Intelligent power plant construction site construction progress safety management method based on model fusion

Abstract

The invention relates to the technical field of construction management, and discloses an intelligent power plant construction site construction progress safety management method based on model fusion. S01, constructing a basic three-dimensional model through static data of the smart power plant; s02, collecting dynamic data of the smart power plant; s03, constructing a dynamic three-dimensional model by using the point cloud data; s04, generating a three-dimensional real scene model; s05, determining an early warning mechanism and early warning prompt information based on the comprehensive deviation rate; according to the method, the static data and the dynamic data are matched with each other, and the three-dimensional live-action model taking the basic three-dimensional model and the dynamic three-dimensional model as fusion objects is constructed, so that the planned construction state of the smart power plant can be intuitively displayed; and multi-dimensional dynamic construction elements such as material stacking, mechanical distribution and structure completion conditions of the current construction site can be reflected in real time, so that an all-dimensional and three-dimensional visual management effect of the construction progress of the intelligent power plant is realized.

Description

Technical Field

[0001] This invention relates to the field of construction management technology, and more specifically, to a method for managing the construction progress and safety of smart power plant construction sites based on model fusion. Background Technology

[0002] The construction of smart power plants is characterized by a wide construction scope, numerous core equipment, complex and overlapping construction procedures, and rigid schedule nodes. Traditional schedule management models can no longer meet the needs of intelligent and refined control. With the development of information technology, the integrated application of building information modeling and 3D modeling technology has provided new tools for engineering construction management and plays a crucial role in improving the real-time performance, accuracy, initiative, and intelligence level of power plant construction engineering management.

[0003] Reference patent application CN116664062A discloses a method, system, and equipment for substation construction management based on BIM and GIS. This includes using a drone equipped with LiDAR and a panoramic camera to simultaneously acquire LiDAR point cloud data and image data of the substation; constructing a 3D reality model of the substation based on the acquired LiDAR point cloud data and image data, combined with the drone's position and attitude data; acquiring the substation's system structure information and establishing a substation BIM model linked to project cost and schedule; establishing a 3D visualization model based on BIM and GIS integration based on the substation's 3D reality model and BIM model to obtain the planned workload at different stages; and monitoring changes in the project workload in real time based on the planned workload at different stages, judging cost and schedule deviations, and issuing timely early warning signals. Existing smart power plants typically focus on periodic static scene construction when managing construction progress. They lack sufficient dimensions for capturing and acquiring dynamic elements such as material stacking and equipment installation that change constantly on the construction site. This makes it difficult to accurately and automatically match and integrate the pre-designed static BIM model with the dynamic construction site data collected in real time by drones. As a result, the constructed 3D model is updated late, fails to accurately reflect the real-time construction status, and is prone to large errors in analysis results and lagging control, thus reducing the effectiveness of real-time dynamic management of the construction progress of smart power plants.

[0004] In view of this, the present invention proposes a model fusion-based method for safe management of construction progress at smart power plant sites to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of existing technologies and to achieve the above objectives, this invention provides the following technical solution: a model fusion-based method for safe management of construction progress at smart power plant sites, comprising: S01: Construct a basic 3D model using static data from a smart power plant, divide the basic 3D model into functional areas, and label the key nodes of the functional areas. The static data includes design parameters and terrain data. S02: Collect dynamic data from the smart power plant, including image data, laser point cloud data, and GPS coordinate data, and process the dynamic data to generate discrete point cloud data; S03: Divide the point cloud data into different construction dynamic elements, and perform Gaussian splashing operation on the point cloud data within the construction dynamic elements to construct a dynamic 3D model; S04: Embed the dynamic 3D model into the basic 3D model for model fusion, generate a 3D reality model, and mark the construction progress. S05: Extract progress parameters from the 3D reality model. Progress parameters include completed area value, construction height value, equipment installation quantity and material consumption progress. Calculate the comprehensive deviation rate and determine the corresponding early warning mechanism and early warning prompt information based on the comprehensive deviation rate.

[0006] Furthermore, the method for constructing the basic 3D model is as follows: Input the design parameters into the BIM simulation software to simulate the BIM design model; Mark the simulated positions with interval distribution on the BIM design model, import the terrain data into the simulated positions one by one, and record the simulated positions with imported terrain data as terrain simulation positions. At the same time, a fusion command is sent to the terrain simulation location to control the terrain simulation location to undergo simulation transformation, upgrading the BIM design model to a basic three-dimensional model.

[0007] Furthermore, the functional areas include the main plant area, boiler installation area, steam turbine area, power distribution equipment area, material storage area, and construction access area; When annotating key nodes, first query the overall progress of the basic 3D model construction. Decompose the overall progress into weekly progress in one-week units. Then, record the weekly progress containing key nodes in any functional area as key progress. Query the completion time of key nodes in each key progress and record it as key time. Use the key time as the node annotation for key nodes.

