A method for generating an irregular hyperboloid prefabricated wall model

Abstract

This invention discloses a method for generating a prefabricated wall model with an irregular hyperboloid surface, belonging to the field of building information modeling (BIM) technology. The method includes acquiring the original design surface data of the irregular hyperboloid building wall, conducting a prefabrication assessment on the data, identifying prefabricable areas and special node areas and generating assessment results, inputting the results into an improved surface segmentation algorithm to complete modular segmentation, obtaining a set of prefabricated unit surfaces containing planar, single-curved, and hyperboloid panels. Each unit surface is parametrically unfolded to obtain two-dimensional planar contour data, based on which a three-dimensional geometric component model is created and connection construction information is added, forming a three-dimensional model set of prefabricated units. After virtual pre-assembly detection of interference and gap generation error reports, the component dimensions and connection point positions are iteratively adjusted, and the integrated and adjusted model generates the final prefabricated wall overall BIM model. This invention optimizes the modeling process for irregular hyperboloid surfaces, improving the rationality of prefabricated unit segmentation and the model assembly accuracy.

Description

Technical Field

[0001] This invention belongs to the field of building information modeling technology, specifically a method for generating irregular hyperboloid prefabricated wall models. Background Technology

[0002] Conventional prefabricated wall model construction often relies directly on the original design surface data for modeling. The surface processing stage does not incorporate specific analysis based on the assembly construction requirements. Surface segmentation often adopts general geometric segmentation methods. Parametric unfolding and 3D component modeling are both directly executed based on the original surface. Error adjustment after virtual pre-assembly often relies on manual experience. The overall modeling process is not combined with the characteristics of prefabricated construction.

[0003] This type of technology has obvious limitations in practical applications. The original curved surface does not distinguish the assembly adaptation area, special node areas are confused with regular areas, the surface segmentation results cannot match the actual assembly conditions, the segmented unit surface shapes are messy, there is no unified basis for the division of planar, single-curved and hyperbolic panels, the contour regularity after parametric unfolding of unit surfaces is insufficient, the connection structure layout of the three-dimensional component model lacks specificity, a large number of interference and gap problems are prone to occur during virtual pre-assembly, there is no clear reference basis for adjusting component size and connection points, the adjustment process is cumbersome and the effect is not good.

[0004] The original design surface data has not undergone assemblability assessment, making it impossible to accurately identify assemblable areas and special node areas on the surface. Surface segmentation cannot be optimized based on assembly features. Conventional surface segmentation algorithms are difficult to adapt to the assembly segmentation requirements of irregular hyperboloids. The assembly unit surface after modular segmentation cannot meet the dual requirements of processing and installation. Summary of the Invention

[0005] This invention aims to solve at least one of the technical problems existing in the prior art; Therefore, this invention proposes a method for generating an irregular hyperboloid prefabricated wall model, comprising: Obtain the original design surface data of the irregular hyperboloid building wall to be processed; Assemblability evaluation is performed on the original design surface data to identify assemblable areas and special node areas on the surface, and assembly feature evaluation results are generated. The assembly feature evaluation results are input into the improved surface segmentation algorithm to perform modular segmentation of the original design surface, generating a set of assembly unit surfaces including planar, single-curved, and hyperbolic panels; Each unit surface in the assembly unit surface set is parametrically expanded to obtain the two-dimensional planar contour data corresponding to each assembly unit. Based on the two-dimensional planar contour data, a detailed three-dimensional geometric component model is created for each assembly unit, and connection construction information is added to form a complete three-dimensional model set of assembly units. The complete 3D model set of the assembly unit is virtually pre-assembled to check for interference and gaps, and an assembly error detection report is generated. Based on the assembly error detection report, the dimensions and connection point positions of the relevant component models in the 3D model set of the assembly unit are iteratively adjusted. All the adjusted component models are integrated according to their spatial relationships to generate the final overall BIM model of the irregular hyperboloid prefabricated wall.

[0006] Furthermore, an assemblability assessment is performed on the original design surface data, including: Extract the Gaussian curvature distribution map and principal curvature direction field of the surface from the original design surface data; Based on preset process parameter thresholds, the Gaussian curvature distribution map is analyzed, and areas with absolute Gaussian curvature values ​​less than the process parameter thresholds are marked as assemblable areas, while the remaining areas are marked as special node areas. Within the special node region, the boundary transition region and the core hyperbolic region are further divided by combining the principal curvature direction field; The boundary information of the assemblable region, the boundary transition region, and the core hyperbolic region, together with the principal curvature direction information within them, are encapsulated to form the assembly feature evaluation result.

[0007] Furthermore, the working principle of the improved surface segmentation algorithm includes: Receive the assembly feature evaluation results, and use the assemblable region, boundary transition region and core hyperbolic region in the evaluation results as the initial segmentation seed region; Within each seed region, guided by the principal curvature direction field within the region, a direction-sensitive growth criterion is used to control the region's growth, ensuring that the boundaries of the generated assembly units are as parallel or perpendicular as possible to the principal curvature direction. At the boundary of adjacent seed regions, a coordination buffer zone is established, and a boundary optimization function is applied within the coordination buffer zone to smooth and align the cell boundaries obtained from the initial growth, ensuring the continuity of adjacent cell boundaries. After all regions have completed growth and boundary optimization, the final segmentation mesh is output, and the corresponding surface patches are cut out from the original design surface to form the assembly unit surface set.

