connect virtual simulation and on-site construction, which greatly improves construction efficiency and safety and reduces reliance on human experience.
[0083] Based on the above technical solution, installation instructions corresponding to the component to be installed are generated according to the optimal six-degree-of-freedom installation pose parameters. This includes: performing dynamic collision detection of the installation process based on the optimal six-degree-of-freedom installation pose parameters and the preset installation path to obtain collision detection results; performing quantitative analysis of the machining allowance of the assembly section based on the model dataset to obtain machining allowance distribution data; determining the optimal collision-free installation path based on the collision detection results and the machining allowance distribution data; and generating installation instructions corresponding to the component to be installed based on the optimal collision-free installation path and the optimal six-degree-of-freedom installation pose parameters.
[0084] Specifically, a multi-level bounding box tree can be constructed for the measured model and surrounding obstacles. The model is moved step by step along a preset path to perform dynamic collision detection, and the collision detection results are obtained by locking the conflict area. The distance of the measured model is measured along the normal direction of the theoretical model to generate color-coded machining allowance distribution data. Combining the collision detection results and allowance data, a collision-free avoidance path is generated by the artificial potential field method to determine the optimal installation path. Based on the optimal path and pose parameters, the hydraulic pump station pressure, the extension stroke of the attitude adjustment jack, and the CNC cutting parameters are calculated to generate standardized installation and construction instructions.
[0085] For example, such as Figure 5As shown, based on the optimal installation pose, the complete dynamic process of component installation is simulated in virtual space to detect collision risks in advance and quantitatively evaluate the processing allowance of the closing section. The specific implementation is as follows: (1) Bounding box tree construction: For the actual three-dimensional model and the obstacle models such as the base, support structure, and adjacent components in the surrounding environment, a multi-level axis-aligned bounding box (AABB) tree is constructed. The specific implementation is as follows: ① The set of triangular facets of the model is used as input, and the space is recursively divided into smaller subsets. The segmentation stops when the number of triangular facets in the subset is less than 50 or the bounding box size is less than 10 mm; ② For each subset, the minimum and maximum values in the X, Y, and Z directions are calculated, and a hexahedral bounding box aligned with the coordinate axes is constructed; ③ All bounding boxes are organized into a tree structure according to the spatial hierarchy. The root node corresponds to the entire model, and the leaf node corresponds to the minimum facet set for subsequent rapid collision detection.
[0086] (2) Collision detection: Real-time collision detection is implemented during the virtual installation process. Specifically, the following steps are taken: ① The measured 3D model is moved step by step along the preset installation path with a fixed step size of 5 mm; ② For each step, a depth-first traversal method is used to traverse the bounding box tree of the measured model and the obstacle model at the same time: if the bounding boxes of two nodes do not intersect, the branches are pruned directly; if they intersect and reach the leaf node, the intersection calculation of the triangular facets in the leaf node is further performed; ③ Once spatial overlap is detected, the conflict area is locked immediately, and the collision depth (the maximum distance between the component and the obstacle) is calculated using the separating axis theorem or the GJK algorithm, providing a quantitative basis for subsequent path adjustment.
[0087] (3) Allowance Analysis: For key parts such as flanges and closing sections with machining allowances, the allowances are quantified and visualized. Specifically, the following steps are taken: ① Mark the boundary of the key parts on the theoretical model and find the corresponding point set on the measured model through the nearest point search algorithm; ② For each sampling position, measure the vertical distance from the measured model to the theoretical model along the normal direction of the theoretical model surface: if the measured point is outside the theoretical model, the gap is positive (too much allowance); if the measured point is inside the theoretical model, the gap is negative (insufficient allowance); ③ Statistically analyze all measured values and generate an allowance distribution chromatogram in the form of color coding: blue represents excessive allowance that needs to be cut, red represents insufficient allowance that needs to be welded, and green represents allowance that meets the tolerance requirements, providing accurate guidance for on-site processing.
[0088] Furthermore, such as Figure 6As shown, the position parameters and interference margin analysis results are transformed into construction instructions that can be directly executed on site, realizing the connection between virtual simulation and physical operation. The specific implementation is as follows: (1) Path planning and decision-making: Generate and select the optimal installation path. Specifically, ① Generate several alternative installation paths through the path planning algorithm to form a candidate path set; ② Construct a multi-criteria decision model with the load-bearing capacity of the hoisting equipment, the swing amplitude limit of the wire rope, and the safe operating radius on site as the core hard constraints; ③ Use the analytic hierarchy process to quantify the safety, accuracy, and efficiency of each alternative path, and finally output the path that meets all hard constraints and has the highest comprehensive score as the optimal execution path.