[0008] Furthermore, the method for generating point cloud data is as follows: Import the laser point cloud data into the point cloud processing software, and use the laser point cloud data as index points to construct a mesh-like spatial index structure; Set the number of neighboring points and standard deviation threshold of the laser point cloud data, match the neighboring points corresponding to the laser point cloud data one by one through the spatial index structure, and calculate the distance threshold from the laser point cloud data to the neighboring points through the statistical filtering algorithm. Laser point cloud data with an average neighborhood distance greater than a distance threshold are recorded as noise points, and the laser point cloud data corresponding to noise points are removed. Randomly select one image data as the reference image, record the point at the center of the reference image as the reference point, and record the remaining image data as the matching image; Set the feature point threshold and matching distance threshold for the image data, and use the reference point as the registration center to perform image registration between the matching image and the reference image using the SIFT feature point matching algorithm. The first iteration number and convergence threshold are set, and the laser point cloud data is rotated and translated by the ICP algorithm. All laser point cloud data are unified into the WGS84 coordinate system. The image data and laser point cloud data are matched and bound with the actual location of the smart power plant by GPS coordinate data to generate discrete point cloud data.

[0009] Furthermore, the construction dynamic elements include material pile elements, construction machinery elements, completed structural elements, and pipeline support elements; The method for dividing construction dynamic elements is as follows: The minimum distance threshold from the point cloud data to the neighboring points is recorded as the density benchmark, and the density benchmark is multiplied by the density conversion coefficient to calculate the density value of the point cloud data. The range between the minimum and maximum density values ​​is denoted as the global range, and the global range is divided into four consecutive sub-ranges to obtain the first sub-range, the second sub-range, the third sub-range, and the fourth sub-range. Point cloud data with density values ​​in the first, second, third, and fourth sub-ranges are aggregated and grouped to form material pile elements, construction machinery elements, completed structure elements, and pipeline support elements.

[0010] Furthermore, the method for constructing dynamic 3D models is as follows: Import the point cloud data from the material pile element, construction machinery element, completed structure element and pipeline support element into the same three-dimensional space in sequence, and mark the spatial coordinates of the point cloud data as (x, y, z) to obtain A three-dimensional spatial points; Gaussian splashing calculations are performed sequentially on A three-dimensional spatial points to calculate the superposition value of the Gaussian splashes at A three-dimensional spatial points; Using Gaussian splash superposition values ​​as the reconstruction benchmark, the Marching Cubes algorithm is used to calculate isosurfaces for A three-dimensional spatial points, extracting surface points on the same plane, and then superimposing and stitching all surface points sequentially to generate a mesh model; Locate the model region on the mesh model that coincides with the spatial location of the image data, and attach the image data to the corresponding model region to map the mesh model texture into a dynamic 3D model.

[0011] Furthermore, the method for generating 3D reality models is as follows: Both the basic 3D model and the dynamic 3D model are converted to the WGS84 coordinate system, and all model points in the basic 3D model and the dynamic 3D model are marked one by one. Mark the spatial coordinates of fixed feature points from the basic 3D model, mark the spatial coordinates of dynamic feature points from the dynamic 3D model, and establish the correspondence between fixed feature points and dynamic feature points; Set the second iteration number and the interior point threshold, and use the RANSAC algorithm to match the fixed feature points and dynamic feature points with corresponding relationships, thereby aligning the spatial coordinates of the model points on the dynamic 3D model with those on the basic 3D model. After removing overlapping meshes that are intersecting from the dynamic 3D model, the remaining dynamic 3D model is embedded into the base 3D model to generate a 3D reality model.

[0012] Furthermore, the calculation method for the overall deviation rate is as follows: The database is used to retrieve the planned area, planned height, planned installation quantity, and planned consumption progress corresponding to the construction progress. The area deviation rate is calculated by taking the absolute value of the difference between the completed area value and the planned area value and comparing it with the planned area value. The absolute value of the difference between the construction height and the planned height is taken and compared with the planned height to calculate the height deviation rate. The installation deviation rate is calculated by taking the absolute value of the difference between the number of equipment installed and the planned number of equipment installed, and comparing it with the planned number of equipment installed. The absolute value of the difference between the real-time consumption progress and the planned consumption progress is taken and compared with the planned consumption progress to calculate the progress deviation rate. The area deviation rate, height deviation rate, installation deviation rate, and schedule deviation rate are weighted and summed to calculate the overall deviation rate.

[0013] Furthermore, the early warning mechanism includes no warning, Level 1 warning, and Level 2 warning; The method for determining the early warning mechanism is as follows: When the overall deviation rate is less than or equal to the first overall deviation threshold, select no warning; When the overall deviation rate is greater than the first overall deviation threshold and less than or equal to the second overall deviation threshold, a Level 1 warning is selected; When the overall deviation rate exceeds the second overall deviation threshold, a level-two warning is selected.

[0014] Furthermore, the method for generating early warning information is as follows: When the early warning mechanism is set to no warning, no early warning message will be generated; When the early warning mechanism is at level one, a system pop-up notification will be issued first, followed by a text message notification to the construction personnel. When the early warning mechanism is at level two, a system pop-up notification is first issued, followed by a text message notification to the construction personnel, then a telephone notification to the construction supervision department, and finally a construction progress analysis report is generated with the comprehensive deviation rate and progress parameters as the report content.