[0008] Furthermore, each unit surface in the assembly unit surface set is parametrically expanded to obtain the two-dimensional planar contour data corresponding to each assembly unit, including: For a target unit surface in the assembly unit surface set, calculate the mapping relationship of its three-dimensional space vertices on a preset two-dimensional parameter domain; By maintaining the boundary length ratio of the target unit surface and minimizing the internal angular distortion, the mapping relationship is solved to obtain the expansion transformation matrix from the three-dimensional surface to the two-dimensional plane. By applying the aforementioned expansion transformation matrix, the coordinates of all three-dimensional vertices of the target unit surface are transformed into corresponding two-dimensional planar coordinates; Using the two-dimensional plane coordinates, a closed polygonal contour describing the shape of the unfolded surface of the target unit is generated, which is the two-dimensional plane contour data corresponding to the assembly unit.

[0009] Furthermore, based on the two-dimensional planar contour data, a detailed three-dimensional geometric component model is created for each assembly unit, including: Read the two-dimensional planar contour data of an assembly unit, and stretch it in the normal direction of the contour according to the preset plate thickness information to generate a cubic base model. On the boundary of the base model, the three-dimensional position and orientation of the connecting holes or connecting entities are calculated according to the preset connecting type and arrangement rules; Perform Boolean operations on the base model to subtract connecting holes or add connecting entities to form a preliminary component model with connecting structures; All faces of the preliminary component model are assigned material property identifiers, and a unique code for the preliminary component model is generated in the entire wall system. These are then combined to form the detailed assembly unit three-dimensional geometric component model.

[0010] Furthermore, the complete 3D model set of the assembly unit is virtually pre-assembled, including: Import all the detailed three-dimensional geometric component models of the assembly units into the same virtual assembly space according to their designed spatial coordinates; In the virtual assembly space, each component model is driven to move from its initial position to the designed target position according to the preset assembly sequence logic. During the movement of the component model, the spatial containment relationship between the component model and other component models is calculated in real time, and the overlap of geometric volumes is detected. After all component models are in place, calculate the distance between corresponding points on the connecting surfaces of adjacent component models, and record all distance values ​​to form a gap data set.

[0011] Furthermore, the process of inspecting for interference and gaps and generating an assembly error detection report includes: Analyze the geometric volume overlap information recorded during the motion of the component model, and record the overlapping component model pairs and their overlapping volumes as hard interference; Analyze the gap data set, and record gaps with a distance value greater than the preset upper tolerance limit as out-of-tolerance gaps, and gaps with a distance value less than the preset lower tolerance limit as compression interference; All recorded hard interference, out-of-tolerance gaps, and extrusion interference, along with their corresponding unique component model codes and occurrence location coordinates, are organized together to generate a structured assembly error detection report.

[0012] Furthermore, based on the assembly error detection report, the dimensions and connection point positions of the relevant component models in the 3D model set of the assembly unit are iteratively adjusted, including: Read one error record related to the target component model from the assembly error detection report; Depending on the type of error record, if it is hard interference or extrusion interference, the interference depth is calculated, and the outline size of the target component model in the corresponding direction is reduced according to the interference depth; if it is an out-of-tolerance gap, the gap width is calculated, and the outline size of the target component model in the corresponding direction is expanded according to the gap width. Based on the changes in the outline dimensions of the target component model, the positions of the connecting holes or connector entities on it are automatically recalculated to maintain the connection points at the preset relative positions on the adjusted outline boundary. After an adjustment is completed, the updated target component model is added back to the virtual assembly space, and the virtual pre-assembly and inspection process is repeated until the assembly error detection report no longer contains error records related to the target component model, or the preset maximum number of iterations is reached.

[0013] Furthermore, the process of integrating all adjusted component models according to their spatial relationships to generate the final overall BIM model of the irregular hyperboloid prefabricated wall surface includes: Collect all component models that no longer exhibit interference or out-of-tolerance gaps after iterative adjustments, and use them as a qualified set of component models; Based on the spatial topological relationships of each unit in the original assembly unit surface set, the adjacency and connection relationships between each component model in the qualified component model set are determined; All component models are spatially positioned according to their design coordinates, and logical association indexes are established between component models based on the adjacency and connection relationships described above. All component models that have completed spatial positioning and established logical association indexes are encapsulated with the original irregular hyperboloid building wall design information and project information, and combined into a single digital model file to form the final irregular hyperboloid prefabricated wall overall BIM model.

[0014] Furthermore, the direction-sensitive growth criterion includes: During the region growth process, the principal curvature direction consistency is calculated for each candidate point on the current growth boundary. The principal curvature direction consistency is the cosine of the angle between the principal curvature direction at the candidate point and the initial principal curvature direction of its seed region. A directional consistency threshold is set, and a candidate point is only allowed to be included in the current growth region if the directional consistency of the candidate point is greater than the directional consistency threshold. Simultaneously, the geodesic distance between the candidate point and the centroid of the current growth region on the curved surface is calculated, and a comprehensive growth priority score is generated by combining the directional consistency. In each growth step, the point with the highest overall growth priority score among all current candidate points is always selected for merging, thereby driving the region to grow along the path guided by the principal curvature direction field.

[0015] Compared with the prior art, the beneficial effects of the present invention are: Assemblability assessment is performed on the original design surface data to identify assemblable areas and special node areas on the surface, generating assembly feature assessment results. The assembly adaptability attributes of different areas of the surface can be clearly defined, and the spatial characteristics and assembly requirements of special node areas can be accurately distinguished. Subsequent addition of connection construction information can be carried out based on the regional assembly attributes. The construction benchmark of the component model is consistent with the actual assembly conditions of the surface, and the matching degree of the spatial position of the component and the surface is improved. The fit of the component connection in the virtual pre-assembly stage is significantly optimized, and the component misalignment and connection deviation caused by the ambiguity of the region division are greatly reduced. The layout logic of the connection structure is consistent with the surface assembly characteristics, and the initial form of the component model is more in line with the actual assembly scenario, avoiding the adaptability problems caused by the untargeted component construction design.

