Power distribution engineering line design construction method and platform supporting multi-scenario simulation

Abstract

The application discloses a power distribution engineering line design construction method and platform supporting multi-scene simulation and relates to the related field of digital twinning. The method comprises the following steps: combining digital twinning to construct a power distribution engineering specification space, calling an explicit structure priori, and initializing an engineering point set; collecting multi-source data on a design construction area, predicting the attributes of the initialized engineering point set by using a cross-attention mechanism; adaptively configuring an initialized power distribution engineering line twin model; simulating a scene based on a design construction scene set, and managing the design construction according to the simulation result. The application solves the technical problems of the disconnection between the design scheme and the dynamic environment, the insufficient use of multi-source data, the lack of full-life-cycle simulation, and the insufficient risk predictability in the existing power distribution engineering line design construction management, and achieves the technical effects of dynamic self-adaptation of design parameters, intelligent fusion of multi-source data, full-life-cycle multi-scene simulation, and pre-control of construction operation risks.

Description

Technical Field

[0001] This application relates to the field of digital twins, and in particular to a method and platform for designing and constructing power distribution lines that supports multi-scenario simulation. Background Technology

[0002] The design and construction quality of power distribution lines directly affects the safe and stable operation of the power system and the reliability of power supply, making it a core and critical issue in the field of power engineering. Currently, the industry mainly relies on traditional CAD design, experience-based construction scheme formulation, and physical simulation under limited scenarios to carry out related work. Existing methods lack the ability to deeply integrate multi-source environmental data and are difficult to achieve full-process dynamic simulation and accurate prediction, resulting in problems such as insufficient adaptability of design schemes, delayed prediction of construction risks, and incomplete scenario coverage.

[0003] At present, the design, construction and management of power distribution lines suffer from technical problems such as the disconnect between design schemes and dynamic environments, insufficient utilization of multi-source data, lack of full life cycle simulation, and insufficient risk foresight. Summary of the Invention

[0004] This application provides a design and construction method and platform for power distribution lines that supports multi-scenario simulation. It employs digital twin technology, integrating explicit prior knowledge such as tower structure, conductor topology, and geological structure to construct a power distribution engineering specification space. This involves initializing the engineering point set and the initial twin model, collecting multi-source data from the design and construction area, predicting the attributes of the initial engineering point set through a cross-attention mechanism, adaptively configuring the initial twin model based on the attribute prediction results to obtain a basic twin model, acquiring a set of design and construction scenarios, conducting scenario simulations on the basic twin model, and managing the design and construction of power distribution lines based on the simulation results. This addresses the technical problems in existing power distribution line design and construction management, such as the disconnect between design schemes and dynamic environments, insufficient utilization of multi-source data, lack of full lifecycle simulation, and insufficient risk foresight. It achieves the technical effects of dynamic adaptive design parameters, intelligent fusion of multi-source data, full lifecycle multi-scenario simulation, and proactive risk prevention and control during construction and operation.

[0005] This application provides a method for designing and constructing power distribution lines that supports multi-scenario simulation. The method includes: constructing a power distribution engineering specification space using digital twins; calling tower structure templates, conductor topology constraints, and geological structures as explicit structural priors; initializing the engineering point set in the power distribution engineering specification space to obtain an initialized engineering point set and an initialized power distribution line twin model; collecting multi-source data on the design and construction area of ​​the target power distribution line to obtain a multi-source regional data set; using a cross-attention mechanism to predict the attributes of the initialized engineering point set based on the multi-source regional data set to obtain a set of engineering point attribute prediction results; adaptively configuring the initialized power distribution line twin model based on the set of engineering point attribute prediction results to obtain a basic power distribution line twin model; acquiring a design and construction scenario set; performing scenario simulation on the basic power distribution line twin model based on the design and construction scenario set; and managing design and construction based on the scenario simulation results.

[0006] In one possible implementation, a power distribution engineering specification space is constructed using digital twins, and the following processes are performed: standard pole and tower models, standard conductor parameters, standard foundation models, and standard equipment models are acquired; basic information of the design and construction area of ​​the target power distribution line is acquired, and a neutral terrain model is matched based on the basic information; the standard pole and tower models, standard conductor parameters, standard foundation models, and standard equipment models are imported into the neutral terrain model using digital twins, and unified descriptive coordinates and a unified attribute structure are added to obtain the power distribution engineering specification space to be commissioned; based on the basic information, a neutral environmental state is matched, and environmental simulation is performed on the power distribution engineering specification space to be commissioned to construct the power distribution engineering specification space.

[0007] In a possible implementation, the tower structure template, conductor topology constraints, and geological structure are used as explicit structural priors to initialize the engineering point set in the power distribution engineering specification space, obtaining an initialized engineering point set and an initialized power distribution line twin model. The following processes are performed: Based on the tower structure template, the topological connection relationship of the tower components is mapped to the engineering point set to obtain the first engineering point set; Based on the conductor topology constraints, the key points of the conductor in the first engineering point set are bound as suspension points, tension points, or segmentation points, and initial tension parameters, sag parameters, and wind, ice, and temperature offset coefficients are written in, and a chain-like elastic topology structure of the conductor points is established to obtain the second engineering point set; Based on the geological structure parameters, the foundation bearing capacity, settlement coefficient, lateral displacement limit, surface friction resistance, and lithological sensitivity are written into the second engineering point set, and the foundation points are associated with the terrain points so that the terrain slope affects the tower foundation tilt direction and the geological properties affect the tower foundation settlement, obtaining the initialized engineering point set; Based on the initialized engineering point set, an initialized power distribution line twin model is constructed.

[0008] In a possible implementation, the following processing is performed: Based on the tower structure template, the tower foot points in the engineering point set are set as fixed constraint points, the tower body nodes are set as rigid connection points, the tower head nodes are set as swingable nodes, and the stiffness matrix, degrees of freedom, and connection adjacency relationships are written to obtain the first engineering point set.

[0009] In a possible implementation, multi-source data is collected from the design and construction area of ​​the target power distribution line to obtain a multi-source regional data set. A cross-attention mechanism is used to predict the attributes of the initial engineering point set based on the multi-source regional data set to obtain a set of engineering point attribute prediction results. The following processing is performed: extracting a first engineering point and a first engineering point query feature from the initial engineering point set; extracting the coordinates of the first engineering point from the first engineering point query feature; matching the multi-source regional data set based on the first engineering point coordinates to obtain a first matched multi-source data set; using a cross-attention mechanism, performing attribute prediction based on the first matched multi-source data set and the first engineering point query feature to obtain a first engineering point attribute prediction result; and adding the first engineering point attribute prediction result to the set of engineering point attribute prediction results.

[0010] In a possible implementation, a cross-attention mechanism is adopted to perform attribute prediction based on the first matching multi-source data set and the first engineering point query features to obtain the first engineering point attribute prediction result, and the following processing is performed: the dynamic weight library of the cross-attention mechanism is retrieved, and dynamic weight matching is performed in combination with the first engineering point query features to obtain the first dynamic weight; the first matching multi-source data set and the first engineering point query features are used as input, and cross-attention attribute prediction is performed in combination with the first dynamic weight to obtain the first engineering point attribute prediction result.