[0015] The technical effects of this invention's model fusion-based smart power plant construction site progress and safety management method are as follows: (1): This invention matches static data with dynamic data and constructs a three-dimensional real-scene model with the basic three-dimensional model and the dynamic three-dimensional model as the fusion objects. It can not only intuitively display the planned construction status of the smart power plant, but also reflect the dynamic construction elements of multiple dimensions such as material stacking, machinery distribution, and structural completion status at the current construction site in real time. This achieves the goal of comprehensive and three-dimensional visual management of the construction progress of the smart power plant, avoiding the problems of low efficiency, large error and lagging control caused by traditional manual inspection and phased management operations, and realizing the real-time dynamic management effect of the construction progress of the smart power plant.

[0016] (2): By introducing statistical filtering, SIFT feature matching and ICP registration into the point cloud data processing, this invention can effectively improve the data quality of multidimensional data by denoising and aligning it, ensuring the fusion accuracy of the basic three-dimensional model and the dynamic three-dimensional model in the model fusion stage, and ensuring that the three-dimensional real scene model can accurately manage the real-time construction progress of the smart power plant, thereby enhancing the adaptability and reliability of the three-dimensional real scene model in complex construction environments. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the method for safe management of construction progress at a smart power plant site based on model fusion, as provided in Embodiment 1 of the present invention. Figure 2 This is a schematic diagram of the modules of the smart power plant construction progress and safety management system based on model fusion provided in Embodiment 2 of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0019] Example 1: Please refer to Figure 1 As shown in this embodiment, the smart power plant construction progress and safety management method based on model fusion includes: S01: Collect static data of the smart power plant, construct a basic 3D model, divide the basic 3D model into functional areas, and label the key nodes of the functional areas. Static data refers to relatively fixed data that is recorded in advance on the construction design drawings during the construction of a smart power plant, as well as data obtained from advance surveys before construction. This data can provide a basis for comparison of the construction progress of a smart power plant at a static level. Specifically, static data includes design parameters and terrain data.

[0020] Design parameters refer to the parameters recorded on the construction design drawings of smart power plants, used to design and plan the construction progress of smart power plants, and provide drawing-level parameter basis for the basic three-dimensional model. In this embodiment, the design parameters include, but are not limited to, seismic resistance level, environmental protection material level, pipeline routing, and load-bearing capacity.

[0021] Topographic data refers to surveying and mapping data on the distribution of topographic and geological structures collected through various surveying equipment, which represents the environmental topography of the smart power plant during its construction. Specifically, topographic data includes, but is not limited to, elevation points, contour lines, topographic maps, and control point coordinates. Topographic maps include scale topographic maps used for macro-site selection and overall layout. Elevation points are used to reflect the elevation information of the earth's surface. Contour lines are used to assist in analyzing the slope and undulation of the terrain, optimizing the layout of the smart power plant and road design. Control point coordinates are used to provide accurate plane and elevation benchmarks to ensure the accuracy of construction layout.

[0022] The basic 3D model is a 3D model constructed using 3D simulation technology based on static data to represent the static dimensions of a smart power plant, enabling the basic 3D model to serve as a model of the overall construction progress of the smart power plant under the initial settings. Specifically, the method for constructing a basic 3D model is as follows: The design parameters are input into the BIM simulation software (Autodesk Navisworks + Civil 3D is used in this embodiment), and the design parameters are simulated and converted one by one to simulate the BIM design model. Mark the simulation positions at intervals on the BIM design model, import the terrain data into the simulation positions one by one, and record the simulation positions where the terrain data is imported as terrain simulation positions; the simulation positions are used to provide location constraints for the import of terrain data, and the interval distribution of simulation positions can ensure that each piece of terrain data remains independent. At the same time, a fusion command is sent to the terrain simulation location to control the simulation transformation of the terrain simulation location, thereby upgrading the BIM design model to a basic 3D model. The fusion command is a control command used to drive the simulation transformation of the BIM design model, ensuring the consistency and synchronicity of the model upgrade process.

[0023] The basic 3D model can only represent the original construction progress of the smart power plant in terms of its overall shape. It cannot distinguish the construction progress of different progress and different areas in detail, which makes the basic 3D model unclear in the detailed areas. Therefore, it is necessary to divide the different functional areas in the basic 3D model and label the key nodes in the functional areas. In this embodiment, the functional area is used as part of the overall structure of the smart power plant and forms part of the basic three-dimensional model. Specifically, the functional area includes the main plant area, boiler installation area, steam turbine area, power distribution equipment area, material storage area, and construction passage area.

[0024] After dividing the basic 3D model into functional areas, it is also necessary to mark the key nodes in each functional area so that the key nodes can serve as the basis for completing the construction schedule of each functional area and as the basis for comparing the subsequent construction progress. In this embodiment, the key nodes of each functional area are unique, and there is no duplication or overlap. For example, the key node in the main plant area is the completion of the main plant roof, and the key node in the boiler installation area is the completion of the boiler body hoisting.

[0025] Specifically, when annotating key nodes in functional areas, the overall progress of the basic 3D model construction is first queried. The overall progress is then broken down into weekly progress, with each week as a unit. The weekly progress containing key nodes in any functional area is then recorded as the key progress. The completion time of key nodes in each key progress is then queried to obtain the key time, and the key time is used as the node annotation for the key node.

