A long-span steel structure index prediction method based on revit

Abstract

The application relates to the technical field of data processing, and discloses a large-span steel structure index prediction method based on Revit, which comprises the following steps: analyzing a target steel structure three-dimensional building information model, and extracting geometric entities and material attribute information; obtaining a key stress component set through structure topology analysis; performing characteristic engineering reconstruction on the components in the key stress component set to obtain cross-section characteristic parameters, connection node types and spatial positioning information; constructing a component relationship topology graph according to the connection node types and the spatial positioning information; parameterizing and packaging the component relationship topology graph, the cross-section characteristic parameters and the material attribute information to obtain a structured characteristic spectrum; correlating and mining the structured characteristic spectrum with a historical engineering case characteristic spectrum to obtain a performance index prediction conclusion, and mapping the conclusion to the corresponding components of the model and visually marking the conclusion; and the application can improve the efficiency of large-span steel structure index prediction based on Revit.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method for predicting the indicators of large-span steel structures based on Revit. Background Technology

[0002] Due to the complexity of the structural system and the diversity of stress relationships, the prediction of indicators for large-span steel structures requires the integration of multi-dimensional engineering information. Traditional prediction methods often rely on manual decomposition and analysis of building information models. The information extraction process is cumbersome and easily affected by subjective factors, making it difficult to fully obtain the geometric, material, and connection characteristics of key components, resulting in excessively long data processing cycles in the early stages.

[0003] Existing technologies lack standardized feature encapsulation and case association mechanisms, making it impossible to efficiently and accurately match target structural features with historical engineering data. This results in insufficient scientific rigor and accuracy in prediction conclusions, and the prediction results are difficult to intuitively map into 3D models, which is detrimental to subsequent engineering decisions. Therefore, improving the efficiency of predicting indicators for large-span steel structures has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a Revit-based method for predicting the performance indicators of large-span steel structures, in order to solve the problems mentioned in the background section.

[0005] To achieve the above objectives, this invention provides a method for predicting the indices of large-span steel structures based on Revit, comprising: S1. Perform feature tree parsing and extraction on the three-dimensional building information model of the target large-span steel structure to obtain the geometric entity information and material attribute information of the target large-span steel structure; S2. Perform structural topology analysis on the geometric entity information to obtain the set of key load-bearing components of the target large-span steel structure; S3. Perform characteristic engineering reconstruction on the components of the key load-bearing component concentration to obtain the cross-sectional characteristic parameters, connection node types and spatial positioning information of the components; S4. Construct a component relationship topology diagram of the target large-span steel structure based on the connection node type and the spatial positioning information; S5. The component relationship topology diagram, the cross-sectional feature parameters and the material attribute information are parametrically encapsulated to obtain the structured feature spectrum of the target large-span steel structure; S6. Correlate the structured feature spectrum with the feature spectrum of historical engineering cases to obtain the performance index prediction conclusion of the target large-span steel structure. S7. Map the performance index prediction conclusions to the corresponding key load-bearing components in the three-dimensional building information model, and perform visual annotation in the three-dimensional building information model.

[0006] In a preferred embodiment, the step of performing feature tree parsing and extraction on the three-dimensional building information model of the target large-span steel structure to obtain the geometric entity information and material attribute information of the target large-span steel structure includes: Traverse the parametric feature tree in the three-dimensional building information model of the target large-span steel structure, and filter out the feature nodes in the parametric feature tree whose feature type is marked as structural component; The feature nodes are instantiated and expanded to obtain the specific geometric entity of the target large-span steel structure; Boundary representation extraction is performed on the specific geometric entities to obtain the geometric entity information of the target large-span steel structure; In the three-dimensional building information model, access the material attribute set bound to the feature node, and extract the material type and physical attribute parameters defined in the material attribute set to obtain the material attribute information of the target large-span steel structure.

[0007] In a preferred embodiment, the step of performing structural topology analysis on the geometric entity information to obtain the set of key load-bearing components of the target large-span steel structure includes: The geometric entity information is separated into component elements to obtain the component element set of the target large-span steel structure; The spatial contact relationships between the component elements in the component element set are derived by performing an association matrix to obtain the component connection matrix of the target large-span steel structure; Mechanical force transmission path deduction is performed on the network connectivity described by the component connection matrix to obtain the main force transmission path set of the target large-span steel structure; The topological importance index of the component element is calculated based on the frequency and position of the component element in the main force transmission path set. Based on the topological importance index, the set of component elements is evaluated and selected to obtain the set of key load-bearing components of the target large-span steel structure.

[0008] In a preferred embodiment, the formula for calculating the topological importance index is as follows: ; In the formula, For component elements Topological importance index, This represents the total number of force transmission paths in the set of main force transmission paths. This is a function for determining path inclusion. The preset shape adjustment coefficient, For component elements In the path The sequential position number in For the path Total length, It is an exponential function.

[0009] In a preferred embodiment, the step of performing characteristic engineering reconstruction on the components of the key load-bearing component concentration to obtain the cross-sectional characteristic parameters, connection node types, and spatial positioning information of the components includes: A cross-sectional profile geometric analysis is performed on the components of the key load-bearing component concentration to obtain the cross-sectional profile data of the components; Based on the typical cross-sectional type of the component, the cross-sectional profile data is subjected to typical geometric fitting to obtain the cross-sectional characteristic parameters of the component; In the three-dimensional building information model, the contact interface of the connection relationship between the component and adjacent components is extracted to obtain the connection geometric features of the component; Based on a pre-stored connection rule knowledge base, the connection geometric features are matched and determined to obtain the connection node type of the component. In the global coordinate system of the three-dimensional building information model, the spatial positioning information of the component is obtained by integrating the positioning base point coordinates, direction vector and geometric length of the component in a spatial state.

[0010] In a preferred embodiment, constructing a component relationship topology diagram of the target large-span steel structure based on the connection node type and the spatial positioning information includes: The components in the key load-bearing component set are used as topological nodes, and corresponding spatial coordinate descriptions are configured for the topological nodes based on the spatial positioning information; When the connection node type indicates that there is a direct physical connection between the topology nodes, a topology edge is established between the topology nodes, and the connection node type is marked as the connection attribute of the topology edge. Based on the spatial coordinate description, the relative orientation relationship between the two topological nodes connected by the topological edge is derived; The topological nodes, spatial coordinate descriptions, topological edges, connection attributes, and relative orientation relationships are integrated into a component relationship topology diagram of the target large-span steel structure.

[0011] In a preferred embodiment, the step of parametrically encapsulating the component relationship topology diagram, the cross-sectional feature parameters, and the material property information to obtain the structural feature spectrum of the target large-span steel structure includes: The topological structure of the component relationship topology graph is deconstructed to obtain the graph structure data of the component relationship topology graph; Based on the globally unique identifier in the graph structure data, the cross-sectional feature parameters and the material attribute information are respectively bound to the corresponding nodes in the graph structure data to obtain the node attribute data of the target large-span steel structure; The graph structure data and the node attribute data are heterogeneously fused to obtain the structured feature spectrum of the target large-span steel structure.