[0011] In a possible implementation, a dynamic weight library across attention mechanisms is retrieved, and dynamic weight matching is performed in conjunction with the query features of the first engineering point to obtain a first dynamic weight. The following processing is then performed: using the query features of the first engineering point as an index, the similarity of the query features corresponding to each dynamic weight in the dynamic weight library is calculated, and the dynamic weights corresponding to the similarity values ​​in the top m positions are selected as the matching dynamic weight set; the mean of the matching dynamic weight set is calculated to obtain the mean of the matching dynamic weights; using the mean of the matching dynamic weights as the starting point, the mean drift filter is performed on the matching dynamic weight set to determine the first dynamic weight.

[0012] In a possible implementation, the initial power distribution line twin model is adaptively configured based on the set of predicted project point attributes to obtain a basic power distribution line twin model. The following processing is then performed: based on the geometric offset, stress offset, environmental sensitivity, geological response, and material state of each project point attribute prediction result in the set of predicted project point attributes, the initial power distribution line twin model is configured independently for each project point. The global optimization configurator is then called to perform global consistency optimization on the initial power distribution line twin model after the independent configuration of project points, to obtain the basic power distribution line twin model.

[0013] In a possible implementation, a set of design and construction scenarios is obtained, and the following processing is performed: using the basic information of the design and construction area of ​​the target power distribution line as an index, big data mining of design and construction scenarios is carried out to obtain a set of mined design and construction scenarios; the mined design and construction scenario set is aggregated to obtain multiple aggregated mined design and construction scenario sets; the multiple aggregated mined design and construction scenario sets are traversed to perform representative filtering within the set, and the filtering results are added to the design and construction scenario set.

[0014] This application also provides a power distribution line design and construction platform supporting multi-scenario simulation, including: a power distribution engineering specification space construction module, used to construct a power distribution engineering specification space in conjunction with digital twins, calling tower structure templates, conductor topology constraints, and geological structures as explicit structural priors to initialize the engineering point set of the power distribution engineering specification space, obtaining an initialized engineering point set and an initialized power distribution line twin model; an attribute prediction module, used to collect multi-source data on the design and construction area of ​​the target power distribution line, obtain a multi-source regional data set, and use a cross-attention mechanism to predict attributes of the initialized engineering point set based on the multi-source regional data set, obtaining a set of engineering point attribute prediction results; an adaptive configuration module, used to adaptively configure the initialized power distribution line twin model based on the set of engineering point attribute prediction results, obtaining a basic power distribution line twin model; and a scenario simulation module, used to acquire a design and construction scenario set, perform scenario simulation on the basic power distribution line twin model based on the design and construction scenario set, and perform design and construction management based on the scenario simulation results.

[0015] The proposed method and platform for designing and constructing power distribution lines, supporting multi-scenario simulation, firstly constructs a power distribution engineering specification space using digital twins. It then uses tower structure templates, conductor topology constraints, and geological structures as explicit structural priors to initialize the engineering point set in the specification space, obtaining an initialized engineering point set and an initialized power distribution line twin model. Next, it collects multi-source data from the design and construction area of ​​the target power distribution line, obtaining a multi-source regional data set. A cross-attention mechanism is used to predict the attributes of the initialized engineering point set based on the multi-source regional data set, obtaining a set of predicted engineering point attributes. Then, based on the predicted engineering point attributes, the initialized power distribution line twin model is adaptively configured to obtain a basic power distribution line twin model. Finally, a design and construction scenario set is obtained, and scenario simulations are performed on the basic power distribution line twin model based on these scenario sets. Design and construction management is then performed based on the simulation results. Through this process, the proposed method and platform achieve the technical effects of dynamic adaptive design parameters, intelligent fusion of multi-source data, full lifecycle multi-scenario simulation, and proactive risk prevention and control during construction and operation. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a flowchart illustrating a power distribution line design and construction method that supports multi-scenario simulation, as provided in an embodiment of this application.

[0018] Figure 2 This is a structural diagram of a power distribution engineering line design and construction platform that supports multi-scenario simulation, provided in an embodiment of this application.

[0019] Figure labeling: Module 10 for power distribution engineering specification space construction, Module 20 for attribute prediction, Module 30 for adaptive configuration, and Module 40 for scenario simulation. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0021] This application provides a method for designing and constructing power distribution lines that supports multi-scenario simulation, such as... Figure 1 As shown, the method includes:

[0022] Step S100: Combine digital twin to construct the power distribution engineering specification space, call the tower structure template, conductor topology constraints and geological structure as explicit structural priors, initialize the engineering point set of the power distribution engineering specification space, and obtain the initialized engineering point set and the initialized power distribution engineering line twin model.

[0023] Specifically, this involves establishing a standardized digital infrastructure framework for power distribution engineering and initially setting up the engineering point set. Digital twin refers to constructing a virtual model corresponding to the physical power distribution project using digital means. The power distribution engineering specification space is a virtual working environment integrating industry standards and design specifications. Explicit structural priors refer to known and clearly defined engineering structural constraints, including tower structure, conductor topology, and geological structure. The engineering point set is a collection of digital nodes representing key components such as towers, conductors, and foundations in the virtual model. The initialization process involves assigning initial attributes and relationships to these nodes, ultimately forming an optimizable initial virtual model—the initialization of the power distribution engineering line twin model.

[0024] In one possible implementation, combining digital twins to construct a power distribution engineering specification space, step S100 further includes step S110, acquiring standard tower models, conductor standard parameters, foundation standard models, and equipment standard models. Specifically, preset models and parameters are retrieved by accessing the power industry standard database or the company's internal technical specification library, and key information is extracted and converted into a unified data format using structured data parsing. For example, standard tower models can be obtained from the State Grid's "General Design Specifications for 110 kV and Below Distribution Line Towers" database, including three-dimensional structural models of different types such as straight towers, tension towers, and angle towers, containing parameters such as tower height, crossarm length, and tower material specifications; conductor standard parameters include industry standard values ​​such as the cross-sectional area, rated breaking force, and weight per unit length of steel-cored aluminum stranded wire; foundation standard models include dimensional parameters and bearing capacity standard values ​​for types such as cast-in-place concrete foundations and precast pile foundations; equipment standard models include data such as installation dimensions and interface specifications of power distribution equipment such as circuit breakers and disconnect switches. All data is converted into an extended markup language format that can be recognized by the digital twin platform and stored.

[0025] Step S120: Obtain basic information about the design and construction area of ​​the target power distribution line, and match a neutral terrain model based on the basic information. Specifically, retrieve macro-geographic data of the target area through a geographic information system interface, such as latitude and longitude range and overall terrain category. Combined with the basic requirements for terrain adaptation in the engineering design specifications, select a neutral terrain model that conforms to the definition of the blank base plate in the standardized terrain template library. Use a coordinate calibration algorithm to ensure that the model accurately matches the spatial range of the actual area. The neutral terrain model is a pre-constructed general terrain template. Its core features include no vegetation cover, no artificial structures such as buildings, slopes controlled within preset values, no extreme slopes, elevation smoothed using regional average elevation, foundation set as a uniform hard soil layer, no soft soil, karst caves or other unstable geological features, and only retain the basic flatness and macro-spatial outline of the terrain.

[0026] For example, the basic information extracted focuses on the boundary coordinates of the design and construction area, the overall terrain trend, and the outline of the area without extreme terrain. Using a spatial range matching algorithm, the boundary coordinates of the neutral terrain model are aligned with the boundary coordinates of the target area. Linear stretching or compression techniques are then used to adjust the model size, ensuring that the neutral terrain model completely covers the design and construction area, and that the basic terrain conditions are uniform without any localized special terrain interference. This provides a consistent blank base for subsequent import of the standard model.