[0026] It should be noted that when marking key nodes, relevant indicators in the construction progress of the smart power plant can also be set in advance and used as the data basis for real-time comparison of construction progress. For example, the pre-set indicator is the daily construction area, and the value of the daily construction area is 500 square meters.

[0027] S02: Collect dynamic data of the smart power plant through drones, and preprocess the dynamic data to generate discrete point cloud data; Dynamic data refers to the dynamically changing data collected in real time by data acquisition equipment carried on drones during the construction of smart power plants, which can provide a basis for comparison of the construction progress of smart power plants at a dynamic level; Specifically, dynamic data includes image data, laser point cloud data, and GPS coordinate data; Image data refers to construction images of different locations in a smart power plant, captured in real time by high-definition cameras mounted on drones, serving as image-level data during the construction of the smart power plant.

[0028] Laser point cloud data refers to the point cloud data of different locations in a smart power plant obtained by real-time scanning of the LiDAR carried on a drone. It serves as the data at the point cloud data dimension level during the construction of a smart power plant.

[0029] GPS coordinate data refers to the coordinates of different locations in a smart power plant obtained in real time by the GPS positioning module carried on the drone, serving as data at the spatial location dimension during the construction of the smart power plant.

[0030] In this embodiment, when collecting dynamic data, it is necessary to first plan the flight path of the UAV using ArcGIS Pro and a basic 3D model, draw the flight trajectory of the UAV, and set the collection period of dynamic data. The UAV is then controlled to collect dynamic data of different locations in the smart power plant along the flight trajectory during the collection period, thereby obtaining sufficient image data, laser point cloud data and GPS coordinate data.

[0031] After obtaining the dynamic data, the dynamic data can only be used as real-time data of the construction progress of the smart power plant during the current collection period. Since the collected dynamic data is raw and unprocessed, there will be additional negative phenomena such as noise interference in the dynamic data. Therefore, it is necessary to preprocess the dynamic data to improve the overall quality of the dynamic data and generate point cloud data in a discrete state. Specifically, the method for generating point cloud data is as follows: The laser point cloud data is imported into the point cloud processing software (CloudCompare is used in this embodiment), and a mesh spatial index structure is constructed using the laser point cloud data as index points. The spatial index structure is a data structure based on the geometric features or positional relationships of spatial objects, which is used to improve the efficiency of spatial data retrieval. The mesh spatial index structure can ensure the correlation between data. The method sets the number of neighboring points and the standard deviation threshold for the laser point cloud data. It then uses a spatial indexing structure to match each neighboring point in the laser point cloud data one by one, and calculates the distance threshold from the laser point cloud data to the neighboring points using a statistical filtering algorithm. Statistical filtering algorithms are methods used to remove outliers or noise from a dataset. They identify and remove data points that do not conform to these statistical characteristics by analyzing the data's statistical properties, thereby cleaning and optimizing the dataset. Statistical filtering algorithms are existing technologies and will not be elaborated upon here. When the average neighborhood distance of a laser point cloud data point to its neighboring points is greater than a distance threshold, the laser point cloud data point is recorded as a noise point, and the laser point cloud data point corresponding to the noise point is removed. Randomly select one image data as the reference image, record the point at the center of the reference image as the reference point, and record the remaining image data as the matching image; The feature point threshold and matching distance threshold of the image data are set. With the reference point as the registration center, the SIFT feature point matching algorithm is used to register the matching image and the reference image, and to align the image data at multiple view positions at the same time. The SIFT feature point matching algorithm is a classic local image feature extraction and matching algorithm used to match feature points in image data. The SIFT feature point matching algorithm is existing technology and will not be described in detail here. The first iteration number and convergence threshold are set, and the laser point cloud data is rotated and translated by the ICP algorithm to stitch all the laser point cloud data together. The ICP algorithm is a classic method for point cloud registration, which is used to align two overlapping 3D point clouds, that is, to estimate the rigid body transformation between them so that the source point cloud coincides with the target point cloud as much as possible. The ICP algorithm is existing technology and will not be described in detail here. All laser point cloud data are unified into the WGS84 coordinate system. The image data and laser point cloud data are matched and bound to the actual location of the smart power plant through GPS coordinate data. The GPS coordinate data is calibrated and discrete point cloud data is generated.

[0032] It should be noted that discrete point cloud data not only contains point cloud position information in three-dimensional space, but also image information and coordinate information corresponding to the threshold, which can provide necessary data support for subsequent real-time dynamic simulation models of smart power plants.

[0033] S03: Divide the point cloud data into different construction dynamic elements, and perform Gaussian splashing operation on the point cloud data within the construction dynamic elements to construct a dynamic 3D model; Discrete point cloud data is usually in a relatively discrete state, which means that point cloud data with different location and shape correspond to different construction dynamic elements in the construction of smart power plants. Since the degree of compactness of different construction dynamic elements in point cloud data is not consistent, and the construction location and objects corresponding to different degrees of compactness are also different, it is necessary to identify and distinguish the construction dynamic elements in point cloud data. Specifically, construction dynamic elements include material pile elements, construction machinery elements, completed structural elements, and pipeline support elements.