[0089] (2) Path avoidance: When a collision risk is detected, a collision-free avoidance path is automatically generated. This is achieved using an artificial potential field method, specifically: a potential field is constructed in virtual space: a gravitational field is generated at the target installation location, with the gravitational force pointing towards the target and its magnitude proportional to the distance. The formula is: ;in, 1. Here, d is the gravitational gain coefficient, and d is the distance from the current position to the target position. 2. A repulsive field is generated around the obstacle. The direction of the repulsive force is away from the obstacle, and its magnitude is inversely proportional to the distance to the obstacle. It only takes effect when the distance is less than the influence radius ρ0. The formula is: ;in, ρ is the repulsive force gain coefficient, and ρ is the distance from the current position to the obstacle. ③ Calculate the resultant force: The component is guided to move along the direction of the resultant force, and the calculation is iteratively performed with a fixed step size to finally generate a collision-free avoidance path.
[0090] (3) Command generation: Based on the final determined installation path and position parameters, three types of core construction commands are calculated and generated. The specific calculation process is as follows: ① The pressure setpoint of the hydraulic pump station is calculated based on the component weight G, lifting height, and hydraulic system parameters to determine the required pressure of the hydraulic cylinder. The formula is: Where n is the number of parallel hydraulic cylinders, A is the effective area of a single hydraulic cylinder piston, and η is the mechanical efficiency of the hydraulic system. ② The extension / retraction stroke of the attitude adjustment jacks is calculated based on the translation amounts Δx, Δy, and Δz of the optimal posture and the rotation matrix R. Specifically: Let the coordinates of the support point of the j-th jack in the initial coordinate system of the component be... After pose transformation, the new position of this point in the global coordinate system is: ;in, Let be the translation vector. The extension or retraction of the jack is . The change in vertical distance from the fixed point of its base is converted into a specific extension stroke based on the mechanical structure. ③ The CNC cutting process parameters are based on the allowance distribution chromatogram to identify the area to be cut, extract the area boundary and allowance value; according to the component material, tool type and machine tool performance, the cutting depth, feed rate and tool path are determined, and standard format CNC code or a machining notification sheet that can be directly referenced by the fitter on site is generated.
[0091] The technical solution of this invention involves acquiring measured data of the component corresponding to the component to be installed, and determining high-precision point cloud data corresponding to the component based on the measured data. Based on the measured 3D model corresponding to the high-precision point cloud data component and the component to be installed, the measured 3D model and the theoretical 3D model are spatially registered to obtain a model dataset. Based on the model dataset, the optimal six-degree-of-freedom installation pose parameters corresponding to the component to be installed are determined. Installation instructions corresponding to the component to be installed are generated based on the optimal six-degree-of-freedom installation pose parameters. Based on the above technical solution, by acquiring measured data corresponding to large-size components and determining the installation pose parameters and generating corresponding control instructions based on the model dataset obtained by registering the 3D model and the theoretical 3D model, key features are prioritized for alignment. Furthermore, the entire process dynamically simulates and predicts installation risks, achieving autonomous control of accuracy through closed-loop feedback. This improves the installation accuracy and one-time assembly success rate of large-size components, reduces construction risks, and decreases reliance on manual experience.
[0092] In one possible implementation of the present invention Figure 7 This is a flowchart of a large-size component installation simulation method provided by an embodiment of the present invention. The embodiment further describes the technical solution of the large-size component installation simulation method, such as... Figure 7 As shown, the method includes:
[0093] S710. Calculate the installation accuracy index of the component to be installed based on the installation simulation results.
[0094] The installation accuracy indicators include the physical clearance of the connection interface, the coincidence of the fastening bolt holes, and the form and position tolerances of the components. The installation simulation results can be virtual installation results based on the optimal pose and path. The installation accuracy indicators can be parameters used to quantify the installation quality. The physical clearance of the connection interface can be understood as the maximum vertical distance between the measured and theoretical model mating surfaces. The coincidence of the fastening bolt holes is the spatial Euclidean distance between the center points of the corresponding bolt holes. The form and position tolerances of the components are geometric error values such as straightness and flatness.