[0012] In a preferred embodiment, the step of correlation mining between the structured feature spectrum and the feature spectrum of historical engineering cases to obtain the performance index prediction conclusion of the target large-span steel structure includes: The structured feature spectrum is mapped to the feature spectrum library of historical engineering cases to obtain a subset of candidate cases for the target large-span steel structure; The structured feature spectrum is compared with the candidate case subset using multidimensional features to obtain the similarity comparison result of the target large-span steel structure; The historical case with the highest similarity comparison result in the candidate case subset is taken as the best matching case; Extract the archived performance metrics associated with the best-matching case from the historical engineering case feature spectrum library; Based on the archived performance indicators, the performance indicator prediction conclusion of the target large-span steel structure is obtained by performing engineering logic reasoning on the feature differences between the structured feature spectrum and the best matching case.

[0013] In a preferred embodiment, the step of using the archived performance indicators as a benchmark to perform engineering logic reasoning on the feature differences between the structured feature spectrum and the best-matching case to obtain the performance indicator prediction conclusion of the target large-span steel structure includes: Multidimensional difference feature extraction is performed on the structured feature spectrum and the feature spectrum of the best matching case to obtain the structured difference descriptor of the target large-span steel structure; The structured difference descriptor is subjected to rule-driven derivation to obtain the correction factor for the archived performance metric; Based on the correction factor, the archived performance indicators are corrected and integrated to obtain the predicted performance indicators of the target large-span steel structure.

[0014] In a preferred embodiment, mapping the performance index prediction conclusion to the corresponding key load-bearing components in the three-dimensional building information model and performing visual annotation in the three-dimensional building information model includes: The logical correspondence between the performance index prediction conclusions and the set of key load-bearing components is indexed and organized to obtain the mapping relationship table of the target large-span steel structure; Based on the mapping table, the three-dimensional building information model is traversed to locate the key load-bearing components that match the unique identifier in the component; The performance index prediction conclusions of the key load-bearing components are visually encoded and converted to obtain the visualization rendering parameters of the three-dimensional building information model. Based on the visualization rendering parameters, the graphics rendering engine of the three-dimensional building information model is driven to update the visual state of the key load-bearing components.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention extracts feature trees and performs structural topology analysis on the 3D building information model of a target large-span steel structure to accurately obtain geometric entities, material attribute information, and key load-bearing component sets. Then, through feature-based engineering reconstruction and component relationship topology graph construction, it achieves systematic integration of multi-dimensional data. By parametrically encapsulating and forming a structured feature spectrum, it achieves standardized collection of various engineering information, providing comprehensive and reliable data support for performance index prediction, and effectively improving the basic accuracy and data processing efficiency of prediction work.

[0016] 2. This invention achieves accurate derivation of performance indicators by associating and mining structured feature spectra with feature spectra from historical engineering cases, combined with engineering logic reasoning based on feature differences and optimization using correction factors. Simultaneously, the predicted conclusions are mapped to a 3D model and visualized, allowing for intuitive presentation of the prediction results. This facilitates rapid understanding of the performance status of key load-bearing components, providing clear guidance for engineering design and decision-making, and further enhancing the practicality and application value of predicting indicators for large-span steel structures. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a Revit-based method for predicting the performance indicators of large-span steel structures, as provided in an embodiment of the present invention. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides a Revit-based method for predicting the performance indicators of large-span steel structures. The execution entity of this Revit-based method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the Revit-based method for predicting the performance indicators of large-span steel structures can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0020] Reference Figure 1 The diagram shown is a flowchart illustrating a Revit-based method for predicting the performance indicators of large-span steel structures, according to an embodiment of the present invention. In this embodiment, the Revit-based method for predicting the performance indicators of large-span steel structures includes: S1. Perform feature tree parsing and extraction on the three-dimensional building information model of the target large-span steel structure to obtain the geometric entity information and material attribute information of the target large-span steel structure; In this embodiment of the invention, the step of performing feature tree parsing and extraction on the three-dimensional architectural information model of the target large-span steel structure to obtain the geometric entity information and material attribute information of the target large-span steel structure includes: Traverse the parametric feature tree in the three-dimensional building information model of the target large-span steel structure, and filter out the feature nodes in the parametric feature tree whose feature type is marked as structural component; The feature nodes are instantiated and expanded to obtain the specific geometric entity of the target large-span steel structure; Boundary representation extraction is performed on the specific geometric entities to obtain the geometric entity information of the target large-span steel structure; In the three-dimensional building information model, access the material attribute set bound to the feature node, and extract the material type and physical attribute parameters defined in the material attribute set to obtain the material attribute information of the target large-span steel structure.

[0021] In the parametric feature tree of a 3D Building Information Model (BIM), the feature tree starts from the root node and forms a clear hierarchical structure from top to bottom. The root node is divided into functional nodes, including structure, building, and equipment, etc. Each functional node is further subdivided into multiple levels of child nodes. Each node has a fixed feature type marker field, which is one item from a preset enumeration set of values. The nodes are accessed sequentially in the order of "root node → first-level functional node → second-level child node → ... → terminal node." For each node accessed, the specific content of its feature type marker field is read and compared with the preset string "structural component." Strict case sensitivity and no extra spaces or character differences are considered. Only nodes whose marker content is completely consistent with "structural component" are retained; those that do not match are directly excluded. The final set of feature nodes contains basic information such as its unique identifier, its hierarchical path in the feature tree, and its feature type marker, facilitating accurate retrieval in subsequent steps.

[0022] For each selected feature node, abstract structural description information is retrieved from its dedicated data storage area using its unique identifier. This information includes the basic shape definition of the component, the clear geometric shape (e.g., beams correspond to cuboids, columns to cylinders), component association rules, connection position constraints between the main body and end connectors, assembly sequence requirements, and specific dimensional parameters such as length, width, height, and radius. When reconstructing the actual spatial form, the overall outline of the component is first constructed based on the basic shape definition and dimensional parameters. Then, according to the component association rules, each component, main body, and connector is spliced ​​and integrated according to preset position constraints, clarifying the relative positional relationships of each part. The central axis of the end connector coincides with the central axis of the main body and is located at a specified distance from both ends of the main body. This ultimately forms a three-dimensional geometric shape corresponding to each feature node. This shape has a clearly identifiable outline, components, and spatial position—that is, the specific geometric entity of the target large-span steel structure. Each geometric entity is bound to a unique identifier corresponding to the feature node.