[0034] The method for dividing construction dynamic elements is as follows: The minimum distance threshold between the point cloud data and its neighboring points is set as the density baseline. The density baseline is then multiplied by the density conversion factor to calculate the density value of the point cloud data. The density conversion factor is the conversion ratio required to convert the density baseline to the density value, ensuring accurate numerical conversion between the density baseline and the density value. The range between the minimum and maximum density values ​​is denoted as the global range, and the global range is divided into four consecutive sub-ranges to obtain the first sub-range, the second sub-range, the third sub-range, and the fourth sub-range. Point cloud data with density values ​​in the first, second, third, and fourth sub-ranges are aggregated and grouped to form material pile elements, construction machinery elements, completed structure elements, and pipeline support elements.

[0035] Once the construction dynamic elements are obtained, Gaussian splashing calculations can be performed on the point cloud data within the construction dynamic elements to construct a dynamic 3D model that adapts to the real-time dynamic construction progress of the smart power plant. When performing Gaussian splash calculations on point cloud data, it is necessary to first determine the required parameters for various Gaussian splash calculations corresponding to the point cloud data in material pile elements, construction machinery elements, completed structure elements, and pipeline support elements. According to existing technical data and database content, the specific values ​​of the required parameters for various Gaussian splash calculations of material pile elements, construction machinery elements, completed structure elements, and pipeline support elements are shown in Table 1 below. Table 1 Based on the contents of Table 1 above, material elements include, but are not limited to, sand, gravel, and steel; construction machinery elements include, but are not limited to, cranes and excavators; completed structural elements include, but are not limited to, concrete frames and foundations; and pipeline support elements include, but are not limited to, small components.

[0036] Specifically, the method for constructing dynamic 3D models is as follows: Import the point cloud data from the material pile element, construction machinery element, completed structure element and pipeline support element into the same three-dimensional space in sequence, and mark the spatial coordinates of the point cloud data as (x, y, z) to obtain A three-dimensional spatial points; Gaussian splashing calculations are performed sequentially on A three-dimensional spatial points to calculate the superposition value of the Gaussian splashes at A three-dimensional spatial points; The formula for calculating the superposition value of Gaussian splashes is: ; In the formula, This represents the superposition value of Gaussian splashes. The weights of the Gaussian kernel, It is a natural exponential function. For the first Spatial coordinates of a three-dimensional point Where is the splash radius, High weight; Using Gaussian splash superposition values ​​as the reconstruction benchmark, the Marching Cubes algorithm is used to calculate isosurfaces on A three-dimensional spatial points, extracting surface points on the same plane. All surface points are then superimposed and stitched together to generate a mesh model. The Marching Cubes algorithm is a classic isosurface extraction method, widely used to reconstruct continuous triangular mesh surfaces from a three-dimensional scalar field. The Marching Cubes algorithm is existing technology and will not be elaborated on here. Locate the model region on the mesh model that coincides with the spatial location of the image data, and attach the image data to the corresponding model region to map the mesh model texture into a dynamic 3D model.

[0037] It is important to note that the dynamic 3D model changes continuously over time. Therefore, each dynamic data collected by the drone corresponds to a unique dynamic 3D model, which enables dynamic and continuous simulation of the construction progress of the smart power plant in real time.

[0038] S04: Embed the dynamic 3D model into the basic 3D model for model fusion, generate a 3D real scene model, and mark the construction progress of the dynamic construction elements. After the basic 3D model and the dynamic 3D model are constructed, the two models can be merged, allowing the basic 3D model and the dynamic 3D model to be compared one by one in real time and planned construction progress within the same model framework. In this embodiment, after the basic 3D model and the dynamic 3D model are fused, a 3D real-scene model that can support scaling, rotation, and sectioning can be obtained, and the real-time status of dynamic construction elements can be marked. This allows the 3D real-scene model to serve as a simulation model that can accurately and comprehensively represent the construction progress of the smart power plant in real time.

[0039] Specifically, the method for generating 3D reality models is as follows: Both the basic 3D model and the dynamic 3D model are converted to the WGS84 coordinate system, and all model points in the basic 3D model and the dynamic 3D model are marked one by one. The spatial coordinates of fixed feature points are marked in the basic 3D model, and the spatial coordinates of dynamic feature points are marked in the dynamic 3D model. A correspondence between fixed feature points and dynamic feature points is established. The correspondence refers to the result of the association and similarity between fixed feature points and dynamic feature points in the actual construction progress, and serves as the prerequisite for subsequent feature point matching. Set the second iteration number and the interior point threshold, and use the RANSAC algorithm to match the fixed feature points and dynamic feature points with corresponding relationships, thereby aligning the spatial coordinates of the model points on the dynamic 3D model with those on the basic 3D model. Mesh nodes that intersect are called overlapping meshes. After removing all overlapping meshes from the dynamic 3D model, the remaining dynamic 3D model is embedded into the base 3D model to generate a 3D reality model.

[0040] It should be noted that after generating the 3D reality model, the construction dynamic elements in the 3D reality model also need to be labeled so that the real-time position, size and size of the construction dynamic elements can be observed intuitively and accurately from the 3D reality model. Based on the real-time position, size and size, the construction progress at the current moment can be deduced, which will facilitate accurate and comprehensive analysis and management of the construction progress of the smart power plant in the future.