[0095] Specifically, the installation accuracy index of the component to be installed is calculated based on the installation simulation results. For example, based on the installation simulation results, the physical gap of the mating surface can be calculated using the nearest point search algorithm, the overlap can be obtained by solving the spatial distance of the bolt hole center point, and the form and position tolerance can be calculated by fitting the ideal geometric elements using the least squares method to form a complete installation accuracy index.
[0096] S720. The installation accuracy index is compared with the preset error threshold to obtain the accuracy evaluation result.
[0097] The preset error threshold can be a critical accuracy value set according to the assembly process requirements, such as 0.5 mm. The accuracy evaluation result can be a judgment result used to characterize whether the installation accuracy meets the standard, including two categories: qualified and out of tolerance.
[0098] Specifically, the installation accuracy indicators are compared with preset error thresholds to obtain the accuracy assessment results. For example, the calculated physical gap of the connection interface, the overlap of fastening bolt holes, and the form and position tolerance of components can be compared with preset error thresholds respectively. If all indicators do not exceed the threshold, they are judged as qualified; otherwise, they are judged as out of tolerance, and the accuracy assessment results are obtained.
[0099] S730. If the accuracy assessment result is out of tolerance, determine the source of error based on the accuracy assessment result and recalculate the control parameters corresponding to the source of error.
[0100] The sources of error can be the reasons that cause installation accuracy to deviate, including measurement data deviations and insufficient pose calculation accuracy. Control parameters can be understood as the core adjustment parameters corresponding to the sources of error, including measurement data and feature point weights.
[0101] Specifically, when the accuracy assessment result is out of tolerance, the source of error can be located according to the type of out-of-tolerance index: if the gap or form and position tolerance is out of tolerance, it is determined to be a measurement data deviation; if the hole position coincidence is out of tolerance, it is determined to be insufficient pose calculation accuracy. For the source of error, the corresponding module is traced back, data is re-acquired or the feature point weights are adjusted, and the control parameters are recalculated.
[0102] For example, such as Figure 6As shown, the accuracy of the simulated installation results is verified, and self-correction and closed-loop control are achieved. The specific implementation is as follows: (1) Accuracy evaluation: After the simulated installation is completed, the core accuracy index is calculated as the residual error, which includes: ① Physical gap of the connection interface: The nearest point search algorithm is used to calculate the vertical distance between the actual model and the corresponding surface of the theoretical model, and the maximum value is taken as the gap value; ② Overlap of fastening bolt holes: The spatial Euclidean distance between the center point of the corresponding bolt hole in the actual model and the theoretical model is calculated as the hole position deviation value; ③ Overall form and position tolerance of the component: Based on the feature point set of the actual model, the least squares method is used to fit the ideal geometric elements (straight lines, planes, etc.), and the distance from each point to the fitted element is calculated to obtain the form and position error values such as straightness and flatness.
[0103] (2) Backtracking control: A preset error threshold of 0.5 mm is set. If any residual error exceeds this threshold, the simulation result is deemed unqualified. The backtracking mechanism is automatically triggered according to the type of deviation: ① If the gap or geometric tolerance is out of tolerance, the error source is determined to be the deviation of the original measurement data. Backtracking is performed to the multi-source high-precision geometric information acquisition module and the environmental parameter real-time compensation module to re-acquire data and complete the compensation correction; ② If the hole position deviation is out of tolerance, the error source is determined to be insufficient pose calculation accuracy. Backtracking is performed to the six-degree-of-freedom spatial pose calculation module to readjust the feature point weights and solve the optimal installation pose; The system records the number of backtracking attempts. If the preset maximum number of attempts is reached and the accuracy requirement is still not met, the backtracking is stopped and an alarm is issued to prompt manual intervention.
[0104] This invention integrates three types of equipment—laser scanning, visual measurement, and total station—to achieve full-size, high-precision data acquisition. Simultaneously, through atmospheric refractive index correction and thermal deformation compensation, it eliminates environmental interference from both the measurement process and the component itself, improving the accuracy of the original point cloud data to the 0.1 mm level. This fundamentally solves the core problem of "inaccurate measurement" of large-size components in complex environments.