[0027] Step S130: Using digital twins, the standard tower model, conductor standard parameters, foundation standard model, and equipment standard model are imported into the neutral terrain model. Unified descriptive coordinates and a unified attribute structure are added to obtain the specification space for the power distribution project to be commissioned. Specifically, the batch import interface of the digital twin platform is used to map the design coordinates of each standard model to the spatial coordinate system of the neutral terrain model through unified coordinate system transformation. A unified attribute field system is created using a structured attribute definition tool, and the model is spatially bound to the terrain base plate.

[0028] For example, a unified coordinate system is used, employing a three-dimensional Cartesian coordinate system. The X / Y axes correspond to planar coordinates, and the Z axis corresponds to elevation. All standard models are calibrated according to this coordinate system during import to ensure complete alignment between the tower foundation model and the neutral terrain model. The unified attribute structure includes core fields such as model category, standard number, dimensional parameters, material properties, installation constraints, and adaptation specifications. For instance, standard tower models require parameters such as tower type number, tower height, and crossarm spacing, while standard conductor parameters require information such as model, cross-sectional area, and rated tensile force. During import, a collision detection algorithm prevents spatial overlap between models. Models are arranged according to the minimum safe distance specified in the power distribution engineering design specifications, ultimately forming a power distribution engineering specification space containing standard models and unified attributes, with all models arranged based on a unified state of the neutral terrain base.

[0029] Step S140: Based on the basic information, a neutral environmental state is matched, and environmental simulation is performed on the specification space of the power distribution project to be commissioned to construct the power distribution project specification space. Specifically, based on the basic information of the target area, such as climate zoning and altitude, neutral environmental parameters without extreme weather are retrieved from the environmental standard database, i.e., a generalized benchmark environmental state. The environmental load simulation is performed on the specification space of the power distribution project to be commissioned using finite element simulation tools, and the environmental response data is integrated into the model to form a specification space that conforms to a unified environmental benchmark.

[0030] For example, the core parameters of a neutral environment include an air temperature of 25 degrees Celsius, a wind speed of 3 m / s, no precipitation, no icing, an atmospheric pressure of 101.3 kPa, and no extreme natural phenomena such as earthquakes or typhoons. An environment-structure interaction model is established using finite element analysis software to simulate the initial sag of conductors at normal temperature, the slight force exerted on towers by a light wind, and the structural stress distribution under standard atmospheric pressure. The initial state parameters of each model under a neutral environment are calculated, such as conductor sag and tower stress values. These parameters are then written into the model attributes, ensuring a unified environmental response benchmark for the power distribution engineering specifications to be commissioned, thus avoiding inconsistencies in initial states due to environmental differences.

[0031] In one possible implementation, the tower structure template, conductor topology constraints, and geological structure are used as explicit structural priors to initialize the engineering point set in the power distribution engineering specification space, obtaining the initialized engineering point set and the initialized power distribution engineering line twin model. Step S100 further includes step S150, which maps the topological connection relationship of the tower components to the engineering point set according to the tower structure template, obtaining the first engineering point set. Specifically, according to the tower structure template, the tower foot points in the engineering point set are set as fixed constraint points, the tower body nodes are set as rigid connection points, and the tower head nodes are set as swingable nodes. The stiffness matrix, degrees of freedom, and connection adjacency relationships are then written into the set. Specifically, the three-dimensional model of the tower structure template is extracted through topological structure analysis. Node constraint types are defined according to component function. The stiffness matrix of each node is calculated using material mechanics formulas to determine the degree of freedom parameters of each node, including translation and rotation permissions. The connection relationships between nodes are recorded using an adjacency list data structure.

[0032] For example, key engineering points such as tower feet, tower body segment nodes, crossarm endpoints, and tower head suspension points are extracted from the 3D model of the tower, and each node is assigned a unique identifier. Tower feet are set as fixed constraint points, with degrees of freedom defined as: translational degrees of freedom in the X / Y / Z directions = 0, rotational degrees of freedom around the X / Y / Z axes = 0, i.e., completely fixed. Tower body nodes are set as rigid connection points, with degrees of freedom defined as: translational degrees of freedom in the X / Y / Z directions = 0, rotational degrees of freedom around the X / Y / Z axes = 0, i.e., no relative displacement or rotation between nodes, maintaining tower rigidity. Tower head nodes are set as swingable nodes, with degrees of freedom defined as: translational degrees of freedom in the X / Y / Z directions = 0, rotational degrees of freedom around the horizontal axis (e.g., Y-axis) = 1, rotational degrees of freedom around the vertical axis (e.g., Z-axis) = 0, i.e., swinging is only allowed along the conductor suspension direction. The stiffness matrix is ​​calculated using material mechanics formulas. For example, the axial stiffness of tower body nodes is calculated using the formula: Stiffness = Elastic Modulus × Cross-sectional Area / Component Length. Bending stiffness and torsional stiffness parameters are simultaneously entered into the matrix. The adjacency list records the associated node number and connecting component of each node. For example, tower foot node 1 is associated with tower body node 2, and the connecting component is the tower leg angle steel, ensuring a complete mapping of node topology relationships.

[0033] Step S160: Based on the conductor topology constraints, bind the key points of the conductor in the first engineering point set as suspension points, tension points, or segmentation points, write the initial tension parameters, sag parameters, and wind, ice, and temperature offset coefficients, and establish a chain-like elastic topology structure for the conductor points to obtain the second engineering point set. Specifically, key points connected to the conductor are selected from the first engineering point set through conductor topology identification, such as tower head suspension points, conductor segmentation points, and terminal fixing points. Node types are assigned according to the power distribution engineering design specifications. Conductor physical parameters and environmental offset coefficients are entered through the parameter configuration tool. Based on the elasticity model, a chain-like connection relationship between nodes is constructed, and the degree of freedom constraints of each conductor point are determined.

[0034] For example, the suspension point of a straight-line tower is a hanging point, bearing only vertical loads, with X / Y translational degrees of freedom = 0 and Z translational degrees of freedom = 1. The suspension point of a tension tower is a tension point, bearing both horizontal and vertical loads, with X / Y / Z translational degrees of freedom = 0. For long-distance lines, a segmentation point is set every 500 meters to distribute tension, with X / Y translational degrees of freedom = 0 and Z translational degrees of freedom = 1. Initial tension parameters are set according to the conductor type. For example, the initial tension of LGJ-300 / 40 steel-cored aluminum stranded wire is 50,000 N. The initial sag is calculated at 4 meters based on a span of 200 meters and a normal temperature of 25℃. The wind offset coefficient is set so that for every 1 m / s increase in wind speed, the horizontal conductor offset increases by 0.1 meters. The ice offset coefficient is set so that for every 1 mm increase in ice thickness, the vertical conductor offset increases by 0.05 meters. The temperature offset coefficient is set so that for every 10℃ change in temperature, the sag changes by 0.3 meters. The chain-type elastic topology discretizes the conductor into several elastic elements, with each element having conductor points at both ends. The elements are connected by elastic forces, and the degree-of-freedom parameters of the conductor points within the element are matched with the node type to ensure that the conductor deformation conforms to the laws of elasticity.