[0016] The assembly feature evaluation results are input into the improved surface segmentation algorithm to modularly segment the original design surface, generating a set of assembly unit surfaces including planar, single-curved, and hyperbolic panels. The surface segmentation process can be precisely processed by combining the assembly feature evaluation results. Surface regions with different curvature characteristics can be divided into suitable assembly unit types. The geometric shape of the assembly unit surface is more in line with the standard requirements of prefabricated processing and installation. The regularity of the two-dimensional planar contour after parametric unfolding of the unit surface is improved, the conversion accuracy between two-dimensional contour data and three-dimensional geometric component model is guaranteed, the geometric dimension standardization of the assembly unit is significantly enhanced, and the interference and unevenness of components during virtual pre-assembly are effectively reduced. The adjustment of component size and connection point position can be carried out on the corresponding assembly unit, making the adjustment operation more targeted and reducing redundant steps in the adjustment process. The spatial layout of the integrated overall BIM model is more regular, and the model forming effect of the irregular hyperbolic prefabricated wall is further improved in terms of adaptability to actual assembly construction. Attached Figure Description

[0017] Figure 1 This is a state diagram of a method for generating an irregular hyperboloid prefabricated wall model according to the present invention; Figure 2 A flowchart for performing assemblability assessment; Figure 3 A flowchart illustrating the work done on the improved surface segmentation algorithm. Detailed Implementation

[0018] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] See Figure 1 This invention provides a method for generating a model of an irregular hyperboloid prefabricated wall panel. The overall implementation scheme is as follows: The original design surface data of the irregular hyperboloid building wall panel to be processed is obtained. This data typically originates from a 3D surface created by Building Information Modeling (BIM) software. An assemblability assessment is performed on the original design surface data. This process identifies areas on the surface suitable for standardized assembly and node areas requiring special processing, generating an assembly feature assessment result. This assembly feature assessment result is input into an improved surface segmentation algorithm. Based on the assessment result, the algorithm intelligently and modularly segments the original design surface, generating a set of assembly unit surfaces containing various types such as planar, single-curved, and hyperboloid panels. Parametric unfolding calculations are performed on each unit surface in the assembly unit surface set, mapping the 3D surface to a 2D plane, thereby obtaining the 2D plane contour data corresponding to each assembly unit. Based on these 2D plane contour data, a detailed 3D geometric component model is created for each assembly unit, and connection construction information is added to the model, forming a complete 3D model set of assembly units. In a digital environment, this complete set of 3D pre-assembled assembly units is virtually pre-assembled to simulate the actual installation process, check for interference and installation gaps between components, and generate an assembly error detection report. Based on this report, the dimensions and connection point positions of relevant component models in the 3D model set are iteratively adjusted to eliminate errors. All adjusted component models are then integrated according to their designed spatial relationships to generate the final, usable BIM model of the irregular hyperboloid prefabricated wall panel, which can guide production and construction.

[0020] In one embodiment of the present invention, the original design surface data of the irregular hyperboloid building wall to be processed is obtained, and an assemblability assessment is performed on the original design surface data. See also... Figure 2 The Gaussian curvature distribution map and principal curvature direction field of the surface are extracted from the original design surface data. Based on preset process parameter thresholds, the Gaussian curvature distribution map is analyzed, and regions with absolute Gaussian curvature values ​​less than the process parameter thresholds are marked as assemblable regions, while the remaining regions are marked as special node regions. Within the special node regions, boundary transition regions and core hyperbolic regions are further subdivided by combining the principal curvature direction field. The boundary information of the assemblable regions, boundary transition regions, and core hyperbolic regions, along with their internal principal curvature direction information, are encapsulated to form the assembly feature evaluation result.

[0021] In practical implementation, the original design surface data of the irregular hyperboloid building wall to be processed is acquired. An assemblability assessment is performed on the original design surface data. The Gaussian curvature distribution map and principal curvature direction field are extracted from the original design surface data. The Gaussian curvature distribution map is generated by calculating the Gaussian curvature value at each point on the surface, and the principal curvature direction field is generated by calculating the principal curvature direction vector at each point on the surface. Based on preset process parameter thresholds, the Gaussian curvature distribution map is analyzed. Areas with an absolute Gaussian curvature value less than the process parameter threshold are marked as assemblable areas, and the remaining areas are marked as special node areas. The process parameter thresholds are set according to the physical bending limit of the assembled panels. Within the special node areas, combined with the principal curvature direction field, boundary transition areas and core hyperboloid areas are further divided. The boundary transition areas are characterized by a gentle change in the principal curvature direction, while the core hyperboloid areas are characterized by a sharp change in the principal curvature direction. The boundary coordinate information of the assemblable areas, boundary transition areas, and core hyperboloid areas, along with their internal principal curvature direction information, are encapsulated to form the assembly feature assessment result, which is stored in a structured data format.