[0105] This invention assigns differentiated weights to meet the precision requirements of different assembly features, prioritizing core mating parts such as bolt holes. It solves for the optimal pose using a weighted least squares objective function, achieving priority alignment of key features and completely resolving the problem of "overall alignment but out-of-tolerance in key areas" inherent in traditional solutions, significantly improving the first-time assembly success rate. This invention achieves rapid collision detection throughout the installation process through a multi-level bounding box tree, preemptively identifying collision risks. Simultaneously, by measuring the allowance in the normal direction, it generates a allowance distribution chromatogram, fully quantifying and visualizing the machining allowance of the joining section, replacing the traditional on-site manual grinding and trial-and-error mode, greatly improving construction efficiency and precision. This invention presets a 0.5 mm precision control threshold. Based on the out-of-tolerance type of the simulation results, it automatically locates the source of error and traces back to the corresponding module for correction, achieving system self-verification and self-correction, forming a complete closed-loop precision control, ensuring controllable final installation precision. This invention can directly calculate and generate executable construction commands such as hydraulic pump station pressure, jack extension stroke, and CNC cutting parameters based on simulation results, without the need for manual secondary calculations. It completely opens up the link between digital twin virtual simulation and on-site physical construction, and greatly reduces the reliance on human experience.
[0106] Figure 8 This is a structural schematic diagram of a large-size component installation simulation device provided in an embodiment of the present invention. Figure 8 As shown, the device includes: a data acquisition module 810, a three-dimensional model building module 820, a pose parameter determination module 830, and a construction instruction generation module 840.
[0107] The data acquisition module 810 is used to acquire the measured data of the component corresponding to the component to be installed, and to determine the high-precision point cloud data corresponding to the component to be installed based on the measured data of the component. The measured data of the component includes the original point cloud data, measurement environment parameters and component temperature distribution data.
[0108] The 3D model building module 820 is used to obtain a model dataset by spatially registering the measured 3D model with the theoretical 3D model based on the high-precision point cloud data component and the component to be installed.
[0109] The pose parameter determination module 830 is used to determine the optimal six-degree-of-freedom installation pose parameters corresponding to the component to be installed based on the model dataset.
[0110] The construction instruction generation module 840 is used to generate installation construction instructions corresponding to the components to be installed based on the optimal six-degree-of-freedom installation pose parameters.
[0111] Based on the above technical solution, a 3D model building module is used to perform statistical outlier removal and noise reduction on high-precision point cloud data to obtain the first point cloud data, perform voxel mesh downsampling on the first point cloud data to determine the second point cloud data, and perform Poisson surface reconstruction based on the second point cloud data to obtain the measured 3D model corresponding to the component to be installed.
[0112] Based on the above technical solution, the 3D model building module is used to perform coarse registration processing on the measured 3D model and the theoretical 3D model through the principal inertial axis alignment algorithm to obtain the first model data; to perform fine registration on the first model data using the iterative nearest point algorithm to obtain the second model data; to determine the geometric feature deviation between the measured 3D model and the theoretical 3D model based on the second model data; and to use the geometric feature deviation and the second model data as the model dataset.
[0113] Based on the above technical solution, the pose parameter determination module is used to extract a set of key feature points from the model dataset and assign weight coefficients to each feature point in the set of key feature points. The set of key feature points includes sampling points on the docking surface, center points of fastening bolt holes, and centroid points of components. A weighted least squares objective function is constructed based on the set of key feature points and the weight coefficients, and the optimal six-degree-of-freedom installation pose parameters are obtained by solving the weighted least squares objective function.
[0114] Based on the above technical solution, the construction instruction generation module is used to perform dynamic collision detection of the installation process based on the optimal six-degree-of-freedom installation posture parameters and the preset installation path to obtain collision detection results; perform quantitative analysis of the machining allowance of the installation closure section based on the model dataset to obtain machining allowance distribution data; determine the optimal collision-free installation path based on the collision detection results and machining allowance distribution data; and generate installation construction instructions corresponding to the component to be installed based on the optimal collision-free installation path and the optimal six-degree-of-freedom installation posture parameters.
[0115] Based on the above technical solution, the device also includes: a feedback control module, used to calculate the installation accuracy index of the component to be installed based on the installation simulation results, wherein the installation accuracy index includes the physical gap of the connection interface, the overlap of the fastening bolt holes, and the form and position tolerance of the component; the installation accuracy index is compared with a preset error threshold to obtain the accuracy evaluation result; if the accuracy evaluation result is out of tolerance, the source of error is determined based on the accuracy evaluation result, and the control parameters corresponding to the source of error are recalculated.