[0035] Step S170: Based on geological structural parameters, the foundation bearing capacity, settlement coefficient, lateral displacement limit, surface friction resistance, and lithological sensitivity are written into the second engineering point set. The foundation points are then linked to topographic points, allowing the topographic slope to influence the tower foundation tilt direction and the geological attributes to influence the tower foundation settlement, thus obtaining the initial engineering point set. Specifically, a geological parameter standardization tool is used to convert data from the geological survey report of the target area into engineering point attribute parameters. Spatial correlation is used to establish the coordinate correspondence between foundation points and topographic points in the neutral topographic model, determining the freedom constraint adjustment rules for the foundation points, and binding the geological parameters to the node attributes.

[0036] For example, geological structural parameters include foundation bearing capacity of silty clay 180 kPa, settlement coefficient 0.001 mm / kN, lateral displacement limit 30 mm, surface friction 25000 N / m², granite sensitivity 0.1, and sandstone sensitivity 0.3. The foundation point and topographic point are precisely matched according to X / Y coordinates. The topographic slope is calculated based on the elevation difference between adjacent 10-meter topographic points; for example, if the elevation difference is 0.5 meters, the slope is 5%. The tower base tilts downwards along the slope. The degree of freedom parameters of the foundation point are adjusted according to the slope: when the slope is ≤5%, the rotational degree of freedom around the horizontal axis is 0.1; when the slope is >5%, the rotational degree of freedom is 0.2. The smaller the foundation bearing capacity in the geological properties, the greater the settlement of the tower foundation. For example, under a load of 500 kN, the settlement is 500 × 0.001 = 0.5 mm. The settlement is achieved by adjusting the Z coordinate of the foundation point. At the same time, parameters such as settlement coefficient and lateral displacement limit are written into the node properties to form an initial engineering point set that includes geological constraints and degree of freedom adjustment.

[0037] Step S180: Based on the initial engineering point set, construct an initial power distribution line twin model. Specifically, using the node modeling function of the digital twin modeling engine, all nodes in the initial engineering point set are spatially modeled according to topological connection relationships, stiffness matrices, degrees of freedom parameters, and geological properties. A visual model is generated through 3D rendering technology, integrating node constraint relationships and physical parameters to form a complete initial virtual model.

[0038] For example, the 3D structure of the tower's base, body, and head is generated sequentially according to the node coordinates. The conductor is generated as a curve model based on a chain-like elastic topology, and the foundation model is aligned with neutral terrain points. The model embeds the degree of freedom parameters, stiffness matrix data, and geological response parameters of each node, supporting subsequent attribute prediction and scene simulation. The 3D rendering preserves the blank base features of the neutral terrain model, i.e., no vegetation, no buildings, etc. The tower uses gray steel structure material, and the conductor uses silver metal material, ensuring a unified initial state and complete parameters for the model.

[0039] Step S200: Multi-source data is collected from the design and construction area of ​​the target power distribution project line to obtain a multi-source regional data set. A cross-attention mechanism is used to predict the attributes of the initial project point set based on the multi-source regional data set to obtain a project point attribute prediction result set.

[0040] Specifically, diverse data on the design and construction area are collected through aerial survey images, laser point clouds, GIS data, and geological exploration data, including topographic details, climate measurements, and geological surveys. Using a cross-attention mechanism, the actual attributes of each node in the initial engineering point set are predicted, such as stress, deformation, and degree of freedom adaptation values, providing data support for the adaptive configuration of the model.

[0041] In one possible implementation, multi-source data is collected from the design and construction area of ​​the target power distribution line to obtain a multi-source regional data set. A cross-attention mechanism is then used to predict the attributes of the initial engineering point set based on the multi-source regional data set, resulting in a set of engineering point attribute prediction results. Step S200 further includes step S210, extracting a first engineering point and its query features from the initial engineering point set. Specifically, any node is selected from the initial engineering point set as the first engineering point, and its core attributes are extracted using a feature encoding tool and converted into a standardized query feature vector.

[0042] For example, the first engineering point selects a node in the middle section of the tower. The extracted query features include: three-dimensional coordinates, node type, initial constraints, degree of freedom parameters, associated component parameters, and initial physical parameters. One-hot encoding is used to convert the node type and constraints into numerical features, which are then integrated with continuous features such as coordinates, degree of freedom parameters, and physical parameters to form a standardized query feature vector for multi-source data matching.

[0043] Step S220: Extract the coordinates of the first engineering point from the query features of the first engineering point, and match the multi-source region data set according to the coordinates of the first engineering point to obtain a first matched multi-source data set. Specifically, a coordinate parsing tool is used to extract the three-dimensional coordinates from the query feature vector of the first engineering point, and R-tree spatial indexing technology is used to retrieve data within a preset range around the coordinates in the multi-source region data set to ensure that the data accurately corresponds to the spatial location of the engineering point.

[0044] For example, the multi-source regional dataset includes a 1-meter precision digital elevation model, meteorological data from the past 5 years, borehole geological data, satellite imagery vegetation data, etc., all with spatial coordinate labels. By inputting the coordinates of the first engineering point, local data such as the terrain slope, average annual maximum wind speed, and measured values ​​of foundation bearing capacity at that location are retrieved and integrated into the first matching multi-source dataset, reflecting the actual environmental characteristics of the engineering point.

[0045] Step S230: Employing a cross-attention mechanism, attribute prediction is performed based on the first matching multi-source data set and the first engineering point query features to obtain the first engineering point attribute prediction result. This first engineering point attribute prediction result is then added to the engineering point attribute prediction result set. Specifically, a cross-attention prediction model based on the Transformer architecture is constructed. The first matching multi-source data set is encoded as keys and values, the first engineering point query features are used as queries, attention weights are calculated to fuse multi-source features, and the attribute prediction result is output through a fully connected network.

[0046] For example, slope and elevation difference features are extracted from topographic data, maximum wind speed and extreme temperature features from meteorological data, and measured bearing capacity and compression modulus features from geological data. These are then encoded into a unified dimension vector using a CNN. A cross-attention mechanism is used to calculate the similarity between the query vector and the key vector, highlighting the influence of key data, and a weighted sum is applied to obtain the fused features. The fully connected network outputs the actual stress value, deformation, and degree-of-freedom adaptation value of the first engineering point, storing these results in the engineering point attribute prediction result set according to node number.

[0047] In one possible implementation, a cross-attention mechanism is employed. Attribute prediction is performed based on the first matching multi-source data set and the first engineering point query features to obtain the first engineering point attribute prediction result. Step S230 further includes step S231, which involves retrieving the dynamic weight library of the cross-attention mechanism and performing dynamic weight matching in conjunction with the first engineering point query features to obtain the first dynamic weight. Specifically, the dynamic weight library is a parameter library trained using a large number of power distribution engineering samples. It contains the correspondence between engineering point features and attention weights. A cosine similarity algorithm is used to find the sample weights most similar to the first engineering point query features, and the first dynamic weight is obtained through weighted averaging and fine-tuning.

