Electric power capital construction process early warning method and system based on high-precision three-dimensional model

Abstract

The invention discloses an electric power capital construction process early warning method and system based on a high-precision three-dimensional model, and relates to the technical field of electric power capital construction, and the method comprises the steps: collecting multi-source data of an electric power capital construction site, and constructing a high-precision three-dimensional model comprising equipment space coordinates, construction time sequence labels and environmental parameters; inputting the space-time diagram into a space-time diagram convolutional network, outputting a risk propagation probability matrix of a three-dimensional space and a construction time sequence dimension, and identifying high-risk nodes and associated construction stages; and dynamically marking root cause coordinates and a risk diffusion range in the high-precision three-dimensional model, generating an early warning instruction containing a visual path, a root cause report and a disposal priority, and pushing the early warning instruction to a terminal in real time through an edge node. Based on the hierarchical space-time attention-physical constraint graph convolutional network, the inter-node cross-stage association weight is dynamically allocated, the convolution kernel parameters are optimized in combination with the physical connection strength, and the space-time association of high-risk node identification is significantly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power infrastructure, and in particular to an electric power infrastructure process early warning method and system based on a high-precision three-dimensional model. Background Art

[0002] In recent years, the risk warning technology of power infrastructure projects has gradually developed towards the integration of three-dimensional modeling and spatiotemporal data analysis. In the existing technology, the monitoring system based on BIM (Building Information Model) has realized the three-dimensional visualization of the construction scene, but it focuses more on static equipment status monitoring and lacks modeling of the dynamic correlation of construction sequence; at the same time, the application of spatiotemporal graph neural network (ST-GNN) in fields such as traffic flow prediction has verified its spatiotemporal correlation modeling capabilities, but it has not yet been deeply coupled with construction physical constraints (such as equipment connection strength and process connection logic). In addition, Bayesian networks are used for risk causal reasoning, but traditional methods rely on expert experience to construct topology and are difficult to adapt to the dynamic evolution characteristics of high-dimensional spatiotemporal data.

[0003] The existing technologies mainly have the following deficiencies: First, the traditional high-precision three-dimensional model does not dynamically bind the spatial coordinates of the equipment with the construction sequence labels and environmental parameters, resulting in limited accuracy in modeling the risk propagation path. For example, the existing method judges equipment abnormalities through fixed thresholds, but ignores the impact of construction stage dependencies and physical connection strength on risk transmission, which is prone to false positives or omissions. Secondly, the risk prediction model based on supervised learning lacks a counterfactual reasoning mechanism and cannot effectively distinguish the contribution of equipment abnormalities and normal state fluctuations to the overall risk, resulting in root cause location deviations. For example, the existing Bayesian network has difficulty decoupling the causal chain in multi-process coupling scenarios because it does not integrate spatiotemporal graph convolution features, which affects the pertinence of the disposal strategy. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an electric power infrastructure process early warning method based on a high-precision three-dimensional model to solve the problems of inaccurate dynamic risk propagation modeling, large root cause equipment positioning deviation and delayed early warning instruction generation in the electric power infrastructure process.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for early warning of a power infrastructure process based on a high-precision three-dimensional model, which includes collecting multi-source data of a power infrastructure construction site and constructing a high-precision three-dimensional model including equipment spatial coordinates, construction timing labels, and environmental parameters; The grid topology is mapped into a time-space graph, the node attributes encode the construction timestamp and the time-series change of the equipment status, and the edge weights are associated with the construction connection relationship and physical connection strength of adjacent nodes; The space-time graph is input into the space-time graph convolutional network, and the risk propagation probability matrix in the three-dimensional space and construction time sequence dimensions is output to identify high-risk nodes and associated construction stages; Construct a causal Bayesian network to locate the root cause equipment and process by calculating the risk evolution difference after the abnormality of the computing equipment is eliminated through counterfactual simulation; The root cause coordinates and risk diffusion range are dynamically marked in the high-precision three-dimensional model, and early warning instructions with visual paths, root cause reports and disposal priorities are generated and pushed to the terminal in real time through the edge node.

[0007] As a preferred solution of the electric power infrastructure process early warning method based on high-precision three-dimensional model described in the present invention, wherein: the multi-source data includes equipment space coordinate data, construction time sequence label data, environmental parameter data, equipment status time sequence data and construction connection relationship data; The specific steps of constructing a high-precision three-dimensional model including equipment space coordinates, construction sequence labels and environmental parameters are as follows: Perform spatiotemporal alignment and cleaning on the collected equipment space coordinates, construction sequence labels and environmental parameter data, and output a standardized spatiotemporal data set; Extract the spatial coordinates of the device from the standardized spatiotemporal data set, combine it with the point cloud data collected by LiDAR, perform denoising and registration, and generate cleaned point cloud data; Convert the cleaned point cloud data into a 3D mesh model, optimize the number of facets and map the device space coordinates to generate a high-precision basic model; Bind construction timing labels and environmental parameters to equipment space nodes to form a high-precision three-dimensional model with timing and environmental attributes.

[0008] As a preferred solution of the electric power infrastructure process early warning method based on high-precision three-dimensional model described in the present invention, wherein: the grid topology is mapped into a time-space graph, the node attribute encodes the construction timestamp and the equipment status time sequence change, and the edge weight associates the construction connection relationship of adjacent nodes with the physical connection strength. The specific steps are as follows: The high-precision three-dimensional model is divided into grid units at a fixed resolution, and each grid unit serves as a node in the space-time graph; Extract construction timestamps and equipment status time series data from the standardized spatiotemporal data set, bind them to the corresponding grid nodes, and form node attributes; According to the equipment space coordinates and construction connection relationship data, the connection relationship between adjacent grid nodes is determined, and the physical connection strength is calculated as the edge weight; The node attributes and edge weights are integrated to generate a spatiotemporal graph, and the graph structure is optimized using a graph optimization algorithm.