[0116] Based on the above technical solution, the data acquisition module is used to complete multi-source point cloud stitching based on the measured data of the component to obtain the original point cloud data of the component to be installed; based on the measured environmental parameters and component temperature distribution data, environmental compensation is performed on the original point cloud data to obtain high-precision point cloud data.
[0117] The technical solution of this invention involves acquiring measured data of the component corresponding to the component to be installed, and determining high-precision point cloud data corresponding to the component based on the measured data. Based on the measured 3D model corresponding to the high-precision point cloud data component and the component to be installed, the measured 3D model and the theoretical 3D model are spatially registered to obtain a model dataset. Based on the model dataset, the optimal six-degree-of-freedom installation pose parameters corresponding to the component to be installed are determined. Installation instructions corresponding to the component to be installed are generated based on the optimal six-degree-of-freedom installation pose parameters. Based on the above technical solution, by acquiring measured data corresponding to large-size components and determining the installation pose parameters and generating corresponding control instructions based on the model dataset obtained by registering the 3D model and the theoretical 3D model, key features are prioritized for alignment. Furthermore, the entire process dynamically simulates and predicts installation risks, achieving autonomous control of accuracy through closed-loop feedback. This improves the installation accuracy and one-time assembly success rate of large-size components, reduces construction risks, and decreases reliance on manual experience.
[0118] The large-size component installation simulation device provided in the embodiments of the present invention can execute the large-size component installation simulation method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method.
[0119] Figure 9 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.
[0120] like Figure 9 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.
[0121] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.
[0122] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as large-scale component mounting simulation methods.
[0123] In some embodiments, the large-scale component mounting simulation method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the large-scale component mounting simulation method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the large-scale component mounting simulation method by any other suitable means (e.g., by means of firmware).
[0124] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.
[0125] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.
[0126] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.
[0127] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).
[0128] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.
[0129] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.
[0130] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.
[0131] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.
Claims
1. A method for simulating the installation of large-size components, characterized in that, include: Acquire the actual measured data of the component corresponding to the component to be installed, and determine the high-precision point cloud data corresponding to the component to be installed based on the actual measured data of the component, wherein the actual measured data of the component includes the original point cloud data, measurement environment parameters and component temperature distribution data; Based on the high-precision point cloud data component and the measured three-dimensional model corresponding to the component to be installed, the measured three-dimensional model and the theoretical three-dimensional model are spatially registered to obtain the model dataset. Based on the model dataset, determine the optimal six-DOF installation pose parameters corresponding to the component to be installed; Based on the optimal six-degree-of-freedom installation pose parameters, an installation command corresponding to the component to be installed is generated. The step of determining the high-precision point cloud data corresponding to the component to be installed based on the measured data of the component includes: correcting the original point cloud data by atmospheric refractive index correction and thermal deformation compensation to obtain high-precision point cloud data. The original point cloud data is the coordinate data of discrete points on the surface of the component collected by multi-source measurement equipment. The measurement environment parameters are the temperature, humidity and air pressure data of the measurement site, which are used to eliminate the interference of the environment on the measurement accuracy. The step of spatially registering the measured 3D model with the theoretical 3D model to obtain the model dataset includes: The first model data is obtained by performing coarse registration between the measured 3D model and the theoretical 3D model using the principal inertial axis alignment algorithm; The second model data is obtained by finely registering the first model data using the iterative nearest point algorithm; Based on the second model data, determine the geometric feature deviation between the measured 3D model and the theoretical 3D model; The geometric feature deviation and the second model data are used as the model dataset; The determination of the optimal six-DOF installation pose parameters corresponding to the component to be installed based on the model dataset includes: The model dataset is used to extract sampling points of the docking surface, center points of fastening bolt holes, and centroid points of components to form a set of key feature points, and weight coefficients are assigned according to their importance. A weighted least squares objective function is constructed, and the optimal rotation matrix and translation vector are solved by singular value decomposition, which is then converted into the optimal six-degree-of-freedom installation pose parameters.
2. The method according to claim 1, characterized in that, The measured 3D model based on the high-precision point cloud data component and the component to be installed includes: Statistical outlier removal and denoising are performed on the high-precision point cloud data to obtain the first point cloud data. Voxel grid downsampling is then performed on the first point cloud data to determine the second point cloud data. Based on the second point cloud data, a Poisson surface reconstruction is performed to obtain a measured three-dimensional model corresponding to the component to be installed.