[0041] S05: Extract progress parameters from the 3D reality model, calculate the overall deviation rate, select the corresponding early warning mechanism based on the overall deviation rate, and generate the corresponding early warning message; Once the 3D reality model is obtained, the actual situation and the planned situation of the construction progress of the smart power plant in the 3D reality model can be analyzed and compared, so as to intelligently manage the construction progress of the smart power plant from the virtual model level. The progress parameter is used to comprehensively represent the difference between the real-time construction progress and the planned construction progress in the 3D reality model at the current moment, which can provide data support for the actual construction progress of the smart power plant from multiple different dimensions. Specifically, the progress parameters include the completed area value, construction height value, equipment installation quantity, and material consumption progress.

[0042] The completed area value refers to the horizontal projected area of ​​the constructed structure in the dynamic three-dimensional model, which can represent the construction area of ​​the smart power plant at the current moment.

[0043] The construction height value refers to the difference between the height value of the top of the structure in the dynamic three-dimensional model and the height value of the top of the structure in the foundation three-dimensional model.

[0044] The number of installed equipment refers to the number of qualified mechanical equipment that has been installed in the dynamic three-dimensional model.

[0045] Material consumption progress refers to the ratio between the actual volume of material consumed in a material pile and the original volume, which can be used to represent the degree of material consumption.

[0046] In this embodiment, when extracting the material consumption progress, the outer surface of the material stacking area is identified in the three-dimensional real scene model using computer vision technology. The volume of the model area located inside the outer surface is calculated and recorded as the retention volume. The original volume of the material stacking area is subtracted from the retention volume, and the difference is compared with the original volume to obtain the real-time consumption progress.

[0047] The overall deviation rate is used to represent the overall deviation between the completed area value, construction height value, equipment installation quantity and material consumption progress and the preset corresponding values. It can be used as a direct basis for determining the severity of the deviation between the real-time construction progress and the planned construction progress of the smart power plant. Specifically, the calculation method for the overall deviation rate is as follows: The database is used to retrieve the planned area, planned height, planned installation quantity, and planned consumption progress corresponding to the construction progress. The area deviation rate is calculated by taking the absolute value of the difference between the completed area value and the planned area value and comparing it with the planned area value. The formula for calculating the area deviation rate is: ; In the formula, This refers to the area deviation rate. To complete the area value, This is the planned area value; The absolute value of the difference between the construction height and the planned height is taken and compared with the planned height to calculate the height deviation rate. The formula for calculating the height deviation rate is: ; In the formula, For the height deviation rate, This is the construction height value. This is the planned height value; The installation deviation rate is calculated by taking the absolute value of the difference between the number of equipment installed and the planned number of equipment installed, and comparing it with the planned number of equipment installed. The formula for calculating the installation deviation rate is: ; In the formula, Installation deviation rate For the number of equipment installed, The planned installation quantity; The absolute value of the difference between the real-time consumption progress and the planned consumption progress is taken and compared with the planned consumption progress to calculate the progress deviation rate. The formula for calculating the schedule deviation rate is: ; In the formula, This refers to the schedule deviation rate. To consume progress in real time, The planned time consumption; The area deviation rate, height deviation rate, installation deviation rate and schedule deviation rate are weighted and summed to calculate the comprehensive deviation rate. The formula for calculating the overall deviation rate is: ; In the formula, The overall deviation rate, , , , These are the weighting factors for area deviation rate, height deviation rate, installation deviation rate, and schedule deviation rate, respectively. , , , All are greater than 0. , , , The sum of is 1.

[0048] The early warning mechanism is a working display mode of the 3D real scene model selected based on the actual magnitude of the comprehensive deviation rate, thereby helping construction management departments and personnel to quickly and accurately understand the actual construction progress at the current moment and providing a prerequisite for subsequent adjustments to the construction progress. Specifically, the early warning mechanism includes no warning, Level 1 warning, and Level 2 warning; and the severity of the abnormal construction progress is from low to high for no warning, Level 1 warning, and Level 2 warning.

[0049] In this embodiment, when determining which early warning mechanism to select, the comprehensive deviation rate is compared with the first comprehensive deviation threshold and the second comprehensive deviation threshold. The first comprehensive deviation threshold and the second comprehensive deviation threshold are used to represent the critical values ​​of the comprehensive deviation rate corresponding to no early warning, first-level early warning and second-level early warning, so as to achieve accurate differentiation between no early warning, first-level early warning and second-level early warning. In this embodiment, the first comprehensive deviation threshold is smaller than the second comprehensive deviation threshold. When the overall deviation rate is less than or equal to the first overall deviation threshold, the deviation between the current construction progress and the planned construction progress of the smart power plant is extremely small, so no warning is selected. When the overall deviation rate is greater than the first overall deviation threshold and less than or equal to the second overall deviation threshold, the deviation between the current construction progress and the planned construction progress of the smart power plant is relatively small, and a first-level warning is selected. When the overall deviation rate exceeds the second overall deviation threshold, the deviation between the current construction progress and the planned construction progress of the smart power plant is extremely large, and a level-two warning is selected.