[0009] As a preferred solution of the electric power infrastructure process early warning method based on a high-precision three-dimensional model described in the present invention, the specific steps are as follows: the space-time graph is input into the space-time graph convolutional network, the risk propagation probability matrix of the three-dimensional space and construction timing dimensions is output, and the high-risk nodes and associated construction stages are identified. Based on the node attributes and edge weights of the spatiotemporal graph, a hierarchical spatiotemporal attention-physical constraint graph convolutional network is constructed, integrating causal spatiotemporal attention units, physical constraint convolution kernels, and counterfactual reasoning layers as the network architecture; Based on the causal spatiotemporal attention unit in the network architecture, the temporal dependency is divided in combination with the construction sequence labels, and the spatial attention distribution is adjusted using the node timestamp difference and physical connection strength; Dynamically allocate cross-stage node association weights through the Softmax function, strengthen the spatiotemporal correlation of high-risk nodes in the same process, and output dynamically calculated attention weights; The physical parameters in the edge weights are spliced ​​with the node device status, input into the multi-layer perceptron to generate the physical perception convolution kernel, and the kernel weight is dynamically adjusted according to the environmental parameters; By integrating attention weights and physical constraint convolution kernels, multi-head graph convolution is performed on node features to capture local physical connections, global process dependencies, short-term state fluctuations, and long-term environmental trends, and output spatiotemporal hidden features. Perform counterfactual intervention on the device status of each node, recalculate the graph convolution output, and generate a risk difference matrix by comparing the Euclidean distance between actual and counterfactual features; The spatiotemporal hidden features are weightedly fused with the risk difference matrix, mapped into a risk propagation probability matrix in the three-dimensional space-time dimension through the Sigmoid function, and the risk value of each grid node in each construction stage is marked; Calculate the dynamic threshold of risk probability according to the construction stage, set the warning threshold, screen the nodes exceeding the warning threshold as high-risk targets, and trace back to the construction stage corresponding to the risk peak, and output the high-risk nodes and related construction stages.

[0010] As a preferred solution of the electric power infrastructure process early warning method based on the high-precision three-dimensional model of the present invention, wherein: the counterfactual intervention is performed on the state of each node device, the graph convolution output is recalculated, and the risk difference matrix is ​​generated by comparing the Euclidean distance between the actual and counterfactual feature vectors. The specific steps are as follows: Based on the spatiotemporal hidden feature matrix, the actual feature vector of each node is extracted as the benchmark for counterfactual comparison; Perform counterfactual intervention on the device state of each node and generate counterfactual state data by setting the device state as a baseline value; Based on the counterfactual state data, re-execute the graph convolution operation and output the counterfactual feature vector; Compare the actual feature vector with the counterfactual feature vector, calculate the Euclidean distance, and generate the node-level risk difference value; The risk difference values ​​of all nodes are mapped according to the grid topology to generate a risk difference matrix, which is then normalized and output as a standardized risk difference matrix.

[0011] As a preferred solution of the electric power infrastructure process early warning method based on a high-precision three-dimensional model described in the present invention, the causal Bayesian network is constructed to locate the root cause equipment and process by counterfactual simulation and calculation of the risk evolution difference after the abnormality of the equipment is eliminated. The specific steps are as follows: Based on high-risk nodes and associated construction stages, a Bayesian network of causal dependencies between equipment status and construction processes is constructed; Generate counterfactual intervention scenarios through Bayesian networks, simulate the state changes after the equipment anomaly is eliminated, and output counterfactual state data; Input the spatiotemporal graph convolutional network to calculate the risk propagation probability matrix under the counterfactual scenario; Compare the risk propagation probability matrix of the actual scenario with the counterfactual scenario to generate the risk evolution difference matrix; Based on the risk evolution difference matrix, identify the root cause devices that contribute most to the overall risk; Trace back the impact path of abnormal status of root-cause equipment on adjacent nodes and construction processes, and locate key processes.

[0012] As a preferred solution of the electric power infrastructure process early warning method based on high-precision three-dimensional model described in the present invention, wherein: the root cause coordinates and risk diffusion range are dynamically marked in the high-precision three-dimensional model, and an early warning instruction containing a visualization path, a root cause report and a disposal priority is generated, and is pushed to the terminal in real time through the edge node. The specific steps are as follows: Based on the located root cause device, extract its spatial coordinate data and dynamically mark the root cause coordinates in the high-precision 3D model; Starting from the root cause coordinates, the risk diffusion range of the root cause equipment is calculated according to the risk evolution difference matrix, and the diffusion boundary is visually marked in the high-precision 3D model; Combined with the risk propagation probability matrix, the risk propagation path of high-risk nodes is extracted and visualized in a high-precision 3D model; Integrate the root cause equipment and related process information to generate a detailed root cause report including equipment status, abnormal reasons, impact range and related processes; Calculate the disposal priority of each root cause equipment based on the risk value and diffusion range of the nodes in the risk evolution difference matrix; Integrate visualization paths, root cause reports, and treatment priorities to generate structured warning instructions; The warning instructions are pushed to the terminal devices in real time through the edge nodes, and countermeasures are taken.

[0013] In a second aspect, the present invention provides an electric power infrastructure process early warning system based on a high-precision three-dimensional model, including a data acquisition module, a graph mapping module, a risk identification module, a root cause location module and an early warning push module; The data acquisition module is used to collect multi-source data from the power infrastructure construction site and construct a high-precision three-dimensional model including equipment spatial coordinates, construction timing labels and environmental parameters; The graph mapping module is used to map the grid topology into a spatiotemporal graph, the node attributes encode the construction timestamp and the equipment status time series change, and the edge weights associate the construction connection relationship of adjacent nodes with the physical connection strength; The risk identification module is used to input the space-time graph into the space-time graph convolutional network, output the risk propagation probability matrix in the three-dimensional space and construction time sequence dimensions, and identify high-risk nodes and associated construction stages; The root cause location module is used to construct a causal Bayesian network, locate the root cause equipment and process by calculating the risk evolution difference after the abnormality of the computing equipment is eliminated through counterfactual simulation; The warning push module is used to dynamically mark the root cause coordinates and risk diffusion range in the high-precision three-dimensional model, generate warning instructions containing visualization paths, root cause reports and disposal priorities, and push them to the terminal in real time via the edge node.

[0014] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the power infrastructure process early warning method based on a high-precision three-dimensional model as described in the first aspect of the present invention is implemented.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the power infrastructure process early warning method based on a high-precision three-dimensional model as described in the first aspect of the present invention.