3. The method according to claim 1, characterized in that, The step of determining the optimal six-DOF installation pose parameters corresponding to the component to be installed based on the model dataset includes: Extract a set of key feature points from the model dataset and assign a weight coefficient to each feature point in the set of key feature points. The set of key feature points includes sampling points on the mating surface, center points of fastening bolt holes, and centroid points of components. Based on the key feature point set and the weight coefficients, a weighted least squares objective function is constructed, and the optimal six-degree-of-freedom installation pose parameters are obtained by solving the weighted least squares objective function.
4. The method according to claim 1, characterized in that, The step of generating installation instructions corresponding to the component to be installed based on the optimal six-degree-of-freedom installation pose parameters includes: Based on the optimal six-degree-of-freedom installation pose parameters and the preset installation path, dynamic collision detection of the installation process is performed to obtain the collision detection results. Based on the model dataset, quantitative analysis of the machining allowance of the assembly section was performed to obtain the machining allowance distribution data. The optimal collision-free installation path is determined based on the collision detection results and machining allowance distribution data. Based on the optimal collision-free installation path and the optimal six-degree-of-freedom installation pose parameters, an installation command corresponding to the component to be installed is generated.
5. The method according to claim 1, characterized in that, The method further includes: The installation accuracy index of the component to be installed is calculated based on the installation simulation results. The installation accuracy index includes the physical clearance of the connection interface, the overlap of the fastening bolt holes, and the form and position tolerance of the component. The installation accuracy index is compared with a preset error threshold to obtain the accuracy evaluation result; If the accuracy assessment result is out of tolerance, the source of error is determined based on the accuracy assessment result, and the control parameters corresponding to the source of error are recalculated.
6. The method according to claim 1, characterized in that, Based on the measured data of the component, high-precision point cloud data corresponding to the component to be installed is determined, including: Based on the measured data of the components, multi-source point cloud stitching is completed to obtain the original point cloud data of the components to be installed; Based on the measured environmental parameters and component temperature distribution data, environmental compensation is performed on the original point cloud data to obtain the high-precision point cloud data.
7. A large-size component installation simulation device, characterized in that, include: The data acquisition module is used to acquire the measured data of the component corresponding to the component to be installed, and to determine the high-precision point cloud data corresponding to the component to be installed based on the measured data of the component. The measured data of the component includes the original point cloud data, measurement environment parameters and component temperature distribution data. The 3D model building module is used to spatially register the measured 3D model with the theoretical 3D model based on the high-precision point cloud data component and the measured 3D model corresponding to the component to be installed to obtain the model dataset. The pose parameter determination module is used to determine the optimal six-degree-of-freedom installation pose parameters corresponding to the component to be installed based on the model dataset. The construction instruction generation module is used to generate installation construction instructions corresponding to the component to be installed based on the optimal six-degree-of-freedom installation pose parameters. The data acquisition module is used to correct errors in the original point cloud data through atmospheric refractive index correction and thermal deformation compensation to obtain high-precision point cloud data. The original point cloud data is the coordinate data of discrete points on the surface of the component collected by multi-source measurement equipment; the measurement environment parameters are the temperature, humidity and air pressure data of the measurement site, which are used to eliminate the interference of the environment on the measurement accuracy. The 3D model building module is used to perform coarse registration processing on the measured 3D model and the theoretical 3D model using the principal inertial axis alignment algorithm to obtain the first model data; to perform fine registration on the first model data using the iterative nearest point algorithm to obtain the second model data; to determine the geometric feature deviation between the measured 3D model and the theoretical 3D model based on the second model data; and to use the geometric feature deviation and the second model data as the model dataset. The pose parameter determination module is used to extract key feature point sets from the model dataset, including sampling points of the docking surface, center points of fastening bolt holes, and centroid points of components, and assign weight coefficients according to their importance; construct a weighted least squares objective function, solve for the optimal rotation matrix and translation vector through singular value decomposition, and convert it into the optimal six-degree-of-freedom installation pose parameters.
8. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor to enable the at least one processor to perform the large-size component installation simulation method according to any one of claims 1-6.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the large-size component installation simulation method according to any one of claims 1-6.
Citation Information
Patent Citations
Tunnel pipe roof construction dynamic deviation correction method based on laser ranging and optical fiber sensing
CN120928706A
Closed-loop control method and equipment for sectional construction of special-shaped curtain wall steel structure and storage medium
CN121381914A