[0050] The early warning information is based on the early warning mechanism and sends a notification to the construction management department of the smart power plant indicating abnormal construction progress. This ensures that the construction management department can identify problems as soon as possible and make subsequent progress adjustment measures, thereby achieving efficient management of the construction progress of the smart power plant and improving the pertinence and real-time nature of the construction progress management. The method for generating early warning information is as follows: When the early warning mechanism is set to no warning, there is no need to generate early warning information. When the early warning mechanism is at the level of Level 1, a milder warning is needed to adjust the construction progress. In this case, a system pop-up notification is first issued, followed by a text message notification to the construction personnel, so as to effectively adjust the construction progress. When the early warning mechanism is at level two, a more serious warning is needed regarding the construction progress. In this case, a system pop-up notification is first issued, followed by a text message notification to the construction personnel, then a telephone notification to the construction supervision department, and finally a construction progress analysis report is generated with the comprehensive deviation rate and progress parameters as the report content, thereby enabling effective adjustments to the construction progress.

[0051] It should be noted that after each calculation and analysis of the comprehensive deviation rate, the selected early warning mechanism and the generated early warning information need to be recorded and stored, and the specific measures taken to adjust the construction progress need to be updated in the database. This will enable the 3D reality model to be updated and upgraded in real time, ensuring that the 3D reality model can be consistent and synchronized with the actual construction progress of the smart power plant.

[0052] Example 2: Please refer to Figure 2 As shown, the parts not described in detail in this embodiment are described in Embodiment 1. A smart power plant construction progress safety management system based on model fusion is provided to implement a smart power plant construction progress safety management method based on model fusion. It includes a first model building module, a dynamic data processing module, a second model building module, a model fusion annotation module, and a construction progress analysis module. The modules are connected to each other via wired or wireless networks. The first model construction module is used to construct a basic three-dimensional model from the static data of the smart power plant, divide the basic three-dimensional model into functional areas, and label the key nodes of the functional areas. The dynamic data processing module is used to collect dynamic data from the smart power plant and process the dynamic data to generate discrete point cloud data. The second model building module is used to divide the point cloud data into different construction dynamic elements and perform Gaussian splashing calculations on the point cloud data within the construction dynamic elements to build a dynamic three-dimensional model. The model fusion and annotation module is used to embed dynamic 3D models into basic 3D models for model fusion, generate 3D reality models, and annotate the construction progress. The construction progress analysis module is used to extract progress parameters from the 3D reality model, calculate the overall deviation rate, and determine the corresponding early warning mechanism and early warning information based on the overall deviation rate.

[0053] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for safe management of construction progress at smart power plant construction sites based on model fusion, characterized in that, include: S01: Construct a basic 3D model using static data from a smart power plant, divide the basic 3D model into functional areas, and label the key nodes of the functional areas. The static data includes design parameters and terrain data. S02: Collect dynamic data from the smart power plant, including image data, laser point cloud data, and GPS coordinate data, and process the dynamic data to generate discrete point cloud data; S03: Divide the point cloud data into different construction dynamic elements, and perform Gaussian splashing operation on the point cloud data within the construction dynamic elements to construct a dynamic 3D model; S04: Embed the dynamic 3D model into the basic 3D model for model fusion, generate a 3D reality model, and mark the construction progress. S05: Extract progress parameters from the 3D reality model. Progress parameters include completed area value, construction height value, equipment installation quantity and material consumption progress. Calculate the comprehensive deviation rate and determine the corresponding early warning mechanism and early warning prompt information based on the comprehensive deviation rate.

2. The method for safe management of construction progress at smart power plant construction sites based on model fusion as described in claim 1, characterized in that, The basic method for constructing a 3D model is as follows: Input the design parameters into the BIM simulation software to simulate the BIM design model; Mark the simulated positions with interval distribution on the BIM design model, import the terrain data into the simulated positions one by one, and record the simulated positions with imported terrain data as terrain simulation positions. At the same time, a fusion command is sent to the terrain simulation location to control the terrain simulation location to undergo simulation transformation, upgrading the BIM design model to a basic three-dimensional model.

3. The method for safe management of construction progress at smart power plant construction sites based on model fusion as described in claim 2, characterized in that, The functional areas include the main plant area, boiler installation area, steam turbine area, power distribution equipment area, material storage area, and construction access area; When annotating key nodes, first query the overall progress of the basic 3D model construction. Decompose the overall progress into weekly progress in one-week units. Then, record the weekly progress containing key nodes in any functional area as key progress. Query the completion time of key nodes in each key progress and record it as key time. Use the key time as the node annotation for key nodes.

4. The method for safe management of construction progress at smart power plant construction sites based on model fusion as described in claim 3, characterized in that, The method for generating point cloud data is as follows: Import the laser point cloud data into the point cloud processing software, and use the laser point cloud data as index points to construct a mesh-like spatial index structure; Set the number of neighboring points and standard deviation threshold of the laser point cloud data, match the neighboring points corresponding to the laser point cloud data one by one through the spatial index structure, and calculate the distance threshold from the laser point cloud data to the neighboring points through the statistical filtering algorithm. Laser point cloud data with an average neighborhood distance greater than a distance threshold are recorded as noise points, and the laser point cloud data corresponding to noise points are removed. Randomly select one image data as the reference image, record the point at the center of the reference image as the reference point, and record the remaining image data as the matching image; Set the feature point threshold and matching distance threshold for the image data, and use the reference point as the registration center to perform image registration between the matching image and the reference image using the SIFT feature point matching algorithm. The first iteration number and convergence threshold are set, and the laser point cloud data is rotated and translated by the ICP algorithm. All laser point cloud data are unified into the WGS84 coordinate system. The image data and laser point cloud data are matched and bound with the actual location of the smart power plant by GPS coordinate data to generate discrete point cloud data.