[0016] The beneficial effects of the present invention are as follows: the present invention realizes multi-dimensional feature encoding of the construction process by constructing a high-precision three-dimensional model that integrates equipment spatial coordinates, construction timing labels and environmental parameters, and mapping it into a spatiotemporal graph structure. Based on the hierarchical spatiotemporal attention-physical constraint graph convolutional network, the cross-stage association weights between nodes are dynamically allocated, and the convolution kernel parameters are optimized in combination with the physical connection strength, which significantly improves the spatiotemporal correlation of high-risk node identification. The risk difference matrix is ​​generated through the counterfactual intervention mechanism to quantify the contribution of equipment anomalies to the overall risk, and the root cause equipment is more accurately located by tracing back the anomaly propagation path in combination with the causal Bayesian network. In addition, the root cause coordinates and risk diffusion range are dynamically labeled, and a visual early warning instruction with disposal priority is generated, which is pushed in real time through the edge node, greatly shortening the risk response time. This method breaks through the limitations of traditional static monitoring and experience-driven causal reasoning, realizes accurate modeling of risk propagation in the three-dimensional space-time dimension, and provides a full-link closed-loop solution from risk identification, root cause tracing to intelligent decision-making for power infrastructure projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use 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 ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0018] Figure 1 This is a flow chart of the electric power infrastructure process early warning method based on a high-precision three-dimensional model in Example 1.

[0019] Figure 2 This is a flow chart for generating a risk difference matrix in Example 1. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0022] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0023] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides an electric power infrastructure process early warning method based on a high-precision three-dimensional model, comprising the following steps: S1. Collect multi-source data from power infrastructure construction sites and build a high-precision three-dimensional model that includes equipment spatial coordinates, construction timing labels, and environmental parameters.

[0024] Specifically, the following steps are included: Multi-source data includes equipment spatial coordinate data, construction time sequence label data, environmental parameter data, equipment status time sequence data and construction connection relationship data; Specifically, drone oblique photography, laser scanning (LiDAR) and ground three-dimensional laser scanning technology are used to collect multi-perspective and multi-resolution data in the power infrastructure project area, ensuring millimeter-level accuracy of terrain, landforms, buildings, roads and other elements, obtaining multi-source data, and providing a reliable data foundation for the subsequent construction of high-precision three-dimensional models.

[0025] Construct a high-precision 3D model that includes equipment space coordinates, construction sequence labels, and environmental parameters. The specific steps are as follows: Perform spatiotemporal alignment and cleaning on the collected equipment space coordinates, construction sequence labels and environmental parameter data, and output a standardized spatiotemporal data set; Among them, the spatiotemporal alignment methods include timestamp alignment and spatial coordinate alignment; the data cleaning methods include outlier detection, missing value processing and redundant data elimination.

[0026] Extract the spatial coordinates of the device from the standardized spatiotemporal data set, combine it with the point cloud data collected by LiDAR, perform denoising and registration, and generate cleaned point cloud data; Specifically, the RANSAC algorithm is used to preliminarily separate the device-related point cloud from the point cloud data, and the sampling times and the internal point threshold are set; then statistical filtering is used to remove low-density outliers, and the noise is further eliminated in combination with adaptive radius filtering; then the FPFH feature descriptor is used for coarse registration to optimize the initial pose; in the fine registration stage, the ICP algorithm with physical constraints is introduced, and the nearest neighbor search is accelerated through KD-Tree; finally, the distance between the registered point cloud and the device coordinates is calculated to ensure that the error is within the allowable range, and the surface continuity is verified, and the cleaned point cloud data that meets the quality requirements is output.

[0027] Convert the cleaned point cloud data into a 3D mesh model, optimize the number of facets and map the device space coordinates to generate a high-precision basic model; Specifically, first use point cloud processing software (such as CloudCompare or MeshLab) to triangulate the cleaned point cloud data to generate a preliminary 3D mesh model to ensure that the 3D mesh model can accurately express the spatial structure of the point cloud; then use mesh optimization algorithms (such as edge collapse or vertex merging) to reduce the number of facets in the model and reduce computational complexity while retaining key geometric features; then spatially align the device spatial coordinate data with the optimized mesh model, and use the coordinate transformation matrix to accurately map the device position information to the corresponding position of the mesh model to ensure the consistency of the 3D mesh model with the actual space; finally, use texture mapping or color coding technology to attach environmental parameters or equipment status information to the surface of the 3D mesh model to generate a high-precision basic model, providing an intuitive and accurate 3D visualization foundation for subsequent construction simulation and data analysis.

[0028] Bind construction timing labels and environmental parameters to equipment space nodes to form a high-precision 3D model with timing and environmental attributes; Among them, the equipment space node refers to the abstract node representing the equipment location in the three-dimensional grid model or space-time graph. It not only contains the equipment space coordinates, but may also be bound to other attributes (such as construction sequence labels, environmental parameters, equipment status, etc.).

[0029] For example: add attribute fields {Construction stage: foundation pouring, ambient humidity: 65%, red line distance: 2.3m} to each device node; Specifically, in the process of building a high-precision 3D model, it is necessary to use 3D modeling software (such as ContextCapture, Bentley, etc.) to fuse the collected multi-source data to generate a high-precision 3D model containing a geographic coordinate system, which supports the refined expression of terrain elevation, underground pipelines, and ground facilities.

[0030] Furthermore, the high-precision 3D model is seamlessly connected to the GIS system, and vector data such as land redline, planning scope, and construction boundary are integrated through the spatial database (such as ArcGIS GeoDatabase) to form a spatial-attribute integrated model. Based on the spatial analysis function of GIS, the spatial relationship between the construction area and the land redline is automatically calculated, and potential conflict areas (such as redline expansion buffer) are marked.

[0031] S2, according to the grid topology mapping to a space-time graph, the node attributes encode the construction timestamp and the equipment status time series change, and the edge weight associates the construction connection relationship and physical connection strength of adjacent nodes.