[0023] When analyzing the surface features of each specific geometric entity, a fixed sequence is followed: "vertex-by-vertex → edge-by-edge → outer surface-by-outer surface → connection-by-connection". First, all vertices of the entity are identified, and the X, Y, and Z coordinates of each vertex in the global coordinate system of the 3D Building Information Model are recorded to ensure accurate correspondence between the coordinate data and the entity's spatial location. Next, each edge is identified, and the two endpoints of each edge, i.e., the recorded vertices, are determined. The straight-line distance between the two points is calculated as the edge's length, and the difference between the two coordinates determines the edge's extension direction. Edges from coordinates 0,0,0 to 1000,0,0 extend in the positive X-axis direction. Then, each outer surface is identified, recording the shape and contour of each face (rectangle, circle, regular polygon, etc.), identifying which edges enclose each face, and simultaneously determining the connections between each edge. The normal vector direction of each face is defined, with the normal vector direction of the front face of the cuboid being the positive X-axis direction. Finally, the various connection relationships are sorted out, clarifying the adjacency of each face with other faces. The front face of the cuboid is adjacent to the top, bottom, left, and right faces, and is relative to the back face. The connection of each edge with other edges is also defined. The two ends of the vertical edges of the cuboid are connected to the top and bottom horizontal edges respectively. The relationship of each vertex with other vertices is also defined. A vertex of the cuboid is connected to three edges in different directions at the same time. All recorded vertex coordinates, edge information, surface information, and connection relationships are systematically integrated into a unified format to form complete and standardized geometric entity information.

[0024] In the storage module of the 3D Building Information Model, each feature node's attribute field stores a unique association identifier. This identifier forms a one-to-one binding relationship with the material attribute set in the material attribute library. The material attribute library is a pre-set dedicated database for the model, containing standard material attribute sets corresponding to various components. During search and location, the unique association identifier in the feature node's attribute field is read first, and then a precise search is performed in the material attribute library based on this identifier to determine the corresponding material attribute set. After opening the material attribute set, the specific name marked in the "Material Type" field is read, such as standard material names like Q355 low-alloy high-strength steel and 304 stainless steel. At the same time, all pre-set required items and corresponding values ​​under the "Physical Property Parameters" column are extracted, including density in kg / m³, tensile strength in MPa, compressive strength in MPa, elastic modulus in GPa, Poisson's ratio, etc., ensuring that no attribute item is omitted. The material type, all physical property parameters, and unique identifier of each feature node are associated and organized. After being categorized and summarized by node, the material property information of the target large-span steel structure is formed. This information is accurately matched with the geometric entity information through the unique identifier of the node.

[0025] The beneficial effects of this invention are that, through clear feature tree traversal rules, standardized node matching standards, and a refined information extraction process, it ensures accurate selection of feature nodes for structural components, accurate restoration of specific geometric entities, and comprehensive and unbiased extraction of geometric entity information and material attribute information. The two types of information form a complete data association through the unique identifier of the node, providing accurate, standardized, and traceable basic data for subsequent steps such as structural topology analysis and identification of key load-bearing components. This further ensures the reproducibility of the entire large-span steel structure index prediction method, while improving the accuracy and reliability of the prediction results.

[0026] S2. Perform structural topology analysis on the geometric entity information to obtain the set of key load-bearing components of the target large-span steel structure; In this embodiment of the invention, the step of performing structural topology analysis on the geometric entity information to obtain the set of key load-bearing components of the target large-span steel structure includes: The geometric entity information is separated into component elements to obtain the component element set of the target large-span steel structure; The spatial contact relationships between the component elements in the component element set are derived by performing an association matrix to obtain the component connection matrix of the target large-span steel structure; Mechanical force transmission path deduction is performed on the network connectivity described by the component connection matrix to obtain the main force transmission path set of the target large-span steel structure; The topological importance index of the component element is calculated based on the frequency and position of the component element in the main force transmission path set. Based on the topological importance index, the set of component elements is evaluated and selected to obtain the set of key load-bearing components of the target large-span steel structure.

[0027] The formula for calculating the topology importance index is as follows: ; In the formula, For component elements Topological importance index, This represents the total number of force transmission paths in the set of main force transmission paths. This is a function for determining path inclusion. The preset shape adjustment coefficient, For component elements In the path The sequential position number in For the path Total length, It is an exponential function.

[0028] When separating component elements from geometric entity information, the independent closed boundaries and spatial non-overlapping characteristics of geometric entities are used as the judgment criteria. Each independent closed geometric unit with complete load-bearing function is identified one by one. By detecting whether there is spatial overlap or shared boundary between units, units without shared boundaries and spatially independent are determined as individual component elements. At the same time, combined with the functional attributes of the components, geometric units that can independently bear loads or transmit forces are collected and organized. Each unit is assigned a unique identifier and corresponding geometric feature information, and finally, a set of component elements of the target large-span steel structure is formed.

[0029] When deriving the association matrix for the spatial contact relationships between component elements in the component element set, the component element set is first numbered and sorted. The unique identifier of each component element is used as the row and column indices of the association matrix, and the matrix dimension is consistent with the number of elements in the component element set. By detecting whether there is an overlapping area on the outer surfaces of any two component elements, a threshold of 1 square centimeter is set. When the overlapping area of ​​the outer surfaces of two component elements is greater than or equal to 1 square centimeter, it is determined that there is a spatial contact relationship between them, and 1 is filled in at the intersection of the corresponding row and column indices in the association matrix. When the overlapping area is less than 1 square centimeter, it is determined that there is no spatial contact relationship, and 0 is filled in at the corresponding position. This rule is followed to traverse all component element pairs to complete the filling of the entire matrix, thus obtaining the component connection matrix of the target large-span steel structure.

[0030] When performing force transmission path derivation on the network connectivity described by the component connection matrix, the starting point of the force transmission path is first defined as the component element corresponding to the load application point of the target large-span steel structure, and the ending point is the component element corresponding to the structural support. Starting from the starting component element, the adjacent component element is found according to the column index marked as 1 in the component connection matrix. This adjacent component element is taken as the next node of the path, and the operation of finding adjacent component elements is repeated from this node until the ending component element is reached, forming a complete force transmission path. All combinations of starting and ending points are traversed to generate all possible force transmission paths. Then, a filtering condition is set to retain only paths in which the number of component elements does not exceed 50% of the total number of component elements. Finally, the main force transmission path set of the target large-span steel structure is obtained, and each path contains a unique number and a unique identifier for all component elements on the path.

[0031] When calculating the topological importance index based on the frequency and location of component elements in the main force transmission path set, the total number of times each component element appears in all main force transmission paths is first counted as the basic frequency data. This frequency is reflected by the path inclusion judgment function; the function takes a value of 1 when a component element exists in a path, and a value of 0 when it does not exist. Then, for each path containing each component element, its sequential position within the path is determined. The sequential position of the path's starting point is 1, increasing sequentially, and the sequential position of the path's ending point is the total number of component elements in that path. This sequential position number, along with the total path length, serves as the positional feature parameter. The shape adjustment coefficient is a fixed value of 2.0 obtained from statistical fitting of 1000 historical engineering cases of large-span steel structures of different types and spans. The position deviation is smoothed and quantified by an exponential function. The frequency and the position weight after exponential operation are fused together. That is, the judgment function result of each path is multiplied by the exponential operation result, and then the product results of all paths are summed to obtain the topological importance index of the component element. This formula realizes the quantitative characterization of the criticality of the component element in the force transmission network by integrating various parameters. The index value is a specific numerical result.

[0032] When selecting key components based on topological importance indices, the average topological importance index of all components in the set is first calculated, and 1.2 times this average is set as the key assessment threshold. Each component element's topological importance index is compared to this threshold. If the topological importance index is greater than or equal to the threshold, it is identified as a key load-bearing component; if the index is less than the threshold, it is identified as a non-key load-bearing component. All elements identified as key load-bearing components are collected, and their unique identifiers, geometric features, and topological importance indices are retained to form the key load-bearing component set for the target large-span steel structure.