5. The method for safe management of construction progress at smart power plant construction sites based on model fusion as described in claim 4, characterized in that, Construction dynamic elements include material pile elements, construction machinery elements, completed structural elements, and pipeline support elements; The method for dividing construction dynamic elements is as follows: The minimum distance threshold from the point cloud data to the neighboring points is recorded as the density benchmark, and the density benchmark is multiplied by the density conversion coefficient to calculate the density value of the point cloud data. The range between the minimum and maximum density values ​​is denoted as the global range, and the global range is divided into four consecutive sub-ranges to obtain the first sub-range, the second sub-range, the third sub-range, and the fourth sub-range. Point cloud data with density values ​​in the first, second, third, and fourth sub-ranges are aggregated and grouped to form material pile elements, construction machinery elements, completed structure elements, and pipeline support elements.

6. The method for safe management of construction progress at smart power plant construction sites based on model fusion as described in claim 5, characterized in that, The method for constructing dynamic 3D models is as follows: Import the point cloud data from the material pile element, construction machinery element, completed structure element and pipeline support element into the same three-dimensional space in sequence, and mark the spatial coordinates of the point cloud data as (x, y, z) to obtain A three-dimensional spatial points; Gaussian splashing calculations are performed sequentially on A three-dimensional spatial points to calculate the superposition value of the Gaussian splashes at A three-dimensional spatial points; Using Gaussian splash superposition values ​​as the reconstruction benchmark, the Marching Cubes algorithm is used to calculate isosurfaces for A three-dimensional spatial points, extracting surface points on the same plane, and then superimposing and stitching all surface points sequentially to generate a mesh model; Locate the model region on the mesh model that coincides with the spatial location of the image data, and attach the image data to the corresponding model region to map the mesh model texture into a dynamic 3D model.

7. The method for safe management of construction progress at smart power plant construction sites based on model fusion as described in claim 6, characterized in that, The method for generating 3D reality models is as follows: Both the basic 3D model and the dynamic 3D model are converted to the WGS84 coordinate system, and all model points in the basic 3D model and the dynamic 3D model are marked one by one. Mark the spatial coordinates of fixed feature points from the basic 3D model, mark the spatial coordinates of dynamic feature points from the dynamic 3D model, and establish the correspondence between fixed feature points and dynamic feature points; Set the second iteration number and the interior point threshold, and use the RANSAC algorithm to match the fixed feature points and dynamic feature points with corresponding relationships, thereby aligning the spatial coordinates of the model points on the dynamic 3D model with those on the basic 3D model. After removing overlapping meshes that are intersecting from the dynamic 3D model, the remaining dynamic 3D model is embedded into the base 3D model to generate a 3D reality model.

8. The method for safe management of construction progress at smart power plant construction sites based on model fusion as described in claim 7, characterized in that, The calculation method for the overall deviation rate is as follows: The database is used to retrieve the planned area, planned height, planned installation quantity, and planned consumption progress corresponding to the construction progress. The area deviation rate is calculated by taking the absolute value of the difference between the completed area value and the planned area value and comparing it with the planned area value. The absolute value of the difference between the construction height and the planned height is taken and compared with the planned height to calculate the height deviation rate. The installation deviation rate is calculated by taking the absolute value of the difference between the number of equipment installed and the planned number of equipment installed, and comparing it with the planned number of equipment installed. The absolute value of the difference between the real-time consumption progress and the planned consumption progress is taken and compared with the planned consumption progress to calculate the progress deviation rate. The area deviation rate, height deviation rate, installation deviation rate, and schedule deviation rate are weighted and summed to calculate the overall deviation rate.

9. The method for safe management of construction progress at smart power plant construction sites based on model fusion as described in claim 8, characterized in that, The early warning mechanism includes no warning, Level 1 warning, and Level 2 warning; The method for determining the early warning mechanism is as follows: When the overall deviation rate is less than or equal to the first overall deviation threshold, select no warning; When the overall deviation rate is greater than the first overall deviation threshold and less than or equal to the second overall deviation threshold, a Level 1 warning is selected; When the overall deviation rate exceeds the second overall deviation threshold, a level-two warning is selected.

10. The method for generating early warning information in the smart power plant construction progress safety management method based on model fusion according to claim 9 is as follows: When the early warning mechanism is set to no warning, no early warning message will be generated; When the early warning mechanism is at level one, a system pop-up notification will be issued first, followed by a text message notification to the construction personnel. When the early warning mechanism is at level two, a system pop-up notification is first issued, followed by a text message notification to the construction personnel, then a telephone notification to the construction supervision department, and finally a construction progress analysis report is generated with the comprehensive deviation rate and progress parameters as the report content.

Images