[0022] In some embodiments, the extraction of the Gaussian curvature distribution map is based on a discretized surface mesh model. The original design surface data is triangulated to generate mesh vertices and patches, and then the Gaussian curvature of each mesh vertex is calculated. In a specific implementation, the Gaussian curvature G for any mesh vertex is calculated using the following formula:

[0023] Where: G represents the Gaussian curvature of the grid vertices. Let represent the corresponding interior angle of the j-th triangle surrounding that vertex, m represent the number of triangles surrounding that vertex, and A represent the local area of ​​that vertex. The principal curvature direction field is obtained by solving the curvature tensor eigenvector of each grid vertex. The principal curvature direction field describes the maximum and minimum curvature directions of the surface at each vertex. When performing analysis based on a preset process parameter threshold, the process parameter threshold T is a positive real constant. For any point on the surface, if the absolute value of the Gaussian curvature of the point satisfies... If the condition is met, the point belongs to the assemblable region; otherwise, the point belongs to the special node region.

[0024] Optionally, the special node region can be further divided into a boundary transition region and a core hyperbolic region. This is achieved by calculating the local rate of change of the principal curvature direction field. Regions where the local rate of change of the principal curvature direction field is below a set threshold are marked as boundary transition regions, and regions where the local rate of change of the principal curvature direction field is above the set threshold are marked as core hyperbolic regions. The encapsulation process integrates the boundary polygon vertex sequences of the assemblable region, boundary transition region, and core hyperbolic region, the region type labels, and the principal curvature direction vector array of uniformly sampled points within the region into a single data file. This data file is the assembly feature evaluation result. It can be understood that the assembly feature evaluation result is directly used as input to the subsequently improved surface segmentation algorithm to guide the segmentation process of the assembly unit.

[0025] In one embodiment of the present invention, see [reference] Figure 3The system receives the assembly feature evaluation results and uses the assemblable region, boundary transition region, and core hyperbolic region from the evaluation results as initial segmentation seed regions. Within each seed region, guided by the principal curvature direction field within the region, a direction-sensitive growth criterion is used to control region growth. The direction-sensitive growth criterion includes: calculating the principal curvature direction consistency for each candidate point on the current growth boundary during region growth; this principal curvature direction consistency is the cosine of the angle between the principal curvature direction at the candidate point and the initial principal curvature direction of its respective seed region. A direction consistency threshold is set; a candidate point is only allowed to be included in the current growth region if its direction consistency is greater than this threshold. Simultaneously, the geodesic distance between the candidate point and the centroid of the current growth region on the surface is calculated, and combined with the direction consistency, a comprehensive growth priority score is generated. In each growth step, the point with the highest comprehensive growth priority score among all current candidate points is always selected for merging, thereby driving the region to grow along the path guided by the principal curvature direction field, making the boundary of the generated assembly unit as parallel or perpendicular to the principal curvature direction as possible. At the boundaries of adjacent seed regions, a coordination buffer zone is established, and a boundary optimization function is applied within the coordination buffer zone to smooth and align the boundaries of the initially grown elements, ensuring the continuity of adjacent element boundaries. After all regions have completed growth and boundary optimization, the final segmented mesh is output, and corresponding surface patches are cut from the original design surface to form an assembly element surface set.

[0026] In practical implementation, the assembly feature evaluation results are received, and the assemblable region, boundary transition region, and core hyperbolic region from the evaluation results are used as initial segmentation seed regions. Each initial segmentation seed region is defined as an independent set of surface patches. Within each seed region, guided by the principal curvature direction field within the region, a direction-sensitive growth criterion is used to control region growth. The direction-sensitive growth criterion includes calculating the principal curvature direction consistency for each candidate point on the current growth boundary during the region growth process. The principal curvature direction consistency is the cosine of the angle between the principal curvature direction at the candidate point and the initial principal curvature direction of its seed region. A direction consistency threshold is set; a candidate point is only allowed to be included in the current growth region if its direction consistency is greater than the threshold. Simultaneously, the geodesic distance between the candidate point and the centroid of the current growth region on the surface is calculated, and combined with the direction consistency, a comprehensive growth priority score is generated. The formula for calculating the comprehensive growth priority score P is:

[0027] Where: P represents the dimensionless comprehensive growth priority score of the candidate point, C represents the consistency of the principal curvature direction of the candidate point, R is a preset reference distance constant with length dimensions, and D represents the geodesic distance from the candidate point to the centroid of the current growth region. It is a preset maximum distance constant. and It is a weighting coefficient used to balance the influence of directional consistency and distance. In each growth step, the point with the highest comprehensive growth priority score P value among all current candidate points is always selected for merging. This drives the region to grow along the path guided by the principal curvature direction field, so that the boundary of the generated assembly unit is as parallel or perpendicular to the principal curvature direction as possible.

[0028] In some embodiments, the centroid location of the initial segmented seed region is determined by calculating the average coordinates of all grid vertices within the region. In a specific implementation, the geodesic distance D of the surface is calculated using a fast traversal algorithm based on a triangular mesh model. Orientation consistency threshold and weighting coefficients are also considered. and Reference distance constant R, maximum distance constant All of these are used as preset input parameters for the improved surface segmentation algorithm. A coordination buffer zone is established at the boundary of adjacent seed regions. This zone is a narrow strip with a preset width, and a boundary optimization function is applied within it. The boundary optimization function aims to smooth the curvature of the shared boundary lines between adjacent cells and reduce the total length of the boundary lines. It smooths and aligns the initially grown cell boundaries to ensure the continuity of adjacent cell boundaries.