[0033] The beneficial effects of this invention are that it achieves accurate acquisition of component element sets, component connection matrices, main force transmission path sets, and key load-bearing component sets by using clear component element separation standards, quantified spatial contact judgment thresholds, standardized force transmission path screening rules, and a topological importance index calculation method that integrates historical engineering data to determine parameters. The formulas and processes are closely linked, and the quantitative parameters objectively reflect the criticality of component stress. The entire process is logically clear and reproducible, ensuring that the selection results of key load-bearing components are objective and accurate. This provides core data support for subsequent characteristic engineering reconstruction and performance index prediction, and ensures the scientificity and reliability of the overall prediction method.

[0034] S3. Perform characteristic engineering reconstruction on the components of the key load-bearing component concentration to obtain the cross-sectional characteristic parameters, connection node types and spatial positioning information of the components; In this embodiment of the invention, the step of performing characteristic engineering reconstruction on the components of the key load-bearing component concentration to obtain the cross-sectional characteristic parameters, connection node types, and spatial positioning information of the components includes: A cross-sectional profile geometric analysis is performed on the components of the key load-bearing component concentration to obtain the cross-sectional profile data of the components; Based on the typical cross-sectional type of the component, the cross-sectional profile data is subjected to typical geometric fitting to obtain the cross-sectional characteristic parameters of the component; In the three-dimensional building information model, the contact interface of the connection relationship between the component and adjacent components is extracted to obtain the connection geometric features of the component; Based on a pre-stored connection rule knowledge base, the connection geometric features are matched and determined to obtain the connection node type of the component. In the global coordinate system of the three-dimensional building information model, the spatial positioning information of the component is obtained by integrating the positioning base point coordinates, direction vector and geometric length of the component in a spatial state.

[0035] When performing cross-sectional profile geometric analysis on components with concentrated critical load-bearing members, the cross-sectional location is first determined. The middle cross-section perpendicular to the length of the component is selected as the standard analysis cross-section, and the distance between this cross-section and both ends of the component is not less than 1 / 4 of the total length of the component. A point-by-point scanning method is used to collect the coordinates of profile points along the cross-sectional edge. The spacing between adjacent profile points is set to 1 mm during scanning to ensure complete capture of the geometric shape of the cross-sectional edge. During the acquisition process, the global coordinate system of the 3D building information model is used as a reference to record the X, Y, and Z coordinate values ​​of each profile point. All collected coordinate values ​​are arranged in clockwise order according to the cross-sectional edge to form an ordered coordinate sequence, ultimately obtaining the cross-sectional profile data of the component.

[0036] When performing typical geometric fitting on cross-sectional profile data based on the typical cross-sectional types of structural members, it is pre-defined that the commonly used typical cross-sectional types for large-span steel structures include five types: rectangular, circular, I-shaped, channel, and angle steel. First, the cross-sectional profile data is initially compared with the standard profile shape of each typical cross-section. The percentage of overlap points between the profile data and each standard shape is calculated, and the typical cross-sectional type with the highest percentage of overlap points is selected as the fitting benchmark type. Using the geometric parameters of this benchmark type as initial values, the geometric parameters are adjusted to control the deviation between the fitted profile and the cross-sectional profile data within 0.1 mm. The deviation is calculated based on the sum of the squares of the coordinate differences of corresponding profile points. Adjustment stops when the deviation meets the requirements, and the geometric parameters are extracted: length and width for rectangular cross-sections; radius for circular cross-sections; flange thickness, flange width, web thickness, and web height for I-shaped cross-sections; and length and thickness of corresponding sides for channel and angle steel cross-sections. These parameters together constitute the cross-sectional characteristic parameters of the structural member.

[0037] When extracting contact interfaces between components in a 3D building information model (BIM), the process begins by filtering out adjacent components with a spatial distance of less than 5 millimeters based on the component's spatial location information. Then, each surface of the current component and each adjacent component is inspected face-by-face to determine if there are overlapping areas. A minimum area threshold of 2 square centimeters is set for overlapping areas; when an overlapping area greater than or equal to 2 square centimeters is detected, that area is identified as a contact interface. The geometric information of the contact interface is recorded, including its shape, area, position coordinates in the global coordinate system, and normal vector direction. This information is then integrated to obtain the connection geometric features of the component.

[0038] When performing rule matching on connection geometric features based on a pre-stored connection rule knowledge base, the knowledge base contains explicit rule entries. Each rule entry includes a specific combination of connection geometric features and a corresponding connection node type. Examples of rule entries are: a circular contact interface with an area between 5 and 10 square centimeters, and a normal vector perpendicular to the component's length direction, corresponds to a bolted connection node; a rectangular contact interface with an area greater than 20 square centimeters, and a normal vector parallel to the component's length direction, corresponds to a welded connection node. The extracted connection geometric features are compared one by one with the rule entries in the knowledge base. When a connection geometric feature completely matches all the feature conditions of a rule entry, the connection node type corresponding to that rule entry is determined as the connection node type of the current component, ensuring a unique and clear matching result.

[0039] When integrating the spatial positioning information of components in the global coordinate system of a 3D Building Information Model (BIM), the positioning base point of the component is first determined. The geometric center of the component is selected as the positioning base point, and the X, Y, and Z coordinates of the positioning base point are obtained by calculating the average of the coordinates of all vertices of the component. The direction vector of the component is determined by the coordinate difference between its two endpoints. Taking one endpoint as the starting point and the other endpoint as the ending point, the difference between the coordinates of the two points on the X, Y, and Z axes is calculated to form the direction vector. The geometric length of the component is calculated by the straight-line distance between its two endpoints in the global coordinate system. The distance between two points in space is calculated using the method for calculating the distance between two points, i.e., by taking the square root of the sum of the squares of the differences in the X, Y, and Z coordinates of the two endpoints. The positioning base point coordinates, direction vector, and geometric length are then linked and integrated in a unified format, with each data point accompanied by a global coordinate system identifier, ultimately yielding the spatial positioning information of the component.

[0040] The beneficial effects of this invention are that it achieves accurate acquisition of cross-sectional contour data, cross-sectional feature parameters, connection geometric features, connection node types and spatial positioning information by clearly defining the cross-sectional analysis location, scanning interval, fitting deviation threshold, contact interface judgment criteria, rule matching logic and spatial positioning parameter calculation method. The whole process is clear in steps, the judgment criteria are quantified and the operation is reproducible, ensuring that the extracted information is complete and meets the actual needs of engineering. It provides high-quality core data support for subsequent component relationship topology diagram construction, structured feature spectrum encapsulation and performance index prediction, effectively ensuring the scientificity and practicality of the overall prediction method.