[0032] Specifically, the following steps are included: The high-precision three-dimensional model is divided into grid units at a fixed resolution, and each grid unit serves as a node in the space-time graph; Extract construction timestamps and equipment status time series data from the standardized spatiotemporal data set, bind them to the corresponding grid nodes, and form node attributes; Specifically, firstly, the construction timestamp and equipment status time series data are filtered out from the standardized spatiotemporal data set through data processing tools (such as Python's Pandas library) to ensure that the time series of the construction timestamp and equipment status time series data are complete and without missing; then, the extracted timestamp and equipment status data are spatiotemporally matched with the nodes of the three-dimensional grid model, and the spatial index algorithm (such as KD-Tree or R-Tree) is used to quickly locate the grid nodes corresponding to each timestamp and equipment status data; then, the construction timestamp and equipment status time series data are bound to the matching grid nodes as attribute information, and the mapping relationship between nodes and attributes is stored through data structures (such as dictionaries or attribute tables); finally, the bound node attributes are verified to ensure that the timestamp and equipment status data are consistent with the spatiotemporal position of the grid nodes, forming complete node attribute information, and providing dynamic and accurate data support for subsequent construction process simulation and status analysis.

[0033] According to the equipment space coordinates and construction connection relationship data, the connection relationship between adjacent grid nodes is determined, and the physical connection strength is calculated as the edge weight; The equipment spatial coordinates refer to the specific position of the equipment in three-dimensional space, usually expressed in the form of a three-dimensional coordinate system (such as X, Y, Z coordinates). It describes the precise spatial position of the equipment at the construction site.

[0034] It should be noted that the position of the equipment in the grid is located through the equipment spatial coordinates, and the adjacent nodes are identified in combination with the construction process and interaction data; the physical connection strength is dynamically calculated based on the actual contact frequency of the equipment, the construction sequence interval and the spatial distance, and the degree of collaboration between the equipment is quantified; the generated weighted connection relationship diagram can clearly display the key paths and high-frequency interaction nodes in the construction network, providing a topological structure basis for subsequent resource allocation optimization, construction conflict detection and progress simulation, ensuring that the scheduling strategy accurately matches the equipment collaboration requirements in actual construction.

[0035] Furthermore, automatic collision detection algorithms (such as Navisworks collision detection modules based on BIM) are embedded to analyze the spatial conflicts between construction equipment, temporary facilities and red lines, underground pipelines, and existing buildings in real time, generate conflict heat maps and associate them with corresponding time and space nodes. The collision detection results are encoded as additional attributes of the nodes (such as conflict type and severity level) for subsequent risk propagation probability calculations.

[0036] The node attributes and edge weights are integrated to generate a spatiotemporal graph, and the graph structure is optimized using a graph optimization algorithm.

[0037] S3. Input the space-time graph into the space-time graph convolutional network, output the risk propagation probability matrix in the three-dimensional space and construction timing dimensions, and identify high-risk nodes and associated construction stages.

[0038] Specifically, the following steps are included: Based on the node attributes and edge weights of the spatiotemporal graph, a hierarchical spatiotemporal attention-physical constraint graph convolutional network is constructed, integrating causal spatiotemporal attention units, physical constraint convolution kernels, and counterfactual reasoning layers as the network architecture; Preferably, based on the node attributes and edge weights of the spatiotemporal graph, a hierarchical spatiotemporal attention-physical constraint graph convolutional network is constructed, integrating causal spatiotemporal attention units, physical constraint convolution kernels and counterfactual reasoning layers as the network architecture, which can effectively improve the accuracy and intelligence level of modeling of the power infrastructure construction process.

[0039] Specifically, by combining node attributes (such as construction timestamps and equipment status time series data) with edge weights (such as construction connection relationships), the hierarchical spatiotemporal attention-physical constraint graph convolutional network can capture the dynamic spatiotemporal dependencies in the construction process; the causal spatiotemporal attention unit enhances the recognition of key construction events by analyzing the causal relationship between nodes; the physical constraint convolution kernel combines the construction physics rules (such as equipment motion constraints and environmental restrictions) to ensure that the model output meets the actual construction conditions; the counterfactual reasoning layer simulates the results under different construction conditions to provide a scientific basis for optimizing construction decisions. This network architecture can not only achieve high-precision construction process simulation, but also provide intelligent support for construction efficiency improvement, risk prediction and resource optimization, significantly improving the scientificity and efficiency of power infrastructure construction management.

[0040] Based on the causal spatiotemporal attention unit in the network architecture, the temporal dependency is divided in combination with the construction sequence labels, and the spatial attention distribution is adjusted using the node timestamp difference and physical connection strength; Specifically, in the causal spatiotemporal attention unit, the causal relationship and spatiotemporal dependency in time series data will be captured through the attention mechanism; The expression based on the attention mechanism is: ; ; in, Indicates the current node For neighbor nodes The attention weight, For the current node For neighbor nodes The attention score between represents the learnable parameter vector, It is used to traverse the current node The variables of all neighbor nodes of Indicates the current node The neighbor set of represents the learnable weight matrix, and The front nodes are For neighbor nodes The characteristic vector of represents the concatenation operation of vectors, represents the activation function; It should be noted that the attention mechanism formula is based on the framework of the Graph Attention Network (GAT). Specifically, by introducing the attention mechanism, it can dynamically allocate the association weights between nodes, thereby capturing the temporal dependencies and physical connection strengths between nodes.

[0041] Dynamically allocate cross-stage node association weights through the Softmax function, strengthen the spatiotemporal correlation of high-risk nodes in the same process, and output dynamically calculated attention weights; Preferably, the Softmax function is used to dynamically allocate the association weights of nodes across stages, strengthen the spatiotemporal correlation of high-risk nodes, and generate dynamic attention weights. This can accurately capture the dependencies between key construction nodes, provide precise input for subsequent causal spatiotemporal attention units and counterfactual reasoning layers, optimize risk prediction and resource scheduling, and improve construction modeling accuracy and intelligent decision-making capabilities.