[0029] Optionally, the boundary optimization function can be solved using an iterative Laplace smoothing method, gradually adjusting the positions of boundary vertices falling within the coordination buffer zone. After all regions have completed growth and boundary optimization, the algorithm outputs the final segmented mesh, which consists of a series of topologically connected surface patches. In specific implementation, corresponding surface patches are trimmed from the original design surface according to the boundaries of the final segmented mesh, forming an assembly unit surface set. Each unit surface in the assembly unit surface set inherits from the original design surface and carries its region type and orientation field information from the growth process. It can be understood that the assembly unit surface set, as the result of modular segmentation, is directly used in the subsequent parametric unfolding step.

[0030] In one embodiment of the present invention, for a target unit surface in the assembly unit surface set, the mapping relationship of its three-dimensional space vertices on a preset two-dimensional parameter domain is calculated. By maintaining the boundary length ratio and minimizing the internal angular distortion of the target unit surface, the mapping relationship is solved to obtain the unfolding transformation matrix from the three-dimensional surface to the two-dimensional plane. Applying the unfolding transformation matrix, the coordinates of all three-dimensional space vertices of the target unit surface are transformed into the corresponding two-dimensional plane coordinates. Using the two-dimensional plane coordinates, a closed polygonal contour describing the unfolded shape of the target unit surface is fitted and generated, which is the two-dimensional plane contour data corresponding to the assembly unit. The two-dimensional plane contour data of an assembly unit is read, and according to the preset sheet thickness information, the contour is stretched in the normal direction to generate a cubic base model. On the boundary of the base model, according to the preset connector type and arrangement rules, the three-dimensional position and orientation of the connecting holes or connector entities are calculated. Boolean operations are performed on the base model to subtract connecting holes or add connector entities to form a preliminary component model with connection structure. Assign material property identifiers to all faces of the preliminary component model and generate a unique code for the preliminary component model in the entire wall system, combining them to form a detailed three-dimensional geometric component model of the assembly unit.

[0031] In practical implementation, for a target element surface in the assembly element surface set, the mapping relationship of the three-dimensional vertices of the target element surface on a preset two-dimensional parameter domain is calculated. This preset two-dimensional parameter domain is typically a unit square region. By maintaining the boundary length ratio and minimizing the internal angular distortion of the target element surface, the mapping relationship is solved to obtain the unfolding transformation matrix from the three-dimensional surface to the two-dimensional plane. The solution process can be constructed as an energy minimization problem. The energy function E, used to measure the degree of unfolding deformation, is defined as:

[0032] Where: E represents the total deformation energy of the parametric unfolding process. This represents the set of all vertices of the target element surface. This represents the scaling factor (i.e., the area stretching) of the associated area of ​​vertex i before and after mapping. It is the area deformation penalty weight. This represents the set of all interior edges of the target element surface. Represents the interior angle of two triangular faces adjacent to side j in the two-dimensional parameter domain. This represents the corresponding interior angle in the original 3D surface. This refers to the angular deformation penalty weight. By minimizing the energy function E through an iterative optimization algorithm, the two-dimensional coordinates that flatten the 3D surface while minimizing geometric feature loss can be obtained, thus yielding the unfolding transformation matrix. Applying the unfolding transformation matrix, the coordinates of all 3D vertices of the target unit surface are transformed into corresponding two-dimensional planar coordinates. Using the calculated two-dimensional planar coordinates, by connecting the boundary vertices and forming closed loops in sequence, a closed polygonal contour describing the unfolded shape of the target unit surface is fitted and generated. This closed polygonal contour is the two-dimensional planar contour data corresponding to the assembly unit. The two-dimensional planar contour data is stored in vertex sequence format and used to drive subsequent component model creation.

[0033] In some embodiments, the two-dimensional planar contour data of an assembly unit is read, and the contour is stretched along the normal direction according to preset sheet thickness information to generate a cubic base model. The sheet thickness information is a preset constant value representing the physical thickness of the wall panel. The stretching operation uses the two-dimensional planar contour as its base and extends along the average normal vector direction of the contour in three-dimensional space by a distance equal to the sheet thickness, thereby generating a three-dimensional entity with a top surface, a bottom surface, and multiple side surfaces—the cubic base model. On the boundaries of the base model, the three-dimensional position and orientation of connecting holes or connecting entities are calculated according to preset connector types and arrangement rules. The preset connector types and arrangement rules define the connection methods and their geometric parameters to be used on different boundary types of the component. The calculation of the connection position is based on the equal division of the side length of the two-dimensional planar contour or determined by a fixed interval, and then the two-dimensional coordinates are combined with the height information of the base model to convert them into three-dimensional spatial coordinates. The connection direction is typically set perpendicular to the boundary and points to the outside or inside of the base model.

[0034] Optionally, the arrangement rules for the connectors can be based on a preset parameter table, which associates boundary geometry features with connection parameters. See Table 1: Table 1: Reference Table for Connector Layout Rules

[0035] Boolean operations are performed on the base model to subtract connecting holes or add connector entities, forming a preliminary component model with connection structures. Boolean operations are a common entity combination operation in 3D modeling. Material property identifiers are assigned to all faces of the preliminary component model. These identifiers distinguish different areas such as plate surfaces, connector surfaces, and inner surfaces of holes, facilitating subsequent quantity calculations and rendering. A unique code for the preliminary component model is generated within the entire wall system. This unique code typically consists of the wall partition number, unit row and column number, and component type abbreviation, which combine to form a detailed assembly unit 3D geometric component model. In essence, the detailed assembly unit 3D geometric component model is a complete digital component containing precise geometry, connection structures, material properties, and an identification code, which can be directly used for production and virtual assembly.