[0041] S4. Construct a component relationship topology diagram of the target large-span steel structure based on the connection node type and the spatial positioning information; In this embodiment of the invention, constructing a component relationship topology diagram of the target large-span steel structure based on the connection node type and the spatial positioning information includes: The components in the key load-bearing component set are used as topological nodes, and corresponding spatial coordinate descriptions are configured for the topological nodes based on the spatial positioning information; When the connection node type indicates that there is a direct physical connection between the topology nodes, a topology edge is established between the topology nodes, and the connection node type is marked as the connection attribute of the topology edge. Based on the spatial coordinate description, the relative orientation relationship between the two topological nodes connected by the topological edge is derived; The topological nodes, spatial coordinate descriptions, topological edges, connection attributes, and relative orientation relationships are integrated into a component relationship topology diagram of the target large-span steel structure.

[0042] Each component in the set of critical load-bearing components is treated as an independent topological node, and a unique node identifier is assigned to each topological node. This identifier is consistent with the unique identifier of the component in the set of critical load-bearing components. Based on the spatial positioning information of the components, the X, Y, and Z coordinates of the positioning base point, the direction vector, and the geometric length of each component are extracted. These data are organized in a fixed order of "positioning base point coordinates - direction vector - geometric length". All data are based on the global coordinate system of the 3D building information model as a unified reference, forming a unique spatial coordinate description for each topological node, ensuring that the spatial state of each topological node can be completely represented by accurate data.

[0043] First, clearly define the specific categories of direct physical connections among the connection node types, including bolted connections, welded connections, and pin connections. These connections represent direct contact between two components without intermediate transition components. Verify the connection node types of any two topological nodes. When the connection node type belongs to one of the three direct physical connection categories, establish a topological edge between these two topological nodes, assigning a unique edge identifier to each edge. Fully label the corresponding connection node type on this topological edge as its connection attribute, ensuring that the connection attribute is completely consistent with the actual connection type of the components.

[0044] Obtain the coordinates of the positioning base points in the spatial coordinate descriptions of the two topological nodes connected by the topological edge, and calculate the coordinate differences between the two positioning base points on the X-axis, Y-axis, and Z-axis. That is, subtract the X-coordinate of the first node from the X-coordinate of the second node, subtract the Y-coordinate of the first node from the Y-coordinate of the second node, and subtract the Z-coordinate of the first node from the Z-coordinate of the second node, respectively, to obtain three sets of coordinate difference data. The orientation criteria are set as follows: when the absolute value of the difference between the X-axis coordinates is greater than the absolute value of the difference between the Y-axis and Z-axis coordinates, and the X-axis difference is positive, the relative orientation is "along the positive X-axis direction"; when the X-axis difference is negative, the relative orientation is "along the negative X-axis direction". Similarly, when the absolute value of the difference between the Y-axis coordinates is the largest, the orientation is determined as "along the positive Y-axis direction" or "along the negative Y-axis direction" based on the sign of the Y-axis difference. When the absolute value of the difference between the Z-axis coordinates is the largest, the orientation is determined as "along the positive Z-axis direction" or "along the negative Z-axis direction" based on the sign of the Z-axis difference. If the absolute values ​​of the differences between any two axes are equal and both are the largest, the relative orientation is determined according to the rule of "X-axis takes precedence over Y-axis, and Y-axis takes precedence over Z-axis". Finally, the relative orientation relationship between the two topological nodes is derived.

[0045] The system integrates node and edge lists. The node list sequentially records the unique identifier and corresponding spatial coordinates of each topological node, ensuring complete and traceable information for each node. The edge list records the unique identifier of each topological edge, the unique identifiers of the two topological nodes it connects to, the corresponding connection attributes, and their relative orientation, making the association information of each edge clear and unambiguous. The node and edge lists are organized in a unified data format to construct a complete topological structure containing nodes, edges, spatial coordinates, connection attributes, and relative orientations, forming a component relationship topology diagram of the target large-span steel structure. This topology diagram can intuitively and accurately reflect the connection relationships and spatial positions between all key load-bearing components.

[0046] The beneficial effects of this invention are that it achieves the accurate construction of component relationship topology diagrams through clear definition rules for topological nodes and edges, unified spatial coordinate benchmarks, quantified relative orientation derivation standards, and standardized integration formats. This ensures that the topology diagram can completely and accurately present the connection type, spatial position, and relative orientation of key load-bearing components, providing logically clear and data-standardized topological structure support for the subsequent parametric encapsulation of structured feature spectra, and further guaranteeing the systematicness and reliability of the large-span steel structure index prediction method.

[0047] S5. The component relationship topology diagram, the cross-sectional feature parameters and the material attribute information are parametrically encapsulated to obtain the structured feature spectrum of the target large-span steel structure; In this embodiment of the invention, the step of parametrically encapsulating the component relationship topology diagram, the cross-sectional feature parameters, and the material attribute information to obtain the structured feature spectrum of the target large-span steel structure includes: The topological structure of the component relationship topology graph is deconstructed to obtain the graph structure data of the component relationship topology graph; Based on the globally unique identifier in the graph structure data, the cross-sectional feature parameters and the material attribute information are respectively bound to the corresponding nodes in the graph structure data to obtain the node attribute data of the target large-span steel structure; The graph structure data and the node attribute data are heterogeneously fused to obtain the structured feature spectrum of the target large-span steel structure.

[0048] When deconstructing the topological structure of the component relationship topology graph, data is extracted in a fixed order of "nodes first, then edges." First, all relevant information about the topological nodes is extracted. Each topological node must include a spatial coordinate description consisting of a globally unique identifier, the coordinates of its positioning base point, its direction vector, and its geometric length. All topological nodes are arranged in ascending order of their globally unique identifiers to form a node list. Next, relevant information about the topological edges is extracted. Each topological edge must include its own unique identifier, the globally unique identifiers of the two topological nodes it connects to, its connection attributes, and its relative orientation relationship. These edges are arranged in ascending order of their unique identifiers to form an edge list. Finally, the node list and the edge list are combined in the order of "node list first, edge list second," and data verification identifiers are added to form standardized and traceable graph structure data.

[0049] When performing binding operations based on globally unique identifiers in graph structure data, the globally unique identifier of each topological node is first extracted from the node list of the graph structure data. This identifier is completely consistent with the unique identifier of the corresponding component in the key load-bearing component set. Then, based on this globally unique identifier, the cross-sectional feature parameters and material attribute information of the corresponding component obtained from the previous feature engineering reconstruction are retrieved. The cross-sectional feature parameters and material attribute information are associated one-to-one with the corresponding globally unique identifier in the order of "cross-sectional feature parameters first, material attribute information second". Each globally unique identifier is only bound to the parameters and information of its corresponding component, without overlap or omission, and finally, the node attribute data of each node is formed.

[0050] When performing heterogeneous data fusion of graph structure data and node attribute data, a globally unique identifier is used as the core link to establish a unified fusion rule: For each globally unique identifier, first associate the node information and edge information corresponding to the identifier in the graph structure data, and then associate the cross-sectional feature parameters and material attribute information corresponding to the identifier in the node attribute data; according to the fixed data structure of "globally unique identifier - node spatial coordinate description - associated edge information - cross-sectional feature parameters - material attribute information", all data corresponding to each identifier are integrated, and all integrated data are arranged in ascending order according to the globally unique identifier to form a dataset with unified structure, logical coherence, and complete data. This dataset is the structured feature spectrum of the target large-span steel structure.