[0048] Step S232: Using the first matching multi-source data set and the first engineering point query feature as input, cross-attention attribute prediction is performed in conjunction with the first dynamic weight to obtain the first engineering point attribute prediction result. Specifically, the vectors obtained after feature encoding of the first matching multi-source data set are used as the key vector and value vector of the cross-attention mechanism, respectively. The first engineering point query feature is used as the query vector. The key vectors corresponding to each type of data are weighted and adjusted according to the first dynamic weight, such as multiplying the geological data key vector by the geological weight and the meteorological data key vector by the meteorological weight. The attention score of the query vector and the weighted key vector is calculated by dot product operation. The value vector is weighted and summed according to the score to obtain the fusion feature. The fusion feature is input into a preset neural network model to output the specific attribute prediction value of the first engineering point, including geometric offset, force offset, degree of freedom adaptation value, etc.

[0049] For example, the first dynamic weights are: topography 0.25, meteorology 0.3, geology 0.35, vegetation 0.05, and buildings 0.05. During cross-attention calculation, the key vectors corresponding to topography data are multiplied by 0.25, meteorological data by 0.3, geological data by 0.35, and vegetation and building data by 0.05 respectively, highlighting the influence of geological and meteorological data. The feature vectors are then fused and input into the neural network model to predict specific attribute results such as the actual sag and wind offset of a mid-span node of a conductor, the rotational degree of freedom adaptation value of a tower head node, and the settlement of a foundation node.

[0050] In one possible implementation, a dynamic weight library across attention mechanisms is retrieved, and dynamic weight matching is performed in conjunction with the query features of the first engineering point to obtain the first dynamic weight. Step S231 further includes step S2311, which uses the query features of the first engineering point as an index to calculate the similarity of the query features corresponding to each dynamic weight in the dynamic weight library, and selects the dynamic weights corresponding to the top m similarity values ​​as the matching dynamic weight set. Specifically, a cosine similarity algorithm is used to calculate the similarity between the query feature vector of the first engineering point and the query feature vectors of all samples in the dynamic weight library one by one. The similarity value ranges from 0 to 1, with the closer to 1 indicating greater similarity. All similarity values ​​are sorted in descending order using a quick sorting algorithm, and the dynamic weight vectors corresponding to the top m similarity values ​​are selected to form the matching dynamic weight set.

[0051] For example, the dynamic weight library stores 10,000 sets of sample data. Each set of samples contains an engineering point query feature vector and a corresponding dynamic weight vector. The cosine similarity between the first engineering point query feature vector and the feature vector of each set of samples is calculated, resulting in 10,000 similarity values. After quick sorting, the dynamic weight vectors corresponding to the top 20 similarity values ​​are selected.

[0052] Step S2312: Calculate the mean of the matching dynamic weight set to obtain the matching dynamic weight mean. Specifically, an arithmetic mean algorithm is used to calculate the average value for each weight dimension in the matching dynamic weight set, and the average values ​​of each dimension are combined to form the matching dynamic weight mean vector.

[0053] For example, the matching dynamic weight set contains 20 dynamic weight vectors, each with 5 dimensions. The sum of the terrain weights in each of the 20 vectors is calculated and divided by 20 to obtain the mean terrain weight. Similarly, the mean meteorological weight, mean geological weight, mean vegetation weight, and mean building weight are calculated sequentially. Assuming the calculated mean terrain weight is 0.24, the mean meteorological weight is 0.28, the mean geological weight is 0.36, the mean vegetation weight is 0.06, and the mean building weight is 0.06, then the matching dynamic weight mean vector is (0.24, 0.28, 0.36, 0.06, 0.06). This vector reflects the common characteristics of the weight distribution of similar engineering points.

[0054] Step S2313: Starting from the mean of the matching dynamic weights, perform mean-shift filtering on the set of matching dynamic weights to determine the first dynamic weight. Specifically, a mean-shift clustering algorithm is used, with the mean vector of the matching dynamic weights as the initial cluster center. The density gradient direction of all weight vectors in the set of matching dynamic weights is iteratively calculated, and the cluster center is moved towards the region with the highest density. This iteration is repeated until the change in the cluster center is less than a preset threshold. The cluster center at this point is the first dynamic weight.

[0055] For example, the initial cluster centers are matched with the dynamic weight mean vector (0.24, 0.28, 0.36, 0.06, 0.06). The Euclidean distance between each weight vector in the set and the initial center is calculated. Weight vectors with a distance less than 0.05 are selected to form a local subset, and the mean of this local subset is calculated as the new cluster centers. The Euclidean distances between all weight vectors and the new centers are calculated again, the local subsets are updated, and the mean is recalculated. This process continues until the change in each dimension of the cluster centers after two iterations is less than 0.001, finally yielding the first dynamic weight. This process filters out abnormal weight samples, making the first dynamic weight more consistent with the weight distribution pattern of most similar project points, thus improving the stability and accuracy of cross-attention attribute prediction.

[0056] Step S300: Based on the set of predicted results of the engineering point attributes, the initialization of the power distribution engineering line twin model is adaptively configured to obtain the basic power distribution engineering line twin model.

[0057] Specifically, based on the actual attributes of each engineering point in the set of engineering point attribute prediction results, such as geometric offset, stress offset, degree of freedom adaptation value, geological response, etc., the spatial location, stress parameters, constraint conditions, degree of freedom settings, etc. of the corresponding engineering points in the initialization of the power distribution engineering line twin model are adjusted in a targeted manner. Then, global optimization is used to ensure the overall mechanical balance, deformation coordination and design compliance of the model, forming a twin model that can reflect the foundation state in the actual engineering environment.

[0058] In one possible implementation, the initial power distribution line twin model is adaptively configured based on the set of predicted engineering point attributes to obtain a basic power distribution line twin model. Step S300 further includes step S310, which involves configuring the initial power distribution line twin model independently for each engineering point based on the geometric offset, stress offset, environmental sensitivity, geological response, and material state of each engineering point attribute prediction result in the set of predicted engineering point attributes. Specifically, a node attribute editing tool is used to read the prediction parameters of each engineering point in the set of predicted engineering point attributes one by one, and the spatial coordinates, stress parameters, environmental response thresholds, geological adaptation parameters, and material performance parameters of the corresponding engineering points in the initial power distribution line twin model are adjusted according to the parameter type.

[0059] For example, the predicted results for a certain tower body node are as follows: geometric offset +15 mm in the X direction, -8 mm in the Y direction, stress offset increased by 30,000 N in horizontal load, environmental sensitivity of a high-temperature sensitivity coefficient of 0.8, geological response of no additional settlement, and material condition of sufficient stress reserve. Using modeling tools, the X coordinate of this node is increased by 15 mm, the Y coordinate is decreased by 8 mm, the horizontal load parameter is updated to the initial value +30,000 N, and a high-temperature environmental stress monitoring threshold is added. Each engineering point is adjusted independently in this manner to ensure that the configuration of a single node is completely consistent with the predicted state under actual conditions.

[0060] Step S320: The global optimization configurator is invoked to perform global consistency optimization on the initialized power distribution engineering line twin model after the independent configuration of engineering points, thereby obtaining the basic power distribution engineering line twin model. Specifically, the global optimization configurator has a built-in deep learning-based neural network optimization model, adopting an encoder-decoder architecture. The optimization objectives are overall model mechanical balance, design specification compliance, node degree of freedom coordination, and engineering feasibility. The input parameters are the model parameters after independent configuration, including the coordinates, forces, stiffness, degrees of freedom, and material properties of each node. Through feature extraction, parameter mapping, and iterative optimization of the neural network, the globally optimal model configuration parameters are output, ensuring that the local configuration of each node is coordinated and unified with the overall system.