[0042] The physical parameters in the edge weights are spliced ​​with the node device status, input into the multi-layer perceptron to generate the physical perception convolution kernel, and the kernel weight is dynamically adjusted according to the environmental parameters; Specifically, the physical parameters in the edge weights (such as the physical connection strength between devices or the construction connection relationship) are spliced ​​with the node device status time series data (such as the operating status or performance indicators of the equipment) through data preprocessing tools to form a multidimensional feature vector; then the multidimensional feature vector is input into the multi-layer perceptron (MLP), and the MLP extracts and transforms the multidimensional feature vector layer by layer through the fully connected layer and nonlinear activation function (such as ReLU) to generate preliminary convolution kernel weights; then the convolution kernel weights are dynamically adjusted in combination with environmental parameter data (such as temperature, humidity or wind speed), and the correlation between environmental parameters and convolution kernel weights is calculated to ensure that the convolution kernel can adapt to the physical constraints under different environmental conditions; finally, the dynamically adjusted convolution kernel weights are applied to the graph convolution operation to capture the physical perception relationship between nodes, and provide convolution kernel support that conforms to actual physical rules for the subsequent spatiotemporal graph convolution network.

[0043] By integrating attention weights and physical constraint convolution kernels, multi-head graph convolution is performed on node features to capture local physical connections, global process dependencies, short-term state fluctuations, and long-term environmental trends, and output spatiotemporal hidden features. The better one is to integrate the attention weight and the physical constraint convolution kernel, capture the local physical connection, global process dependency, short-term state fluctuation and long-term environmental trend respectively through multi-head graph convolution, comprehensively extract the multi-dimensional spatiotemporal information of node features, and output the spatiotemporal hidden features, so as to provide multi-dimensional data support for the subsequent precise modeling of construction process dynamics, optimization of risk prediction and resource scheduling.

[0044] Perform counterfactual intervention on the device status of each node, recalculate the graph convolution output, and generate a risk difference matrix by comparing the Euclidean distance between actual and counterfactual features; Specifically, based on the spatiotemporal hidden feature matrix, the actual feature vector of each node is extracted as a benchmark for counterfactual comparison; counterfactual intervention is performed on the device state of each node, and counterfactual state data is generated by setting the device state as a benchmark value; based on the counterfactual state data, the graph convolution operation is re-executed to output the counterfactual feature vector; the actual feature vector is compared with the counterfactual feature vector, the Euclidean distance is calculated, and the node-level risk difference value is generated; the risk difference values ​​of all nodes are mapped according to the grid topology to generate a risk difference matrix, which is normalized and output as a standardized risk difference matrix.

[0045] Furthermore, based on the risk difference matrix, the Monte Carlo method is combined to dynamically simulate the construction behavior of high-risk nodes (such as equipment movement path, material stacking range), and compare them with the land red line in GIS in real time to calculate the conflict probability (such as crossing the boundary probability, intrusion area). At the same time, multi-level warning thresholds are set (such as distance threshold from the red line, conflict area threshold). When the simulation results show that the construction activity approaches or exceeds the red line, a graded warning (yellow warning, red warning) is triggered.

[0046] It should be noted that in actual construction scenarios, equipment may suffer from abnormal conditions (such as failure or performance degradation), which may lead to reduced construction efficiency or increased risks. The actual feature vector is the key data that captures these real conditions, providing a basis for subsequent risk analysis and modeling. The counterfactual feature vector provides a benchmark for comparison with the actual feature vector, which can quantify the impact of equipment abnormalities on the construction process, reveal the risk differences caused by abnormal conditions, and provide a scientific basis for optimizing construction decisions and risk prevention and control.

[0047] Preferably, the comparative analysis between the actual feature vector and the counterfactual feature vector can accurately quantify the impact of equipment abnormalities on the construction process, generate node-level risk difference values, help identify high-risk nodes and their associated construction stages, and provide a data-driven scientific basis for construction risk prediction, resource optimization, and decision support, significantly improving the level of intelligent construction management.

[0048] The spatiotemporal hidden features are weightedly fused with the risk difference matrix, mapped into a risk propagation probability matrix in the three-dimensional space-time dimension through the Sigmoid function, and the risk value of each grid node in each construction stage is marked; Specifically, the spatiotemporal graph convolutional network is used to perform multi-level feature extraction on the spatiotemporal graph to capture local physical connections, global process dependencies, short-term state fluctuations, and long-term environmental trends.

[0049] Spatiotemporal graph convolution, the expression is: ; in, and Respectively represent Layer and The node feature matrix of the layer, Indicates A spatiotemporal adjacency matrix, Indicates The degree matrix of the spatiotemporal adjacency matrix, Indicates Tier The weight matrix of the convolution kernel, ReLU activation function. The last layer of temporal and spatial hidden features Enter the fully connected layer through Function generates risk propagation probability matrix , the expression is: ; in, represents the normalization operation, which is used to convert the output into a probability distribution; Hiding features of space and time Risk Difference Matrix Perform weighted fusion to generate a fusion feature matrix, the expression is: ; in, represents the fusion feature matrix, Represents the weight parameter, which is used to balance the contribution of spatiotemporal hidden features and risk difference matrix, and its value range is ; The fused feature matrix Input the Sigmoid function and map it into a risk propagation probability matrix in the three-dimensional space-time dimension , the expression is: in, and They represent the learned weight matrix and bias vector respectively, Represents the activation function, which is used to limit the output to within the range.

[0050] Calculate the dynamic threshold of risk probability according to the construction stage, set the warning threshold, select the nodes exceeding the warning threshold as high-risk targets, and trace back the construction stage corresponding to the risk peak, and output the high-risk nodes and related construction stages; Specifically, the specific steps for setting the warning threshold are as follows: first, based on the hidden characteristics of time and space, use probabilistic statistical methods (such as kernel density estimation or Gaussian mixture model) to calculate the node risk probability distribution according to the construction stage, and dynamically determine the risk probability threshold of each stage; then, combine historical construction data and personnel experience, and dynamically adjust the risk probability threshold through a sliding window or adaptive algorithm to ensure that the risk probability threshold can reflect the actual situation of the current construction stage; then, set the warning line according to the risk probability distribution and threshold, and mark the nodes beyond the warning line as potential high-risk targets, providing a basis for subsequent risk screening.

[0051] Furthermore, nodes whose risk probability exceeds the warning threshold are screened out as high-risk targets; then, the risk change curves of these nodes during the construction process are analyzed through a backtracking algorithm to locate the construction stage corresponding to their risk peak; finally, the high-risk nodes and their associated construction stages are integrated and output to form a risk warning report, which provides accurate risk positioning and stage-related information for construction management, and supports risk prevention and control and construction optimization decisions.