[0051] The beneficial effects of this invention are that by deconstructing the topological structure in a fixed order, accurately binding based on globally unique identifiers, and fusing heterogeneous data with core identifiers as the link, it ensures the standardized extraction and effective integration of graph structure data and node attribute data. The resulting structured feature spectrum fully contains key information such as the topological relationships, geometric features, and material properties of components. The data logic is clear, the format is unified, and the traceability is strong. This provides a high-quality and standardized data foundation for subsequent correlation mining with feature spectra of historical engineering cases, and effectively guarantees the accuracy and efficiency of performance index prediction.

[0052] S6. Correlate the structured feature spectrum with the feature spectrum of historical engineering cases to obtain the performance index prediction conclusion of the target large-span steel structure. In this embodiment of the invention, the step of correlating and mining the structured feature spectrum with the feature spectrum of historical engineering cases to obtain the performance index prediction conclusion of the target large-span steel structure includes: The structured feature spectrum is mapped to the feature spectrum library of historical engineering cases to obtain a subset of candidate cases for the target large-span steel structure; The structured feature spectrum is compared with the candidate case subset using multidimensional features to obtain the similarity comparison result of the target large-span steel structure; The historical case with the highest similarity comparison result in the candidate case subset is taken as the best matching case; Extract the archived performance metrics associated with the best-matching case from the historical engineering case feature spectrum library; Based on the archived performance indicators, the performance indicator prediction conclusion of the target large-span steel structure is obtained by performing engineering logic reasoning on the feature differences between the structured feature spectrum and the best matching case.

[0053] The process of using the archived performance indicators as a benchmark, and performing engineering logic reasoning on the feature differences between the structured feature spectrum and the best-matching case to obtain the performance indicator prediction conclusions for the target large-span steel structure includes: Multidimensional difference feature extraction is performed on the structured feature spectrum and the feature spectrum of the best matching case to obtain the structured difference descriptor of the target large-span steel structure; The structured difference descriptor is subjected to rule-driven derivation to obtain the correction factor for the archived performance metric; Based on the correction factor, the archived performance indicators are corrected and integrated to obtain the predicted performance indicators of the target large-span steel structure.

[0054] When mapping the structured feature spectrum to the historical engineering case feature spectrum library, the storage structure of the library is first defined. Each historical case in the library contains a complete structured feature spectrum, including a component relationship topology diagram, cross-sectional feature parameters, material attribute information, and a unique case identifier. Mapping filtering conditions are set: the number of nodes in the component relationship topology diagrams of the target structured feature spectrum and the historical case feature spectrum differs by no more than 20%; materials with a main material type accounting for more than 70% are completely identical; and the cross-sectional type matching rate for key load-bearing components accounting for more than 80% of the total number of key load-bearing components is no less than 80%. All cases in the historical engineering case feature spectrum library are traversed, and each case is checked to see if it meets the above filtering conditions. All historical cases that meet the conditions are organized and summarized according to their case identifiers to form a subset of candidate cases for the target large-span steel structure.

[0055] When performing multi-dimensional feature comparison between the structured feature spectrum and the candidate case subset, the comparison dimensions are clearly defined as topological structure dimension, cross-sectional feature dimension, and material attribute dimension, with the topological structure dimension accounting for 40%, the cross-sectional feature dimension for 30%, and the material attribute dimension for 30%. The topological structure dimension compares the matching of the number of nodes, the number of edges, and connection attributes. A perfect match in the number of nodes and edges earns full marks; a difference within 10% earns 80 marks; and a difference between 10% and 20% earns 60 marks. A perfect match in connection attributes earns full marks; and a matching rate of at least 90% earns 80 marks. The cross-sectional feature dimension compares key cross-sectional parameters such as length, width, and radius. A perfect match in these parameters earns full marks; a difference within 5% earns 80 marks; and a difference between 5% and 10% earns 60 marks. The material attribute dimension compares major physical parameters such as density and tensile strength. A perfect match in these parameters earns full marks; a difference within 3% earns 80 marks; and a difference between 3% and 5% earns 60 marks. Calculate the score of each candidate case in each dimension, and sum them according to the weights to obtain the total similarity score. The total similarity scores of all candidate cases together constitute the similarity comparison result of the target large-span steel structure.

[0056] When selecting the historical case with the highest similarity comparison result from the candidate case subset as the best matching case, the total similarity score of all candidate cases is first extracted to determine the highest score value. If only one candidate case has a score equal to the highest value, then that case is directly determined as the best matching case; if multiple candidate cases have scores equal to the highest value, then the similarity of the force transmission path distribution in the secondary feature dimension is further compared, and the overlap rate of the force transmission paths of these cases with the target structured feature spectrum is calculated. The case with the highest overlap rate is selected as the best matching case, ensuring that the best matching case is unique and accurate.

[0057] When extracting archived performance indicators associated with the best-matching case from the historical engineering case feature spectrum library, a one-to-one association is established between the unique case identifier of each historical case and the archived performance indicator. The archived performance indicators include preset core indicators such as load-bearing capacity, stiffness coefficient, stability safety factor, and fatigue life. Based on the unique case identifier of the best-matching case, a precise search is performed in the associated data table of the historical engineering case feature spectrum library to retrieve the specific values ​​of all core indicators corresponding to that identifier, ensuring that no indicators are omitted and the values ​​are accurate, thus obtaining the archived performance indicators associated with the best-matching case.

[0058] When performing multi-dimensional difference feature extraction between the structured feature spectrum and the feature spectrum of the best-matching case, topology, cross-sectional features, and material properties are the core difference extraction dimensions. The specific data of the two feature spectra in each dimension are compared one by one. The topology dimension records the differences in the number of nodes, the differences in the number of edges, and the specific locations and types of mismatched connection attributes. The cross-sectional feature dimension records the specific difference values ​​of each key cross-sectional parameter, the target feature spectrum parameter minus the best-matching case feature spectrum parameter, and the unique identifier of the corresponding cross-section. The material property dimension records the specific differences in the physical parameters of each major material and the material type identifier. All the difference information from all dimensions is organized in a fixed order of "topology difference - cross-sectional feature difference - material property difference," forming a standardized dataset containing difference types, specific values, and associated identifiers, thus obtaining the structured difference descriptor for the target large-span steel structure.

[0059] When using rule-driven derivation to obtain correction factors for archived performance indicators from structured difference descriptors, a pre-stored rule-driven library contains corresponding rules for various difference features and performance indicator correction coefficients. These rules are based on difference-performance correlation data from 500 sets of historical large-span steel structure projects. For example, a 10% increase in the number of nodes corresponds to a bearing capacity correction coefficient of 1.05, a 5% increase in cross-sectional width corresponds to a stiffness coefficient correction coefficient of 1.03, and a 3% increase in density corresponds to a stability safety coefficient correction coefficient of 1.02. Each difference feature in the structured difference descriptor is matched with a rule in the rule-driven library to extract the corresponding correction coefficient. Then, weights are assigned based on the importance of the difference features: topological difference weight 50%, cross-sectional feature difference weight 30%, and material attribute difference weight 20%. Each correction coefficient is multiplied by its corresponding weight and summed to obtain the final correction factor for the archived performance indicators.