[0036] In one embodiment of the present invention, all detailed three-dimensional geometric component models of the assembly units are imported into the same virtual assembly space according to their designed spatial coordinates. In the virtual assembly space, each component model is driven to move from its initial position to its designed target position according to a preset assembly sequence logic. During the movement of the component models, the spatial containment relationship between the component models and other component models is calculated in real time to detect whether there is geometric volume overlap. After all component models are in place, the distance between corresponding points on the connecting surfaces of adjacent component models is calculated, and all distance values ​​are recorded to form a gap data set. The geometric volume overlap information recorded during the component model movement is analyzed, and overlapping component model pairs and their overlapping volumes are recorded as hard interference. The gap data set is analyzed, and gaps with distance values ​​greater than the preset upper tolerance limit are recorded as out-of-tolerance gaps, and gaps with distance values ​​less than the preset lower tolerance limit are recorded as compression interference. All recorded hard interference, out-of-tolerance gaps, and compression interference are organized together with their corresponding unique component model codes and occurrence position coordinate information to generate a structured assembly error detection report.

[0037] In practice, a virtual pre-assembly is performed on the complete 3D model set of assembly units. All detailed 3D geometric component models of the assembly units are imported into the same virtual assembly space according to their designed spatial coordinates. The virtual assembly space is a digital environment in a 3D coordinate system. Within the virtual assembly space, each component model moves from its initial position to its designed target position according to a preset assembly sequence logic, which defines the order in which the components are installed. During the movement of the component models, the spatial containment relationship between the component models and other component models is calculated in real time to detect any overlap in geometric volumes. The calculation of spatial containment relationships is achieved through intersection tests using the bounding boxes of the components or precise triangular meshes. After all component models are in place, the distance between corresponding points on the connecting surfaces of adjacent component models is calculated, and all distance values ​​are recorded to form a gap data set. This gap data set is a structured list containing distance values, corresponding component model codes, and measurement point coordinates.

[0038] In some embodiments, the gap distance is calculated based on predefined pairs of corresponding sampling points on adjacent component models. For each pair of sampling points, the Euclidean distance in three-dimensional space is calculated. Calculate the Euclidean distance between point pairs. The formula is:

[0039] in: and These represent the 3D coordinates of corresponding sampling points from two adjacent component models. The geometric volume overlap information recorded during component model movement is analyzed, and overlapping component model pairs and their overlapping volumes are recorded as hard interference. The gap data set is analyzed, and gaps with distance values ​​greater than the preset upper tolerance limit are recorded as out-of-tolerance gaps, while gaps with distance values ​​less than the preset lower tolerance limit are recorded as compression interference. The preset upper and lower tolerance limits are determined based on construction technology and sealing material performance, and are used to determine whether the installation gaps are qualified. All recorded hard interference, out-of-tolerance gaps, and compression interference, along with their corresponding unique component model codes and occurrence location coordinates, are compiled to generate a structured assembly error detection report.

[0040] Optionally, the upper and lower tolerance limits can be set differently depending on the type of connection edge. Specific values ​​can be found in the preset tolerance standard table. See Table 2: Table 2: Assembly Gap Tolerance Standards

[0041] The structured assembly error detection report records every error in detail in a list format, including the error number, type, the code of the component model A involved, the code of the component model B involved, the global coordinates of the location where the error occurred, and the measured distance value or overlap volume. In essence, the assembly error detection report comprehensively records all geometric conflicts and dimensional deviations discovered during the virtual pre-assembly process, serving as a direct input basis for driving subsequent model adjustments.

[0042] In one embodiment of the present invention, an error record relating to the target component model is read from the assembly error detection report. Depending on the type of error record, if it is hard interference or extrusion interference, the interference depth is calculated, and the contour dimension of the target component model in the corresponding direction is reduced according to the interference depth. If it is a gap exceeding tolerance, the gap width is calculated, and the contour dimension of the target component model in the corresponding direction is expanded according to the gap width. Based on the change in the contour dimension of the target component model, the positions of connecting holes or connector entities on it are automatically recalculated to maintain the connection points at a preset relative position on the adjusted contour boundary. After one adjustment, the updated target component model is added back to the virtual assembly space, and the virtual pre-assembly and detection process is repeated until the assembly error detection report no longer contains error records related to the target component model, or the preset maximum number of iterations is reached. All component models that no longer exhibit interference or gaps exceeding tolerance after iterative adjustments are collected as a qualified component model set. Based on the spatial topological relationships of each unit in the original assembly unit surface set, the adjacency and connection relationships between each component model in the qualified component model set are determined. All component models are spatially positioned according to their design coordinates, and logical association indexes are established between component models based on adjacency and connection relationships. All component models that have completed spatial positioning and established logical association indexes are encapsulated with the original irregular hyperboloid building wall design information and project information, and combined into a single digital model file to form the final overall BIM model of the irregular hyperboloid prefabricated wall.