[0052] S4. Construct a causal Bayesian network to locate the root cause equipment and process through counterfactual simulation of the risk evolution difference after the abnormality of computing equipment is eliminated.

[0053] Specifically, the following steps are included: Based on high-risk nodes and associated construction stages, a Bayesian network of causal dependencies between equipment status and construction processes is constructed; Specifically, the equipment status time series data and construction process information are extracted from the high-risk nodes and related construction stage data output by the spatiotemporal graph convolutional network to clarify the correlation between the high-risk nodes and the construction process; then, the potential causal relationship between the equipment status and the construction process is defined through historical data analysis, and the node and edge structure of the Bayesian network is determined; then, the network parameters are trained using maximum likelihood estimation or Bayesian learning methods to quantify the conditional probability distribution between the equipment status and the construction process; finally, by verifying the accuracy of the network structure and the rationality of the causal relationship, it is ensured that the Bayesian network can accurately reflect the causal dependency between the equipment status and the construction process, providing a causal reasoning basis for subsequent counterfactual simulation and risk evolution analysis.

[0054] Generate counterfactual intervention scenarios through Bayesian networks, simulate the state changes after the equipment anomaly is eliminated, and output counterfactual state data; It should be noted that the counterfactual state data is generated by simulating the operating state of the equipment after the abnormality is eliminated through the Bayesian network. The specific process is to build a causal Bayesian network based on actual construction data, define the abnormal state of the equipment and its influencing factors; simulate the operation of the equipment in a normal state in the Bayesian network through intervention operations (such as "assuming that the equipment has not been abnormal"), and generate counterfactual scenarios; then use Bayesian reasoning to calculate the state changes of the equipment in the counterfactual scenario and output the counterfactual state data. These data reflect the operating state of the equipment under assumed normal conditions, provide benchmark data for the calculation of the subsequent risk propagation probability matrix, and help quantify the impact of equipment abnormalities on the construction process.

[0055] Input the spatiotemporal graph convolutional network to calculate the risk propagation probability matrix under the counterfactual scenario; Compare the risk propagation probability matrix of the actual scenario with the counterfactual scenario to generate the risk evolution difference matrix; It should be noted that the actual scenario is generated based on real data from the construction site, including equipment status time series data, environmental parameter data, construction timestamps, node attributes and other information, reflecting the status of the nodes under real conditions and risk propagation during the construction process; the counterfactual scenario is generated by simulating the operating status of the equipment after the abnormality is eliminated through the Bayesian network, reflecting the status of the equipment under assumed normal conditions and risk propagation. By comparing the risk propagation probability matrix of the actual scenario and the counterfactual scenario, a risk evolution difference matrix is ​​generated, which can accurately quantify the impact of equipment abnormalities on the overall risk, identify the root cause equipment with the highest risk contribution, and trace back the impact path of its abnormal status on adjacent nodes and construction processes.

[0056] Preferably, the counterfactual state data and the comparative analysis between the actual scenario and the counterfactual scenario can accurately locate the root cause equipment and key processes, reveal the risk impact path of equipment abnormalities on the construction process, and significantly improve the level of intelligent construction management.

[0057] Based on the risk evolution difference matrix, identify the root cause devices that contribute most to the overall risk; Specifically, the risk contribution value of each equipment node is extracted from the risk evolution difference matrix. These risk contribution values ​​reflect the difference in risk changes before and after the equipment abnormality is eliminated. Then, the equipment nodes with the highest risk contribution values ​​are screened out through a sorting algorithm (such as Top-K sorting) as potential root cause equipment. Then, the abnormal status of these root cause equipment during the construction process and their impact range are analyzed in combination with the equipment status time series data and construction process information to verify their rationality as root cause equipment.

[0058] For example, if the risk contribution value of a crane in multiple construction stages is significantly higher than that of other equipment, and its abnormal state (such as failure or performance degradation) directly leads to construction delays or safety accidents, then the crane can be determined to be the root cause equipment; finally, the identified root cause equipment and its risk contribution value are integrated and output to provide accurate target equipment for construction risk prevention and control and resource optimization, and support subsequent decision-making and risk mitigation measures.

[0059] Trace back the impact path of abnormal status of root-cause equipment on adjacent nodes and construction processes, and locate key processes.

[0060] S5. Dynamically mark the root cause coordinates and risk diffusion range in the high-precision three-dimensional model, generate early warning instructions containing visual paths, root cause reports and disposal priorities, and push them to the terminal in real time through the edge node.

[0061] Specifically, the following steps are included: Based on the located root cause device, extract its spatial coordinate data and dynamically mark the root cause coordinates in the high-precision 3D model; It should be noted that dynamically marking the root cause coordinates can intuitively display the specific location of the root cause equipment, providing accurate spatial reference for construction risk positioning, resource scheduling and decision optimization.

[0062] Starting from the root cause coordinates, the risk diffusion range of the root cause equipment is calculated according to the risk evolution difference matrix, and the diffusion boundary is visually marked in the high-precision 3D model. Preferably, by visually marking the diffusion boundary, the regional scope of risk impact can be intuitively displayed, providing an accurate spatial basis for construction risk prevention and control, resource optimization and decision-making, and significantly improving the scientificity and efficiency of construction management.

[0063] Combined with the risk propagation probability matrix, the risk propagation path of high-risk nodes is extracted and visualized in a high-precision 3D model; Integrate the root cause equipment and related process information to generate a detailed root cause report including equipment status, abnormal reasons, impact range and related processes; Specifically, through data fusion and knowledge graph technology, the status data of the root cause equipment, the cause of the abnormality, the scope of impact and the related process information are structured and integrated to generate a detailed root cause report, providing comprehensive and accurate data support for construction risk analysis and decision optimization.

[0064] Calculate the disposal priority of each root cause equipment based on the risk value and diffusion range of the nodes in the risk evolution difference matrix; Integrate visualization paths, root cause reports, and treatment priorities to generate structured warning instructions; Furthermore, data visualization and interaction support will be provided by a 3D visualization platform based on WebGL / Three.js, which will support real-time loading of high-precision models, conflict heat maps, and warning information in browsers or mobile devices.

[0065] The interactive functions provided are as follows: multi-dimensional view switching: overlay / hide land use red lines, planning scope, and construction progress layers; conflict penetration query: click on the conflict area to view the specific equipment number, responsible unit, and disposal time limit; early warning disposal tracking: mark the processed areas and retain the operation log to form a closed-loop management.