[0061] For example, the encoder of the neural network model uses a 3-layer convolutional neural network to extract the global mechanical features and design compliance features of the model, converting the features into 256-dimensional vectors; the decoder uses a 3-layer fully connected network with ReLU activation function and linear activation for the output layer, mapping the global parameters that need to be adjusted. During the optimization process, the loss function is set as follows: all tower stress ≤170 MPa, conductor sag 3-5 meters, foundation settlement ≤3 mm, no degree of freedom conflict, and safe distance meets the standard. The model is iteratively trained using the Adam optimizer. Each iteration outputs a set of optimization parameters and updates the twin model until the loss function value is lower than the preset threshold, ultimately obtaining the twin model of the basic power distribution engineering line with the best overall performance.

[0062] Step S400: Obtain the design and construction scenario set, perform scenario simulation on the twin model of the basic power distribution engineering line based on the design and construction scenario set, and perform design and construction management based on the scenario simulation results.

[0063] Specifically, by identifying and screening typical scenarios covering the entire design and construction process, the structural response, performance, and potential risks under different scenarios are simulated on the twin model of basic power distribution engineering lines. Based on the simulation results, design schemes are optimized, construction plans are formulated, and management strategies are adjusted to ensure the safety, reliability, and economy of engineering design and construction.

[0064] In one possible implementation, obtaining a set of design and construction scenarios, step S400 further includes step S410, which uses the basic information of the design and construction area of ​​the target power distribution line as an index to perform big data mining of design and construction scenarios to obtain a set of mined design and construction scenarios. Specifically, the Apriori association rule mining algorithm is used, with the basic information of the design and construction area of ​​the target power distribution line, including terrain type, climate zone, geological conditions, project scale, voltage level, and surrounding environment, as the retrieval index, to extract historical project scenarios, industry standard scenarios, potential risk scenarios, and key construction scenarios with similar characteristics from the power distribution project scenario database. The power distribution engineering scenario database contains massive amounts of structured scenario data, covering natural environment scenarios, construction process scenarios, geological change scenarios, and design optimization scenarios. Natural environment scenarios include rainstorms, strong winds, icing, high temperatures, earthquakes, and mudslides; construction process scenarios include equipment hoisting, conductor erection, foundation pouring, and emergency repair; geological change scenarios include uneven foundation settlement, landslides, and groundwater level changes; and design optimization scenarios include tower selection adjustments, conductor model replacement, and foundation type changes.

[0065] For example, using mountainous terrain, subtropical monsoon climate, granite geology, 110 kV power distribution line, 5 km in length, and surrounding woodland as indexes, the excavated scenarios include: a rainstorm scenario with 200 mm of daily precipitation, a gale-force wind scenario, a scenario with 10 mm of ice on the conductor, a scenario with limited space for tower foundation construction, a scenario with a small-scale landslide, a scenario with insufficient safe distance between the conductor and the woodland, and a scenario with a narrow equipment hoisting site, etc., forming a set of excavation design and construction scenarios. Each scenario includes specific scenario parameters, impact range, evaluation criteria, and possible risk consequences.

[0066] Step S420: Aggregate the excavation design and construction scenario set to obtain multiple aggregated excavation design and construction scenario sets. Specifically, a hierarchical clustering algorithm is used, with scenario type, influencing factors, target objects, and risk level as clustering features, to classify and aggregate all scenarios in the excavation design and construction scenario set, grouping scenarios with similar features into the same category to form multiple aggregated excavation design and construction scenario sets.

[0067] For example, clustering features are defined as: scene type, influencing factors, target audience, and risk level. A similarity matrix is ​​constructed by calculating the Euclidean distance between scenes, and hierarchical clustering is performed according to the similarity from high to low, ultimately forming multiple core aggregate sets. Each aggregate set contains all scenes with similar features under that category, which are used for subsequent representative screening and efficient simulation.

[0068] Step S430: Traverse the multiple aggregated mining design and construction scenario sets to perform representative screening within each set, and add the screening results to the design and construction scenario set. Specifically, a scene similarity aggregation algorithm is used to calculate pairwise similarity for all scenes within each aggregated mining design and construction scenario set, constructing a scene similarity matrix. By calculating the average similarity between each scene and all other scenes in the set, the scene with the highest overall similarity is selected as the representative scene of that set, and the representative scenes of all aggregated sets are added to the design and construction scenario set.

[0069] For example, a certain aggregate contains four scenarios: rainstorm, strong wind, icing, and high temperature. Each scenario's feature vector includes dimensions such as occurrence probability, impact range, intensity of effect, and risk type. The cosine similarity algorithm is used to calculate the similarity between scenarios. For instance, the rainstorm scenario has a similarity of 0.6 with the strong wind scenario, 0.7 with the icing scenario, and 0.3 with the high temperature scenario, with an overall average similarity of (0.6 + 0.7 + 0.3) / 3 = 0.53. The strong wind scenario has similarities of 0.6, 0.5, and 0.4 with other scenarios, with an overall average similarity of 0.5. The icing scenario has similarities of 0.7, 0.5, and 0.3 with other scenarios, with an overall average similarity of 0.5. The high temperature scenario has similarities of 0.3, 0.4, and 0.3 with other scenarios, with an overall average similarity of 0.33. The rainstorm scenario with the highest overall similarity is selected as the representative scenario of this aggregate. Similarly, other aggregate sets are filtered using the same logic, and finally all representative scenarios are combined to form a design and construction scenario set, ensuring that the representative scenarios in each set can reflect the core characteristics of all scenarios under that category to the greatest extent.

[0070] This application employs digital twin technology, integrating explicit prior knowledge such as tower structure, conductor topology, and geological structure to construct a power distribution engineering specification space. It initializes the engineering point set and initial twin model, collects multi-source data from the design and construction area, predicts the attributes of the initialized engineering point set through a cross-attention mechanism, adaptively configures the initial twin model based on the attribute prediction results to obtain a basic twin model, acquires a set of design and construction scenarios, conducts scenario simulations on the basic twin model, and uses simulation results to manage the design and construction of power distribution lines. This addresses the technical problems of existing power distribution line design and construction management, such as the disconnect between design schemes and dynamic environments, insufficient utilization of multi-source data, lack of full lifecycle simulation, and insufficient risk foresight. It achieves the technical effects of dynamic adaptive design parameters, intelligent fusion of multi-source data, full lifecycle multi-scenario simulation, and proactive risk prevention and control during construction and operation.

[0071] In the above text, refer to Figure 1 This paper describes in detail a power distribution line design and construction method supporting multi-scenario simulation according to embodiments of the present invention. Next, reference will be made to... Figure 2 This invention describes a power distribution line design and construction platform that supports multi-scenario simulation according to embodiments of the present invention.

[0072] The power distribution line design and construction platform supporting multi-scenario simulation according to embodiments of the present invention addresses the technical problems existing in the design and construction management of power distribution lines, such as the disconnect between design schemes and dynamic environments, insufficient utilization of multi-source data, lack of full lifecycle simulation, and insufficient risk foresight. It achieves the technical effects of dynamic adaptive design parameters, intelligent fusion of multi-source data, full lifecycle multi-scenario simulation, and proactive prevention and control of construction and operation risks. The power distribution line design and construction platform supporting multi-scenario simulation includes: a power distribution engineering specification space construction module 10, an attribute prediction module 20, an adaptive configuration module 30, and a scenario simulation module 40.