[0043] In practice, based on the assembly error detection report, the dimensions and connection point positions of relevant component models in the 3D model set of the assembly unit are iteratively adjusted. An error record involving the target component model is read from the assembly error detection report. Depending on the type of error record, if it is a hard interference or extrusion interference, the interference depth is calculated, and the contour dimensions of the target component model in the corresponding direction are reduced according to the interference depth; if the error record is an out-of-tolerance gap, the gap width is calculated, and the contour dimensions of the target component model in the corresponding direction are expanded according to the gap width. Dimensional adjustment amount. The calculation is based on the relationship between the measured value and the tolerance boundary, and the specific formula is as follows:

[0044] in: This indicates the amount of size adjustment needed (positive numbers indicate enlargement, negative numbers indicate shrinkage). This refers to the measured distance or interference depth value recorded in the assembly error inspection report. This indicates the corresponding preset upper tolerance limit (for gaps exceeding tolerance) or lower tolerance limit (for compression interference). For hard interference, The penetration depth of the overlapping area. This can be considered as 0. Based on changes in the target component model's outline dimensions, the positions of connecting holes or connector entities on the target component model are automatically recalculated to maintain the connection points at preset relative positions on the adjusted outline boundary. These preset relative positions are defined by a percentage or fixed distance from the connection point to the outline boundary during the initial design. After each adjustment, the updated target component model is re-added to the virtual assembly space, and the virtual pre-assembly and inspection process is repeated until the assembly error detection report no longer contains error records related to the target component model, or the preset maximum number of iterations is reached.

[0045] In some embodiments, the preset maximum number of iterations is an integer constant to prevent infinite loops. When the number of adjustments reaches this maximum value, the iteration process will be forcibly terminated, and the current latest model state will be output. All component models that no longer exhibit interference or out-of-tolerance gaps after iterative adjustments are collected as a qualified component model set. The qualified component model set is a data list containing all component models that have passed virtual pre-assembly verification. Based on the spatial topological relationships of each element in the original assembly unit surface set, the adjacency and connection relationships between each component model in the qualified component model set are determined. The spatial topological relationships describe the relative positions and shared boundary information of each element in the original surface patch before segmentation.

[0046] Optionally, adjacency and connection relationships are determined by finding component model pairs with shared connection surfaces. If the distance between the connection surfaces of two component models in three-dimensional space is less than a very small tolerance threshold, the two component models are determined to be adjacent components, and a connection relationship record is established. All component models are spatially located according to their design coordinates, and logical association indexes are established between component models based on adjacency and connection relationships. The logical association indexes are stored in the form of a data table, recording the unique code, spatial location matrix, and a list of codes for all adjacent components connected to each component model. In essence, all component models that have completed spatial positioning and established logical association indexes are encapsulated with the original irregular hyperboloid building wall design information and project information, combined into a single digital model file, forming the final overall BIM model of the irregular hyperboloid prefabricated wall. The original irregular hyperboloid building wall design information includes surface geometry definitions, material specifications, and performance requirements; the project information includes project name, version number, and creation date. This metadata, along with the component model set and logical association indexes, is written into a BIM file format conforming to industry standards, such as IFC format.

[0047] The above embodiments are only used to illustrate the technical methods 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 methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for generating an irregularly shaped hyperboloid prefabricated wall model, characterized in that, The method includes: Obtain the original design surface data of the irregular hyperboloid building wall to be processed; Assemblability evaluation is performed on the original design surface data to identify assemblable areas and special node areas on the surface, and assembly feature evaluation results are generated. The assembly feature evaluation results are input into the improved surface segmentation algorithm to perform modular segmentation of the original design surface, generating a set of assembly unit surfaces including planar, single-curved, and hyperbolic panels; Each unit surface in the assembly unit surface set is parametrically expanded to obtain the two-dimensional planar contour data corresponding to each assembly unit. Based on the two-dimensional planar contour data, a detailed three-dimensional geometric component model is created for each assembly unit, and connection construction information is added to form a complete three-dimensional model set of assembly units. The complete 3D model set of the assembly unit is virtually pre-assembled to check for interference and gaps, and an assembly error detection report is generated. Based on the assembly error detection report, the dimensions and connection point positions of the relevant component models in the 3D model set of the assembly unit are iteratively adjusted. All the adjusted component models are integrated according to their spatial relationships to generate the final overall BIM model of the irregular hyperboloid prefabricated wall.

2. The method for generating an irregular hyperboloid prefabricated wall model according to claim 1, characterized in that, Perform an assemblability assessment on the original design surface data, including: Extract the Gaussian curvature distribution map and principal curvature direction field of the surface from the original design surface data; Based on preset process parameter thresholds, the Gaussian curvature distribution map is analyzed, and areas with absolute Gaussian curvature values ​​less than the process parameter thresholds are marked as assemblable areas, while the remaining areas are marked as special node areas. Within the special node region, the boundary transition region and the core hyperbolic region are further divided by combining the principal curvature direction field; The boundary information of the assemblable region, the boundary transition region, and the core hyperbolic region, together with the principal curvature direction information within them, are encapsulated to form the assembly feature evaluation result.

3. The method for generating an irregularly shaped hyperboloid prefabricated wall model according to claim 2, characterized in that, The working principle of the improved surface segmentation algorithm includes: Receive the assembly feature evaluation results, and use the assemblable region, boundary transition region and core hyperbolic region in the evaluation results as the initial segmentation seed region; Within each seed region, guided by the principal curvature direction field within the region, a direction-sensitive growth criterion is used to control the region's growth, ensuring that the boundaries of the generated assembly units are as parallel or perpendicular as possible to the principal curvature direction. At the boundary of adjacent seed regions, a coordination buffer zone is established, and a boundary optimization function is applied within the coordination buffer zone to smooth and align the cell boundaries obtained from the initial growth, ensuring the continuity of adjacent cell boundaries. After all regions have completed growth and boundary optimization, the final segmentation mesh is output, and the corresponding surface patches are cut out from the original design surface to form the assembly unit surface set.