[0066] The warning instructions are pushed to the terminal devices in real time through the edge nodes, and countermeasures are taken.

[0067] The present embodiment also provides an electric power infrastructure process early warning system based on a high-precision three-dimensional model, including: a data acquisition module, a graph mapping module, a risk identification module, a root cause positioning module and an early warning push module; the data acquisition module is used to collect multi-source data of the electric power infrastructure construction site, and construct a high-precision three-dimensional model including equipment space coordinates, construction time sequence labels and environmental parameters; the graph mapping module is used to map into a space-time graph according to the grid topology, the node attributes encode the construction timestamp and the time sequence change of the equipment status, and the edge weights are associated with the construction connection relationship and the physical connection strength of adjacent nodes; the risk identification module is used to input the space-time graph into the space-time graph convolution network, output the risk propagation probability matrix of the three-dimensional space and construction time sequence dimensions, and identify high-risk nodes and associated construction stages; the root cause positioning module is used to construct a causal Bayesian network, calculate the risk evolution difference after the equipment anomaly is eliminated through counterfactual simulation, and locate the root cause equipment and process; the early warning push module is used to dynamically mark the root cause coordinates and the risk diffusion range in the high-precision three-dimensional model, generate an early warning instruction containing a visualization path, a root cause report and a disposal priority, and push it to the terminal in real time through the edge node.

[0068] This embodiment also provides a computer device, which is suitable for the case of an electric power infrastructure process early warning method based on a high-precision three-dimensional model, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the electric power infrastructure process early warning method based on a high-precision three-dimensional model as proposed in the above embodiment.

[0069] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0070] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for realizing an early warning method for a power infrastructure process based on a high-precision three-dimensional model as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic storage device, a flash memory, a disk or an optical disk.

[0071] In summary, the present invention realizes multi-dimensional feature encoding of the construction process by constructing a high-precision three-dimensional model that integrates equipment spatial coordinates, construction timing labels and environmental parameters, and mapping it into a spatiotemporal graph structure. Based on the hierarchical spatiotemporal attention-physical constraint graph convolutional network, the cross-stage association weights between nodes are dynamically allocated, and the convolution kernel parameters are optimized in combination with the physical connection strength, which significantly improves the spatiotemporal correlation of high-risk node identification. The risk difference matrix is ​​generated through the counterfactual intervention mechanism to quantify the contribution of equipment anomalies to the overall risk. The root cause equipment is more accurately located by tracing back the anomaly propagation path in combination with the causal Bayesian network. In addition, the root cause coordinates and risk diffusion range are dynamically labeled, and a visual warning instruction with disposal priority is generated, which is pushed in real time through the edge node, greatly shortening the risk response time. This method breaks through the limitations of traditional static monitoring and experience-driven causal reasoning, realizes accurate modeling of risk propagation in the three-dimensional space-time dimension, and provides a full-link closed-loop solution from risk identification, root cause tracing to intelligent decision-making for power infrastructure projects.

[0072] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for early warning of power infrastructure construction process based on high-precision three-dimensional model, characterized by: include, Collect multi-source data from power infrastructure construction sites and build high-precision three-dimensional models that include equipment spatial coordinates, construction sequence labels, and environmental parameters; The grid topology is mapped into a time-space graph, the node attributes encode the construction timestamp and the time-series change of the equipment status, and the edge weights are associated with the construction connection relationship and physical connection strength of adjacent nodes; The space-time graph is input into the space-time graph convolutional network, and the risk propagation probability matrix in the three-dimensional space and construction time sequence dimensions is output to identify high-risk nodes and associated construction stages; Construct a causal Bayesian network to locate the root cause equipment and process by calculating the risk evolution difference after the abnormality of the computing equipment is eliminated through counterfactual simulation; The root cause coordinates and risk diffusion range are dynamically marked in the high-precision three-dimensional model, and early warning instructions with visual paths, root cause reports and disposal priorities are generated and pushed to the terminal in real time through the edge node.

2. The electric power infrastructure process early warning method based on a high-precision three-dimensional model as claimed in claim 1, characterized in that: The multi-source data includes equipment space coordinate data, construction time sequence label data, environmental parameter data, equipment status time sequence data and construction connection relationship data; The specific steps of constructing a high-precision three-dimensional model including equipment space coordinates, construction sequence labels and environmental parameters are as follows: Perform spatiotemporal alignment and cleaning on the collected equipment space coordinates, construction sequence labels and environmental parameter data, and output a standardized spatiotemporal data set; Extract the spatial coordinates of the device from the standardized spatiotemporal data set, combine it with the point cloud data collected by LiDAR, perform denoising and registration, and generate cleaned point cloud data; Convert the cleaned point cloud data into a 3D mesh model, optimize the number of facets and map the device space coordinates to generate a high-precision basic model; Bind construction timing labels and environmental parameters to equipment space nodes to form a high-precision three-dimensional model with timing and environmental attributes.

3. The electric power infrastructure process early warning method based on a high-precision three-dimensional model as claimed in claim 2, characterized in that: The grid topology is mapped into a time-space graph, the node attributes encode the construction timestamp and the time series change of the equipment status, and the edge weights are associated with the construction connection relationship and physical connection strength of adjacent nodes. The specific steps are as follows: The high-precision three-dimensional model is divided into grid units at a fixed resolution, and each grid unit serves as a node in the space-time graph; Extract construction timestamps and equipment status time series data from the standardized spatiotemporal data set, bind them to the corresponding grid nodes, and form node attributes; According to the equipment space coordinates and construction connection relationship data, the connection relationship between adjacent grid nodes is determined, and the physical connection strength is calculated as the edge weight; The node attributes and edge weights are integrated to generate a spatiotemporal graph, and the graph structure is optimized using a graph optimization algorithm.