[0060] When integrating archived performance indicators based on correction factors, each core indicator value in the archived performance indicators is multiplied by the correction factor to obtain a preliminary correction value for each indicator. If the same indicator is affected by multiple differential characteristics, i.e., there are multiple corresponding correction coefficients, the multiple correction coefficients are first integrated according to their weights before being multiplied by the indicator value. All preliminary correction values ​​are subjected to consistency verification to ensure that the corrected indicator values ​​conform to the engineering mechanics laws of large-span steel structures, such as the load-bearing capacity not being lower than the industry minimum standard value. After the verification is passed, all corrected indicator values ​​are arranged in the order of "load-bearing capacity - stiffness coefficient - stability safety factor - fatigue life" to obtain the performance indicator prediction conclusion of the target large-span steel structure.

[0061] The beneficial effects of this invention are that it achieves standardized operation of the entire process from candidate case screening to performance index prediction through clear mapping screening conditions, multi-dimensional weighted comparison rules, unique best case selection method, and precise difference extraction and rule-based correction process. Each step has specific judgment criteria and execution methods, and the screening threshold, weight allocation, rule source and verification standards are clearly defined to ensure that the prediction conclusions are based on historical engineering data and fit the actual characteristics of the target structure, thereby improving the accuracy and reliability of performance index prediction and providing a scientific and reproducible decision-making basis for the engineering design and safety assessment of large-span steel structures.

[0062] S7. Map the performance index prediction conclusions to the corresponding key load-bearing components in the three-dimensional building information model, and perform visual annotation in the three-dimensional building information model.

[0063] In this embodiment of the invention, mapping the performance index prediction conclusion to the corresponding key load-bearing components in the three-dimensional building information model and performing visual annotation in the three-dimensional building information model includes: The logical correspondence between the performance index prediction conclusions and the set of key load-bearing components is indexed and organized to obtain the mapping relationship table of the target large-span steel structure; Based on the mapping table, the three-dimensional building information model is traversed to locate the key load-bearing components that match the unique identifier in the component; The performance index prediction conclusions of the key load-bearing components are visually encoded and converted to obtain the visualization rendering parameters of the three-dimensional building information model. Based on the visualization rendering parameters, the graphics rendering engine of the three-dimensional building information model is driven to update the visual state of the key load-bearing components.

[0064] When organizing the logical correspondence between performance index predictions and key load-bearing component sets using an index, the unique identifier for each component in the key load-bearing component set is first extracted. This identifier is consistent with the unique identifiers of components from previous steps. Next, the performance index predictions for each component are extracted, including the specific values ​​of all core indicators such as load-bearing capacity, stiffness coefficient, stability safety factor, and fatigue life. Following a fixed column order of "unique identifier - predicted load-bearing capacity - predicted stiffness coefficient - predicted stability safety factor - predicted fatigue life," the identifier and corresponding indicator values ​​for each component are entered into a table one by one. Each row of data in the table corresponds to the complete association information of a single key load-bearing component. A data validation column is also added to indicate the validation results of each association, ultimately forming a mapping table for the target large-span steel structure.

[0065] When traversing the 3D Building Information Model based on the mapping table, the traversal follows a fixed hierarchical path of "root node - structural component node - critical load-bearing component child node". For each component visited, its unique identifier is read from its attribute field. This unique identifier is then matched against all data in the "unique identifier" column of the mapping table. When the unique identifier of a component in the model completely matches the identifier in a row of the mapping table, the component is identified as the corresponding critical load-bearing component. The hierarchical path and positioning coordinates of this component in the model are recorded to ensure accurate positioning of each critical load-bearing component without omissions or misjudgments.

[0066] When visually encoding the predicted performance indicators of key load-bearing components, fixed visual encoding rules are pre-defined. These rules are based on the visual recognition requirements of large-span steel structure engineering design. The encoding rules for load-bearing capacity indicators are as follows: when the predicted value is ≥ 1.2 times the design standard value, the corresponding RGB value is 255,0,0 with 30% transparency; when the predicted value is between 0.9 and 1.2 times the design standard value, the corresponding RGB value is 255,255,0 with 50% transparency; when the predicted value is < 0.9 times the design standard value, the corresponding RGB value is 0,0,255 with 70% transparency. The encoding rules for stiffness coefficient indicators are as follows: when the predicted value is ≥ 1.1 times the design standard value, a solid border is added; when the predicted value is between 0.8 and 1.1 times the design standard value, a dashed border is added; when the predicted value is < 0.8 times the design standard value, a dotted-dash border is added. The encoding rules for stability safety factors and fatigue life indicators follow the same logic, corresponding to different texture fill styles. Based on the predicted performance indicators of key load-bearing components, the corresponding colors, transparency, border styles, and texture fill styles are matched, and these visual parameters are integrated to obtain the visualization rendering parameters of the 3D building information model.

[0067] When a graphics rendering engine for a 3D building information model is driven by visual rendering parameters, it first binds the visual rendering parameters to the corresponding components using the recorded hierarchical paths and positioning coordinates of key load-bearing components within the model. Upon receiving the binding command, the rendering engine replaces the visual attributes of the components one by one in the order of "color attribute update - transparency attribute update - border attribute update - texture attribute update". Color attributes are directly assigned the corresponding RGB values; transparency attributes are adjusted by percentage values ​​to increase the component's transparency; border attributes generate outlines according to the set line type and line width; and texture attributes fill the component surface with a specified style. After the attribute updates are complete, the rendering engine performs a screen refresh operation, ensuring that the visual state of the key load-bearing components is presented in real-time according to the rendering parameters, thus completing the visual state update.

[0068] The beneficial effects of this invention are that, through fixed-format indexed organization, precise identifier matching and positioning, standardized visual coding rules, and an orderly rendering-driven process, it achieves accurate correlation and visualization of performance index prediction conclusions and key load-bearing components in the 3D model. Each step has clear execution standards and judgment criteria, ensuring that the visualization annotations are clear, accurate, and in line with engineering cognitive habits. This allows relevant personnel to quickly and intuitively grasp the performance status of key components, providing convenient and efficient support for engineering decision-making and safety assessment, and further improving the practicality and operability of the prediction method for large-span steel structures.

[0069] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0070] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for predicting the performance indicators of large-span steel structures based on Revit, characterized in that, The method includes: S1. Perform feature tree parsing and extraction on the three-dimensional building information model of the target large-span steel structure to obtain the geometric entity information and material attribute information of the target large-span steel structure; S2. Perform structural topology analysis on the geometric entity information to obtain the set of key load-bearing components of the target large-span steel structure; S3. Perform characteristic engineering reconstruction on the components of the key load-bearing component concentration to obtain the cross-sectional characteristic parameters, connection node types and spatial positioning information of the components; S4. Construct a component relationship topology diagram of the target large-span steel structure based on the connection node type and the spatial positioning information; S5. The component relationship topology diagram, the cross-sectional feature parameters and the material attribute information are parametrically encapsulated to obtain the structured feature spectrum of the target large-span steel structure; S6. Correlate the structured feature spectrum with the feature spectrum of historical engineering cases to obtain the performance index prediction conclusion of the target large-span steel structure. S7. Map the performance index prediction conclusions to the corresponding key load-bearing components in the three-dimensional building information model, and perform visual annotation in the three-dimensional building information model.