4. The method for generating an irregularly shaped hyperboloid prefabricated wall model according to claim 1, characterized in that, The assembly unit surface set is parametrically expanded to obtain the two-dimensional planar contour data corresponding to each assembly unit, including: For a target unit surface in the assembly unit surface set, calculate the mapping relationship of its three-dimensional space vertices on a preset two-dimensional parameter domain; By maintaining the boundary length ratio of the target unit surface and minimizing the internal angular distortion, the mapping relationship is solved to obtain the expansion transformation matrix from the three-dimensional surface to the two-dimensional plane. By applying the aforementioned expansion transformation matrix, the coordinates of all three-dimensional vertices of the target unit surface are transformed into corresponding two-dimensional planar coordinates; Using the two-dimensional plane coordinates, a closed polygonal contour describing the shape of the unfolded surface of the target unit is generated, which is the two-dimensional plane contour data corresponding to the assembly unit.

5. The method for generating an irregularly shaped hyperboloid prefabricated wall model according to claim 4, characterized in that, Based on the aforementioned two-dimensional planar contour data, a detailed three-dimensional geometric component model is created for each assembly unit, including: Read the two-dimensional planar contour data of an assembly unit, and stretch it in the normal direction of the contour according to the preset plate thickness information to generate a cubic base model. On the boundary of the base model, the three-dimensional position and orientation of the connecting holes or connecting entities are calculated according to the preset connecting type and arrangement rules; Perform Boolean operations on the base model to subtract connecting holes or add connecting entities to form a preliminary component model with connecting structures; All faces of the preliminary component model are assigned material property identifiers, and a unique code for the preliminary component model is generated in the entire wall system. These are then combined to form the detailed assembly unit three-dimensional geometric component model.

6. The method for generating an irregular hyperboloid prefabricated wall model according to claim 1, characterized in that, Virtual pre-assembly of the complete 3D model set of assembly units includes: Import all the detailed three-dimensional geometric component models of the assembly units into the same virtual assembly space according to their designed spatial coordinates; In the virtual assembly space, each component model is driven to move from its initial position to the designed target position according to the preset assembly sequence logic. During the movement of the component model, the spatial containment relationship between the component model and other component models is calculated in real time, and the overlap of geometric volumes is detected. After all component models are in place, calculate the distance between corresponding points on the connecting surfaces of adjacent component models, and record all distance values ​​to form a gap data set.

7. The method for generating an irregular hyperboloid prefabricated wall model according to claim 6, characterized in that, The process of inspecting for interference and gaps, and generating an assembly error detection report, includes: Analyze the geometric volume overlap information recorded during the motion of the component model, and record the overlapping component model pairs and their overlapping volumes as hard interference; Analyze the gap data set, and record gaps with a distance value greater than the preset upper tolerance limit as out-of-tolerance gaps, and gaps with a distance value less than the preset lower tolerance limit as compression interference; All recorded hard interference, out-of-tolerance gaps, and extrusion interference, along with their corresponding unique component model codes and occurrence location coordinates, are organized together to generate a structured assembly error detection report.

8. The method for generating an irregular hyperboloid prefabricated wall model according to claim 7, characterized in that, Based on the assembly error detection report, iterative adjustments are made to the dimensions and connection point positions of the relevant component models in the 3D model set of the assembly unit, including: Read one error record related to the target component model from the assembly error detection report; Depending on the type of error record, if it is hard interference or extrusion interference, the interference depth is calculated, and the outline size of the target component model in the corresponding direction is reduced according to the interference depth; if it is an out-of-tolerance gap, the gap width is calculated, and the outline size of the target component model in the corresponding direction is expanded according to the gap width. Based on the changes in the outline dimensions of the target component model, the positions of the connecting holes or connector entities on it are automatically recalculated to maintain the connection points at the preset relative positions on the adjusted outline boundary. After an adjustment is completed, the updated target component model is added back to the virtual assembly space, and the virtual pre-assembly and inspection process is repeated until the assembly error detection report no longer contains error records related to the target component model, or the preset maximum number of iterations is reached.

9. The method for generating an irregular hyperboloid prefabricated wall model according to claim 1, characterized in that, The process of integrating all adjusted component models according to their spatial relationships to generate the final overall BIM model of the irregular hyperboloid prefabricated wall surface includes: Collect all component models that no longer exhibit interference or out-of-tolerance gaps after iterative adjustments, and use them as a qualified set of component models; Based on the spatial topological relationships of each unit in the original assembly unit surface set, the adjacency and connection relationships between each component model in the qualified component model set are determined; All component models are spatially positioned according to their design coordinates, and logical association indexes are established between component models based on the aforementioned adjacency and connection relationships; All component models that have completed spatial positioning and established logical association indexes are encapsulated with the original irregular hyperboloid building wall design information and project information, and combined into a single digital model file to form the final irregular hyperboloid prefabricated wall overall BIM model.

10. The method for generating an irregular hyperboloid prefabricated wall model according to claim 3, characterized in that, The direction-sensitive growth criteria include: During the region growth process, the principal curvature direction consistency is calculated for each candidate point on the current growth boundary. The principal curvature direction consistency is the cosine of the angle between the principal curvature direction at the candidate point and the initial principal curvature direction of its seed region. A directional consistency threshold is set, and a candidate point is only allowed to be included in the current growth region if the directional consistency of the candidate point is greater than the directional consistency threshold. Simultaneously, the geodesic distance between the candidate point and the centroid of the current growth region on the curved surface is calculated, and combined with the directional consistency, a comprehensive growth priority score is generated; In each growth step, the point with the highest overall growth priority score among all current candidate points is always selected for merging, thereby driving the region to grow along the path guided by the principal curvature direction field.

Images