[0073] The power distribution engineering specification space construction module 10 is used to construct the power distribution engineering specification space in conjunction with digital twins. It calls the tower structure template, conductor topology constraints, and geological structure as explicit structural priors to initialize the engineering point set of the power distribution engineering specification space, thereby obtaining the initialized engineering point set and the initialized power distribution engineering line twin model. The attribute prediction module 20 is used to collect multi-source data on the design and construction area of ​​the target power distribution engineering line to obtain a multi-source regional data set. It uses a cross-attention mechanism to predict the attributes of the initialized engineering point set based on the multi-source regional data set, thereby obtaining a set of engineering point attribute prediction results. The adaptive configuration module 30 is used to adaptively configure the initialized power distribution engineering line twin model based on the set of engineering point attribute prediction results, thereby obtaining a basic power distribution engineering line twin model. The scenario simulation module 40 is used to obtain a set of design and construction scenarios, perform scenario simulation on the basic power distribution engineering line twin model based on the set of design and construction scenarios, and perform design and construction management based on the scenario simulation results.

[0074] The detailed description of the specific configuration of the power distribution engineering specification space construction module 10 is as follows: As mentioned above, in conjunction with digital twin construction of the power distribution engineering specification space, the power distribution engineering specification space construction module 10 may further include: a basic data acquisition unit for acquiring standard tower models, conductor standard parameters, foundation standard models, and equipment standard models; a neutral terrain model matching unit for acquiring basic information of the design and construction area of ​​the target power distribution engineering line, and matching a neutral terrain model based on the basic information; a power distribution engineering specification space generation unit for importing the standard tower models, conductor standard parameters, foundation standard models, and equipment standard models into the neutral terrain model in conjunction with digital twin, adding unified descriptive coordinates and unified attribute structures to obtain the power distribution engineering specification space to be commissioned; and an environment simulation unit for matching a neutral environment state based on the basic information, performing environmental simulation on the power distribution engineering specification space to be commissioned, and constructing the power distribution engineering specification space.

[0075] The process involves using tower structure templates, conductor topology constraints, and geological structures as explicit structural priors to initialize the engineering point set in the power distribution engineering specification space, thereby obtaining an initialized engineering point set and an initialized twin model of the power distribution engineering line. The power distribution engineering specification space construction module 10 may further include: a first engineering point set generation unit used to map the topological connection relationships of tower components to the engineering point set based on the tower structure template, thus obtaining the first engineering point set; and a second engineering point set generation unit used to bind key conductor points in the first engineering point set as suspension points, tension points, or segmentation points based on conductor topology constraints, and write them into the initial tension... Force parameters, sag parameters, and wind, ice, and temperature offset coefficients are used to establish a chain-like elastic topology of the conductor points, thus obtaining a second engineering point set. The initialization engineering point set generation unit is used to write the foundation bearing capacity, settlement coefficient, lateral displacement limit, surface friction resistance, and lithological sensitivity into the second engineering point set according to the geological structure parameters, and to establish a connection between the foundation points and the terrain points, so that the terrain slope affects the tower foundation tilt direction and the geological properties affect the tower foundation settlement, thus obtaining an initialization engineering point set. The initialization power distribution line twin model construction unit is used to construct an initialization power distribution line twin model based on the initialization engineering point set.

[0076] The first engineering point set generation unit may further include: setting the tower foot points in the engineering point set as fixed constraint points, setting the tower body nodes as rigid connection points, setting the tower head nodes as swingable nodes, and writing the stiffness matrix, degrees of freedom and connection adjacency relationships according to the tower structure template to obtain the first engineering point set.

[0077] The attribute prediction module 20 is described in detail below: As mentioned above, multi-source data is collected from the design and construction area of ​​the target power distribution line to obtain a multi-source regional data set. A cross-attention mechanism is used to predict the attributes of the initial engineering point set based on the multi-source regional data set to obtain an engineering point attribute prediction result set. The attribute prediction module 20 may further include: an engineering point extraction unit for extracting a first engineering point and a first engineering point query feature from the initial engineering point set; a multi-source regional data set matching unit for extracting the coordinates of the first engineering point from the first engineering point query feature, and matching the multi-source regional data set based on the first engineering point coordinates to obtain a first matched multi-source data set; and an attribute prediction unit for using a cross-attention mechanism to perform attribute prediction based on the first matched multi-source data set and the first engineering point query feature to obtain a first engineering point attribute prediction result, and adding the first engineering point attribute prediction result to the engineering point attribute prediction result set.

[0078] Specifically, a cross-attention mechanism is employed to perform attribute prediction based on the first matching multi-source data set and the first engineering point query features to obtain the first engineering point attribute prediction result. The attribute prediction unit may further include: a dynamic weight matching subunit for retrieving the dynamic weight library of the cross-attention mechanism, performing dynamic weight matching in combination with the first engineering point query features to obtain the first dynamic weight; and a cross-attention attribute prediction subunit for using the first matching multi-source data set and the first engineering point query features as input, combining the first dynamic weight to perform cross-attention attribute prediction to obtain the first engineering point attribute prediction result.

[0079] Specifically, the dynamic weight library across attention mechanisms is retrieved, and dynamic weight matching is performed in conjunction with the query features of the first engineering point to obtain the first dynamic weight. The dynamic weight matching subunit may further include: a similarity calculation component used to calculate the similarity of the query features corresponding to each dynamic weight in the dynamic weight library using the query features of the first engineering point as an index, and selecting the dynamic weights corresponding to the similarity values ​​in the top m positions as the matching dynamic weight set; a mean calculation component used to calculate the mean of the matching dynamic weight set to obtain the mean of the matching dynamic weights; and a mean drift filtering component used to perform mean drift filtering on the matching dynamic weight set starting from the mean of the matching dynamic weights to determine the first dynamic weight.

[0080] The adaptive configuration module 30 is described in detail below: As mentioned above, the initial power distribution line twin model is adaptively configured based on the set of engineering point attribute prediction results to obtain the basic power distribution line twin model. The adaptive configuration module 30 may further include: an engineering point independent configuration unit for configuring the initial power distribution line twin model independently based on the geometric offset, stress offset, environmental sensitivity, geological response, and material state of each engineering point attribute prediction result in the set of engineering point attribute prediction results; and a global consistency optimization unit for calling the global optimization configurator to perform global consistency optimization on the initial power distribution line twin model after the engineering point independent configuration is completed to obtain the basic power distribution line twin model.

[0081] The specific configuration of the scenario simulation module 40 is described in detail below: As mentioned above, to obtain the design and construction scenario set, the scenario simulation module 40 may further include: a design and construction scenario big data mining unit used to perform design and construction scenario big data mining using the basic information of the design and construction area of ​​the target power distribution line as an index, to obtain a set of mined design and construction scenarios; a similar aggregation unit used to perform similar aggregation on the mined design and construction scenario set to obtain multiple aggregated mined design and construction scenario sets; and a representative filtering unit within the set used to traverse the multiple aggregated mined design and construction scenario sets to perform representative filtering within the set, and add the filtering results to the design and construction scenario set.