2. The method for predicting the performance indicators of large-span steel structures based on Revit as described in claim 1, characterized in that, The feature tree parsing and extraction of the three-dimensional building information model of the target large-span steel structure yields the geometric entity information and material attribute information of the target large-span steel structure, including: Traverse the parametric feature tree in the three-dimensional building information model of the target large-span steel structure, and filter out the feature nodes in the parametric feature tree whose feature type is marked as structural component; The feature nodes are instantiated and expanded to obtain the specific geometric entity of the target large-span steel structure; Boundary representation extraction is performed on the specific geometric entities to obtain the geometric entity information of the target large-span steel structure; In the three-dimensional building information model, access the material attribute set bound to the feature node, and extract the material type and physical attribute parameters defined in the material attribute set to obtain the material attribute information of the target large-span steel structure.

3. The method for predicting the performance indicators of large-span steel structures based on Revit as described in claim 1, characterized in that, The structural topology analysis of the geometric entity information yields the set of key load-bearing components of the target large-span steel structure, including: The geometric entity information is separated into component elements to obtain the component element set of the target large-span steel structure; The spatial contact relationships between the component elements in the component element set are derived by performing an association matrix to obtain the component connection matrix of the target large-span steel structure; Mechanical force transmission path deduction is performed on the network connectivity described by the component connection matrix to obtain the main force transmission path set of the target large-span steel structure; The topological importance index of the component element is calculated based on the frequency and position of the component element in the main force transmission path set. Based on the topological importance index, the set of component elements is evaluated and selected to obtain the set of key load-bearing components of the target large-span steel structure.

4. The method for predicting the performance indicators of large-span steel structures based on Revit as described in claim 3, characterized in that, The formula for calculating the topology importance index is as follows: ； In the formula, For component elements Topological importance index, This represents the total number of force transmission paths in the set of main force transmission paths. This is a function for determining path inclusion. The preset shape adjustment coefficient, For component elements In the path The sequential position number in For the path Total length, It is an exponential function.

5. The method for predicting the performance indicators of large-span steel structures based on Revit as described in claim 1, characterized in that, The process of performing characteristic engineering reconstruction on the components of the key load-bearing component concentration to obtain the cross-sectional characteristic parameters, connection node types, and spatial positioning information of the components includes: A cross-sectional profile geometric analysis is performed on the components of the key load-bearing component concentration to obtain the cross-sectional profile data of the components; Based on the typical cross-sectional type of the component, the cross-sectional profile data is subjected to typical geometric fitting to obtain the cross-sectional characteristic parameters of the component; In the three-dimensional building information model, the contact interface of the connection relationship between the component and adjacent components is extracted to obtain the connection geometric features of the component; Based on a pre-stored connection rule knowledge base, the connection geometric features are matched and determined to obtain the connection node type of the component. In the global coordinate system of the three-dimensional building information model, the spatial positioning information of the component is obtained by integrating the positioning base point coordinates, direction vector and geometric length of the component in a spatial state.

6. The method for predicting the performance indicators of large-span steel structures based on Revit as described in claim 1, characterized in that, The step of constructing a component relationship topology diagram of the target large-span steel structure based on the connection node type and the spatial positioning information includes: The components in the key load-bearing component set are used as topological nodes, and corresponding spatial coordinate descriptions are configured for the topological nodes based on the spatial positioning information; When the connection node type indicates that there is a direct physical connection between the topology nodes, a topology edge is established between the topology nodes, and the connection node type is marked as the connection attribute of the topology edge. Based on the spatial coordinate description, the relative orientation relationship between the two topological nodes connected by the topological edge is derived; The topological nodes, spatial coordinate descriptions, topological edges, connection attributes, and relative orientation relationships are integrated into a component relationship topology diagram of the target large-span steel structure.

7. The method for predicting the performance indicators of large-span steel structures based on Revit as described in claim 1, characterized in that, The step of parametrically encapsulating the component relationship topology diagram, the cross-sectional feature parameters, and the material attribute information to obtain the structural feature spectrum of the target large-span steel structure includes: The topological structure of the component relationship topology graph is deconstructed to obtain the graph structure data of the component relationship topology graph; Based on the globally unique identifier in the graph structure data, the cross-sectional feature parameters and the material attribute information are respectively bound to the corresponding nodes in the graph structure data to obtain the node attribute data of the target large-span steel structure; The graph structure data and the node attribute data are heterogeneously fused to obtain the structured feature spectrum of the target large-span steel structure.

8. The method for predicting the performance indicators of large-span steel structures based on Revit as described in claim 1, characterized in that, The step of correlating and mining the structured feature spectrum with the feature spectrum of historical engineering cases to obtain the performance index prediction conclusions of the target large-span steel structure includes: The structured feature spectrum is mapped to the feature spectrum library of historical engineering cases to obtain a subset of candidate cases for the target large-span steel structure; The structured feature spectrum is compared with the candidate case subset using multidimensional features to obtain the similarity comparison result of the target large-span steel structure; The historical case with the highest similarity comparison result in the candidate case subset is taken as the best matching case; Extract the archived performance metrics associated with the best-matching case from the historical engineering case feature spectrum library; Based on the archived performance indicators, the performance indicator prediction conclusion of the target large-span steel structure is obtained by performing engineering logic reasoning on the feature differences between the structured feature spectrum and the best matching case.

9. The method for predicting the performance indicators of large-span steel structures based on Revit as described in claim 8, characterized in that, The process of using the archived performance indicators as a benchmark, and performing engineering logic reasoning on the feature differences between the structured feature spectrum and the best-matching case to obtain the performance indicator prediction conclusions for the target large-span steel structure includes: Multidimensional difference feature extraction is performed on the structured feature spectrum and the feature spectrum of the best matching case to obtain the structured difference descriptor of the target large-span steel structure; The structured difference descriptor is subjected to rule-driven derivation to obtain the correction factor for the archived performance metric; Based on the correction factor, the archived performance indicators are corrected and integrated to obtain the predicted performance indicators of the target large-span steel structure.

10. The Revit-based long-span steel structure index prediction method of claim 1, wherein, The step of mapping the performance index prediction conclusions to the corresponding key load-bearing components in the three-dimensional building information model and performing visual annotation in the three-dimensional building information model includes: The logical correspondence between the performance index prediction conclusions and the set of key load-bearing components is indexed and organized to obtain the mapping relationship table of the target large-span steel structure; Based on the mapping table, the three-dimensional building information model is traversed to locate the key load-bearing components that match the unique identifier in the component; The performance index prediction conclusions of the key load-bearing components are visually encoded and converted to obtain the visualization rendering parameters of the three-dimensional building information model. Based on the visualization rendering parameters, the graphics rendering engine of the three-dimensional building information model is driven to update the visual state of the key load-bearing components.

Images