4. The electric power infrastructure process early warning method based on a high-precision three-dimensional model as claimed in claim 3 is characterized by: The spatiotemporal graph is input into the spatiotemporal graph convolutional network, and the risk propagation probability matrix of the three-dimensional space and construction time sequence dimensions is output to identify high-risk nodes and associated construction stages. The specific steps are as follows: Based on the node attributes and edge weights of the spatiotemporal graph, a hierarchical spatiotemporal attention-physical constraint graph convolutional network is constructed, integrating causal spatiotemporal attention units, physical constraint convolution kernels, and counterfactual reasoning layers as the network architecture; Based on the causal spatiotemporal attention unit in the network architecture, the temporal dependency is divided in combination with the construction sequence labels, and the spatial attention distribution is adjusted using the node timestamp difference and physical connection strength; Dynamically allocate cross-stage node association weights through the Softmax function, strengthen the spatiotemporal correlation of high-risk nodes in the same process, and output dynamically calculated attention weights; The physical parameters in the edge weights are spliced ​​with the node device status, input into the multi-layer perceptron to generate the physical perception convolution kernel, and the kernel weight is dynamically adjusted according to the environmental parameters; By integrating attention weights and physical constraint convolution kernels, multi-head graph convolution is performed on node features to capture local physical connections, global process dependencies, short-term state fluctuations, and long-term environmental trends, and output spatiotemporal hidden features. Perform counterfactual intervention on the device status of each node, recalculate the graph convolution output, and generate a risk difference matrix by comparing the Euclidean distance between actual and counterfactual features; The spatiotemporal hidden features are weightedly fused with the risk difference matrix, mapped into a risk propagation probability matrix in the three-dimensional space-time dimension through the Sigmoid function, and the risk value of each grid node in each construction stage is marked; Calculate the dynamic threshold of risk probability according to the construction stage, set the warning threshold, screen the nodes exceeding the warning threshold as high-risk targets, and trace back to the construction stage corresponding to the risk peak, and output the high-risk nodes and related construction stages.

5. The electric power infrastructure process early warning method based on a high-precision three-dimensional model as claimed in claim 4 is characterized by: The counterfactual intervention is performed on the device state of each node, the graph convolution output is recalculated, and the risk difference matrix is ​​generated by comparing the Euclidean distance between the actual and counterfactual feature vectors. The specific steps are as follows: Based on the spatiotemporal hidden feature matrix, the actual feature vector of each node is extracted as the benchmark for counterfactual comparison; Perform counterfactual intervention on the device state of each node and generate counterfactual state data by setting the device state as a baseline value; Based on the counterfactual state data, re-execute the graph convolution operation and output the counterfactual feature vector; Compare the actual feature vector with the counterfactual feature vector, calculate the Euclidean distance, and generate the node-level risk difference value; The risk difference values ​​of all nodes are mapped according to the grid topology to generate a risk difference matrix, which is then normalized and output as a standardized risk difference matrix.

6. The electric power infrastructure process early warning method based on a high-precision three-dimensional model as claimed in claim 5, characterized in that: The causal Bayesian network is constructed to locate the root cause equipment and process by counterfactual simulation of the risk evolution difference after the abnormality of the computing equipment is eliminated. The specific steps are as follows: Based on high-risk nodes and associated construction stages, a Bayesian network of causal dependencies between equipment status and construction processes is constructed; Generate counterfactual intervention scenarios through Bayesian networks, simulate the state changes after the equipment anomaly is eliminated, and output counterfactual state data; Input the spatiotemporal graph convolutional network to calculate the risk propagation probability matrix under the counterfactual scenario; Compare the risk propagation probability matrix of the actual scenario with the counterfactual scenario to generate the risk evolution difference matrix; Based on the risk evolution difference matrix, identify the root cause devices that contribute most to the overall risk; Trace back the impact path of abnormal status of root-cause equipment on adjacent nodes and construction processes, and locate key processes.

7. The electric power infrastructure process early warning method based on high-precision three-dimensional model according to claim 6 is characterized by: The root cause coordinates and risk diffusion range are dynamically marked in the high-precision three-dimensional model, and an early warning instruction containing a visualization path, a root cause report and a disposal priority is generated and pushed to the terminal in real time through the edge node. The specific steps are as follows: Based on the located root cause device, extract its spatial coordinate data and dynamically mark the root cause coordinates in the high-precision 3D model; Starting from the root cause coordinates, the risk diffusion range of the root cause equipment is calculated according to the risk evolution difference matrix, and the diffusion boundary is visually marked in the high-precision 3D model; Combined with the risk propagation probability matrix, the risk propagation path of high-risk nodes is extracted and visualized in a high-precision 3D model; Integrate the root cause equipment and related process information to generate a detailed root cause report including equipment status, abnormal reasons, impact range and related processes; Calculate the disposal priority of each root cause equipment based on the risk value and diffusion range of the nodes in the risk evolution difference matrix; Integrate visualization paths, root cause reports, and treatment priorities to generate structured warning instructions; The warning instructions are pushed to the terminal devices in real time through the edge nodes, and countermeasures are taken.

8. An electric power infrastructure process early warning system based on a high-precision three-dimensional model, based on the electric power infrastructure process early warning method based on a high-precision three-dimensional model according to any one of claims 1 to 7, characterized in that: Including data collection module, graph mapping module, risk identification module, root cause location module and warning push module; The data acquisition module is used to collect multi-source data from the power infrastructure construction site and construct a high-precision three-dimensional model including equipment spatial coordinates, construction timing labels and environmental parameters; The graph mapping module is used to map the grid topology into a spatiotemporal graph, the node attributes encode the construction timestamp and the equipment status time series change, and the edge weights associate the construction connection relationship of adjacent nodes with the physical connection strength; The risk identification module is used to input the space-time graph into the space-time graph convolutional network, output the risk propagation probability matrix in the three-dimensional space and construction time sequence dimensions, and identify high-risk nodes and associated construction stages; The root cause location module is used to construct a causal Bayesian network, locate the root cause equipment and process by calculating the risk evolution difference after the abnormality of the computing equipment is eliminated through counterfactual simulation; The warning push module is used to dynamically mark the root cause coordinates and risk diffusion range in the high-precision three-dimensional model, generate warning instructions containing visualization paths, root cause reports and disposal priorities, and push them to the terminal in real time via the edge node.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the electric power infrastructure process early warning method based on a high-precision three-dimensional model as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the electric power infrastructure process early warning method based on a high-precision three-dimensional model as described in any one of claims 1 to 7 are implemented.

Images