[0082] The power distribution line design and construction platform supporting multi-scenario simulation provided in the embodiments of the present invention can execute the power distribution line design and construction method supporting multi-scenario simulation provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0083] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0084] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A power distribution engineering line design construction method supporting multi-scenario simulation, characterized in that, The method includes: By combining digital twins to construct a power distribution engineering specification space, and calling tower structure templates, conductor topology constraints and geological structures as explicit structural priors, the engineering point set of the power distribution engineering specification space is initialized to obtain the initialized engineering point set and the initialized power distribution engineering line twin model; Multi-source data is collected from the design and construction area of ​​the target power distribution project line to obtain a multi-source regional data set. A cross-attention mechanism is used to predict the attributes of the initial project point set based on the multi-source regional data set to obtain a set of project point attribute prediction results. Based on the set of predicted project point attributes, the initialization of the power distribution line twin model is adaptively configured to obtain the basic power distribution line twin model. Obtain a set of design and construction scenarios, perform scenario simulation on the twin model of the basic power distribution engineering line based on the set of design and construction scenarios, and carry out design and construction management based on the scenario simulation results; Specifically, the initial power distribution line twin model is adaptively configured based on the predicted result set of the engineering point attributes to obtain the basic power distribution line twin model, including: Based on the geometric offset, stress offset, environmental sensitivity, geological response, and material state of each engineering point attribute prediction result in the engineering point attribute prediction result set, the engineering points of the initial power distribution line twin model are configured independently. The global optimization configurator is invoked to perform global consistency optimization on the initialization power distribution engineering line twin model after the independent configuration of the engineering points is completed, so as to obtain the basic power distribution engineering line twin model. This includes obtaining a set of design and construction scenarios, including: Using the basic information of the design and construction area of ​​the target power distribution project line as an index, big data mining of design and construction scenarios is carried out to obtain a set of design and construction scenarios. The set of excavation design and construction scenarios is aggregated to obtain multiple aggregated sets of excavation design and construction scenarios. The multiple aggregated mining design and construction scenario sets are traversed to perform representative filtering within the sets, and the filtered results are added to the design and construction scenario set.

2. The electric power engineering line design and construction method supporting multi-scenario simulation according to claim 1, characterized in that, The construction of a power distribution engineering specification space using digital twins includes: Obtain standard tower models, standard conductor parameters, standard foundation models, and standard equipment models; Obtain basic information about the design and construction area of ​​the target power distribution line, and match a neutral terrain model based on the basic information; By combining digital twins, the standard tower model, conductor standard parameters, foundation standard model, and equipment standard model are imported into the neutral terrain model, and unified descriptive coordinates and unified attribute structure are added to obtain the specification space of the power distribution project to be commissioned. Based on the aforementioned basic information and a neutral environmental state, an environmental simulation is performed on the specification space of the power distribution project to be debugged, thereby constructing the specification space of the power distribution project.

3. The power distribution line design and construction method supporting multi-scenario simulation as described in claim 1, characterized in that, By invoking the tower structure template, conductor topology constraints, and geological structure as explicit structural priors, the engineering point set in the power distribution engineering specification space is initialized, resulting in the initialized engineering point set and the initialized power distribution engineering line twin model, including: Based on the tower structure template, the topological connection relationship of the tower components is mapped to the engineering point set to obtain the first engineering point set; Based on the topological constraints of the conductor, the key points of the conductor in the first engineering point set are bound as suspension points, tension points or segmentation points. The initial tension parameters, sag parameters and wind, ice and temperature offset coefficients are written in, and a chain elastic topological structure of the conductor points is established to obtain the second engineering point set. Based on the geological structure parameters, the foundation bearing capacity, settlement coefficient, lateral displacement limit, surface friction and lithological sensitivity are written into the second engineering point set, and the foundation points are linked with the topographic points so that the topographic slope affects the tower foundation tilt direction and the geological properties affect the tower foundation settlement, thus obtaining the initial engineering point set. Based on the initialization project point set, an initialization power distribution line twin model is constructed.

4. The electric power engineering line design and construction method supporting multi-scenario simulation according to claim 3, characterized in that, Based on the tower structure template, the tower foot points in the engineering point set are set as fixed constraint points, the tower body nodes are set as rigid connection points, and the tower head nodes are set as swingable nodes. The stiffness matrix, degrees of freedom, and connection adjacency relationships are written to obtain the first engineering point set.

5. The electric power engineering line design and construction method supporting multi-scenario simulation according to claim 1, characterized in that, Multi-source data is collected from the design and construction area of ​​the target power distribution project line to obtain a multi-source regional data set. A cross-attention mechanism is then used to predict the attributes of the initialized project point set based on the multi-source regional data set, resulting in a project point attribute prediction result set, including: Extract the first engineering point and the first engineering point query features from the initial engineering point set; The coordinates of the first engineering point are extracted from the query features of the first engineering point, and the multi-source regional data set is matched based on the coordinates of the first engineering point to obtain the first matched multi-source data set; A cross-attention mechanism is adopted to perform attribute prediction based on the first matching multi-source data set and the first engineering point query features to obtain the first engineering point attribute prediction result, and the first engineering point attribute prediction result is added to the engineering point attribute prediction result set.

6. The electric power engineering line design construction method supporting multi-scenario simulation according to claim 5, wherein, A cross-attention mechanism is employed to perform attribute prediction based on the first matching multi-source dataset and the first engineering point query features, obtaining the first engineering point attribute prediction result, including: Retrieve the dynamic weight library across attention mechanisms, combine it with the query features of the first engineering point to perform dynamic weight matching, and obtain the first dynamic weight; Using the first matching multi-source data set and the first engineering point query features as input, cross-attention attribute prediction is performed in combination with the first dynamic weight to obtain the first engineering point attribute prediction result.

7. The electric power engineering line design construction method supporting multi-scenario simulation according to claim 6, characterized in that, Retrieve the dynamic weight library across attention mechanisms, combine it with the query features of the first engineering point to perform dynamic weight matching, and obtain the first dynamic weight, including: Using the query features of the first engineering point as an index, the similarity of the query features corresponding to each dynamic weight in the dynamic weight library is calculated, and the dynamic weights corresponding to the similarity values ​​in the top m positions are selected as the matching dynamic weight set. Calculate the mean of the matching dynamic weight set to obtain the mean of the matching dynamic weights; Starting from the mean of the matching dynamic weights, the set of matching dynamic weights is filtered by mean drift to determine the first dynamic weight.

8. A power distribution engineering line design and construction platform that supports multi-scenario simulation, characterized in that, The platform is used to implement the power distribution engineering line design and construction method supporting multi-scenario simulation as described in any one of claims 1-7, and the platform includes: The power distribution engineering specification space construction module is used to construct the power distribution engineering specification space in conjunction with digital twins. It calls the tower structure template, conductor topology constraints and geological structure as explicit structural priors to initialize the engineering point set of the power distribution engineering specification space, and obtain the initialized engineering point set and the initialized power distribution engineering line twin model. The attribute prediction module is used to collect multi-source data on the design and construction area of ​​the target power distribution project line to obtain a multi-source regional data set. It then uses a cross-attention mechanism to predict the attributes of the initial engineering point set based on the multi-source regional data set to obtain a set of engineering point attribute prediction results. An adaptive configuration module is used to adaptively configure the initial power distribution engineering line twin model based on the set of engineering point attribute prediction results to obtain a basic power distribution engineering line twin model. The scenario simulation module is used to acquire a set of design and construction scenarios, perform scenario simulation on the twin model of the basic power distribution engineering line based on the set of design and construction scenarios, and perform design and construction management based on the scenario simulation results.

Images